INDUSTRIAL and ENGINEERING CHEMISTRY

INDUSTRIAL
and
ENGINEERING CHEMISTRY
A N A L Y T IC A L E D ITIO N
W A L T E R J . M U R P H Y , E D IT O R
I S S U E D J A N U A R Y 25, 1944
Editorial Assistant: G . G l a d y s G o r d o n
M anuscript Assistant: S t e l l a A n d e r s o n
.
/ r-j
-
V \ i û ' j " T il,
B. L. C l a r k e
T . R. C u n n i n g h a m
Advisory Board
G. E . F . L u n d e l l
M . G. M e l l o n
V O L . 16, N O . 1
C O N S E C U T IV E N O . 2
M ake-up Assistant: C h a r l o t t e C . S a y r e -
R. H . M ü l l e r
II. H . W i l l a r d
Ph
E m u ls io n P o ly m e riz a tio n of S y n th e tic R u b b e r i n 10G r a m S y s t e m s ..................................... C harles F. Fryling
1
C o lo r o f A g u e o u s P o ta s s iu m D ic h ro m a te S o lu tio n s .
R. E. Kitson with M. G . Mellon
D e te r m in a tio n of T e tr a e th y lle a d in G a so lin e . . . .
Harry G onick and J. J. Milano
4
G ra v im e tric D e te r m in a tio n of T u n g s t e n ......................
John H. Yoe and A. Letcher Jones
Specific G ra v ity of B u t a d i e n e ...................................... .
M. R. Dean an d T. W. Legatski
7
D ic h ro m a te D e te r m in a tio n o f Ir o n , U sin g S ilv er
R e d u c t o r ......................J. L. H enry and R. W. G elbach 49
P h y sic a l M e th o d s of A n a ly sis of S y n th e tic a n d
N a tu r a l R u b b e r ................................. R. Bowling Barnes,
Van Zandt Williams, A. R. Davis, and Paul Gieseoke
9
50
E th y lb is - 2 ,4 -D in itro p h e n y la c e ta te , N ew p H I n d ic a ­
t o r .................................. E. A. Fehnel and E. D. Amstutz 53
D e te r m in a tio n of R a te of C u re fo r N a tu r a l a n d S y n ­
th e t ic R u b b e r .........................................................................
Leonard H. Cohan an d M artin Steinberg
15
D e te r m in a tio n of A lp h a ,P a ra - D irq e th y ls ty re n e in
.P re s e n c e o f P a ra - M e th y ls ty r e n e , S ty r e n e , a n d
P a ra - C y m e n e . . John H. Elliott and Evelyn V. Cook
20
D e te r m in a tio n of S o lu b le P e c tin a n d P e c tic A cid b y
E l e c t r o d e p o s i t i o n ................................................................
Kenneth T. Williams and C larence M. Johnson
P re c ise M e a s u r e m e n t of V o lu m e in T itr im e tr ic
A n a ly s is .................................. . William M. Thornton, Jr.
45
23
P y c n o m e te r fo r V o la tile L iq u id s. C o n tro l of D iffu sio n
a s A id in P re c is io n P y c n o m e try . . . . M. R. Lipkin,
J, A. Davison, W. T. Harvey, and S. S. Kurtz, Jr.
55
A p p a ra tu s fo r M e a s u rin g G a s P e r m e a b ility of F ilm
M a te ria ls of Low P e r m e a b ility . A. Cornwell Shum an
58
D e sig n of L a rg e -S iz e L a b o ra to ry E x tr a c tio n G lass
A p p a r a t u s ........................................... Raymond Jonnard 61
A n a ly sis of P e tr o le u m O il-S o lu b le S o d iu m S u lfo ­
n a te s b y A d s o r p t i o n .............................. John M. Koch 25
C o n tin u o u s L iq u id -L iq u id E x tr a c to r . . Irwin A. Pearl
Q u a n tita tiv e D e te r m in a tio n of d -G a la c to s e b y S e ­
le c tiv e F e r m e n ta t io n w ith S p e c ia l R e fe re n c e to
P l a n t M u c ila g e s . Louis E. Wise and John W. A ppling
M IC R O C H E M IS T R Y
62
28
D e te r m in a tio n of S m a ll A m o u n ts of A c ry lo n itrile
i n A i r ......................G. W. Petersen and H. H. Radke 63
U se o f D is c r im in a n t F u n c tio n in C o m p a ris o n of
P ro x im a te C o al A n a l y s e s ...............................................
W . D. Baten and C. C. DeWitt
32
S p e c tr o p h o to m e tric D e te r m in a tio n o f Io d in e
L ib e ra te d in O x id a tio n of C a rb o n M o n o x id e b y
Io d in e P e n to x id e . B ernard Smaller an d J. F. Hall, Jr.
64
C o lo rim e tric A n a ly sis of X a n th o n e S p ra y R e sid u e s .
C. C. Cassil and J. W. H ansen
35
P h e n o l S t u d i e s ........................................Wm. B. Deichmann
37
Y e a st M ic ro b io lo g ic a l M e th o d s fo r D e te r m in a tio n
o f V i t a m i n s ........................ Lawrence Atkin, William
L. Williams, Alfred S. Schultz, an d C harles N. Frey
67
P o la ro g ra p h ic D e te r m in a tio n o f C o p p e r, L e a d , a n d
C a d m iu m i n H ig h - P u r ity Z in c A l l o y s .................
R. C. Hawkings and H. G . Thode
71
D e te r m in a tio n of P e c tin i n B io lo g ical M a te ria ls .
Edwin F. Bryant, G ran t H. Palmer, and G . H. Joseph
74
S ta b ility of S ta n d a r d S o lu tio n s of C o p p e r P e r c h lo r a te
a n d P o ta s s iu m I o d a t e .............................. Joseph J. Kolb 38
S e m ia u to m a tic P re s s u re C o n tro l in L o w -P re ssu re ,
L o w - T e m p e ra tu re L a b o ra to ry F r a c tio n a tio n . . .
D. R. Douslin and W. S. Walls
40
T h e A m erican C hem ical Society assum es no resp o n sib ility for th e s ta te m e n ts an d opinions advanced b y c o n trib u to rs to its p u b lications.
C op y rig h t 1944 b y A m erican C hem ical Society.
30,900 copies of th is issue p rin ted .
P ublished b y th e A m erican C hem ical S o ciety a t E a s to n , P en tia. E d i­
to ria l Office: 1155 10th S tre e t, N . W ., W ash in g to n 6, D . C.; telephone,
R epublic 5301; cable, Jiech em (W ash in g to n ). Business Office: A m erican
C hem ical Society, 1155 16th S tree t, N . W .t W ashington 6, D . C . A d v e rtis­
ing Office: 332 W est 42nd S tre e t, New Y o rk 18, N. Y .; telep h o n e, B ry a n t
9-4430.
E n te re d as second-class m a tte r a t th e P o st Office a t E a s to n , P e n n a .,
u n d e r th e A ct of M arch 3, 1879, as 24 tim es a year— In d u stria l E d ition
m o n th ly on th e 1st, A n aly tical E d itio n m o n th ly on th e lo th . A cceptance
for m ailing a t special ra te of p o stag e p ro v id ed for in Section 1103, A ct of
O c to b e r s , 1917, au th o rized Ju ly 13, 19 1 8 .i!
.
R em itta n ce s an d o rd ers for su b scrip tio n s an d for single copies, notices of
changes of add ress an d new professional connections, and claim s for m issing
n u m b ers should b e se n t to th e A m erican C hem ical Society, 1155 16th S treet,
N . W ,, W ashin g to n 0, D . C . C hanges of ad d ress for th e In d u s tria l E d itio n
m u st h e received on o r before th e 18th of th e preceding m o h th a n d for th e
A n a lytical E d itio n n o t la te r th a n th e 3 0th of th e preceding m o n th . C laim s
for missing num bers will n o t be allow ed (1) if received m ore th a n 60 d ays
from d a te of issue (owing to th e h a za rd s of w artim e d elivery, no claim s van be
honored from subscribers o utside of N o rth A m erica), (2) if loss w as d u e to
failure of notice of change of address to be received before th e d a te s .specified
in the preceding sentence, or (3) if th e reason for claim is “ m issing from
files".
A nnual su b sc rip tio n — In d u s tria l E d itio n a n d A n a lytical E d itio n sold
only as a u n it, m em bers S3.00, nonm em bers S4.00. P o stag e to countries n o t
in th e Pan-A m erican U nion S2.25; C an ad ian postage SO.75. Single copies—
cu rre n t issues, In d u stria l E d itio n §0.75, A nalytical E d itio n S0.50; back
num bers. In d u stria l E d itio n 30.80, A nalytical E d itio n prices on request;
special ra te s to m em bers.
T h e A m erican C hem ical S ociety also publishes Chemical and Engineering
Newe, Chemical Abstracta, and Jo u rn a l o f the Am erican Chemical Society.
R ates on req u est.
A L U N D U M Laboratory W are has high heat transfer properties, is chem i­
ca lly inert and assures long service. C leaning is a sim ple process of wash­
ing and ign iting to constant weight. From your regular laboratory supply
house you can obtain N orton tubes, cores and muffles for electric furnaces;
crucibles and dishes for igniting, incinerating, melting,- dishes, cones,
discs, thimbles, crucibles for filtering; combustion boats, pyrometer tubes
and refractory cements. A Norton Laboratory W are catalog w ill be sent
free upon request.
NORTON C O M P A N Y , W o rc e ster 6, Mass
E L .E G ff.R -IC
FU RN ACE
/FU SED
January 15, 1944
ANALYTICAL
With todays increased production demands taxing
the m etallurgist, chemist, and physicist—the
method used should facilitate obtaining informa­
tion concerning the submicroscopic structure of
any sort of material in the shortest possible time.
The G-E XRD Unit permits the registration of
patterns with exposure times which are from three
to thirty times faster than was previously possible.
In routine diffraction studies of steel, four to eight
hour tests were formerly necessary—with the G-E
XRD Unit it is possible to conduct the same study
----------------
EDITION
with a fifteen to thirty minute exposure. Corre­
sponding reductions in exposure times have
materially speeded up the control of other metals.
G-E X-Ray offers a complete line of diffraction
cameras—each a precision instrument designed to
meet the exacting standards required in analytical
procedures. The XRD Unit, along with the cameras,
offer a solution for many of the numerous problems
encountered in the control and investigation of
materials used in manufacturing processes. For
complete information, write to Department NN41.
GENERAL i f f ELECTRIC
X R A Y CORPORATION
2012
JACKSON
BLVD.
C H I C A G O (12), ILL., U. S. A.
6
INDUSTRIAL
AND
ENGINEERING
Vol. 16, No. 1
CHEMISTRY
a c c u ra te
m in u te s
h o u r s .
SPECTROGRAPH
'k The
tim e tod ay is all im p o rta n t, p artic u larly th e tim e o f highly train ed
technicians.
T h a t is w hy m a n u fac tu rers th ro u g h o u t th e co u n try are u sin g th e A . R . L .D ie te rt L arg e G ra tin g S p e ctro g rap h in th e ir
f|
testin g lab o rato ries. T h e y h av e fo u n d th a t one
tra in e d o p e ra to r an d th e S p ectro g rap h w ill do th e
H
The Spectrograph
work t h a t fo rm erly re q u ired th e co m b in ed skill o f 1 2
illustrated includes
op erato rs usin g th e w et m e th o d s o f a n aly sis.
spark, D. C. Arc,
N o t only is th e re a defin ite sav in g o f m a n -h o u rs, b u t th e
and A . C. A rc ,
A . R . L .-C )ietert G ra tin g S p ectro g ra p h w ith its larg e linear
source units.
dispersion a n d exceptionally h ig h reso lu tio n is a n id eal in s tru ­
m e n t fo r all types o f research a n d co n tro l analysis.
A t th e p re s e n t tim e m a n u fa c tu re rs a re u sin g it to analyze iro n , steel, m agnesium , alum inum ,
b ronze, b rass, zinc, le a d —in fa c t, alm o st e v ery th in g .
F o r com plete in fo rm atio n a b o u t th e A . R . L .-D ie te rt S p ectro g rap h w rite for o u r M o d ern A nalysis
C atalo g u e N o . 128 to d a y .
APPLIED RESEARCH LABORATORIES
4336 SAN FERN ANDO RD., GLENDALE, CALIF.
HARR Y W
D I E T E R T CO.
9330 ROSELAWN AVE., DETROIT, MICH.
January 15, 1944
ANALYTICAL
EDITION
7
B e g i n n i n g w it h P y r e x b r a n d L a b o r a to r y G la s s w a r e in 1 9 1 5 , C o r n in g R e s e a r c h h a s k e p t
p a c e w it h t h e g r o w t h o f t h e m o d e r n la b o r a t o r y b y d e v e lo p in g n e w lin e s o f g la s s w a r e t o m e e t e a c h n e w
la b o r a t o r y n e e d . T o d a y t h e r e a r e f iv e s e p a r a te lin e s m e e t in g e v e r y la b o r a t o r y r e q u ir e m e n t.
1. Pyrex brand Laboratory Glassware
THE A L L - A R O U N D
3 . Pyrex brand Fritted Ware-FOR
W ARE FOR A LL-A R O U N D
S P E E D , R E T E N T IV IT Y
A N D FREEDOM FROM CH EM ICA L R E A C T IO N
USE
F a b ric a te d fro m “ P y re x ” C h em ical G lass N o. 774— “ P y re x ”
W a re com bines th e essen tial p ro p e rtie s o f ch em ical sta b ility ,
m ech a n ical stre n g th a n d h e a t resistance, scientifically balanced
— th e standard glassw are fo r gen eral la b o ra to ry use.
2 . Pyrex brand Lifetime Red Low Actinic Glassware
FO R L IG H T - S E N S IT IV E S U BSTA N C ES
T liis colored glassw are affords hig h p ro te c tiv e v alu e. F a b ri­
c a te d from “ P y re x ” C hem ical G lass N o . 774, its l i f e t i m e b e d
color is a n in te g ra l p a r t o f th e glass. M e ch an ical stre n g th ,
chem ical sta b ility a n d h e a t re sista n c e a re com b in ed w ith
a b ility to r e ta r d d e te rio ra tio n fro m lig h t influence.
T h e follow ing d a ta w ill serv e a s a guide to its efficiency:
A p p ro x im ate p e rc e n t w av e­
le n g th tran sm issio n in A ng­
stro m u n its o f “ P y re x ” Low
A ctinic W are
3000
4000
5000
6000
A n g stro m s 0 %
A n g stro m s 1 %
A n g stro m s 4 %
A n g stro m s 12%
G lass p a rtic le s a re fr itte d to form discs w hich a re sealed in to
n o n -p o ro u s bodies— b o th o f “ P y re x ” C h em ical G lass N o . 774.
W ith five p o ro sities a v ailab le— from u ltra fine to e x tra coarse—
filtra tio n o f v a rio u s ty p e s o f p re c ip ita te s c a n be acco m p lished
a t m ax im u m speed. D iscs a re m ech a n ically stro n g , th e rm a lly
a n d ch em ically re s ista n t.
4 . Vycor brand Glassware
FOR H IG H TEM PERATURE R E A C T IO N S , R A P ID A N A L Y S IS
F a b ric a te d fro m 9 6 % Silica G lass N o. 790, th is w are possesses
ex cep tio n al ch em ical sta b ility , h a s a h ig h so ften in g p o in t a n d
ex trem ely low coefficient o f expansion.
5 . Corning brand Alkali-Resistant Glassware
FOR
B O R O N D ETER M IN A TIO NS A N D S IM ILA R A P P L IC A T IO N S
D esigned specifically for use w h ere resistan ce to alk alies is im ­
p o rta n t. F a b ric a te d from “ C o rn in g ” A lk a li-R e sista n t G lass N o.
728— su b sta n tia lly boron-free (ap p ro x . B 2 O 3 C o n te n t, 0 .0 6 % ).
C O M P A R I S O N O F P R O P E R T I E S — (F or o th e r physical properties consult C atalog LP21 an d Supplem ents)
N um ber and Type
Coef, o f
Expansion
Annealing Softening
Point
Point
Strain
Point
Specific
Gravity
Loss in Weight—-100 lbs. Steam
Pressure 96 IJrs. (grams per sg. cm.)
I^oss in Weight— Boiling 6 IJrs. in
5% N a O ll (grams per sg. cm.)
Pyrex brand C hem ical
Glass No. 774
0.0000032
560°C
820°C
510°C
2.23
0.0005
0.002
96% Silica Glass No. 790
0.0000008
910°C
1500°C
820°C
2.18
0.0001
0.0015
“ C orning” A lkali-R esistant Glass
(B oron-Free) No. 728
0.0000063
620°C
870°C
585°C
2.60
0.0005
0.00004
PYREX BRAND
(PYREX
LABORATORY WARE
P Y R E X ” and ‘‘V Y C O R " are registered
trade-marks and indicate manufacture by
m e a n s ——
-
Research- in Glass
CORNING GLASS WORKS
CO RN IN G, N. Y.
I
8
INDUSTRIAL
AND
ENGINEERING
CHEMISTRY
Vol. 16, No. 1
Official U.S. Army Photograph
H O W
IS
THE
W A T E R ?
Testing w ater supply is one of the jobs of the Army M edical Corps.
The Army alw ays p lays safe with the health of the soldiers, who must be
protected against contaminated w ater. In this ever-constant watchfulness
over Army w ater supply, H ellige Sim plex Testers are helping in provid­
ing a quick and accurate means for securing true chlorine determinations.
3718 Northern Blvd.
Lo n s Island City 1
New York
January 15, 1944
ANALYTICAL
EDITION
9
ÉXAX-
a n d ASSURANCE throng)
The Visible Guarantee o f In v is ib le Q u a lity
INDUSTRIAL
10
AND
ENGINEERING
CHEMISTRY
Vol. 16, No. 1
the original
building of O hio
W esleyan University
— founded in
w a s the n am e of the cou rse in
the first classroom instruction in chemistry at Ohio W esleyan University.
The course began in 1845 and w a s housed in the old M ansion House —
later nam ed Elliott Hall
Edgar Hall now houses the modern fire­
proof chem ical laboratories w here Multiple Unit Organic Combustion
Furnaces are used for instruction and research in organic chemistry.
See your Laboratory Supply Dealer or send fo r Bulletin H D -135 A
H E V I
D U T Y
E Ł E C T R IC
COM P A N Y
T «4 0 E m a rk
LABORATORY FURNACES MULTIPLE UNIT ELECTRIC EXCLUSIVELY
REG. U. S. RAT. OFF.
iffe
PÄS
JBBMRMEB
MILWAUKEE,
WISCONSIN
1»
-1
January 15, 1944
ANALYTICAL
L in d b e r g L a b o r a to r y fu r n a c e s , h o t p la te s a n d
c o n tro l, m o d e r n in d e sig n , a r e e n g in e e r e d a n d b u i l t to
m e e t th e e x a c tin g e v e r y d a y r e q u i r e m e n ts o f y o u r
la b o r a to r y .
L in d b e r g C o m b u s tio n T u b e F u r n a c e —T e m p e r a tu r e s
to 2500° F . F o r f a s t m o d e r n v o lu m e tr ic a s w e ll a s
g r a v im e tr ic ty p e s o f d e te r m in a tio n s . E q u ip p e d w ith
b u i l t - i n tr a n s f o r m e r , a ll n e c e s s a r y c o n tro l, G lo b a r
e le m e n ts a n d t u b e a d a p to r s .
L in d b e r g B o x F u r n a c e s —T e m p e r a tu r e s to 2000° F .
E x p r e s s ly b u i l t f o r th e m a n y u s u a l la b o r a to r y jo b s,
s u c h a s : a s h d e te r m in a tio n s , fu s io n s, e tc. E q u ip p e d
w ith h e a v y , lo w v o lta g e , r o d ty p e e le m e n ts . B u ilt in
tw o siz e s w ith c h a m b e rs 4 1/£>" x 10" x 4" a n d ‘7 % " x 14"
x 5% ".
L in d b e r g H o t P la t e s —T e m p e r a tu r e s to 900° F . T h e s e
h o t p la te s , a v a ila b le in th r e e siz es, 10" x 12", 12" x 20"
a n d 12" x 30", h a v e sp e c ia l c a s t m e ta l to p s . C o ile d
n ic k e l- c h r o m iu m e le m e n ts a r e p la c e d to g iv e u n if o r m
f a s t h e a tin g .
T h e In d ic a tin g P y r o m e te r a n d I n p u t C o n tro l
m o u n te d in th e s a m e c a s e p r o v id e th e la b o r a to r y w ith
“s te p le s s ” te m p e r a t u r e s a n d e m b o d y th e fa m o u s
L in d b e r g “p e r c e n t of c u r r e n t in p u t ” c o n tro l.
EDITION
11
Box Type Furnaces
Combustion Tube Furnace
Hot Plates
Lindberg Input Control
and Pyrometer
F o r com plete details and information get in touch
with y o u r lab o ra to ry equipm ent dealer today.
LINDBERG E N G IN E E R IN G COM PANY
2 4 5 0 W EST H U B B A R D
STREET
-
C H IC A G O
1 2 , IL L IN O IS
W ell-known Throughout the
W orld as the Leaders in D eveloping
and Manufacturing Industrial Heat
Treating Equipment.
12
INDUSTRIAL
AND
ENGINEERING
CHEMISTRY
when you need chemicals
b y the pound or carload ...
come to us with confidence
★ A 32 page listing of industrial
chemicals supplied by Harshaw
is available to you in booklet
form. Write for your free copy.
HARSHAW CHEMICAL
1945 East 97th Street, Cleveland 6, Ohio
B R A N C H E S I N P R I N C I P A L C IT IE S
Vol. 16, No. 1
January 15, 1944
ANALYTICAL
EDITION
13
booking F o r w a r
W a r tim e a c h ie v e m e n ts in scien ce a re
d e v e lo p in g a n e ra o f p ro g r e s s w h i c h
c h a lle n g e s th e im a g in a t io n .
A ll t h a t S p e n c e r is d o in g n o w — p r o ­
A t th e w a r ’s e n d , S p e n c e r w ill be
re a d y to se rv e sc ie n tific o p tic a l n e e d s o n
a f a r b r o a d e r sc a le t h a n e v e r b e fo re .
d u c in g m ic ro s c o p e s , p e ris c o p e s , te le ­
sc o p e s, a ir c r a f t a n d a n ti - a i r c r a f t g u n s i g h ts , p ris m b in o c u la r s , a z im u th in ­
s tr u m e n ts f o r d ir e c tin g a r t i l l e r y fire,
ta n k s i g h ts , te le s c o p ic a lid a d e s fo r n a v i­
g a t i o n , p r o je c to r s f o r in s t r u c t i o n — w ill
r e a p p e a c e tim e re w a r d s in a d v a n c e d
Spencer
k n o w le d g e , b e tt e r m a n u f a c tu r in g te c h ­
S C I E N T IF I C IN S T R U M E N T D IV IS IO N O F
n iq u e s , fin e r in s tr u m e n ts .
AMERICAN
-L
L EN S C O M P A N Y
B U FFA LO , NEW Y O R K
OPTICAL COMPANY
14
INDUSTRIAL
AND
ENGINEERING
CHEMISTRY
Vol. 16, No. 1
B EC K M A N P H O T O E L E C T R IC
QUARTZ S P E C T R O P H O T O M E T E R
A self-contained, precision instrument with quartz prism, operating on an electronic circuit, for the
rapid measurement of percentage transmission and density
, W av e le n g th Scale; B, B u ilt-in E lectro n ic In d ic a tin g M eter: C , Slits w ith precision a d ju stm e n t; D , L ig h t Source; E , C o m p a rtm e n t for tw o
P h o to tu b es; F , H older for fo u r 10 m m A b so rp tio n Cells; G, F ilte r Slide:; H , C o m p a rtm e n t fo r A b sorption Cells; J , P h o to tu b e Selector;
an d K , Sw itch fo r checking d a rk c u rre n t
QUARTZ SPECTROPHOTOMETER, Beckman Photoelectric. A self-contained unit with which a
wide variety of research and control work can be conducted with convenience, speed and precision.
Consisting of a quartz monochromator, with light source, holder for absorption cells, twin phototubes
and built-in electronic meter for translating phototube currents into direct readings of percentage
transmission and density. See Cary and Beckman, "A Quartz Photoelectric Spectrophotometer,”
Journal of the Optical Society of America, Vol. 81, No. 11 {November, 1941), p. 682.
M onochrom ator. Auto collim ating type, w ith 30° q uartz
prism of selected crystal w hich provides high dispersion in
th e ultra-violet. W avelength scale approx. 100 cm long,
graduated from 200 m m u to 2000 mmu, readable to 0.1 m m u
in th e ultra-violet and to 1.0 m m u in th e red, w ith a scale
accuracy of 1 mmu.
Optical p a rts rigidly m ounted in a massive h eat-treated
iron block w ithin a dust-proof steel case.
m arked “0.1” provides a ten-fold expansion of th e tra n s­
mission scale for m ore accurate readings on solutions below
10% transm ission.
Slits. P rotected by quartz windows, w ith stra y light ef­
fects reduced to a minim um. Sim ultaneously and continu­
ously adjustable from 0.01 to 2.0 m m b y a precision m echa­
nism. Slits can not be dam aged by closing too far. F ull
scale reading w ith nominal b and w idth less th a n 2 m m u over
all b u t th e extrem e ends of th e spectrum .
Sam ple H olders. A bsorption cells are accom m odated in
a rem ovable holder w hich is inserted in a lig ht-tight com­
p a rtm en t and is operated from th e front of th e in strum ent
by m eans of a sliding rod. Cells and holders arc available
for 10, 20, 50 and 100 m m liquid lengths.
Electronic Indicating M eter. A built-in potentiom eter
and electronic amplifier m akes possible direct readings in
percentage transm ission an d density. T he sw itch position
9101.
Light Source. A stan d ard 32 c. p. 6-volt, tu n g sten lam p
serves as a light source for th e range 320 to 1000 mmu.
F or th e ultra-violet range below 320 m m u a sm all hydrogen
discharge lam p is offered w ith power supply.
P hototubes. Tw o phototubes are furnished in a com part­
m ent which adjoins th e cells. A sliding rod brings either
tu b e into position and sim ultaneously switches th e electrical
connections.
Q u a rtz S p ectro p h o to m eter, B eck m an P h o to electric, M o d el D , ra n g e 320 to 1000 m illim icrons. C onsisting of m ono ch ro m ato r w ith
q u a rtz prism an d tw o slits, b u ilt-in electronic m eter, 6-voLt tu n g ste n la m p in d e ta ch a b le housing, one each caesium -oxide a n d b luesensitive p h o to tu b es, an d holder w ith selected s e t of fo u r Corex glass ab so rp tio n cells fo r 10 m m liq u id le n g th . W ith d ry cells for
op eratin g th e m e te r b u t w ith o u t 6 -v o lt sto rag e b a tte ry as req u ired for o p e ra tin g th e tu n g ste n lam p a n d electronic tu b e filam ents.
O v erall dim ensions, 30 inches long X 9 inches high X 9>/i inches deep; shipping w eight, 110 lbs...........................................................764.00
9101-B. D itto , M o d el DUY, ran g e 220 to 1000 m illim icro n s, id en tical w ith above b u t w ith ultrav io let-sen sitiv e p h o to tu b e in place of th e bluesensitive tu b e and w ith th e ad d itio n of accessories for fa r u ltra v io le t consisting of one p a ir of F used Silica A bsorption C ells for 10 m m
liquid length, H y drogen D ischarge L am p, housing for sam e, an d a pow er su p p ly u n it, 110 volts, 5 0 /6 0 cycles, for m a in ta in in g the
discharge a t c o n sta n t in te n sity . Shipping w eight, 165 lb s..................................................................................................
1052.00
Copy of 7-pp detailed description sent upon request.
ARTHUR H. T H O M A S COMPANY
R E T A IL — W H O LE SA LE — E X P O R T
L A B O R A T O R Y A P P A R A T U S AND R E A G E N T S
W E S T W A S H I N G T O N S Q U A R E , P H I L A D E L P H I A 5, U . S . A .
Cable Address, “Balance,” Philadelphia
IN D U STRIAL
and
EN G IN EER IN G CHEM ISTRY
P U B L I S H E D BY
THE A M E R I C A N
CHEMICAL
WALTER
SOCIETY
J. M U R P H Y ,
EDITOR
Emulsion Polymerization of Synthetic Rubber in 10-Gram
Systems
A n Exp e rim e n tal T e c h n iq u e
C H A R L E S F. F R Y L IN G , The B. F. G oodrich Com pany, A k ro n , O h io
The experimental procedure consists of sealing the ingredients of a polymerization
recipe into a test tube and rotating the tube at a constant temperature, following the
course of the reaction by noting the decrease in volume of the system. The latex is
removed for coagulation when the polymerization has proceeded as far as desired,
and the y ie ld is determined by weighing the dried stabilized coagulum.
C ontrol testing of raw m aterials. T h e particu lar properties
essential for polym erization of all shipm ents of m aterials intended
for large-scale production can be tested quickly. A good corre­
lation can be obtained between the experim ental results of this
m ethod and behavior on a m anufacturing scale.
D E V E L O P IN G practical recipes for th e m anufacture of
INsynthetic
rubber, it was desirable to investigate the effects
on the polym erization process of a large v ariety of highly puri­
fied substances. Only sm all q u antities of m any m aterials were
available. F urtherm ore, in order to obtain valid comparisons
betw een experim ents conducted over a period of time, it was
necessary to keep standardized sam ples of th e m ajor compo­
nents and to use them as econonomically as possible. These
considerations led to the developm ent of a sm all-scale polym eri­
zation technique w hereby from 10 to 20 gram s of monomers
could be employed. B y this m ethod an extraordinary am ount
of valuable inform ation was obtainable. T his included:
T he m ost serious lim itation to this technique is th a t m aterials
cannot be added to or su b tracte d from th e system once poly­
m erization has started . [Balandina el al. described a sim ilar
technique, th e details of which are n o t readily available to
English-speaking investigators (I).]
RECIPES FOR POLYM ERIZING SYNTHETIC RUBBER
The developm ent of satisfactory polym erization recipes is one
of the im p o rtan t objectives of synthetic ru b b er research. P a te n t
literature contains m any examples of such. The following, from
a U. S. p a te n t issued to W ollthan and Becker ([5), and recalculated
to th e scale of this technique, is perhaps typical:
1. Yield
2. A polym erization reaction curve from which to estim ate
length of induction period, if an y ; rate of polym erization a t any
desired tim e; and over-all conversion a t any desired time
3. K ind of emulsion— i.e., w hether fluid, viscous, gelatinous,
o r heterogeneous, a t any stage in the process
4. pH of emulsion a t end of process
5. Q ualitative observations on coagulation
G. P roduction of a sam ple of sy nthetic rubber sufficiently
large to determ ine solubility, milling characteristics, and cured
properties by procedures described by G arvey (3)
B u tad ien e
S ty ren e
Isohexyl m e rc a p tan
W a te r
Sodium oleate
A m m onium p e rsu lfa te
T e m p e ra tu re
T im e
Y ield
T his small-scale technique has been employed for investigating
polym erization of a large num ber of m onom ers an d comonomer
m ixtures, for evaluating em ulsifying agents, for determ ining the
effect of im purities in the reagents employed, for varying the
comonomer ratio and the ratio of hydrocarbons to aqueous phase,
an d for investigating behavior o f various initiators, inhibitors,
an d other ingredients of th e polym erization recipe. I t proved to
be especially advantageous in providing inform ation on the
effect of some one ingredient over a range of concentrations.
T h e influence of reaction tem perature was readily determ ined.
In general, th e advantages of th e technique were particularly ap­
p a ren t in:
7 .5 gram s
2 .5 gram s
0 .0 5 g ram
1 8 .0 gram s
2 .0 gram s
0 .0 3 gram
30° C.
“ S everal d a y s”
“ E x c e lle n t”
An understanding of th e function of each ingredient is essential
In th e above example, the butadiene and styrene are th e mono­
mers, which, by copolymerizing, form syn th etic rubber. T he
isohexyl m ercaptan is described as a substance exerting a “regu­
lating effect” —i.e., it possibly decreases th e branching charac­
teristics of th e resulting polym er. Sodium oleate is th e emulsify­
ing agent, an d th e am m onium persulfate acts as th e polym eriza­
tion initiator, also called th e “polym erization c ataly st” .
T he possibilities of research on such a system are g reat and
are increased by the fact th a t a v ariation introduced in an y one
ingredient m ay require a concom itant change in some oth er in­
gredient. F o r example, su b stitu tin g an o th er substance for the
in itiato r m ight require a change in th e ty p e of em ulsifying agent
employed in order to get satisfactory results.
Prelim inary surveys, where wide areas of investigation had to
be covered in the sh o rtest possible tim e. T h e m ethod is am en­
able to sim ple labor-saving tricks, such as filling a large num ber
of reaction tubes a t the sam e tim e w ith solutions of a given emul­
sifying agent.
1
2
INDUSTRIAL
AND
ENGINEERING
GEN ERA L CONSIDERATIONS O N TECHNIQUE
Polym erization, while a science, is also an a rt. T he w ay in
which things are done— th a t is, th e niceties of experim ental tech­
nique em ployed—is of equal im portance to th e scientific aspects
of the subject. Unless this view point is k ep t clearly in m ind, the
investigator is frequently confronted by baffling failures. E m ul­
sion polym erization is particularly susceptible to the influence of
traces of contam inants. T he equipm ent of a research laboratory
m ay be covered w ith d u st which contains inhibitors o r accelera­
tors of polym erization, indeed, some substances, which under
certain conditions inhibit polym erization, m ay under sightly dif­
ferent conditions a c t as catalysts. N evertheless, th e difficulties
confronting the investigator can be avoided w ith a little care and
forethought.
W eighing Butadiene
In general, solutions or other substances to be employed in
polym erization experim ents should n o t be exposed to atm os­
pheric contam ination for longer periods than necessary. Glass
stoppers afford adequate protection. T he contents of a flask
m ay be tem porarily protected by covering the m outh w ith a
sheet of clean dry tinfoil. Cork stoppers should be avoided;
b u t if necessary, they can be covered w ith tinfoil. In no case
should the alkaline contents of flasks come in con tact w ith tin ­
foil.
Glass bottles and flasks m ust be clean. In m ost cases tre a t­
m ent w ith chromic acid cleaning solutions, followed by rinsing
w ith tap w ater and then distilled w ater, is adequate. R eaction
tubes, however, require more effective cleaning.
In one operation it is convenient to pour th e volatile contents
of a D ew ar flask through a sh o rt length of rubber tubing. Al­
though rubber contains antioxidants, accelerators, and other
chemicals, no trouble is experienced from th is source if th e rubber
tubing is first extracted by boiling in several changes of acetone.
T his can be done (on a steam plate) in a covered beaker if a sizable
piece of dry ice is placed on th e w atch crystal, which thereby be­
comes a convenient reflux condenser. T h e extracted tubin g can
be k ep t in a stoppered w ide-m outhed b o ttle for future use.
C ertain monomers can be efficiently separated from powerful
inhibitors added as stabilizers by d istillation through relatively
simple equipm ent. T he practice of distilling monomers in the
absence of an inhibitor should be avoided because of th e danger
of explosions due to the accum ulation of peroxides in the distillation
residue.
I t has been custom ary in th e laboratory to prepare fresh sam ­
ples of butadiene by condensation in clean glassware from a
stream of gas taken from a large cylinder of hnuefied m aterial.
H igher bofling monom ers are freshly prepared b y atm ospheric,
vacuum , or steam distillation as required. All-glass distillation
equipm ent is m ost satisfactory. T h e unstabilized monom ers
can be k ept stoppered in a refrigerator a t —30° C. for several days
w ithout detectable deterioration.
Sometimes repetition of an operation is advisable. M etallic
CHEMISTRY
Vol. 16, No. 1
polym erization vessels, no m a tte r how carefully cleaned, m ay be
inhibited; merely em ptying and recharging are often sufficient
to ensure a satisfacto ry reaction.
T here is no su bstitu te for constant care and cleanliness on th e
p a rt of th e investigator. A careful experim enter can easily a d ­
ju s t th e w eight of sm all portions of certain m onom ers using a
clean m edicine dropper, while a careless experim enter (perform ­
ing th e sam e operation) can ruin a large num ber of experim ents
by allowing th e m onom er to come in to contact w ith the ru b b er
bulb of th e medicine dropper.
EXPERIMENTAL PROCEDURE
T he ingredients of a polym erization recipe are sealed into a test
tube, which is ro ta te d a t a constant tem perature. T he course of
the reaction is followed by noting th e decrease in volum e of th e
system , and th e latex is rem oved for coagulation when th e poly­
m erization has proceeded as far as desired. T h e yield is deter­
m ined by w eighing th e dried, stabilized coagulum.
Pyrex reaction tubes, 22 mm. in diam eter, approxim ately 55
ml. in capacity, to th e upper end of which are sealed 10-mm.
d iam eter Pyrex tubes, m ay be obtained in gross lots from the
Corning Glass Com pany, according to the following specifica­
tions: "G lass tubes, Pyrex, 22 mm. O.D. X 1.5 mm . walls (uni­
form), 215-mm. body to neck, 22-mm. tapered shoulder, 145mm. neck. 10-mm. neck X 1-mm. wall thickness” . These can
be m ade by th e experim enter, b u t it has been found cheaper to
purchase them .
New tubes are cleaned by rinsing w ith distilled w ater and an­
hydrous c .p . sy nthetic m ethanol, in th a t order, and evacuating
un til dry. E vacu atio n m ay be accomplished w ith a Cenco
H yvac pum p connected in train through a d ry ice-acetone trap ,
which condenses the m ethanol and prevents diffusion of any vola­
tile in h ib ito r back into th e tube. A cetone-extracted rubber
tubing is used to a tta c h the reaction tubes to the vacuum system .
Old tubes, after being cleaned in a chrom ic acid b ath , are re­
paired b y sealing new necks of 10-mm. Pyrex tubing, 10 cm.
long, to the shoulder of the tubes. T he open ends are fire-pol­
ished. T h e tubes are then subjected to the cleaning treatm en t
recom mended by Suess, Pilch, and R udorfer (4). Clean con­
centrated n itric acid is poured into th e tubes and allowed to
stan d from 16 to 24 hour's. If the tubes are reauired a t once,
they are filled to th e shoulder w ith n itric acid and gently heated
from 15 to 30 m inutes, then rinsed three tim es w ith tap w ater.
One rinse should completely fill the tube to displace all th e fumes.
A fter two additional rinses w ith distilled w ater, the tubes m ay be
dried w it Ip synthetic m ethanol and evacuated. If desired, the
m ethanol rinse and evacuation can be dispensed w ith by allowing
the distilled w ater to drain from th e inverted tubes overnight.
T h e d ry tubes are stoppered w ith No. 0 corks, which have been
covered w ith fresh tinfoil, and th e tubes are kept in a clean place
until used.
Solids m ay be added to the reaction tubes in q u an tities as low
as 0.5 mg. from small alum inum foil weighing scoops. If dupli-
A d d in g Butadiene to Reaction Tube
January IS, 1944
ANALYTICAL
catc quantities are to be used throughout a series of experim ents,
it is convenient to dissolve organic compounds in the less volatile
comonomer and inorganic compounds in the aqueous solution of
emulsifying agent.
EDITION
3
visable to do th e sealing close to th e open end of the neck in an
oxidizing atm osphere; otherwise a carbon m irror m ay form on
the interior surface of th e tubing and prevent a tig h t seal. Oc­
casionally, if condensation is n o t complete, a slow blue flame tra v ­
els from the heated glass down into th e reaction tube. I t has
been impossible, however, to dem onstrate th a t th is brings about
any variation in the ensuing polym erization.
POLYM ERIZATION
T h e sealed reaction tu b e is brought to reaction tem perature
by immersion in w ater. T h e height of th e m eniscus is d eter­
mined and recorded, using a millim eter rule. T h e tu b e is then
rotated a t a constant tem perature and readings of th e meniscus
height are m ade periodically. I t is from these readings th a t re­
action curves such as Figure 1 can be plotted.
If the polym erization is complete in less th a n 5 hours, a w ater
therm o stat is required to prevent tem perature buildup. How­
ever, a therm ostatically controlled air cabinet, provided w ith
shafts for ro tatin g the tubes, is more convenient.
Rotation of Tubes in Constant-Temperature Cabinet
' T he solution of em ulsifying agent is generally prepared sepa­
rately and added to the reaction tubes from a pipet or a g raduated
cylinder. Since slight v ariatio n s in th e ratio of m onom ers to
aqueous phase have little effect, convenience generally d ictates
the use of th e la tte r m ethod. T h e volum e of em ulsifying agent
employed m ay vary from 10 to 30 ml. for 10 gram s of m onom ers.
T he reaction tubes containing the em ulsifying agent and other
ingredients are placed in a refrigerator, inclined a t an angle of
2° 8 ' from horizontal, and frozen a t —30° C. T his inclination
m ay be obtained by placing a piece of 10-mm. glass tubing under
th e necks of th e tubes. Unless chilled in th is position, th e tubes
will crack when placed in a d ry ice-acetone bath. (The corkstoppered glass reaction tubes m ay be laid on a table a t the
proper inclination and covered w ith d ry ice. However, any
contam ination of the contents by carbon dioxide will alter the
pH of th e soap solution and affect th e reaction ra te seriously.)
If care is exercised, the aqueous phase can be frozen by direct
im mersion of th e tubes in a d ry ice-acetone b ath , w ith frequent
w ithdraw als and nearly horizontal rotations, so th a t th e aqueous
phase freezes in contact w ith th e glass in th e form of a hollow
cylinder. T he losses due to cracking of th e glass are high and
th e operation is tim e consum ing; therefore th e m ethod of freezing
first described is preferred in m ost cases.
In the next step the frozen tubes are individually cooled further
in a dry ice-acetone m ixture contained in a quart-size straig h t­
sided D ew ar flask. A rubber dam is fitted over th e tu b e and
th e D ew ar flask, the neck of th e tube extending through a. hole
in the dam. T his minim izes contam ination of the tu b e w ith es­
caping carbon dioxide, and holds it in a convenient vertical po­
sition.
T he higher boiling comonomer is weighed to 0.1 gram in to a
tared, clean m icrobeaker and poured into th e reaction tube. A
b u ret is som etim es used for th is operation b u t contam ination by
stopcock grease should be avoided. In either case th e opera­
tion m ust be performed in a rigorously clean manner.
Freshly distilled butadiene is tem porarily contained in a pintsize D ew ar fitted w ith a two-hole extracted ru b b er stopper
through w hich extend two sh o rt glass tubes arranged for con­
venient pouring. T he tem perature of th e butadiene is held a t
—30° C. I t is weighed to 0.1 gram into a sm all silvered D ewar
weighing flask using a torsion balance. T he small D ew ar is then
attached to th e reaction tube by a sh o rt length of extracted rub­
b er tubing and th e butadiene is poured into th e reaction tube,
allowing ab o u t 45 seconds for condensation of vapor before re­
moving the rubber tube. T h e weighing flask has a 10-mm. neck,
and is 11.5 cm. from bottom to shoulder, 35 mm. in outside di­
am eter, 25 mm. in inside diam eter, and ab o u t 16 cm. in oyer-all
length. Such flasks have been m ade in the lab oratory b u t it has
been found more satisfactory to have them m ade by professional
glass blowers.
T h e reaction tube is scaled, using a hand blow torch. I t is ad­
Tl
Figure 1.
Polym erization Curve
A reaction tube, prepared as described, using th e W ollthan and
Becker recipe (5), will give an initial meniscus height of approxi­
m ately 110 mm . D u rin g the course of polym erization, it will
drop 11 m m . Since the height can be read to 0.5 m m ., this pro­
cedure provides a simple m ethod of following th e reaction rate
w ith an accuracy of a b o u t 5% . T able I shows for o th er recipes
th a t th e drop in height of th e meniscus is directly proportional
to th e yield of polym er, w ithin th e lim its of error of th e m ethod.
If th e to ta l decrease of m eniscus height is m easured ju s t before
opening th e tube, th e yield, as measured on th e d ry polymer, will
Table I.
Correlation between Percentage Polym erized and Fall of
Meniscus for Three Polymerization Recipes
F a ll of
M eniscus
Mm.
------------------ Y ie ld C alculated
from m eniscus
%
19
44
65
89
Series 1
3 .0
7 .0
10.4
14.1
Series II
3 .0
7 .0
10.9
14.4
M easured
%
18
39
61
89
21
43
70
93
Series I I I
2 .0
4 .0
4 .5
5 .8
7 .6
7 .8
11.0
12.6
14.7
9
20
26
38
44
47
67
86
94
D ifference
%
+ 1
+ 5
+ 4
19
45
70
93
_2
+2
0
13
26
29
37
49
50
70
81
94
+ 4
4-6 ■
+ 3
-1
4-5
4-3
4-3
—5
4
INDUSTRIAL
AND
ENGINEERING
CHEMISTRY
Vol. 16, No. 1
essary to break th e tip and direct the violently expelled foam into
a large beaker. T he synthetic rubber latex m ust next be sta­
bilized by th e addition of an antioxidant. Two per cent of dis­
persed phenyl-beta-naphthylam ine has been found convenient
and satisfactory. T he dispersion of the stabilizer can be ob­
tained by aqueous dilution of an alcoholic solution. The latex
can be coagulated by any m ethod custom arily employed for
breaking emulsions or coagulating n atu ral rubber latex. Follow­
ing coagulation, th e rubber is w ashed free of soap and electro­
lytes, using a B uchner filter and filter paper, and dried in air,
preferably a t a low tem perature. T h e yield, accurate to ± 2 % ,
can be obtained by weighing to 0.1 gram.
W ashing Coagulated Synthetic Rubber Samples
give an accurate basis for calculating th e p artia l yields a t any
given tim e during the process. R eaction curves p lotted using
these figures will n o t be affected by errors due to v ariations in
d iam eter of the individual tubes.
The form ation of foam, w hich breaks w ith difficulty, frequently
interferes w ith th e m easurem ent of meniscus height. In such a
case the tubes can be centrifuged by swinging in a suitable tube
on the end of a 90-cm. (3-foot) rope. [According to R eynolds (S)
th e accuracy of this m ethod can be im proved by constriction of
the tube a t the position where th e meniscus is read, together w ith
high-speed centrifuging and redispersion of th e emulsion b y shak­
ing after the reading.) If a gel forms, or if there is m uch coagu­
lation during polym erization, the height of th e meniscus cannot
be determ ined accurately.
The end of the induction period, or the beginning of polym eri­
zation, is generally indicated by the appearance of a bluish opales­
cence, in addition to the change in height of th e meniscus.
DETERMINATION OF YIELD
Opening the reaction tubes presents no difficulty except when
low partial conversions are under investigation. T hen it is nec­
SUMMARY
M any m anufacturers of monomers, emulsifying agents, ini­
tiators, modifiers, and oth er ingredients going into polym eriza­
tion reactions find it necessary to have a reliable polym erization
procedure for testing the quality of th eir products. T he proce­
dure presented, despite some shortcomings, has m any advan­
tages. M inim um am ounts of m aterial are required and large
num bers of experim ents can be conducted in a relatively short
time. A serious effort is m ade to point o u t some of the pitfalls
which beset investigators of polym erization regardless of the
type of technique employed. T h e best recom m endation for the
procedure described is th a t it has been used to develop certmn
types of syn th etic ru b b er w hich are now in commercial produc­
tion.
LITERATURE CITED
(1) Balandina, V ., et a l., Bull. acad. sci. U. R . S . S ., d a sse set. m ath.
n ot., S er. chim ., 1936, 397-407.
(2) G arvey, B . S ., J r ., I n d . E n g . C h e m ., 34, 1320-3 (1942).
(3) R eyn old s, W . B ., personal com m unication.
(4) Suess, P ilch, and Rudorfer, Z . ph ysik. Chem., A179, 361 -7 0
(1937).
(5) W ollthan and Becker, U . S. P a ten t 2,281,613 (M ay 5, 1942).
P
resented
of th e
A
before th e D ivision of R u b b e r C h em istry a t th e 105th M eeting
C h e m i c a l S o c i e t y , D e tro it, M ich.
m e r ic a n
Determination of Tetraethyllead in Gasoline
H A R R Y G O N I C K AN D J. J . M I L A N O , Shell O il Com pany, Incorporated, M arlinez, Calif.
A method for the determination of tetraethyllead in gasoline is described in which the
tetraethyllead Is decomposed with iodine and the lead subsequently titrated by a new
acidimetric method em ploying 8-hydroxyqulnoline. The method is rapid and is ap­
plicable to all types of gasolines.
T IL recently
U Ndeterm
ination
th e m ost widely used m ethod for the
of tetraethyllead in gasoline was the
bromirration m ethod described by E dgar and C alingacrt (S)
in which the tetraethyllead was decomposed by th e action of
bromine. Although this m ethod was rapid and convenient for
the determ ination of tetraethyllead in straig h t-ru n gasolines,
difficulties were encountered w ith cracked gasolines owing to
the rapid absorption of bromine by the olefins present, in com­
petition w ith the tetraethyllead. W ith gasolines of high olefiu
content it was necessary to brom inate the gasoline completely7
to ensure complete decomposition of th e tetraethyllead. E ven
so, low results were frequently7 obtained. M oreover, th e q u an ­
tity of bromine required for complete brom ination of a gasoline
of high olefin content was ra th e r large (frequently in excess of
200 gram s) and added substantially7 to th e cost of th e analysis.
In addition, th e brom ination reaction was violent and was
accom panied by th e evolution of corrosive vapors w hich caused
considerable hazard to the operator. F or these reasons, the
brom ination m ethod w as n o t suited to th e routine analysis of
cracked fuels.
More recently other m ethods have been devised in w hich the
gasoline is treated w ith hydrochloric acid and th e load determ ined
in the acid extracts. T h e best known of these is th e m ethod of
C alingacrt and G am brill (3), recently adopted by th e American
Society for T esting M aterials as a ten ta tiv e stan d ard (I). Al­
though this m ethod gives satisfactory results w7ith all types of
gasolines, the over-all tim e required for a determ ination is some­
w h at lengthy7and specialized equipm ent is required.
T he m ethod described in this paper was developed in an
effort to reduce th e tim e required for the tetraeth y llead d eter­
m ination. B y th e proposed method, a single determ ination can
be completed in one hour and for determ inations in q u an tity only7
a small fraction of th is tim e is required p er determ ination. The
accuracy7 of the m ethod appears to be equal to previous m ethods
and docs n o t appear to be affected by7 th e ty p e or composition
of th e gasoline. More th a n three thousand sam ples of gasoline,
January 15, 1944
ANALYTICAL
representing all the principal brands and grades sold in th e
western states, have been analyzed successfully.
PRINCIPLE OF PROPOSED METHOD
As in previous m ethods, the determ ination of the te tra e th y l­
lead divides itself into two distinct parts: (1) th e decom position
of th e tetraethyllead to yield an inorganic lead salt, and (2) th e
determ ination of the lead.
EDITION
rate, c .p ., crystals. H ydrochloric acid, C .P ., dilute solution;
1 to 1. Sodium chloride, c .p ., crystals. 8-H ydroxyquinoline,
0.065Ar in GO per cent isopropyl alcohol. S ta n d a rd sodium
hydroxide, 0.0624AL S tandard hydrochloric acid, 0.0624Ar.
(Standard 0.0624Y acid an d base were selected since these re­
agents are in general use in oil laboratories.)
M ethyl red indicator; dissolve 1 gram in 600 ml. of alcohol
and dilute to 1 liter w ith w ater. Phenol red indicator, 0.2 gram
per liter of water.
S e p a ra tio n op L ead .
M easure exactly 100 ml. (corrected to
G0° F.) of the gasoline to be tested into a 500-ml. Erlenm eyer flask.
Add 50 ml. of th e iodine solution and allow to sta n d for a t least
5 minutes. Place the flask on a h o t plate under a h o t air stream
and evaporate the gasoline to dryness. T he velocity of th e air
stream and the tem perature of the h o t p late should be regulated
so as to secure the maximum rate of evaporation w ithout sp a t­
tering or bumping. T he air stream effectively suppresses the
tendency tow ard bum ping which is alm ost unavoidable w ithout
its use. The evaporation ordinarily takes from 15 to 20 m inutes.
Add 25 to 50 ml. of concentrated nitric acid (depending on the
am ount of th e organic residue) and ro tate th e flask over a burner
until dense fumes of iodine and nitrogen dioxide cease. Should
any organic m a tte r adhere to th e walls of the flask continue
rotating the flask until it is completely dislodged.
Figure 1.
Effect of Sodium Chloride on Neutralization Curve
of Solutions Containing Lead Ions
0.137 gram of IPb(N03)j In 100 ml. of water. (1) No NaCI added, (2)
5 grams of NaCI added
1. Experim ents conducted in this lab oratory showed t h a t
tetraethyllead is rapidly an d quantitativ ely decomposed by th e
action of free iodine. U nsaturated hydrocarbons do n o t interfere,
as they do n o t iodinate so rapidly as to compete w ith th e te tr a ­
ethyllead reaction. T he gasoline is rem oved by evaporation
under a h o t air stream , and an y organic m a tte r rem aining is
subsequently oxidized w ith nitric acid and potassium chlorate.
T he inorganic residue rem aining th en contains all th e lead in the
form of inorganic salts including lead iodate, which is insoluble
in w ater. T he lead salts so obtained are converted to th e more
soluble chloride by treatm e n t w ith hydrochloric acid.
2. T he lead is determ ined b y an acidim étrie titra tio n m ethod ;
it is therefore necessary to neutralize th e solution exactly before
titra tin g the lead. T he neutralization of a solution containing
lead salts ordinarily presents difficulties owing to th e hydrolysis
of th e lead. T hese difficulties are obviated, however, b y the
resence of sufficient chloride ions which effectively suppress the
ydrolysis of the lead. T h e effect of sodium chloride on the
neutralization of a lead solution is shown by th e curves in Figure 1.
A fter th e solution has been neutralized, an exceas of 8-hydroxy­
quinoline is added. T his reagent reacts w ith th e lead ions to
liberate an equivalent q u an tity of acid which is th e n titra te d
w ith stan d ard alkali. T he reactions are assumed to be as follows :
i... i
'
Air in /e A
E
HOC„H,N = -OC»H„NH+
(1)
P b+ + + 2 "O C dH jN H + = P b(O C ,H ,N H + )2
(2)
Pb(O C tH jN H +)j + 2 0 H - = Pb(O C 8H ,N )2 + 211,0
(3)
As indicated in E quation 1, 8-hydroxy quinoline is am photeric
and goes over to th e ionic form. W hen 8-hydroxyquinoline re­
agent is added to a solution containing lead ions, lead 8-hydroxyquinolinium ions are formed according to E quation 2. These
are quantitatively titra te d w ith stan d ard alkali to p H 7 according
to E quation 3. O ther equilibrium reactions are undoubtedly
involved, including reaction between lead 8-hydroxyquinolinium
ions and excess reagent; however, the equations given indicate
the essential result.
DETAILS OF METHOD
The h o t a ir-je t evaporator (Figure 2 ) is designed
to direct a hot air stream into four Erlenm eyer flasks sim ul­
taneously during evaporations. A lthough th e design shown
has proved very satisfactory in actual practice, o th er designs
which will accomplish the same purpose m ay be used.
R e a g e n ts .
Iodine, sa tu ra te d solution in carbon tetrachloride
(technical). N itric acid, c.r., concentrated. Potassium chlo­
A p p a ra tu s .
Figure 2. H o t A lr - J e t Evaporator
A ll interior metal parts must be heat-resistant
To the actively boiling solution add crystals of potassium
chlorate until th e organic m a tte r is com pletely destroyed. T he
potassium chlorate should be added cautiously. T he solution
should n o t be perm itted to evaporate to dryness while th ere is
visible organic m a tte r present as indicated by a brow nish colora­
tion of the solution- otherwise spontaneous ignition will occur
and cause losses of lead. Usually all organic m atter disappears
after th e first 2 or 3 gram s of potassium chlorate have been added;
however, one or tw o additional portions of 2 or 3 gram s each are
added in order to complete th e oxidation. W ith practice, the
oxidation of residues from cracked gasolines w ith potassium
chlorate can be effected in 1 to 3 m inutes.
E vaporate th e clear nitric acid solution to com plete dryness.
Should the solution exhibit any darkening during the ev ap o ratio n ,
INDUSTRIAL
Table I.
AND
ENGINEERING
Determination of Tetraethyllead in Synthetic Gasoline
Blends by the Proposed M ethod’’
Sam ples
S0% cracked gasoline + 2 0% stra ig h t-ru n
gasoline
100% s tra ig h t-ru n gasoline
50% cracked gasoline - f 50% iso p en tan c
50% cracked gasoline + 50% alcohol
T e trae th y lle a d C o n te n t
C alcu lated
D eterm in ed
M il/gallon
3 .0 7
3 .0 7 ,3 .0 6 ,3 .0 7
1.54
1 .5 4 ,1 .5 4 , 1 .56
0 .2 0
0 .1 9 ,0 .2 0 ,0 .1 9
3 .0 0
3 .0 0 ,3 .0 0
1.50
1 .5 1 ,1 .5 0
0 .3 0
0 .2 9 ,0 .2 9
1.54
1.54
1.54
1.55
° T hese blends w ere m ade using eth y l fluid o b tain ed from th e E th y l
Corp.
add m ore potassium chlorate. The use of the hot air stream is
n o t recommended during this operation, as the cooling effect
of the air stream impedes the oxidation of possible traces of or­
ganic m atter. To ensure the complete destruction of organic
m atter, heat the residue over a burner until it is completely fused.
T he residue after fusion should be w hite (see note below on the
use of potassium cldorate).
Allow the flask to cool som ewhat, and add sufficient 1 to 1
hydrochloric acid to dissolve th e residue completely after 2 or 3
m inutes’ boiling. Usually 10 to 20 ml. of the dilute acid are
sufficient. A fter complete solution is effected evaporate to
dryness. Special care should be exercised tow ard th e end of the
evaporation, as the potassium chloride formed has a tendency to
spatter. Remove the rem aining acid as completely as possible
by thoroughly heating th e flask while blowing a h o t air stream
into it.
D e te rm in a tio n o f L ead .
Dissolve th e residue in the flask
in 150 to 200 ml. of distilled w ater, add 2 or 3 drops of m ethyl red
indicator, and exactly neutralize the solution w ith 0.0624;V
sodium hydroxide to the alkaline (yellow-) end p o in t of the
indicator. Usually there will be a sufficient concentration of
chlorides as a result of the preceding operations to suppress the
hydrolysis of the load. A deficiency of chlorides will render the
n eutral point indefinite, in which case 5 to 10 gram s of sodium
chloride should be added. A t th e neutral point one drop of
0.0624Ar acid should suffice to revert th e indicator color from a
canary yellow to a definite pink. Occasionally the m ethyl red
indicator will show a fading tendency, owing to remaining traces
of oxidizing substances. T his fading ten d en cy is readily over­
come by the addition of a few milliliters of 0.1 N sodium thiosulfate solution.
W here the q u an tity of lead present is approxim ately known,
th e titratio n of the lead is carried o u t as follows: T o the neu­
tralized solution add a 2- to 3-ml. excess of the 8-hydroxyquino­
line reagent and a sim ilar excess of the stan d ard sodium hy­
droxide solution. Stopper th e flask and shake vigorously for a
few seconds to break up th e precipitate and liberate any occluded
substances. Add sufficient phenol red indicator (about 3 ml.)
to produce a definite pink color and b ack -titrate the excess alkali
w ith standard hydrochloric acid. T he hydrochloric acid should
be added dropwise tow ard the end of the titratio n and th e end
point taken on the yellow (acid) side of the indicator change.
A gitate the flask when observing the end point and ignore any
pink fluorescence which m ay appear after settling of the pre­
cipitate. I t is advisable to redeterm ine th e end point by adding
a further excess of alkali and repeating th e b ack-titration w ith
acid. M ake certain th a t there is a 2- to 3-ml. excess of th e 8hydroxyquinoline reagent over th e n et volume of alkali con­
sumed.
W hen the approxim ate q u an tity of lead is n o t previously
known, th e titratio n m ust be carried ou t stepwise in order to
secure the correct excess (2 to 3 ml.) of the 8-hydroxyquinoline
reagent. Add about 3 ml. of phenol red indicator to th e solution
which has been previously neutralized to m ethyl red. Now add
th e S-hydroxyquinoline in 3-ml. increm ents an d after each addi­
tion add an equal volume of 0.0624Ar sodium hydroxide. An
excess of the S-hydroxyquinoline reagent is indicated when the
addition of the alkali increm ent renders the solution alkaline
(pink) to the phenol red indicator. A t th is point stopper the
flask, shake vigorously for a few seconds, and b ack -titra te the
excess alkali w ith standard acid to determ ine th e approxim ate
consum ption of alkali. A djust th e volume of 8-hydroxyquinoline
added, so th a t there is an excess of 2 to 3 ml. over the n et volume
of alkali consumed (volume of stan d ard alkali m inus volume of
standard acid). A larger excess of 8-hydroxyquinoline should
be avoided, as this reagent exhibits a slight buffering effect which
interferes w ith the end-point determ ination. R edeterm ine the
CHEMISTRY
Vol. 16, No. 1
end point after the addition of a further 2 to 3 ml. of standard
alkali by back-titration w ith stan d ard acid as already described.
Two equivalents of titra ta b le acid are formed for each mole of
lead. The solutions should be standardized against a known
q u a n tity of lead, using the same titratio n procedure as in the
analysis. P ure te st lead or lead n itrate m ay be used as a standard.
T he result expressed in milliliters of tetraeth y llead per gallon
of gasoline is obtained by m ultiplying th e PbO equivalent (ex­
pressed in gram s) of the n et volume of sodium hydroxide con­
sum ed b y 33.24.
U se o f P o ta s s iu m C h lo r a te .
In order to determ ine the
explosion hazard attending th e use of potassium chlorate-nitric
acid m ixtures for th e oxidation of organic residues, the effects of
various conditions were investigated. I t was found th a t a mild
explosion would sometimes occur in the vapor if th e liquid was
not k ep t actively boiling during the oxidation process. In every
case th e explosion was preceded by a dense accum ulation of
greenish yellow vapors (probably a m ixture of chlorine and or­
ganic vapors). T he explosion hazard appeared to be completely
elim inated by m aintaining th e solution in an actively boiling
condition during the oxidation process, in which case th e accum u­
lation of greenish yellow fumes was prevented. By following
this procedure more th an 3000 sam ples have been analyzed w ith­
o u t an explosion. In spite of this record it is suggested th a t a
protective m ask be w orn by the operator during the oxidation.
Table II.
Comparison of A .S .T .M . and Proposed Methods
Sam ples
C om p e titiv e Q gasolines
B ra n d
B rand
B rand
B ra n d
B ra n d
I
II
III
IV
V
T e tra e th y lle a d C o n te n t
A .S .T .M . m ethod
P roposed m ethod
M l./aaU on
0 .2 6
0 .0 6
1.26
0 .3 7
0 .5 3
0 .2 6
0 .0 7
1.25
0 .3 8
0.5 3
1.35
1.78
1.92
1.87
1.58
2.01
1.06
1.35
1 .8 0
1 .9 3 ,1 .9 4 ,1 .9 3
1.87
1.57
2.01
1.08
3.0 4
3.0 5
2 . 9 7 ,2 .9 7 ,2 .9 7
3 .0 4
3 .0 6
3 . 0 0 ,3 .0 0 ,2 .9 9
C om p e titiv e E th y l gasolines
B ra n d
B ra n d
B ra n d
B ra n d
B ra n d
B rand
B ra n d
I
II
III
IV
V
VI
V II
A viation gasolines
B ra n d I
B ra n d II
B ra n d I I I
ACCURACY OF THE METHOD
A series of synthetic blends of tetraethyllead in cracked and
straig h t-ru n gasolines was prepared to check th e accuracy of
the m ethod. Blends were aLso prepared w ith th e addition of
isopentane and alcohol and analyzed by th e described m ethod.
T he results of these experim ents are shown in T able I.
A comparison of results obtained by the proposed m ethod and
by the A .S.T.M . m ethod (!) is shown in Table II.
LITERATURE CITED
(1) A m . Soc. T esting M aterials, D esignation D 526-41T .
(2) Calingaert and G am brill, I n d . E n o . C h e m ., A n a l . E d ., 11, 324
(1939).
(3) Edgar and Calingaert, Ib id ., 1, 221 (1929).
American Society for Testing Materials Meetings
T he 1944 Spring M eeting and C om m ittee W eek of A .S.T.M .
is to be held in C incinnati, Ohio, a t th e N etherland Plaza from
Feb ru ary 28 to M arch 3. T he 47th A nnual M eeting will be
held in New York, N. Y., a t th e W aldorf-Astoria June 26 to 30,
1944.
Specific G ravity of Butadiene
M . R. D E A N A N D T. W . L E G A T S K I
Phillips Petroleum Com pany, Research Department, Bartlesville, O k la .
The specific gravities ( f/ 6 0 ° F.) of 1,3-butadiene of 9 9 .6
mole per cent purity were determined experimentally for
the temperature range of — 1 7 .7 8 ° to 6 0 ° C. ( 0 ° to 1 4 0 ° F.),
by means of a specially constructed steel pycnometer of
approximately 4 5 0 0 -cc. capacity and capable of with­
standing the resultant vapor pressures without under­
going a permanent change in volume. The experimentally
determined specific gravities were smoothed graphically
and the experimental and smoothed values were compared
to show the magnitude of the probable error in the smoothed
data. The densities reported here were compared with
densities found in the literature, and it was concluded that
the final interpolated values determined in this work were
probably correct to ± 0 .0 0 0 1 5 . The best value for the
specific gravity ( 6 0 ° / 6 0 ° F.) of pure 1,3-butadiene was
estimated to be 0 .6 2 7 4 ± 0 .0 0 0 1 5 .
rren t in terest in 1,3-butadiene as a sy nthetic rubber
T HrawE mcuaterial
has created a need for m ore accurate and more
com plete inform ation on the physical properties of this hydro­
carbon.
L andolt and B ornstein {2) give a table of liquid densities for 1,3butadiene for the tem perature range —20° to 20° C. ( —4° to
6S° F.), Prevost (4) has reported a value a t 6.22° C. (21.2° F.),
and Doss (1) lists a value a t 68° F . Because these d a ta cover a
range of tem perature too sm all to m eet m ost requirem ents, it
seemed advisable to check previous values an d extend th e tem ­
perature range. T his paper reports experim entally determ ined
liquid specific gravities for 1,3-butadiene for th e tem perature
range -1 7 .7 8 ° to 20° C. (0° to 140° F .).
METHOD
T he procedure consisted essentially of com paring th e d eter­
m ined w eights of known volum es of b u tadiene under a num ber of
tem perature conditions and under pressures approxim ately equal
to the vapor pressures w ith th e w eights of identical volumes of
w ater a t 15.56° C. (60° F .) and a t atm ospheric pressure. T his
procedure has been employed previously b y th e w riters to arrive
a t sim ilar d a ta for propane, iso- and n-butane, th e various b u ty l­
enes, and a num ber of commercial products falling in the classifi­
cation of liquefied petroleum gases (S).
COM POSITION OF BUTADIENE USED
T he 1,3-butadiene used in th e investigation was obtained from
th e by-product butadiene p la n t of th e Phillips Petroleum Com­
pany and was representative of th e commercial p roduct of th e
p lan t a t th e tim e of sam pling. T h e sam ple was inhibited
against oxidation w ith 0.02 w eight per cent of phenvl-betanaphthylam ine. N o solvent for th e in h ib ito r w as used. T he
q u a n tity of added inhibitor was calculated to increase th e specific
g rav ity by no m ore th an 0.00005 and its presence in th e sam ple
could, therefore, be ignored.
T he com position of th e sam ple was ascertained by two different
analytical techniques, b o th based upon th e well-known chemical
reaction betw een maleic anhydride and 1,3-butadiene. Analyses
by th e tw o techniques gave a p u rity of 99.6 mole per cent. T he
im purities present were believed to consist of 1-butene and the
high- and low-boiling 2-butenes.
APPARATUS
T he apparatus consisted of tw o steel pycnom eters of approxi­
m ately 4500-cc. capacity fitted w ith expansion cham bers to fa­
cilitate m easurem ents a t tem peratures below room tem perature.
A constant-tem perature b ath, a centrifugal pum p for stirrin g th e
b ath liquid, a torsion balance w ith calibrated w eights, and a
calibrated mercury-in-glass therm om eter were also provided.
T he details of one of the pycnom eter u n its are shown in Figure
1, where A is th e pycnom eter and B is th e expansion cham ber.
C is a high-pressure stainless steel needle valve of such construc­
tion th a t when fully opened th e pressure of th e m aterial in the
cham bers is held by a m etal-to-m etal seat instead of by valve
packing, th u s reducing th e chance for errors due to leakage.
Valve D is a brass body steel needle valve. B oth pycnom eter
un its were tested before use w ith hydrogen gas a t 27-kg. (600
pounds) pressure to assure absolute freedom from leaks. T h e two
units were used sim ultaneously for check determ inations.
T he therm om eter used for the m easurem ent of b a th tem pera­
tures was graduated in 0.2° F . divisions. I t was checked before
use against a B ureau of S tandards calibrated m ercury-in-glass
therm om eter.
T h e torsion balance w as checked before use for accuracy, sensi­
tiv ity , stab ility , and equality of length of balance arm s. I t was
tested during use for reproducibility of w eights b y weighing an
iron w eight of a b o u t 9 kg. (20 pounds) a t various tim es during th e
day and on successive days. These tests indicated th a t a w eight
in th e desired range— i.e., 7.7 to 8.6 kg. (17 to 19 pounds)—
could be reproduced to 68 mg. (± 0.0015 pound). T h e set of
brass w eights used were calibrated to 9 mg. (± 0.0002 pound).
C a l ib r a tio n o f P y c n o m e te rs .
A lter being carefully cleaned
and dried, b o th internally and externally, th e tw o pycnom eter
units were evacuated an d th e ta re w eights determ ined and
checked b y repeated weighings to th e n earest 22 mg. (0.0005
pound). T h e volumes of th e pycnom eter cham bers were then
ascertained for tem peratures of 4.44°, 15.56°, 26.67°, 37.78°,
48.89° and 60° C. (40°, 60°, 80°, 100°, 120°, and 140° F .), and
w ith no internal pressure on the chambers, b y weighing th e w aterfilled cham bers a t th e various tem peratures and th en m aking cor­
rections for th e changing density of w ater. Freshly boiled dis­
tilled w ater was used. T h e effect of internal pressure on pyc­
nom eter cham ber volumes was ascertained for a tem perature
condition of 15.56° C. (60° F .) and pressures of 0,14, an d 28 kg.
p er sq. cm. (0, 200, and 400 pounds per square inch) gage, re­
spectively. T h e final results of the calibrations expressed in
term s of volum e for various conditions of tem perature and in­
tern al pressure, were p lo tted to arrive a t a sm ooth relationship
for use in th e subsequent experiments. I t is believed th a t the
finally assigned volum es for the
various conditions were know n to
± 0 .2 cc. for th e en tire tem perature
range.
MEASUREMENT OF DENSITIES
The p y c n o m e t e r u n i t s w e r e
evacuated to an absolute pressure
of less th a n 1 mm. of m ercury.
S u f f i c i e n t b u t a d i e n e was then
charged in to th e units to fill A com­
pletely and B to half its capacity.
Valve D was closed and C was left
open. T he u n its were th en placed
in a constant-tem perature b a th in
such a m anner th a t cell A was
to tally im mersed in th e b a th liquid,
b u t w ith no p a rt of cell B immersed.
H e a t was applied externally to B
b y m eans of an electrically heated
rem ovable jack e t to m aintain its
tem perature, 6° to 9° C. (10° to
15° F .) higher th a n th e b ath tem ­
perature. T h e tem perature of B
was m easured by a therm ocouple on
th e outside surface of th e cell a t a
p o in t below th e liquid level in th e
cell. In a prelim inary series of ob­
servations i t was determ ined th a t
approxim ately one hour was re­
quired to bring th e tem perature of
th e pycnom eter an d its charge of
butadiene to th e b a th tem perature.
T h e pycnom eters were consequently
held in th e co n stan t tem perature
b a th for 1.5 hours before being
8
INDUSTRIAL
Tabic I.
AND
ENGINEERING
Summary of Experimentally Determined Specific Gravities
of 1,3-Butadiene
T his W ork (99.6 M ole P e r C e n t 1,3-B utadicne)
T e m p e ra ­
P y cn o m eter
tu re
N o.
Specific g ra v ity
° F.
(i1/600 P .)°
L ite ra tu re ,
Specific G ra v ity
(t/6 0 ° F .)a
0.66671
0.66671
0 .6 6675
0 .6 6675
0 .6 6 5 9 (2)b
0.65414
0 .6 5 4 1 0
0 .6 6 4 0 (2)b
40
0.64097
0.64097
0 .6 4070
0.6 4065
0 .6 4 0 9 (£)&
60
0.6 2732
0.6 2725
0 .6 2 7 3 (2)b
20
21.2
0 .6 4 9 3 (4)
68
0.6 2 1 3 ( /)
80
0.61361
0.61371
100
0.59933
0 .5 9944
0.59944
0.59932
0.59937
0.59947
120
0 .5 8445
0 .5 8422
140
0.56920
0.56890
0.56850
0.56S70
<* Specific g ra v ity a t te m p e ra tu re t w ith reference to w a ter a t 60° F.
&In te rp o la te d values.
removed for weighing. D uring th a t tim e, th e b a th tem p era­
ture w as held constant to 0.06° C. (± 0 .1 ° F .). J u s t before
rem oving a pycnom eter from th e b ath , C was closed. Im m e­
diately upon removal, th e outside surface was dried and B was
evacuated. D was then closed, C was opened, and th e u n it was
weighed. In those instances where th e b a th tem perature was be­
low th e dew point tem perature of th e room air, the tem perature
of the pycnom eter was raised to above room tem peratures to
avoid errors in weighing occasioned by condensation of m oisture
on the surface. In such cases B served as a receiver for th e liquid
butadiene displaced from A .
RESULTS
T he specific gravities of th e butadiene were subsequent^’ ar­
rived a t by dividing th e determ ined w eights of the butadiene
contained in A a t the various tem peratures by th e w eight of
th e sam e volume of w ater when a t 60° F . an d atm ospheric pres­
sure. T he value for the density of w ater a t 60° F. used in these
calculations was tak en as 0.999017 gram p er cc. All experi­
m ental results are presented in T able I together w ith comparable
d a ta from Doss (1), L andolt an d B ernstein (2), and Prevost (/,).
An analysis of the m ethod used showed th a t errors in deter­
m ined gravities traceable to air buoyancy effects were of sm all
m agnitude. T he buoyancy correction, calculated to be
+0.00003, was n o t applied since it was too sm all in comparison
w ith the experim ental error to be significant. W ithin the experi­
m ental error it has been concluded th a t th e determ ined specific
gravity values can be accepted as equivalent to those taken in a
vacuum .
T he specific gravities presented in T able I were p lotted against
tem perature, and, from a sm ooth curve draw n through th e points,
th e specific gravity values for th e various interm ediate tem pera­
tures were determ ined. These sm oothed specific grav ity values
are presented in Table II . Of th e tw enty-tw o different specific
g rav ity m easurem ents m ade at. tem peratures of 48.89° C.
(120° F.) an d below, only those m ade a t 4.44° and 26.67° C. (40°
Table II.
im pcratu rc
° F.
0
2
4
6
8
10
12
14
16
18
20
22
24
26
28
30
32
34
36
38
40
42
44
46
48
50
52
54
56
58
60
62
64
66
68
70
CHEMISTRY
Vol. 16, No. 1
Smoothed Specific G ravity V alues for Com m ercially
Pure 1,3-Butadiene
(09.6 molo p e r c en t 1,3-butadicnc)
Specific
T e m p e ra ­
G ra v ity
tu re
(it/6 0 ° F .)a
0 F.
0.6668
0.6655
0.6643
0.6631
0 .6618
0.6606
0.6593
0.6580
0.6567
0.6554
0.6541
0.6529
0.6515
0.6502
0.6489
0.6476
0.6463
0 .6449
0 .6436
0.6423
0.6409
0.6396
0.6382
0.6369
0.6356
0.C342
0 .6328
0.6314
0 .6300
0.6286
0 .6 2 7 3 b
0.6259
0 .6245
0 .6232
0.6218
0.6205
72
74
76
78
80
82
84
86
88
90
92
94
96
9S
100
102
104
106
108
110
112
114
116
118
120
122
124
126
128
130
132
134
136
138
140
Specific
G ra v ity
(it/6 0 ° F.)'
0.6191
0.6177
0.6163
0.6149
0 .6135
0.6121
0.6106
0.6091
0.6077
0.6063
0.6049
0.6035
0.6021
0.6007
0.5993
0.5978
0.5963
0.5949
0 .5934
0.5919
0.5904
0.5S89
0.5874
0.5859
0.5844
0 .5828
0.5813
0.5798
0.5783
0.5767
0.5751
0.5736
0 .5720
0 .5704
0.5689
o Specific g ra v ity a t te m p e ra tu re t w ith reference to w a ter a t 60° F .
& P robable Bpccific g ra v ity (6 0 °/6 0 ° F .) for p u re 1,3-butadiene, a rriv ed
a t b y applying corrections for assum ed im p u ritie s, waa estim ated to be
0.027*1 * 0.00015.
and 80° F.) in pycnom eter u n it 2 differed from the sm oothed
value b y m ore th a n th e predicted probable am ount of 0.00018.
I t was believed th a t th e 0.4 mole p er cent of im purities present
in th e sam ple consisted of 1-butene and th e high- and low-boiling
2-butenes. B y m aking certain assum ptions, it was possible to
predict the specific grav ity a t 60° F. of pure 1,3-butadiene. T hus,
if it were assumed th a t th e three probable im purities were present
in su bstantially equal proportions, th e com puted value of pure
1,3-butadiene would be higher th an the determ ined specific
g rav ity of th e te st m aterial b y 0.00007. On adding th e buoy­
ancy correction of 0.00003 and sub tractin g th e correction of
0.00005 for th e am o u n t of inhib ito r present, th e final value
rounded off to 0.6274 was obtained for th e specific grav ity (6 0 °/
60° F.) of pure 1,3-butadiene. T h is was considered to be the
m ost nearly correct value for pure 1,3-butadiene a t 60° F.
ACKNOW LEDGMENT
T h e w riters wish to acknowledge th e helpful suggestions m ade
b y various m em bers of th e Phillips Petroleum C om pany R e­
search D e p artm en t in th e developm ent of th e experim ental
m ethod and, in particular, th e assistance rendered b y L. R. F ru it
on all experim ental m easurem ents. Acknowledgm ent is also
m ade b o th to Phillips Petroleum C om pany an d to th e B. F .
Goodrich C om pany for th e use of certain analytical techniques.
LITERATURE CITED
(1) D oss, “P hysical C on stan ts of th e Principal H ydrocarbons", 4 th
ed., p. 47, T exas Co., 1943.
(2) L andolt, H ans, and B ornstein, Richard, "Physikalisch-C hem ische
T ab ellen ” , 2nd Supplem ent, P art 1, p. 207, B erlin, Julius
Springer, 1931.
(3) N atural G asoline A ssoc. Am erica, I n d . E n o . C h e m ., 34, 1240-3
(1942).
(4) P revost, C ., Compt. Rend.., 186, 1209 (1928).
Physical Methods of A n alysis of Synthetic and Natural
Rubber
R. B O W L IN G B A R N E S , V A N Z A N D T W IL L IA M S , A . R. D A V I S , a n d P A U L G IE S E C K E
Stamford Research Laboratories, Am erican Cyanamid C o ., Stamford, Conn.
In a compounded rubber stock, the ratio of natural to synthetic rub­
ber can be estimated approximately from a knowledge of the phos­
phorus content of the rubber hydrocarbon. A considerably more
exact analysis can be carried out by means of infrared spectroscopic
methods, which permit a determination of the type as well as the
amount of rubber present. Complete details and comparative results
of these two methods of analysis are given, as well as a simple pro­
cedure for separating the rubber hydrocarbon of a rubber stock from
the carbon black and other com pounding ingredients.
th e am o u n t present. In conjunction w ith th e la tte r analysis,
a m ethod was devised for separating from com pounded rubber
products a sam ple of rubber hydrocarbon free of pigm ent and
filler. T he details of th e preparation and analysis of sam ples are
given, together w ith a typical set of d a ta obtained in connection
w ith the tire and tu b e studies.
A lthough these m ethods m ay no t necessarily be successful in
solving every type of ru b b er m ixture analysis encountered, th eir
value in this p articu lar problem w arrants careful consideration
in connection w ith sim ilar problems which m ay arise in th e future.
PO R T S have been m ade th a t m ore th a n 3000 kinds of
R Erubberlike
synthetics have already been prepared. Obviously,
DETERMINATION OF RATIO O F N A TU R A L TO SYNTHETIC RUBBER
BY PHOSPHORUS CONTENT
n o t all are of practical im portance, nor can they legitim ately be
The m etabolic processes of plants bring about, w ithin th e vari­
referred to as “synthetic rubbers” , b u t a certain select group of
ous p a rts of th e p lant, th e deposition of a g reat m any of the
these, as well as still other new synthetics, will survive and prove
m etals commonly found in th e soil. In con trast w ith this, the
m eritorious. Uses will be found for these new synthetic rubbers
m etals found in any sy nthetic p roduct are lim ited to those pur­
both alone and when blended w ith o th er synthetics or w ith n a t­
posely added an d those accidentally introduced by con tact w ith
ural rubber. T he rubber chemists of th e future will therefore be
the pieces of processing equipm ent.
faced w ith the problem of analyzing such m ixtures. Because th e
In order to determ ine w hether this difference m ight prove to be
com ponents of these m ixtures are often complex and th eir deg­
valuable
for analytical purposes, sam ples of n a tu ra l an d syn­
radation products are n o t well known, it will be difficult, by
thetic rubbers were subjected to an u ltraviolet spectrochem ical
conventional analytical m ethods, to establish th eir identities or
analysis, using a large Iiilg er E - l spectrograph. T able I shows
th e proportions in which they are present.
th e results of such a com parative emission analysis. I t m ay be
In anticipation of these difficulties, it was deemed advisable
seen a t once th a t, whereas th e synthetics contain little or no
to study th e applicability of various physical tools to this and
phosphorus, th is elem ent is present in th e n atu ra l rubbers in
other problem s of the rubber industry. In a previous publica­
readily detectable quantities. T his clue was followed up in
tion (I) the authors called a tte n tio n to th e fact th a t infrared
spectroscopy could be of value in differentiating between rubbers
great detail and th e exact values for th e phosphorus contents
and in analyzing rubber m ixtures, and pointed o u t th a t th e in­
were determ ined through th e use of an o th er spectrochem ical
frared absorption spectra of th e v ari­
ous rubbers are unique and offer one
m eans of attacking th e analytical
Table I. Ultraviolet Spectrochemical A n a lysis of the Metal Content of Rubber
problem outlined. However, in this
Hydrocarbons
prelim inary paper th e analytical possi­
S y n th e tic R u b b e r------ »
,----------------------- N a tu ra l R u b b e r----------------------- *
bilities were merely noted w ith o u t any
R eD oB una S
Sm oked
C repe
T rea d
C arcass claim ed m estic
G erm an
tre a d
a tte m p t to reduce them to practice.
sh e e t
ru b b e r
sto ck
sto ck
tires
B una S B una S
stock
Shortly a fter th e completion of this
-P e r C en t A sh (6 .3 )
(0 .2 7 )
(0 .1 7 )
(3 .6 )
(3 2 .5 )
(2 3 .2 )
(1 .4 9 )
(2 .1 1 )
prelim inary study, th e authors were
232
A lum inum
2222+
24called upon by th e W ar Production
A ntim o n y
2
0
0
0
1
0
0
0
2
B oard to analyze a series of captured
B ariu m
23141+
14141422B oron
21
1
+
141
+
14Germ an tires and inner tubes for the
C adm ium
20
1
0
0
1
114Calcium
2
3 32
2242424A rm y Ordnance D epartm en t, an d to
22C hrom ium
1
1
0
14141.4determ ine, if possible, th e am ounts
22222C o p p er
1+
141422
Iro n
2
2
2
2
22
and types of synthetic rubbers which
2
Lead
1
21
1414141433 M agnesium
332+
24242+
had been blended w ith n atu ral rubber
2
1
M anganese
1
1
1
14141+
in the m anufacture of these tires. As
20
0
0
M oly b d en u m
0
0
0
0
2N ickel
1
0
1
14141414a result of this investigation, two
2
2
2
0
0
Pho sp h o ru s
142+
24222
2
0
P o tassiu m
0
0
14satisfactory m ethods of analysis were
2
3 Silicon
2
2
2
242424developed.
32
2
2
2
2
Sodium
24242
1
S tro n tiu m
1
0
14141+
1+
T he first, the determ ination of the
1
1
T in
1
1
1
1
1
142
2
1
T itan iu m
1
1
1
141+
phosphorus content, furnishes a sim­
20
0
0
0
0
V anadium
0
0
2332
223
ple m ethod for m easuring the rela­
Zinc
24R an g es for q u a lita tiv e estim ates:
tive am ounts of n atu ral and synthetic
2100 to 1.0 p.p .m .
3
»
100
to 1.0%
rubbers present in an unknow n; the
10 to 0 .1 0 p.p.m .
3 10 to 0 .1 0 %
141
1 .0 to 0 .0 1 p.p.m .
2 4- 1 .0
to 0 .0 1 %
second, the application of infrared
1
less
th a n 0 .1 p.p .m .
2
«
0 .1
to 0 .0 0 1 %
m etal n o t d etected
spectroscopy, perm its determ ination
0
of both the type of each rubber and
9
10
INDUSTRIAL
AND
ENGINEERING
tool, the visible light spectrophotom eter. T he details of the
exact procedure followed are given below.
T able I I gives the phosphorus contents of a v ariety of n atu ral
and synthetic rubbers. These values show a considerable v aria­
tion for n atu ral rubbers of different origins, b u t all m ay be char­
acterized by high phosphorus (average 400 p.p.m .) -when com­
pared w ith typical synthetics (average 20 p.p.m .). T hus, an
exact phosphorus determ ination should m ake possible a d eter­
m ination of the ratio of n atu ra l to syn th etic rubber in an unknow n
sample.
Tabic 1!.
Phosphorus
P .p .m .
N a tu ra l ru b b e r
Sm oked sh ee t A
Sm oked sh ee t B
Sm oked sh eet C
S m oked sh e e t D
Sm oked sh ee t E
C repe
P a le crepe
R eclaim ed tu b e
G u a y u le
690
380
320
500
400
490
350
390
200
P .p .m .
S y n th e tic ru b b e r
B u n a S (A m erican ;
B u n a A (G erm an ;
C o tto n tir e cord
Viscose tir e cord
15
30
210
10
CHEMISTRY
Vol. 16, No. I
versus sy nthetic m aterials m ay arise in which this general
m ethod m ight be of help.] Therefore, it is advisable th a t all
cords, if present, be rem oved b y th e following procedure: A
section of th e rubber (tire) is shredded in a two-roll mill, then
mixed w ith w ater in a W aring Blendor. A fter 5 to 10 m inutes’
stirring, the cord can be partially separated by d écantation. T he
rem aining cord is digested a t a b o u t 4° C. for 18 hours w ith an
excess of cupram m onium solution [see Clibbens and Gaeke (5)
and the com m ittee of th e A m e r i c a n C h e m i c a l S o c i e t y (4) for
preparation of th is solution]. T h e m aterial is w ashed free of
cupram m onium w ith w ater and then dried. If th e sam ple is a
prepared rubber stock, all phosphate-containing plasticizers
m ust be rem oved b y refluxing th e shredded rubber for 8 hours in
a Soxhlet extractor w ith a solvent composed of 68% chloroform
and 32% acetone.
F o r m axim um photom etric accuracy, a w eight of rubber
should be taken such th a t a phosphorus con ten t up to 0.080 mg.
is contained. W ith syn th etic rubbers, a 1-gram sam ple is suit­
able, whereas, w ith n atu ra l rubber, approxim ately 0.1 gram is
used. In order to ensure a rep resentative sam ple, a larger weight
is tak en and subjected to the prelim inary preparation. The
color value of a 1-gram eq uivalent is then m easured to furnish a
rough idea of th e phosphorus content. W ith th is as a basis, a
w eight of sam ple is chosen according to th e above criterion and is
m easured accurately.
Figure 1 was plotted in order to show th a t th e phosphorus
could be determ ined accurately enough to allow th e content of
tlxis elem ent to be used as a y ardstick and as a direct m easure of
th e am ount of n atu ral rubb er present. (If th e origin of th e
n a tu ral rubber is unknown, th e slope of this curve is n o t determ in­
able, and the m ethod is only roughly q u an titativ e. On th e other
hand, if a sample of the n atu ra l rubber used in compounding is
available for a phosphorus test, a m uch higher degree of accuracy
m ay be expected.) In spite of the variation of the phosphorus
content of n atu ral rubbers of different origin, a reasonable first
guess a t a m ixture of unknow n constitution m ay be based upon
th e following generalization:
Figure 1.
H igh pho sp h o ru s (300 to 450 p .p.m .)
M edium phosphorus (100 to 250 p.p.m .)
Low phosphorus (0 to 50 p.p.m .)
n a tu ra l
n a tu ra l 4- sy n th etic
all sy n th etic
As will be seen from T able IV, th e above m ethod provides a
very simple approxim ate analysis. T he agreem ent between this
an d the m ore exact infrared m ethod is indeed surprising. How­
ever, th e m ethod is alw ays open to th e u ncertainty th a t phos­
phorus m ay have been introduced in processing and n o t been
elim inated in the sam pling procedure.
A lthough any sensitive m ethod for determ ining th e phosphorus
will be satisfactory, th e spectrochem ical procedure detailed below
was found to be m ost suitable. T his colorimetric m ethod is based
upon m any literature references of which Zinzadze (10) and
Goodloe (6) are typical. An intense blue color is produced in
solutions containing phosphorus b y reduction of phosphom olybdate, the intensity of th e color being directly proportional to th e
phosphorus content of the solutions. In th is laboratory, th e color
values or per cent transm issions were m easured on a General
Electric recording spectrophotom eter. T he blue color which
develops has a rath e r broad spectral w idth -which can be m eas­
ured satisfactorily by m eans of an y ordinary comparison color­
im eter, although the increased precision of th e m ore accurate
photoelectric instrum ents is to be preferred. Figure 2 shows a
series of transm ission curves for stan d ard phosphorus solutions.
P r e p a r a t i o n o f S a m p l e . All glassware and reagents m u st be
free of phosphorus and a suitable blank or control m u st be carried
through the complete procedure.
O ther possible sources of extraneous phosphorus are th e cord
and plastieizer used in fabrication. [It is interesting in this
connection to note (Table II ) th a t cotton, o r n atu ral, fibers are
high in phosphorus, whereas sy nthetic fibers, rayons, etc., have
a low content, th u s enabling a distinction to be m ade between
these two types of fibers. U ndoubtedly o th er cases of n atu ral
Phosphorus in Carcass Stock Com pounded
with Natural Rubber and Buna S
Formula oF Stock
Rubber hydrocarbon
100
50
Zinc oxide
Stearic acid
1
Pine tar
1.5
1.5
Agerite retln
Sulfur
2.0
Altax
1.25
Preferably, the phosphorus values should be calculated to the
w eight of ru b b er hydrocarbon, th e carbon, zinc oxide fillers, etc.,
having been rem oved previously. T his was n o t alw ays done in
th e au th o rs’ work, a s th e carbon analysis w as done elsewhere;
hence, the phosphorus con ten t of th e sam ples reported here was
based on the entire ru b b er sample.
A s h i n g o f S a m p l e . T he weighed sam ple of extracted rubber
is ashed in a suitable crucible in a controlled furnace a t a tem ­
perature n o t higher th an 600° C. u n til free of organic m aterial.
T his m ust be done slowly enough to p rev en t th e m aterial from
kindling. I t was n o t found necessary to add a phosphorus fixa­
tive in th e case of ru b b er m aterials, as no increased phosphorus
values were obtained by adding zinc oxide to prevent th e vola­
tilization of phosphorus. A fter cooling, th e residue is dissolved
by boiling in dilute sulfuric acid in a q u a n tity ju s t sufficient to dis­
solve th e soluble m aterials. Excess acid is to be avoided, since
th e final acid ity is very im portant, b u t an excess of acid m ay be
added an d la te r neutralized. A slight tu rb id ity w hich m ay be
rem oved b y filtering rem ains in some cases as a result of the
presence of a siliceous m aterial. T h e clear solution of th e ashes
is transferred to a 50-ml. Pyrex volum etric flask an d th e volume
is ad ju sted to a b o u t 40 ml.
D e v e l o p m e n t o f C o l o r . A lthough m any procedures are de­
scribed in th e literatu re, th e following modification was used
here.
Standard Solidions. Solution I, am m onium m olybdate, 5.45
gram s [(N H 4)5M o j0 2(.4H20 ], dissolved b y w arm ing w ith w ater
and m ade to 100 ml.
Solution II , lOiV sulfuric acid (282 ml. of concentrated sulfuric
acid diluted to 1000 ml.), checked b y titratio n .
January 15, 1944
ANALYTICA L EDITION
Solution I I I , stannous chloride. A stock solution of stannous
chloride dehydrate is 40 gram s dissolved in concentrated h y ­
drochloric acid (density 1.18) and m ade up to 100 ml. (stable
for m onths). F o r use, it is diluted 200 tim es w ith distilled w ater
(stable for one day).
Analytical Procedure. To th e solution prepared according to
instructions above, 2.5 ml. of Solution I and 5 ml. of Solution I I
are added and mixed thoroughly, and 2 ml. of Solution I I I are
added w ith constant swirling. T h e solution is m ade to volum e
(50 ml.) and m ixed thoroughly. (T he final acidity of th e 50-ml.
sam ple should be 1 N sulfuric acid. If excess acid is used in dis­
solving the ash, it m ay be neutralized a t th a t p o in t or th e am o u n t
of Solution I I m ay be reduced to give th e final acidity stipulated.)
A blank is carried through the sam e procedure to check for th e
presence of phosphorus in the reagents.
D e te r m in a tio n o f P h o sp h o ru s C o n te n t.
T h e per cent
transm ission of the solution in a 1-cm. thick cell a t 700 mu is
m easured exactly 20 m inutes after adding Solution I I I . T he
am ount of phosphorus in an unknow n m ay be determ ined from
a calibration curve which shows th e per cent transm ission plotted
as ordinates versus th e phosphorus con ten t of known stan d ard
solutions as abscissas. T he calibration curve m ay be prepared
from a scries of transm ission curves such as those of Figure 2.
N o t e s o n P h o s p h o r u s A n a l y s i s . W hen a m olybdate is
added to a solution containing orthophosphate according to th e
m ethod described above, an insoluble phosphom olybdate is
formed. However, because of th e high dilution, a p recipitate is
n o t apparent. T h e addition of a reducing agent (stannous chlo­
ride) causes a reduction of th e phosphom olybdate and gives an
intense blue color. U nder th e proper conditions, th e m olyb­
denum reagent necessary as an excess is n o t reduced.
T he factors which affect th e color are:
1. T he acid concentration is very im p o rtan t. In th e presence
of too m uch acid, a light color is produced. Conversely, if too
little acid is used, the m olybdate reagent itself will be reduced,
causing dark colors. In th e m ethod used, a 20% decrease in
acidity results in a 10% increase in color intensity, while a 10%
increase in acidity results in a 10% decrease in color intensity.
Soluble silica also produces a blue color w ith th e above reagent
if the silica concentration is high or th e acidity is insufficient to
suppress th e ionization of th e silicic acid. T he acidity chosen as
optim um in this investigation ( I N sulfuric acid) is such th a t silica
up to 2000 p.p.m . does n o t interfere.
2. T he m olybdate concentration is th e n ex t m ost critical
factor. An _increased m olybdate concentration results in a
higher sen sitivity to phosphorus, b u t also an increased blank. A
high blank reading is not desirable when low concentrations of
phosphorus are present. A 50% increase in th e m olybdate con­
centration, as used above, results in approxim ately a 15% in­
crease in the color intensity. A 20% decrease in m olybdate con­
centration results in approxim ately a 10% decrease in color in­
tensity.
3. T he stannous chloride concentration is n o t very critical.
A ± 5 0 % change in the am o u n t of stannous chloride influences the
phosphorus result n o t m ore th an ± 5 % .
4. U nder the given conditions, th e color increases for 5 to 10
m inutes after the stannous chloride is added and bleaches slowly
thereafter. I t is recom m ended therefore th a t th e color be m eas­
ured 20 m inutes after addition of th e stannous chloride.
5. T he following sources of interference have been considered,
and m ethods of reduction or elim ination are recom m ended where
necessary:
Ferric ion up to 6 p.p.m . does no harm . F ifteen p arts per
million slightly in h ib it color developm ent, while larger am ounts
of ferric ion cause very rapid fading of the color. F errous ion
causes no harm . A Jones rcductor w ith m etallic cadm ium gives
best results for prevention of interference from ferric ion (7).
T he presence of m ore th an 20 p.p.m . of titan iu m causes in­
terference by retarding the ra te of color developm ent.
A rsenates give th e same color as phosphates and th e intensities
are inversely proportional to the molecular weights. R eduction
w ith sodium bisulfite elim inates th e influence of arsenates (20
mg. of arsenic pentoxide per 50 ml. m ay be taken care of b y re­
duction to arsenic trioxide).
N itrates up to 100 p.p.m . have no effect; 200 p.p.m . reduce
the color ab o u t 10%.
Sulfates in large am ounts interfere, presum ably by depressing
the ionization of the sulfuric acid.
T arta ric and citric acids interfere above 20 p.p.m ., w ith in­
hibition of m axim um color. T h ey m ay be rem oved b y oxidation
witli perm anganate.
Aluminum an d m anganese in reasonable am ounts do n o t in ­
terfere.
Calcium an d m agnesium u p to 1000 p.p.m . have no effect.
N ickel up to 1000 tim es th e phosphorus co n ten t does n o t in­
terfere except for its own color.
11
Trichloroacetic acid begins to interfere w ith the m axim um color
developm ent a t concentrations above 4 % in th e final m ixture.
Acetic acid shows practically no effect.
H ydrochloric acid has a tendency to lessen color stab ility and
retard s or inhibits m axim um coloration.
DETERMINATION O F TYPES A N D AM OUNTS O F RUBBERS BY
INFRARED SPECTROSCOPY
R ecent publications (I, 8, 3, 8) have described in g reat detail
the m ethods an d applications of infrared spectroscopy to the
identification and analysis of m any ty p es of organic m aterials.
I t is sufficient here to point o u t th e tw o salient characteristics of
infrared absorption in order th a t th e basis for analysis can be
understood.
In the first place, th e infrared absorption spectrum (a p lot of
per cent transm ission as ordinate versus frequency in cm .-1 ) of a
m aterial is a unique characteristic of th e m aterial an d cannot be
duplicated by an o th er compound. Some of th e absorption
bands can be ascribed to particular atom ic groups w ithin the
molecule while others, characteristic of th e molecule as a whole,
are particularly useful for such studies as differentiating isomers.
T hus it is to be expected th a t th e phenyl group in B una S would
give rise to absorption bands which would n o t be present in
n atu ral rubber, while conversely, th e m ethyl groups of n a tu ral
rubber would cause a characteristic absorption which would not
be observed in B una S. Moreover, it is to be expected th a t there
will be fu rth er bands characteristic of th e molecule as a whole
which will assist in th e differentiation.
X IN ANGSTROMS
Figure 2. Spectrophotometric Determination” of2Phosphorus
Calibration of phoiphorui at reduced photphomolybdate
Final toluUon 1N In sulfuric acid, MoOi 0.13 gram par 50.ml., SnCli.SHiO
0.006 gram par 50 ml.
PhotNet
T
- lo g 7"
—log T
phorus
—log T
Blank
%
94.5
73
58
34
12.5
0.024
0.137
0.237
0.469
0.903
0.024
0.024
0.024
0.024
0.113
0.213
0.445
0.879
M icro ­
g ra m *
M icrog ram
Nona
8
15
30
60
0.0141
0.0142
0.0148
0.0146
12
INDUSTRIAL
AND
ENGINEERING
In the second place, so long as no interm olecular action occurs,
th e spectrum of a m ixture of rubbers will be simply th e spectra of
th e pure com ponents combined in the proportion in which the
m aterials them selves are present. Hence, it is to be expected
th a t an unknow n m ixture can be analyzed by direct m easure­
m ent of th e strength of absorption bands unique to each compo­
n ent, or by comparison of th e absorption spectra of th e unknow n
w ith those of a series of known prepared standards.
1800
1600
1400
1200
1000
800
FREQUENCY in cm : 1
Figure 3.
CHEMISTRY
Vol. 16, No. 1
this series by a comparison of the absorption intensities
1500 cm.-1
, 917 cm.“ 1
■orin „
and r,„r r . B y this m ethod, an unknow n such as
1380 cm. 1
83o cm. 1
J
th a t shown in Figure 4 could be estim ated to be 85% n atu ral
rubber, 15% B una S. All absorption spectra were tak en w ith
sam ples sm eared on salt plates and a high resolution spectrom ­
eter (2) was employed. T h e accuracy of th e m ethod is limited
to 5 to 10% because of th e inability to m ake sm ear sam ples of
1800
1600
1400
1200
FREQUENCY in
1000
800
cm:'
Infrared A bso rp tion Spectra of Rubber Hydrocarbons
T he infrared absorption spectra of various common types of
rubber are shown in Figure 3 (I). T he differences in these spec­
tra readily show th eir applicability in identifying an unknown
rubber or in estim ating the content of a m ixture. In th e course
of this investigation, only m ixtures of na tu ral ru b b er and B una S
were encountered. A consideration of these tw o spectra in Figure
3 or Figure 4 shows several m ajo r points of difference— B una S
has a strong arom atic ring frequency n ear 1500 cm.-1 and two
strong bands a t 970 cm .-1 and 917 cm .-1 which are n o t present in
th e n atu ral rubber. N atu ra l rubber, on th e other hand, shows a
m ethyl band a t 1380 cm.-1 an d a strong band a t 835 cm .-1
which are n o t observed in B una S. T aking advantage of these
differences, each unknow n sam ple was analyzed b y comparison
of its spectrum w ith the spectra of a series of known m ixtures
com pounded and treated in approxim ately th e same w ay as th e
unknow n m aterials. T he m ethod is outlined roughly in Figure 4,
showing the absorption spectra of n atu ral rubber, 50% n a tu ral
and 50% B una S, pure B una S, an d an unknow n tire tread. I t
can be seen from the 50-50 m ixture th a t the unique bands m en­
tioned above are all present w ith a stren g th proportional to the
com ponent concentration.
In the actual analysis, a g reat m any interm ediate stan d ard
spectra were obtained. An unknow n w as m atched w ith one of
constant thickness. G reater accuracy, if desired, could be ob­
tained by m aking per cent transm ission m easurem ents a t the
chosen frequencies according to some of th e m ore involved
analytical m ethods referred to above {2, 8, 8). In order to justify
th e careful treatm en t, it would be necessary to stu d y th e un­
knowns and standards as solutions of known concentration in a
suitable solvent which would tran sm it infrared radiation a t the
analytical frequencies chosen. T his more careful trea tm en t
would still require a prelim inary complete absorption spectrum of
each unknown, in order to furnish a qualitativ e analysis for the
various ru b b er m aterials which are present.
S a m p le P r e p a r a t i o n .
T h e m ethods described above are
stan d ard procedures and are given in detail in th e references.
Considerable time, however, had to be devoted to finding a m eans
of converting a piece of tire stock to a sam ple whose infrared
absorption spectrum could be m easured. All a tte m p ts to obtain
th e spectra of th e tire stocks directly, either b y m icrotom ing a
th in section o r by a sm ear from a solution of the stock m aterial,
were unsuccessful. In order to o b tain a satisfactory spectrum ,
it is necessary to e x tract th e ru b b er hydrocarbon free of plasticizer, filler, etc. T h e m ethod finally devised was com pletely satis­
factory for th e au th o rs’ purpose and should be of g reat value in
January 15, 1944
ANALYTICAL
any situation whore it is desirable to separate filler and rubber
w ithout destruction of the rubber hydrocarbon. Therefore, this
separation by m eans of p-cym ene and xylene solution is com­
pletely described below.
A fter this separation, the rubber is present as a solution in the
hydrocarbon solvents. These solvents are rem oved by vacuum
distillation until the rubber is a gum m y solid th a t will n o t flow
a t room tem perature. T his gum is th en w ashed four or five tim es
w ith h o t acetone to remove th e last traces of plastieizcr and
solvent.
In m ost cases the state of th e rubber is such th a t it can easily
be sm eared on a rock salt plate. T he desired thickness is ob­
tained by scraping the film w ith a razor blade u n til th e 1450
cm .-1 C II band shows a transm ission of a b o u t 10%. I f th e ru b ­
ber does n o t spread easily, it m ay be softened w ith a volatile
solvent such as carbon tetrachloride. T h e film is th en dried in
a vacuum oven to remove th e last traces of solvent. Finally,
th e absorption spectrum of th e sam ple is obtained throughout
th e spectral region of interest.
EDITION
showed clearing in a few m inutes an d could be filtered free of the
carbon black a fte r stan d in g 10 m inutes. T his m ethod was fu rther
altered slightly by using 250 cc. of n-hexane instead of 300 cc. of
70° B6. gasoline for diluting th e solution. Again th e carbon black
separated easily. In order to determ ine w hether or n o t th e sam e
dilution technique would work w ith xylene as th e solvent, a
solution of th e sam e tread stock was m ade w ith 30 cc. of xylene
plus a sm all am o u n t of thio-jS-naphthol, b u t on diluting w ith
benzene and hexane, th e carbon black would n o t separate from
th e solution.
F u rth e r experim ents showed th a t th e combined use of pcymene and xylene gave th e m ost successful results and led to a
satisfactory m ethod.
M e t h o d I I (adopted for solution of ru b b er an d separation of
pigm ents). Sheet th e sam ple to a thickness of approxim ately
0.375 cm. (0.15 inch) on a tig h t cold 15 X 30 cm. (6 X 12 inch)
lab oratory mill.
E x tra c t the plasticizer, etc., from 4 gram s of th e sheeted sam ple
w ith a m ixture of 32% b y volum e of acetone and 68% by volum e
of chloroform for a m inim um of 7 hours, or until the extracting
liquid no longer shows color. T h is size of sam ple is sufficient to
perm it rep eat analyses.
Place 1.0 gram of the dried, extracted sam ple in a 400-cc. rub­
ber extraction flask w ith 25 cc. of p-cym ene and 5 cc. of xylene.
H e a t on a steam b a th a t ab o u t 70° to 80° C. for 1 hour, then on a
DISSOLVING RUBBER HyD RO CARBO N A N D SEPARATING IT FROM
COM POUNDING INGREDIENTS
I n the case of vulcanized n a tu ral rubber compositions, various
solvent m ethods of determ ining th e am o u n t of ru b b er hydrocar­
bon have been used w ith some success. These m ethods, however,
are som ew hat involved and time-consuming. E ven th e A.S.T.M .
m ethod requires th a t the solution stan d overnight to allow the
m ineral fillers and pigm ents to settle before filtration. I f the
rubber composition contains highly dispersed, exceedingly fine
pigm ents, such as carbon black, separation of th e pigm ents by
filtration or centrifuging is usually impossible.
Since m any present-day ru b b er products, p articularly tire
tread compositions or compounds, contain large am ounts of
carbon black, none of th e above-m entioned m ethods appeared
prom ising for separation of th e rubber from the carbon black in
suitable condition for infrared analysis. Accordingly, research
on this question was initiated.
Fearing th a t destruction o r decomposition of th e ru b b er hydro­
carbons m ight result from th e use of th e conventional high-boiling
solvents used in the above separation m ethods, a num ber of lowboiling solvents were tried: ethylene dichloride, toluene plus
piperidine, o-nitroanisole, D ispersing Oil N o. 10 (B arrett) plus
xylene, and C ircolight Process Oil (Sun Oil Co.) plus xylene.
These experim ents were unsuccessful.
M ethod I. In the next attem p ts, which were som ew hat more
successful, xylene plus a sm all am ount of thio-/3-naphthol boil­
ing under reflux was found to dissolve n a tu ra l or syn th etic tire
tread stocks in 5 to 6 hours. T o accom plish this, 1 gram of chloro­
form -acetone-extracted tread stock was heated in 100 cc. of
xylene plus 0.3 gram of thio-/3-naphthol a t 140° C., u n til solu­
tion was complete.
However, this solution did n o t perm it easy separation of the
carbon black unless a com bination of centrifuging and slight
alcohol precipitation was used. (By adding alcohol to th e solu­
tion, a sm all am ount of rubber is p recipitated in order to drag
dow n the carbon black during centrifuging. T here is some danger
th a t this precipitation m ay be selective and thus change the
composition of the residual rubber.) A sufficient separation of
th e carbon black can be m ade in this m anner to perm it fairly
satisfactory infrared analysis. In the case of carcass stocks con­
taining only zinc oxide as a filler, the pigm ent was easily separated
b y settling or centrifuging. T h is m ethod w as n o t completely
satisfactory, however, because of carbon black troubles and the
difficulty of rem oving th e peptizer w ith o u t loss of th e rubber
itself.
T h e search for a b e tte r m eans of carbon black separation was
therefore continued. T he cresol m ethod (9) was tried w ith some
success, although there was considerable oxidation of th e rubber
a t the high reflux and distillation tem perature required. T h e
use of p-cym ene in place of the oresol w ith digestion a t ab ou t 160°
to 170° C. gave a satisfactory solution in ab o u t 4 hours and caused
no noticeable oxidation of th e rubber. On dilution w ith th e ben­
zene and 70° B6. rubber solvent gasoline, th e su p e rn a ta n t liquid
13
Figure 4. O u tline of Infrared A n a ly tica l M ethod
Spectra of prepared standards and unknown sample
14
INDUSTRIAL
AND
ENGINEERING
BSELKSa CORD
CARGOES CORD A. STOCK
Figure 5. First Step in Sam pling Procedure
Labeled cross section of tire to be analyzed for rubber content
Sm oked sheets
Buna S
Zinc oxide
C arb o n black
S tearic acid
P in e ta r
T herm oflex A
N eozone D
B ard o l
S ulfur
C ap ta x
A lta i
A gerite resin D
50.
50.
50.
1.
1 .5
2.
---------S tocks------A-912
ACKNOW LEDGM ENT
In conclusion, th e authors would like to ex­
press th eir appreciation to th e several o th er m em bers of the
staff of th is lab oratory who helped m aterially in th e establish­
m e n t of these analytical m ethods.
Com position of Typical Rubber Stocks
87-1
Vol. 16, No. 1
cushion, breaker, and carcass stocks were re­
m oved from each tire, extracted, prepared for
study, an d th en subjected to th e tw o an a­
lytical procedures.
In T able IV, th e results obtained on a few
typical G erm an tires an d inner tubes are p re­
sented. In view of th e spread in th e phos­
phorus contents of n atu ral rubbers from v ari­
ous sources (Table II), a m ean value of 400
p.p.m . was used in m aking th e analytical
calculations from phosphorus determ inations.
T he agreem ent betw een th e infrared an d the
phosphorus m ethods is therefore all th e more
gratifying.
SPEAKER ł CUSHION STOCK
Table III.
CHEMISTRY
A l-141
100.
5.
50.
2 .5
1 .5
1 .4
0 .6
3.
1 .2
100.
5.
50.
Table IV .
T ire N o.
5.
2.
1 .5
1 .2 5
1 .5
h o t plate under reflux a t 150° to 160° C. u n til th e ru b b er stock is
com pletely dissolved. T hree to 4 hours are usually sufficient.
Cool and dilute w ith 20 cc. of benzene an d then w ith 150 cc.
of hexane. Allow th e carbon black an d oth er pigm ents to settle
for a t least 10 m inutes, then decant th e clear su p e rn a ta n t liquid
through a No. 1 filter paper. If desired, filtration m ay be done
w ith suction, using an asbestos pad.
W ash the flask, pigm ents, and filter w ith a m ixture of 5 cc. of
benzene and 45 cc. of hexane.
T he above treatm en t easily removes th e carbon black and
o th er opaque compounding ingredients, giving a clear solution of
th e ru b b er from which th e solvent m ay be rem oved b y vacuum
distillation.
In order to m ake valid comparisons between know n an d un­
known samples, m any m ixtures of n atu ral and B una S ru b b er in
v arying proportions were com pounded according to th e typical
m ethods of preparing tread, carcass, and tu b e stocks. T he ru b b er
content of these known sam ples w as th en extracted according to
th e final m ethod discussed above and was used for infrared com­
parison. T he compounding ingredients used in three such typical
stan d ard s are given in T able I I I . T he infrared absorption spec­
tru m of th e 50-50 m ixture is shown in Figure 4.
TYPICAL A N A LY T IC A L RESULTS O N A C TU A L TIRE A N D TUBE STOCKS
T his entire project was initiated in order to provide a suitable
m ethod for determ ining the composition of tires and tubes, w ith
particular reference to the types an d relative am ounts of the
various rubbers present. T he analytical m ethods which were
evolved were applied successfully to a large num ber of samples.
T he ty p e of inform ation obtained is well illu strated b y th e results
shown below, which were selected a t random from th e samples
examined.
T h e tires, w hen received, were cross-sectioned, photographed,
a n d labeled as shown in Figure 5, th u s ensuring a perm anent
record of th e tires and a ready m eans of identifying th e various
sam ples chosen. W herever possible, sam ples of th e tread,
A n a ly sis of Captured German Tires
P hosphorus
N a tu ra l
P .p .m .
%
In fra re d
N a tu ra l
Buna S
%
%
T rea d S tocks
FT
61
99
100
101
102
103
104
105
106
107
280
30
20
10
10
30
25
20
27
3
FT
61
99
100
101
102
103
104
105
106
107
340
340
380
430
130
190
230
200
305
190
FT
61«
99«
101«
102&
103&
104&
105*
106&
107 &
290
270
450
200
255
285
295
260
290
75
0
0
0
0
0
0
0
0
0
100
0
0
0
0
0
0
0
0
0
0
100
100
100
100
100
100
100
100
100
100
100
100
100
20
50
50
80
100
50
0
0
0
0
80
50
50
20
0
50
100
100
0
0
*75
75
85
100
100
100
*25
25
15
0
0
0
C arcass Stocks
100
100
100
100
25
50
60
50
80
50
C ushion an d T ubes
75
70
100
50
65
70
75
70
75
“ C ushion.
b T ubes.
LITERATURE CITED
(1) Barnes, R . B ., Liddel, Urner, and W illiam s, V an Zandt, I n d .
E n q . C h e m ., A n a l . E d ., 15, 8 3 - 9 0 (1 9 4 3 ).
(2) Ibid., p p . 6 5 9 -7 0 9 .
(3 ) B rattain , R . R ., R asm ussen, R . S., and C ravath, A . M ., J.
A p p lied P h y s., 14, 4 1 8 - 2 8 (1 9 4 3 ).
(4) Carver, E . K ., B ingham , E . C., Bradshaw , H ., and Venable,
C . S., I n d . E n g . C h e m ., A n a l . E d ., 1, 4 9 -5 1 (1 9 2 9 ).
(5) Clibbens, D . A ., and G aeke, Arthur, J . T extile In st., 19, 7 7 9 2 T (1 9 2 8 ).
(6) Goodloe, Paul, I n d . E n g . C h e m ., A n a l . E d ., 9 , 5 2 7 - 9 (1 9 3 7 ).
(7) K u ttn er, Theodore, and Cohen, H . R ., J . Biol. Chem., 75, 5 1 7 31 (1 9 2 7 ).
(8) N ielsen, J . R ., and Sm ith, D . C ., I n d . E n g . C h e m ., A n a l . E d ,,
15, 6 0 9 -1 5 (1 9 4 3 ).
(9) R oberts, J. B ., Jr., Rubber Chem. Tech., 14, 2 4 1 - 8 (1 9 4 1 ).
(10) Z in z a d z e , C h ., I n d . E n g . C h e m ., A n a l . E d ., 7, 2 2 7 - 3 0 (1 9 3 5 ).
Determination of Rate of Cure for Natural and
Synthetic Rubber
LEO N A R D H. C O H A N
and
M A R T IN S T E IN B E R G
Continental Carbon Com pany, Research Laboratories, 6 1 3 0 W est 51st St., Chicago, III.
The tensile strength at a given marked undercure divided b y the
maximum tensile appears to be a convenient index of rate of cure.
This index is called the tensile ratio. In comparing two stocks, the
one having the higher tensile ratio is the faster curing. The proper
undercure to use in calculating the tensile ratio has been found to be
in the range of roughly 6 0 % of the maximum tensile, which for many
stocks occurs at about one fourth the time to reach maximum tensile.
The tensile ratio is very easy to determine, requires no special equip­
ment, and can be determined fairly accurately. In addition, it
agrees well with generally accepted indexes of cure rate for typical
natural rubber and G R -S formulations.
H E m ethods th a t have been proposed for determ ining the
ra te of cure of n atu ral ru b b er stocks include:
T
Tim e to reach some m aximum physical p roperty, such as te n ­
sile strength, tensile product (12, 16), modulus a t a definite
strain, aged tensile, or some prop erty of practical im portance for
a given compound.
T ests depending on tem p eratu re susceptibility like T-50 (5,
6, 7 ,1 1 ,1 4 ) or zero degree set (S).
Chemical tests, such as determ ination of th e am ount of com ­
bined sulfur (12).
Special tests based on physical m easurem ents, such as tim e
to reach the break in the m odulus a t a given strain vs. cure-tim e
curve (4), tim e to reach th e cure giving “ reasonable snap w ith
substantially unim paired te a r” (optim um hand te a r m ethod) (2),
and tim e to reach m inim um reduced residual elongation (8).
R a te of cure m ay be defined as th e tim e required to reach a
given sta te of cure com pared to some stan d ard or control stock.
Specifically, ra te o f cure referred to a given sta te of cure equals
tim e to reach th a t sta te for the control divided by th e tim e for
th e sam ple to reach th e sam e state. In th e previous example,
if th e stan d ard stock reaches a sta te of cure of 0.94 based on re­
bound in 20 m inutes, th e ra te of cure of the sam ple tread stock
would be 20/30 = 0.67, since th e sample stock required 30
m inutes to reach th e 0.94 state.
O ptim um cure an d sta te of cure depend on th e physical prop­
erty used to determ ine curing characteristics. R ate of cure m ay
depend in addition on th e state of cure a t which th e rate is m eas­
ured. From a practical view point th e basic physical property
is th e useful life of a satisfactory commercial p roduct m ade from
the com pound in question. If we th en tak e useful life as the
physical pro p erty for determ ining sta te of cure and use a state
of cure of 1.0 (optim um cure) a t which to measure ra te of cure,
we arrive a t th e following definition of ra te of cure a t a given
CONTINENTAL A A IN NATURAL RUBBER
3500
Of these m ethods, th e T-50 te s t and th e tim e to reach m axi­
m um tensile have been m ost widely accepted and have proved
very useful, b u t leave som ething to be desired. T h e T-50 te s t
requires special equipm ent n o t alw ays available in sm aller rubber
laboratories, while the tim e to reach m axim um tensile, although
obtainable from tensile d ata w hich in m ost cases would have to
be determ ined anyhow in order to evaluate th e compound, is
difficult to determ ine accurately, as th e tensile vs. cure-tim e
curve usually has a flat m aximum .
In GR-S th e need for a convenient index of ra te of cure is more
urgent. T he T-50 te st is n o t applicable to GR-S compounds (9)
and as a rule the tensile vs. cure-tim e curve has an even flatter
m axim um th a n in rubber, so th a t th e accuracy of determ ining
the tim e to reach m axim um tensile is even lower.
I n considering the question of obtaining a satisfactory index
of ra te of cure, it ■null be helpful to define th e term s “optim um
cure” , and “sta te of cure” , “ra te of cure” .
O ptim um cure for a given physical property m ay be defined as
the tim e required to reach th e m axim um or optim um value for th a t
p roperty— for example, th e optim um cure w ith respect to re ­
bound would be th e cure tim e a t a specified cure tem perature to
bring th e stock to its m aximum rebound. T h e p ro p erty m ost
generally used is tensile stren g th and th e optim um cure for ten ­
sile stren g th is often referred to sim ply as th e optim um cure.
S tate of cure has been defined as th e position of th e cure in
question in a series of cures (16). I t seems more precise to de­
fine sta te of cure w ith respect to a given physical property as
th a t fraction of the m aximum value of th e property shown by
th e cure in question (16). F o r example, if th e 30-m inute cure
of a certain tread stock has a rebound of 47% and if th e m axi­
m um rebound for this tread stock is 50% , then th e sta te of cure
of the 30-m inute cure w ith respect to rebound can be expressed as
47/50 = 0.94. I t m ay be useful in some instances to distinguish
undercures from overcures; we m ay, therefore, indicate over­
cures by placing a negative sign before th e state of cure.
Figure 1
I3 %
■■64 S
40
50
CURE TIME
60
70
(MIN. AT 3 0 7' f )
Figure 2
60
16
INDUSTRIAL
AND
ENGINEERING
CHEMISTRY
Vol. 16, No. 1
cure tem perature— rate of cure for an y stock is th e cure tim e
required to reach m axim um useful life for a stan d ard control
stock divided b y th e cure tim e required to reach m axim um useful
life for th e sam ple stock.
Table 1.
Natural Rubber
F o rm u la
G um
P p td .
W h itin g
Sm oked sheet
Zinc oxide
P ine ta r
S tearic acid
Sulfur
M ercap to b en zo th iazple
P re c ip ita te d w h itin g
M edium th e rm a l (M T )
C o n tinex (S R F )
C o n tin e n tal AA (E P C )
C o n tin e n ta l D (M P C )
C o n tin e n ta l R-40 (CC)
100
7 .8 5
3
3 .3
2.8 1
100
7 .8 5
3
3 .3
2 .8 1
0 .7 4 3
0 .7 4 3
78
MT
100
7 .8 5
3
3 .3
2 .8 1
0 .7 4 3
C o n ti­
nex
100
7 .8 5
3
3 .3
2 .8 1
0 .7 4 3
W hile th e foregoing definition is precise, it is n o t very useful in
actual labo rato ry work. Some index of ra te of cure is therefore
desired which correlates closely w ith ra te of cure as defined above
b u t can be determ ined conveniently
and accurately. In th e absence of
________________
d a ta connecting any proposed index
of ra te of cure w ith actu a l useful life
tests it m ay be considered satisfac­
C o n ti­
C o n ti­
C o n ti­
to ry to shorv th a t th e proposed index
n e n ta l
n e n ta l
n e n ta l
D
R-40
AA
correlates closely w ith indexes of rate
100
100
100
of cure, th e usefulness of w'hich has
7 .8 5
7 .8 5
7 .8 5
3
3
3
already been established.
3 .3
2.81
3 .3
2.81
3 .3
2 .8 1
0.7 4 3
0 .7 4 3
0.7 4 3
50
50
50
50
50
C u re a t 280° F.
Min.
0
8
15
30
60
90
180
180
280
300
390
390
T ensile a t break , l b ./
sq. in.
0
8
15
30
60
90
180
2 4 i6 ‘
3580
3800
3490
2670
2450*
2400
2300
2175
2000
E lo n g atio n a t b reak ,
0
8
15
30
60
90
180
8ÖÖ*
765
755
695
680
135
145
185
M odulus _ a t
elongation,
in.
400%
lb ./sq .
%
B re ak in g set,
0
8
15
30
60
90
180
0
340
870
960
1100
900
780
300
720
1060
1250
1350
1320
900
1420
1920
2120
2200
2300
350
1070
1620
2100
2330
2675
250
720
1060
1430
2360
2630
180
370
770
1690
1800
2470
100
70
1620
2650
2700
2820
2550
2450
200
1950
2900
3200
3340
3200
2450
250
600
2400
3340
3750
3900
3750
370
750
2040
2660
3290
3880
3780
380
460
1200
2600
3300
3360
3170
975
900
6i0*
600
580
580
570
670*
600
585
550
550
750
600
600
550
540
450
450
860
775
600
600
600
565
530
975
800
650
640
630
570
535
1100
850
800
720
600
600
490
180
10
12
16
21
7
7
220
120
17
25
30
27
21
205
105
24
25
27
32
25
275
180
....
7
10
10
5
0
1.5
17
12
17
15
i.7*
17
22
15
11
*35*
27
30
17
0
8
15
30
60
90
180
i7 0 ‘
180
220
225
180
170
25
100
260
250
285
200
200
25
100
300
325
275
230
230
35
150
480
485
485
370
300
40
65
440
800
865
770
635
45
90
135
250
600
730
730
50
65
125
200
6S0
680
540
D u ro m c te r a t 25° C.
0
8
15
30
60
90
180
24
26
34
39
41
42
41
34
39
48
54
56
55
55
30
38
45
50
51
53
51
37
47
54
58
60
62
62
40
45
52
58
60
65
69
39
45
51
60
62
69
70
45
48
51
59
63
66
70
D u ro m c te r a t 100° C.
0
8
15
30
60
90
180
8
35
36
41
41
42
40
15
44
50
56
57
55
53
14
38
45
50
52
52
49
18
40
49
56
58
61
60
21
28
42
56
59
63
65
20
28
43
54
55
65
67
30
37
41
50
61
61
64
R ebound a t 25° C .,
% , B ashore
0
8
15
30
60
90
180
46
49
51
57
60
58
56
35
39
42
52
55
53
53
46
51
52
55
58
59
58
43
38
42
48
47
47
45
31
31
30
33
37
36
34
30
32
33
34
32
32
32
23
23
0
8
15
30
60
90
180
28
54
63
70
75
68
65
27
58
65
69
71
68
67
28
60
62
73
75
77
74
26
48
53
63
60
61
60
23
33
43
48
52
50
47
23
34
40
48
46
46
47
*33*
48
48
41
0
8
15
30
60
90
180
2 3 .0
5 .0
- 1 .0
-1 S .0
-4 6 .5
-4 7 .5
-4 7 .8
2 2 .5
7 .0
0 .0
-1 2 .0
-2 4 .0
-2 7 .0
-2 7 .0
2 2 .5
1 1 .5
4 .5
-6 .5
-2 2 .8
-2 6 .5
-2 6 .5
2 1 .0
1 2 .5
5 .5
-6 .5
— 1 8 .0
-2 2 .0
- 2 5 .5
2 1 .0
18 .5
15.3
3 .0
-7 .2
- 1 0 .0
-1 5 .8
1 9 .8
2 0 .2
18.0
9 .5
3 .3
-0 .5
-3 .5
T e a r, lb ./in .
R eb o u n d a t 100° C.,
% , B ashore
T -50. ° C.
2 2 .0
*i 2 *5*
1 .5
-9 .3
-1 3 .7
-1 9 .0
*27*
34
32
28
23
25
TENSILE RATIO
In order to determ ine which physi­
cal properties m ost closely fulfill
th e above criteria for a useful index
of ra te of cure, th e indexes suggested
by previous investigators and other
com binations of physical properties
m entioned below were examined in
n atu ral ru b b er and GR-S. T h e best
index was found to be th e tensile
ratio, which is defined as th e tensile
strength a t a m arked undercure
divided by th e m axim um tensile
stren g th . T he undercure best suited
for this purpose w’as found to be the
cure giving a tensile stren g th of
roughly 60% of th e maxim um tensile.
F o r m any tread type stocks th e de­
sired cure is roughly one fourth the
tim e to reach m aximum tensile. F or
example, for a series of stocks of this
ty p e which reach m axim um tensile
in ab o u t 60 m inutes—say between
30 and 90 m inutes—the proper
undercure for calculating th e tensile
ratio wrould be a b o u t a 15-minute
cure.
A very definite objection to the
use of th e tim e to reach some m axi­
m um p ro p erty such as maxim um
tensile as an index of ra te of cure
is th e inaccuracy involved in d eter­
m ining th is tim e when th e tensile vs.
cure-tim e curve has a flat m aximum .
T h e tensile ratio is, of course, not
subject to th is difficulty, as the
flatter th e m axim um in th e tensile
vs. cure-tim e curve, th e easier it is to
determ ine accurately th e value of
th e m axim um tensile an d hence the
tensile ratio.
In Figures 1 and 2 the effect of the
form of th e tensile m aximum on the
accuracy of determ ining the tim e to
reach m axim um tensile is illustrated.
In both figures th e solid line gives
th e tensile stren g th as determ ined
from th e average of four identical
stocks cured sim ultaneously a t the
indicated cures. F rom the devia­
tions of th e individual stocks from
th e average, th e probable error (0.67
tim es th e stan d ard deviation) a t each
cure was calculated an d th e tw o broken lines correspond to th e average
value m inus th e probable error and
th e average plus th e probable error.
January IS, 1944
ANALYTICAL
T hus the two broken lines represent th e 50% confidence lim its
for the tensile a t various cures. T he stock used for this purpose
in Figure 1 was C ontinental AA in the n atu ral rubber form ulation
given in Table I, while in Figure 2 C ontinental AA in th e GR-S
form ula given in T able I I was used. F or th e possible tensile vs.
cure-time curves th a t can be draw n w ithin th e broken lines, the
maxima can be anyw here in th e shaded areas, so th a t th e time
to reach maxim um tensile can vary from <miD. to G,». T h e ex­
trem e per cent variation is therefore, as shown on th e figures,
35% for the rubber stock and 64% for th e flatter G It-S curve.
From the same figures the extrem e per cent variation for th e te n ­
sile ratio can also be calculated as indicated on th e figures and is
10% for rubber an d 13% for G R-S. (These values are only
approxim ate and illustrative. Since th e experim ental errors
responsible for th e variation in th e individual curves are n o t en­
tirely independent, th e confidence limits for th e various calcula­
tions m ay be som ewhat different and the figures should n o t be
considered as indicating a precise comparison of the accuracy of
the tensile ratio and tim e to reach m axim um tensile.)
DETERMINATION OF RATE OF CURE OF N A TU RA L RUBBER
In order to determ ine how satisfactory an index of rate of cure
was furnished by the tensile ratio, various pigm ents were milled
in to a typical n atu ral rubber form ula, 26 volum es of pigm ent
Figure 3
EDITION
17
per hundred volum es of ru b b er being used in all cases. In Table
I th e formulas of th e stocks used an d th e experim ental test results
are given. T he results include tests on uncured stock an d on a
range of cures from 8 to 180 m inutes and perm it com parison of the
tensile ratio and various indexes of ra te of cure which have been
used or recom m ended for n atu ra l rubber. T he tim e to reach
m axim um tensile, tensile p roduct (12, 16) (tensile m ultiplied by
elongation), tear, and rebound were determ ined from th e d ata.
Likewise the tim e to reach m inim um reduced residual elongation
(breaking se t divided by tensile) (5), break in m odulus a t 400% '
elongation (4), and optim um cure as determ ined b y h and te a r
(2) were determ ined. In addition to tab u latin g th e tim e to reach
m axim um values th e tim e to reach various percentages (60, 70,
75, 80, 85, 90) of m axim um values w as determ ined for tensile
strength and also for th e o th er physicals.
We have defined sta te of cure as th e ratio of the p hysical
property a t a given cure divided b y the m axim um value of th e
physical property. I t was therefore th o u g ht of interest to see if
th e state of cure for an y given cure, such as 15 m inutes, could
offer a useful index of ra te of cure. F o r this purpose the ratioof tensile a t all undercures to m axim um tensile w as determ ined
and sim ilar ratio s for tensile product, rebound, m odulus, a n d
durom etcr. (In th e case of m odulus and durom eter the 180m inute cure was tak en as th e m axim um value.) Since th e re­
bound of th e uncured stock is appreciable, th e ratio of the in­
crease in rebound over uncured rebound a t all undercures to the
increase in rebound over uncured rebound for the m axim um re­
bound was also calculated.
Finally, it was thought, th e slope of th e curve for the various
physical properties vs. cure tim e a t m arked undercures m ig h t
furnish an index of cure. Therefore the difference in tensile for
various undercures— e.g., tensile a t 15 m inutes m inus tensile a t 8
m inutes— divided by m axim um tensile was tab u la te d and sim ilar
ratios for tensile product, tear, and rebound.
Inasm uch as the T-50 test is probably th e m ost generally ac­
cepted index of rate of cure for n atu ra l rubber, th e indexes
enum erated above were tested b y determ ining th eir correlation
w ith T-50. T he tensile ratio w hich for this series of stocks was
tak en as the tensile a t 15 m inutes divided by th e m axim um ten­
sile was found to give th e best correlation.
T he results are sum m arized in Figures 3 and 4, w here the ten ­
sile ratio an d oth er indexes of rate of cure are p lo tted against
T-50 a t th e 60-m inute cure. T h e o th er indexes are those found
to give th e closest correlation w ith T-50 or which have been pre­
viously suggested in th e literature.
T he prediction indexes (P .I.) indicating th e degree of correla­
tion betw een th e various indexes of ra te of cure and T-50 are
listed on th e figures. (Prediction index = 1 — \ / l — r 2, w here r
is the coefficient of correlation calculated according to stan d ard
statistical m ethods. P .I. is 0 for no correlation and 1 for perfect
correlation. F o r six pairs of values as in the present instance
r > 0.8 or P .I. > 0.4 shows significant correlation w ithin th e
95% confidence lim its.) T he prediction index of 0.84 for tensile
ratio was th e highest found. Likewise the fact th a t tensile ratio,
is the only function which can be fairly well approxim ated by a.
single straig h t line confirms the indication given b y th e predic­
tion indexes.
T he tensile ratio has been used in the authors’ labo rato ry as;
an index of ra te of cure for various channel black stocks. I n at
series of tw enty-seven experim ental blacks in th e sam e base for­
m ula used for the pigm ents in T able I the tensile ratio an d tim e
to reach m axim um tensile were both m easured. T h e ra te o f
cure as defined previously was then determ ined b y dividing th e
tensile ratio of each experim ental stock by the tensile ra tio for
the stan d ard control which was cured a t th e sam e tim e as th e
experim ental sample. T h e sam e control was, of course, used
th roughout th e series. Likewise, th e rate of cure was also esti­
m ated b y dividing the tim e to reach m axim um tensile of th e con­
trol b y th e tim e to reach m axim um tensile for the sam ple. The.
INDUSTRIAL
Table II.
G um
P p td .
W h it­
ing
100
5
5
100
5
5
1.5
2
1 .5
2
MT
100
5
5
1 .5
o
ÄND
ENGINEERING
CHEMISTRY
G R -S
C o n tinex
HMF
100
5
5
100
5
5
1 .5
2
1 .5
2
C o n ti­
n e n ta l
AA
C o n ti­
ne n ta l
D
C o n ti­
n e n ta l
R-40
100
5
5
100
5
5
100
5
5
1 .5
2
1 .5
2
1.5
2
78
50
Vol. 16, No. 1
prediction index indicating the corre­
lation between the ra te of cure de­
term ined in these two w ays was then
found to be 0.24. Since for tw entyseven pairs of values a prediction
index g reater th a n 0.08 shows a sig­
nificant correlation, there is a p p ar­
ently a definite relation between these
two m ethods of determ ining ra te of
cure. Of the two m ethods, tensile
ratio can be d e t e r m i n e d m o r e
precisely.
50
DETERMINATION O F RATE OF
CURE IN GR-S
50
50
50
50
C u re a t 307« F .
Min.
<*s.
0
8
240*
240
240
240
240
250
90
210
320
400
410
420
220
580
850
1130
1130
1190
210
790
1270
1420
1450
1470
200
510
900
1230
1320
1480
260
360
800
1180
1260
1500
310
390
410
970
1100
1250
0
160
190
190
190
190
190
0
700
790
760
760
760
500
0
520
940
1130
1010
1010
840
40
650
1340
1630
1550
1460
1400
50
560
1880
2200
2200
2170
2100
100
300
1570
2380
2620
2590
2570
720*
950
2320
2550
2760
2600
80
500
740
1820
2360
2580
2570
460
315
400
300
315
340
225
300
590
560
550
550
545
455
275
880
805
665
625
590
520
415
685
590
520
390
385
350
410
600
550
450
415
395
385
490
640
625
575
465
450
425
550
675
660
600
490
490
430
550
600
660
650
550
540
480
12
10
4
6
4
5
5
15
10
9
6
5
5
4
17
7
4
5
4
4
4
15
10
9
5
4
2
2
32
22
17
12
5
5
5
37
16
19
15
5
5
5
31
29
26
20
15
14
10
20
70
90
90
85
80
80
20
115
135
100
100
110
85
i 55
225
185
155
150
150
20
125
265
225
220
175
170
30
50
230
325
285
265
270
10
33
36
35
35
35
35
15
42
44
44
45
43
43
14
28
35
41
42
42
43
15
33
43
49
51
52
51
16
33
44
51
52
55
54
21
38
45
55
59
58
60
23
37
45
54
58
58
61
33
41
47
57
61
61
66
8
32
35
35
34
34
35
11
44
44
44
43
41
41
6
33
36
41
41
41
40
9
33
42
48
51
54
52
11
33
44
50
51
53
51
11
29
40
50
56
55
57
14
29
42
51
55
56
59
23
35
44
50
58
57
62
43
53
51
52
53
53
52
28
39
40
40
39
40
38
33
44
47
48
46
46
46
20
38
40
43
43
42
41
28
38
41
42
43
42
42
25
30
31
34
35
33
32
23
28
30
31
31
30
31
19
22
25
25
26
26
26
15
30
60
90
180
26
56
57
58
60
59
58
33
4S
48
49
47
48
46
35
48
50
57
55
54
54
32
40
45
51
53
52
51
30
40
43
48
49
51
50
30
32
33
39
40
40
39
27
29
32
38
39
38
39
22
24
27
32
34
35
35
S
15
30
60
90
180
7 0 .5
8 4 .8
9 6 .9
9 8 .4
9 8 .8
9 S .0
8 9 .5
9 0 .6
4 3 .3
6 4 .0
7 S .4
8 6 .5
9 0 .0
7 0 .0
4 5 .5
6 5 .6
3 7 .5
6 6 .3
8 3 .2
3 2 .6
5 4 .2
7 2 .4
8 7 .4
8 9 .2
9 0 .4
15
30
60
90
ISO
0
8
15
30
60
90
ISO
0
8
15
30
60
90
180
0
8
15
30
60
90
180
0
8
15
30
60
90
180
0
8
15
30
60
90
180
0
8
15
30
60
90
180
0
8
15
30
60
90
180
0
8
9 5 .0
9 0 .8
7 2 .5
85! 6
*69 .*8
7 7 .6
iö ö ’
200
320
380
380
290
*4 3 .6
5 9 .0
8 5 .2
8 9 .9
9 3 .9
55
145
180
265
280
315
305
3 1 .6
3 6 .2
50.1
7 0 .0
82 1
8 8 .0
As in th e case of n atu ra l rubber
various pigm ents were studied in a
typical form ulation, 26 volumes of
pigm ent per 100 volumes of polym er
being used as before. T h e form ula
and te st results appear in Table II.
Since T-50 does n o t show m uch
v ariation w ith cure for GR-S, this
te st was om itted. In place of the
T-50 as a reference test w ith which
o ther indexes could be compared,
combined sulfur was determ ined by
analysis. T h e sodium sulfite m ethod
(1, 10) was used to determ ine free
sulfur and th e combined sulfur cal­
culated by sub tractin g th e free sul­
fur from th e to ta l sulfur. T he com­
bined sulfur values for all cures are
tabu lated in T able I I . I t is w orth
noting th a t the figures for the faster
curing stocks pass through a maxi­
mum. M oreover, if the slower curing
channel black stocks were examined
a t even longer cures th a n 180 m inutes,
it is possible th a t the combined sulfur
for these stocks too m ight pass through
a maximum. T he reason for th is be­
havior is no t clear.
T he sam e indexes of rate of cure
studied in n atu ral ru b b er were d eter­
mined. In F igures 5 and 6 those
indexes suggested by oth er investiga­
tors or found to correlate well with
combined sulfur are p lotted against
° A n in te re stin g c o n tra s t a p p ea rs in the
b eh av io r of elongation as a fu n c tio n of cure
tim e for n a tu ra l ru b b e r and G R -S.
In
n a tu ra l ru b b e r, elongation is hig h e st in the
uncured s ta te for all stocks an d decreases
reg u larly as cure proceeds (T ab le I).
In
G R -S, how ever, elongation increases w ith
cure reaching a m axim um for m oat stocks in
th e vicin ity of th e 8 -m in u te cure.
b In connection w ith com bined sulfur
values for th e gum stock, w hich ap p ro ach al­
m o st 100% of th e original su lfu r, for a gum
fo rm u la tio n differing from th e above stock
on ly in th a t 10 p a rts of zinc oxide an d 0.75
p a r t of sulfur w ere used in ste ad of 5 of zinc
oxide a n d 2 of Bulfur, th e m axim um com ­
bined su lfu r w as less th a n 80% of the
original su lfu r. T h e com bined su lfu r values
for th is sto ck are:
C u re, m in u tes
C om bined
sulfur, %
0
8
15
30
90
180
0
42
58
09
77
63
T h is behavior for G R -S is sim ilar to th a t
observed b y T h o rn h ill a n d S m ith (13) for
n a tu ra l ru b b e r. T h e y found th a t for higher
su lfu r loading th e p e r c e n t of th e original
sulfur com bined a t a given cure is higher.
January 15, 1944
ANALYTICAL
EDITION
19
Prediction indexes giving th e correlation of th e indexes of cure
ra te w ith cure tem p eratu re are listed a t the to p of Figure 7. T he
best correlation w ith cure tem perature is obtained w ith tensile
ratio and combined sulfur. T im e to reach m axim um tensile
also shows a significant correlation w ith cure tem p eratu re b u t
does n o t correlate as closely as th e o th er tw o indexes. (For the
eight stocks used in th is case a prediction index > 0.3 or a corre­
lation coefficient > 0.7 indicates a significant correlation w ithin
95% confidence limits.)
DISCUSSION
Studies w ith G R -I stocks also indicate th a t the tensile ratio is
a useful index of cure for this synthetic. I t is planned to pre­
sent these results in a separate paper.
I n dealing w ith certain special problem s th e tim e required to
obtain a satisfactory index of ra te of cure m ay possibly be short­
ened. F o r example, in com paring a series of stocks all of winch
have very sim ilar m axim um tensiles, th e tensile of th e proper
undercure is approxim ately proportional to th e tensile ratio and
can be used sim ilarly as an index of ra te of cure. A case of this
so rt would occur in th e control testing of rubber channel blacks
or other pigm ents. On th e o th er hand, although th e m axim um
tensiles for different stocks in a series m ay be different, th e cure
ra te m ay be sufficiently sim ilar so th a t th e tensile o f a fixed cure
in th e m axim um tensile range, say th e 60-m inute cure, m ay be
fairly close to th e m axim um tensile for all stocks. In th is event
th e tensile ratio w ould be approxim ately equal to th e tensile a t
15 m in u tes/tensile a t 60 m inutes, w hich could th en be used as an
index of ra te of cure an d would only require th e m aking of two
cures. A possible example of th is ty p e m ight be a series of stocks
containing different am ounts of petroleum -type softeners.
T h e rate-of-cure results for b o th n atu ral ru b b er and GR-S
indicate th a t th e tensile ratio is a satisfactory index of ra te of cure
even in a series of stocks containing a widely different v ariety of
pigm ents. However, it seems likely th a t th e g reatest usefulness
for th e tensile ratio will occur in th e comparison of sim ilar pig­
m ents or oth er compounding ingredients such as carbon blacks,
inorganic pigm ents, softeners, antioxidants, etc., in th e sam e test
formula.
T he above sh o rt cuts as well as tensile ratio itself are satisfac­
to ry indexes of ra te of cure only w ithin a set of stocks milled and
cured side b y side, such as th e stocks in T ables I and II. In the
comparison of stocks n o t m illed an d cured together a reference
control is custom arily run a t th e sam e tim e as each stock. The
04 J
(/)
z
a
COMBINED
SU LFU R AT IS’ CURE |% OF ORIGINAL SULFUR|
Figure 5
Figure 6
th e combined sulfur a t th e 15-minute cure expressed
as per cent of original sulfur. An index analogous
to tensile ratio b u t utilizing tensile p roduct instead
of tensile was also found to correlate w ith combined
sulfur alm ost as well as tensile ratio b u t was n o t in­
cluded in th e graph.
Prediction indexes listed on th e figures indicate the
best correlation for tensile ratio, tim e to reach 85%
of m axim um tensile, tim e to reach m axim um tensile,
and tim e to reach break in m odulus. T h e highest
value for the prediction index is obtained w ith tensile
ratio.
Since ra te of cure increases as tem p eratu re of cure
is increased, a good te st of an index of ra te of cure is
th e behavior of th e index a t different cure tem pera­
tures. In Figure 7, tensile ratio, combined sulfur a t
th e 15-m inute cure, and tim e to reach m aximum
tensile are plotted against cure tem perature. For
these tests C ontinental AA in th e G R-S form ula given
in T able I I was used. In addition to th e stocks cured
a t various tem peratures, four duplicate stocks were
cured a t 307° F . in order to give some indication of
the reproducibility of these th ree m ethods.
,
,
,------
CONTINENTAL A A IN G R -S
P. I.
T E N SILE RATIO
O ---------- — 0.73
COMBINED S U LFU R A
— 0.69
TIME TO MAX. T E .
X ------------ 0.36
TEN SILE
RATIO
TIME TO
MAX. TE.
(MIM)
Tt&. AT Igr\
I MAX. TE J
COMBINED
SULFUR
AT 15' 70
\SULFUftJ
60
310
CURE TEMP (°F)
Figure 7
20
INDUSTRIAL
AND
ENGINEERING
tensile ratio of the individual stocks divided by th e tensile ratio
for th e corresponding control should th en be tak en as th e index of
ra te of cure.
ACKNOW LEDGMENT
T he authors wish to th an k M iss B. C. M yerson and Miss M.
Sohn for th eir assistance in obtaining th e experim ental d ata.
LITERATURE CITED
(1) A m . Soc. T estin g M aterials, C om m ittee D - l l , D 297-41T -16b,
A .S .T .M . Standards on Rubber P roducts, p. 10 (1943).
(2) Columbian Colloidal Carbons, IV, 26 (1943).
(3) Fielding, J. I I ., I n d . E n g . C h e m ., A n a l . E d ., 12, 4 -9 (1940).
(4) G arvey, B . S ., Jr., and Freese, J. A ., Jr., Sum m ary of U npub­
lished W ork, Offico of Rubber D irector M em orandum , p. 5
(Aug. 7, 1943).
(5) G ibbons, W . A ., G erke, R . H ., and C uthbertson, G . R ., Proo.
R ubber T ech . Conference, London, P a p er 36, pp. 8 81-7 (1938).
CHEMISTRY
Vol. 16, No. 1
(6) G ibbons, W . A ., Gerke, R . H ., and T ingey,
C h e m ., A n a l . E d ., 5, 279-83 (1933).
H. C ., I n d . E n o .
(7) H aslam , G . S., and K lam an, H . C ., Ibid., 9, 5 5 2 -8
(1937).
( 8 ) K u sov, A ., Kautschuk, 11, N o . 2, 2 4 -7 (1935);
Rubber Chem.
Tech., 8 , 548 -5 3 (1935).
(9) N au gatu ck C hem ical, R eport N-40-1, B una S, p . 9 (Jan. 27,
1942).
(10) O ldham , E . W ., Baker, L. M ., and Craytor, M . W ., I n d . E n o .
C h e m ., A n a l . E d ., 8 , 4 1 -2 (1936).
(11) R oberts, G . L., Proc. Rubber T ech . Conference, L ondon, P aper
54, p p . 5 06 -1 5 (1938).
(12) S teven s, H . P ., In d ia Rubber J ., 52, 4 0 1 -3 (1916).
(13) Thornhill, F. S ., and S m ith, W . R ., I n d . E n q . C h e m ., 34, 2 1 8 24 (1942) p . 222.
(14) Villa, G. R „ I n d ia Rubber W orld, 101, N o . 2, 3 4 -8 (1939).
(16) W hitby, G . S., " P lan tation Rubber and the T estin g of R ubber",
p p . 320, 340, London, L ongm ans, G reen & C o., 1920.
(16) W iegand, W . B ., I n d . E n g . C h e m ., 18, 1157-63 (1926).
P
before th e N ew Y ork G roup, D ivision of R u b b e r C h em istry ,
resented
Am
e r ic a n
C
h e m ic a l
S o c ie t y .
Determination of Alpha^ara-Dim ethylstyrene
In the Presence of Para-M ethylstyrene, Styrene, and Para-Cymene
J O H N H . E L L IO T T a n d E V E L Y N V . C O O K
H ercules Experiment Station, H ercules Powder Com pany, W ilm ington, D el.
a,p-Dim ethylstyrene can be determined in the presence of p-m ethylstyrene, styrene, and p-cym ene by two independent methods. The
chemical method depends upon addition of hydrogen chloride or
bromide to the styrenes with subsequent estimation of the tertiary
halide formed by a,p-dimethylstyrene. The other method involves
an analysis of the ultraviolet absorption curve of such mixtures.
Results obtained by these independent methods are in good
agreement.
CHEM ICAL METHOD
T he chemical m ethod of analysis depends upon th e following
reaction. If th e addition of H X to a,p-dim ethylstyrene an d pm ethylstyrene follows M arkownikoff’s rule, where X is either Br
or Cl, th e following compounds will be formed:
X
X
C H .— ¿ — C H i
E E D has recently arisen for a m ethod for th e determ ination
of a-jp-dimethylstyrcne in th e presence of p-m ethylstyrene,
styrene, an d p-cymene. Since no satisfactory procedures were
available for th e analysis of such m ixtures, chemical m ethods,
ultraviolet absorption, an d polarographic analysis were consid­
ered.
Prelim inary experim ents on th e last, using th e technique of
L aitincn an d W awzonek (I) were n o t prom ising an d this approach
w as n o t studied further.
I t w as found possible, however, to analyze such m ixtures by a
chemical m ethod and also by m eans of ultraviolet absorption.
B oth procedures were readily ad ap tab le to control use an d have
been successfully applied to th e analysis of a large num ber of
samples.
H—
CHj
N
In order to te st these m ethods it was first necessary to obtain
pure sam ples of the various styrenes and p-cymene. T he a,pdim ethylstyrene and p-m ethylstyrene were synthesized and pu ri­
fied in this laboratory. A good grade of commercial styrene
was vacuum -distilled several tim es before use. T he p-cymene
was distilled through a 40-plate column packed w ith 0.23-cm.
(’/«-inch) stainless steel helices a t 100-mm. pressure and a reflux
ratio of approxim ately 60/1, a middle cu t being coEected and used
in th is work.
T he halogen in (1), formed from a,p-dim ethylstyrene, is a t­
tached to a te rtia ry carbon atom , while th a t in (2), formed from
p-m ethylstyrene, is attach ed to a secondary carbon atom . The
tertia ry halogen in (1) is m uch m ore readily hydrolyzed th a n the
secondary in (2), and this fact offers a m eans of analytically de­
term ining <x,p-dimcthylstyrene in th e presence of p-m ethylstyrene
a n d styrene. T here is no reaction of ILX w ith p-cymene.
I t has been reported th a t hydrogen brom ide adds rapidly to
styrene (2, 10), and th a t th e secondary brom ide formed by nor-
Table I.
C om ­
po u n d
p-C ym ene
T he constants of these compounds are given in T able I. The
brom ine num bers were run by a modification of th e m ethod of
U hrig and Levin (3).
T he refractive index results are in good agreem ent w ith the
In ternational C ritical Tables values, th e tem perature being con­
sidered. T he sam ples of th e various styrenes were alw ays
freshly vacuum-dlstiUed before being used.
S tyrene
p -M e th y l8ty re ne
ci,p-Dim eth y latyrene
Constants of Materials Used in Testing M ethods
F o rm u la
p -C H i.C iH « .C H (C H ,),
C ,H ,.C H :C H t
p -C H j. C«H i. C II :CH,
p -C H 1 .CflH 4.C (C H ,):C H ,
r ’d ' 5
1.4913
1.5467
1.5421
1.5356
no
(I.C .T .)
1.4947
(15° C.)
1.5467
(20° C .)°
1.5447
(1 6 .4 ° C.)
1.5344
(1 8 .7 ° C.)
B rom ine N o.
O b­
served
T h e o ry
<1
0
151
154
137
135
117
121
* V alue o t M oore, B u rk , a n d L an k elm a (7).
1
January 15, 1944
Table II.
ANALYTICAL
Determination of a,p-Dim ethylstyrene in Known Mixtures
(U sing m eth o d inv o lv in g a d d itio n of H B r)
W eig h t of a ,p -D im e th y lsty rc n e
O th e r M ate rials P re se n t
P re se n t
R ecovered
Grain
Gram
N one
N one
0.2080 g ra m p -m e th y lsty re n e
N one
N one
0.1 m l. of benzene, 0.2 m l. of
s ty re n e, 0.2 m l. of p -m e th y l­
sty re n e, 0.5 ml. of p-cym ene
0.2731
0.2131
0 .2 4 0 8
0.2 6 9 3
0 .2 4 1 4
0.4 1 4 7
0 .2 6 1 3
0.2 5 4 5
0.1 9 7 7
0.2 2 1 5
0.2 4 7 7
0 .2 2 4 5
0.3942
0.2 5 2 2
R ecovery
%
9 3 .3
9 2 .8
9 1 .9
9 2 .0
9 4 .0
9 5 .2
9 6 .7
25.6% p -m e th y lsty re n e an d
48.8% p-cym ene
0 .0 8 5 4
0.1 0 0 6
0 .0 7 9 6
0 .0 9 2 6
9 4 .2
9 2 .0
m al addition m ay be hydrolyzed under relatively mild condi­
tions, indicating th a t th e conditions of hydrolysis of th e te rtiary
halide m ust bo chosen so as to avoid interference from th e second­
ary halides present. B ased on these facts, th e following m eth ­
ods have been developed, involving th e addition of IIX , rem oval
of excess, followed by hydrolysis, and estim ation of th e tertiary
halide formed from a,?j-dim ethylstyrene under conditions where
there is no appreciable hydrolysis of th e secondary halide formed
from p-m ethylstyrene and styrene.
A sam ple weighing approxi­
m ately 0.5 to 1.0 gram is accurately weighed o u t into a 125-ml.
Erlenm eyer flask containing 25 ml. of carbon tetrachloride.
Gaseous hydrogen brom ide, generated by dropping brom ine onto
naphthalene and purified by passing through several towers of
naphthalene and finally one containing D rierite, is bubbled
through the sample for 30 m inutes a t a rate of 3 to 4 bubbles per
second. A t the end of this tim e th e addition tube is washed
down w ith 10 ml. of carbon tetrachloride and a brisk stream of
nitrogen is passed through the solution for 30 m inutes to remove
unreacted hydrogen bromide. T he flask is then rem oved and
chilled in an ice-salt b ath for 10 m inutes. T h irty milliliters of
H y d r o g e n B ro m id e M e th o d .
EDITION
21
previously chilled 90% alcohol are then added, th e flask is swirled
for a few seconds, tw o drops of m ethyl red indicator are added,
and the cold solution is rapidly titra te d w ith stan d ard O.liV alco­
holic potassium hydroxide to a b right yellow end point lasting
for 10 seconds. T able II gives the results obtained by this m ethod
on known samples.
H y d r o g e n C h l o r i d e M e t h o d . T h e procedure is th e sam e as
th a t used in th e hydrogen brom ide m ethod w ith th e following
changes:
Benzene instead of carbon tetrachloride is used as th e solvent.
F o rty milliliters of 80% alcohol are used to effect hydrolysis
of the te rtia iy halide.
The titra tio n is performed a t room tem perature and th e sample
titra te d with 0.1 A 'alcoholic potassium hydroxide to a 30-second
bright yellow end point. H ydrogen chloride was generated by
dropping concentrated hydrochloric acid into concentrated sul­
furic acid. T h e evolved gas was passed through a tow er filled
w ith concentrated sulfuric acid before being passed into the solu­
tion being analyzed.
T able I I I gives th e results obtained b y this m ethod on known
samples. T he end points are considerably sharper th a n those
given by th e hydrogen brom ide m ethod.
Table III.
Determination of a,p-Dim ethylstyrene in Known
Mixtures
(U sing m ethod involving a d d itio n of HC1)
W eight of a ,p -D im e th y lsty re n e
S ubstances P rese n t
P rese n t
R ecovered
Gram
Gram
p-C ym cno
S ty ren e
p -M e th y lsty re n e
a ,p -D im eth y lsty ren e
a ,p -D im e th y lsty re n e plus 0.5
ml. each of p-cym ene, s ty ­
rene, and p -m e th y lsty re n e
N one
N one
N one
0 .2936
0.4739
0 .4116
0.3077
0.5693
0.4015
0.4052
N one
N one
N one
0.2740
0.4508
0.3903
0 .2940
0.5415
0 .3855
0 .3902
R ecovery
%
93 '.h
9 5 .3
9 4 .7
95.5 *
9 5 .3 °
9 6 .0
9 6 .2
a HC1 gas a dded for 1 ho u r in ste ad of 30 m inutes as in a ll o th e r cases.
Table IV .
Sam ple
A
B
C
D
E
F
G
H
A n alyse s by H ydrogen Bromide and H ydrogen
Chloride Methods
H B r m ethod
%
1 9 .6 ,2 0 .2
2 9 .2 ,2 9 .5
2 0 .2 ,1 9 .9
2 0 .5 ,2 0 .4
2 2 .6 ,2 2 .0
1 0 .6 ,1 0 .7
3 2 .8 ,3 2 .6
3 2 .9 ,3 2 .6
H C l m ethod
%
1 9 .8 ,2 0 .6
2 8 .2 ,2 9 .3
2 0 .2 , 19.8
2 0 .6 ,2 0 .6
2 2 .6 , 2 3 .3
9 .5 , 9 .5
3 2 .2 ,3 2 .0
3 1 .8 ,3 2 .5
DISCUSSION OF RESULTS
W AVELEN G TH
IN
M IL L IM IC R O N S
Figure 1. Ultraviolet A bsorption Curves
For a, p-dimethylstyren«,p-methylstyrene, styrene,end p-cymene In
ethanol solutions
An exam ination of Tables I I and I I I shows th a t som ew hat low
recoveries (about 95% ) of a,p-dim ethylstyrene are obtained in
every case. Therefore, to correct for th is a factor of 1.05 is used
in calculating th e results. T his low recovery of a,p-dim ethylstyrcne is n o t raised by a longer tim e of addition of th e hydrogen
halide. I t m ay be due to th e fact th a t a sm all fraction of the
addition takes place contrary to M arkownikoff's rule. A con­
sideration of the results obtained by use of these m ethods in the
analysis of over fifty sam ples of know n an d unknow n composition
indicates th a t th e precision of th e m ethods is approxim ately =*=3 %
of th e a,p-dim ethylstyrene present.
Table IV shows the agreem ent betw een th e hydrogen brom ide
an d hydrogen chloride m ethods in th e analysis of eight different
sam ples of unknow n composition.
T he agreem ent betw een th e results obtained b y th e tw o m eth­
ods is seen to be good. Since th e hydrogen chloride m ethod is
som ew hat easier to use and th e end p o in t is sharper, it is recom­
mended for use.
22
INDUSTRIAL
AND
ENGINEERING
CHEMISTRY
Vol. 16, No. 1
ULTRAVIOLET ABSORPTION METHOD
\
T he spectrophotom etric m ethod has been successfully applied
to the q u an titativ e determ ination of compounds having charac­
teristic absorption bands in th e ultraviolet region of th e spec­
trum . Good examples of th is m ethod used in calculating two,
three, and four constituents of a m ixture have been reported in
th e literature (S-6 ).
In m ixtures of pure styrene, «,p-dim ethylstyrene, p-m ethylstyrene, and p-cym ene it was found possible to determ ine th e
am ount of each styrene present w ithin 2 % of th e know n value.
T he absorption d a ta were obtained from m easurem ents made
w ith a B eckm an quartz spectrophotom eter. T he solvent was
ethanol in all cases. T he form ulas used in m aking the calulations
use th e term specific a :
,
cl
where a = absorption coefficient
70 = inten sity of radiation tran sm itted b y th e solvent
I — intensity of radiation tran sm itted b y th e solution
c = concentration of solute in gram s per 1000 ml.
I = length in centim eters of solution thro u g h w hich the
radiation passes
T he sam ples of pure a,p-dim ethylstyrene, p-m ethylstyrene,
styrene, and p-cym ene used to obtain th e specific absorption
coefficients of th e pure compounds were th e sam e as those used
in th e developm ent of th e chemical m ethod previously described.
Per Cent Com position
C om position of
M ix tu re of P u re
C om pounds
M ix tu re
C alculated
from
U ltra v io le t
A b so rption
D a ta
% .
3 0 .8
4 2 .4
2 6 .0
E rro r
%
1 .2
0 .4
0
1
<x,p-D im ethylstyrene
p -M e th y lsty re n e
S ty ren e
K now n
%
32
42
26
2
a ,p -D im eth y lsty ren e
p -M e th y lsty re n e
S ty ren e
5
5
90
6 .0
5 .3
8 8 .1
1 .0
0 .3
1 .9
3
« .p -D im e th y lsty re n e
p -M e th y lsty re n e
S ty ren e
5
90
5
3 .9
9 0 .3
5 .3
1.1
0 .3
0 .3
4
a .p -D im eth y lsty ren e
p -M e th y lsty re n e
S ty ren e
p-C ym ene
25
26
25
24
2 4 .3
2 6 .3
2 5 .9
0 .7
0 .3
0 .9
5
or.p-D im ethylstyrene
p -M e th y lsty re n e
S ty ren e
p-C ym ene
20
30
35
15
2 1 .9
2 9 .4
3 5 .8
1 .9
0 .6
0 .8
h
logro y
Specific a =
Tabic V .
Table V I .
Samples of Unknown Composition
a .p -D im e th y lsty re n e
U ltra v io le t
a b sorption
m ethod
C hem ical
m ethod
Sam ple
T he values a t different wave lengths are shown in Figure 1.
T he curve for styrene agrees very closely w ith previously pub­
lished d a ta concerning this com pound (5).
%
%
2 3 .0 .2 2 .9
2 0 . 0 . 2 0 .2
3 2 .5 , 3 2 .3
2 7 .3 ,2 7 .4
2 1 .7 ,2 2 .3
19.8, 2 0 .3
3 2 .8 , 3 2 .5
2 7 .8 ,2 7 .5
wave lengths there are sufficient differences in a of th e pure com­
pounds to perm it q u an titativ e calculations. I t is also evident
th a t p-cym ene h as its absorption values below 280 m p and th ere­
fore does n o t appreciably affect absorption values above 280 mp,
even though present in q u antities up to 70 or 80% .
Q u an titativ e determ inations on three unknow ns require the
selection of th ree suitable wave lengths in order to set up three
sim ultaneous equations. A t a given wave length, components
a, b, and c present in a solution in percentages of x, y, an d z, re­
spectively, will give a to ta l absorption value w hich can be repre­
sented by th e equation:
M x +
Figure 2. Ultraviolet Abso rp tion Curves
For a, p-dlmcthylstyrcnc, p-mtUiyltlyrcn«. styrene end p-cymene In ethanol solutions
(enlarged scale of spectral region ¡75 to 300 mp)
T he styrenes have a strong absorption b and in th e spectral
region 248 to 252 m p w ith less pronounced b u t y e t d istin c t bands
in th e region to 300 mp. F o r th e th ree styrenes th e bands in th e
region 248 to 252 mp are so sim ilar th a t no a tte m p t w as m ade to
use them for q u an titativ e determ inations. A larger scale graph
of the region 280 to 300 mp (Figure 2), shows th a t a t certain
+
= “ solution
a ., a b , a e represent th e specific absorption coefficients for each
of th e pure compounds a t th a t particu lar wave length. B y set­
tin g u p th ree equations a t different wave lengths, th e percent­
ages (x, y, and z) of each com ponent can be calculated.
Since sim ultaneous equations w ith three or m ore unknowns
are custom arily solved b y th e m ethod of determ inants, th e cri­
terion for th e selection of wave lengths is th a t th e determ inants
in th e expressions for th e unknow n concentrations be as large as
possible. Likewise, th e individual values of a should be large
enough to be m easured w ith accuracy on th e spectrophotom eter.
I n know n m ixtures of pure compounds, w ith or w ith o u t pcymene, it was found possible to calculate th e per cent composi­
tion by using th e absorption d a ta a t wave lengths 285, 291, and
295 m p (Table V). Values calculated from th e absorption d a ta
differ from know n values by less th a n 2% . T h e presence of pcymene did n o t interfere w ith th e calculations. T he wave
lengths 283, 287, an d 291 m p were found to give ju s t as accurate
values as th e previously m entioned wave lengths. I t is th ere­
fore possible to use tw o sets of wave lengths to serve as checks on
each other.
F o u r sam ples of unknow n compositions were analyzed for
a,p-dim ethylstyrene b y b o th th e chemical and ultraviolet ab­
sorption m ethods (Table V I).
T he agreem ent betw een these independent m ethods is seen to
be good.
January 15, 1944
ANALYTICAL
EDITION
23
ACKNOWLEDGMENTS
(5) M iller, E . S., M ackinney, G ., and Zacheilo, F . P ., Ib id ., 10, 375
T he authors wish to express th eir th an k s for the assistance in
running analyses given by Jordan P. Snyder and M iss M ary
T hom as M ontgom ery.
( 6 ) M itchell, J . H., K ray bill, H . K ., and Zscheile, F . P ., I n d . E n g .
LITERATURE CITED
(1) Laitinen, H . A ., and W aw zonek, S., J . A m . Chem. Soc., 64, 1765
(1942).
(2) M ayo, F . I t . , and W alling, C ., Chem. Rev., 2 7 , 351 (1940).
(3) M iller, E . S ., J . A m . Chem. Soc., 57, 347 (1935).
(4) M iller, E . S., P la n l P hysiol., 12, 667 (1937).
(1 9 3 5 ).
C h e m ., A n a l . E d ., 15, 1 (1 9 4 3 ).
(7) M oore, B urk, and Lankeltna, J . A m . Chem. Soc., 63, 2 9 5 4 (1 9 4 1 ).
(8) P estem er, M ., and W iligut, L., M onalsh., 6 6 , 1 19 (1 9 3 5 ).
(9) Uhrig, K ., and L evin, H., I n d . E n g . C h e m ., A n a l . E d ., 13,
9 0 (1 9 4 1 ).
(10)
W alling, C ., K harasch, M . S., and M ayo, F . R ., J . A m . Chem.
Soc., 61, 2 6 9 3 (1 9 3 9 ).
before th e D ivision of A n a ly tic al a n d M icro C h em istry a t th e
100th M eeting of th e A m e r i c a n C h e m i c a l S o c i e t y , P ittsb u rg h , P enna.
P r e s e n te d
Determination of Soluble Pectin and Pectic A c id by
Electrodeposition
K E N N E T H T. W I L L I A M S A N D C L A R E N C E M .
Western Regional Research Laboratory, A lb a n y ,
In a new approach to the determination of pectin, solutions
of pectin or pectates are deashed by the use of ion-exchange
resins. The pectin or pectic acid is electrolytically de­
posited at a platinum anode in a weighable form. The
method requires less of the analyst's time and attention
than does the calcium pectate method and, with partially
purified solutions, gives results of the same order of accuracy.
The method is especially applicable to amounts of pectin
ranging between 5 and 50 mg.
ST IG A T IO N S designed to develop new an d extended uses
INforV Epectin
have m ade desirable an analytical m ethod capable
Since th e soluble pectinous m aterials are negatively charged
colloids, it was decided to investigate th e possibility of collecting
th e pectin on th e anode of a suitably arranged electrolysis system .
Brown (2) unsuccessfully a ttem p ted to precip itate pectin
electrolytically from aqueous solutions containing large am ounts
of electrolytes. Griggs and Jo h n stin (5) observed flocculation
of pectin a t th e anode on electrolysis w ith 110-volt d irect cur­
rent. T h e authors have found th a t pectin can be q u an titativ ely
determ ined by electrodeposition, provided th e electrolyte con­
centration of th e solution is low. Ion-exchange resins can be
used when it is necessary to rem ove th e electrolytes before electro­
deposition.
APPARATUS
of determ ining sm all am ounts of pectin. A pproxim ate estim a­
tions of pectin can be m ade b y p recipitation w ith 70 to 80 per
A conventional electrolysis ap p a ra tu s supplied w ith 220-volt
cent alcohol or 50 per cent acetone (1, 3, 8, 9). Alcohol and
d irect current from a rectifier-transform er u n it was used. A
m ercury cathode cell was constructed from a 250-ml. G riffin-type
acetone of these strengths precipitate gums, some proteins, and
beaker, in to th e side of which was fused a platinum wire (Figure
calcium and potassium salts of some of th e organic acids as well
1). A mercury-filled side arm was added to provide flexibility
as pectin. However, when d ilute solutions are used th e results
in th e connection to th e negative binding post. Clean m ercury
are likely to be low because of incom plete precipitation of the
completely covering th e platinum wire in the vessel served as
the cathode. T h e anode was a disk of 45-mesh p latinum gauze
pectin and difficulties in handling sm all am ounts of th e precipi­
6.25 cm. (2.5 inches) in diam eter, edged w ith 0.075-cm. (0.03-inch)
ta te. T h e alcohol p recipitate is difficult to filter and wash
p latinum wire to give rig idity and supported by a 15-cm.
thoroughly. T his filtration m ay be im proved b y precipitation
(6-inch) piece of 0.127-cm. (0.05-inch) p latin u m wire a t­
w ith acidified alcohol, b u t a fte r several w ashings enough acid
tached to th e center (Figure 1). Sim ilar electrodes are availm ay rem ain w ith the pectin to cause charring
when th e precipitate is dried. F ifty per cent
acetone yields a precipitate which is m ore easily
filtered and w ashed b u t affords little im prove­
m e n t in accuracy.
T he pectic
je ct to error
acid. T h e calcium p ectate m ethod of E m m et
and Carr6 (4), although tedious and tim econsuming, is probably th e m ost reliable of the
m ethods in use a t th e present tim e, b u t is no t
readily adaptable to very sm all am ounts of
pectin because of the large num ber of m anipu­
lations required. I t has been used throughout
this investigation for the analysis of stock solu­
tions for com parisons w ith the proposed m ethod.
T h e proposed m ethod is th e resu lt of a n a t­
te m p t to devise an accurate m ethod w hich re­
quires less of the an aly st’s tim e and atten tio n
and can be used to determ ine sm aller am ounts of
pectin th an the m ethod of E m m et and Carr6.
Figure 1. Electrolysis V essel, A n o d e and A n o d e s in H o ld er
24
INDUSTRIAL
Table I.
AND
ENGINEERING
CHEMISTRY
A system of commercial
ion-exchange resins has been
used to rem ove electrolytes
from pectin solutions in
th is laboratory. (The use
of exchange m aterials for
th e rem oval of ash constitu­
ents from pectin solutions
was first investigated in this
laboratory by W. D . M aclay
and co-workers who are in­
vestigating th e efficiencies
of various comm ercial prod­
ucts for th is purpose.) T he
I R 100 resin removes the
cations b y exchange w ith
hydrogen ion, while the IR
4 resin rem oves th e acids
which would interfere, by
conductance, w ith th e clectrodeposition of pectin. No
pectin was lost during this
process when th e conditions
described below were fol­
lowed. T h e capacity and
regeneration of these resins
are described by th e m anu­
factu rer (7).
Electrodeposition of Pectin from Solutions Containing
V arious Percentages of A lc o h o l
(25-m l. aliq u o ts of sto ck so lu tio n .
T im e, 5 hours)
A lcohol
Content
W eig h ts of D eposits
M g.
M g. M g.
M g.
% by volume
30
IT .2
1 8 .9
2 0 .6
40
50
60
2 1 .8
22.1
7 .9
7 .6
2 2 .2
8 .5
9 .2
7 .6
7 .1
Table II.
T im e
H r.
2
3
4
5
6
2 1 .9
5 .7
9 .8
R em arks
S olutions becam e h o t. O ne of 4
d eposits d id n o t ad h ere to
anode
C o m p a ct deposits
D eposits n o t co m p act
B u lk y deposits
Time Required for Electrodeposition of Pectin from
40 Per Cent A lc o h o l Solutions
(25-m l. aliq u o ts of sto ck solution)
W eights of D eposits
A verage
M g.
M g.
M g.
M g.
1 9 .9
1 8 .9
2 0 .8
2 1 .8
2 1 .9
2 0 .4
2 0 .5
2 0 .3
22 .1
2 1 .7
1 9 .4
19 .6
2 0 .1
2 2 .2
2 2 .2
1 9 .0
2 0 .3
2 0 .1
2 1 .9
2 2 .1
1 9 .7
1 9 .8
2 0 .3
2 2 .0
2 2 .0
R em arks
P recision
P recision
P recision
Precision
P recision
poor
poor
fair
good
good
able from platinum m anufacturers. T his electrode system has
proved to be m uch m ore satisfactory th an one using either
cylindrical or spiral electrodes, since th e m ercury cathode-disk
anode system m akes it possible to have all th e solution under
the influence of the field w ithout stirring.
Ion-exchange resin columns were prepared from 24 X 190 mm.
glass tubing constricted to a 6 -mm. o u tlet (Figure 2). T en
gram s drained w eight of ion-exchange resin, supported on a glass
wool m at, w as used in each tube. T h e cation-exchange column
was m ounted so th a t the effluent dripped directly into th e acidabsorbing bed.
ELECTRODEPOSITION O F PECTIN FROM LO W -A SH SOLUTIONS
T h e following procedure w as found to give a firm deposit th a t
adhered very well to the anode:
An aliquot of th e aqueous solution containing approxim ately
5 to 50 mg. of pectin was placed in the electrolysis vessel and
diluted to 0 8 ml. w ith distilled w ater, and 42 ml. of 95 p er cent
ethyl alcohol were added w ith stirring. T o o btain q u an tita tiv e
deposition in the sh o rtest tim e, th e alcohol concentration m ust
be close to 40 per cent by volum e, as indicated in T able I. P re­
cipitation of th e pectin a t th is point indicates th a t the electrolyte
content is too high an d th a t th e solution m u st be deashed.
D uring th e electrolysis the vessel was immersed in a water-ice
b a th to prevent an undue rise in th e tem perature of th e solution.
T he anode was lowered into th e solution so th a t th e gauze was
barely covered and 220-volt d irect cu rren t was applied. The
cu rren t varied between 5 and 20 milliam peres, depending on the
electrolyte con ten t and th e tem perature of the solution. N o
stirring of any kind was employed during th e rim. Occasionally
it was necessary to ad ju st th e anode level if the solution had
expanded or contracted as a result of changes in tem perature.
A t the end of 5 hours th e anode was w ithdraw n from th e vessel
and im m ersed in 99 per cent eth y l alcohol for 3 m in u tes an d then
in anhydrous ether for 3 m inutes.
A very com pact deposit w as obtained, which was fu rth er de­
h y d rated by th e alcoliol-ether tre atm e n t to a v ery thin uniform
film. T his alcohol-ethcr tre a tm e n t p erm itted complete drying
in o n e h o u r a tl0 5 ° C. T he electrodes were cleaned by immersion
in boiling w ater for a few m inutes, followed by rinsing in distilled
w ater. If th e deposit w as pectic acid, th e addition of a little
d ilute alkali aided in the solubilization. R em oval of the deposit
by ignition in a flame w as avoided because of th e possible danger
of changing th e electrode surface.
T able I I shows the relation between the tim e of deposition and
the weight of the deposit.
RE M O V A L OF ELECTROLYTES FROM H IG H -A SH SOLUTIONS
Pectin deposition from ex tracts of fru it was unsuccessful
because the current was transferred b y th e electrolytes present.
T he solution heated up and gas was evolved from b o th electrodes.
N o deposit was obtained.
Vol. 16, No. 1
Solutions containing 0.4
to 1 .0 m g. of pectin per ml.
were deashed in th e follow­
ing m anner: T h e columns
were rinsed w ith two 1 0 - to
15-ml. portions of th e solu­
Figure 2. Deashing System
tion, w ith thorough drain­
Upper, cation exchange column,
age betw een rinses. T he
lower, add adsorbing column
rinsings were discarded and
th e m ain portion of th e
solution w as deashed by
percolation through the columns a t th e ra te of approxim ately 3 ml.
per m inute.
T he pectin concentration of th e solution w as unchanged by
this trea tm e n t (Table II I ) . T h e pectin solutions trea ted in this
m anner were low in ash, ranging from 0.5 to 4 mg. per 100 m l.,
and successful electrodepositions could be m ade from such
solutions. T h e resin system has been used successfully on neu­
tral, acid, and alkaline aqueous ex tracts of pectin.
Table III.
Pectin Content of Solutions before and after Ion-Exchange
Treatment
Source of P e c tin
B efore“
M g ./m l.
A fte r“
M g ./m l.
E x tra c t of whole g ra p e fru it
1 .0 3
1.03
E x tra c t of whole orange
0 .6 4
0 .6 4
E x tra c t of whole apple
0 .4 1
0 .4 0
H igh-ash com m ercial p e ctin
0 .3 5
0 .3 5
o A verago of th re e analyses b y m eth o d of E m m e t a n d C arré (4).
Table IV .
E x tra c t
of W hole
G ra p e fru it
Electrodeposition of Pectin from A c id Extracts of W hole
Fruit after Removal of Electrolytes“
A liquot of
O riginal
S olution
M l.
M g.
—P e c tin D epositedM g.
M g.
M g.
A verage
Mg./rrd.
2 .5
14.2
14.1
14.1
5 .6 5
2 8 .4
5 .0
2 8 .1
2 8 .2
2 8 .0
5 .6 4
10.0
5 7 .8
5 7 .1
5 8 .3
5 7 .7
5 .7 7
O range
5 .0
13 .8
14.4
13.7
1 3 .8
2.7 8
10.0
2 9 .0
2 8 .9
2 8 .7
2 8 .5
2 .8 8
A pple
5 .0
10.9
11.1
11.1
11.1
2 .2 2
10.0
2 1 .4
21 .4
2 1 .8
2 2 .0
2 .1 7
2 5 .0
5 5 .2
5 4 .8
5 6 .7
5 6 .2
2 .2 3
a A verages of trip lica te analyses b y m e th o d of E m m e t and C arré (4)
yielded 5.16 m g. of p e ctin p er m l. of g ra p e fru it e x tra c t, 2.57 mg. pe r ml. of
orange e x tra c t, an d 1.99 m g. p e r ml. of apple e x tra c t.
January IS, 1944
ANALYTICAL
Table V . Electrodeposition of Pectic A c id from Extracts of W hole
A p p le and Grapefruit after A ce to n e Purification, H yd ro lysis, and
Removal of Electrolytes'*
E x tra c t
ot W hole
A pple
G ra p e fru it
A liq u o t of
O riginal
S olution
M l.
1 0 .0
2 0 .0
5 .0
1 0 .0
------- P e c tin A cid D ep o sited ------ .
M g.
M g.
M g.
M g.
1 0 .7
2 2 .1
10 .6
2 1 .1
1 0 .6
2 2 .1
1 0 .4
2 1 .0
1 0 .7
2 2 .1
10.7
2 1 .5
1 0 .7
1 0 .7
2 1 .2
A verage
M g ./m l.
1 .0 7
1 .10
2 .1 2
2 .1 2
a A verages of trip lica te analyses b y m e th o d of E m m e t a n d C arré (4)
yielded 1.07 m g. of p eotie acid p er ml. of ap p le e x tra c t an d 2.13 mg. of pec­
tic acid p e r m l. of g ra p e fru it e x tra c t.
A N A LY SE S OF EXTRACTS OF GRAPEFRUIT, O RA N G ES, A N D APPLES
T h e flavedo was rem oved from oranges and grapefruit, leaving
m ost of the albedo in tact. T h e apples were n o t peeled. T he
following m odification of the extraction procedure of M yers and
B aker (5) was used: ,
T he fru it was cu t up an d disintegrated in a W aring Blendor
w ith the addition of sufficient w ater and hydrochloric acid to
give a th in slurry of pH 1.7 to 2.0. T h e slu rry was th en heated
in a w ater b a th a t 85° C. for 30 m inutes an d filtered through
analytical grade Celite. T he residual pulp was extracted and
filtered twice, m ore in th e sam e m anner. T h e clear extracts were
com bined and diluted, if necessary, to a p ectin content of a p ­
proxim ately 1 mg. per ml. A liquots of 100 ml. were tak en for
analysis by th e calcium p ecta te m ethod of E m m et an d Carré.
E lectrolytes were rem oved from o th er portions of th e solution
25
EDITION
b y ion exchange and the pectin was determ ined by electro­
deposition and by th e calcium pectate m ethod (T able IV ).
A nalyses b y th e proposed m ethod gave résulta from 9 to 13
per cent higher th a n those obtained b y the m ethod of E m m et
and Carré.
N one of th e extracts used in th is stu d y contained starch or
dextrins, as indicated by th e absence of an y color form ation
w ith iodine solution. I t is probable th a t such interference
could be obviated, when necessary, b y am ylolytic tre atm en t
prior to deashing.
E lectrodepositions of pectic acid were also m ade from deashed
aliquots of fru it extracts w hich h ad been subjected to th e sam e
prelim inary acetone purification an d hydrolysis as used for
calcium pectate determ ination. T hese results w ere in good
agreem ent w ith those obtained by th e m ethod of E m m et and
C arré (T able V).
LITERATURE CITED
(1) Assoc. Official Agr. C hem ., “Official and T e n ta tiv e M eth od s of
A n alysis” , 5 th ed., p. 340 (1940).
(2) B row n, J. C., A m . Vinegar In d ., 2, 14 (1923).
(3) Carré, M . H ., and H ayn es, D ., Biochem. J ., 16, 60 (1922).
(4) E m m et, A . M „ and Carré, M . H „ Ibid., 20, 6 (1926).
(5) Griggs, M . A ., and Johnstin, R u th , I n d . E n o . C h e m ., 18, 623
(1926).
( 6 ) M yers, P . B ., and B aker, G. L., D el. Agr. E xp t. S ta., B ull. 160,
T ech . N o . 10 (1929).
(7) R esinou s P roducts and C hem ical Co., P hiladelphia, P enna.,
“T h e A m berlites” .
( 8 ) W ichm ann, H . J., J . Assoc. Official Agr. Chem., 6 , 34 (1922).
(9) Ibid., 7, 107 (1923).
A nalysis of Petroleum O il-So lu b le Sodium Sulfonates
by Adsorption
JO H N M. KO CH
The A tlan tic Refining Co¡, Philadelphia, Penna.
A n adsorption method for separating and estimating the
components of oil-soap mixtures has been developed,
which compares favorably with aqueous alcoh olic extrac­
tion analyses in accuracy and precision. It is rapid and
convenient and entirely free from emulsion difficulties.
H E m anufacture of medicinal w hite oil b y th e tre a tm e n t
INofTpetroleum
stocks w ith fum ing sulfuric acid, sulfonic acids
are formed, some of which rem ain dissolved in th e oil layer after
it is separated from the sulfuric acid sludge. U pon neutralizing
th e “acid oil” layer w ith alkali, oil-soluble sodium sulfonates are
produced. These “m ahogany soaps” are extracted w ith aqueous
alcohol an d are fu rth er refined to produce commercially im por­
ta n t emulsifiers.
Products of this ty p e are m arketed principally as a blend con­
sisting of approxim ately equal am ounts of oil and soap, contam ­
inated w ith sm all am ounts of inorganic m atter. T o control the
ratio of oil to soap in these m ixtures, it is necessary to estim ate a t
least one of these com ponents. T his m ay be done by separating
th e sulfonates from the hydrocarbons b y m eans of some physical
process. D istillation is not suitable because of th e relatively high
boiling points of th e hydrocarbons, coupled w ith th e unstable
character of the sulfonates a t tem peratures in th e neighborhood
of 130° C. Since neither of th e com ponents crystallizes readily
from their m ixtures or from solutions of th eir blends, crystalliza­
tion cannot be used to separate them .
Because of th e greater solubility of th e soaps in a water-alcohol
mixed solvent, unsaponifiable m a tte r o r oil is usually extracted
selectively w ith petroleum ether, from soap dissolved in 50 per
cent aqueous alcohol solution. M ost of th e oil dissolves in the
petroleum e th er an d th e aqueous alcohol retains in solution th e
m ajo r portion of th e soap. B y a system atic process of m ultiple
extractions, soap an d oil m ixtures can be sep arated com pletely
if th e proper conditions are chosen. A rchibald an d B aldeschwieler (2) have described a m ethod of analyzing petroleum sul­
fonates b y th is ty p e of extraction. Since some of th e separations
were found to be incom plete, a fte r em ploying th e standardized
procedure, th e au th o rs have described a m ethod of correcting the
oil fraction for th e soap contained in it.
T h e large num ber of extractions an d m anipulations ordinarily
required by th is m ethod are tim e-consum ing, an d severe em ul­
sion difficulties are encountered w ith some ty p es of sam ples.
Selective adsorption can be used to separate m ixtures composed
of m aterials having widely different adsorption characteristics,
ju s t as selective solvent action is used to separate m aterials of
different solubilities. T he adsorption process is often a d v a n ta ­
geous because of its convenience, speed, and freedom from emulsion
troubles. An adsorbent m u st be found which will adsorb sodium
sulfonates, since oil is n o t easily adsorbed, and th e proper solvent
an d displacing ag en t m u st be chosen. Simple adsorption, like
simple extraction, is usually inadequate for sh arp sep aration of
tw o com ponents. F ractio n atio n of some kind m u st be em ployed
to th e b est advantage to achieve q u a n tita tiv e results.
T h is p aper presents a new adsorption m ethod which has th e
advantages of speed, convenience, an d freedom from emulsion
troubles, and which has been found to be especially useful in th is
lab oratory for p la n t control work.
26
INDUSTRIAL
AND
ENGINEERING
PRELIMINARY EXPERIMENTS
Percolation experim ents showed th a t
sodium sulfonates could be selectively ad­
sorbed from n ap h th a solution, an d th a t
th ey could, in tu rn , be com pletely displaced
from th e adsorbent b y m ethanol.
APPARATUS A N D M ATERIALS
T he percolation apparatu s, shown in
Figure 1, consisted of a 250-ml. separatory
funnel, A, a 250-ml. extraction flask, B,
and a percolation tube, C. T his tu b e had
a 1.5-cm. inside diam eter and was 40 cm.
long. _ A delivery tu b e 5 cm. long and 3
mm. in inside diam eter was sealed to the
b ottom of th e percolator. In to it a glass
wool plug was packed for th e purpose of
supporting the adsorbent. J u s t before
startin g the percolation, fresh A ttapulgus
clay was packed into the percolator by
tam ping successive 5-cm. sections of clay
tig htly in place w ith a rod. T he clay was
30- to 60-mcsh size an d had been calcined
a t 482° C. (900° F.). (Silica gel of 28- to
200-mesh, from th e D avison Chemical
Corporation, and Fisher adsorption alum ina
for chrom atographic analyses, were tried as
sodium sulfonate adsorbents, b u t were found
to be less, efficient th a n A ttapulgus clay.
T he clay was supplied by th e A ttapulgus
Clay Co., 260 South B road St., Philadelphia,
Penna. A pproxim ately 35 gram s were used
for an analysis.)
T he volatility of th e petroleum n a p h th a
solvent used in these experim ents was suffi­
ciently above th a t of the oil in th e sample,
Figure 1. Per­
so th a t the solvent could be completely
colation A p p a ­
ratus
separated from the oil by evaporation.
N ap h th a of 30° to 80° C. (S6° to 176° F.)
boiling range, distilled from a paraffinbase crude oil, and A.S.T.M . precipitation n a p h th a (J) of 50°
to 130° C. (122° to 266° F .)’ boiling range, were found to be
satisfactory.
Vol. 16, No. 1
CHEMISTRY
is less th a n 0.01 gram. Usually, 15 to 30 m inutes of heating are
sufficient.
Sim ilar tre a tm e n t is applied to th e soap fraction dissolved in
m ethyl alcohol, except th a t a n oven k e p t a t 120° to 130° C. is
used to save tim e. T h e w e i g h t s of oil and soap are reported as
per cent of sam ple tak en .
TEST OF THE METHOD
Tw o tests were applied to th is m ethod. F o r th e first test, 40
grams of a refined petroleum sulfonate (sample A, T able II) were
separated into an oil and a soap fraction b y th e m ethod of Archi­
bald an d Baldeschwieler (2). S ettling tim es of from 8 to_ 16
hours were perm itted for emulsions of n a p h th a and aqueous iso­
propyl alcohol to separate sharply into two layers. T h e oil frac­
tion, however, had a n ash co n ten t of 0.14 per cent; so it was
filtered through A ttapulgus clay to produce a colorless oil which
contained no ash, and was, therefore, free of soap. A series of
m ixtures of th is oil and th e soap fraction w as prepared, and then
analyzed by th e adsorption procedure. T able I contains these
analyses and th e corresponding known blend values. Sample _5
consisted of th e extracted sulfonates themselves, w ith o u t th e addi­
tion of an y oil.
In th e second test, a sim ilar series of know n m ixtures wTas pre­
pared b y blending oil and soap fractions which had been sepa­
rated from a group of sam ples by th e adsorption m ethod. T he
adsorption analyses of these known m ixtures are also presented
in T able I.
These d a ta indicate th a t th e adsorption procedure gives ac­
curate analyses of oil-soap m ixtures of widely varying composi­
tion. As showm by th e analysis of sam ple 5, th e sodium sulfo­
nates obtained b y aqueous isopropyl alcohol extraction contained
only approxim ately 0.3 per cent of oil. T he oil fractions obtained
in these adsorption analyses were practically colorless an d free of
soap, as indicated b y th e absence of ash following th eir ignition.
PRECISION A N D COM PARISON WITH EXTRACTION METHOD
A group of five refined soap sam ples was analyzed in duplicate
to te st th e precision of th e adsorption procedure (Table II ).
These d a ta , together w ith th e results shown in T able I indicate
satisfactory reproducibility for m ost purposes.
PROCEDURE
A pproxim ately 2.0 gram s of sample are accurately weighed
into extraction flask B , and dissolved in 25 ml. of petroleum
n ap h th a. The solution is transferred to separatory funnel A,
■which is then stoppered. T he stem of the funnel is placed inside
.th e open end of percolator C, so th a t it ju s t touches the clay, as
shown in Figure 1. U pon opening th e stopcock the solution will
percolate down through th e clay, and th e stoppered separatory
-funnel will a c t as an autom atic feeding device. T h e percolate
issuing from th e bottom of th e clay colum n is caught in th e tared
extraction flask.
As soon as th e last drop of solution has entered the percolator,
th e stem and inside of the funnel are wrashed w ith petroleum
n aphtha. These w ashings are charged to th e percolator. The
top of the percolator is also w ashed clean of a n y rem n an ts of
sample. T hen 100 ml. of n a p h th a are charged to th e separatory
funnel and percolated through th e clay using th e same au to ­
m atic feeding arrangem ent described for the solution.
If the total n aphtha percolate is clear, it is set aside for evapora­
tion. If it is cloudy, indicating the presence of unadsorbed soap
o r salts, th e combined percolate is ru n through a second percola­
to r packed with fresh clay. A 100-ml. n ap h th a wash is used
as before. T he combined n ap h th a percolate in the tared extrac­
tion flask is very carefully evaporated on a steam b ath . Oil is
prevented from creeping to the outside wall of th e flask by blow­
ing a gentle stream of a ir into th e flask during th e evaporation.
While this operation is taking place, 100 ml. of absolute m ethyl
alcohol are percolated through each of the clay colum ns used to
adsorb the soaps. T he resulting percolates are collected in a
tared extraction flask, and the solvent is evaporated from the
soaps in the sam e m anner as was described for th e naphtha.
W hen practically all of the n ap h th a has been evaporated from
th e oil fraction on th e steam b ath , th e extraction flask is placed in
a n oven and kept a t 100° to 105° C. for 15 m inutes, th e n cooled
and weighed. T his process of heating in th e oven for 15-minute
periods, followed by weighing, is repeated until th e loss in weight
Table I.
Sam ple
N o.
1
2
3
4
5
6
7
8
9
10
11
12
A n alyse s of Known Blends of W hite O il and Sodium
Sulfonates by the A dso rp tion Procedure
A ccording to B lend
Soap
Oil
%
%
8 0 .4
19.6
6 2 .4
3 7 .6
5 8 .7
4 1 .3
8 2 .5
17.5
100.0
0 .0
10.1
8 9 .9
8 0 .4
19.6
3 5 .0
6 5 .0
5 0 .0
5 0 .0
3 5 .0
6 5 .0
7 8 .7
2 1 .3
8 9 .4
10.6
C om position of S a m p le F o u n d b y A nalysis
Soap
Oil
%
%
8 0 .0
2 0 .0
3 8 .0
6 2 .2
5 9 .2
4 0 .9
8 2 .5
17.6
100.0
0 .3
8 9 .7
10.6
19.8
8 0 .1
3 4 .7
6 5 .0
4 9 .7
4 9 .9
6 4 .2
35.1
7 8 .2
2 1 .8
11.2
8 8 .7
------- ,
D e v iatio n
Soap
Oil
%
%
-0 .4
+ 0 .4
-0 .2
+ 0 .4
- 0 .4
+ 0 .5
0 .0
+ 0 .1
+ 0 .3
0 .0
-0 .2
+ 0 .5
+ 0 .2
-0 .3
-0 .3
0 .0
-0 .3
-0 .1
+ 0 .1
-0 .8
-0 .5
+ 0 .5
-0 .7
+ 0 .6
S am ples 1 to 5 w ere blended from com ponen ts p rep ared b y aqueouR
isopropyl alcohol e x tractio n . S am ples 6 to 12 w ere blended from compon en ts p re p a re d b y A tta p u lg u s clay a d so rp tio n .
Table II.
Sam ple
A
Duplicate A dsorption A n alyse s of Refined Soaps
C om position b y A nalysis
Soap
Oil
%
%
4 6 .4
5 3 .5
4 6 .5
5 3 .8
D e viation
Soap
Oil
%
%
0.1
0 .3
B
5 0 .6
4 9 .6
46 .2
4 6 .0
1.0
0 .2
C
5 0 .4
5 0 .8
4 7 .0
4 7 .3
0 .4
0 .3
D
5 1 .2
5 1 .3
4 9 .0
48.7
0.1
0 .3
4 9 .0
4 9 .8
5 0 .3
4 9 .3
0 .8
1.0
January 15, 1944
Table III.
Sam ple
F
G
H
I
J
ANALYTICAL
Comparison of A n alyse s of Refined Soaps b y the
A dsorption and Extraction M ethods
/----------- ---------------C om position b y A n aly sis—
A d so rp tio n M eth o d
E x tra c tio n M eth o d
Soap Oil
T o ta l
Soap Oil
T o ta l
%
%
%
%
%
%
5 6 .8 4 2 .7
9 9 .5
5 6 .6 4 3 .1
9 9 .7
6 0 .3 3 9 .1
9 9 .4
5 9 .8 3 9 .6
9 9 .4
5 5 .7 4 4 .4 100.1
5 4 .0 4 4 .8
9 8 .8
5 6 .6 2 4 .8
8 1 .4
5 8 .1 2 3 .4
8 1 .5
6 7 .4 3 2 .8 100.2
6 6 .6 3 3 .4 1 0 0 .0
Table IV .
D e v iatio n
Soap
Oil
%
%
+ 0 .2
-0 .4
+ 0 .5
-0 .5
+ 1 .7 - 0 . 4
-1 .5
+ 1 .4
+ 0 .8
-0 .6
Resin-D isplacing Power of a Series of Trial Eluants
T ria l E lu a n t
W eig h t of M a te ria l
D isplaced
Grama
A cetone
E th y l a ce ta te
D ie th y l e th e r
C hloroform
E th y len e dichloride
N itro m eth a n e
C arb o n te tra c h lo rid e
C arb o n disulfide
1.4 9 0
0 .7 8 8
0 .7 8 6
0 .0 5 2
0 .6 1 2
0 .5 6 2
0 .1 3 6
0 .0 3 2
R atio of M ate rial
D isplaced to
R esin on C lay
1 .9
1.0
1. 0
0.8
0.8
0 .7
0. 2
EDITION
27
prepared clay column following th e m ethod described in th e a d ­
sorption procedure. A fter th e clay w as w ashed w ith 100 ml. of
petroleum naph th a, 150 ml. of a trial resin elu an t were percolated
through th e colum n. T his percolate was collected separately,
and th e w eight of m aterial displaced w as determ ined by the
m ethod described for soaps. T he ratio of this w eight to 0.781
gram, th e am o u n t of resin in th e percolator charge, was calcu­
lated.
Table IV contains th e d a ta obtained in th is m anner for a series
of organic liquids.
Im p u re solvents were percolated through an excess of A tta p u l­
gus clay to rem ove sm all am ounts of w ater and o th er easily
adsorbed im purities.
Acetone displaced practically all of th e resin and th e sulfonates.
E th y l acetate and diethyl e th er displaced a q u a n tity of m aterial
practically equal to th e w eight of resin adsorbed. Since th e dis­
placed m aterial contained negligible am ounts of ash, it was as­
sum ed to be resinous m atter. E th y l acetate an d diethyl ether
were chosen as being th e best resin eluants of th e group tested.
0.0
MODIFIED ADSORPTION PROCEDURES
F ive o th er refined soap sam ples were analyzed by b o th th e ex­
traction m ethod of A rchibald an d Baldescbwieler (2) and th e ad ­
sorption procedure an d the results are compared in T able II I.
Samples F and G were from th e same m anufacturer, b u t samples
H , I, and J came from three different sources. Sample I was
found to contain 18.6 per cent of m aterial which was volatile a t
130° C.
T he agreem ent betw een th e tw o m ethods appears to be sa tis­
factory. T h ey are probably capable of giving equal accuracy and
precision on sam ples of this type.
SO A PS CONTAINING RESINS
Soaps th a t are produced in th e course of medicinal w hite oil
m anufacture, as described above, have been found to be p ra c ti­
cally free of resinous (oxygenated hydrocarbon) components.
T here are some products on th e m arket, however, which do con­
tain resinous m aterials. These sam ples, containing resins in
addition to sodium sulfonates an d oil, present a more difficult
analytical problem for both th e extraction and th e adsorption
m ethods, th a n th e sim pler case of soap an d oil m ixtures. W hereas
th e extraction m ethod usually gives som ew hat high results for oil
on these sam ples, th e adsorption m ethod gives high results for
soap, due to th e adsorption of resins as well as sodium sulfonates
on th e clay.
Since there is a difference in m olecular stru ctu re and polarity
betw een resin and sodium sulfonate molecules, there should also
be an appreciable difference in th eir adsorption affinities. If all
the sulfonate molecules in a given sam ple are more strongly ad­
sorbed by clay th a n all the resin molecules, th e la tte r can be se­
lectively displaced, by th e proper eluant, from a solid on which
b oth com ponents are adsorbed. B y a series of tria l experiments,
ethyl acetate was found to be a good resin eluant. Several pu b ­
lications (S, 4, 5) have been helpful in choosing eluants.
C H O O SIN G THE RESIN ELUANT
A sam ple of soap, prepared by treatin g a very naphthenic pe­
troleum stock w ith concentrated sulfuric acid, followed by neu­
tralization w ith sodium hydroxide, was separated into its chief
com ponents: oil, resins, and sulfonates. T his was done by ex­
tracting the soap from th e oily m a tte r (oil-resin m ixture) by the
aqueous isopropyl alcohol extraction m ethod (2), and th en sepa­
rating th e resin from the oil by percolating a petroleum nap h th a
solution of these com ponents through a column of A ttapulgus
clay. T he oil-free resin was recovered from th e clay b y displac­
ing it w ith absolute m ethyl alcohol.
A know n te s t m ixture consisting of 34.5 p er cent oil, 33.1 per
cent resin, and 32.4 per cent soap was prepared and then dis­
solved in petroleum naphtha, an d 25 ml. of this solution, contain­
ing 2.359 gram s of th e test m ixture, were percolated through a
To check these conclusions, and to te st a modified adsorption
procedure for soaps of th is type, a group of m ixtures containing
known am ounts of oil, resin, and sulfonates was analyzed. The
procedure em ployed was th e sam e as th e one previously described,
except th a t after th e 100-mi. petroleum n a p h th a wash was per­
colated through the clay, 100 ml. of th e resin elu an t were perco­
lated in th e sam e m anner. T h e m ethyl alcohol percolation fol­
lowed th a t of th e resin eluant. T he results of these analyses are
shown in T able V.
Table V .
A n alyse s of Known Blends of O i l , Resin, and Sodium
Sulfonates b y a M odified A dsorption Procedure
Sam ple C om position of K now n
N o.
Oil
R esin
Soap
1
I
1
2
3
4-
%
%
%
%
%
%
3 4 .0
3 4 .6
3 4 .0
5 5 .0
4 9 .8
5 9 .2
3 3 .0
3 3 .0
3 3 .0
5 .0
1 5 .5
10.9
3 2 .4
3 2 .4
3 2 .4
4 0 .0
3 4 .7
2 9 .9
3 4 .0
3 4 .8
3 4 .2
5 7 .3
5 0 .5
6 0 .5
2 0 .4
3 2 .0
60 .3
5 .7
14.5
4 1 .3
3 3 .0
Table V I.
Sam ple
1
2
3
4
5
6
C om position b y A nalysis
Oil
R esin
Soap
11.1
8.8
3 8 .4
3 4 .5
2 9 .5
R esin E lu a n t
C hloroform
D ie th y l e th e r
A cetone
D ie th y l e th er
E th y l a ce ta te
D ie th y l e th e r
Comparison of A n alyse s of Soaps Containing Resins by
the Extraction and A dsorption M ethods
E x tra c tio n M ethod
Oily
Soap m a tte r T o ta l
%
%
%
9 .9 8 9 .7
9 9 .6
20.1 8 0 .3 100.4
3 0 .5 7 0 .5 101.0
37.1 6 3 .4 100.5
4 2 .9 5 7 .4 100.3
4 9 .8 5 1 .4 101.2
A d sorption M ethod
O ily
Soap m a tte r T o ta l
%
%
%
9 .8 9 0 .0
9 9 .8
99.1
19.9 7 9 .2
9 9 .4
2 9 .7 6 9 .7
9 9 .9
3 6 .8 6 3.1
4 2 .5 5 7 .9 100.4
5 0 .2 4 9 .8 100.0
D e v iatio n
Oily
m a tte r
Soap
%
-0 .1
-0 .2
-0 .8
-0 .3
-0 .4
+ 0 .4
%
+ 0 .3
-1 .1
-0 .8
-0 .3
+ 0 .5
-1 .6
D iethyl eth er and eth y l acetate were again found to be th e best
resin eluants, an d th e modified adsorption procedure using either
of these liquids gave analyses of satisfactory accuracy.
In th e analysis of petroleum sulfonates containing resins, it is
usually sufficient to determ ine th e sodium sulfonate con ten t and
th e to ta l am o u n t of oily or inactive m a tte r in th e sam ple. T he
aqueous isopropyl alcohol extraction m ethod (2) accomplishes
this on sam ples of this type, b u t it is tim e-consum ing. T h e ad­
sorption m ethod can also be applied to obtain only th e soap and
oily m a tte r in these sam ples by sim ply replacing th e petroleum
n ap h th a, in th e original procedure, w ith a suitable resin eluant.
T h e first percolate will then contain b o th th e oil and th e resins
and th e m ethyl alcohol percolate will again contain th e sulfo­
nates.
A series of soaps prepared from a very naphthenic stock, as
described above, an d containing varying am ounts of oil and resins,
was analyzed by th e isopropyl alcohol extraction m ethod (#) and
28
INDUSTRIAL
AND
ENGINEERING
also by the modified adsorption m ethod. E th y l acetate was used
as the solvent for th e oily m atter, and m ethyl alcohol as th e eluant
for the sodium sulfonates. A comparison of th e analyses is given
in T able V I. B oth m ethods gave essentially th e sam e analyses,
b u t the adsorption procedure was more rapid an d convenient.
A know n blend containing 49.8 per cent oil, 15.5 per cent resin
(65.3 per cent oily m atter), and 34.7 per cent sulfonates was an a­
lyzed by the eth y l acetate-m eth y l alcohol adsorption procedure;
66.2 per cent oily m a tte r and 34.0 per cent soap was obtained.
In the case of petroleum sulfonates which are n o t know n to be
free of resins, it is therefore advisable to introduce an ethyl ace­
ta te percolation a t th e end of th e petroleum n a p h th a wash, and
before the m ethyl alcohol percolation, to te s t for th e presence of
resins. If th e ethyl acetate displaces m aterial which leaves
practically no ash after ignition, th e presence of resins is indi­
cated, and one of the modified adsorption procedures described
above should be adopted.
CHEMISTRY
Vol. 16, No. 1
ous alcohol from caustic neutralized acid oil in th e m anufacture of
medicinal w hite oil, it will be necessary to introduce an eth y l ace­
ta te percolation as a te s t for th e presence of resins.
T h e adsorption procedure can be applied to crude oil-soluble
sodium sulfonate products containing appreciable am ounts of
w ater and inorganic salts, after first rem oving these components.
T his rem oval can be accomplished very conveniently in th e d eter­
m ination of salt and w ater by th e usual m ethods. F or salt, a
precipitation-type m ethod such as A .S.T.M . M ethod D91-40
(I) m ay be used, and for w ater a modified D ean and S tark m ethod
is very convenient.
A dsorption m ethods should find applications in grease and as­
p h alt analyses, and in th e separation of m ixtures containing or­
ganic salts or acids an d hydrocarbons.
ACKNOWLEDGMENT
T he au th o r wishes to express his appreciation to S. W . Ferris,
W . K . Griesinger, and C. E . H eadington for advice an d assistance
in th e preparation of th is paper.
DISCUSSION
T he chief advantages of th e adsorption procedure are freedom
from em ulsion difficulties, rapid convenient physical operations,
and sharp separations of oil and sodium sulfonate com ponents.
Experience in this laboratory has dem onstrated th a t a given
adsorption procedure m ay fail to give correct analyses on differ­
e n t types of sam ples. T he procedure used will depend upon th e
adsorption characteristics of th e com ponents in th e refined oilsoluble sodium soap. Unless th e pro d u ct to be analyzed is
known to consist entirely of oil and sodium sulfonates, as is nor­
mally the case, th e m ahogany soaps which are ex tracted b y aque­
LITERATURE CITED
(1) Am . Soc. T estin g M aterials, C om m ittee D -2, R ep ort 1941,
M ethod D 91-10, p. 236.
(2) Archibald, F . M ., and Baldcschwieler, E . L., I n d . E n o . C h e m .,
A n a l . E d ., 13, 6 08 (1 9 4 1 ).
(3) Strain, H . H ., "Chrom atographic Adsorption A n alysis” , N ew
York, Interscience Publishers, 1942.
(4) Strain, H . H ., I n d . E n o . C h e m ., A n a l . E d ., 14, 245 (1942).
(5) Zechm eister, L., and C holnoky, L ., “P rinciples and P ractice of
Chrom atography” , N ew York, John W iley & Sons, 1941.
Quantitative Determination of c/-Galactose by Selective
Fermentation
W ith S P ecial Reference to Plant M ucilages
L O U I S E. W IS E a n d J O H N W . A P P L I N G
The Institute of Paper Chemistry, A p p le to n , W is.
A simple method has been developed which permits the
determination of small amounts of d-galactose in the
presence of mannose, glucose, fructose, xylose, arabinose,
and glucuronic acid with an accuracy of 9 2 to 9 8 per ce n t
ft depends on differential fermentations with two yeasts,
Saccharomyces carlsbergensis (N .R .R .L . N o . 3 79) which
ferments galactose, and S. bayanus (N .R .R .L . N o . 9 6 6 )
which leaves galactose unfermented. The yeasts have little
action on xylose, arabinose, or glucuronic acid, and these
C E N T L Y , interest in th e determ ination of galactose has
R Ebeen
revived because of th e newer technological applications
of certain m annogalactan mucilages. T he estim ation of d-galac­
tose, w hich th u s assum es a new im portance, has alw ays presented
difficulties, especially when o th er carbohydrates were present in
q u an tity . T he v an der Ila a r modification of th e K ent-T ollensC reydt m ethod (4), w hich depends on th e oxidation of galactose
to m ucic acid, is not q u an titativ e. O nly when rigorous precau­
tions are taken, and when relatively large am ounts of galactose
are present, does the procedure give fairly accurate results (/.)).
E ver since the earlier investigations of ICluyver (7 ), th e quan­
tita tiv e estim ation of d-galactose by ferm entation w ith certain
yeasts has interested chem ists and microbiologists. K luyver
compounds do not interfere with the determination. Reduc­
ing values of galactose, mannose, and d-glucurone were
determined using the M unson-W alker method of analysis.
The fermentation techniques were successfully applied
to the hydrolysis products of lactose and to certain plant
mucilages. The possible application of the method to
galactose in the presence of galacturonic acid is being
studied with a view toward its use in the analysis of other
natural products.
found th a t galactose was ferm ented by tw o varieties of Saccharomyces cerevisiae and by a “m ilk sugar y e a st". He also
showed th a t Schizosaccharomyces pombe, Torula monosa, an d T.
daltila did n o t ferm ent galactose. On th e basis of these differ­
ences, he proposed a proxim ate m ethod for th e microbiological
determ ination of galactose in th e presence of o th er sugars.
Among those who used (and modified) K lu y v er’s procedures were
Schm idt, Trcfz, an d Schnegg (11), ITopkins, Peterson, an d F red
(6), Sherrard and Blanco (IS ), K u rth an d R itte r (9), K u rth (8),
S co tt an d W est (12), H arding, N icholson, and G ra n t (5), and,
very recently, M enzinsky (10). T he last au th o r showed th a t the
strain of Saccharomyces cerevisiae which he used in galactose fer­
m entation required preconditioning by culturing th e y east on a
January 15, 1944
ANALYTICAL
galactose-containing m edium . W ith o u t such p retre a tm en t th e
y east was unable to utilize galactose.
Besides th e yeasts m entioned above, th e following are known
to ferm ent galactose: Saccharomyces pastorianus (S), S . m arxianus (3, 6), and S . fragilis {10). O ther yeasts know n to ferm ent
th e common hexoses other th a n galactose are S . produdivus (3),
S . apiculatus (3), and “Honey B ” yeast (3). These lists are n o t
exhaustive.
K luyver used the evolution of carbon dioxide as a m easure of
ferm entable sugars. L a te r investigators (3, 8) showed th a t a
q u an titativ e m easure of the reducing value simplified th e analysis
of ferm ented sugar solutions.
N otw ithstanding th e extensive w ork on th e selective fermen­
ta tio n of galactose, the results of relatively few experim ents w ith
pure sugar m ixtures have been published an d no a tte m p t has
been m ade to determ ine w hether th e m ethods could be applied in
th e presence of uronic acids. T h e lim itations and general appli­
cability of th e ferm entation m ethods are, therefore, indeterm inate.
T he presen t stu d y shows th e usefulness of th e microbiological
m ethod.
EXPERIMENTAL
T he objects of the presen t investigation were to examine, more
critically th a n heretofore, th e application of differential fermen­
tatio n s to know n sugar m ixtures, noting th eir lim itations, and
to develop a proxim ate biochem ical m ethod for th e determ ina­
tion of galactose in the hydrolyzates obtained from mucilages
and hemicelluloses.
Several strain s of S. cerevisiae, cultured in th e laboratories of
T he In s titu te of P ap er C hem istry, h ad been shown to ferm ent
galactose q u an titativ ely in 1937 b y K u rth (3). In 1942, how­
ever, q ualitative experim ents w ith these sam e strain s showed
Figure 1
EDITION
29
th a t none ferm ented 1 p er cent galactose solutions completely
w ithin 190 hours. I t is evident th a t strain s of S . cerevisiae m ay
lose th eir potency as galactose ferm enters w ith tim e, unless spe­
cial precautions are tak en to recondition them .
Torula datlila appeared a t first to give fairly prom ising results
as a nonferm enter of galactose. A bout 92 to 95 p er cent of the
galactose (in a 1 per cent solution), when treated w ith Torula
datlila, could be recovered a fte r 2 days a t 30° C., b u t only about
80 per cent rem ained a fte r a 5-day ferm entation period. W hen­
ever th e concentrations of galactose dropped to 0.1 p er cent, the
sugar was rapidly destroyed.
T he use of these organisms was discontinued in favor of tw o in­
teresting yeasts obtained from L. J . W ickerham , N o rth ern R e­
gional R esearch L aboratory, Peoria, 111. These were Saccharo­
myces carlsbergensis var. mandshuricus, N .R .R .L . N o. 379 (origi­
nally obtained from C harles N . F rey of th e Fleischm ann L abora­
tories in 1940), a highly ferm entative strain acting on dextrose,
galactose, and some of th e common disaccharides, and S . bayanus,
N .R .R .L . No. 966 (also originally obtained from F rey in 1940),
which was known to ferm ent dextrose, sucrose, and m altose, b u t
which, qualitativ ely a t least, had no effect on galactose. N either
yeast had (during th e p ast three years) been k e p t on galactose
media.
These organism s proved entirely satisfactory. N either had
more th a n a slight effect on arabinose, xylose, an d glucuronic
acid. No. 379 ferm ented d-glucose, m annose, fructose, and
galactose alm ost q u an titativ ely w ithin 48 hours. N o. 966 fer­
m ented the first three readily w ithin th e sam e tim e period and
showed no action w hatsoever on galactose. These differences led
to th e developm ent of a satisfactory proxim ate m ethod based on
differential ferm entations, in which th e M unson-W alker reduc­
tion m ethod (3) was used w ith o u t m odification. In aliquot
portions, ta k en from ferm entation m ixtures, 20 to 125 mg. of
galactose could be determ ined w ith an accuracy of 92 to 98 per
cent, even when o th er sugars were present originally in g reat ex­
cess.
T he yeasts were m aintained in good condition by m onthly
transfers on glucose ag ar (Bacto-D extrose agar, dehydrated).
T hey have shown no decrease in potency over a period of 8
m onths. A gar sla n t cultures, 2 to 7 days old, were used for the
preparation of th e suspensions required in inoculating th e sugar
solutions. T he following procedure w as th e sam e for either
yeast.
A bout 2 ml. of sterile w ater were pipetted into th e tube con­
taining th e agar slan t culture, and th e surface grow th was re­
moved gently oy m eans of th e pipet, which also served to stir the
suspension briskly. F o r each b o ttle slant of glucose agar re­
quired, 0.5 ml. of th e suspension was rem oved and spread over
th e surface by tilting th e b o ttle back and forth. Eight-ounce,
narrow -m outh, square, flint glass bottles w ith molded screw
caps were used as containers for sterile w ater, y east suspensions,
and b o ttle slants. Thereupon, th e slants were incubated for
about 48 hours a t 30° C.
One b ottle slan t easily furnished
enough inoculum for four subsequent sugar analyses, because a
dense grow th of y east cells coated th e agar surface a t th e end of
th e 48-hour period.
A bout 10 ml. of sterile w ater were added to th e b o ttle slant,
and th e b ottle was tilted back and forth to loosen th e growth.
A bout 5 ml. of th e dense y east suspension were p ip etted into a
sterile dilution bottle, and 20 to 30 ml. of sterile w ater were added.
T he exact am ount of w ater depended upon the tu rb id ity shown
b y a Ceneo-Sheard-Sanford photelom eter. Suspensions whose
readings fell w ithin th e range of 10 to 16 on this instrum ent,
when distilled w ater gave a reading of 90, subsequently ferm ented
sugar m ixtures satisfactorily. If the initial photelom eter reading
fell below 10, more w ater was added and thoroughly mixed w ith
th e suspension. W henever th e photelom eter reading exceeded
16, more y east suspension w as added. In practice it was found
b e tte r to w ork w ith too dense th a n w ith too light a suspension.
T h e density range listed above was shown b y p late counts to
approxim ate 35,(KK>,000 cells p er ml. T h is density also corre­
sponded roughly to th a t of barium sulfate suspension prepared by
mixing 5 ml. of a stock solution of 10 gram s of barium chloride di­
h y d rate p er liter of w ater w ith 55 ml. of 1 p er cent sulfuric acid
and allowing th e m ixture to sta n d a t least a week in a sealed con-
INDUSTRIAL
30
Table I.
ENGINEERING
A c tio n of Yeasts on Pentoses after a 6 -D a y Incubation
Period
Pentoses P rese n t
in A liq u o t
M g.
12.5
12.5
12.5
12.5
12.5
12.5
AND
(xylose)
(xylose)
(xylose)
(arabinose)
(arabinose)
(arabinose)
T re a tm e n t
C o n tro l
O rganism
O rganism
C o n tro l
O rganism
O rganism
CuaO O b tain ed in
M unson-W alker
D e term in a tio n s
Mg.
2 6 .3
2 0 .8
2 2 .3
2 8 .6
2 6 .7
2 6 .7
379
966
379
966
tainer. I t was found expedient to use 25 ml. of sugar solutions
in th e ferm entations and to carry o u t all experim ents in 125-ml.
Erlenm eyer flasks. T he to tal reducing sugar in such solutions
never exceeded 2 per cent, and the galactose concentration was
ordinarily k ept w ithin th e range of 40 to 250 mg. per 25 ml. To
the sugar solution were added 15 ml. of a filtered y east extract.
(Red S ta r starch-free yeast cake was mixed w ith sufficient dis­
tilled w ater to yield a 10 per cent suspension. T his was heated
for 1 hour in an Arnold sterilizer a t about 100° C. and subse­
quently for 20 m inutes a t 15 pounds’ pressure (a t 121° C.).
T he cooled suspension was filtered several tim es through fluted
filter paper using Cellite to clarify the solution. O rdinarily the
y east extract formed a slightly turbid solution. T his was dis­
pensed in 80-ml. portions into Erlenm eyer flasks, w hich were
plugged w ith cotton and heated a t 15 pounds’ pressure for 20
m inutes. T he cooled flasks of sterile y east extract can be stored
for m onths w ithout change in a refrigerator.]
T h e sugar and y east extract m ixtures, w hich showed a p H of
a bout 5 to 6 (alkacid paper), were then sterilized a t 15 pounds’
pressure for 15 m inutes, cooled to about 30° C., inoculated under
aseptic conditions w ith 10 ml. of the appropriate y east suspen­
sion, and incubated a t 30° C. for a minim um of 48 hours.
Tlie ferm entations were always run concom itantly in pairs,
under identical conditions, one flask being inoculated w ith organ­
ism 966 and the other w ith organism 379. T hree or four times
during the incubation period, the flask was ro tate d gently to bring
the bottom yeasts into intim ate con tact w ith the sugar solution.
A t the end of the fennentatio n period, each solution was diluted
to 100 ml. w ith distilled w ater, thoroughly mixed, and filtered
through tw o No. 50 W hatm an filter papers. Twenty-five to 50ml. aliquot portions of the clear, pale yellow filtrates were tak en
for analysis by th e usual M unson-W alker technique. (C are
m ust be tak en to digest th e asbestos used in Gooch crucibles
thoroughly w ith h o t Fehling solution, an d w ith concentrated
n itric acid. T o prevent la te r clogging of Gooch crucibles, such
treatm en ts should be continued u n til asbestos filter pads perm it
the rapid filtration of h o t Fehling solution containing th e filtered
y east extract referred to above.)
T h e w eight of cuprous oxide resulting from the ferm entation
w ith organism 379 was subtracted from th a t obtained by th e use
of organism 966. T he galactose equivalent was calculated by the
use of the galactose-cuprous oxide graph (Figure 1) draw n from
d a ta obtained experim entally with pure galactose. (The mannose
values given in Figure 1 were obtained from pure dknannose;
the glucose values were taken from M unson and W alker’s tables.)
T he effects of S . carlsbergensis an d S . bayanus on sm all am ounts
of xylose an d arabinose were shown to be relatively unim portant,
even a fter a n incubation period of 144 hours instead of th e usual
48-hour period. T his is indicated in T able I (and T able II).
Inasm uch as the n e t reducing values in all differential galactose
determ inations were obtained b y su b tractin g th e w eight of cu­
prous oxide found a fter ferm entation w ith organism 379 from
th a t found after a ferm entation w ith organism 966, th e over-all
error resulting from th e presence of pentoses is negligible. F u r­
therm ore, th e au th o rs’ q u an tita tiv e ferm entation periods seldom
exceeded 2 days, w hich presum ably would resu lt in a lessened
action on th e pentoses.
Because glucuronic acid m ay be a m inor com ponent of hemicellulose hydrolysis, the effect of d-glucurone in th e galactose
analysis w as determ ined. Figure 2 shows th e reducing values
in milligrams of cuprous oxide p lotted ag ain st th e weights of
glucurone (m elting point 174-5° C.) tak e n for analysis. Over a
fairly wide range, this curve is practically coincident w ith th a t
of glucose. T he action of organisms 379 and 966 on glucurone
%
CHEMISTRY
Vol. 16, No. 1
was found to be very slight an d th eir over-all effect in differential
ferm entations of galactose was alm ost negligible. T his is indi­
cated in th e la st row of T able II.
O rientating experim ents were also carried o u t w ith purified
galacturonic acid, which rem ains v irtually u n attack ed by either
organism, even when only sm all am ounts of th e acid are present
in th e usual ferm entation m ixture. T he reducing value (M unson-W alker) of 12.5 mg. of galacturonic acid was shown to be
23.0 mg. of cuprous oxide; when neutralized, sterilized, inocu­
lated, and incubated in th e usual m anner, 20.7 mg. and 20.2 mg.
of cuprous oxide were obtained w ith organisms 379 and 966.
H ere again th e “ error” is cancelled.
T he value an d lim itations of th e selective differential fermen­
tatio n s in th e determ ination of galactose are clearly shown in
T able II . T h e first and last columns of th is tab le should be com­
pared. Invariably, in th e higher concentrations, galactose shows
a slight b u t persistent residual reduction a fte r ferm entation w ith
organism 379. W hether this is due to very sm all am ounts of un­
ferm ented galactose or (w hat is m ore probable) to th e slight re­
ducing pow er of th e products of th e ferm entation is n o t known.
T he error is never very appreciable, b u t it accounts for th e fact
th a t galactose recoveries are usually som ew hat low. Organism
966 is w ith o u t effect on galactose.
Fisure 2
In order to determ ine w hether galactose could be satisfactorily
determ ined in th e hydrolyzate of a disaccharide, pure lactose was
heated w ith 2 per cent sulfuric acid a t th e boiling po in t for 2.75
hours. T he solutions wrere nearly neutralized w ith solid sodium
carbonate (pH a b o u t 5, alkacid te s t paper). A liquot portions
representing, in each ease, 125 mg. of lactose (hydrate) were sub-
January 15, 1944
Table II.
1
G alactose
T ak en
M g.
125
125
125
150
(F erm en ted
72 hrs.)
ANALYTICAL
EDITION
Galactose Determination A lo n e and in Mixtures of Pure Sugars
2
O th e r
C om pon en ts
M o.
N one
N one
N one
N one
3
CusO O b tain ed
a fte r F e rm e n ­
ta tio n w ith
N o. 966
Mg.
24 7 .0
25 0 .1
2 4 7 .3
29 2 .6
N one
N one
M annose 125
M an n o se 125
6 2 .5
45
25
20
M annose
M annose
M annose
M annose
62.5
80
100
230
N one
Negligible
<1
126.6
9 1 .2
4 9 .5
37 .1
4°
C ujO O btained
a fte r F e rm e n ­
ta tio n w ith
N o. 379
M g.
2 .6
5 .7
4 .5
6 .0
N one
Negligible
<1
1 .7
1 .6
0 .5
N one
5
G alactose
N et
R ecovered
W eight of (C alcd. from
C ujO ,
5, b y U se of
3 -4
Figure 1)
Mo.
24 4 .4
24 4 .4
24 2 .8
286.6
N one
N one
M o.
123.5
123.5
123.0
146.5
N one
N one
124.9
8 9 .6
49
3 7.1
61 .5
4 2 .5
2 3 .5
18.5
40
( M annose 40
■<G lucose 40
( X ylosc 5
88
8 .8
(due to xylose)
7 9 .2
3 7 .5
40
( M annose 20
<G lucose 20
( X ylose 45
173
9 2 .0 6
(due to xylose)
8 1 .0
3 8 .5
1 .6
4 9 .6
31
by control experim ents w ith pure sugar solutions
th a t had been tre a te d w ith 2 p er cent sulfuric
acid for 6 hours. One h undred m illigram s of
m annose yielded 217 mg. of cuprous oxide be­
fore an d 2 1 0 mg. of cuprous oxide a fte r acid
treatm en t. T he m annose reversion p ro d u c ts,'
however, h ad no effect on th e galactose d eter­
m ination. T he acid -treated m annose ferm ented
com pletely w ith o u t reducing value. G alactose
appeared to be unaffected by th e action of 2 per
cent sulfuric acid; 1 0 0 mg. of galactose yielded
200.5 mg. of cuprous oxide before an d 201.0 mg.
of cuprous oxide after acid treatm en t.
T able I I I gives th e yields of galactose (cal­
culated as galactan on th e oven-dry, ash-free
basis) of several endosperm mucilages, some of
which have d istin ct technological interest.
DISCUSSION
Assuming a m odicum of microbiological con­
trol, th e relative sim plicity of th e differential
ferm entation m ethod for galactose has obvious
f F ru c to se 3 0 .0
30
6 0 .5
N one
6 0 .5
2 9 .0
\ M annose 7 5 .5
advantages over th e older m ucic acid procedure.
3 7 .5
G lucurone 12.6
101 .5 5
2 5 .7 5 c
7 5 .8
3 6 .0
I t requires less m aterial for analysis, less a tte n ­
(duo to glucurono)
tion on th e p a rt of the analyst, and is less subject
« U nless otherw ise s ta te d , figure in colum n 4 rep resen ts e rro r d u e to in com plete fe rm e n ta ­
to fluctuations w ith slight variations in technique.
tio n of galactose o r to presence of sm all a m o u n ts of red u cin g su b sta n c es a m ong galactose
fe rm e n ta tio n p ro d u c ts.
I t also appears to be more accurate.
b R ecovered (using A llih n ’s fa c to r), 40.7 m g. of xyloBe.
S . carlsbergensis (organism N .R .R .L . N o. 379)
c R ecovered, 13 mg. of gluourone.
_____
requires no reconditioning to galactose, such as
th a t required by certain strain s of S . cerevisiae. Organism N .R .R .L . N o. 379 ferm ents
jected to differential ferm entation. T h e galactose values found
galactose readily, despite th e fa c t th a t th is sugar has n o t been
were 62.5 and 62.5 mg. ( o 126.5 and 126.6 mg. of cuprous oxide);
used as a n u trie n t in its culture. In th e a u th o rs’ brief experience,
th e calculated value is 62.5 mg. of galactose.
S . carlsbergensis does n o t lose its potency on transfer. I t was as
T he effect of fructose on th e galactose determ ination is neg­
active in ferm enting galactose after 8 m onths of culture as it was
ligible. Fructose, when present in large am ount, shows a per­
on receipt from Peoria. S . bayanus, although a powerful hexose
sistent reducing value a fter ferm entation w ith eith er organism.
ferm enter, leaves galactose practically untouched. T h is differ­
Inasm uch as these copper values were practically identical for
ence in behavior should m ake for a n ideal com bination of m icro­
b oth N o. 379 and No. 966, th e errors cancelled each other. In
organisms. On th e oth er hand, th e galactose ferm entation by
a set of ferm entations w ith fructose alone, 250 mg. of this sugar
N o. 379 leads to products w hich have a slight reducing value (as
yielded 7.1 mg. of cuprous oxide a fter ferm entation w ith No. 966,
shown in T able II ). A lthough v ery slight, th e error th u s caused
an d 6.9 mg. of cuprous oxide a fte r tre atm e n t w ith No. 379.
m u st be tak en into account. T he reduction is also su b ject to
Acid hydrolyzates of polysaccharides (such as gums or hemim inor fluctuations. In general, it leads to galactose values th a t
celluloses) are ordinarily neutralized w ith purified barium car­
are slightly low.
bonate. To avoid the introduction of barium ion, which m ay or
T he usefulness of the m ethod in its application to certain gums
m ay n o t be deleterious to yeasts, sodium carbonate w as used in
and
mucilages is m anifest. In the case of th e endosperm muci­
lowering th e acidity of th e sugar solutions. T hese were never
lages, independent determ inations of m annose (7), coupled with
rendered alkaline. Alkacid te st paper showed th a t th e p H was
th e figures given in T able I I I , account for 94 to 97 p er cent of
a b o u t 5 to 6 , w hich experience had shown to be satisfactory for
th e to ta l hydrolyzates. T he possible application of th e galactose
y east ferm entations. R elatively large am ounts of sodium sul­
fate had practically no effect on eith er th e ra te of ferm entation of
m ethod to th e hydrolysis products of pectins, in w hich galacthe sugar or on the results of th e M unson-W alker determ ination.
turonic acid residues predom inate, is under investigation. How25
( F ru cto se 25.0
\ M annose 62.5
5 1 .2
T he above analytical m ethod was applied to a
series of m annogalactan mucilages isolated from
the endosperm s of various seeds b y extraction
w ith hot w ater, filtration, and precipitation w ith
ethanol. T he dried mucilages were hydrolyzed by
boiling for about 6 hours w ith 25 ml. of 2 per
cent sulfuric acid. T he solutions were cooled,
brought to a p H of 5 to 6 writh solid sodium car­
bonate, and subjected to the differential ferm enta­
tion w ithout rem oving the solution from th e 125-ml.
flask. T h e analyses were made as usual on 25- or
50-ml. filtered aliquots tak en from th e ferm enta­
tion m ixtures th a t had been diluted to 1 0 0 ml.
T he sugar yields found in such hydrolyses should
be considered m inim al values. Significant am ounts
of m annose were lost on hydrolysis, b u t galactose
appeared to be largely unaffected. T his was shown
2 3 .8
Table III.
Galactan Content of Plant M ucilages
M ucilage
G a lac ta n , %
M ucilage
G a lac ta n , %
G uar 1
(Cyam opsis tetragonalobus)
3 7 .8 ,3 8 .2 °
3 7 .4 ,3 7 .6 °
A rab o g alactan
(from W . L archw ood)
7 9 . 4 , 7 9 .5 »
80.1
P alo v erd e
(C ercidium torrcyanum )
2 1 .3 ,2 1 .9 «
T a ra
(Cacsalpinia spinosa)
2 6 .2 , 26 .4 «
1 8 .2 ,1 8 .9 °
H uizache
( C aesalpinia cacaloca)
2 7 . 7 , 27.3»
2 6 .2 ,2 6 .8 °
2 5 .4 ,2 5 .6 °
Sophora japonica
G uar 2
3 3 .8 , 3 4 .4 °
L ocust b ean
(Ceratonia ailiqua)
2 0 .0
19.9
H oney lo cu st
(Gleditsia Iriacanthos)
2 6 .0
2 5 .9
F la m e tre e
(D elonix regia)
K e n tu ck y coffee bean
( Gymnocladus dioica)
15.6
° D u p lic a te d ete rm in a tio n s on aliq u o ts fro m sam e fe rm e n ta tio n a re given on sam e line.
32
INDUSTRIAL
ANÖ
ENGINEERING
ever, a few prelim inary experim ents indicate th a t galacturonic
acid is affected b u t little by either organism.
ACKNOW LEDGM ENT
G rateful acknow ledgm ent is m ade to L. J. W iekcrham of the
N orthern R egional R esearch L ab o rato ry for his cooperation in
supplying the organisms, and to E d a N ihlen of T he In s titu te of
P aper C hem istry for draw ing Figures 1 and 2 of th e text.
LITERATURE CITED
(1) Anderson, Ernest, U n iversity of Arizona, unpublished data.
(2) A ssoe. Official Agr, Chem ., Official and T en ta tiv e M eth od s of
A nalyses, 5th ed., p. 500, 1940.
(3) Fisher, E ., and Thierfelder, H ., B er., 27, 2031 (1894).
(4) H aar, A. W . v a n der, “A nleitung zum N achw eis, zur T rennung
und B estim m u ng der M onosaccharide und A ldehydsaüren” ,
p. 123, Berlin, Gebrüder Borntraeger, 1920.
CHEMISTRY
Vol. 16, No. 1
(5) H arding, V . J., N ich olson , T . F ., and G rant, G . A ., J . B iol.
Chcrn., 99, 625 (1933).
( 6) H opkins, E . W ., Peterson, W . H ., and Fred, E . B ., J . A m . Chem.
Soc., 5 2 ,3 6 5 9 (1 9 3 0 ).
(7) K luyver, A . J., dissertation, “B iochem ische Suikerbcpalingen",
T echnische H oogeschool, D elft, 1914.
( 8 ) Ivurth, E . F ., unpublished data, T he In stitu te of Paper C hem istry.
(9) K urth, E , F ., and R itter, G . J., J . A m . Chem. Soc., 56, 2720
(1934).
(10) M enzinsky, G ., Svensk P apperslidn ., 45, 421 (1942).
(11) Schm idt, E ., Trefz, F ., and Schnegg, II., B er., 59, 2635 (1926).
(12) S eott, M ., and W est, E . S., Proc. Soc. E xptl. B iol. M ed., 34, 52
(1936).
(13) Sherrard, E . C ., and B lanco, G . W ., I n d . E n g . Chem ., 15, 611
(1923).
(14) W ise, L . E ., and P eterson, F . C ., Ibid., 22, 362 (1930).
P
resented
th e
A
before th e D ivision of S ugar C h em istry a t th e 106th M eeting of
C h e m i c a l S o c i e t y , P ittsb u rg h , P enna.
m e r ic a n
Use of the Discriminant Function in the Comparison of
Proximate Coal Analyses
W . D. B A T E N A N D C. C . D e W IT T
M ichigan State C o llege , East Lansing, M ich.
Comparing a number of analyses of a material from one
source with analyses of similar material from another source
has heretofore presented a very real problem. The present
paper applies statistical methods to the comparison of
proximate analyses and B.t.u. per pound of coal from two
mines. This statistical analysis b y random sampling shows
the probability that two coals come from the same source.
The discriminant function developed b y Fisher is used. By
means of this function a comparison may be obtained
between analytical data based on material from one source
and data on similar material from a different source.
In
addition, the order of significance of the several analytical
constituents in a single series of samples may be deter­
mined. The method Is capable of general application.
cance betw een th e constituents of w hich th e compounds are
formed.
T h e com pound for the first m ine is:
X = &1X1 + hx? - f 63X3 + biXt
where Xi, Xj, x 3, an d x t represent, respectively, B .t.u ., per cent of
volatile m aterial, p er cent of fixed carbon, an d p er cent of ash,
and 61, 62, b3>an d 6, are constants to be found. T h e variables xi,
X2, Xz, and xt m ay be correlated. T h e com pound for th e second
m ine is
X ' = b & l + 62X3 + ¡13X3 + b tx i
where x j, x^, x'3, and xi represent, respectively, th e above
sim ilar m easurem ents; th e coefficients are th e sam e as in com­
pound X . T he difference betw een th e m eans of th e above two
comptmnds m ade up of th e four m easurem ents is
E R (8), in 1936, developed th e discrim in an t function for
FIStheH com
parison of m ultiple m easurem ents obtained in taxo­
nomic problems. Since th en three papers (3, 4 , 5) have appeared
which, m ake use of F isher’s technique. T h e present paper is
concerned w ith th e application of th e discrim inant function to
th e differentiation of two series of proxim ate coal analyses and th e
B .t.u. per pound of coal. E ach series of an alytical d a ta is from a
different mine.
E ach series consists of 100 sam ples of coal. T h e proxim ate
analysis, covering the volatile m a tte r, fixed carbon, per cent of
ash, as well as th e B .t.u. per pound of coal, is used in m aking the
comparison of th e coal from these mines. T he sam ples were
taken from cars of coal sent to th is college over a period of several
years, an d th e analyses are reported on sam ples dried a t 105° C.
T h e analytical d ata, while accurate, indicate th a t th e m ethods of
sam pling m ay be questioned. T h e value of the present approach
is, however, n o t in th e d a ta reported b u t in th e application of
statistical procedures to comparison of sim ilar chemical data.
T h e discrim inant function enables one to o btain a numerical
com parison of th e coals by th e use of tw o linear compounds or
equations in w hich th e effects, in th e present instance, of all four
o f these m easurem ents are combined. F u rth er, th e application
of th is fimction to these analytical d a ta perm its a te st for signifi­
I) — 61'h -f- b^di T b$d3 T btdt
(1 )
w here d 1 — x l — x ', d , = x a — x 3, d3 = x 3 — x 3, and
= x^ —
x i, and x t , x,2, x3, an d x 4 represent th e arithm etic m eans of the
respective m easurem ents m ade of th e coal from th e first m ine
an d x ', x ', x ', and x i represent th e m eans of sim ilar m easure­
m en ts m ade on th e coal from th e second m ine.
Table I. Measurements of B.t.u., Per Cent of V o la tile M aterial,
Per Cent of Fixed Carbon, and Per Cent of A s h for M ines A and B
B.t.u.,
XI
13,000
13,700
12,800
12,300
14,100
13,900
Av. 13,110
----- Mine A--------Vola­
tile
Fixed
matter. carbon,
Xt
Per
cent
ash,
xt
Xi
25.7
25.0
23.0
22.8
33.5
04.3
64.2
60.2
59.9
60.2
7 .9
S .4
10.0
12.9
5.9
2 7 !3
28.33
59.4
61.16
8^3
8.71
------Mine B--------Per
Vola­
tile
Fixed
cent
ash.
B.t.u., matter, carbon,
<
<
13,600
33.0
59.6
0 .5
14,300
36.9
6.2
56.3
35.5
12.9
13,000
50.9
34.9
14,000
58.7
5 .8
13,700
30.1
59.3
10.0
13,800
13,596
35.2
34.00
5Ü 8
56.28
9 .2
8.65
January 15, 1944
ANALYTICAL
EDITION
33
T able I contains a few m easurem ents from m ines A and B,
taken a t random from the sets of observations. T h e m eans of th e
1 0 0 m easurem ents on each of four properties are given in th e last
line of th a t table.
(C)
6, - 0 . 04927762 +
(D)
- 6 , - 0.0040536, -
L et
(A ) -
S = 6 Js,i + b%$i2 + &3S33 + 6 Js„ -f- 2 &1&2S12 + 26,6,Si, +
(J?) - (C)
(B )
0.0894186, -
— 0 .0 0 9 7 5 8 & 2 +
0.0410916, =
-0 .0 0 0 2 3 6
0.0060316, + 0.0107826, =
0.000000
0 .0 0 8 7 7 3 6 , +
0 .0 0 0 8 8 5 6 , =
-0 .0 0 0 0 3 3
0.0623106, - 0.097644&3 + 0.0364816, =
0.000282
26 , 6,S i, + 2626,32, + 26261.521 + 26363831
(C) + (D)
where
100
s<i = Y
k =1
_
~ *»)(*/.» -
_
100
x i) + Y
(A)
where k is the variable of sum m ation.
1 and j — 1 is
_
100
(*<■* ~ x i ) ( x f.k ~ x 'i)
k= 1
T he value of s,,- when i =
100
(*>•* ~ x *)*+ Y
sii = Y
k=1
- O. O 5 3 3 3 O62 + 0.0833876, - 0.0303096, =
-0 .0 0 0 2 3 6
D ividing eacli equation by th e absolute value of th e coef­
ficient of 62 in it gives:
(*!•* “ I >')2
k= 1
-6 2
+ 0.8990576, + 0.0906956, = -0 .0 0 3 3 8 2
(F) 62 - 1.5670686, + 0.5854766, =
(G) - 6 2 + 1.5636046, - 0.5683296, =
(F ) + (F) - 0.6680116, + 0.6761716,
(F) + (G) - 0.0034646, + 0.017146,
0.004526
-0 .0 0 4 4 2 5
=
0.001143
=
0.000101
In our case th is q u a n tity is (using values in T able I) found as
follows:
D ividing each equation b y th e absolute value of th e coefficient
of 6 , in it gives:
s„ = (13,000 - 13,110)* + (13,700 - 13,110)* + . . . + (13,900 13,110)* + (13,600 - 13,596)* + (14,300 - 13,596)* + . . . +
(13,800 - 13,596)* = 20,110,000 + 17,718,400 = 37,828,400
(H) - 6 , + 1.0122156, = 0.001711
(7) - 6 , + 4.9500586, = 0.029157
(if) - (7) -3 .9 3 7 8 4 3 6 , = -0 .0 2 7 4 4 6 ; 6 , = 0.006970
T he value of % where i = 2 and j = 3 is
_
100
«23 =
Y
-
_
Xl)(.X3.k — Xi) +
k =1
100
Y
(X^ k -
x D ( x i.k -
x i)
k= 1
B y using the values in T able I th is is
s., = (25.7 - 28.33) (64.3 - 61.16) + (25.0 - 28.33) (64.2 61.16) + . . . + (27.3 - 28.33) (5 9 .4 - 61.16) + (33.0 34.00) (59.6 - 5 6 .2 8 ) + (36 .9 - 3 4 . 0 0 ) (56.3 - 5 6 . 2 8 ) + . . . +
(35.2 - 34.00) (54.8 - 56.28) = - 645.82 - 373.49 =
-1 ,0 1 9 .3 1
T he q u an tity S is equal to th e sum of squares w ithin com­
pounds. B y maximizing th e q u a n tity D 2/ S th e following equations arise. These equations are actually proportionalities,
b u t for the _purpose
urpose of evaluating constants 61, 6 2, 63, and 64
th ey m ay be used.
S1161 +
«1262 + 81363 +
sh&i
= di
= di
«1261 +
S2262 + S2363 +
82161
«1361 +
S2362 + 83363 +
S3464 = d i
«1461 +
S2462 + S3463 +
S4464 = <f<
From these equations th e values of th e &’s, of
in com pound X , can be found. T he solution
tions gives, of all linear compounds in Xi, X2, x3,
compound, X , which m ost clearly discrim inates
th e other. These equations a re :
From (77)
From (7)
6,
6,
= 0.005344
= 0.005345; 6 , = 0.005345 (average)
From (F )
From (F)
From (G)
62
6:
62
= 0.008820
= 0.008821
= 0.008821; 62 = 0.008821 (average)
From
From
From
From
6,
6,
61
6,
=
=
=
=
(A)
(B)
(C)
(D)
0.000007
0.000007
0.000007
0.000007; 6 , = 0.000007 (average)
Therefore
6,
= 0.000007, 6 , = 0.008821, 6 , = 0.005345, 6 , = 0.006970
T his m ethod of solving sim ultaneous equations is easy to
follow and easy to explain to th e average com puter.
The linear com pound which enables one to detect th e greatest
difference betw een th e mines in relation to these four m easure­
m ents is
X
= 0.000007xi + 0.008821xj + 0.005345x, + 0.006970x,
T he m ean com pound pertaining to mine A is
X
th e coefficients
of these equa­
and x,, th e one
one mine from
37.828.400.006, + 123,906.0062 + 20,685.206, 140,927.0064.= 486.00
123.906.006,+ 1,614.8762 - 1 ,0 1 9 .3 1 6 ,- 571.2164 = 5 .6 7
20.685.006, - 1,019.3162 + 1,849.626, - 849.976, = - 4 . 8 8
-1 4 0 ,927.006, - 571.216, - 849.976, + 1,519.426, = - 0 . 0 6
T here are several ways of solving these equations. One way
(2), using the com puting machines, is as follows: D ividing
each equation by the absolute value of th e coefficient of 6 , in it
gives:
= 0.000007xi + 0.00S821X2 + 0.005345x, + 0.006970x,
= 0.000007 (13,100) + 0.008821 (28.33) +
0.005345 (61.16) + 0.006970 (8.71)
or
X = 0.7293
T he m ean com pound pertaining to mine B is
X ' = 0.000007(13,596) + 0.008821 (34.00) +
0.005345 (56.28) + 0.006970 (8.65) = 0.7562
T he difference D betw een these two m eans is 0.7562 — 0.7293 =
0.0269. T his can be found directly from E q uation 1 as follows:
D = 0.000007 (486) + 0.008821 (5.67) +
0.005345 ( - 4 . 8 8 ) + 0.006970 ( - 0 . 0 6 )
D = 0.003402 + 0.050015 - 0.026083
0.000418 = 0.0269
(2)
as before.
(A)
6,
+ 0.00327562 +
0.0005476, -
0.0037256, =
0.000013
(B)
6,
+ 0.0130336, -
0.0082266, -
0.0046106, =
0.000046
T h e question arises as to w hether or n o t th e m eans of these
compounds are statistically significantly different. T able II
contains an analysis of variance of these compounds and enables
34
INDUSTRIAL
AND
ENGINEERING
one to te s t for a significant difference betw een these two means.
T he tw o asterisks in th e la st colum n indicate th a t these com­
pounds are highly significantly different. T his m eans th a t the
probability of getting by chance such a large value of D is less
th an 0.01. T his is found from a table of F values (I, 7). T he
odds in favor of getting by chance such a large difference between
the compounds pertaining to th e two m ines are less th a n 1 to 99.
T his indicates th a t the coal from m ine A is definitely different
from the coai from m ine B, as far as these four m easurem ents are
concerned.
B y exam ining th e four term s in E q u atio n 2 one can determ ine
which characteristic of th e coal (B .t.u., per cent of volatile
m aterial, per cent of fixed carbon, or per cent of ash) is
m ost im p o rtan t for differentiating th e two coals, which is of n ext
im portance, etc. T h e second term , 0.050015, in th is equation is
greater th a n th e absolute value of each of th e o th er term s; hence
th e per cent of volatile m aterial is th e m o s t im p o rtan t factor, in
th is casé, for determ ining w hether or n o t the coal from one m ine
is different from th e coal from th e o th er m ine. T h e absolute
value of the th ird term in th is equation, 0.026083, is second in
size; hence the per cent of fixed carbon is next in im portance for
revealing a difference between th e coal from th e two m ines. T he
factor B .t.u. is th ird in im portance for differentiating these two
coals. P er cent of ash is of th e least im portance for these mines.
T he order of im portance of these factors m ay change for other
coals. T able I I I gives th e ranks of im portance of these charac­
teristics when various com binations are used to calculate the
com pounds.
T able I I I gives th e ranks of th e ch aracteristics of th e coals for
various, com binations of the sets of m easurem ents. In column 2
it is seen th a t per cen t of fixed carbon is th e m ost im p o rtan t
characteristic for differentiating one m ine from th e oth er as far
as per cent of volatile m aterial, per cen t of fixed carbon, and per
cent of ash are concerned.
CHEMISTRY
Vol. 16, No. 1
Table IV . A rithm etic A verages of B.t.u., Per Cent of V o la tile
M aterial, Per Cent of Fixed Carbon, and Per Cent of A s h
Measurements from M ines C and D
M ines
B .t.u .
P e r C en t of
V olatile
M a tte r
P e r C en t of
Fixed
C arb o n
P e r C en t
of Ash
C
D
13,700
13,205
3 5 .0 9
2 9 .9 9
57 .1 8
59.81
6 .5 6
8 .7 2
495
5 .1 0
-2 .6 3
-2 .1 6
D ifference
Table V . A verages of B.t.u., Per Cent of V o la tile M aterial, Per
Cent of Fixed Carbon, and Per Cent of A s h for M ines E and F
M inea
E
F
D ifference
B .t.u .
P e r C e n t of
V olatile
M a tte r
P e r C e n t of
Fixed
C arbon
P e r C en t
of Ash
14,012
12,134
1,878
3 4 .1 5
35 .7 8
-1 .6 3
60.31
4 8 .1 4
12.17
4 .2 1
13.25
-9 .0 4
m ost im p o rtan t, per cent of fixed carbon is second, per cent of
volatile m aterial is third, and per cent of ash is la st in im portance
in testing w hether or n o t th e m ines differ as fa r as a com pound of
these four sets of m easurem ents is concerned.
T able V contains averages of m easurem ents pertaining to
B .t.u., volatile m aterial, fixed carbon, and per cent of ash for
m ines E an d F.
T h e difference betw een th e m eans of the two discrim inant func­
tions pertaining to th e m ines is
D = 0.000083 (1878) - 0.001356 ( - 1 . 6 3 ) + 0.006124 (12.17) +
0.009580 ( - 9 . 0 4 ) = 0.155874 + 0.002210 +
0.074529 - 0.086603 = 0.146010
In th is case B .t.u. is first, per cent of ash is second, per cent of
fixed carbon is th ird , and per cent of volatile m aterial is fo urth in
im portance for differentiating betw een th e coal from these mines.
DISCUSSION
Table II.
Source of V ariatio n
A n a ly sis of V ariance of the Compounds
D egrees of
Freed o m
T o ta l
B etw een com pound m ean s
W ith in
Table 111.
199
4
195
Sum of S quares
M ean S q u are
50D* - 0.0362
D - 0.0269
0.00905**
0.00014
Rank of Characteristics for V arious Compounds Pertaining
to Coals A and B
C om pound Com posed of
B .t.u.
V olatile m a tte r
Fixed carbo n
P e r c e n t ash
P e r cen t
of v o latile
B ,t.u .
m a tte r
P e r cen t
of v o latile P e r c en t
of fixed
m a tte r
B .t.u .
carb o n
P e r c en t
P e r cen t
P e r cen t of v o latile
of fixed
carbon
m a tte r
of ash
2
3
2
1
1
2
1
3
P e r cen t
of carb o n
P e r c en t
of ash
P e r cent
of v o latile
m a tte r
P e r cent
of fixed
carbon
Ï
2
2
1
T able IV contains th e m eans of m easurem ents pertaining to
B .t.u., volatile m aterial, fixed carbon, an d per cent of ash for
two other mines, C and D . T he discrim inant function pertain ­
ing to these m ines for th e value D is
I) = 0.000020 (495) - 0.006S60 (5.10) 0.001559 ( - 2 . 6 3 ) - 0.000468 ( - 2.16)
or
D = 0.00990 - 0.00350 + 0.00410 + 0.00101
D = 0.01151
An analysis of variance table (not given) shows th a t the two
coals are significantly different. T h e characteristic B .t.u. is
T he discrim inant function enables one to te st for a statistical
difference betw een two linear compounds m ade up of several
variables or m easurem ents. I t has m any advantages because it
furnishes one m easurem ent pertaining to a com bination of several
measurem ents.
In the present instance th e statistical analysis of th e d ata
shows a difference betw een th e coal from th e two mines. I t m ay
be useful in th e future to compare an alytical d a ta from other
mines. Such an effort would involve the selection of a standard
coal, an d it seems ap p aren t th a t o th er analytical d ata, such as
per cent of sulfur, per cent of m oisture, and the fusing tem perature
of the ash, should be included in th e m ore accurate statistical
comparison.
CONCLUSION
Fisher’s discrim inant function has been applied to the dif­
ferentiation of analytical d a ta obtained from one hundred coal
sam ples tak en from each of two m ines. T h e intercom parison of
the relative significance of the elem ents of th e analytical d ata
relating to coal has been accom plished. T h e probability th a t two
coals corrfe from the sam e source b y random sam pling is given.
LITERATURE CITED
( 1 ) B aten , W . D ., “ E lem entary M ath em atical S ta tistics” , pp. 2 4 9 62, N ew York, John W iley & Sons, 1938.
(2) B aten, W . D ., J . A gr. Research, 61, 237—10 (1940).
(3) Brier, G . W ., S ch ott, R . G ., and Sim m ons, V. L., P roc. A m . Soc.
A n im a l Production, 1, 153-60 (1940).
(4) Cox, G . M ., and M artin, W . P ., Iow a Stale College J . S ci., 11,
N o. 3, 323-31 (1937).
(5) D ay, B . D ., Sandom ire, M . M ., J . A m . S lat. Assoc., 37, 4 6 1 -7 2
(1942).
(6 ) Fisher, R . A ., A n n . Eugen., 7, 179-88 (1936).
(7) Snedecor, G . W ., “ S tatistical M eth od s A pplied to E xperim ents
in Agriculture and B io lo g y ” , p p . 184-7, A m es, Iow a, Col­
legiate P ress, 1940.
Colorimetric A nalysis of Xanthone Spray Residues
C.
C. C A S S IL a n d J. W . H A N S E N
U. S. Department of A griculture, Bureau of Entom ology and Plant Quarantine, Yakim a, Wash.
represent 100, 250, and 500 micrograms of xanthone. Prepare
a sta n d a rd graph b y plotting th e q u antities of xanthone in the
stan d ard solution against th e logarithm s of th e corresponding
photom eter readings. T h e light transm ission of 14S micrograms
of xanthone read under th e conditions above in a 2.5-cm. (1-inch)
cell is 50 per cent.
A n a l y s i s o f S a m p l e s . A statistical analysis has shown th a t
20 to 25 apples, tak en from different p a rts of a tree, con stitute a
satisfactory sam ple for residue determ ination. Place th e apples
in a tared glass jar, weigh, and calculate the surface area from the
weight, by the use of a previously established relationship. Add
from 100 to 250 ml. of toluene, depending on th e q u a n tity of
xanthone and th e size of th e apples, and shake for 5 m inutes in a
m achine by th e process described by F ahey et al. (1). F ilter a
portion of th e solution, and use an aliquot n o t to exceed 10 ml. of
the filtrate for th e analysis as described under th e procedure for
standards. R ead th e am ount of xanthone p er aliquot from the
stan d ard graph.
Apple plugs m ay be tre a te d in a sim ilar m anner. If th ey are
used, a sm aller volum e of toluene can be used for stripping.
A colorimetric method for the determination of xanthone
spray residues consists in adding a measured quantity of
toluene to a sample of apples or apple plugs in a glass jar
and shaking for 5 minutes. The resulting solution of
xanthone and apple waxes is filtered and an aliquot of the
filtrate taken for analysis. The xanthone is reduced to
xanthydrol b y refluxing with sodium amalgam in toluene
and methanol. A fte r removal of the methanol b y a water
extraction, an aliquot of the supernatant toluene solution is
swirled in concentrated hydrochloric acid, effecting an
equilibrium transfer of the xanthydrol to the acid layer to
give a yellow color, which is measured photometrically.
X
A N T H O N E has been used experim entally as an insecticide
against codling m oth larvae and o th er insects. T he p ur­
pose of th is study w as to develop a satisfactory m ethod for deter­
mining xanthone residues on sprayed apples, and th e following
colorim etric procedure is recom mended. I t is based on a proce­
dure used by th e L aurel H ill L aboratory of th e G eneral Chemical
Com pany for determ ining xanthone spray residues. T h e w riters
are indebted to several of th e staff of this lab oratory for construc­
tive criticism and advice in the preparation of this paper.
DISCUSSION
S o lv e n ts f o r X a n th o n e .
In selecting a solvent for th e re­
m oval of xanthone residues, consideration was given to solubility,
efficacy of wax rem oval from apples, an d th e extraction of inter­
fering substances. Acetone, alcohols, benzene, a n d petroleum
ether were n o t satisfactory because of poor solvent pow er for
xanthone an d too g reat solvent power for interfering substances.
A N A L y T IC A L PROCEDURE
Toluene. Toluene m ay contain an im purity,
probably a thiophene derivative, which upon shaking w ith con­
cen trated hydrochloric acid yields a slig h t yellow color in the
acid layer. T his im purity can be rem oved by adding 50 ml. of
concentrated sulfuric acid p er liter of toluene, allowing to stand
over 24 hours, separating th e layers, and distilling th e toluene.
T he first m ilky portion of the distillate is discarded. Toluene
from all operations m ay be recovered and reused if treated in this
m anner. W hite crystals of p-toluenesulfonic acid m ay appear in
th e toluene layer during the sulfuric acid treatm en t, b u t th ey do
n o t distill and do not interfere, c.p. toluene usually does not
contain th is im purity.
M ethanol (absolute).
Sodium amalgam. C autiously melt" 9 gram s of sodium in 20
ml. of toluene in a round-bottom ed flask and add 750 gram s of
m ercury, drop by drop a t first an d m ore rapidly after a few milli­
liters have been added. M ost of th e toluene will volatilize, b u t
some should be k ep t over th e amalgam when it is transferred to
an airtight bottle.
H ydrochloric acid (c.p. concentrated).
P r e p a r a tio n o f S ta n d a rd s .
Carefully weigh 50.0 mg. of
pure xanthone, transfer to a 250-ml. volum etric flask, and m ake
to volume w ith toluene. K eep tig h tly stoppered to p revent loss
of toluene. One m illiliter of th is solution contains 200 micro­
gram s of xanthone. If pure xanthone is n o t available for stan d ­
ards, it can be prepared by distilling a crude xanthone product and
recrystallizing the distilled m aterial several tim es from dioxane
or other suitable solvent to a co n stan t m elting point of 174° C.
M easure 2 ml. of th e stand ard solution and sufficient toluene to
m ake a to ta l of 20 ml. into a 125-ml. flask fitted w ith a groundglass joint. A dd 10 ml. of m ethanol and 0.5 to 1.0 ml. of sodium
amalgam , connect w ith a water-cooled condenser, and reflux for
30 m inutes. Before rem oving th e flask from the condenser, cool
to prev ent loss of toluene. Add 20 ml. of w ater and shake vigor­
ously to rem ove th e m ethanol from th e toluene. P our into a tall
tube, such as a 50-ml. N essler tube, retain in g th e am algam in the
flask. Allow th e layers to separate, and p ipet 5 ml. of the tolu­
ene layer (containing the xanthydrol) into a 1 0 0 - to 150-ml. flask
th a t contains exactly 1 0 ml. of concentrated hydrochloric acid.
Develop th e color by swirling th e m ixture gently for approxi­
m ately 1 m inute. P our into a te s t tu b e or cell and m easure the
color of th e acid layer in a photom eter. A glass color filter hav­
ing m aximum transm ission a t 424 millimicrons gave satisfactory
results in a T y p e F Aminco photom eter.
R epeat for 5- an d 10-ml. aliquots of stan d ard xanthone solu­
tion. T he 5/20 aliquots of toluene used for color developm ent
R e a g e n ts .
10
20
30
40
SO
TIME OF REFLUX - M INU TES
Figure 1.
Rate of Xanthone Reduction to Xanthydrol
in Toluene-M ethanol M ixture
T he solubility of xanthone in toluene is 1.43 gram p er 100 mL
a t 30° C., w hich is far in excess of any concentration norm ally en­
countered in residue analysis. Since th e apple w ax appears to be
com pletely dissolved, an y xanthone th a t m ay be covered w ith
wax is also obtained. T h e am ount of interference introduced by
shaking even m atu re waxy apples w ith toluene for 30 m inutes is
negligible. On th e o th er hand, 5 m inutes’ shaking is sufficient
for com plete rem oval of xanthone residues. T oluene was there­
fore selected as th e m ost desirable solvent for this m ethod.
R e d u c tio n o f X a n th o n e .
I t is best to use an aliq u o t of the
residue solution containing from 0.4 to 2.0 mg. of x anthone for
reduction. If th e aliq u o t contains larger quantities, u p to 50
mg., reduction will be complete, b u t fu rth er aliquoting and dilu­
35
INDUSTRIAL
36
AND
ENGINEERING
tion after rem oval of th e m ethanol will be necessary. T he term
"com plete reduction” as used here m eans th a t, under th e condi­
tions of th e m ethod, a m ore intense yellow color cannot be ob­
tained w ith a given am ount of xanthone even if th e tim e of reduc­
tion is doubled. R eduction of xanthone in a m ixture of 20 ml. of
toluene and 1 0 ml. of m ethanol reaches a m axim um , under the
reflux condition of th e m ethod, in 25 m inutes (Figure 1), as
judged by color developm ent.
C o lo r D e v e lo p m e n t.
A fter reduction th e addition of w ater
followed by vigorous shaking rem oves th e m ethanol from th e
toluene layer, w hich retu rn s to its original volum e of 2 0 ml.
W hen th e m ixture is transferred to a container for the separation
of the tw o layers, it is convenient to w ithdraw th e am algam for
subsequent recovery. T he 5-ml. aliq u o t of th e toluene layer
needed for color developm ent can be tak en before all th e toluene
has separated; it is n o t even necessary to filter to rem ove slight
w ater tu rb id ity . If sm aller aliquots are used, sufficient pure
toluene to m ake 5 ml. m u st be added before tre atm e n t w ith acid.
W hen xanthydrol is treated w ith hydrochloric acid, chlorine
is sub stitu ted for the hydroxyl group. T h e resulting com­
pound is colorless in dilute acid, b u t in th e presence of concen­
tra te d hydrochloric acid an intense yellow color is produced.
T he intensity of yellow color is proportional to th e q u a n tity of
xanthone used in the determ ination. I f th e toluene is rem oved
from th e acid layer to p reven t a sh ift in th e d istribution ratio,
th e solution can be diluted w ith concentrated hydrochloric acid
and th e color still conforms to B eer’s law.
N o decomposition of xanthydrol has been observed in toluene
up to 1 2 hours after reduction, b u t low recoveries were obtained
on some sam ples when th e toluene layer was allowed to stan d for
a longer tim e. I t is possible th a t th e observed decomposition is
due p artly to oxidation of xan thydrol to xanthone, because a re­
newed reduction increases th e yellow color upon subsequent acid
treatm en t, b u t n o t to its original intensity. S tan d ard solutions
m ade directly from xanthone and toluene are stable for a t least
60 days.
Table I.
Rcco'
I
V a rie ty
J o n a th a n apples
W inesap apples
W inesap plugs
of Known Am ounts of Xanthone A d d e d to
les Sprayed with Le ad Arsenate0
W eig h t of
Apples
Grams
X a n th o n e
A dded
M 0.
X a n th o n e
R ecovered
Mg.
467
467
586
453
524
323
142 sq. cm .
10 .0
10.0
2 5 .0
2 .0 0
5 .0 0
50. Q
5 .0 0
9 .4 0
1 0 .0
2 5 .5
2 .0 0
5 .0 0
4 9 .6
4 .9 0
A v.
R ecovery
%
94.
100.
102.
100.
100.
99.
98.
9 9 .0
° E a c h sam p le c o n tain ed 25 apples or 80 plugs.
X a n th y d r o l D is tr ib u tio n R a tio s .
In th is m ethod there
are tw o distributions of xanthydrol, betw een toluene and w aterm ethanol solution and betw een toluene an d hydrochloric acid.
B oth distribution ratios have been found to be co n stan t for any
to tal am ount of xanthydrol up to 50 mg. for th e form er and 0.6
mg. for the latter, when the volum es are k e p t as specified in the
m ethod. T he distribution ratio betw een toluene and hydro­
chloric acid w as n o t studied beyond 0 .6 m g., because th e in ten ­
sity of color a t th e corresponding concentration was m ore th an
sufficient for th e m ethod. Two p e r cen t of th e x an thydrol re­
m ains in th e w ater-m ethanol solution, and 77 p er cent of th e
to ta l is transferred to the hydrochloric acid. T he concentration
of th e hydrochloric acid is n o t too critical, since experience has
shown no color differences in th e range of 34 to 36.5 p e r cent acid,
and whereas th e use of lower stren g th acid will lead to lesser color
intensities, constant results will be obtained if th e stan d ard s are
tre a te d w ith the sam e acid. If conditions require th e use of
volum es other th an those specified, a new standardization curve
CHEMISTRY
Vol. 16, No. 1
2.0
100
• ALIQUOTS FROM ONE
REDUCED SOLUTION
o SOLUTIONS EACH OF
WHICH WAS REDUCED
2
o
90
80
70
CO
60 2
CO
so in
O
2
<
co
2
40 C
O
ce 1.6
I-
2
<
2
30
UJ
ir i.4
h
2
LU
20 (x
iu
Q.
5 ML T olu en e-1 0 Ml. H C I
1.2
1.0
BLUE F IL T E R W ITH
MAXIMUM TRANSMISSION
AT 4 2 4 M ILLIM ICRONS
100
200
300
M IC R O G R A M S
Figure 2.
OF
400
500
10
XANTHONE
Representative Standard Xanthone Graph
m u st be established. Since distribution ratio s are involved in
th e m ethod, it is im perative th a t all volum es be m easured accu­
rately.
W hen an aliquot is tak en from th e su p e rn a ta n t toluene layer,
there is some tendency for redistribution of th e x anthydrol upon
standing 2 hours o r longer. T his effect is th e m ore pronounced
th e larger th e concentration of xanthydrol. N o explanation can
be given for th is change in distribution, b u t it is co n stan t for any
co n stan t volum e ratio. If th e procedure is followed as described,
good results are obtainable. If th e toluene solution is to be re­
tained for checking color developm ent on th e same day, it should
be separated from th e w ater-m ethanol solution before a n aliquot
is removed.
T he d istribution of x anthydrol betw een toluene an d hydro­
chloric acid rapidly comes to equilibrium , b u t as a check one
should m ake duplicate determ inations a t th is p o in t in th e proce­
dure. A fter color developm ent no change in inten sity occurs,
even a fte r 24 hours, if precautions are tak en w ith respect to
apple-wax concentration in th e toluene as described in th e next
section.
In te rfe re n c e s .
Apples sprayed w ith lead arsenate alone
a n d stripped for 5 m inutes in toluene gave a photom eter reading
corresponding to th a t given w ith 0 .1 m icrogram of xanthone per
square centim eter. Sim ilar sam ples stripped for 30 m inutes did
n o t show an interference g reater th a n 0.15 microgram per square
centim eter. W hen th is ty p e of interference was tested on Delici­
ous, Rome, an d W inesap varieties, no significant differences were
found. T h e blank on apple plugs is also negligible. X anthone
deposits of 2 2 m icrogram s per square centim eter dropped to 1 .6
m icrogram s a fte r w eathering for 4 weeks; therefore, if a n y de­
composition products are formed, th ey eith er do n o t rem ain on
th e apple or do n o t cause an y detectable interference w ith the
m ethod.
A pple wax does n o t interfere in th e reduction of xanthone.
I t also causes no interference w ith th e color developed in hydro­
chloric acid when th e aliquot of strip solution is 1 0 ml. or less.
I t is possible to use a 15-ml. aliq u o t or even 20-ml. w ith earlyseason apples, b u t if tu rb id ity occurs in th e acid phase, th e wax
concentration in toluene m u st be reduced in some way. T he
yellow color developed in hydrochloric acid is stable for 24 hours
or more, b u t should be read w ithin an hour, since tu rb id ity m ay
develop on long standing. T his effect is m ost m arked when th e
January 15, 1944
ANALYTICAL
toluene is in con tact w ith th e acid layer overnight and is prob­
ably due to tem perature changes.
Ground-glass joints are preferred for th e reflux reduction.
R ubber stoppers m ay be used if th ey arc boiled in norm al alkali
for 15 m inutes, th en in norm al sulfuric acid for 15 m inutes longer,
and finally rinsed well w ith distilled w ater.
A c c u r a c y a n d P r e c i s i o n . One stan d ard graph obtained in
th is investigation is given in Figure 2. T he absorption cells used
were stan d ard te s t tubes (2-cm. inside diam eter). T he line was
fitted by th e m ethod of least squares. T he stan d ard error of
estim ate for th e points on th is line is =*=6.7 micrograms. Aliquots
can be so chosen as to perm it readings on 200 to 500 m icrograms
of xanthone, th u s allowing reduction of th e percentage stan dard
error to 3.3 or less. M ore accurate and precise results can be
obtained if the colored solutions are read in cells w ith flat optical
windows, since th e te s t tubes used in th is w ork v a ry a b o u t 0.5 per
cent in light transm ission.
EDITION
37
R ecovery experim ents .were run by adding known am ounts of
xanthone to apples th a t had been sprayed w ith lead arsenate, and
analyzing b y th e described procedure (T able I). T h e average
recovery from seven sam ples w as 99.0 p er cent, w hich is n o t
statistically different from complete recovery.
T he completeness of rem oval of xanthone from W inesap and
Rome apples (collected 6 weeks before h arvest) was also studied
by subm itting sam ples to a second stripping w ith toluene. T he
am o u n t of xanthone rem oved by th e second tre a tm e n t w as cal­
culated a fte r allowing for th e am o u n t of toluene left on th e apples
from th e first stripping. Six sam ples, having from 6 to 13 mg. of
xanthone per sample, tre ate d in th is m anner, showed a rem oval of
99 per cent or b e tte r w ith th e first 5-m inute stripping treatm ent.
LITERATURE CITED
*
(1) F ahey, J. E ., C assil, C . C ., and R u sk , H . W ., J . Assoc. Official
A gr. Chem., 26, 150-5 (1943).
P H E N O L S T U D IE S
Q ualitative Tests for Phenol and o-,
m-,
and p-Cresol
W m . B. D E IC H M A N N
Kettering Laboratory of A p p lie d Physio logy, C o llege of M edicine, University of Cincinnati, Cincinnati, O h io
Q ualitative tests for phenol em ploying ferric chloride,
hypochlorite, or the reagents of M e lze r, M illo n , Lie be rmann, Guareschi, and Cotton have been modified to permit
differentiation between phenol and o-, m-, ar^d p-cresol.
H E m anufacture and use of phenol an d cresol on a large
scale have furnished an incentive for investigating th e
toxicity and m etabolism of these compounds, an d these investiga­
tions, in tu rn , have called for a review of analytical m ethods
useful in th eir detection an d estim ation. A review of q u an tita tiv e
m ethods for th e estim ation of phenol in biological m aterial, in ­
cluding a spectrophotom etric procedure for th e q u a n tita tiv e esti­
m ation of free, conjugated, an d to ta l phenol in tissues and
fluids, has been published (2, S). In this paper q u alitativ e tests
for o-, 771-, an d p-cresol w hich are modifications of well known
q u alitativ e tests for phenol are described. T he hypochlorite test
for o-cresol offered here has ap p aren tly n o t been recorded before.
A single te st or a com bination of several of these color tests
can be used very effectively for th e identification of phenol or
o-, mr, or p-cresol, if th e unknow n solution contains only one
of these com pounds. I f th e unknow n contains tw o or m ore of
these substances, positive identification of each com pound is n o t
alw ays possible; m-cresol, w hen present in m ixtures in a low con­
centration, is particularly a p t to escape recognition.
I n analyzing for these compounds, even though each substance
gives very sim ilar color reactions in high and in low concentra­
tions, it is best to prepare and te s t dilutions th a t approach th e
ranges of sensitivity. These concentrations will furnish in addi­
tion some rough idea of the q u an tities present.
T
T he phenol used w as M erck’s reagent quality, and th e o-, m-,
an d p-cresol, obtained from th e E astm an K odak Com pany, was
believed to be from 96 to 98 per cent pure. T h e m elting points
o f these cresols are 30-31 °, 10-11°, and 32-34° C., respectively.
P h e n o l an d D e r iv a t iv e s
G r o u t i n g . T he sensitivity
M o d i f i c a t i o n o f M e l z e r ’ s B e n z a l d e h y d e T e s t (â) f o r
D e t e c t i o n o f P h e n o l a n d o-, m - , a n d p - C r e s o l .
M ix 1 ml. of
th e aqueous solution to be tested w ith 2 ml. of concentrated sul­
furic acid (sp. gr. 1.84), and a fte r cooling under th e ta p add 2
drops of benzaldehyde. H ea t over a flame, cool, and add 10
ml. of w ater and 20 ml. of 40 per cent potassium hydroxide.
T he sensitivity of th is te s t is a b o u t 1 mg. per ml. of solution.
M o d ific a t io n o f G u a r e s c h i’s T e s t (4) f o r D is t in g u is h in g
P h e n o l a n d p - C r e s o l f r o m o - a n d t o - C r e s o l . A dd ab o u t 0.5
gram of solid potassium hydroxide an d 3 drops of w ater to 3
ml. of a chloroform ex tract containing phenol or cresol, th en warm
and observe.
Straw-colored (yellow) globules will rise and th e potassium
hydroxide and w ater layer will assume a yellowish tinge on
warming if phenol or p-cresol is present. In th e presence of oor m-cresolj th e potassium hydroxide and th e w ater layer will
assume a pinkish or rose-red color. T he sensitivity for each of
these com pounds is a b o u t 4 mg. in 3 ml. of extract.
F e r r ic
C h lo r id e T e s t
(S) f o r D e t e c t i o n o f
Table I.
P h en o l
o- an d
p-
Co lo r Changes
o -C h e b o l
m -C re s o l
p -C re so l
Changes observed in 5 minutes after addition of sulfuric acid and
benzaldehyde
Cloudy olive
Cloudy orange
Cloudy yellow
Milky white
After heating
Cloudy
Cloudy
reddish-brown
brownish-red
Cloudy
Cloudy
yellowish-brown brownish-green
I
Q U A LITA TIVE TESTS
D e t e c t io n
of
P h en o l- IIy d ro x y
used and th e m anner of preparing th e reagent. F o r these
studies th e la tte r was prepared by dissolving 497 gram s of m er­
cury in 700 ml. of nitric acid (sp. gr. 1.42) and diluting this
solution w ith 2 volum es of w ater. One milliliter of th e test
solution is added to 2 ml. of M illon’s reagent.
Phenol and o-cresol when present to a b o u t 1.0 mg. per ml.,
and 77i- and p-cresol when present to a b o u t 0.5 mg. p er m l.,
produce a red color alm ost im m ediately a t room tem perature.
W hen reduced to ab o u t 0.05 mg. per m l., each of these com­
pounds produces in th e cold or on careful heating a straw-yellow
color which is destroyed by fu rther heating.
After cooling and addition of water and potassium hydroxide
C o n t a in in g
of Millon’s
te s t (7) depends to some ex ten t upon th e q u an tity of mercury
Blue or violet solution and prccipitate
Colorless or faintly tan colored
solution and precipitate
38
INDUSTRIAL
AND
ENGINEERING
C r e s o l . Add 2 drops of a freshly prepared 10 per cent aqueous
solution of ferric chloride to 5 ml. of th e te st solution.
Phenol and rn-cresol produce clear bluish-purple colors, ocresol produces in about 10 m inutes a slightly cloudy urine
yellow or brown solution, while p-cresol produces a cloudy blue
solution. T he sensitivity for each com pound is a b o u t 10 mg. in
5 ml.
M o d i f i c a t i o n o p L i e b e r m a n n ’s T e s t (5 ) f o r D e t e c t i o n o p
j j - C r e s o l . T o 3 ml. of th e unknow n aqueous solution add
slowly and w ith shaking 1 ml. of th e reagent (6 p er cent solution
of sodium n itrite in concentrated sulfuric acid). A cloudy orange
solution develops in about 6 m inutes if p-cresol is present.
Phenol and m-cresol yield clear, and o-cresol very slightly cloudy
brow n or yellowish-brown solutions. T he sensitivity of this
te s t is about 5 m g. in 3 ml.
M o d i f i c a t i o n o f C o t t o n ’s T f . s t ( 1 ) f o r D e t e c t i o n o f p C r b s o l . T o 3 ml. of an aqueous solution, add 1 ml. of concen­
tr a te d am m onium hydroxide (sp. gr. 0.901) an d 4 drops of the
freshly prepared reagent (10 ml. of concentrated hydrochloric
acid and 0.5 gram of potassium chlorate added to 40 ml. of
w ater).
Phenol and o- and m-cresol produce in 5 to 10 m inutes clear
light blue colors; p-cresol produces a clear light straw-yellow
color. T h e sensitivity is abo u t 10 mg. in 3 ml. of solution.
H y p o c h lo r ite T e s t f o r D e t e c t i o n o f o -C re s o l.
T o 5 ml.
of the te st solution, add one drop of sodium hypochlorite solution.
In th e presence of o-cresol th e solution will im m ediately tu rn a
cloudy yellowish-w hite; in th e presence of phenol, and m - and pcresol, it will rem ain clear and colorless. T he sensitivity is
a b out 4 mg. in 5 ml. of solution. (Excess of hypochlorite m ust
be avoided because it m ay produce faint cloudiness w ith pcresol.)
CHEMISTRY
Vol. 16, No. 1
DISCUSSION
I t is reasonable to assum e th a t all reagents discussed in this
p aper will also react w ith some com pounds other th a n phenol, or
o-, m~, or p-cresol. T herefore one m u st m ake certain th a t the
te s t solution is com paratively free from compoimds related to
phenol or cresol. T h is m ay require prelim inary precipitation, ex­
traction, or d istillation procedures.
M illon’s test, even though it m akes specific differentiation be­
tween phenol an d the three cresois difficult, is of value because of
its sim plicity, as a prelim inary test. T h is should be followed by
the modified te sts of M elzer an d G uareschi, which will identify
the com pound. T h e conclusions draw n from these la tte r two
tests m ay be checked b y th e hypochlorite and ferric chloride tests
or b y the modified procedures of L ieberm ann and of C otton.
LITERATURE CITED
(1) C otton , S., Bull. soc. chim ., 21, 8 (1874).
(2) D eichm ann, W m , and Schafer, L. J., Am . J . C lin. P ath ., 12, 129
(1942).
(3) D eichm ann, W m ., and S cott, E . W „ I n d . E n o . C h e m ., A n a l .
E d ., 11, 4 2 3 (1 9 3 9 ).
(4) G uareschi, cited by H . SchifT in Correspondenzen, Ber., 5, 1055
(1872).
(5) Lieberm ann, C., B er., 7, 248, 1098 (1874).
(6) M elzer, H „ Z . anal. Chem., 37, 345 (1898).
(7) M illon, M. E., Compt. rend., 28, 40 (1849).
(8) SchifT, H „ A n n ., 1 5 9,158 (1871).
Stability of Standard Solutions of Copper Perchlorate and
Potassium lodate
JO S E P H J. K O L B
Lankenau Hospital Research Institute, Philadelphia, Penna.
W
H E N E V E R sodium thiosulfate is used in titra tio n work
of high accuracy frequent restandardization is neces­
sary. In order to avoid the troublesom e and wasteful necessity
of preparing for each standardization a fresh solution of a prim ary
sta n d ard [iodine, potassium iodate (S), copper perchlorate (2)],
a standard in the form of a solution of a prim ary substance of de­
pendable stability is desirable. T he work reported here deals
w ith th e possibility of using potassium iodate or copper perchlo­
rate (5) for such a purpose.
In considering the stability of such solutions it is necessary to
distinguish betw een changes due to chemical instability, pre­
sum ably resulting in decrease of active concentration, and
changes due to evaporation from th e container, which will cause
increases of concentration. In th e stu d y of B erm an (I), for in­
stance, on th e stability of potassium iodate, it is impossible to
distinguish the role of these tw o factors. However, his d a ta sug­
gest th a t evaporation has been an appreciable factor in bis results,
and th a t, in some eases, an ap p aren t stab ility has resulted from
the opposing effects of evaporation and decomposition.
In a recent paper (4) on th e stab ility of sodium thiosulfate solu­
tions no account was taken of th e possible effect of evaporation.
A graphical study of th e d a ta on stability shows th a t in th e first.
60 days there is approxim ately a 0.3 per cent increase in norm al­
ity . A fter th a t the values drop. Again two antagonistic tend­
encies, evaporation and decomposition, tend to produce a false
picture of the stab ility of thiosulfate solutions.
T he evaporation factor can be elim inated if, a t th e beginning
of th e experiment, sam ples of the solution to be examined are
p ipetted into separate flasks, an d some of these are titra te d a t
once, while others are titra te d after a suitable lapse of tim e. The
e x ten t of evaporation, on th e other hand, can be m easured by
suitable weighings of th e vessels containing stock solutions. If
th e solution is chemically stable, it is then possible to calculate the
theoretical norm ality a t any tim e from th e initial norm ality and
th e loss of weight th a t has occurred.
APPARATUS A N D REAGENTS
A 50-cc. bu ret an d one 10-cc. p ip et were carefully calibrated
and were used thro u g h o u t th e experim ents. D etails of th e
preparation and use of th e copper perchlorate and potassium
iodate are given in (2) an d (5), respectively.
EXPERIMENTAL
All titra tio n s were carried o u t in duplicate. T h e a m o u n t of
active substance present in solutions when th ey were fresh, an d
a fter th ey h ad stood for various lengths of tim e, was alw ays de­
term ined by titratio n w ith thiosulfate (0.025V), new ly sta n d a rd ­
ized ag ain st tw o freshly prepared cupric perchlorate solutions (2).
Table I.
Solution
Cu(C10*)j
K IO i
D ays
S tan d in g
565
565
289
454
565
565
565
289
454
Stability of 0.1 N Copper Perchlorate and
Potassium lodate Solutions
N o rm a lity
In itia l
Final®
0 .1007
0 .1007
0.0993
0.0993
0.1027
0.1027
0 .1027
0 .1007
0 .1007
0 .10 0 2
0 .1 0 0 4
0.0992
0 .0993
0 .10 2 1
0 .1013
0 .0983
0 .10 02
C o nditions
T h y m o l, glass sto p p ers
G lass sto p p ers
C o rk sto p p ers, spores on corks
G lass sto p p ers
Thymol*», glass sto p p ers
C ork sto p p ers, spores on corks
0 .1004
° F in a l n o rm a lity c alcu la te d on b a sis of its in itia l volum e.
& O nly one sam ple ; o th e rs a re a v erag e v a lu e fo u n d for tw o sam ples.
January 15, 1944
ANALYTICAL
Table II.
Loss of W eig h t p e r
100 D ay s
% o f weight of
solution present
Grams
D ays
EDITION
39
Stability of Solutions in Glass-Stoppered Bottles
N o rm a lity
C alcu la te d
Found
C o n ta in e r
B o ttle
C ap a c ity
A p p ro x im a te
V olum e of
S olution
Cc.
Cc.
C o p per P e rc h lo ra te
0
0-8 5
85-290
290-355
0.7 4 0
0 .6 7 9
0 .4 4 5
1.2 1 8
0 .8 9 8
0 .5 3 4
0 .1 0 0 5
0 .1 0 1 3
0 .1 0 2 5
0.1 0 3 4
G lass-stoppcrcd P y rex a b o u t 12 years old
250
155
0 .1 0 2 6
0 .1 0 3 3
0 .6 7 5
0 .4 8 8
0 .5 0 6
1 Ü Î2
0.671
0 .5 7 6
0ÜÔÔ9
0.1 0 1 9
0 .1 0 2 8
0.1 0 0 3
0 .1 0 0 9
0.1 0 2 3
0.1 0 3 4
G lass-stoppered, flint
500
165
0 ’. 0 i 8
0!021
0*.iÓÓ7
0.1 0 0 7
0.1 0 0 7
R u b b er s to p p er, P yrex
250
12a
o !<348
0 .0 7 5
o ! 0994
250
145
0 .10 0 1
0 .0 9 9 3
0.0 9 9 4
0.0 9 9 7
G lass s to p p er, P y rex , sealed w ith paraffin
0*033
0 .0 8 3
500
265.
B row n, glass sto p p e r, paraffined
500
195
S ta n d a rd in te rc h a n g ea b le glass sto p p e r, P yrex
250
160-
0 .10 11
0
0-85
85-290
290-455
0
0-185
0
0-373
373-1159
P o ta ssiu m Io d a te
0
0-8 5
85-453
o!o53
0 .0 9 5
0Ü 41
0 .1 5 6
o!iÓÓ7
0 .1 0 0 7
0.1 0 0 7
0 .10 11
0 .10 10
0.' 032
0.0 2 1
0.0 6 4
0 .0 3 6
o!iÓ 29
0 .1 0 3 0
0.0 5 1
0 .0 3 5
0 .0 8 3
0 .0 5 0
0 *.i 002
0 .10 0 1
0 .10 0 0
0.1 0 0 3
0.0 9 9 7
0!021
0!Ô40
0 . i 028
0.1 0 2 7
0.1 0 2 7
G lass sto p p e r an d g ro u n d -jo in t cap (" e th e r
b o ttle ")
500
185.
0 .Ó22
0 . 036
0 *. i 028
0.1 0 2 7
0 .1 0 1 3
R u b b e r s to p p er
250
165.
o! 130
o! i 13
0.1 0 4 0
0 .1 0 2 8
0.1 0 3 7
F lin t, glass s to p p er, tu rb id , d e p o sit of inorganic
m a te ria l on w alls
1 lite r
85
0
0-368
368-768
0
0-157
157-557
0
0-238
0
0-165
0
0-855
0 .1 0 2 8
0.1 0 2 6
S olution
D ay s
Effect of Storage under Conditions of Minimum
Evaporation
Loss of W eig h t
p e r 100 D a y s
% of #
weight of
solution
present
Grams
N o rm a lity
In itia l
F in a l
m old a fte r several
0 .10 2 1
In the experim ents sum m arized in T able I, 10-cc. sam ples of
freshly prepared cupric perchlorate and potassium iodate solu­
tions were pipetted into 125-cc. Erlenm eyer flasks. Some solu­
tions were titra te d a t once, while o th er flasks were closed w ith
either glass stoppers or fresh cork stoppers and protected against
du st by paper caps. These flasks were stored in th e labo rato ry in
a cabinet th a t was opened frequently an d th u s provided no pro­
tection against possible contam ination from th e lab oratory a t­
mosphere. T he tem perature was 22° to 35 ° C. T he last of these
solutions were titra te d a fter 19 m onths. As a possible preven­
tion of mold grow th, ab o u t 10 mg. of thym ol were added to some
of th e flasks.
Table 111.
G lass s to p p er, flint glass;
m onths
titra te d as above. T he bottles were th en reweighed an d theabove procedure was repeated a t intervals. F rom th e d a ta given,
one can readily calculate th e initial w eight of th e solutions. In
some cases, portions of th e solutions were rem oved for o th er p u r­
poses. T he bottles were weighed before an d a fte r such rem ovals
and due corrections were applied in th e calculations.
In order to keep evaporation a t a m inim um th e following pro­
cedure was tried. A tared 250-cc. glass-stoppered b ottle contain­
ing ab o u t 100 cc. of copper perchlorate solution was weighed.
T he bottle, placed in a d ry beaker, was stored in a desiccator o v er
a portion of th e sam e solution. One y ear la te r th e b o ttle was
taken from th e desiccator, wiped, and carefully weighed as be­
fore. D uplicate 10-cc. sam ples were th e n titra te d . A p otas­
sium iodate solution was trea ted in th e sam e m anner. A com­
parison of th e results obtained (T able I I I ) w ith those shown in
Table I I shows th a t by th e use of good glass-stoppered or rubberstoppered bottles nearly th e sam e results can be attain ed as by
the desiccator m ethod.
C on d itio ns
CONCLUSIONS
C u(C lO 0»
368
0 .0 1 8
0 .0 1 9
0 .10 22
0 .1 0 1 8
KlO a
392
0 .0 0 1 8
0 .0 0 1 8
0 .1 0 0 2
0.0 9 9 9
G lass sto p p er.
P y rex
G lass sto p p er,
flint
T able I shows the results. I t appears th a t copper perchlorate
solutions possess a high degree of stability, while potassium iodate
solutions show a pronounced tendency to become weaker. In
either case thym ol has a harm ful effect, presum ably because it is
oxidized, while in th e copper perchlorate solutions, even the
presence of black, sporeliko spots on th e cork stoppers was no t
associated w ith any loss of titer.
In the experim ents sum m arized in T able II, sam ples of freshly
prepared copper perchlorate and potassium iodate solutions were
titra te d w ith thiosulfate. T h e rem aining portions of th e solu­
tions were transferred to clean, dry, tared , glass-stoppered re­
ag ent bottles and weighed w ith an accuracy of ± 5 mg. A fter
standing for varying periods under th e conditions described
above, the bottles were carefully dusted an d a b o u t half an hour
later weighed. A p air of 10-cc. sam ples were w ithdraw n and
Solutions in glass-stoppered bottles m ay lose weight by evapo­
ration. T his loss in weight has a tendency to give some types o f
solutions an appearance of stability. R esults of experim ents
indicate th a t solutions of copper perchlorate are more stable than,
those of potassium iodate. F o r use as perm anent standards
copper perchlorate solutions should be stored in good glassstoppered or rubber-stoppered bottles (evaporation losses should
be determ ined by weighing, and norm alities should be corrected
accordingly). Io d ate solutions m u st be stored in glass-stoppered,
bottles. Thym ol should n o t be used as a preservative for either
copper perclilorate or potassium iodate solutions.
ACKNOW LEDGM ENT
T he a u th o r wishes to th a n k G. Toennies of this in stitu te forhis encouragem ent and suggestions.
LITERATURE CITED
(1) Berm an, S. M ., J . Assoc. Official A gr. Chem., 20, 590 (1937).
(2) K o lb , J . J ., I n d . E x q . C h e m ., A n a l . E d ., 11, 1 97 (1 9 3 9 ).
(3) K o lth o ff , I . M ., and S a n d e ll, E . B ., “ T e x tb o o k o f Q u a n tita tiv e -
A n alysis", p. 593, N ew York, M acm illan Co., 1936.
(4) R u e , S . O ., I n d . E n g . C h e m ., A n a l , E d ., 14, 8 0 4 (1 9 4 2 ).
Semiautomatic Pressure Control in Low-Pressure,
Low-Temperature Laboratory Fractionation
D . R. D O U S L IN a n d W . S. W A L L S
Phillips Petroleum Com pany, Research Department, Bartlesville, O k la .
E necessity of using laboratory personnel m anpow er to the
T Hbest
advantage and th e desire continually to im prove the
o ther simply by m oving th e nitrogen flask. N o other changes
in th e pressure control mechanism are necessary.
T h e principal features of the control device are shown in
Figure 1, in relation to th e fractionating colum n. T hree es­
sential p a rts of the mechanism are shown in detail: th e th ro ttlin g
valve in Figure 2, and th e nitrogen dispenser and nitrogen ex­
pander in Figure 3.
q uality of fractional analysis have led to th e installation of
autom atic (£) and sem iautom atic control (1, 2, 4) on fractio n at­
ing columns in m any laboratories. In th e past, even semi­
autom atic pressure control (2, 3) has entailed th e use of an elabo­
rate device installed a t no sm all expense. T h e pressure control
devices described by B osschart (I) and Podbielniak (2) employ
compressed air for dispensing liquid nitrogen to th e colum n head
by m eans of electrically operated valves controlled b y electrical
contacts in the column m anom eter.
F ig u re JI.
THROTTLING V A L V E
T h e th ro ttlin g valve (Figure 2) is an extension of th e outer
arm of th e column m anom eter, an d is actu a ted b y m ercury w hich
rises ou t of the m anom eter and into th e valve, causing th e needle,
a, to float on th e meniscus of th e m ercury column. T h e ball
on to p of the needle will move in the th ro a t, 6, of th e valve w ith
rise and fall of m ercury. T his m otion will restrict th e freedom
of a stream of air which norm ally exhausts through th e vent, c,
on th e valve an d m ust, as a result of the th ro ttlin g effect of the
needle, create sufficient pressure w ithin th e nitrogen-dispensing
bo ttle to cause a discharge of liquid nitrogen. T h e needle will
seek an equilibrium position in th e th ro a t of th e valve, effecting
a sm all b u t constant discharge of nitrogen ju s t sufficient to
control th e pressure in th e column and m ain tain it w ith very
slight fluctuation. T h e am ount of m ercury in th e m anom eter
m ay be varied to produce any pressure
plateau desired from 0 to 760 mm.
T h e dim ensions of th e th ro ttlin g valve are
som ew hat a rb itra ry ; b u t 7-mm. glass tubing
for th e column m anom eter and lower sec­
tion of th e control valve in which th e needle
rides was found to give satisfactory results.
T h e following dim ensions are suggested:
length of th ro a t section, 7.5 cm.; diam eter
a t narrow est section of th ro a t, 2 to 2.5 m m .;
length of needle, 9 cm. T h e th ro a t section
can be draw n down to proper size from
10- or 11-mm. glass tubing; th e m ain pre­
cautions are to keep th e glass circular and
as thick as possible. A fter a satisfactory
th ro a t section has been draw n, it should be
scaled to tubing of th e proper size (7 m m .).
T h e ball on to p of th e needle should ju s t
pass through the narrow est section of the
th ro a t w ith o u t sticking, and th e bottom of
th e needle m u st ride freely in th e tu b e on
the m ercury meniscus. T h e m ost satisfac­
to ry m ethod for connecting th e th ro ttlin g
valve to th e column m anom eter is shown
in Figure 4. T h e three-w ay stopcock should
be a t least 25 m m . below th e 760-mm. m ark
on th e m eter stick to allow a length of m an­
om eter tu b e to ta k e care of pressure
build-up when recharging th e nitrogen
bottle.
T h e needle in th e th ro ttlin g valve de­
scribed above never seats, b u t allows ex­
h au st air to flow through a t all tim es. T he
m ovem ent of th e needle in th e tapered sec­
tio n varies th e size of th e orifice formed by
th e thro ttlin g valve, which acts to control
th e flow of exhaust air, thereby producing
control of pressure w ithin the nitrogendispensing system . T h e position of th e ball
on th e needle in th e lower tapered section
of the th ro a t of th e valve restricts th e flow
of a ir th ro u g h th e valve ju s t enough to
Figure 2. Throt­
create sufficient pressure w ithin th e n itro­
tling V a lv e
g e n -d is p e n s in g s y s t e m t o f o r c e t h e
a. Floating needle
am ount of liquid nitrogen o u t of th e
b. Throat section
liquid nitrogen flask needed to control
c. Air exhaust
d. Tungsten elec­
column pressures and to m ain tain liq­
trical contact
uid reflux in th e column. As th e n itro ­
e. Constriction in
gen level in th e flask decreases, m ore
tubing
Semiautomatic Pressure Control D evice
A . ThrottJing valve
B . Liquid nitrogen dispenser
C. Column condenser
D . Manual control
E . Electrical buzzer
T his paper describes a control device developed in this labora­
tory from m aterials usually available, which provides a simplified
m eans of sem iautom atic control for low -tem perature, low-pressure
laboratory fractionating equipm ent, w ithout sacrificing excellence
o f control o r ease of operation. I ts o u tstanding a ttrib u tes are
sim plicity of construction and low cost. T h e device m ay be
used w ith good results on either th e stan d ard low -tem perature
laboratory fractionating colum n or th e Podbielniak H eli-Grid
type (3); and if these tw o types o f columns are connected to th e
sam e manifold, th e control m ay be shifted from one column to the
40
January 15, 1944
ANALYTICAL
EDITION
as th e large resistance offered to th e flow of nitrogen m akes it
necessary to m ain tain an unduly large pressure in th e D ew ar
flask in order to force over sufficient nitrogen.
T he heating elem ent in th e liquid nitrogen dispenser is useful
for reducing th e fluctuation in column pressure during the
fractionation of m ethane an d ethane, since it tends to cause de­
livery of the.cooling agent to th e colum n head in th e form of a
more nearly continuous stream of dropwise increm ents of liquid
nitrogen autom atically adjusted to m eet th e column require­
m ents. F or separating the com ponents propane through hexane
the th ro ttlin g valve usually affords satisfactory control of column
pressure. T he use of th e heating elem ent in th e liquid nitrogen
dispenser appears to cause very little if an y significant increase
in the liquid nitrogen requirem ents for fractionating m ethane
and possibly ethane. Since only a sm all portion of th e liquid
nitrogen is vaporized by the h eater while fractionating m ethane,
Figure 3.
Liq u id N itrogen Dispenser
{ (right)
Dewar fla»k
Porcelain conduit
Rubber stopper
Air intake tube
P.
Heating element
<7.
Electrical lead wires
(left) end N itrogen Expander
Discharge tube
Nitrogen expander
Asbestos string packing
Liquid nitrogen intake tube
Spent nitrogen exhaust tube
V
Compressed
Air
pressure is required in th e flask to force th e liquid nitrogen
ou t; this ex tra pressure is furnished autom atically by th e valve,
since th e ball on the needle merely assum es a position fu rth er up
th e th ro at, fu rth er restricting th e flow of air. W ithin reason­
able lim its, th e ra te of furnishing compressed a ir to th e nitrogendispensing system produces no difficulties to operation, since the
needle in th e th ro ttlin g valve autom atically ad ju sts itself to a
position suitable for operation a t a given ra te of a ir flow. T he
rate of air flow is usually set to perm it th e ball of th e needle to
operate a t a point ab o u t 1 to 1.5 cm. below th e narrow est section
of th e th ro at. Since th e nitrogen expander has been designed
to perm it rapid response of column pressure to th e addition of
cooling agent, the th ro ttlin g valve is capable of autom atically
controlling th e column p ressu re 'w ith in very narrow lim its of
fluctuation.
LIQUID NITROGEN DISPENSER
T he nitrogen-dispensing u n it shown in detail in Figure 3 (left)
is contained in a quart-size, w ide-m outhed D ew ar flask. An air
intake tube, t, a nitrogen discharge tube, m , and electrical leads,
k, are sealed into th e flask by m eans of a rubber stopper, h,
and a porcelain conduit, g, for th e electrical lead wires. Air
which has been induced to flow into th e D ew ar flask as a result
of th e action of the throttling valve will force liquid nitrogen up
m into contact w ith the heating elem ent, j, which being relatively
ho t flash-vaporizes a sm all portion of the liquid nitrogen, and
th e sudden expansion of th e vapor forcibly ejects th e liquid
nitrogen from the discharge end of the tube. T he increm ents
of liquid nitrogen discharged in th is m anner are of sm all size,
since the flash-vaporization occurs before an y large q u an tity
can find its w ay into th e upper portion of th e discharge tube.
T he heating elem ent in the discharge tu b e will continue to expel
nitrogen in sm all bursts, a t rapid, regular intervals, which are
regulated as regards am ount of nitrogen and tim e interval by
th e pressure brought to bear b y th e th rottling valve, which in
turn reflects th e need of the column for cooling agent.
T he heating elem ent can be m ade from 8.75 cm. (3.5 inches) of
No. 26 Chromel resistance wire, sufficient to m ake a heater
which extends from ju s t below th e rubber stopper to th e bend
a t th e top of the discharge tube. D uring an analysis it is neces­
sary to v ary th e am ount of current in the heating elem ent; this
can be done w ith a sm all 2-ohm, 5-am pere rh eo stat if a potential
of 6 to 9 volts is being used. T h e discharge tube, m, should be
m ade of 9-mm. glass tubing sealed as it enters th e rubber stoDper
to capillary tubing having 2.5- to 3-mm. inside diam eter. C ap­
illary tubing sm aller th an 2.5 mm. cannot be used successfully,
s
Figure 4. Connection of Throttling
V a lve to Column Manometer
r. Pinch clamp
». Three-way stopcock
t
Throttling valve
v. Valve on compressed air line
y , Manual control button
x. Electrical buzzer
th e more uniform m anner of dispensing nitrogen to th e column,
head enables b e tte r operation of th e colum n an d more efficient
use of th e cooling agent, th ereby partially com pensating for
the nitrogen requirem ents of th e heater.
NITROGEN EXPANDER
In designing th e nitrogen expander (Figure 3, rig h t) and th e
nitrogen intake tube, it is im p o rtan t to keep in m ind th a t th e
liquid nitrogen should be conducted from th e D ew ar flask to
make con tact w ith th e distilling tube b y th e sh o rtest possible
42
INDUSTRIAL
AND
ENGINEERING
CHEMISTRY
Vol. 16, No. 1
route and should in no case pass through a difficult p a th before
m aking contact. T he intake tube, p, should be a continuation of
tube m in th e nitrogen flask, and should have th e sam e inside
diam eter (2.5 m m .). E xpander n is m achined from a solid
piece of m etal, and is supported by sm all pieces of ru b b er or
felt inside th e column condenser. T his type of expander allows
th e nitrogen to con tact th e distilling tu b e im mediately, and
conducts any excess nitrogen aw ay from th e distilling tu b e where
it m ay vaporize w ithout causing excessive cooling near the
tube. T he packing, o, is asbestos strin g wound around the
distilling tu b e and pressed in to place. Precautions m u st be
tak en to p rev en t w ater from accum ulating in th e nitrogen ex­
pansion cham ber or around the distilling tube.
Soon after th e th ro ttlin g valve controller is cu t in, the nitrogen
b o ttle should be inspected for leaks around th e stopper an d special
care taken to see th a t a sm ooth connection is m ade between
th e discharge tu b e in th e nitrogen b o ttle and the tu b e to the
column head.
If the sam ple being analyzed contains m ethane, it m ay be de­
sirable to build up reflux in th e column w ith th e m anual control
b u tto n , and c u t in th e th ro ttlin g valve controller a fter th e column
h as become com pletely wet.
An operator m u st be presen t to a d ju st th e take-off rate valve
a t th e various c u t points and to record th e necessary data.
T h e buzzer, z, signals th e operator when th e nitrogen b o ttle is
em pty, or if for any oth er reason th e column condenser is n o t
being supplied w ith sufficient nitrogen. W ith . a suitable ar­
rangem ent of laboratory equipm ent it is frequently possible for
one o perator to operate two columns sim ultaneously.
TECHNIQUE OF OPERATION
ACKNOWLEDGMENTS
D uring th e tim e th a t a sam ple is introduced into th e fractionate
ing column, the th ro ttlin g valve controller should be cu t o u t by
turning the stopcock (s, Figure 4). A fter sam pling is completed
a n d the k ettle allowed to w arm up, clam p r should be released
from the leveling bottle and m ay be closed again when th e de­
sired pressure is reached. A t som e tim e during th e period when
pressure is building up, valve v should be opened, causing a small
stream of air to flow through th e th ro ttlin g valve. T h e condenser
head on th e colum n is next brought to proper tem perature,
using the m anual control b u tto n , y, and should be a t tem perature
when th e th ro ttlin g valve controller is cu t in.
T he authors are indebted to T . A. M atthew s, C. F . W einaug,
and D . E . S m ith of Phillips Petroleum C om pany L aboratories
for assistance o r suggestions and to th e Phillips Petroleum Com­
p any for permission to publish this paper.
LITERATURE CITED
(1)
(2)
(3)
(4)
B o s s c h a r t, R . A . J ., I n d . E n g . C h e m ., A n a l . E d ., 6, 2 9 (1 9 3 4 ).
P o d b ie ln ia k , W . J ., Ib id ., 5, 172 (1 9 3 3 ).
I b id ., 13, 6 39 (1 9 4 1 ).
R o se , A r th u r , I b id ., 8, 4 7 8 (1 9 3 6 ).
Color of A queous Potassium Dichromate Solutions
R. E. K IT S O N WITH M . G . M E L L O N
Purdue University, Lafayette, Ind.
A spectrophotometric study has been made of the color
of aqueous potassium dichromate solutions, including the
effect of p H and of the concentration of acid, base, and d i­
chromate. A l l these variables have an effect on the color
of the solutions, and any recommendations of potassium
dichromate solutions as permanent colorimetric standards
should specify the exact nature of the solution.
A Q U E O U S solutions of potassium dichrom ate a n d /o r potasI \ sium chrom ate are recom m ended as p erm anent colorimetric
standards in a num ber of colorim etric m ethods involving un­
stable yellow colors. Exam ples are th e following procedures:
silica as molybdisilicic acid, residual chlorine in w ater, carotene,
an d varnish. In sp ite of these w idespread uses, little work has
been done on the color of potassium dichrom ate and potassium
chrom ate solutions, especially the effect on them of such variables
as pH an d th e kind and am o u n t of acid or base added.
H antzsch and C lark (S), N euss an d Itiem an (4), Sherrill (5),
Vosburgh an d Cooper (7), and others have studied th e chrom atedichrom ate relationship from the stan d p o in t of th e ionic equilibria
involved rath e r th a n th e color of th e solutions. Swank and
Mellon (G) pointed o u t th e necessity of buffering potassium chro­
m ate solutions a t pH 9 to secure color m atches w ith molybdisilicic
acid.
T h e present spectrophotom etric stu d y was und ertak en to de­
term ine th e effect of pH , various acids an d bases, and dichrom ate
concentration on the color of aqueous potassium dichrom ate solu­
tions.
EXPERIMENTAL WORK
A p p a ra tu s a n d S o lu tio n s .
T ran sm ittan cy m easurem ents
were m ade in 1.000-cm. cells w ith a G eneral E lectric recording
spectrophotom eter, ad ju sted for a spectral band w idth of either
5 or 10 m/i. In case of colored reagents, these solutions were
used in th e reference cell. Otherwise, redistilled w ater served.
All p H m easurem ents were m ade w ith a glass electrode assembly.
Stock solutions of potassium dichrom ate were prepared from
twice recrystallized sa lt an d redistilled w ater. Tw o solutions,
containing 10.00 gram s p er liter and 12.50 gram s per 100 ml.,
were used. A series of C lark and Lubs buffer solutions was pre­
pared according to th e directions of C lark (2).
E f f e c t o f pH o n t h e C o l o r o f P o ta s s iu m D ic h r o m a te
S o lu tio n s .
T o study th e effect of p H on th e chrom ate-dichrom ate system , th é desired am ount of potassium dichrom ate solu­
tion was pipetted into a 50-ml. volum etric flask, and th en di­
luted to th e m ark w ith a buffer of th e desired pH . T h e contents
were mixed and allowed to sit a few m inutes to ensure equilibrium .
T he spectral transm ission curve and th e p H were th en d eter­
mined.
In general, dichrom ate solutions a t a low p H have an orange
hue, while those a t a high pH are yellow. T hese colors are com­
m only associated w ith th e dichrom ate and chrom ate ions, re ­
spectively. An interm ediate range exists in which th e hue is ex­
trem ely sensitive to sm all changes in pH . A t low dichrom ate
concentrations this in term ediate range extends from pH 5 to 7,
as shown in Figure 1. A t higher concentrations, th e range shifts
upw ard, being from pH 6 to 8 a t 2.0 mg. of potassium dichrom ate
per ml.
A t low concentrations of sa lt th e series of solutions a t various
pH values has an isobestic p o in t (I) (Figure 1). W ith different
concentrations of dichrom ate, th e transm ittancies of this point
v ary according to B eer’s law. T h e w ave length a t th is p o in t de­
creased from 444 m a a t 0.2 mg. to 440
a t 1.0 m g. of dichro­
m ate p er ml. A t concentrations g reater th a n 1.0 m g. per ml. the
absorption is too high a t these wave lengths to show th e point.
E f f e c t o f A c i d s . In studying th e effect of various acids, and
in all subsequent work, th e following procedure was used:
January 15, 1944
ANALYTIC
The dichrom ate solution was p ip etted into a 50-ml. volum etric
flask, the acid or base added, an d th e solution diluted to th e m ark
a t room tem perature w ith w ater. In case th e w ater could n o t be
added to the concentrated acid, approxim ately th e desired
am ount was added to th e flask before the acid. T h e tran sm ittancy curves were determ ined w ithin an ho u r after m aking the
final volum e adjustm ents.
W ith single acids two dichrom ate concentrations were used:
0.4 and 5.0 mg. per ml. T he form er was low enough to show th e
relatively flat portion of the tran sm ittan cy curve near 440 m/i,
a n d to observe the effect of acid upon it. T he la tte r was high
enough so th a t this p a rt of th e curve did n o t appear. T he acid
concentration varied from none to concentrated acid.
L EDITIO N
43
where th e m inim um should be. E ven a sm all am o u nt of acid
shifts th e entire curve tow ard longer wave lengths. T h en acidi­
ties on up to 9M have little effect on th e color; b u t betw een 9
and 12M th e color changes from orange-red to brown. Above
12il/ there is little further color change.
N itric Acid. W ith a low dichrom ate concentration solutions
containing nitric acid show th e first of th e tw o color changes noted
w ith sulfuric acid-potassium dichrom ate solutions. T he m ini­
mum near 440 mp appears when th e acidity reaches 5M , and the
intensity of th e minim um increases w ith increasing acidity. W ith
a high dichrom ate concentration th e entire curve shifts to longer
wave lengths as th e nitric acid concentration increases. T he
brown color noted w ith sulfuric acid does n o t appear.
Phosphoric Acid. Q ualitatively phosphoric acid has m uch the
same effect on th e dichrom ate color as nitric acid, b u t q u a n tita­
tively it differs considerably. W ith a low dichrom ate concen­
tratio n th e m inim um n ear 440 m/u is definite w ith 0.3Ï1Î acid.
Increasing th e acidity to l.OAf increases th e intensity greatly,
b u t higher acidities cause little further change. T h e tran sm it­
tancy curve for high dichrom ate concentrations shifts to longer
wave lengths w ith sm all am ounts of acid, but, as before, am ounts
greater th a n I .Oil/ have little additional effect.
Perchloric A cid, W ith a low dichrom ate concentration, this
acid has little effect on th e color. T he minim um in th e trans­
m ittancy curves does n o t appear, b u t th e percentage tra n sm it­
tancy of th e horizontal portion of th e tran sm ittan cy curve in­
creases slightly w ith increased acid concentration. In th e pres­
ence of larger am ounts of potassium dichrom ate, a precipitate,
presum ably potassium perchlorate, appears.
Acetic Acid. D ichrom ate solutions containing acetic acid
show changes sim ilar to those w ith nitric acid. T h e m agnitude
of th e changes is greater th an those w ith nitric acid, although not
so large as th e corresponding changes w ith sulfuric acid.
Hydrochloric A d d . Although it is generally assumed th a t po­
tassium dichrom ate does n o t oxidize hydrochloric acid in aqueous
solution, th e yellow color of solutions which contain 0.4 mg. of
potassium dichrom ate per ml. and w hich are 911/ or stronger in
acid disappears w ithin 5 m inutes a fte r addition of th e acid.
Because of th is reaction, which proceeds a t a slower ra te in
more weakly acidic solutions, hydrochloric acid solutions of po­
tassium dichrom ate fade and should n o t be used as p erm anent
Figure 1. Effect of p H on the Color of Potassium
Dichromate Solution
pH on curve
m g . K2C r t 0 7 p e r m l.
I
A q u e o u s s o lu t io n
2 . 0..8 I M Hz S O ,
W ith mixed acids only one concentration of dichrom ate was
used: 0.4 mg. per ml. T he to ta l acid concentration did n o t ex­
ceed th a t likely to be encountered in general analytical practice.
W ith the binary m ixtures 5 ml. of each acid per 50-ml. final volume
was th e m axim um , while w ith th e te rn ary and q u atern ary mix­
tures th e m axim um of each acid was lim ited to 3 ml.
3
1 .6 2
4
. 7 .2 0
■■
5
1 0 .8
"
6
1 6 .2
■■
-
Lack of space prevents presentation or ad equate description of
all the tran sm ittan cy curves. An a tte m p t is m ade in th e follow­
ing paragraphs to sum m arize th e ir characteristics:
Sulfuric Acid. Potassium dichrom ate solutions containing
sulfuric acid showr th e greatest v ariation in color of an y of the
acids used. W ith a low concentration of dichrom ate, small
am ounts of acid give a very definite m inim um in th e curve near
440 mp, th e wave length of th e m inim um being related to the
acidity (Figure 2). L arger am ounts of acid, up to 9M , cause no
great change. Between 9 and 1111/, however, th e hue of th e solu­
tions changes from orange to orange-red, an d th e minim um point
in the curve disappears. As th e acidity is increased further,
there is a shift tow ard an orange-brown hue, w ith no striking
change in the curve.
A high concentration of dichrom ate behaves similarly, b ut
there is no m inim um in the curve a t concentrations m uch greater
th an 1 mg. of dichrom ate per ml. because of complete absorption
500...520 540 560 SR0
W a v e le n g th , m jj
Figure 2.
Effect of Sulfuric A c id on the Color of Potassium
Dichromate Solutions
44
INDUSTRIAL
AND
ENGINEERING
standards. A solution 0.8M in acid fades ab o u t 2 per cent in 48
hours, and one 0.4A/ in acid fades to th e sam e ex ten t in a m onth.
If the fading of th e solution is overlooked, hydrochloric acid
affects the color in a m anner sim ilar to nitric acid. T he minim um
appears a t 0.4A/ acid, and as th e acid concentration increases the
m inim um is more m arked.
B inary M ixtures. B inary m ixtures containing sm all am ounts
of phosphoric acid and one of th e oth er acids show very definite
m inim a near 440 m/i. Changes in th e concentration of th e
second acid have only a sm all effect on th e color of th e solution.
T he m inim um is present in th e tran sm ittan cy curve even if th e
second acid is nitric or perchloric acid, neither of which, alone and
a t low concentrations, causes th e m inim um to appear.
M ixtures of either perchloric or nitric acid and sulfuric acid
show lim ited m inim a form ation a t low, and definite m inim a a t
higher, sulfuric acid concentrations.
M ixtures of nitric and perchloric acid show no m inim a in the
tran sm ittan cy curve.
CHEMISTRY
Vol. 16, No. 1
1.2 m g. of potassium dichrom ate per ml. can be determ ined in
solutions 1.8M in sulfuric acid or 1.5M in phosphoric acid
(Figure 3). B eer’s law is valid over th e entire range in both
cases. W ith 1.6M n itric acid, 1.7M acetic acid, or 0.911/ per­
chloric acid, 0.02 to 0.9 m g. of potassium dichrom ate per ml. can
be determ ined if the m easurem ents are m ade in 1-cm. cells a t 430
m/i. B eer’s law is valid only to 0.6 mg. p er ml. for these solutions.
If th e dichrom ate solution is ad ju sted to p H 2 or 9, and the
tran sm ittan cy m easurem ents are m ade in 1-cm. cells a t 440 m/r
th e m easurable range is 0.02 to 1.0 mg. per m l., B eer’s law being
valid over th e entire range.
E f f e c t o f B a s e . I n this stu d y th e sam e experim ental proce­
dure w as followed as w ith th e acids. A freshly prepared solu­
tion, 8A/ in sodium hydroxide, served as the source of the base.
A fter enough base has been added to convert the dichrom ate
to chrom ate, addition of an excess has little effect. Solutions 8A/
in base show a very slight greenish tin t.
DISCUSSION
Solutions of potassium chrom ate a n d /o r potassium dichrom ate
used as perm anent colorim etric stan d ard s should be buffered to a
definite pH . In th e case of th e interm ediate p H values, the buf­
fer should have a high capacity, since small changes in pH m ake a
large change in th e color. A t either a high or low p H th is is n o t
so im portant.
I t is necessary to specify both th e kind of acid, or acids, an d the
am ount of each when acidified dichrom ate solutions are used.
If the phosphoric acid is m ore th a n 0.3AZ, sm all v ariations in th e
concentration of a second, third, or fo urth acid will n o t seriously
change th e color of th e solution. Solutions containing hydro­
chloric acid are n o t suitable for perm anent standards.
A high concentration of base should be avoided, since it a ttack s
glass and causes tu rb id ity . A low concentration (pH 9 or 10)
gives th e sam e color, w ith m uch less a tta c k on th e container.
T h e isobcstic points probably have some significance in term s
of th e ionic equilibria involved. C lark (1) sta tes th a t an isobestic
po in t is an intersection of all isohydric curves, an d “ consequently
the probability of occurrence is low unless tw o colored com­
pounds and two only have some in tim ate relationship” . Such a
po in t in th e chrom ate-dichrom ate transm ission curves seems to
indicate th a t only tw o ions are significantly involved in the
equilibrium transform ation. T hree ions have been postulated
for this system , CrO» , H C rO t- , an d Cr20 7 , th e reactions in­
volved being
Figure 3. Effect of Potassium Dichromate
Concentration on the Color of Solutions
in 1.5M Phosphoric A c id
Ternary M ixtures. Solutions containing sm all am ounts of
phosphoric acid w ith tw o oth er acids show a very definite mini­
m um near 440 my. in the tran sm ittan cy curve. T he intensity of
the minim um is n o t m uch affected by sm all changes in th e concen­
tratio n of either or both of th e o th er acids.
Solutions n o t containing phosphoric acid show a lim ited mini­
mum form ation, and the intensity of th e m inim um is affected by
changes in the concentration of th e other acids. Increases in
sulfuric acid concentration increase th e minim a, while increases
in nitric a n d /o r perchloric acid decrease them .
Quaternary M ixtures. T h e spectrophotom etric curves for all
th e quaternary m ixtures tested were virtu ally th e same, and all
showed a pronounced m inim um near 440 mji.
E f f e c t o f D ic h ro m a te C o n c e n tr a tio n .
T he range of con­
centration for m easurem ents of dichrom ate solutions in 1-cm.
cells depends on th e am o u n t an d kind of acid used, and th e w ave
length a t w hich the m easurem ents are m ade. W ith 1-cm. cells,
m easured a t the wave length of m inim um transm ission, 0.02 to
C rO r~ + II+ ^ H C rO r
(1)
2 H C r O r ;=± C r20 7— + H 20
(2)
T he isobestic p o in t would seem to indicate th a t R eaction 1 is of
m ajor im portance, since it alone involves hydrogen ions. Calcu­
lations from th e d a ta of N euss an d R iem an (/) show th a t 88 per
cent of th e dichrom ate ion in a solution containing 1 mg. of po­
tassium dichrom ate per ml. should react to give th e HCrO«- ion.
If these d a ta, an d th e facts ab o u t isobestic points, are correct,
the absorption spectra of th e HCrO<~ and C r20 7— ions should be
sim ilar, since conversion of one form to th e o th er has little effect
on th e over-all tran sm ittan cy .
LITERATURE CITED
(1) Clark, “ D eterm in ation of H ydrogen Io n s” , pp. 153, 154, B alti­
more, W illiam s & W ilkins C o., 1928.
(2) Ib id ., pp. 193 ff.
(3) H an tzsch and Clark, Z . ph ysik. Chem., 63, 367 (1908).
(4) N eu ss and R iem an, J . A m . Chem. Soc., 56, 2238 (1934).
(5) Sherrill, Ibid., 29, 1641 (1907).
(6) S w a n k a n d M e llo n , I n d . E n g . C h e m ., A n a l . E d ., 6, 3 4 8 (1 9 3 4 ).
(7) Vosburgh and Cooper, J . A m . Chem. Soc., 63, 437 (1941).
b s t r a c t e d from a thesis p resented b y R . E . K itson to th e G ra d u a te School
of P u rd u e U n iv e rsity in p a rtia l fulfillm ent of th e re q u ire m e n ts for th e
degree of d o c to r of philosophy, F e b ru a ry , 1944.
A
Gravimetric Determination of Tungsten
W ith A n ti-1 /S-d i-ip -m e th o xy p h e n y l)-!-h y d ro xy lam ino-3-oxim ino-4-pentene
JO H N H . Y O E a n d A . LETCH ER JO N E S 1
University of V irgin ia, Charlottesville, V a .
A new organic compound has been developed as a reagent
for the gravimetric determination of tungsten. Its physical
and chemical properties have been investigated and pro­
cedures are given for its use in the determination of tungsten
in ores and alloys. Determinations of tungsten with the
new reagent are equivalent in accuracy to the standard
cinchonine method.
pressly for this work, was observed to form a highly insoluble
yellow precipitate w ith tu n g sta te ions (W 0 4— ) in acid solution.
A detailed investigation revealed th a t th is new com pound com­
bines w ith th e tu n g state ion
CH]0<^~
)>CH— C II2— C— C H = C H < ^
HONH
^>OCH,
NOH
in a definite ratio of one molecule of reag en t to one tu n g state
ion and th a t th is reaction m ay be applied to th e gravim etric
determ ination of tungsten.
procedures for th e gravim etric determ ination of tu n g ­
M OST
sten involve separation of th e m ajor portion of th e ele­
m ent from solution as tungstic acid, H jW O (, by strong m ineral
acid treatm en t, and recovery of th e sm all am o u n t rem aining in
NATU RE O F R EA CTIO N WITH TU NGSTATE IO N
solution by m eans of an organic precipitant. Cinchonine is gen­
erally used for this purpose. P resent governm ental restrictions
T he exact mechanism b y which th e reagent combines w ith the
on th e sale of cinchonine for analytical purposes and th e increas­
tu n g state ion to form th e insoluble complex is n o t know n with
ing num ber of applications of tu n g sten an d tungsten alloys to
certainty. I t is possibly a salt-form ing reactio n :
industrial uses m ake th e introduction of an effective and easily
obtainable reagent for tungsten p articularly desirable a t th is
CH,Q<(
^>CII— C H r - C — C H = C H < ^
)>OCH, + II,WO« —
tim e.
T his paper describes a new organic reagent for tungsten,
HONH
NOH
a n ti -1,5 - di - (p - m ethoxyphenyl) - 1 - hydroxylam ino - 3 - oximino-4 pentene, first synthesized in th is labo rato ry and applied
to th e analysis of a v ariety of tungsten ores and alloys with highly
IIO N H ,
HNOH
satisfactory results. T he com pound m ay now be obtained from
±[WO«]=*=
L aM o tte Chemical Products Co., Towson 4, Baltim ore, M d.
V arious organic reagents for th e gravim etric determ ination of
tun gsten have been proposed since L efort (IS) first reported in
1881 th a t quinine acetate would precipitate tu n g states from solu­
tion. Crem er (1) reported th e cinchonine reaction in 1895, b u t
it w as n o t applied to actu a l analyses u n til a b o u t 20 years later
when Low (14) developed a gravim etric procedure employing
cinchonine as a p recipitant for tungsten. Jan n asch an d B ettges (8) used hydrazine hydrochloride and strong hydrochloric
acid to precipitate tungsten b u t separations were n o t complete by
th is m ethod. K norre (11) found benzidine hydrochloride
slightly b e tte r th a n hydrazine hydrochloride b u t still n o t en­
tirely satisfactory.
O ther investigators have proposed a v ariety of gravim etric or­
ganic reagents: a-naphthylam ine and cumidine (23), te tram ethyl-p-dinm inodiphenylm ethane (10, 20), 1,4-diphenylendanilodihydrotriazole (N itron) (4), quinoline (15), tan n in (16, 21),
phenylhydrazine hydrochloride (2), 8-hydroxyquinoline (5, 9,1 8 ),
vanillylidene-benzidine (7), and rhodam ine B (17). N one of
these reagents has proved sufficiently effective to replace cin­
chonine as th e preferred reagent in stan d ard procedure, although
its use involves certain difficulties (1, 6, 12).
T he reagent belongs to the basic salinogenic group of organic
compounds, m any m em bers of which combine w ith anions such
as nitrates, perchlorates, sulfate, phosphates, m olybdates, v an a­
dates, thiocyanates, etc. (28), in acid m edium . .The “sy n ” iso­
m er of th e organic com pound is com pletely unreactive w ith the
CH,Q<(
)>CH — C II2—C— C H = € H
HONH
)>OCII,
HON
tu n g state ion. I t crystallizes in the form of colorless plates
(m .p. 217° C.).
P H Y SIC A L A N D C H E M IC A L PROPERTIES O F R EA G EN T
T he reagent crystallizes from ethanol in th e form of sm all yel­
low needles which m elt a t 156-157° C. (corrected). I t is soluble
in th e following organic solvents, solutions being yellow in all
instances: ethanol, acetone, eth y l eth er, eth y l acetate, dioxane,
acetic acid, an d benzene. T h e com pound is insoluble in w ater
and petroleum eth er. Solutions of th e reagent in eth an o l are
used in tun g sten analyses. T he solubility is 0.766 gram of re­
agent p e r 100 ml. of ethanol a t 25° C.
S o l u b i l i t y o f T u n g s t e n C o m p l e x . A n experim ent was
conducted to determ ine w hether or n o t th e precip itated organotungsten complex was soluble in w ater to th e e x ten t of one p a rt
per million.
E arly in 1931 a system atic investigation of th e reactiv ity of
organic com pounds, w ith respect to color an d precipitate form a­
tion w ith inorganic ions, was begun in this laboratory. S tan d ard
solutions of a b o u t eighty inorganic ions were prepared and
tested, in b o th acid an d alkaline m edium , w ith various ty p es of
organic com pounds, especially those w ith one or m ore chelating
or salt-form ing groups. T he oximes, as a class, appeared to be
one of th e m ost prom ising ty p es to tr y in th is search for specific
an d highly sensitive reagents for inorganic ions.
D uring the 1930’s several new and im p o rtan t organic analytical
reagents were discover«^ in th is laboratory (19, 24-27). In 1940
an ti - 1,5 - di - (p - m ethoxyphenyl) - 1 - hydroxylam ino - 3 - oxim ino-4-pentene, one of a series of com pounds synthesized ex-
One milligram of finely pulverized dried complex (dried a t
105° C. for 2 hours) was introduced into a flask containing 1 liter
of w ater a t 25° C. T he w ater w as m echanically stirred for one
week. A t th e end of th a t time, th e complex had n o t dissolved;
hence it w as concluded th a t its solubility is less th a n 1 mg. per
liter of w ater—i.e., 1 p.p.m .
1 P rese n t address, C ornell U n iversity , Ith a c a , N . Y .
45
46
INDUSTRIAL
AND
ENGINEERING
CHEMISTRY
Vol. 16, No. 1
f
T he pure reagent is pale
yellow in color b u t is slowly darkened by light. T his discolora­
tion does not affect its reactivity wit h tungstates. N o discolora­
tion occurs if stored in d ark bottles.
R e a c t io n s w it h I n o r g a n ic I o n s
T ests for reactivity with
inorganic ions were m ade on porcelain spot plates by adding a
drop of an ethanol solution of th e reagent to a drop of solution
containing the respective ions (approxim ately 0.05 mg.) in both
acid and alkaline m edium where possible. No reactions were ob­
served betw een th e reagent an d any of th e following ions:
S t a b il it y o p R e a g e n t to L ig h t .
Ag+, A l+++, A s+++, A s0 4
, B 40 7— , B a++, B e++, B i+++,
Br~, CO,— , C a++, C b 0 4
, C d ++, C e++\ C l“ , Co++, C r+++,
C s+, D y +++, E r +++, E u +++, F ~, F e++, G a+++, G d +++,
G e + + + + (aqueous solution of GeOj), IIfO ++, H g +, H g ++, I ,
I n +++, K + , L a+++, L i+, M g++, M n ++, N O r , N O r , N a+,
N d +++, N i++, H P 0 4“ , P b ++. P r +++, P tC l5— , R b +, R e 0 4- ,
R h+++, R u +++, S— , S 0 4— , S b +++, Sc+++, SeO, , S i O , ~ ,
Sm +++, S r++, T a 0 4
, TeO*“ , T iO ++, T h ++++, T l+++,
T m +++, VO+, Y +++, Y b +++, Zn+ +, Z r+ +++ (ZrO ++).
Auric, eerie, and iridic ions produce brown precipitates w ith the
reagent b u t these are extrem ely rare in tun g sten ores and alloys.
Ferric iron reacts to form a sm all am o u n t of brown precipitate
b u t n o t enough to interfere even in th e analysis of steels contain­
ing tungsten.
Stannous and stannic tin s produce an orange and a yellow pre­
cipitate, respectively, w ith the reagent.
C opper reacts in alkaline solution to produce a brown precipi­
ta te b u t does not react a t all in acid medium. In analysis, th e re­
agent is alw ays used in acid solution; hence copper presents no
interference.
M olybdates are precipitated b u t n o t quantitativ ely . W hen
m olybdenum and tungsten are present together, a sm all am o u n t
of m olybdenum m ay precipitate w ith th e tungsten. T he pre­
cipitation of m olybdenum by the reagent is no greater th a n when
cinchonine is used (see procedure for analysis of steels an d alloys;
cf. analysis of N .B .S. steel 132).
T he only other ions found to react w ith th e reagent are .UO,++
an d OsOs , which produce an orange an d brow n precipitate,
respectively.
OPTIMUM EXPERIMENTAL CONDITIONS
P e r m i s s i b l e A c i d i t y . T ungsten is q u an titativ ely precipi­
ta te d by the reagent when th e acidity is less th a n p H 1. P artial
precipitation occurs above p H 1 an d above p H 5 no precipitation
occurs. T he best acidity to use is ab o u t 0.2Ar hydrochloric acid.
Sulfuric and nitric acids m ay also be used, b u t in th e presence
of lead or tin sulfuric acid is objectionable and nitric acid solu­
tions stronger th a n IN begin to a tta c k th e reagent.
T e m p e r a t u r e o f S o l u t i o n . Room tem perature is preferable
for effective precipitation of th e tungsten complex. P recipita­
tion begins im m ediately upon addition of th e reagent in cold
solution, w ith th e precipitate collecting an d settling m uch more
readily th a n if the solution is hot. T he reagent should therefore
n o t be added to hot solutions.
A m o u n t o f R e a g e n t R e q u i r e d . T he reagent combines w ith
tu ngstate ions in a ratio of 1 to 1. However, when th e approxi­
m ate q u a n tity of tungsten is know n, it is recom m ended th a t
twice th e theoretical am oun t of reagent be added to ensure com­
plete an d rapid precipitation. If th e tun g sten content is un­
known, th e am ount of reagent required m ay be judged by ob­
serving the color of th e precipitated complex as it forms.
T i m e N e c e s s a r y f o r C o m p l e t e P r e c i p i t a t i o n . Complete
precipitation is obtained when solutions are allowed to stan d a t
room tem perature for 3 hours a fte r th e addition of excess reagent,
w ith occasional stirring to collect th e precipitate. Allowing
solutions to stan d overnight ensures com plete precipitation b u t is
n o t necessary.
W a s h i n g t h e P r e c i p i t a t e . T he m ost effective wash solution
has been found to be a cold, dilute hydrochloric acid solution
containing a small am ount of th e reagent. I t is prepared b y d i­
luting 20 ml. of concentrated hydrochloric acid to 1 liter and
adding I ml. of a satu rated ethanol solution of th e reagent.
F i n a l T r e a t m e n t o f P r e c i p i t a t e . A ttem p ts to weigh the
organo-tungstate precipitate directly, after washing an d drying,
indicate th a t th e m ethod is im practical. C onditions have to be
tediously controlled to get consistent results. T h e difficulties
arise from th e fact th a t th e reaction m u st occur in acid solution
and th a t th e reagent itself precipitates when added in sm all ex­
cess. Unless th e reagent is added im m ediately after acidifying
the solution, some tungsten m ay p recipitate as tungstic acid and
be occluded in the organic precipitate, preventing a considera­
tion of it as a com pound of definite composition. I t is also diffi­
cu lt to remove precipitated reagent from th e complex w ithout a t
th e sam e tim e dissolving some of th e complex.
T h e complex m ay be ignited easily to tungstic oxide, W Os, in
which case it m akes no difference if tungstic acid an d excess re­
agent are present, because tungstic acid is converted to tungstic
oxide an d excess reagent is com pletely rem oved upon ignition
T he precip itated complex is easily and rap id ly separated by
filtration thro u g h q u a n tita tiv e ashless filter paper. W hatm an
filter paper N o. 40 an d other papers of sim ilar tex tu re and grade
are suitable for q u a n tita tiv e analyses w hen using th is reagent.
F iltratio n is m ost rapidly accomplished if th e p recipitated com­
plex is allowed to settle (about 3 hours is sufficient) an d the
solution filtered b y d ecantation. T here is very little tendency
for th e p recipitate to clog th e pores of th e paper, th u s p erm itting
thorough washing «nth a m inim um tim e consum ption.
T h e washed p recipitate m ay be ignited w ith o u t difficulty.
T here is no tendency for explosion, or decom position rapid enough
to sweep o u t an y of th e contents of th e crucible. T he precipi­
ta te an d th e filter paper decompose a t a b o u t th e sam e rate.
M oisture should be rem oved a t ab o u t 100 ° C. before raising the
tem perature of th e crucible an d contents to ignition tem perature.
T he crucible should n o t be heated above a dull red until all car­
bon has been b urned off.
DETERMINATION OF TUNGSTEN IN SOLUTIONS, STEELS,
A L L O Y S , A N D ORES
Of th e various procedures investigated, a modification of th a t
of H illebrand an d Lundell (6, pp. 553-5) has been found m ost
satisfactory for th e analysis of tungsten ores. F or steels and
tungsten alloys, m odifications of th e procedures of th e American
Society for T esting M aterials (22, pp. 1011-2) are both speedy
a n d accurate.
D e t e r m i n a t i o n i n O r e s . T ran sfer to a 400-ml. beaker 1
gram of sam ple w hich has been ground in an agate m o rtar to
200-mesh or finer an d dried to co n stan t w eight a t 105° C. U n­
less th e sam ple is ground to 200-mesh or finer, a protective layer
of precipitated tungstic acid m ay coat th e particles, preventing
the hydrochloric and nitric acids from coming in con tact w ith all
the unreacted m aterial. T h is is especially tru e w ith ferberite
and w olframite, which are difficult to decompose. If sufficient
hydrochloric acid is used an d the tem perature is held a t approxi­
m ately 75° C., th e form ation of tungstic acid on th e particles
m ay be avoided. Serious error m ay be introduced by this
effect if sam ples of particle size m uch larger th a n 200-mesh are
used. T he m agnitude of th e error depends on th e m ineral being
analyzed. Scheelite, C a W 0 4, an d hubncrite, M n W 0 4, are
easily decomposed by th e hydrochloric-nitric acid treatm en t, in
which case th e error is sm all; ferberite, F e W 0 4, and w olframite,
(F e,M n )W 0 4, being m uch m ore difficult to decompose, require
very sm all particle size.
Add 5 ml. of distilled w ater an d ro tate th e beaker so as to dis­
trib u te th e sam ple evenly over its bottom . A dd 100 ml. of hy­
drochloric acid (sp. gr. 1.19), cover th e beaker w ith a w atch
glass, and h eat for one hour a t a tem perature n o t exceeding
60° C. S tir occasionally to break up form ations of crusts and to
facilitate co n tac t of th e acid w ith all particles of th e sam ple.
Raise th e w atch glass on glass hooks a n d cautiously boil to a
volume of a b o u t 50 ml. B reak up th e m aterial on th e bo tto m of
th e beaker w ith a glass stirring rod. A dd 40 ml. of hydrochloric
acid an d 15 m l. of nitric acid an d again boil th e solution to about
50 ml. S tir up th e caked m a tte r again, add 5 ml. of nitric acid,
a n d boil to 10 or 15 ml. D ilute to 250 ml. w ith h o t w ater and
h e a t ju s t below boiling for 30 m inutes. Allow to cool to room
tem perature and add 25 ml. of alcoholic reagent solution (made
by dissolving 0.7 gram of reagent in 100 ml. of ethanol) slowly
w ith co n stan t stirring. Allow th e p recipitate to settle for about
2 hours an d te s t the su p ern a tan t liquid w ith a few drops of the
reagent for complete precipitation of th e tungsten.
1
!
January IS, 1944
ANALYTIC
Because of th e lim ited solubility of th e reagent in aqueous
solutions, some precipitate m ay be expected to form whenever
th e reagent is added. T he appearance of th e precipitate formed
in th e presence of tungsten is different from th a t form ed when
th e la tte r is absent. T he organic tu n g state is deep yellow-orange
in color while th e precipitated reagent alone is alm ost w hite. I t
is advisable for a n analyst using th e reagent for th e first tim e to
add some of it to a solution of sodium or am m onium tu n g state
(1 mg., of tu n g state per m l.) acidified w ith 8 or 10 volumes of
0.2JV hydrochloric acid. T he same am ount of reagent should be
added to a sim ilar volum e of 0.2N acid alone. T h e striking dif­
ference betw een the color of th e reagent an d th a t of th e organotu n g state m ay be observed by th is simple procedure.
In testing a solution for completeness of precipitation of tu n g Bten, more reagent should be added if an orange precipitate is
formed. If it is w hite or pale yellow, precipitation of th e tu n g ­
sten is complete.
A fter th e precipitate has settled a t least 2 hours (or overnight
if convenient) filter by decantation thro u g h an ll-c m . ashiess
filter paper. W ash th e precip itate several tim es w ith reagent
wash solution.
T ransfer the precipitate to an ashiess filter paper an d m oder­
ately scrub th e beaker by m eans of a ru b b er policeman to remove,
as far as possible, any precipitate adhering to th e walls of the
beaker. I t is n o t necessary a t th is point to a tte m p t to remove
all th e finely divided tungstic acid adhering to th e walls of the
beaker, since the m ain precipitate, a fte r it is washed, is to be
transferred back to th is beaker an d all th e precipitated tungsten
redissolved.
W ash w ith th e prepared w ash solution. R ep eat several times
and set aside the combined filtrate an d washings. T e st th e fil­
tra te w ith a few drops of reagent solution for complete precipita­
tion of the tungsten. I t is rare th a t an y is found in this filtrate.
T ransfer th e filter paper containing the washed precipitate to the
original beaker an d add 6 ml. of concentrated am m onium hy­
droxide. Shred th e filter paper to a uniform pulp b y m eans of a
glass stirring rod, cover the beaker, an d w arm gently for a few
m inutes. S tir th e pasty m ass w ith th e rod, th en w ash down the
inside of th e beaker w ith w arm dilute am m onium hydroxide (1 to
9) containing 10 gram s of am m onium chloride p er liter. W arm
again and stir thoroughly. F ilte r through an ll-c m . ashiess
filter paper and collect th e filtrate in a 400-ml. beaker. W ash the
original beaker and residue several tim es w ith th e w arm dilute
am m onium hydroxide solution; betw een washings, squeeze as
m uch of th e liquid from th e fibers of th e pulp as possible. W ash
th e beaker and residue w ith several sm all portions of h o t 95 per
cent ethanol to dissolve any organo-tungstote th a t has n o t been
decomposed by the am m onium hydroxide tre atm e n t. Follow
w ith a final washing w ith the w arm d ilute am m onium hydroxide
solution. K eep th e volum e of liquid as sm all as possible. R e­
serve th e residue of filter fiber for fu rth e r recovery of traces of
tungsten.
E v ap o rate th e filtrate to a volum e of a b o u t 50 m l.; ad d 20 ml.
of concentrated hydrochloric acid and 10 ml. of concentrated
nitric acid; cover and cautiously boil to a volum e of 10 to 15 ml.
If an y organic residue is still present an d ten d s to adhere to the
walls of th e beaker, it m ay be decomposed by th e addition of more
hydrochloric an d nitric acids in th e sam e ratio as above. D ilute
th e solution to ab o u t 250 ml. w ith w ater an d allow it to cool
slowly to room tem perature. A dd th e alcoholic reagent solution
until th e color of the precipitate, as it forms, indicates complete
precipitation of th e tungsten. Allow th e precipitate to settle and
filter through an ll-c m . ashiess filter p aper; w ash thoroughly
w ith reagent w ash solution. If th e filtrate has a clear yellow
tin t, precipitation is complete. I f it is colorless, m ore reagent
m ust be added to complete th e precipitation. T h e w ashed pre­
cipitate is the m ain precipitate an d is to be ignited w ith th e very
small am ounts of tungsten th a t m ay be obtained in th e procedure
described in th e following paragraphs.
A ny tungsten th a t m ay not have been recovered is contained
in th e reserved residue of filter fiber, in com bination w ith iron
or alum ina or w ith sm all am ounts of reagent th a t were n o t com­
pletely dissolved. T his combined tun g sten m ay be dissolved by
digesting the filter paper and residue of fiber w ith w arm dilute
hydrochloric acid (1 to 9). F ilte r an d w ash th e residue w ith
small am ounts of h o t 0.5 per cent am m onium chloride solution
and the w arm dilute am m onium hydroxide w ash solution, col­
lecting all in th e sam e vessel. Acidify th e filtrate w ith hydro­
chloric acid u n til it is approxim ately 0.2Ar an d th en slowly
add 5 to 10 ml. of alcoholic reagent solution. Any precipitate
L E D ITI O N
47
obtained should bo filtered an d w ashed w ith th e reag en t wash
solution. R eserve th e w ashed precipitate and ignite la te r w ith
th e m ain one already obtained.
The residue th a t now rem ains is usually free from tungsten.
To be positive, ignite it in a porcelain crucible (not a platinum
crucible because tin m ig h t be present), tran sfer th e ash to a
platinum crucible, and volatilize th e silicon by tre a tin g it w ith
hydrofluoric an d sulfuric acids. Fuse th e rem aining residue
w ith as little sodium carbonate as possible, cool, and e x tract the
m elt w ith w ater. F ilte r and acidify th e filtrate w ith hydrochlo­
ric acid, boil to expel carbon dioxide, an d ad d dropwise some of
the reagent solution to te s t for the possible presence of tu n g state
ions. If th e precipitate is orange, upon th e addition of the first
few drops of reagent, filter it off, w ash w ith th e reag en t wash solu­
tion and ignite w ith th e tw o residues already obtained. (The
authors have never obtained any organo-tungstate p recipitate
a t this point.)
Place th e papers containing th e m ain precipitate and th e two
recoveries in a weighed platinum crucible and h e a t a t a tem pera­
ture below a dull red u n til all th e carbon has been burned off.
Cool, add a few drops of hydrofluoric acid to th e residue (enough
to m oisten it com pletely), ad d a drop of sulfuric acid and evapo­
ra te to dryness over a w ater or sand b ath . Reignite in order to
get the weight of W 0 3 free from S i0 2.
T he tungstic oxide obtained a t th is point is not pure b u t m ust
be examined for contam inants. T h e exam ination for contam i­
n ants does n o t involve any changes from th e stan d ard procedures
now in use (3; 6, p. 555; 22). T he contam inants m ost a p t to be
present a t th is po in t are principally iron and m olybdenum , w ith
very small traces of phosphorus. T he iron is separated by fusion
w ith sodium carbonate, dissolving in h o t w ater, and filtering.
T ungsten and m olybdenum (if present) form soluble sodium salts.
If m olybdenum is present, th e am ount of its contam ination m ay
be determ ined colorim etrically (3; 22, pp. 1008-9) in th e fil­
trate. I t is rare th a t th e am ount of phosphorus present justifies
a te st for it. T he w eights of contam inants found are su b tracted
as oxides from th e w eight of im pure tungstic oxide and th e cor­
rected weight is used to calculate the percentage of tungstic oxide
or tungsten in th e sample.
D e t e r m i n a t i o n i n A l l o y s a n d M e t a l s . T re a t 1 gram of
the finely divided m etallic sam ple w ith 5 ml. of hydrofluoric acid
in a large covered p latinum crucible or dish. A fter th e initial
effervescence has ceased, add nitric acid dropwise un til th e m etal
has dissolved. A dd 15 ml. of sulfuric acid (1 to 1), transfer the
vessel to a sand b ath , and h e a t cautiously until dense fum es of
sulfur trioxide are evolved freely. Perchloric acid m ay be sub­
stitu ted for sulfuric acid if desired. In this case, add 15 ml. of
perchloric acid (60 per cent) a fter th e hydrofluoric-nitric acid
tre atm e n t an d h ea t cautiously to dense fumes of perchloric acid.
Cool, dilute, and tran sfer in th e sam e m anner as when sulfuric
acid is used. T h e use of perchloric acid shortens slightly th e tim e
necessary for th is p a rt of th e analysis, b u t otherwise offers no
particu lar advantage.
Allow to cool and tran sfer th e co n ten ts to a 400-ml. beaker by
washing th e p latinum vessel w ith a fine stream of w ater. W ipe
th e vessel w ith a sm all piece of ashiess filter paper and tran sfer
it to th e beaker. R inse th e vessel w ith a little w arm am m onium
hydroxide (1 to 1), a little w ater, th en a little h o t hydrochloric
acid (1 to 1). R ep eat th e treatm en ts with am m onium hydroxide,
w ater, and hydrochloric acid, adding all rinsings to th e 400-ml.
beaker. D ilute th e contents of th e beaker w ith w ater to a b o u t
150 ml. Add 10 ml. of hydrochloric acid (sp. gr. 1.19), cover
w ith a w atch glass supported on glass hooks, and boil cautiously
for a t least 5 m inutes. Rem ove th e source of h e a t an d dilute
th e contents to a b o u t 350 ml. w ith w ater. Allow to cool and add
slowly, w ith co n stan t stirring, 15 to 20 ml. of alcoholic reagent
solution (0.7 gram per 100 ml. of 95 per cent ethanol). Allow
th e p recipitate to settle for a b o u t 2 hours and te s t th e su p e rn a ta n t
liquid for complete precipitation of th e tungsten. If an orange
precipitate forms upon the addition of m ore reagent, precipita­
tion is incom plete. If th e precip itate is alm ost w hite, separation
of tungsten is complete.
W hen the p recipitate has com pletely settled, filter by d ecanta­
tion through an ll-c m . ashiess paper. W ash th e p recipitate
several tim es w ith reagent wash solution, ignite th e p aper and
residue, in th e platinum vessel in w hich th e sam ple was tre ate d
originally, a t a tem perature below a dull red h e a t u n til all of th e
carbon is consumed. A dd a few drops of nitric acid an d evapo­
ra te to dryness on a w ater or sand b ath. Ig n ite to co n stan t w eight
(preferably in an electric muffle furnace) a t a tem p eratu re n o t
exceeding 750° C. T his is th e w eight of im pure tungstic oxide.
Add a b o u t 5 gram s of sodium carbonate and carefully h e at until
a clear m elt is obtained. R o ta te th e fused m ass in th e vessel
until it solidifies around th e wall of th e container. W hen cool,
dissolve th e m elt in ho t w ater, filter through an 11-em. ashiess
paper, and wash thoroughly w ith hot w ater. Place the filter in
48
INDUSTRIAL
AND
ENGINEERING
th e crucible and ignite again. R epeat th e sodium carbonate fu­
sion on the small residue, using a proportionately sm aller am ount
of carbonate th an in th e first fusion. Cool and dissolve in th e
same m anner as before. F ilter and w ash thoroughly to rem ove
all traces of sodium carbonate. Ig n ite in th e same platinum ves­
sel, cool, and weigh. S u b tra c t th e w eight of this oxide residue
from th a t of the original im pure tungstic oxide. Calculate the
percentage of tungsten from the corrected w eight of tungstic
oxide.
If m olybdenum is present, th e am o u n t of its contam ination of
th e tungsten precipitate can be determ ined in th e sodium car­
bonate extracts. T his is best done colorim etricallv (3; 22, pp.
1008-9).
A N A LY SIS OF STANDARD TUNGSTATE SOLUTIONS
A stan d ard tu n g state solution was prepared from c .p . tungstic
acid (HjWO<, 99.84 per cent pure) by fusion w ith sodium hy­
droxide in a silver crucible. T he fusion m ass was dissolved in
w ater and diluted to th e m ark in a volum etric flask. M easured
volumes of this solution were tak en and acidified to 0.2N w ith
hydrochloric acid and the tungsten w as precipitated by th e addi­
tion of an ethanol solution of th e new reagent. In order to com­
pare the tw o reagents, the tu n g sten was also precipitated by cin­
chonine from a sim ilar series of solutions. T he precipitated com­
plexes were filtered off through q u an titativ e paper and ignited
in platinum crucibles to tungstic oxide (Table I).
APPLICATIO N OF PROCEDURES
N atio n al B ureau of S tan d ­
ards sam ples Nos. 50a, 132, and 75 were used to stu d y th e appli­
cability of the procedure to th e determ ination of tungsten in
steels an d alloys (Table II ). One-gram sam ples were used and
th e tungsten was precipitated and separated according to th e
procedure given for th e analysis of tun g sten alloys and m etals.
A n a l y s i s o f O r e s . Samples of scheelite and ferberite (ob­
tained through the courtesy of G. E. F . Lundell, N ational B ureau
of S tandards, W ashington, D . C.), an d w olfram ite (supplied by
th e C allite T ungsten C orporation, U nion C ity, N . J.) were
analyzed by the procedure outlined for th e determ ination of
tungsten in ores (Table I I I ) . One-gram sam ples were used.
A n a ly s is o f S te e ls a n d A llo y s .
DISCUSSION OF RESULTS OF A N A LY SE S
T he results obtained in th e analyses of th e solutions, steels,
alloys, and ores show th a t anti-1,5-di-(p-m ethoxyphenyl)-lhydroxylamino-3-oximino-4-pentene m ay be applied to th e gravi-
Table 1.
Standard Tungsten Solutions
Sam ple
WO» P re se n t
M g.
1
2
9 .1
1 8 .5
1 8 .0
1 8 .0
3
4
Table II.
Sam ple
1
2
3
4
5
WO» F o u n d b y
C inchonine
M 0.
9 .1
1 8 .5
1 7 .9
18.1
Steels and A llo y s
M ate rial
N .B .S . ch ro m c-tu n g sten -v an ad iu m
steel N o . 50a
N .B .S . m o ly b d cn u m -tu n g sten steel
N o. 132
W
P re se n t
%
1 8 .2 5 °
6 .2 9 °
6
7
8
9
10
WO» F o u n d b y
N ew R ea g e n t
M o.
9 .1
1 8 .5
18 .0
18.1
N .B .S . ferro -tu n g aten N o. 75
7 5 .2 °
W
F o u n d D ifference
%
%
18.35
+ 0 .10
18.21
-0 .0 4
18.27
+ 0 .02
6 .5 8
6 .6 5
6 .7 3
6 .3 4
6 .3 4
7 5 .3 6
75 .1 1
+ 0 .2 9 6
+ 0 .3 6 6
+ 0 .4 4 6
+ 0 .0 5 c
+ 0 .0 5 «
+ 0 .2
- 0 .1
“ N a tio n a l B u re au of S ta n d a rd s certificate value.
S N .B .S . N o. 132 co n tain s 7 % (approx.) m o lybdenum . I n sam ples 4,
5, an d 6 , no correction was m ade on ig n ited tu n g s tic oxide for m olybdic
oxide as a c o n ta m in an t.
e I n sam ples 7 an d 8 , m o ly b d en u m co n ta m in atin g oxide was d eterm ined
colorim etrically a n d co rrectio n s applied. H ig h resu lts of sam ples 4, 5,
a n d 6 are ev id en tly d u e to presence of m olybdic oxide.
CHEMISTRY
Vol. 16, No. 1
m etric determ ination of tun g sten in these m aterials. T he varia­
tion of percentage values betw een individual sam ples is well
w ithin th e range of v ariation encountered b y th e use of oth er ac­
ceptable gravim etric m ethods. W ith th e N atio n al B ureau of
S tandards sam ples, th e values found b y th is m ethod are w ithin
th e range of v ariation of values reported by th e various analysts
using o th er m ethods. T h e sm all variations in percentage found
b y th is m ethod are believed to be due to experim ental technique
ra th e r th a n an y inconsistency in th e behavior of th e new reagent.
In th e analysis of N .B.S. steel 132, w hich contains approxi­
m ately equal am ounts of tun g sten and m olybdenum (7 per cent
m olybdenum an d 6 per cent tungsten) it was necessary to d eter­
m ine th e am ount of m olybdenum contam inating th e tungsten
p recip itate and correct for it.
Table III.
O res
1
2
Scheelite
W Oj
P rese n t
%
5 9 .6 °
3
4
5
F e rb e rite
66.0«
6 6 .0 8
6 6 .2 5
65 .9 7
+ 0 .1
+ 0 .3
6
W olfram ite
6 9 .3 9 t
6 9 .4 0
69 .2 3
+ 0 .0 1
-0 .1 6
Sam ple
M ate rial
7
WO»
F ound
%
59 .5 3
5 9 .7 0
D ifference
%
- 0 .1
+ 0 .1
0 .0
« T e n ta tiv e values supplied b y G . E . F . L undell. Sam ples w ere d ried
for 18 hours a t 105° C. before being analyzed b y D r. L u n d ell in 1918. T h e y
w ere d ried u n d e r sam e conditions for analysis w ith new tu n g ste n reagent,
a fte r each was th o ro u g h ly mixed.
&V alue supplied b y C allite T u n g ste n C o rp o ra tio n on basis of air-dried
sam ple of a com m ercial w olfram ite ore.
LITERATURE CITED
(1) Cremer, F ., Eng. M in in g J ., 59, 345 (1895).
(2) D otreppe, G ., B ull. soc. chim. Belg., 38, 385 (1929).
(3) G rim aldi, F . S., and W ells, R . C ., I n d . E n g . C hem ., A n a l . E d .,
15, 315 (1943).
(4) G utbier, A., and W eise, G . L., Z. anal. Chem. 53, 426 (1914).
(5) H alberstadt, S., Ibid., 92, 86 (1933).
(6) H illebrand, W . F., and Lundell, G . E. F ., "Applied Inorganic
A n alysis” , p. 552, N ew Y ork, John W iley & Sons, 1929.
(7) Ilovork a, V ., Collection Czechoslov. Chem. Com m un., 10, 518
(1938).
(8) Jannasch, P ., and B ettges, W ., B er. 37, 2219 (1904).
(9) Jilek, A ., and Ryädnek, A ., Collection Czechoslov. Chem. Commun.,
5, 136 (1933).
(10) K afka, E., Z. anal. Chem., 52, 601 (1913).
(11) Knorre, G. v., B cr., 38, 783, 789, 2512 (1905).
(12) Lam bie, D . A ., A n alyst, 64, 481 (1939).
(13) Lefort, G ., Compt. rend., 92, 1461 (1881).
(14) Low’, A. H ., Eng. M in in g J ., 106, 27 (1918).
(15) M ellet, R ., H d v. Chim . A cta, 6, 656, (1923).
(16) M oser, L., and B laustein, W ., M onatsh., 52, 351 (1929).
(17) O ats, J. T ., Eng. M in in g J ., 144, N o . 4, 72 (1943).
(18) Otero, E., and M ontequi, R ., Anales, soc. espafl. f is . qulm ., 33,
132 (1935).
(19) O vcrholser, L. G ., and Y oe, J. H ., I n d . E n g . Chem ., A n a l .
E d ., 15, 310 (1943).
(20) Papafil, M ., and Cernatesoo, R ., A n n . S ei. Univ. J assy, 16, 526
(1931).
(21) Schoeller, W . R „ and Jahn, C „ A n alyst, 52, 504 (1927).
(22) S cott, W . W ., "Standard M ethods of C hem ical A n alysis” , 5th
ed., edited by N . II. Furm an, V ol. I, pp. 1007-9, N ew York,
D . Van N ostrand Co., 1939.
(23) T schilikin, M „ B er., 42, 1302 (1909).
(24) Y oe, J. H „
A m . Chem. Soc., 54, 4139 (1932).
(25) Y oe, J. H „ and H all, R . T „ Ibid., 59, 872 (1937).
(26) Y oe, J. H ., and Overholser, L. G ., I n d . E n g . Chem ., A n a l . E d .,
15, 73 (1943).
(27) Y oe, J. H ., and O verholser, L . G ., J . A m . Chem. Soc., 61, 2058
(1939).
(28) Y oe, J. H ., and Sarver, L . A ., “ Organic A n alytical R eagen ts” ,
p. 158, N ew Y ork, John W iley & Sons, 1941.
from a d isse rta tio n p re sen te d b y A . L e tc h e r Jones to th e
G ra d u a te F a c u lty of th e U n iv e rsity of V irginia in p a rtia l fulfillm en t of th e
re q u ire m e n ts for th e degree of d o cto r of philosophy, M ay , 1943.
A b s tra c te d
>
Dichromate Determination of Iron, Using the Silver Reductor
J . L . H E N R Y A N D R. W . G E L B A C H , State Co llege of Washington, Pullman, Wash.
a t a rate of a b o u t 30 ml. p er m inute. T he reductor was w ashed
w ith 150 ml. of 1 m olar hydrochloric acid added in sm all portions.
T o th e reduced solution 5 ml. of 85% phosphoric acid and 5 to 6
drops of 0.25% diphenylam ine sulfonic acid were added, an d the
solution was titra te d w ith potassium dichrom ate. T he results,
shown in T able I, deviated from those of th e W alden m ethod by
only 0.06% error.
Sim ilar analyses were m ade by introducing in th e first series
25-ml. portions of 0.01 m olar potassium dichrom ate, in th e second
0.1 molar m anganese chloride, and in th e th ird 0.1 m olar titanium
sulfate. C ontrols were run on each ion by passing th e solution
through th e reductor in th e absence of iron. In each case one
drop of 0.1 norm al potassium dichrom ate gave a d istin ct end
point.
E m ethod of W alden, H am m ett, and E dm onds (4) for th e
T Hdeterm
ination of iron using th e silver reductor and titra tin g
w ith eerie sulfate, although highly satisfactory, is expensive be­
cause of the relatively high cost of th e eerie sulfate and its very
high equivalent w eight. In th e titra tio n large q u an tities of sul­
furic acid are required. T h e dichrom ate m ethod as presented in
this paper overcomes these objections and seems to retain th e more
desirable features.
In the analysis of ores and alloys of iron th e common im purities
which m ay interfere w ith th e reduction and subsequent titra tio n
are chromium, manganese, molybdenum , titan iu m , an d vana­
dium . M olybdenum m ay be separated from iron by precipitating
iron as the hydrous oxide. In the silver reductor, chrom ium is
n o t reduced below the triv alen t sta te , while m anganese and tita ­
nium are n o t reduced a t all. V anadium is reduced to th e vanadyl
ion which is n o t oxidized b y either cerate or dichrom ate ions.
Table I.
Solution A dded
MU
I t has been shown th a t reliable results can be obtained in the
presence of chrom ium , manganese, or tita n iu m w ith acidities
ranging betw een 0.5 and 1.5 m olar. Checks were obtained w ith
titra tio n acidities as low as 0.1 m olar and as high as 2 m olar, b u t
a t th e extremes of concentration th e end points were n o t sharp.
T he results shown in T able I indicate th a t these ions do n o t in ter­
fere w ith th e analysis.
In the presence of high concentrations of vanadium an indis­
tin c t end p o in t was obtained, th e color change being from ligh t
green to gray. T itra tio n s were m ade in the presence of a wide
range of concentrations of vanadyl ion; u p to concentrations of
100 mg. of vanadium per 200 ml. of titra tin g volum e a v ery sharp
end p o in t was obtained. T h e deep violet color wras n o t shown as in
the absence of vanadium b u t th ere was; a d istin c t color change
from light green to deep blue. T h e results of these runs checked
well w ith th e accepted value for iron in th e sta n d ard ferric chloride
solution.
Effect of A d d e d Impurities
(V olum e of F eC li. 25 ml.)
N o. of
D e tn 8.
Fe Found
Gram
F ro m previous sta n d a rd iz a tio n 0
N one
25
O.OlJVf K jC rjO r
25
0 . 1 3 f M n C li
25
O .lA f TiCSOi)*
5
O .litfN a V O i
10
O .lilfN a V O .
15
0 A M N aV O i
20
0 . lJlf N aV O j
40
1 0 m l. of each
0
6
5
10
2
2
3
2
2
0.1 7 5 5
0.1 7 5 0
0.1 7 5 7
0.1 7 6 0
0 .1 7 5 6
0 .1 7 5 5
0 .1 7 5 4
0.1 7 5 4
0 .1 7 5 3
0 .1 7 5 4
A verage
D e v iatio n
P .p . 1000
‘o .6
-
1.1
2.8
0 .6
0
0 .6
0 .6
1.1
0 .6
• F ro m th e s ta n d a rd m eth o d of W ald en , H a m m e tt, a n d E d m o n d s.
T$ro samples obtained from th e B ureau of S tan d ard s were
analyzed: iron ore, B. of S. N o. 27, and ferrovanadium alloy,
B. of S. N o. 61. T h e sam ples were dissolved according to recom­
m ended procedures accom panying th e samples. T he rem oval of
m olybdenum from th e alloy was accomplished by tw ice precipi­
tatin g th e hydrated ferric oxide from am m oniacal solution. In
each instance th e acidity was adjusted to 1 m olar w ith hydro­
chloric acid and reduced as before. R esults of th e analyses are
shown in T able II.
T he oxidation potential in I N hydronium -ion concentration
of the system , VO<
-V O ++, is given as 1.2 volts (1), and th a t
of diphenylam ine sulfonic acid is 0.83 vo lt (8), which according to
W alden, H am m ett, and E dm onds (8) is sufficiently below th e
v anadate-vanadyl potential to give precise results. Adj usting the
a cidity to approxim ately IN and adding 5 ml. of 85% phosphoric
acid, one obtains a condition favorable to the titra tio n of ferrous
ion w ith potassium dichrom ate even in th e presence of vanadium .
Table II.
M ATERIALS
B . of S. S am ple
S tandard solutions of eerie am m onium sulfate and potassium
dichrom ate were prepared in th e usual m anner and standardized
against pure iron wire. T h ey were also checked against a gravim etrically standardized solution of ferric chloride, using stannous
chloride reduction.
A solution of ferric chloride standardized gravim etrically was
used in studying the behavior of th e dichrom ate-diphenylam ine
sulfonic acid titratio n in th e presence of th e ions of chromium,
m anganese, vanadium , and titanium .
T he effect of im purities was studied b y adding definite q u an ti­
ties of 0.01 m olar potassium dichrom ate, 0.1 m olar manganese
chloride, 0.1 m olar sodium vanadate, an d 0.1 m olar titan iu m
sulfate. T he la tte r was prepared by fusing 4 gram s of titan iu m
dioxide w ith 80 gram s of potassium hydrogen sulfate, dissolving
in 55 ml. of concentrated sulfuric acid, and diluting th e resulting
solution to 500 ml.
T he silver reductor was prepared in th e m anner described by
W alden, H am m ett, and Edm onds (4).
Iro n ore 27
F e rro v an a d iu m a llo y 61
Determination of Iron
N o . of
D e tn s .
4
2
Fe
Found
%
6 9 .3 0
52 .8 3
F e , B . of S.
C ertificate
%
69 .2 6
5 2 .8
A verage
D e viation
P .p . 1000
0 .6
SUMM ARY
P otassium dichrom ate, using diphenylam ine sulfonic acid as
indicator, is an oxidizing agent for th e d eterm ination of iron by
th e W alden silver red u cto r m ethod. M anganese, chromium,
and tita n iu m do n o t interfere. V anadium does n o t interfere in
concentrations of 100 mg. or less in 200 ml. of titra tin g solution.
A m ore economical m ethod results from th e use of only IN
hydrochloric acid instead of th e higher concentrations of sulfuric
acid required in th e eerie sulfate m ethod. T h e stab ility , purity,
low equivalent weight, and com paratively low cost of potassium
dichrom ate render it a very desirable oxidizing ag en t in this con­
nection.
METHOD
LITERATURE CITED
Analyses were m ade w ith 25-ml. portions of th e stan d ard ferric
chloride solution by th e m ethod of W alden, H am m ett, and
E dm onds (4), titra tin g w ith eerie am m onium sulfate. Sim ilar
analyses were then m ade by titra tin g w ith potassium dichrom ate.
A 25-ml. sam ple of standard ferric chloride solution was pipet­
ted, the acidity was adjusted to 1 m olar w ith hydrochloric acid,
and the final volume of 50 ml. w as passed through th e reductor
(1) K olthoff and Sandell, “ T extb ook of Q u an titative Inorganic
A n alysis” , p. 483, N ow York, M acm illan C o., 1943.
(2) I b id ., p. 493.
(3) W alden, H am m ett, and Edm onds, J . A m . Chem. Soc., 56, 59
(1934).
(4) Ib id ., p. 350.
49
Precise Measurement of Volum e in Titrimetric
A n alysis
W I L L I A M M . T H O R N T O N , JR., Lo yo la Co llege , Baltimore, M d.
A review of precise measurement of volume in titrimetric
analysis is presented. Detailed descriptions of the burets
and special techniques used by the author are given. Rep­
resentative standardizations indicate the reliability to be e x­
pected under the prescribed conditions.
expansiveness is a m a tte r of some im portance; hence its fu rther
stu d y would seem to be justified.
THERM AL EXPAN SIO N
Thiessen and his collaborators, an d also C happuis, have
determ ined th e density of w ater (“ ordinary w ater-substance” ) a t
various tem peratures w ith g reat exactness.
From th eir d a ta th e expansivity m ay be calcu­
lated. Accordingly, C ircular 19 of th e B ureau
of S tandards (5) enables one to correct th e ob­
served volume to w hat it would be a t 20° C —
the stan d ard tem perature for volum etric analysis
th roughout th e U nited S tates (24). T h e small
num bers th a t are to be added or subtracted
have been com puted on the assum ption th a t the
glass forming th e m easuring utensils has a coef­
~1
ficient of cubical expansion of 0.000025 per de­
T
gree Centigrade. (C ertain borosilicate glasses,
-\S
47, such as Pyrex, exhibit therm al expansions
m uch sm aller th an the foregoing.) These cor­
rections apply n o t only to w ater b u t also,
practically speaking, to sufficiently dilute aque­
ous solutions. F urtherm ore, a supplem entary
table (5, below T able 38) gives the percentage
increase in th e corrections for w ater to be applied
when stan d ard solutions of four common acids
and bases—nam ely, nitric and sulfuric acids
and sodium and potassium hydroxides—are under
consideration. T his increase is ab o u t 5 per cent
for th e before-named reagents when of O.IjV
concentration. In like m anner, it is b u t 3 per
cent for 0.1 Ar hydrochloric acid (44)—an alm ost
negligible increm ent. Y et th e higher th e n o r­
m ality of th e solution th e greater m u st be the
augm entation of th e tem perature corrections
for pure w ater.
As early as 1869, G erlach (13) studied th e ex­
pansion of aqueous solutions of acids and salts;
and in 1877 C asam ajor (6), utilizing M atthiesson’s d a ta for w ater, a ttem p ted to correct the
volume of his stan d ard solution for changes of
tem perature. Some 5 years later, Schulze (39)
determ ined the ra te of expansion for a good m any
-•
of th e b e tte r known voluis
metric solutions, and these
apparently reliable values
have been used to a con­
siderable extent. Schloesser (30, 31) and several
other authorities (7 ,9 ,1 0 ,
20, 25, 88) have contrib­
uted to the subject in
one w ay or another.
F inally, O saka (22) has
made available his exten­
sive investigations.
A S ID E from the choice of a suitable chemical reaction, or
/ \ th e resultant of a series of successive reactions, on which
to base a dependable titrim etric process, one is confronted w ith
th e responsibility of perform ing th e physical m easurem ents w ith
sufficient nicety. M ore specifically, if th e highest accuracy is
to be attained, th e following points m u st be given serious con­
sideration: (1) influence of tem perature change, (2) position of
the meniscus, 1.3) drainage o r afterflow, (4) errors of graduation,
(5) evaporation, and (6) po in t of complete reaction.
E xcept for the last two, these sources of u n certain ty can be
removed by weighing, instead of m easuring by volume, the
standard solutions. F o r this purpose, various forms of “ w eight”
o r “weighing burets” have been invented; and a few investi­
gators (I, 2, 15) have seen fit to tak e account of th e solution of
known concentration by weighing th e reaction vessel both before
an d after the titration.
If the loss of w ater (or oth er solvent) from stan d ard solutions
during long periods of storage be left o u t of account, it is safe
to assume th a t the evaporation taking place w ithin the ,time
required to effect a volum etric determ ination is of no signifi­
cance (19). On the contrary, rendering a decision as to the
equivalence point looms up as an unquestionably difficult ob­
jective in this kind of analysis, an d one in which th e weight
buret, as compared w ith the volum e buret, offers no advantage.
If the volume burets are properly designed, and used w ith
certain precautions, a m uch higher degree of accuracy m ay be
realized in dealing w ith a num ber of th e well known volum etric
solutions th an is commonly supposed. T he very fine w ork of
Ponndorf (26) m ay be cited in support of th e above contention,
and this constitutes by no m eans the sum to tal of the evidence
(cf. 1 1 , 1 4 , 31, 48).
A dm ittedly, when m any titratio n s have to be m ade, it is
expedient to feed th e evaluated solution from a large reservoir
directly into the buret. ¡Moreover, in th e case of solutions th a t
are sensitive to th e oxygen of th e air and have to be kept under
a nearly co n stan t pressure of an indifferent gas (4, 17), the
transfer to a short buret for weighing would probably lead to a
lowering of the tite r (28). W eighing the titra tio n flask
is not alw ays feasible: in some experim ents th e reaction m ixture
m ust be heated; it is often desirable to bubble an inactive gas
(carbon dioxide, for example) through th e te s t solution (43);
it is sometimes necessary to introduce an indicator during the
la tte r stages of the procedure. However, Lee (18) has developed
an ingenious weighing buret, wherein a known solution of titanous
sulfate m ay be obtained by shaking acidulated titan ic sulfate
w ith zinc amalgam . B u t such a device, owing to th e high
density of th e am algam (about 4 p er cent zinc, presum ably),
would seem to bo unduly heavy, besides being restricted to the
p reparation of only a sm all q u a n tity of th e reducing agent a t a
time.
T he chief lim itation to m easurem ent b y volume in precise
analysis is the lack of experim entally established d a ta regarding
th e therm al expansion of num erous very useful solutions. This
1
In th is connection, it
m ay be desirable to call
a tten tio n to a point th a t
is a p t to be overlooked.
W hen a n auxiliary re­
agent is added to a titri­
m etric
solution,
even
though it does n o t enter
into th e stoichiom etric re­
lations, it m u st needs alter
th e therm al expansion of
the liquid. M any such
cases m ight be cited, b u t a
50
< & >
Figure 1
January IS, 1944
ANALYTICAL
few will suffice. A ddition of potassium iodide increases the
solubility of iodine, acidification of am m onium sulfato-cerate
prevents hydrolytic decomposition of th e cerium compound, and
so on. In all candor, there seems to be a d earth of inform ation
concerning th e expansivity of th e m ore complex volum etric
solutions.
Besides dilatation or contraction, certain other physical
properties of an aqueous solution— namely, density, surface
tension, and viscosity (37)— undergo appreciable change when
th e concentration w ith respect to a dissolved substance is m ark­
edly increased. As these properties influence delivery, the
h abitual assum ption th a t th e volum e m easured a t a given tem ­
perature is equal to th a t of pure w ater m ay n o t be fully w ar­
ran ted in the case of the complex solutions discussed above.
BURETS
T he gratifying results obtained by Getz and the w riter (4 2 )
in standardizing 0.1 N barium hydroxide (measured from an
ordinary buret) against benzoic acid and p-nitrobenzoic acid,
respectively, led to the expectation th a t, w ith a special instru­
m en t of com paratively simple construction, it m ight be possible
to a tta in a very high degree of precision in certain titrim etric
analyses.
Two such burets (Figure 1) were procured. T hey were m ade
by E im er and Amend, of New Y ork C ity, in accordance with
definite specifications. E ach instru m en t consists essentially of
a middle wider portion, graduated in intervals of 0.2 m l., and two
narrow er portions, one above and one below, both m arked in
intervals of 0.05 ml. T he to ta l capacity is 50 ml., and the entire
graduated length is approxim ately 70 cm. E ach stopcock, which
is som ew hat larger th an usual, is composed of a blown (and
therefore hollow) plug and a neatly fitting shell. T he former
tu rn s in the la tte r w ith the u tm o st sm oothness. T his lowfriction bearing and the long handle (more th a n 5 cm.) perm it
th e operator to release fractions of drops when nearing th e end
point. These stopcocks do n o t leak a t all.
T he following statem ents and recom m endations regarding
th is ty p e of b u ret are held to be self-evident. No m a tte r w hat
m ay be the m agnitude of the sam ple, a reading will be obtained
in any titratio n , provided the required volum e does n o t exceed
50 m l., for the entire interval is g raduated—even th e shoulders.
If, however, the level of the liquid stands betw een th e 3- and the
45-ml. m arks, the determ ination should be taken as ten tativ e
only. B u t, w ith this d atu m in hand, it is a simple problem to
calculate by m eans of a proportion w h at increase in th e weight
of the sam ple will give rise to a reading somewhere betw een the
45- and th e 50-ml. graduations, where, on grounds of probability,
th e best observations are to oe had. (The linear value corre­
sponding to 1 ml. is ab o u t 34 m m . in th e narrow er portions of the
tube.)
In order to measure volum es som ew hat less th a n 45 m l., a
25-ml. N orm ax b u ret was provided. T his in strum ent is fashioned
from a straig h t tube, th e sm allest division is 0.1 m l., an d the
distance representing 1 ml. is 16.3 mm.
To m ake th e burets ready for use, a uniform , thin layer of a
suitable lubricant is spread betw een th e plug and th e shell of
th e stopcock. W ith th e plug in position B t, Figure 1, a small
am ount of carbon tetrachloride is poured into th e buret. T he
key is then turned through an angle of 90°, B ,t and th e issuing
liquid is caught in a sm all beaker. T h e je t being m om entarily
closed by tire index finger, th e sam e portion of the tetrachloride
is again introduced into the b u ret a t its to p opening; b u t on no
account should the key bo tu rn ed back to position B . a t this
juncture, for fear th a t a furrow will be cu t in th e film of lubricant
by th e solvent. These operations are repeated un til th e small
tubes have been washed free from visible greasy m a tte r; w here­
upon the buret is dried thoroughly b y aspiration.
T he well known “cleaning m ixture” (concentrated sulfuric acid
to which powdered potassium dichrom ate has. been added in
abundance) is then applied, th e preceding instructions as to the
two positions of the stopcock key (more particularly, th e plug
capillary) being adhered to in an equally conscientious m anner.
In calibrating the burets, th e gravim etric m ethod of Peffer
and M ulligan (24) was followed in a general way. However,
g reat care was taken to discharge th e w ater very slowly—not
faster than by rapid dropping from th e tip, even during th e initial
stages of the delivery— and th e settings and rem oval of hanging
drops were m ade w ith all necessary pains.
Experience has shown th a t very uniform volumes of pure w ater
m ay be obtained by slow delivery from a scrupulously clean
EDITION
51
buret. T his observation is in su b stan tial agreem ent w ith th e
words of Osborne and Veazey (28): “ B y lim iting th e rate o f
outflow th e residue and th e afterflow m ay bo m ade negligibly
sm all.”
The results obtained in calibrating th e three burets are given
in T able I. Judging from duplicate determ inations, it is reason­
able to hope th a t, in th e case of the special b u rets (E im er and
Amend 12 and 13, respectively), m easurem ents will agree to
w ithin 0.005 ml., and w ith th e 25-ml. b u ret (N orm ax 830),
to w ithin 0.01 ml.
Tabic 1.
In te rv a l
M l.
Calibration o f Burets
(V olum es corrected to 20° C.)
D elivery
M l.
B ou n d ed
M l.
V a lu e
13. an d A. B u re t N o. 12
O- 3
0-25
0-45
0 -4 7 .5
0-5 0
2.9 9 6
2.995
25 .0 0
4 5 .086
45.086
4 7 .606
4 7 .608
5 0 .080
50.082
2 .9 9 5 °
2 5 .0 0 b
4 5 .0 8 5 °
4 7 .6 0 5 °
5 0 .0 8 0 °
E . an d A. B u re t No. 13
0 - 2 .5
0-25
0-4 5
0 -4 7 .5
0-5 0
2.5 1 7
2.5 1 4
24 .8 4
4 5 .0 6 8
4 5 .068
4 7 .579
47.580
50.068
5 0 .067
2 .5 1 5 °
2 4 .8 4 &
4 5 .0 7 0 °
4 7 .5 8 0 °
5 0 .0 7 0 °
N orm ax B u re t N o. 830
0- 5
0 -10
0 -15
0-20
0 -25
4 .9 9 2
4.994
9 .9 9 5
9.9 9 4
14.998
14.996
19.999
19.999
2 4 .997
25.006
4 .9 9 *
9 .9 9 *
15.00*
20.00*
25.00*
° M ean valu e ro u n d e d off to n e arest 0.005 ml.
* M ean valu e ro u n d e d off to n e arest 0 .0 1 ml.
STA N D A RD IZA TIO N OF SOLUTIONS
B y way of forming a fairly satisfactory estim ate as to the p re­
cision of the physical m easurem ents in really nice volum etric
analysis, wherein the titra tio n s are made w ith volume burets,
three fam iliar titrim etric solutions, of O .lA strength, were
standardized in accordance w ith supposedly accurate m ethods.
T he solutions were: (1) potassium perm anganate, (2) sodium
tetrab o rate, or borax, and (3) sodium hydroxide.
In th e first set of experim ents, O.IA solutions of potassium
perm anganate were evaluated against th e certified sodium
oxalate of th e N atio n al B ureau of S tan d ard s (Standard Sam ple
40c), th e procedure of Fow ler and B rig h t (12) being followed in
all its m inutiae.
T h e solutions were prepared by dissolving high-grade crystals
of potassium perm anganate in th e requisite q u an tity of distilled
w ater, aging for more th a n a m onth, and filtering very slowly
through glass frit of fine tex tu re under th e influence of mild
suction.
T he clear solution was introduced into th e b u re t b y m eans o f
the “ lift” (Figure 2), an ap p aratu s modeled a fte r a sim ilar con­
trivance b y Osborne an d Veazey (23). T h e suction being tu rn ed
on and properly regulated b y obvious m anipulations of the
stopcocks, S an d T , an y desired q u a n tity of th e liquid m ay be
pulled u p from th e storage b o ttle into a large pipet, P , held th ere
a t will, and finally allowed to en ter th e b u ret. T ransferring in
this m anner loaves th e neck of th e b o ttle clean—a condition to
be wished for in th e case of a perm anganate solution.
T he tem perature of th e solution in th e bu ret was taken by
inserting a slender therm om eter w ith enclosed scale, ju s t before
setting th e meniscus upon th e 0-ml. g rad u atio n ; and again a t
the end of th e titratio n , by running th e rem aining liquid in to a
52
INDUSTRIAL
Table II.
Solution
No.
I
ENGINEERING
Potassium Permanganate against Sodium O xalate
D a te
J u n e 25
J u n e 26
J u n e 27
J u ly 8
J u ly 9
NaaCiOi
Gram
KM nO*
ML
N o rm a lity
0.3 1 1
0 .3 1 2
0 .3 1 3
0 .3 1 0
0.3 1 1
4 5 .3 2 5
4 5 .4 3 5
4 5 .5 7 0
4 5 .2 2 5
4 5 .3 2 0
M ean
0 .1 0 2 6 0
0 .1 0255
0 .1 0257
0 .1 0259
0 .1 0 2 5 8
0 .1 0 2 5 8
M ean
0 .0 9967
0 .0 9966
0 .0 9967
0 .0 9966
0 .0 9966
M ean
0 .0 9940
0 .0 9939
0 .0 9940
M ea n
0 .0 9947
0.09951
0 .0 9 9 4 9
55
15
15
85
45
II
M ay 24
M ay 24
A ug. 5
A ug. 5
0 .3 1 6
0 .3 1 5
0 .3 1 6
0 .3 1 6
65
60
85
60
4 7 .4 2 0
4 7 .2 7 0
4 7 .4 5 0
4 7 .4 1 5
III
A ug; 16
A ug. 19
0 .3 1 7 05
0 .3 1 7 85
4 7 .6 0 0
47.7 2 5
IV
AND
J a n . 18
J a n . 20
0 .3 1 7 05
0 .3 1 7 90
4 7 .5 7 0
4 7 .6 7 5
A
Vol. 16, No. 1
A slight lack of uniform ity in th e results w as apparent, owing,
presum ably, to a n end p o in t th a t is som ew hat difficult to judge;
and, because of this u n certain ty , th e tests w ith borax are not
tab u lated in th e present report. N evertheless, as th e average
deviation of a determ ination was less th a n 3 p arts per 10,000,
th is p a rt of th e w ork m ay still be regarded as acceptable.
I n th e th ird set of experim ents, a 0.11V solution of sodium
hydroxide was prepared, reasonably free from carbonate, by the
m ethod of Cowles (5); it was protected against atm ospheric
carbon dioxide in th e usual m anner. A second basic solution, in
all respects sim ilar to th e first, was also em ployed. T h e sodium
hydroxide was standardized against 0.2N hydrochloric acid,
which, in its tu rn , had been obtained by ap propriately diluting
constant-boiling acid.
F o r alkaline Solution I (see T able I I I), a reference color was
established b y adding 1 drop of m ethyl red (0.2 per cent alcoholic
solution) to 100 ml. of w ater. F o r Solution II, 25 ml. of 0.2N
sodium chloride were diluted to 100 m l., and th e before-nam ed
q u a n tity of th e indicator was introduced. T h e la tte r m ixture
is theoretically correct, b u t it differs very little from th e form er
in tin t.
Tw enty-five m illiliters of th e 0.2Ar hydrochloric acid were
m easured from th e small b u ret (N orm ax 830), 25 ml. of w ater
an d a single drop of th e m ethyl red were added, and th e titratio n
was perform ed w ith th e 0.11V sodium hydroxide (in b u re t 12),
th e stan d ard color being m atched as well as possible.
T h e results of th is last series of experim ents, w herein all the
solutions were m easured by volume, are given in T able II I.
Table III.
Y
CHEMISTRY
Solution
No.
Sodium H yd ro xid e against H yd ro ch lo ric A c id
D a te
0.20066 N
HC1
M l.
N aO H
ML
N o rm a lity
I
J u ly 19
J u ly 19
2 4 .9 5
24 .9 4
4 7 .525
4 7 .5 0 0
M ean
0.10534
0 .10536
0.10535
II
O ct. 13
O ct. 13
O ct. 14
24 .9 8
24 .9 8
24 .9 8
4 7 .165
4 7 .145
47.145
M ea n
0.10628
0.10632
0.10632
0.10631
CONCLUSIONS
sm all test tube. T he m ean of the tw o values was arbitrarily
accepted, and th e volum e was ad ju sted on th e assum ption th a t
the therm al expansion is th e sam e as th a t of p u re w ater {22).
E xcellent observations of th e meniscus were m ade by utilizing
a device sim ilar to the one proposed by W hite (46), despite the
purple color of the solution.
I h e tin y excess of perm anganate unavoidably added was
estim ated by m atching the pink color of th e te st solution in th e
usual way. B u t it is doubtless m ore accurate to arrive a t the
correction iodom ctrically, as suggested by B ray (5).
T he results obtained w ith potassium perm anganate are set
forth in T able II.
In the second set of experim ents, th e prim ary stan d ard was
constant-boiling hydrochloric acid (41), and from th is a 0.1 N
acid was obtained. A solution of sodium tetrab o rate, likewise
O.IAT, was prepared from recrystallized borax (16). F o r m aking
the standardizations, the borax solution was weighed in a stop­
pered flask and was then titra te d w ith the hydrochloric acid
(in b u ret 12), m ethyl red serving as indicator.
B arring such rem arkably accurate results as those a ttain ed
in iodom etry by W ashburn (45), who used a w eight b u re t of the
R ipper ty p e (27), th e experim ental d a ta reported herein compare
favorably w ith m any to be found in th e scientific literature,
which were obtained by weighing, an d n o t b y m easuring the
volume of, th e stan d ard solution.
T he delivery of a liquid from a graduated tu b e is a t best a
som ewhat empirical operation, and, naturally, th e best m easure­
m ents m ay be expected when th e physical properties of the
solution under consideration are n ear to those of th e liquid
(usually w ater) employed in calibrating th e instrum ent. M ore
especially, th e therm al expansion of th e titra n t should bo known,
a t least approxim ately. These and m any oth er points have
been discussed and elucidated by such prom inent m etrologists as
Schloesscr (29-37), Osborne an d Veazey (23), and S to tt (40), to
whose writings th e interested reader is referred for fu rth er in­
form ation and advice.
LITERATURE CITED
(1) A shley, S. E . Q., and H u lett, G. A ., J . A m . Chem. Soc., 56,
1275 (1934).
(2) Bezzcnberger, F . K ., Ibid., 39, 1321 (1917).
(3) B ray, W . C„ Ib id ., 32, 1204 (1910).
(4) Brennecke, E ., " N ew er M eth od s of V olum etric A n alysis” , p.
131, N ew York, D . V an N ostran d C o., 1938.
(5) Bur. Standards, C irc. 19, T ab le 38 (1924).
(6) Casaraajor, P ., Chem. N ew s, 35, 96, 130, 160, 170 (1877); 38,
137 (1879).
(7) C ouvee, W . J., Chem. W eekblad, 23, 550 (1926).
(8) Cowles, H . W ., Jr., J . A m . Chem. Soc., 30, 1192 (1908).
(9) D ebrun, C „ A n n . fa ls., 7, 407 (1914).
(10) D ealin g, H . C ., J. I n d . E n q . C h e m ., 8, 451 (1916).
(11) D uranton, B ., A n n . chim. anal. chim. a p p l., 20, 257 (1938).
January 15, 1944
ANALYTICAL
(12) F ow ler, R . M ., and B right, H . A ., J . Research N a tl. B u r. Stan d­
ards, 15, 493 (1935).
(13) Gerlach, G . T ., Z . anal. Chem., 8, 245 (1S69).
(14) G ol’tz, R „ J . Gen. Chem. (.U .S .S.R .), 5, 779 (1935).
(15) H eath , W . P ., U . S. P a ten t 1,358,950 (N o v . 16, 1920).
(16) H urley, F . H ., Jr., I n d . E n g . C h e m ., A n a l . E d ., 8, 220 (1936);
9, 237 (1937).
(17) K necht, E ., and H ibbert, E ., “ N ew R ed u ction M eth od s in
V olum etric A n alysis” , 2nd ed., p. 62, N e w York, Longm ans,
G reen & Co., 1925.
(18) Lee, K . W ., J . Chem. Soc. Ja p a n , 53, 25 (1932).
(19) M ellon, M . G ., P roc. In d ia n a A cad. S ei., 32, 159 (1922).
(20) M eurice, R ., A n n . chim . anal. chim . a p p l., 21,202 (1939).
(21) M ika, J ., Z . anal. Chem., 96, 401 (1934).
(22) O saka, Y ., J . Tokyo Chem. Soc., 32, 450 (1911); 40, 424 (1919).
(23) Osborne, N . S., and V eazey, B . H ., B u r. Standards B ull. 4,
553 (1908).
(24) Peffer, E . L., and M ulligan, G . C ., N a tl. Bur. Standards, Circ.
434 (1941).
(25) P ellet, H ., A n n . chim. anal., 20, 97 (1915).
(26) Ponndorf, W ., Z . anal. Chem., 84, 289 (1931); 85, 1 (1931).
(27) Ripper, M ., Chem .-Ztg., 16, 794 (1892).
(28) R osem an, R ., and T hornton, W . M ., Jr., J . A m . Chem. Soc., 57,
328, 619 (1935).
(29) Schloesser, W ., Chem .-Ztg., 28, 4 (1904).
EDITION
(30)
(31)
(32)
(33)
(34)
(35)
(36)
(37)
(38)
(39)
(40)
(41)
(42)
(43)
(44)
(45)
(46)
(47)
(48)
53
Ibid., 29, 509 (1905).
Ib id ., 33, 1105 (1909).
Schloesser, W ., Z . anal. Chem., 46, 392 (1907).
Schloesser, W ., Z . angew. Chem., 16, 953, 977, 1004, 1061 (1903).
Ib id ., 17, 1608 (1904).
Ib id ., 21, 833 (1908).
Schloesser, W ., Z . Chem. A p p a ra t., 3, 353 (1908).
Schloesser, W ., and G rim m , C., Chem .-Ztg., 30, 1071 (1906).
Schoorl, N „ Chem. W eekblad, 23, 581 (1926).
Schulze, A., Z. anal. Chem., 21, 167 (1882).
S to tt, V ., "V olum etric G lassw are” , L ondon, H . F . and G .
W itherby, 1928.
T hornton, W . M ., Jr., and Christ, C. L ., I n d . E n g . C hem .,
A n a l . E d ., 9, 339 (1937).
T hornton, IV. M ., Jr., and G etz, D oroth y, A m . J . S ci., 9, 176
(1925).
T hornton, W . M ., Jr., and lV ood, A . E ., I n d . E n g . C h e m ., 19,
150 (1927).
T readw ell-H all, “A n alytical C h em istry” , 8th ed., V ol. II, p.
479, N ow Y ork, John W iley & SonB, 1935.
W ashburn, E. W ., J . A m . Chem. Soc., 30, 31 (1908).
W h ite, E. P., I n d . E n g . C hem ., A n a l . E d ., 10, 668 (1938).
W ichers, E., F in n , A. N ., and C labaugh, W. S., Ib id ., 13, 419
(1941).
Zotier, V ., B u ll. sci. ph arm acol, 24, 298 (1917); 25, 274, 357
(1918).
Ethyl b is-2/4-Dinitrophenylacetate/ a New p H Indicator
D eterm in ation o f Sa p o n ific a tio n E q u iv a le n ts in D a rk -C o lo r e d O ils
E.
A . F E H N E L a n d E. D . A M S T U T Z
Wm. H . Chandler Chem ical Laboratory, Lehigh University, Bethlehem, Penna.
A new acid-base indicator, ethylbis-2,4-dinitrophenylacetate, has been studied and
its preparation described. The p H range over which the change from colorless to
deep blue occurs is found to be from 7.5 to 9.1 (p K ca. 8.3), making the indicator
suitable for most titrations which are ordinarily performed with phenolphthalein.
The indicator gives an accurate end point in amber-colored solutions where the
phenolphthalein end point is not visible, and it is therefore recommended for use
in the determination of acid numbers and saponification equivalents of dark-colored oils.
T H E course of a n investigation on th e preparation and
INproperties
of su b stitu ted phenylm alonic esters, th e com­
pound identified by R ich ter (6) as ethylbis-2,4-dinitrophenylacetate was obtained, an d th e opportu n ity w as tak e n to study
suited to th e titratio n of orange- an d red-colored solutions in
which th e phenolphthalein end p o in t is n o t visible. Because of
the frequency w ith w hich dark-colored oils are encountered in
organic analysis an d because of th e inadequacy of present-day
m ethods of perform ing titra tio n s w ith these products (1-6),
d a ta were obtained to determ ine th e applicability of this indicator
for such work. I t was found th a t satisfactory results are o btain­
able even w ith extrem ely dark-colored oils an d th a t th e ac­
curacy of th e m ethod approaches th a t of ordinary phenolphtha­
lein titrations.
PREPARATION A N D PROPERTIES
N 02
its indicator properties. R ich ter noted th e color change of this
com pound from colorless in acid to intense blue in base and ac­
cordingly suggested its possible application as a n acid-base
indicator, b u t he reported no fu rth er studies along this line.
T he au th o rs’ results show n o t only th a t th e com pound m ay serve
as a satisfactory su b stitu te for phenolphthalein in th e titra tio n
of m ost weak acids w ith strong bases, b u t also th a t it is specially
T he indicator is n o t difficult to prepare and, while th e yield
is n o t good, enough of th e com pound m ay be obtained from 0.5
mole of dinitrochlorobenzene to serve for thousands of titratio ns.
T h e following procedure furnished 19.0 gram s (18.1 p er cent of
theoretical) of recrystallized m aterial sufficiently pure for use
as an indicator.
One-half gram -atom (11.5 gram s) of sodium was dissolved in
200 ml. of absolute alcohol in a 1-liter three-necked flask fitted
with a reflux condenser, m otor stirrer, and dropping funnel. T he
solution was cooled and 0.25 mole of diethyl m alonate wras added
dropwisc, w ith stirring, over a 30-m inute period. Stirring was
continued for an o th er 10 m inutes and a h o t solution of 0.5 mole of
2,4-dinitrochlorobenzene in a b o u t 200 ml. of absolute alcohol
was then added over a 30-m inute period. T h e deep red reaction
54
INDUSTRIAL
AND
ENGINEERING
m ixture was refluxed, w ith stirring, for 4 hours and allowed to
sta n d overnight. E nough w ater was th e n added to th e resulting
■olive-brown solution to bring the volum e up to a b o u t 1200 ml.,
a fte r which the_ solution was acidified w ith a little concentrated
sulfuric acid, stirred for 20 m inutes, th en allowed to stan d for 30
m inutes. T h e w ater layer was decanted and th e residual ta r
w ashed twice w ith w ater and finally w ith successive 200-ml. por­
tions of alcohol until a gran ular black m ass was obtained. By
repeated washing w ith h o t alcohol an orange solid was finally
o btained, 'which, after recrystallization from benzene, gave pale
yellow' crystals m elting a t 150-153.5° C. (uncorreeted).
(A sm all am ount of a second compound wfith indicator proper­
ties can be isolated from th e alcohol washings; it appears to be
th e m onophenylated ester, ethyl-2,4-dinitrophenylm alonatc,
which would be expected as an interm ediate in the form ation of
th e diphenylated ester. A fter recrystallization from alcohol this
com pound m elted a t 48° C. and dissolved in d ilute base to give a
deep red color. T h e change from colorless to red was found to
occur over a pH range of ca. 7.2 to 9.0, b u t th e com pound is n o t
recommended as a su b stitu te for phenolphthalein because th e
change is less distinct.)
Table I.
CHEMISTRY
Vol. 16, No. 1
Saponification Equivalents Determined for Colored O ils
Oil
P ropyl benzoate
P ro p y l be n zo a te
P ro p y l benzoate
E th y l-3 -c a rb e th o x y -2 -fu ra n ac e ta te
E th y l-3 -c a rb e th o x y -2 -fu ra n ac e ta te
Oil A
Oil A
Oil A (electrom etric)
Oil B
Oil B
A pproxi­
m ate
C olor
Saponification E q u iv a le n t
D
en sity ,
O bserved
U n­
C alcu ­
*£
colored
C olored
la te d
it
164
164
164
113
113
...
165
167
166
165
166
164
113
115
256
254
256
276
270
1.8 5
1.85
1.8 5
1.85
1.85
1 .5 5
1.55
1.55
6 0 .3
6 0 .3
these conditions, saponification equivalents were determ ined for
sam ples of pure propyl benzoate and ethyl-3-carbethoxy-2furanacetate b o th before and a fte r th e addition of an artificial
coloring m aterial. F or th is purpose th e addition of a few
m illiliters of a highly colored n eu tral caram el solution prepared
from pure cane sugar was found to be satisfactory.
A bout 0.3 gram of th e carefully distilled ester was weighed into
a 125-ml. Erlenm eyer flask an d th e m ixture was saponified w ith
15.00 ml. of a sta n d a rd alcoholic sodium hydroxide solution ac­
cording to th e procedure of Shriner and Fuson (7). T h e saponi­
fication m ixture was titra te d alm ost to n eu trality w ith stan d ard
acid, 5 drops of th e indicator solution were added,_ and th e end
point w'as determ ined b y adding an excess of acid and backtitratin g w ith stan d ard alkali to th e first blue tinge. Several
m illiliters of aqueous caram el were th en added to th e alkaline
solution an d th e end p o in t w as redeterm ined in the dark green
m ixture b y adding acid un til the am ber caram el color ju s t re­
turned. In th e dark-colored solutions the end point w'as ap­
proached from th e basic side, n o t from th e acid side as in th e case
of th e uncolored solutions.
Figure 1.
Indicator Range» in Titration of Potassium A c id
Phthalate with Sodium H yd ro x id e
T h e exact range over which the color change occurs was d eter­
m ined b y titratio n of a buffered hydrochloric acid solution w ith
sodium hydroxide, using th e glass electrode to m easure pH .
A bout 5 drops of a satu rate d solution of th e indicator in 50-50
acetone-ethyl alcohol were added to each 100 ml. of solution to
be titra te d . A t pH 7.5 the first blue tinge appeared in th e solu­
tion, an d a t pH 9.1 the change to deep blue was complete. (The
indicator has been successfully used in th e au th o rs' analytical
laboratories by stu d en ts who, b y reason of deficiencies in color
vision, do n o t easily recognize th e phenolphthalein end point.)
T he pK of th e indicator is therefore ab o u t 8.3. Figure 1
shows the indicator range given by ethylbis-2,4-dinitrophenyla cetate com pared w'ith th a t given by phenolphthalein in the
titra tio n of potassium acid p h th ala te w ith sodium hydroxide.
T h e sam e reversible color change was observed to occur in
anhydrous solvents such as absolute alcohol, d ry ether, and dry
benzene.
TITRATION O F DARK-COLORED OILS
Since the basic color of cthylbis-2,4-dinitrophenylacetate,
in stead of being m asked in amber-colored solutions as is the
case wfith phenolphthalein, becomes m ore pronounced as a result
of th e com plem entary spectral relations, th e end po in t is
easily visible even in titratio n s involving dark-colored oils.
In order to check the accuracy of the end point determ ined under
In T able I th e saponification equivalents th u s determ ined
for uncolored propyl benzoate and for artificially colored propyl
benzoate an d ethyl-3-carbethoxy-2-furanacetate are com pared
w ith th e tru e values calculated from th e m olecular -weights of
th e esters. T h e agreem ent betw een th e calculated an d the
observed values Is sufficiently good to prove th e usefulness of
th e procedure as an analytical m ethod.
Two heat-bodied linseed oils, one of which (oil B, T able I) was
alm ost opaque w hen viewed by daylight in thicknesses greater
th an 5 cm., were studied in order to determ ine the reproducibility
of th e results under actu a l working conditions. Oil A was also
subjected to eloctrom etric titratio n , th u s affording an additional
chock on the accuracy of th e results. T h e following procedure is
recom m ended as a working m ethod.
A sam ple of 0.2 to 0.4 gram of th e oil was saponified as before,
and a fte r reducing th e alkalinity of th e m ixture b y titra tin g al­
m o st to n eu trality w ith 0.25A hydrochloric acid, 5 drops of
sa tu ra te d solution of th e indicator were added. M ore acid was
then run in un til th e d ark green color of th e indicator ju s t dis­
appeared. Two m ore determ inations of th e end point were th en
carried o u t on th e sam e sam ple b y adding a few m illiliters of
sta n d a rd sodium hydroxide solution an d b ack -titratin g w ith acid.
T h e results of th e triplicate determ ination were th en averaged
and th e saponification equivalent was calculated. T h e electrom etric titra tio n on a saponified sam ple of oil A was perform ed in
the usual m anner, using a B eckm an p H m eter and a glass elec­
trode.
As can be seen by reference to T able I, th e values for oil A
obtained by m eans of th e relatively simple indicator m ethod agree
w ith th e value obtained from th e m ore elaborate electrom etric
m ethod, and even th e v ery high color density of oil B does not
seriously interfere w ith th e precision of th e determ ination.
T h e approxim ate density of th e am ber color of b o th th e
n atu rally an d th e artificially colored oils is recorded in th e last
column of T ab le I w ith reference to 0.017.1/ potassium dichrom ate solution arb itrarily assigned a value of unity. The
ANALYTICAL
January 15, 1944
lig ht absorption was m easured by comparison of th e various oils
w ith th e stan d ard on a K lett-Sum m erson photoelectric colorim eter
w ithout a filter. W ith th e in stru m en t ad justed to give a scale
reading, B , of zero for th e reference stan d ard , th e ratio of th e
lig ht absorbed by the oil, i„ to th a t absorbed by th e stan d ard , t.,
w as calculated from the equation
or
l o g ^ = 0.002/E
Ik
4%- —, antilog 0.002/E
h
T h u s oil* B has a light-absorbing power approxim ately 60
tim es th a t of 0.017/1/ potassium dichrom ate solution.
EDITION
55
ACKNOW LEDGM ENT
T he au th o rs are indebted to E arl J . Serfass for m any helpful
suggestions concerning th e analytical procedures.
LITERATURE CITED
(1) Am . Soc. T estin g M aterials, Standards, P art II I, pp. 954, 958
(1942).
(2) M cllh in ey, J. Am. Chem. Soc., 16, 408 (1894).
(3) N ational T echnical Laboratories, pH Bull. BIO.
(4) Pom eranz, Seifensieder-Ztg., 45, 578 (1918); Chem. Zentr., 90, II,
647 (1919).
(5) Pschorr, Pfaff, and Berndt, Z. angew. Chem., 34, 334 (1921).
(6) R ichter, Ber., 21, 2470 (1888).
(7) Shriner and F uson, “S ystem atic Id en tification of Organic Com ­
p ou n ds” , 2nd ed., Procedure B , p. 118, N ew Y ork, John W iley
& Sons, 1940.
Pycnometer for Volatile Liquids
Control of Diffusion as an A id in Precision Pycnometry
M . R. L IP K IN , J . A . D A V I S O N 1, W . T. H A R V E Y , a n d S. S. K U R T Z , J r .
Sun O i l Com pany, Marcus H o o k and N orw ood, Penna.
This paper reports the design of a pycnometer which is especially
well suited for the determination of density on 5 ml. of volatile liquid
with an accuracy of =*=0.0001 gram per ml.
T H E R publication in th e field of density determ inations
FUmRight
seem unnecessary in view of th e large num ber of de­
vices which have already been described. However, m ost of
these instrum ents were designed for a special purpose and a
simple pycnom eter for general use in hydrocarbon analysis, and
particularly for obtaining precise densities on m aterials as vola­
tile as pentane, has n o t previously been described.
M any general references (I, 9, 16, 19, 90, 24, 29) are available
on the determ ination of density, an d Irv in g {16) gives a general
discussion of th e m ethod of floating equilibrium . U nusual m eth­
ods include m easuring th e frequency of acoustic vibratio n {17)
which is dependent on the den sity of th e surrounding m edium.
E specially interesting are th e balanced column m ethods {4, 7 ,1 1 ,
13) w hich elim inate th e use of a n an alytical balance, and which
m erit fu rth er investigation.
T h e m ore conventional pycnom eters, such as th e p ip et types
{8, 10, 33) and th e Sprengel-Ostwald ty p e w ith its modifications
{6, 19, 20, 23, 27), often involve some kind of closure or cap to
prevent rapid vaporization of volatile m aterials. T hese caps are
in p a rt successful, b u t are n o t entirely satisfacto ry w ith m aterials
as volatile as pentane and ether.
In 1884 P erkin {22) recognized th e ad vantage of having both
m enisci rem ote from th e ends of th e pycnom eter arm s, so th a t
vaporization would be som ew hat hindered. T h is principle is in­
volved in the design of several flask-type pycnom eters {2, 3, SO,
31) and m any capillary-arm pycnom eters {6, 12, 21, 22, 25, 26).
However, th e control of diffusion b y th e use of an unfilled capil­
lary arm has n o t been generally recognized and has never, as far
as th e authors are aware, been discussed on a q u a n tita tiv e basis.
P arker and P ark e r {21) also used th e principle of siphoning to
fill th eir pycnom eter. Siphoning has th e definite advantage over
filling by a suction technique th a t it reduces th e loss of th e more
volatile com ponents of gasoline.
B oth upper side arm s of th e U -tube are calibrated in scale divi­
sions from 0 to 8.0 w ith ten interm ediate scale divisions between
each m ajor scale division. E ach m ajor scale division equals 10
mm. T h e capillary tu b e should have a 0.6- to 1.0-mm. bore
and should n o t be over 6 mm. in outside diam eter. T he pyc­
nom eter is sim ilar to those of Shedlovsky and Brown {26) and
R obertson {25), which feature two calibrated stem s an d 25-ml.
capacity. T h e present pycnom eter, however, is of m uch sm aller
capacity and is easily handled on an analytical balance. I t is
constructed w ith a bulb in only one side arm , which allows the
ycnom eter to be filled easily an d avoids loss of volatile products
y th e slow ra te of diffusion through unfilled capillary arm s.
PyCNOMETER
T he pycnom eter which has been developed (Figure 1) has been
in use in these laboratories for a b o u t 4 years. I t is of Pyrex and
consists of a 0.3-, 1-, 3-, or 5-ml. bulb blown in one side arm of a
capillary U -tube. A bout 2 cm. of th e upper end of th e other
side arm are b en t over to form a hook for filling th e pycnom eter
by capillary action and for hanging th e pycnom eter in th e balance._
T his hook is a self-filling device, previously described b y H ennion {14). T h e liquid is first draw n in to th e pycnom eter by
capillary action and th e pycnom eter th en fills by siphoning.
1 P re se n t ad d ress, U . S. R u b b er C o m p an y , P assaic, N . J.
Figure 1. Pycnometer
Capillary bore 0.6 to 1.0 mm., outside diameter 6 mm. maximum,
material Pyrex, total weight not to exceed 30 grams
56
INDUSTRIAL
AND
ENGINEERING
T he d a ta showing the restraining effect of diffusion thro u g h a
capillary on ra te of evaporation, sum m arized in Figure 2, were
obtained using a prelim inary design in which both arm s of the
pycnom eter were straight. R ecent tests w ith th e pycnom eter
having one arm b en t are sim ilar. T o ta l length referred to is
th e sum of the unfilled capillary in b o th arm s of th e pycnom eter
an d the rate of evaporation is th e to ta l rate from b o th arm s of the
pycnom eter in m illiliters per m inute. T he rates given in Figure
2 are based upon d a ta obtained during Ju ly and A ugust in a hot,
d rafty laboratory. T herefore these rates are probably m axi­
m um since heat an d drafts on th e exposed ends of th e pycnom ­
e ter both increase vaporization. Howrever, these rates are suffi­
ciently accurate to establish th e length of th e unfilled capillary
necessary to avoid volatilization. U nder m ore favorable condi­
tions th e evaporation rates are som ew hat lower th a n shown in
Figure 2.
I t is clear from Figure 2 th a t if to ta l length of unfilled capillary
is over 1 0 cm., the rate of diffusion is so low th a t th e vapor loss
from th e pycnom eter becomes negligible.
T he specification for these pycnom eters is 0.6- to 1.0-mm.
inside capillary diam eter. Since th e d a ta were obtained on a
pycnom eter of 1 .0 -mm. capillary, the evaporation losses will n ot
be greater th an indicated in Figure 2 .
CHEMISTRY
Vol. 16, No. 1
dried by suction. T h e final rinse should be w ith good clean ben­
zene. Acetone is n o t recom m ended because it frequently con­
tains nonvolatile residue. If th e outside of the pycnom eter is
d irty or oily, it is rinsed w ith benzene and dried b y gently wiping
w ith cheese cloth or o th er suitable m aterial. [Static charge in­
troduced by wiping th e pycnom eter w ith a dry cloth is a p t to
cause errors in weighing, especially in cold d ry w eather. Before
hanging pycnom eter on wire hook in th e balance, observe w hether
ycnom eter exerts a ttrac tio n for the wire hook by touching the
ook w ith th e pycnom eter and gently draw ing it aw ay. S tatic
charge can usually be nullified by touching th e side of th e pycnom cter lightly w ith one’s fingers. If th is procedure will not
rem ove th e static charge, th e h um idity of th e balance room
should be increased u n til the difficulty disappears (about 60 per
cent relative h um idity).] T h e w eight is th e n obtained w ithin
=±=0 .0 0 0 1 gram on a good analytical balance, a fter allowing the
pycnom eter to come to balance case tem perature.
CALIBRATIO N OF PYCNOMETER
T h e pycnom eter calibration is based on the density of w ater
an d is checked w ith a pure hydrocarbon such as benzene. [Ben­
zene m ay be easily purified Dy th e following procedure: Place 5
gallons of commercial c.p. benzene in a can surrounded w ith 5 cm.
(2 inches) of felt, and place th e can in a cold room a t a b o u t
—30° C. ( —22° F .), or if this is n o t available, in a box chilled
w ith dry ice. Allow the benzene to freeze w ithout ag itatio n until
only about 1000 cc. of benzene rem ain uncrystallized. P our off the
im pure benzene, m elt the crystallized benzene and recrystallize.
E ach crystallization should tak e 3 to 5 days. F iv e or six such
slow crystallizations will usually give benzene having a freezing
p o in t of 5.51° C. {SI), and density and refractive index agreeing
w ith th e b est literatu re values.]
T he volume a t 20° C. of w ater free from a ir (boiled and cooled
w ithout agitation shortly before use) is obtained near th e top, b ot­
tom, and middle of the scales of the pycnom eter. T h e volum e
a t 20° C. is calculated by dividing th e w eight of w ater by
0.99823. A calibration curve is draw n, plottin g th e sum of th e
scale divisions on both arm s as th e abscissa and th e volum e in
m illiliters as the ordinate. All points m u st fall on th e sam e
stra ig h t line, w hich is the calibration curve for th e pycnom eter.
N o correction is m ade for air buoyancy in th e w ater calibration.
In stead, a correction, C, given b y E q u atio n 1, is added to th e
observed value of the density to tak e care of all buoyancy errors
in which D 0 = density in air. T his correction
C = 0 .0 0 1 2 X (1 - D 0)
(1)
is based on the use of an average value of 0 .0 0 1 2 gram per ml. for
th e density of air {29), and is applicable to all types of weights,
provided th a t w eights of the sam e density are used in b o th the
calibration an d the density determ inations.
T h e use of a sealed counterpoise has n o t been recommended,
since th e to tal volume of th e 5-ml. pycnom eter is only ab o u t 15
ml. and a variation in a ir den sity of =±=0.00005 gram per ml. will
give a change in buoyancy of only 0.00075 gram . Since a b o u t 4
gram s of liquid are weighed, this buoyancy v ariation corresponds
to about =*=0.00015 in the density of the sample. T his value is
probably extreme. F o r m ore accurate work, especially w ith pyc­
nom eters of larger size, a sealed counterpoise can be used to com­
pensate accurately for buoyancy of th e a ir on th e pycnom eter it­
self. T h e buoyancy effect can then be corrected using an accu­
ra te value for th e density of air, w hich is dependent on th e tem ­
p eratu re and hum idity.
Check points on th e w’a te r calibration arc obtained w ith ben­
zene. T o obtain volumes, th e w eights of benzene are divided by
0.8788. T he density of benzene in vacuum a t 20° C. determ ined
by use of this calibration curve and E q u atio n 1 should be 0.87893,
which checks th e values of 0.87890 of T im m erm ans and M artin
{28) and 0.87895 of W ojeiechowski (32).
DETERMINATION O F PRECISION DENSITIES
A.
P u r e C o m p o u n d s a n d M i x t u r e s o p M e d iu m a n d L o w
V o la tility .
T h e pycnom eter is cleaned w ith benzene and
T O T A L LEN G TH U N F ILLE D C A P IL L A R Y .
Figure 2. Evaporation Rates
T he pycnom eter is filled in an uprig h t position b y placing the
hooked tip in a vial containing the liquid un til th e liquid level
reaches scale m ark 4 on th e bulb side. L iquid level m u st be on
th e scale when th e pycnom eter contents are a t 2 0 ° C. (6 8 ° F.).
In h o t w eather (100° F .), th e contraction to 6 8 ° F . on a 3-ml.
pycnom eter of sm all bore is approxim ately 10 cm. If room
tem perature is below 6 8 ° F . or th e sam ple is m uch below th a t
tem perature, due allowance should be m ade in th e filling level.
T he level of liquid in the vial m ust be very near th e top. Pentan e will fill a 3-ml. pycnom eter in less th an a m inute. More
viscous sam ples wall fill m ore slowly. T h is pycnom eter is not
recom mended for sam ples more viscous th a n gas oil of 50 seconds
Saybolt viscosity a t 100° F . (about 7.4 centistokes).
T h e pycnom eter is rem oved from the sam ple bottle, and th e
hook which had been im m ersed in th e sam ple is cleaned w ith a
cloth m oistened w ith benzene and wiped dry. T h e w eight is then
obtained w ithin =*=0 .0 0 0 1 gram .
T he pycnom eter is placed for 10 to 15 m inutes in a vertical
position in a glass ja r th erm o sta t held a t 20.00° =*= 0.05° C. and
allowed to reach th a t tem perature. A variatio n of 0.05° C. is
equivalent to approxim ately 0.00004 gram ml. per ml. for gaso­
January 15, 1944
ANALYTICAL
line hydrocarbon. A m ore général expression for the tem pera­
tu re coefficient of density of hydrocarbons m ay be obtained from
the correlation of Lipkin and K u rtz {18).
T h e m eniscus levels are read on b o th arm s of th e in stru m en t
w ithin one half of the sm allest scale u n it w ith o u t rem oving the
pycnom eter from the bath . To avoid undue volatilization losses,
th e elapsed tim e betw een weighing and reading th e meniscus
levels should n o t be over 15 m inutes. T h e volum e in milliliters
corresponding to the sum of the readings is read from th e cali­
b ration curve:
„
w eight of sam ple
D ensity in air = -r —2— 3
¡—1—
observed volum e
(2 )
E q uation 1 gives the corrections to be added to th e determ ined
d ensity to tak e care of air buoyancy and to o b tain density in
vacuum .
B. H i g h l y V o l a t i l e M i x t u r e s . T his procedure is to be
used on highly volatile m ixtures because it prevents or reduces
change in the compositions of th e sam ple b y th e selective vola­
tilization of th e lower boiling com ponents while the pycnom eter
is being filled an d while it a tta in s balance tem perature.
T h e sam ple is cooled to ice tem perature, charged to th e pyc­
nom eter, and th e pycnom eter and contents are allowed to come
to tem perature, 20° C. T he volum e is read, th e in stru m en t re­
moved from th e b ath , th e outside cleaned and dried, and
th e
pycnom eter plus sam ple weighed. T he sam ple is th en flushed
from th e pycnom eter w ith benzene, and the pycnom eter is dried,
an d weighed em pty.
E ith er procedure m ay be used on pure com pounds of high
volatility, since selective loss of one com ponent is n o t involved.
Procedure A is more convenient for sam ples w hich are n o t ex­
trem ely volatile.
In b o th procedures there is a tim e in terv al betw een th e meas­
urem ent of w eight and the m easurem ent of volum e. If volatili­
zation occurs betw een these m easurem ents in th e case of proce­
dure A a high density will result, since w eight is m easured first.
In th e case of procedure B, since volum e is m easured first, low
densities will result from such vaporization. Such losses are
negligible except in the m easurem ent of very volatile m aterials
such as isopentane in pycnom eters of sm all volum e such as 0.3
ml. If vaporization losses are suspected, th e density should be
determ ined by both procedures and averaged.
Table I.
Deviations in Density O bserved Using Pycnometers of
Different Sizes“
O p erato r
M ate rial
E x am in ed
d 2 0 /4
1
1
2
2
Iso p e n tan e
0 .6 1 9 7
Av.
1
1
2
2
E th y l e th e r
0 .7 1 4 0
Av.
1
1
2
2
1
1
2
2
1
1
2
3 00-400° F . 0 .7 9 7 5
n a p h th a
0.32-M I.
Pycnom ­
e te r
1.2-M I.
Pycnom ­
e te r
3-M I.
Pycnom ­
e te r
5-M I.
Pycnom ­
e te r
+ 0 .0 0 1 9
+ 0 .0 0 11
+ 0 .0 0 8 4
+ 0 .0 0 4 0
+ 0 .0 0 1 4
+ 0 .0 0 10
+ 0 .0 0 2 5
+ 0 .0 0 11
+ 0 .0 0 0 6
± 0 .0 0 0 0
+ 0 .0 0 0 5
+ 0.0 0 01
-
+ 0 .0 0 3 8
+ 0 .0 0 1 5
+ 0 .0 0 0 3
- 0.0 0 01
+ 0.0 0 2 4
+ 0.0022
+ 0 .0 0 4 4
+ 0 .0 0 0 6
+
+
+
+
0 .0 0 0 2
0 .0001
- 0 .0 0 0 2
± 0 .0 0 0 0
+ 0.0004
- 0 .0 0 01
- 0 .0 0 0 5
- 0 .0 0 01
± 0 .0 0 0 0
+ 0 .0 0 0 0
+ 0 .0 0 2 4
+ 0 .0 0 0 5
- 0 .0 0 0 2
- 0.0 0 01
—0 .0 0 0 3
- 0 .0 0 0 5
-0 .0 0 0 3
-0 .0 0 0 7
-0 .0 0 0 6
- 0 .0 0 01
- 0 .0 0 0 2
- 0 .0 0 0 3
± 0 .0 0 0 0
+ 0.0 0 01
- 0 .0 0 01
- 0 .0 0 0 3
- 0 .0 0 0 2
+CLÓÓ04
- 0 .0 0 0 4
0 .0 0 0 6
0 .0 0 1 3
0.0 0 01
0 .0 0 01
0 .0 0 0 2
0 .0 0 01
A v.
- 0 .0 0 0 4
- 0 .0 0 0 5
- 0 .0 0 01
- 0 .0 0 0 2
(50% iso­ 0 .7 1 2 8
p en tan e)
50% 3 0 0 400 F .
n a p h th a
•
Av.
- 0 .0 0 1 5
- 0.0022
- 0 .0 0 6 9
+ 0 .0 0 1 9
+
+
+
+
+
+
+
0 .0 0 0 2
0 .0 0 0 2
0 .0 0 0 2
0.0 0 0 3
± 0 .0 0 0 0
- 0 .0 0 01
+ 0 .0 0 0 4
± 0 .0 0 0 0
-0 .0 0 3 1
+ 0 .0 0 0 7
+ 0 .0 0 0 2
+ 0 .0 0 01
G as oil 50
sec. S ayb o lt
at
100° F .
Av.
0 .8 3 0 4
0 .0 0 0 5
0.0 0 0 3
0 .0 0 0 9
0 .0 0 12
-0 .0 0 3 2
-0 .0 0 2 8
-0 .0 0 1 4
+ 0 .0 0 01
- 0 .0 0 0 5
-0 .0 0 0 8
-0 .0 0 0 5
- 0.0 0 10
± 0 .0 0 0 0
- 0 .0 0 0 2
- 0.0001
-0 .0 0 0 3
+ 0.0 0 01
+ 0 .0 0 01
± 0 .0 0 0 0
-0 .0 0 0 3
-0 .0 0 1 9
-0 .0 0 0 7
- 0 .0 0 0 2
- 0.0 0 01
° O b tained using pro ced u re A a n d e arly design of p y c n o m e ter w ith b o th
arm s s tra ig h t. Sim ilar d a ta o b ta in e d w ith b e n t-a rm p y cn o m eter check these
d a ta . E a rlie r te sts a re given, Bince th e y a re m ore com plete th a n re c e n t te sts.
EDITION
57
PRECISION
T ests of th e precision of th e pycnom eter using procedure A are
shown in T able I. (R esults obtained w ith procedure B are of
comparable accuracy.) Two operators determ ined the densities
of isopentane, eth y l ether, a n ap h th a, a m ixture of isopentane
and a n ap h th a, and a gas oil. E ach operator m ade duplicate
determ inations. T h e d a ta showed th a t, w ith th e 3- or 5-ml.
pycnom eter, precision does n o t depend on th e operator, and th a t
the 5-ml. size is precise to ±0.0001, th e 3-ml. to ±0.0002, th e
1-ml. to ±0.001, and th e 0.3-ml. to ± 0.002 gram p er ml. over
the entire viscosity and volatility range betw een p en tan e and
light gas oil.
T he sources of th e larger errors in determ ining densities w ith
the sm aller pycnom eters are shown in T able I I . T h e effect of
th e various possible errors on th e final determ ination of th e den­
sity is given. W ith th e 1-ml. an d th e 0.3-ml. pycnom eters,
errors in the scale readings and weighings are im p o rtan t. E rrors
due to tem perature control an d air buoyancy are th e sam e w ith
all size pycnom eters, as indicated in th e table. T h e 3- and 5-ml.
pycnom eters are recom m ended for routine use.
Table II.
Errors in Determined Densities of Ethyl Ether Due to
Errors of Observation
Volume of
Pycnom eterB “
M l.
1° C. E rro r in
T e m p e ra tu re
O ./m l.
0.5 M g. E r ro r
in W eighing
a ./m l.
0.5 m m . E r ro r in
Scale R eading
O ./m l.
0 .2 5
1
3
5
0.0011
0.0011
0.0011
0 .0011
0 .0 0 2 0
0 .0 0 0 5
0 .0 0 0 2
0.0001
0.0011
0.0003
0.0001
0.0001
“ C ap illary d ia m e te r 1.0 m m .
ACCURACY
T h e d a ta in colum n 3, T able I, on the density of th e m aterials
were obtained by a skilled o perator w ith th e 5-ml. pycnom eter,
and are believed to be accurate w ithin ±0.0001 gram p er ml.
If this is true, th e d a ta in T able I con stitu te a te s t of accuracy as
well as of precision. As a fu rth er te s t of absolute accuracy of
the au th o rs’ m ethod, th e density of a carefully purified sam ple of
benzene was determ ined on a pycnom eter calibrated w ith w ater.
T he density a t 20° C. was found to be 0.87895. T his result
checks th e literatu re value of T im m erm ans {28) of 0.87895 cor­
rected from 15° C. an d th e value of 0.87891 determ ined by W ojciechowski {82) corrected from 25°, using th e tem perature co­
efficient of density values determ ined by Tim m erm ans.
SPEED
E ach determ ination requires 10 to 15 m inutes of th e operator’s
time, and an elapsed tim e of 20 to 30 m inutes. W ith a dam ped
balance, 10 m inutes of o perator’s tim e per sam ple is sufficient.
ACKNOW LEDGMENT
T he au th o rs wish to acknowledge the cooperation of Ace Glass,
Inc., Vineland, N . J., in connection w ith th e developm ent of this
pycnom eter. T h ey also wish to acknowledge th e assistance of
I. W . M ills w ith prelim inary work on this pycnom eter and C. C.
M artin, A. E . W ikingsson, G. H . H ansel, and C. W. D ittric h in
obtaining statistical d ata.
LITERATURE CITED
B earce, H . W ., Proc. A m . Soc. Testing M aterials, 19, 412 (1919).
Bousfield, W . R., J . Chem. Soc., 93, 679 (1908).
B u svold, B „ Chem.-Ztg. 49, 276 (1925).
Ciochina, J., Z . anal. Chem., 107, 108 (1936).
D avis, P . B ., and P ratt, L. W ., J . A m . Chem. Soc., 37, 1199,
(1915).
(6) F reund, Ida, Z . ph ysik. Chem., 66. 569 (1909).
(7) Frivold, O. E „ P h ys. Z „ 39, 529 (1920).
(8) Furter, M „ H d v. Chim . Acta, 21, 1666 (1938).
(1)
(2)
(3)
(4)
(5)
58
INDUSTRIAL
AND
ENGINEERING
(9)
Geiger, H ., and Scheel, K ., “ H andbuch der P h y sik ” , V ol. 2, p.
160, Berlin, Julius Springer, 1926.
(10) G osko-M ulhoim R uhr, Z . Untersuch. N ähr. Genussm., 24, 245
(1 9 1 2 ).
(11)
(12)
(13)
(14)
(15)
H are, J . In st. Brewing, 40, 92 (1934).
H artley, H ., and B arrett, W . H ., J . Chcm. Soc., 99, 1072 (1911).
H eard, L ., J . Chem. Education, 7, 1910 (1930).
H ennion, G . F ., I n d . E n g . C h e m ., A n a l . E d ., 9, 479 (1937).
H ouben, J., “ M eth od en der organischen C hem ie” , V ol. 1, p.
891, Leipzig, G . T h iem e, 1925.
(16) Irving, H ., Science Progress, 31, 654 (1 9 3 7 ).
(17) K alahno, A ., B er. deut. p h ysik. Ges., 16, 8 1 -9 2 (1914).
(18) Lipkin, M . R ., and K urtz, S. S., Jr., I n d . E n o . C hem ., A n a l .
E d ., 13, 291 (1 9 4 1 ).
(19) O stwald, W ., J . prakt. Chem.., 16, 396 (1877).
(20) O stwald, W ., Luther, R ., and D rucker, C., “H an d - und H ilfs­
buch zu . A usführung physiko-chem ischer M essungen” , pp.
2 3 4 -7 , Leipzig, A kadem ische V erlagsgesellschaft, 1931.
(21) Parker, H . C ., and Parker, E . W ., J . P h ys. Chem., 29, 130
(1925).
CHEMISTRY
Vol. 16, No. 1
(22) Perkin, W . H ., J . Chem. Soc., 4 5 , 1, 4 4 3 -5 (1884).
(23) R eidel, R ., Z . ph ysik. Chem., 56, 245 (1906).
(24) R eilly, J., and R ae, W . N ., “ P hysical C hem ical M eth od s” ,
N ew Y ork, D. V an N ostrand C o., 1939.
(25) R obertson, G . R ., I n d . E n o . Chem ., A n a l . E d . , 11 , 464 (1939).
(26) S h edlovsky, T ., and B row n, A . S., J . A m . Chem. Soc., 56, 1066
(1934).
(27) Sprengel, H ., Pogg. A n n ., 150, 459 (1873).
(28) Tim m erm ans, J., and M artin, F ., J . chim. p h ys., 23, 747 (1926).
(29) Ward, A . L ., K urtz, S. S., Jr., and Fulweiler, W . H ., in “S cien ce
of P etroleu m ” , ed. b y D u n stan , N ash, B rooks, and T iza r d ,
V ol. II, p . 1137, N ew York, Oxford U n iversity P ress, 1938.
(30) W ashburn, E . W ., and Sm ith, E . R ., B u r. Standards J . R e­
search, 12, 305 (1934).
(31) W ostberg, J., Tek. Fören. Finland Förh., 58, 314 (1938).
(32) W ojciechow ski, M ., J . Research N a tl. B u r. S tandards, 19, 347
(1937).
(33) Y uster, S. T ., and R eyerson , L . H ., I n d . E n g . C hem ., A n a l
E d ., 8, 61 (1936).
Apparatus for Measuring the G as Permeability of
Film Materials of Low Permeability
A.
CO RN W ELL SH U M A N
Central Laboratories, General Foods Corp., H o b o ke n , N . J.
Apparatus is described (or measuring the gas permeability
of film materials having permeabilities as low as 0.001 cc.
(at standard temperature and pressure) per 1 0 0 square
inches per 24 hours. The low range for previously reported
methods (water vapor permeability and balloon fabric
perm eability) is about 1 0 0 cc. per 1 0 0 square inches per
2 4 hours. The apparatus combines sim plicity of design,
sim plicity of manipulation, and high sensitivity in a unit
which can be fabricated in most machine shops. The
measurements are made under conditions of one atmosphere
pressure differential.
relatively im perm eable to oxygen. P erhaps a process of solution
of gas in th e film on one side, diffusion through th e film, and
evaporation from th e film on th e oth er side takes place in some
cases.
E packaging of some foods, drugs, chemicals, etc., requires
T Hflexible
films and paper coatings of very low perm eability to
fixed gases, particularly oxygen. T h e need for film having a
perm eability of not m ore th a n 0.25 cc. per 645 sq. cm. (100 square
inches) per 24 hours has been pointed o u t by E ld er (S). T his
figure is entirely o u t of th e range usually m easured for balloon
fabrics (5), w here perm eabilities range from 100 cc. p er 100
square inches per hour an d up. T he custom ary m easurem ents
o f w ater vapor perm eability (1, 2, 4) of packaging films are also
in a relatively high range— for example, a “good” w ater vapor
barrier will have a perm eability figure of 0.1 gram or 125 cc. p er
100 square inches per 24 hours. T he ap p a ra tu s described in th is
p aper is designed for m easurem ents in th e range 0.001 to 1000 cc.
per 100 square inches per 24 hours, and prim arily for th e m eas­
urem ent of the fixed gases, although w ith some modification it
m ight be also used for w ater v ap o r perm eabilities. T he purpose
o f this publication is to m ake available to others th e m eans which
th e au th o r has developed for testing packaging films of very low
perm eability.
T he m echanism of gas transm ission through solids has n o t
been explained thoroughly. Oswin discusses th is briefly (4).
If the gas passed through sm all pores or openings in th e solid, th e
situation would be a relatively simple one; th e perm eability of a
film to various gases could be predicted from th e know n laws of
gas flow through orifices. However, such is n o t th e case. I t is
known th a t polyvinyl alcohol film is perm eable to w ater vapor b u t
n o t to oxygen, and Pliofilm is perm eable to carbon dioxide b u t
Figure 1.
Diagram of Apparatus
January 15, 1944
ANALYTICAL
T he m echanism of gas transm ission through film determ ines to
some extent the conditions u n d er which a film should be tested.
Tw o factors are involved: th e difference in to ta l pressure on th e
tw o sides of th e film and the difference in p a rtial pressure of th e
gas being tested on th e tw o sides of th e film. I f perm eability is
due to gas flow through orifices in th e film, th e difference in to ta l
pressure is of prim ary im portance and th e difference in p artial
pressure is of secondary im portance. If th e perm eability is due
to solubility of th e gas in th e film, th e difference in to ta l pressure
is of no im portance, b u t the difference in p a rtia l pressures is im­
p o rtan t. In th e ap p aratu s described in th is article, th e te s t is
m ade under th e conditions of atm ospheric pressure of th e gas
being tested on one side of the film and zero pressure of all gases
on the o th er side.
T he apparatus, shown in Figures 1 and 2, combines simplicity
of design, sim plicity of m anipulation, an d high sensitivity in a
u n it which can be fabricated in m ost m achine shops. Essentially
th e m ethod involves m easuring changes in pressure inside a sm all
evacuated space due to gas passing into th a t space through the
film sam ple being tested.
EDITION
59
stopper w ith a glass tu b e is inserted into each of these holes.
T he space betw een E and F is th en flushed o u t w ith gas to be
m easured by passing th e gas in one opening an d o u t th e other.
I t has been shown th a t th e m oisture conten t of fixed gases m ark­
edly affects th eir ra te of transm ission thro u g h some film m ate­
rials (4). T his factor m ay be controlled by conditioning th e gas
to be tested before it is passed thro u g h th e apparatus.
CONSTRUCTION OF APPARATUS
T he m anom eter, M , is held in place by th e n u t, N , to w hich it
is sealed w ith Cenco Plicene cem ent or other sealing compounds
of good adhesive properties an d low vapor pressure. T h e hole
through the n u t is bored nearly th e same diam eter as th e glass
tube, thereby minim izing the am o u n t of cem ent necessary for this
seal. T he n u t is screwed up against th e th in ru b b er w asher, W,
and the entire jo in t is coated w ith shellac to ensure a vacuum tig h t seal.
M anom eter M serves th e d ual purpose of recording th e pres­
sure change and of evacuating th e space inside th e apparatus.
F o r evacuation, the entire ap p aratu s is tipped so th a t th e m er­
cury runs over into th e reservoir, R ; th en th e a p p aratu s is evacu­
a ted through tu b e B w ith stopcock S in th e open position. A fter
evacuation, S is closed and th e a p p aratu s is tip p ed again so th a t
th e m ercury runs from R into M . In case air leaks through S
during a test, th e m ercury in th a t arm of th e m anom eter will be
depressed, th u s serving as an indicator of leakage a t th is point
w ithout spoiling th e test. T h e center arm of M is m ade of cap­
illary tubing of ab o u t 1.5-mm. internal diam eter for th e purpose
of reducing the volum e of gas space inside th e a p p aratu s and
thereby increasing its sensitivity. T he rest of th e m anom eter
m ay be m ade of tubing of any convenient size. T h e scale used
in m easuring the height of th e m ercury is n o t shown. A piece of
m illim eter cross-section p aper held behind th e m anom eter serves
very well for th e purpose.
T he upper p a rt of th e hole in th e center of th e disk, K , is
covered w ith a sm all m etal disk, D, which has four sm all holes
ab out 1 m m . in diam eter for passing gas in to th e m anom eter
system . T he top surface of th e sm all disk is flush w ith th e top
surface of th e large disk, If, presenting a continuous sm ooth sur­
face for m ounting th e film, F.
T he drying agent absorbs all m oisture w hich passes through F,
th u s elim inating this gas as a source of error in gas transm ission
m easurem ent. D ehydrite (anhydrous m agnesium perchlorate)
diluted w ith a sm all q u an tity of Indicating D rierite (anhydrous
calcium sulfate) serves adm irably. A screen fraction passing 14mesh and retained on 40-mesh has been found to be a convenient
size. In general, an y drying agent w hich will reduce th e w ater
vapor pressure inside th e app aratu s to less th a n 0.5 mm . of m er­
cury will serve. A sm all plug of co tto n over th e to p of th e m an ­
om eter tube will prevent particles of th e drying agent from
falling into th e m anom eter.
MOUNTING SAMPLES
T he film sam ple, F, is m ounted on th e ap p aratu s as showrn in
th e diagram . A piece of filter paper, P , is placed betw een the
film and K to provide a porous medium for gas leaking through
the film to trav el to th e openings in D an d pass thro u g h th e drying
agent into M . T he p a rt of F overlapping P and in d irect contact
w ith K is sealed to th e disk by m eans of a th in film of heavy stop­
cock grease having a low vap o r pressure. T his area of th e film
is held in place by th e rubber gasket, C.
D isk E has tw o sm all holes diam etrically opposite one another
and near the inside circumference of th e m etal ring, 0 . W hen
m easuring air transm ission, these openings are left open so th a t
th e space betw een E and F is filled w ith air. W hen it is desired
to measure the gas transm ission for any o th er gas, a sm all rubber
Courtesy "M o d e rn P a cka g in g "
Figure 2.
Apparatus
In m ounting m ulti-ply film having one or more plies of a porous
m aterial— for example, a film lam inated betw een twro pieces of
paper—it is p ractically impossible to o btain a v acuum -tight seal
using stopcock grease betw een th e p ap er surface of th e sample
and K . In m ounting such films, use h a s been m ade of a plas­
ticized ta r as shown in Figure 3. An edge coating extending
a b o u t 0.6 cm. (0.25 inch) from th e circum ference is applied to
both sides of th e disk of film to be tested b y dipping in th e p lasti­
cized ta r (Figure 3). T h e tar-co ated disk of m aterial is m ounted
directly on K w ith o u t th e use of stopcock grease. W hen cap G
is screwed into place, th e plasticized to r is spread o u t by th e com­
pressive forces and m akes a vacuum seal betw een F an d K .
Carbowax 1500 is a suitable plasticizer for th e ta r when used a t
th e level of a b o u t 5 per cent.
CA LCU LA TIO N S
Gas. transm ission m easurem ents are conveniently recorded as
cubic centim eters of gas a t stan d ard conditions of tem perature
and pressure tran sm itted per 100 square inches of surface p e r 24
hours. Such a figure is given b y th e following expression:
_P_ v v
273
100
24
760 *
X T X A A hours
w’here P = th e absolute pressure inside th e ap p a ra tu s as m eas­
ured on th e m anom eter in m illim eters
V = to ta l volum e of th e inside of th e ap p a ra tu s in cubic
centim eters
T = tem p eratu re in degrees absolute
A — surface area of te s t sam ple (area of filter p ap er P in
square inches)
hours = tim e of te st
The value of V m ay be calculated w ith sufficient accuracy
(w ithin 5 per cent) from th e known dim ensions of th e apparatus,
m aking a correction for th e volume of th e drying agent used.
60
INDUSTRIAL
AND
ENGINEERING
I t is equal to the sum of the volume of th e holes in D , th e volum e
of th e circular opening containing th e drying ag en t m inus th e
volume of the drying agent, an d th e volum e of th e capillary bore
of the center stem of th e m anom eter down to th e m ercury level
in this stem . T h e volum e of th e pore space in P am ounts to
ab out 0.05 cc. and the volum e change due to a m ercury level
change of 20 mm . in th e center stem of th e m anom eter am ounts
to ab o u t 0.04 cc. These errors in volum e are neglected.
CHEMISTRY
Vol. 16, No. 1
urem ents w ith th is ap p aratu s are accurate to w ithin 15 per cent.
T he tem p eratu re, tim e, and m anom eter reading m ay be d eter­
m ined w ithin a few p er cent. T h e principal factors affecting the
accuracy are th e volum e of th e a p p aratu s and th e area of th e test
sam ple. I t is estim ated th a t these are easily determ ined in the
m anner described w ithin 10 p er cent accuracy. T he precision or
reproducibility of m easurem ents on duplicate sam ples of the same
m aterial is generally a b o u t 10 per cent.
EXAM PLES
T he following examples will serve to typify th e d a ta w hich may
be obtained w ith this a p p a ra tu s:
F or the ap p aratu s diagram m ed in Figure 1, th e value of V is
a b o u t 2 cc., and th e value of A is 23 sq. cm. (5 square inches).
A t a tem perature of 30° C., T = 303° absolute. S u b stitu tin g
these values in th e above equation, we get th e following working
equation:
1. Air transm ission for an im perm eable sam ple (a sam ple of
lam inated glassine a t 60 to 80 per cent relative hum idity).
In a period of 36 hours, no change was observed in th e m a n ­
ometer. Assum ing th a t a m inim um change of 0.5 mm . can be
detected on th e m anom eter, th e ra te of change of th e m anom eter
was n o t greater th a n 0.5/36 or 0.0139 mm . per hour. T he gas
transm ission was, therefore, n o t g reater th a n 0.0139 X 1.14, or
0.016 cc. p er 100 square inches per 24 hours.
2. Air transm ission for an im perm eable sam ple containing a
volatile plasticizer (a polyvinyl alcohol coating a t low hum idity).
T he m anom eter changes during th e indicated tim e intervals
were as follows:
T im e In te rv a l
H ours
M an o m e te r C hange
Mm.
R a te of C hange
M m ./h o u r
0-1
1-2
2
1 .5
2
2.0
3 .5
0 .5
0 .5 8
0 .0 1 7
2 -4
4-1 0
10-40
P /h o u rs X 1.14 = cc. of gas a t stan d ard conditions of tem p era­
tu re and pressure tra n s m itte d /100 square inches/24 hours
•
T o ta l
1.5
1.0
9 .5
SENSITIVITY, ACCURACY, A N D PRECISION
T he actual tim e required for a te s t w ith th is a p p aratu s varies
w ith the perm eability of th e film being tested and th e lim its
w ithin which it is desired to determ ine th e gas transm ission rate.
F o r example, it m ight bo desired to know w hether a film sam ple
has a transm ission rate greater or less th a n 0.25 cc. per 100 square
inches per 24 hours. Using th e above equation and constants
for the ap p aratu s herein described, th is transm ission ra te will
produce a pressure change in th e m anom eter of 0.5 m m . in ab o u t
2.5 hours: 0.5-mm. change is easily read on th e m anom eter;
therefore, the failure of the m anom eter to change 0.5 m m . in 2.5
hours’ tim e is sufficient evidence th a t th e gas transm ission ra te is
less th a n 0.25 cc. per 100 square inches per 24 hours. Should the
tim e of the te st be extended five tim es as long, to 12.5 hours, th e
failure of the m anom eter to cliange 0.5 m m . is sufficient evidence
th a t th e gas transm ission ra te is less th a n 0.05 (0.25/5) cc. per
100 square inches per 24 hours. F u rth er extending th e tim e of
th e te s t th u s fu rth er reduces th e figure Tor th e m axim um tra n s­
mission ra te of th e sam ple.
W ith films of liigh perm eability, th e m ercury in th e m anom eter
will drop several m illim eters in a n h o u r’s tim e. W ith such films,
therefore, the te s t need n o t be extended m ore th a n a few hours.
Occasionally film sam ples contain m aterial, such as plasticizers,
which have appreciable vapor pressures. Such m aterials will
vaporize into the ap p aratu s an d depress th e m ercury in th e m an­
om eter. T his m ay bo m istaken for gas transm ission through the
film, b u t m ay be distinguished from gas transm ission b y th e fact
th a t th e pressure will reach a co n stan t value, th e ra te of change
of pressure decreasing as this value is approached. I t is th e re ­
fore necessary to m ake readings of th e m anom eter a t intervals
during the te s t and calculate th e ra te of m anom eter change for
each interval. I f th e ra te of change is constant, it is due to gas
transm ission through th e film, b u t if it is a decreasing ra te of
change, th e change is due a t least p a rtly to vaporization of some
volatile m aterial from th e film.
T he required accuracy is n o t very g reat for m easurem ents of
perm eability of film packaging m aterials where values m ight vary
a millionfold am ong various m aterials and a thousandfold am ong
various sam ples of the sam e m aterial. I t is estim ated th a t m eas­
I t is evident from th e above d a ta th a t a m aterial having a vapor
pressure of a b o u t 9.5 m m . is ev aporating from th e sam ple and
th a t equilibrium is nearly established a fter 10 hours. T he rate
of change of th e m anom eter during th e last 30 hours of th e te st
was 0.017 mm. per hour. Therefore, th e gas transm ission ra te
for th e sam ple was n o t m ore th a n 0.017 X 1.14 or 0.019 cc. per
100 square inches per 24 hours.
3. Air transm ission for a perm eable sam ple (a th in coating of
polyvinyl alcohol).
T he m anom eter changes during th e indicated tim e intervals
were as follows:
T im e In te rv a l
H ours
0-1
1-2
10-18
1 .5
1 .5
3 .0
9 .5
1 1 .5
18
2 7 .0
2 -4
1-10
T o ta l
M an o m e te r C hange
Mm.
R a te of C hange
M m ./h o u r
1 .5
1 .5
1 .5
1.58
1 .4 5
A v. 1 .5
T h e c o n sta n c y o f th e ra te o f ch a n g e figu res in d ic a te s t h a t th is
sa m p le is p erm ea b le t o air. T h e r a te o f g a s tra n sm issio n is
1.5 X 1.14, or 1.71 cc. p er 100 sq u are in c h e s p er 2 4 h ou rs.
LITERATURE CITED
(1)
(2)
(3)
(4)
(5)
A non., M odern Packaging, 1 6 ,7 8 -8 2 ,1 0 0 (1942).
Carson, F . T ., N a tl. B ur. Standards. M isc. P u bl. M 127 (1937).
Elder, L . W ., M odern Packaging, 16, 6 9-71 (1943).
Oswin, C . R ., J . Soc. Chem. In d ., 62, 4 5 -8 (1943).
Sager, T . P ., J . Research N a ll. B u r. Standards, 25, 309 -1 3 (1940).
B ecause of critical shortages, th e A m e r i c a n C h e m i c a l
has been forced to cu t its use of paper to an abso­
lu te m inimum . I t will no longer be possible to p rin t the
custom ary num ber of extra copies to supply dem ands
for volumes an d sets in subsequent years. Therefore, it
is suggested th a t subscribers who do n o t bind their
journals save current issues for later sale.
S o c ie ty
Design of Large-Size Laboratory Extraction G lass Apparatus
R A Y M O N D JO N N A R D
Warner Institute for Therapeutic Research, 113 West Eighteenth St., N ew York, N . Y .
A
S IM P L E large-capacity extraction a p p aratu s has been
described by Smallwood (2).
T he frequent need for extraction ap p ara tu s of larger capacity
th a n th e conventional Soxhlet extractors for research as well as
for sm all pilot production justifies a description of th e im proved
all-glass ap p aratu s which has been in satisfactory operation in
th is laboratory for a num ber of m onths.
tem perature was alw ays lower th a n 39.3° C., th e difference being
probably accounted for by th e presence of noncondensable gas
in th e vapor phase, w hich is difficult to avoid in ap p a ra tu s of
relatively large size. D espite th e length of th e v ap o r line, th e
a p p aratu s appears well suited for use w ith mixed or diluted
solvents.
Some determ inations of th e approxim ate h ea t balance have
been m ade an d one typical resu lt is reported in T able II.
T he a p p a r a t u s is
shown in Figure 1,
after rem oval of the
insulation. I t is de­
signed for operation a t
either ordinary or re­
duced pressures and is
equipped to p erm it the
collection of all engi­
neering d a ta required
for pilot work. I t is
m a d e of s t a n d a r d
Pyrex p arts w ith Ace
spherical joints, size
35/25. T he boiler is
a three-neck 12-liter
flask i n s t a l l e d i n a
Glas-Col h eate r p ro­
vided w ith proper in­
p u t control (rheostat
. and a m m e t e r ) a n d
pyrom eter. T he ex­
tra c to r has a to tal
capacity of 8 liters.
In order to o btain a
sufficient speed of ex­
tractio n the h o t sol­
v e n t vapor is fed a t
the top of th e 600-mm.
Allihn condenser, in­
stead of th e bottom ,
as in th e conventional
Soxhlet. E xperim ent
has shown th a t flood­
ing occurs w ith the
conventional design a t
a distillation ra te of 3
liters per hour a t ordi­
n ary pressure and 1.5
Figure 1
liters per hour a t 74.1
mm . h g . ( 2 7 - i n c h )
vacuum in the pres­
ence of alcohol vapor, w ith a condensing area of 640 sq. cm.
cooled w ith brine, whereas th e present design perm its an extrac­
tion of 5.5 liters per hour a t ordinary pressure and over 4 liters
per hour in vacuum w ith a condensing area of only 350 sq. cm.
W hen used under vacuum , excessive evaporation of the solvent
in the extractor is prevented by a oack-condenscr inserted in the
vacuum line (350-sq. cm. Allihn Pyrex vertical condenser) a t the
level of th e side connection of the 8-liter bell ja r form ing the
extractor.
Inasm uch as the sudden em ptying of the large extractor into
th e h o t boiler could be dangerous, particularly w ith v ery volatile
solvents, th e retu rn U -tube is provided w ith a glass stopcock
which perm its either controlled periodical or continuous circula­
tion of the solvent. A sam pling o u tlet is provided a t th e bo tto m
of the extractor.
T h e condensers can be cooled w ith eith er w ater or brine, and
th e piping circuit includes th e necessary valves and m easuring
devices for determ ination of the h e at balance.
Table I.
H ead
B oiler
T em ­
T em p e ra tu re p e ra tu re
° C.
° C.
32
36
29
35
Performance of Apparatus
E x tra c t
T em ­
p e ra tu re
° C.
A bsolute
P ressu re
M m . Hij
O ver­
head
M ole %
B o tto m
Mole %
27
26
109
7 3 .0
6 2 .8
2 3 .5
2 3 .5
10 0
P la te
N o .°
1 .4
1.1
° O ver-all th e o re tic al p la te n u m b e r. A lthough a c e rta in reflux ta k e s place
from th e to p of th e v a p o r line dow nw ard, no liquid is ever carried b ack to th e
still p o t w hen th e lic a t in p u t is c orrectly a d ju ste d . T h e re fo re th e phases in
equilibrium considered for th e calculation of th e "o v e r-a ll th e o re tic al p la te
n u m b e r” w ere th e s till p o t liquid com position an d th e d is tilla te com posi­
tion, respectively. N o flow m e a su re m e n t is involved in th is c alculation.
H ow ever, th e ra te of distilla tio n can be d e te rm in e d b y m easuring th e tim e
req u ired to fill th e e x tra c to r w hose cap a c ity is know n; th is op eratio n re ­
q uires s h u ttin g off th e stopcock on th e re tu rn U -tube for only a s h o rt tim e.
T he d a ta reported show th e excellent perform ance of the
ap p aratu s. T hey dem onstrate th e possibility of obtaining a rate
of extraction m uch faster th a n h ith erto reported w ith large all­
glass lab oratory ap p aratu s, while a t th e sam e tim e m aintaining
a satisfactory an d economical h e a t balance so th a t th e extraction
can be carried o u t under conditions m ore closely resembling
those of p la n t operation.
Table II.
H eat Balance
H e a t In p u t from D istilla te
H e a t O u tp u t (to B rine)
A. H e a t of condensation
K a te of d istilla tio n , ml. p er
m in., 80
D istilla te , sp. g r., 0.840
E th a n o l, m ole % , 62.8
W a te r, m ole % , 37.3
Sp. h e a t of eth an o l, 204.26
Sp. h e a t of w a ter, 579.3
A *= 23,103.96 calories p e r m in.
K a te of b rin e flow, lite rs p e r m in.,
5.6
B rin e , sp. g r„ 1.20
B rin e , sp. h e a t, 0.7
I n p u t te m p e ra tu re , ° C ., 8.0
T e m p e ra tu re increase, 0 C., 8.1
W eight, gal. of w a ter, lb ., 8.33
H e a t o u tp u t, 81.4 B .t.u . or 20,462.4
calories p e r m in.
B . H e a t of of cooling of distillate
T e m p e ra tu re decrease, ° C.,
2 9 -2 6 ° » 3°
Sp. h e a t of 62.8 m ole % of
e th a n o l a t 23° C .f calories
p e r gram , 72.0
B «■ 145.15 calories p e r m in.
T o ta l h e a t in p u t, 23,249.11 calories
p e r m in.
Balance®, 2787 calories p e r m in. or
11.9%
° “ B alance” rep resen ts sum m ed effect of h e a t losses from th e s till p o t
and th e v a p o r line, th e th e rm o m e te rs, v arious experim ental u n c erta in tie s,
an d som e sim plifying assu m p tio n s m ad e in calculation of th e q u a n titie s
involved.
LITERATURE CITED
(1) Beebe, A . H ., Coulter, K . E ., L in d s a y , R . A ., a n d Baker, E . M .#
I n d . E n g . C h e m ., 3 4 , 1501 (1 9 4 2 ).
(2 ) S m a llw o o d , E ., I n d . E n g . C h e m ., A n a l E d ., 14, 9 03 (1 9 4 2 ).
T able I indicates th e perform ance of th e a p p aratu s in two
different experim ents.
T he volatility involved in calculating th e over-all theoretical
p late num ber has been calculated from th e recen t d a ta of Beebe
an d co-workers (f): F o r Xf = 24.10 an d x<> = 56.75 th e overhead
r e s e n t e d before th e D ivision of In d u s tria l an d E n g in ee rin g C h em istry
a t th e 105th M eetin g of th e A m e r i c a n C h e m i c a l S o c i e t y , D e tro it, M ich.
P
61
A Continuous Liquid-Liquid Extractor
IR W IN A . P E A R L
The Institute of Paper Chemistry, A p p le to n , W is.
T L Y Kieselbach (I) introduced a new principle into
R EthC Ee Nfield
of liquid-liquid extractors for use w ith m ixtures
stan d ard -tap er jo in t was used for th e connection to th e boiling
flask, C. T ubing of a t least 10-mm. outside diam eter should be
used for th e glass p a rt of the trap , D; otherwise, th e passage of
very slowly breaking emulsions is impeded. T he tra p its e lf was
m ade of a sh o rt length of fairly thick-walled neoprene tubing.
T he inner tube, E , was m ade by sealing a piece of 8-mm. outside
diam eter glass tubing to a stock gas inlet adapter. T he length
of th e tu b e depended upon th e size of th e flask containing th e
m aterial to be ex tracted. If a num ber of gas inlet ad ap ters are
no t available, a 10/30 stan d a rd jo in t m ay be sealed to th e stock
ad ap ter and a num ber of 8-mm. tubes w ith stan d ard 10/30 joints
of lengths to fit various flasks or bottles m ay be used.
which tend to emulsify. H is extractor, using a ir agitation and a
settling cham ber for the separation of an y emulsion formed, m ade
possible the rapid extraction of solutions form erly tak in g as long
as several weeks. However, K ieselbach's ex tracto r has a num ber
of disadvantages, chief am ong w hich are th e facts th a t a separate
ex tractor is necessary for each desired volum e of solution to be
ex tracted an d th a t very stable emulsions do n o t separate in th e
narrow settling cham ber.
T he app aratu s herein described was an a tte m p t to increase the
u tility of K ieselbach’s extracto r for use w ith strongly emulsifying
sulfite waste liquor reaction m ixtures varying in volume from
250 cc. to several gallons. T he e x tracto r u n it is actu ally an
ad ap te r to be used w ith sta n d ard -tap er glass bottles or flasks.
I t is m ade from readily available stock glassware and requires
a m inim um of glass blowing.
T h e completion of th e extraction is usually determ ined by th e
change in color tak in g place in th e settling cham ber. However,
in th e case of extractions of colorless substances, sam ples of th e
solvent layer in th e settling basin m ay be tak en periodically by
lowering th e solvent solution interface (by w ithdraw ing a little
solution through th e gas inlet tube, E ) an d th e appropriate use
of several screw clam ps on th e ru b b er trap . (R ichard Kieselbach, a fte r reviewing th e paper, suggested th a t th e settling
cham ber m ight be provided w ith a tu b u latu re, close to th e point
which used to be the neck of th e Erlenm eyer flask, w hich would
perm it sam pling w ithout interru p tio n of the operation of th e ex­
tracto r. T his should prove a valuable modification of th e present
ap p a ra tu s when dealing w ith colorless solutions.)
A glass T in th e ru b b er tra p m ay be used if a num ber of such
solutions are to be encountered. W hen th e extraction is com­
plete, th e solvent in th e settling basin m ay be draw n oS in the
m anner described above. These advantages m ake th e rubber
tra p preferable to th e all-glass U -trap. In addition, th e glass
blowing is greatly simplified. An Erlenm eyer flask was used for
th e settling cham ber because m any fairly stable emulsions did
n o t break in th e sm all-diam eter settling basin of th e earlier ap­
p aratu s. T h e shape of th e E rlenm eyer flask is adm irably suited
for th is purpose because th e slopes of its base an d sides (when in
th e position shown in Figure 1) facilitate rapid separation of the
tw o liquid phases. F urtherm ore, th is design is relatively com­
p act. F o r m axim um applicability, th e seals a t b o th ends of the
E rlenm eyer flask should be a t th e sam e level. T he size of the
flask depends upon th e n atu re of the solutions encountered. F o r
fast-breaking emulsions a 250-cc. flask m ay be used, thereby
holding up less solvent in th e settling cham ber. I f a large mixing
cham ber is used w ith a correspondingly increased a ir stream , a
larger settling cham ber should be used. T he specifications of the
entire ap p a ratu s are flexible.
W hen extracting m aterials su b ject to oxidation by a ir a stream
of in e rt gas should be used for agitation. Furtherm ore, w hen ex­
tractin g g as-saturated solutions (such as bicarbonate o r bisulfite
solutions), carbon dioxide or sulfur dioxide, respectively, m ay be
used advantageously as th e ag itatin g gas. In extractions of large
volumes of solutions w ith corresponding increases in tim e, th e
loss of solvent due to ex trainm ent in th e exit gases m ay become
appreciable, and more solvent m ay have to be added to th e boiling
flask. In th is case advantage m ay be tak e n of a two-necked
boiling flask. A long efficient reflux condenser should alw ays be
used. An active carbon tra p is useful for recovering larger
am ounts of solvents.
A large ap p aratu s, using sem iball joints, for use w ith 22-, 50-,
and 72-liter flasks was fabricated according to specifications by
th e Scientific Glass A pparatus Co., Bloomfield, N . J.
Figure 1
T he operation of the complete ex tracto r assem bly draw n to
scale in Figure 1 is identical to K ieselbach’s. T h e ex tracto r u n it
was used successfully for extracting solutions contained in vessels
ranging from 250-cc. flasks to 12-liter bottles.
T he mixing cham ber, A , was m ade by sealing together two
stan d ard -tap er glass joints. T h e upper jo in t in this case was
24/40, although any size com patible w ith o th er equipm ent is sa t­
isfactory. T he lower joint was 29/42; because of the possibility
of constriction, a sm aller stan d ard -tap er jo in t should n o t be used.
B oth stan d ard joints were connected to 28-mm. outside diam eter
tubing. A 500-cc. Erlenm eyer flask served as th e settling cham ­
ber, B , and was connected to th e mixing cham ber b y a short
piece (30 mm.) of 25-mm. outside diam eter tubing. A 24/40
LITERATURE CITED
(1)
62
K ie s e lb a c h , R ., I n d . E n g . C h e m ., A n a l . E d ., 15, 2 23 (1 9 4 3 ).
Determination of Small Am ounts of Acrylonitrile in A ir
G . W . PETERSEN a n d H . H . R A D K E
Industrial H y g ie n e Laboratory, The B. F. G oodrich Com pany, A k ro n , O h io
E recent w idespread use of acrylonitrile in th e m anu­
T Hfacture
of one type of synthetic ru b b er has m ade it im pera­
Table I.
Concentrations
tive th a t a qu an titativ e m ethod for its determ ination in air be
developed. T he toxicity experim ents conducted b y th e U. S.
Public H ealth Service (1) indicate th a t th e m axim um allowable
lim it for acrylonitrile is in th e neighborhood of 20 p.p.m . Con­
sequently, an analytical m ethod m u st be accurate to a t least
th a t concentration. As far as is known no analytical m ethod
has y et been reported. T he Raleigh-Jeans gas interferom eter
has been used for concentrations above 90 p.p.m . Below this
figure the results are questionable (2). In addition, th e gas in ter­
ferom eter is not very satisfactory for field use where m ixtures
of vapors are likely to be encountered.
T heoretical
S am pling
C o n cen tratio n
R a te
T im e
P .p .m .
L ./m x n .
M in .
0 .4
0 .4
0 .4
0 .4
0 .4
400
200
100
50
25
20
35
75
150
180
.------Mg.
—Y ie ld -
6 .5 4
5.51
394
189
9 7 .9
5 0 .6
3 0 .0
P .p .m .
6 .12
6 .2 9
4 .5 0
9 8 .5
9 4 .6
9 7 .9
101.2
120.0
w ater, and the excess sulfuric acid is titra te d w ith 0.0LY sodium
hydroxide, using m ethyl red as th e indicator.
T he concentration of acrylonitrile is calculated as follows:
METHOD
P.p.m . = (25.00A1'a -
T he m ethod developed in th is laboratory depends upon a
modified K jeldahl reaction in which th e absorbing solution
containing th e acrylonitrile is m ade strongly alkaline w ith sodium
hydroxide and then oxidized w ith hydrogen peroxide. Upon
refiuxing, the acrylonitrile is
converted q u an titativ ely to
am m onia, which is distilled
j,
over and collected in a stan d ­
ard acid solution.
The
am ount of am m onia evolved
is determ ined by titratio n of
th e excess acid. T he acrylo­
nitrile vapors are absorbed
in cold concentrated sulfuric
acid contained in- a suitable
absorption trap.
T his
m ethod is lim ited by th e fact
th a t there can be present in
t h e c o n t a m i n a t e d a ir no
nitrogen-bearing compounds
oth er th a n acrylonitrile.
when N a
Vb
Vs
Nb
C
=
=
=
=
—
N BV B) X ^ X C
norm ality of I I 2SOi
ml. of N aO H used in titra tio n
volum e of air sampled, liters
norm ality of N aO H used in titratio n
22,400 corrected to sam pling pressure and t emperature
DISCUSSION
T he analysis is based upon Radziszewski’s reaction (3).
mechanism is as follows:
T he
O
H 20 2
N aO H
/•
R C N — > R — C—H N 2— > N H S + R — C
A
\
A
PROCEDURE
S a m p lin g .
T w o absorp­
tion trap s (Figure 1) are filled
to a d epth of about 2.5 cm.
(1 inch) w ith glass beads, and
2 ml. of concentrated sul­
furic acid are added to each
trap . T he tra p s are con­
Figure 1. A bsorption Trap
nected in series and p u t in
A . Glass outlet tube
a water-ice b ath. T he air
B . One-hole rubber stopper
sample is draw n through a t
C. 0.25-Inch test tube
a m aximum rate of 0.4 liter
D. Glass beads
E. Glass inlet tube
per m inute. T he ra te of
sam pling is accurately m eas­
ured by m eans of a ro tam ­
eter. T he volume of air sam pled should be such th a t the
to ta l sam ple consists of approxim ately 6 mg. of acrylonitrile.
T he sam ple is th en washed into th e reflux flask (Figure 2), and
0.2 gram of copper acetate is added as an inhibitor to p revent
polym erization.
A n a l y s i s . Tw enty-five milliliters of 0.0251V sulfuric acid are
p ipetted into th e titratio n beaker an d diluted w ith distilled w ater
un til the bubbler is a t least 1.25 cm. (0.5 inch) below th e surface.
T he reflux flask containing th e sample is p u t into place and th e
system is completely closed. T he acid sam ple is m ade alkaline
by adding 50 ml. of 50 per cent sodium hydroxide to th e closed
system by m eans of the separatory funnel. T h e residual sodium
hydroxide in th e reflux condenser is w ashed down w ith 10 ml.
of distilled w ater, and 20 ml. of 30 per cent hydrogen peroxide
are then added slowly from th e separatory funnel. W hen the
addition is completed, th e sam ple is refluxed gently for 30 m inutes.
A t the end of th e reflux tim e, th e w ater is drained from th e reflux
condenser and approxim ately one half of th e sam ple is distilled
over. T he second condenser is then w ashed down w ith distilled
Figure 2. Apparatus for D e ­
termination of A crylo n itrile
A . 200-cc. balloon flaslc
B. Ground-glass joint
C . Reflux condenser
D . Separatory funnel
E. No. 2 condenser
F . Sintered-slass bubbler
63
O Na
T he use of hydrogen
peroxide reduces th e re­
flux tim e from 4 hours
to 0.5 hour, and also
drives th e reaction to
completion. Low yields
due to polym erization
of th e acrylonitrile are
prevented by th e addi­
tion of copper acetate
as an inhibitor, and by
lim itation of th e size of
th e sample. A sam ple
containing approxi­
m ately 6 mg. of acrylo­
nitrile proved to be m ost
satisfactory because it is
dilute enough to p re­
v e n t polym erization and
y e t large enough to give
a n accu rate analysis.
T h e maxim um sam pling
ra te of 0.4 liter per
m inute is very critical
for th e typ e of absorp­
tion equipm ent de­
scribed in th is article.
A higher ra te will re­
su lt in loss of sample.
64
INDUSTRIAL
AND
ENGINEERING
T able I contains the results of th e analyses of known concen­
tratio n s of acrylonitrile vapors, prepared in a gas cham ber of
1000 liters’ capacity by the introduction of m easured am ounts
of liquid acrylonitrile. T he various concentrations were chosen
to indicate the accuracy of the m ethod over a wide range.
Over the entire range of concentrations th e only variable is
th e tim e of sam pling. T he large relative error in th e analysis
of the 25 p.p.m . sam ples is misleading, since th e actu a l error
is only 5 p.p.m . T he principal reason for th e error in the
results of the low concentrations is th e difficulty of m easuring
th e exceedingly small am ounts of liquid acrylonitrile required
in m aking up these concentrations in the 1000-liter gas cham ber
available.
CHEMISTRY
Vol. 16, No. 1
ACKNOWLEDGMENTS
T he authors gratefully acknowledge th e cooperation of V.
M igrdichian of th e A m erican C yanam id C om pany for his
suggestions as to th e sam pling m edium, and of IV. P. T yler, T.
R . Steadm an, and J. C. M cCool of th e B. F. G oodrich Com pany
in th e developm ent of th e analysis.
LITERATURE CITED
(1) D u d ley, N . C ., and N eal, P . A ., J . In d . H yg. Toxicol., 24, 2 7 -3 6
(1942).
(2) P roceedings, Conference on H ealth H azards in R ubber Industry,
M ay 29, 1942.
(3) R adziszow ski, B er., 18, 335 (1SS5).
Spectrophotometric Determination of Iodine Liberated in
the Oxidation of Carbon M onoxide by Iodine Pentoxide
B E R N A R D S M A L L E R a n d J O H N F. H A L L , J r .
A e ro M e d ical Laboratory, A rm y A ir Forces, W right Field, Dayton, O h io
A spectrophotometric method for measuring the Iodine
liberated in the oxidation of carbon monoxide b y iodine
pentoxide Is described. The Iodine concentration was
measured at 3 50 millimicrons against a water blank, using a
Coleman photoelectric spectrophotometer. Results in terms
of carbon m onoxide were calculated b y the use of a for­
mula derived from the calibration curve and the chemical
reactions Involved. The method Is sensitive, convenient
(requiring the preparation of only one reagent, 1 per cent
potassium iod ide), and reliable for concentrations of
carbon m onoxide as low as 0 .0 0 5 to 0.001 per cent.
The range of ap plicability is 0.001 to approximately 0.2
per cent. Interfering substances (gasoline vapor, water,
etc.) are effectively removed b y the use of chromic acid,
silica gel, and phosphorus pentoxide as the absorbing or
"scouring" agents. A s described, the method has an
accuracy of ± 1 0 per cent In the range 0.001 to 0 .0 1 0 per
cent carbon monoxide. For analyses of carbon monoxide
concentrations above 0 .0 1 0 per cent and within the limits
of usefulness of the apparatus (0 .2 per cent) the accuracy
of the method is Increased to ± 3 to 5 per cent.
O
N E of the stan d ard m ethods and probably th e m ost widely
used q u an titativ e procedure a t present available for th e de­
term ination of sm all am ounts of carbon monoxide in a ir is th e
iodine pentoxide m ethod. N um erous w orkers, including Edell
(J), Sendroy (S), and V andaveer and Gregg (4, 5) have suggested
m odifications or have added im provem ents to th e m ethod as
originally introduced by de la H arfe an d R everdin (2).
I n these m ethods th e gas sam ple is usually first passed through
chrom ic acid to rem ove volatile hydrocarbons, th en through po­
tassium hydroxide and phosphorus pentoxide to rem ove w ater
vapor, and finally over d ry solid iodine pentoxide a t 150° to 160 0 C.
to produce q u an titativ ely th e volatile products carbon dioxide
an d iodine. T hese are absorbed or collected in barium hydrox­
ide in case carbon dioxide is m easured, or in potassium iodide in
case th e iodine is to be determ ined. Iodom etric m easurem ent by
titratio n w ith sodium thiosulfatc, using starch as th e indicator,
has been generally adopted as the m ethod of choice.
However, the necessity for frequent determ inations in this
laboratory of low (0.005 to 0.001 p er cent) carbon monoxide
concentrations disclosed several difficulties and disadvantages
w hich m ade desirable some oth er m eans of m easuring th e iodine
liberated. T h e following m ay be m entioned specifically: (1) the
very dilute (0.001A0 sodium thiosulfate solutions used, because
of a slow ra te of decom position, require periodic checks upon
th eir concentration; (2) slight o v ertitratio n of th e end point is a
p o ten tial source of error; an d (3) duplicate titratio n s on th e same
sam ple are som etim es impossible.
APPARATUS A N D REAGENTS
Iodine pentoxide ap p aratu s. Colem an photoelectric spectro­
photom eter, M odel 10S. One per cent potassium iodide solution
(M erck’s c.p. g ranular potassium iodide is satisfactory). Pipets,
10.0, 1.0, and 0.2 cc.
METHODS A N D PROCEDURE
T h e above-m entioned difficulties could be elim inated, it was
found, by th e use of a spectrophotom etric m ethod of m easuring
th e iodine concentration. F o r this purpose a Coleman double
m onochrom ator photoelectric spectrophotom eter (M odel 10S)
having a range of 350 to 1000 millim icrons was used. D eter­
m inations of th e spectral
transm ittance of iodine in 1
per cent potassium iodide,
the calibration curve relating
concentrations to logarithm
of th e percentage tran sm it­
tance, as well as all m eas­
urem ents of iodine concen­
tratio n in blanks and un­
know n sam ples were per­
formed w ith th is instrum ent.
The
spectrophotom etric
m ethod . of m easuring the
a m ount of iodine in a 1 per
c e n t potassium iodide solu­
tion requires no chemical
reaction of th e iodine b u t is
dependent only upon the
absorption of lig h t b y the
iodine in solution. F o r th e
range of iodine concentra­
tions used th is absorption
has been observed to follow
B eer’s law, in th a t th e rela­
tionship between the
logarithm of th e percentage
tran sm ittan ce and th e con-
January 15, 1944
ANALYTIC
L EDITI O N
65
Blank analyses are carried o u t by filling calibrated containers
w ith nitrogen (commercial), passing the gas through the ap p ara­
tus, and flushing exactly as w ith unknow n sam ples. F rom the
am ount of iodine m easured th e equivalent carbon monoxide is
then calculated. B lank analyses are always perform ed prior to
th e analysis of unknow n sam ples and the blank value in term s of
percentage carbon monoxide is always su b tracte d from the un­
known value to o btain the tru e or “ corrected” carbon monoxide
percentage concentration. T he only reagent required for an
analysis is the 1 per cent potassium iodide solution. T his should
be prepared from a pure grade of potassium iodide (since im pure
preparations readily liberate appreciable iodine) and should be
renewed about once a week. T he solution should bo stored in a
dark glass bottle.
RESULTS
j
350
45 0
550
1------ 1------1------
650
75 0
WAVE LENGTH-MILLIMICRONS
Figure 2.
Spectral Transmission Curve of Iodine
(5 X 10 -» N )
In 1 percent potassium Iodide solution as determined with Coleman
photoelectric spectrophotometer (Model 10S) using a 5-mm. slit.
Cuvettes were circular tubes of 16.5-mm. diameter and comparison
tube was a blank solution (10 cc.) of 1 per cent potassium iodide
centration of absorbing m aterial (iodine in 1 per cent potassium
iodide) is linear and th e calibration curve is a stra ig h t line.
T his m ay be expressed as:
In Figure 2 the spectral tran sm ittan ce curve of iodine (5 X
10~W ) in 1 per cent iodide solution is presented. M aximum
absorption of light b y iodine (5 X 10~SAT) in 1 p er cent po tas­
sium iodide solution occurs in th e region of th e ultraviolet.
However, a t 350 millim icrons (th e lower lim it of th e Colem an in­
strum ent) th e absorption is sufficiently g reat (90 per cent) to pro­
vide a m eans of m easurem ent th a t is as sensitive an d accurate as
th e routine sodium thiosulfate titra tio n m ethod. B oth th e ti­
tratio n m ethod an d th a t described here have been em ployed in
this lab oratory for determ ining carbon monoxide in air samples.
However, th e au th o rs have found th e titra tio n m ethod less satis­
factory for concentrations of carbon monoxide below 0.005 per
cent. Below this value th e titra tio n m ethod was less convenient
and more tim e-consum ing th a n th e spectrophotom etric.
T h e relationship betw een th e concentration of iodine over a
wide concentration range and th e logarithm of percentage trans­
m ittan ce is linear. T h e value of th e calibration constant, K ',
which expresses this relationship betw een concentration in equiv­
alents per liter an d logarithm of percentage transm ittance, is
5.1 X 1 0 "‘.
C oncentration = K ' X —^ r °
wherc To equals the intensity of incident light, T equals the in­
ten sity of tran sm itted light (after passing through sam ple), and
K ' equals the calibration co n stan t expressing the relationship
betw een concentration of iodine in 1 per cent potassium iodide
solution and th e logarithm of the percentage tran sm ittan ce. Ab­
sorption of light by iodine (5 X 10“ bV) in 1 per cent potassium
iodide increases as the wave length of light used decreases into
th e ultraviolet. As a result of this observation all m easurem ents
were m ade a t 350 millimicrons, th e point of g reatest absorption
of th e w ave-length range available w ith th is ty p e spectrophotom ­
eter.
G as sam ples are collected in calibrated m etallic containers of
approxim ately 1600- to 1700-cc. (STP) capacity; sm aller 200to 250-cc. (STP) glass tonom eters m ay be used for sam pling p ur­
poses, b u t give less satisfactory results. T he sam ples are then
passed through th e iodine pentoxide ap p aratu s by displacem ent
w ith w ater, th is m ethod being preferable from the stan d p o in t of
safety, ease of adjustm ent, and convenience to th e suction m ethod
usually employed. I t was experim entally observed th a t th e
displacem ent of the 1600- to 1700-cc. sam ple in 60 m inutes fol­
lowed by a 30-m inute flushing w ith nitrogen gas a t th e same ra te
of flow gave m ost accurate results. However, size of sam ple, its
concentration, and degree of accuracy dem anded are all factors
which m ay m odify the tim e necessary to complete an analysis.
F o r the collection of the liberated iodine th e original Gomberg
tu be used routinely w ith th e iodine pentoxide ap p aratu s was re­
placed by an absorption tub e designed and adap ted by J. W.
H eim of this laboratory for the present m ethod of analysis (Figure
1). T he iodine passes through the collecting tube, which extends
alm ost to the bottom of the absorption vessel, and bubbles o u t
through the sintered-glass filter a t its end. T h e am ount of iodide
solution used (10 cc.) is sufficient, of course, to keep th e filter well
below the liquid level. T his tube has m arked advantages over
th e Gomberg type, in th a t its w ide-m outhed ground-glass stopper
perm its ready access to the iodide solution for sam pling a t the
com pletion of an analysis; it is easily washed ou t or cleansed;
an d q u an titativ e results obtained w ith its use have proved it an
efficient type of absorption vessel.
0 10
co
<2
1
<
8
cc
Y-
Z
Id
o
z
<
X
o
O
T IM E -H O U R S
Figure 3. Changes in Percentage Transmission of Iodine
in 1 Per Cent Potassium Iodide as a Function of Time
T he change in th e percentage light transm ission of iodine in 1
p er cen t potassium iodide solution as a function of tim e is illus­
tra te d in Figure 3 and th e q u an titativ e effects of such changes
are presented in T able I. D ilutions of 1 to 50 showed changes
equivalent to only a 6 p er cent decrease in th e percentage of car­
bon monoxide after one hour. If spectrophotom etric readings
are m ade w ithin 5 to 10 m inutes following th e com pletion of the
sam ple run, th e error due to any change in lig h t transm ission m ay
be considered negligible.
B lank analyses using nitrogen gas (commercial) an d performed
as previously described gave fairly consistent values, th e average
being eq uivalent to a carbon monoxide concentration of 0.0004
per cent. T hese results are sum m arized in T ab le II.
INDUSTRIAL
Tabic I.
AND
ENGINEERING
Effect of Changes in Percentage Transmission of Iodine
in 1 Per Cent Potassium Iodide Solution with Time
Original^
C o n ce n tratio n
D ilu tio n
CO
%
1 :1 0
1 :2 0
1 :5 0
0.0 1 3 7
0 .0 1 3 6
0 .0 1 1 6
1 H our
5 H o u rs
23 H o u rs
CO
C h an g e
CO
C h an g e
CO
C h ange
%
%
%
%
%
%
0 .0 1 2 0 - 1 2 . 4 0 .0 1 0 9 - 2 0 . 4 0 .0 1 0 3 - 2 4 . 8
0.0120 -1 1 .8 0 .0 1 1 2 - 1 7 . 6 0 .0 1 0 7 - 2 1 . 4
0 .0 1 0 9 - 6 .0 0 .0 0 9 1 - 2 1 . 6 0 .0 0 8 5 - 2 6 . 8
Table II.
Blank A n alyse s
Sam ple V olum e
Cc.
1 2 /5 /4 2
1 2 /5 /4 2
1 2 /8 /4 2
1 2 /1 2 /4 2
1 2 /1 2 /4 2
1 2 /1 2 /4 2
1 2 /1 9 /4 2
1 /2 /4 3
1 /6 /4 3
1660
1610
1630
1590
1600
1580
1640
1630
1650
0 .0 0 5 0
4 /1 2 /4 2
4 /1 2 /4 2
4 /1 2 /4 2
4 /1 2 /4 2
4 /1 1 /4 2
K ' X (0.056 X 1000) X 100 X 10 X (2 - log % T)
cc. of sam ple (STP)
where K ' is 5.1 X 10- t and 10 ml. of 1% potassium iodide are
used.
Of m
0 .0 0 0 5
0.0 0 0 4
0.0 0 0 3
0 .0 0 0 5
0.0 0 0 4
F inally
Carbon M o no xide
(C om p ariso n of B u re a u of S ta n d a rd s a n d W rig h t F ie ld A nalyses)
D a te
pn
/0
0 .0 0 0 2
0 .0 0 0 2
A n opportunity to check th e accuracy and sensitivity of this
m ethod was offered when known sta n d a rd sam ples of carbon m on­
oxide m ixtures prepared an d analyzed b y th e N atio n al B ureau of
S tandards, W ashington, D . C., were sen t to W right Field for
check analyses. T he com parative results are presented in T able
I I I . In general, values w ith th e spcctrophotom etric m ethod
varied from 4 to IS per cen t below those reported b y th e B ureau
of S tandards, w ith results averaging 10 to 11 per cent lower.
While the m ethod is capable of detecting concentrations as low as
0.001 per cent, th e lim it of accuracy is approxim ately 5 to 10 per
cent for concentrations below 0.010 per cent.
B u re au of
S ta n d a rd s A nalyses
%
0 .0 0 7 6
th e chemical reactions involved 1.0 ml. of O.OOliV iodine
is equivalent to 0.056 ml. of carbon monoxide m eas­
0° an d 760 m m . Tlie volum e of carbon monoxide a t 0°
mm. is th en calculated as follows:
M po
5.1 X 1 0 -6 X (0.056 X 1000) X 1000 X (2 - log % T)
/0
“
cc. of sam ple (STP)
CO (as Io d in e E
%
0.0 0 0 3
0 .0 0 0 7
A v. 0 .0 0 0 4 M D « ( ¿ 0 0.0001
Table III.
W rig h t Field
A nalyses
%
0 .0 0 7 0
0 .0 0 0 5
0 .0 0 6 8
0 .0 0 6 9
0.0 0 4 1
0 .0 0 4 8
Vol. 16, No. 1
substituting,
[N itro g en (C om m ercial) G as Sam ples]
D a te
F rom
solution
ured a t
and 760
CHEMISTRY
T he analysis of sam ples whose concentration is below 0.010 per
cen t requires consideration of several factors. As previously
m entioned, ra te of sam ple flow should be sufficiently slow to en­
sure com plete oxidation of all carbon monoxide b y th e iodine
pentoxide. F o r concentrations of 0.010 p er cen t an d below to ta l
sam pling periods of 80 to 90 m inutes gave satisfactory results.
Sample volum e is another factor to be given consideration when
accurate results are desired. C om parative results of analyses
of a stan d ard carbon monoxide m ixture when using sm all and
large sam pling volum es are shown in T able IV . A ccuracy and
dependability of analysis are obviously increased when larger
volumes are used, particularly w hen th e concentration of carbon
monoxide is below 0.010 per cent.
T he photoelectric colorim eter used w ith a filter w hich tran sm its
the greater portion of its lig h t as nearly as possible w ithin th e re­
gion of th e maxim um absorption of iodine in 1 p er cen t potassium
iodide solution—i.e., ab o u t 350 millimicrons— can be used if
the spectrophotom eter is n o t available. A calibration curve
m ust be determ ined and th e sen sitiv ity of th e photoelectric colori­
m etric procedure is som ew hat less th a n th a t obtained w ith th e
spectrophotom eter.
% CO =
X dilution used
T he value 2.86 is term ed K in the final equation below and is
determ ined as indicated from bo th th e calibration curve and the
q u an tita tiv e relationship betw een th e concentration of carbon
m onoxide an d iodine released in the oxidation reaction. The
final general equation is then:
% CO = K X (2 ~
X dilution
cc. of sam ple (STP)
As an exam ple of th e use of the above equation in calculating
results, th e following analytical d a ta are presented:
A 10-ec. portion of th e 1 per cent potassium iodide was used to
collect th e iodine liberated when a gas sam ple of 2360 cc. (STP)
was passed through th e apparatus. A 1 to 100 dilu tio n of the
iodine collected in the 10-cc. portion of th e 1 per cent potassium
iodide solution was m ade and when m easured (a t 350 m illim i­
crons) w ith th e spectrophotom eter against a 10-cc. 1 p er cent po­
tassium iodide blank set a t 100 per cent lig h t transm ission, gave
a percentage lig h t transm ission of 44.5. T h e calculation, b y sub­
stitu tio n of th e above da ta , is illustrated as follows:
D ifference
%
- 7 .9
-1 4 .5
-1 0 .5
- 9 .1
A v. - 1 0 . 5
-1 8 .0
- 4 .0
A v. —1 1 .0
= 5.1 X 10-s X 56,000 X (2 - log % T )
cc. of sam ple (STP)
%CO =
Table IV .
2.86 X (0.352)
X 100 = 0.043
2360
Standard Carbon M o no xide M ixture
[C om parison of A nalyses U sing Sm all a n d L arge Sam ple Volum es
of S ta n d a rd CO (0.012 p e r c en t)]
D a te
G as V olum e
Cc.
1 0 /2 3 /4 2
1 0 /2 3 /4 2
1 0 /2 3 /4 2
1 0 /2 4 /4 2
1 0 /24/42
1 0 /2 4 /4 2
1 0 /2 4 /4 2
264
204
264
270
270
270
270
1 0 /2 6 /4 2
1 0 /2 6 /4 2
1 0 /2 6 /4 2
1 0 /2 7 /4 2
1 0 /2 7 /4 2
1 0 /2 7 /4 2
1 0 /3 0 /4 2
1 0 /3 1 /4 2
1 1 /1 3 /4 2
1660
1660
1660
1660
1660
1660
16001060
1645
CO F o u n d
%
0 .0 11
0.009
0.0 1 3
0 .0 0 6
0 .0 10
0 .0 0 6
0 .0 1 5
A v. 0 .0 1 0
0.0 1 3
0 .0 10
0 .0 11
0 .0 12
0 .0 11
D e v iatio n
%
+ 0 .0 0 1
- 0 .0 0 1
+ 0.0 0 3
-0 .0 0 4
0.000
-0 .0 0 4
+ 0 .0 0 5
M D - ( a .) 0 .0020
+ 0 .0 0 2
- 0 .0 0 1
0.000
+ 0 .0 0 1
0.000
0 .0 1 3
+ 0 .002
0 .0 11
0 .0 11
0 .0 12
0.000
0.000
+ 0.0 0 1
A v. 0.011
MD -
( * ) 0 .0 0 0 8
LITERATURE CITED
(1) E d ell, G . M ., I n d . E n o . C h e m ., 20 , 275 (1928).
(2) H arfe, C. de la, and R everdin, F ., Chem .-Ztg. 1 2 , 1726 (1888).
(3) Sendroy, J., in Peter3 and V an Slyke, “Q u an titative Clinical
C hem istry, V ol. II, M eth od s” , B altim ore, W illiam s & W ilkins
Co., 1932.
(4) V andaveer, F . E ., Gas, 18, 24 (1942).
(5) V andaveer, F . E ., and G regg, R . C ., I n d . E n o . C h e m ., A n a l .
E d ., 1, 129 (1 9 2 9 ).
»
Yeast M icrobiological Methods for Determination
of Vitamins
Pan to th e n ic A c i d
L A W R E N C E A T K IN , W I L L I A M
L . W IL L IA M S , A L F R E D S. S C H U L T Z , a n d C H A R L E S N . F R E Y
The Fleischmann Laboratories, Standard Brands Incorporated, N ew Yo rk 51 , N . Y .
The yeast growth method for determination of pyridoxine
is modified for the determination of pantothenic acid.
The basal medium contains ammonium sulfate as a nitrogen
source and in addition sufficient asparagine to prevent
interference due to 0-alanine. Extracts of substances to
be assayed are prepared b y aqueous extraction under
pressure (1 5 pounds for 15 minutes) at p H 5.6 to 5 .7, by
E y east microbiological m ethod recently described for the
T Hdeterm
ination of pyridoxine (0) can, w ith certain modifica­
tions, be used to determ ine pan to th en ic acid. T h e y east grow th
factor activ ity of pantothenic acid was know n before its need in
anim al n u tritio n w as established (72). T h a t th e y east m ethod
was n o t further developed as an assay m ethod-by the discoverers
of pantothenic acid was due in p a rt to th e interference of 13alanine (S). 0-Alanine is a stru ctu ra l p a rt of th e vitam in mole­
cule b u t is itself w ithout vitam in activ ity for higher anim als, al­
though under certain conditions it can replace pantothenic acid
as a yeast grow th factor. T h e inclusion of asparagine in th e
y east grow th m edium tends to reduce th e effect of 0-alanine (73),
b u t ap p aren tly it was not realized th a t th e presence of sufficient
asparagine fu rth er reduces th e interference to an insignificant
level. A sparagine does not, however, affect th e activ ity of p an ­
tothenic acid under th e conditions employed.
T he assay m ethod described here has been used b y th e authors
for some tim e and m ay possess certain advantages over th e micro­
biological m ethod of Pennington, Snell, an d W illiams (3). T he
m ethod is rapid, 16 to 18 hours being allowed for y east grow th,
an d it is especially adapted for turb id im etric m easurem ent of
th e y east grow th w ith a photoelectric colorim eter. Furtherm ore,
it offers an opportunity for checking assay results w ith a different
ty p e of microorganism .
APPARATUS
^ T he apparatus em ployed has previously been described (7, 2).
T h e utility of the E velyn photoelectric colorim eter was also
studied. Using the te st tubes usually supplied w ith th e E velyn
and the 660 (red) filter, th e absorption curves were found to be
essentially the sam e as w ith th e L um etron instru m en t.
SOLUTIONS
One lite r contains 200 gram s of
gram s of m onopotassium phos­
p hate, 1.7 gram s of potassium chloride, 0.5 gram of m agnesium
sulfate (M gS0(.7H 20 ) , 0.5 gram of calcium chloride (CaCl2.2H 20 ) ,
0.01 gram of m anganese sulfate, an d 0.01 gram of ferric chloride.
P o ta s s iu m C i t r a t e B u f f e r .
One liter contains 100 gram s of
potassium citrate (K SC«HS0 7 .H 20 ) an d 20 gram s of citric acid
(H 3C ,H 50 7.H 20 ).
S u g a r a n d S a lts S o lu tio n .
c .p . dextrose (anhydrous), 2.2
T h i a m i n e S o l u t i o n , 10 m i c r o g r a m s p e r m l.
P y r i d o x i n e S o l u t i o n , 1 0 m i c r o g r a m s p e r m l.
I n o s i t o l S o l u t i o n , 1 m g . p e r m l.
B io tin S o lu tio n .
S. M . A. Corp. biotin
concentrate N o.
5000, diluted so th a t it contains approxim ately 0.8 m icrogram of
biotin per ml. On one occasion th is m aterial was found to con­
ta in appreciable am ounts of p an to th en ate and as a consequence
a high blank, 15 to 20 per cent absorption, w as obtained. T he
p antothenate was readily destroyed by alkaline h eat tre a tm en t
and th e blank or zero value retu rn ed to th e previous value (less
th a n 10 per cent). P ure biotin was also tried when it recently
enzyme digestion at the same p H , or by enzym e digestion
followed b y aqueous extraction (1 5 pounds for 15 minutes).
The choice of extraction method depends upon the sub­
stance, since some have pantothenate in a bound form
whereas others do not. The results of assays of a number
of representative substances compare favorably with results
obtained b y other methods.
became available and th e results indicate th a t it m ay be sub­
stitu te d for th e crude concentrate.
A m m o n iu m S u l f a t e S o l u t i o n , 150 mg. p er ml.
A sp a ra g in e S o lu tio n .
One hundred m illiliters contain 3
gram s of (-asparagine. H eatin g to 100° C. is necessary to dis­
solve th e asparagine.
T he solutions are sterilized by heating in flowing steam for 30
m inutes on th ree consecutive days, an d m ay th en be stored a t
room tem perature u n til used. T h e pyridoxine solution is care­
fully p ro tected from light.
P a n t o t h e n a t e S t a n d a r d S o l u t i o n . d-Calcium p an to th e­
n a te (synthetic) is used as a p rim ary stan d ard . A freshly dis­
solved solution containing 1 mg. p er ml. is k e p t in th e refrigerator
and used as a working stan d a rd for n o t m ore th a n 2 to 3 weeks.
Im m ediately before use a portion of th e w orking sta n d a rd is di­
lu ted w ith distilled w ater, to contain 0.1 m icrogram (100 millimicrograms) p er ml.
YEAST INOCULUM
Fleischm ann culture 4228, a strain of Saccharomyces carlsbergensis, is grown on m alt agar slants (Difco) for 24 hours a t 30°
and th en m ay be sto red in th e refrigerator for n o t m ore th an one
m onth. T h e day before an assay run, a fresh tran sfer is made
an d incubated a t 30°. Y east is tran sferred from th is sla n t to a
tube of sterile saline u n til th e lig h t absorption indicates th a t th e
concentration is 3 mg. p er ml. T h e calibration (50 p er cent ab­
sorption w ith th e au th o rs’ instrum ent) is based on a known sus­
pension of m oist b aker’s y east. T he absolute q u a n tity of yeast
in th e inoculum is n o t critical an d hence th is approxim ation is
satisfactory. A 10-ml. aliquot of th is suspension is added to 90
ml. of sterile saline contained in an Erlenm eyer flask. T he final
suspension contains 0.3 mg. of m oist y e ast p er ml. and is ready
for use.
PREPARATION OF SAM PLES FOR A S S A Y
K u h n an d W endt (7) have reported th a t p an to th e n a te (filtrate
factor) m ay occur as p a rt of a nondialyzable complex of high
m olecular weight. T h e p an to th en ate con ten t of tissues an d yeast,
as m easured b y microbiological m ethods, is increased by
autolysis or enzyme digestion and th e increase appears to be due
to a decom position of th e complex. P a n to th e n a te is a relatively
u nstable com pound, being readily hydrolyzed by e ith er acid or
base. C onsequently it is desirable to prepare ex tracts for assay
b y th e m ildest an d m ost direct m eans. I n general, th ere are
three extraction m ethods available:
1. Enzym e digestion by clarase or o th er suitable enzyme
preparations. A utolysis (self-digestion) m ay be considered a
form of enzym e digestion b u t is inapplicable to m ost substances
and u n certain w ith some tissues.
2. W ate r extraction a t high tem p eratu re (autoclave) and a t
th e m ost stable p H range.
3 . Enzym e digestion followed by w ater extraction a t high
tem p eratu re.
68
INDUSTRIAL
AND
ENGINEERING
Table 1.
CHEMISTRY
Vol. 16, No. 1
Typ ical Protocol
T o each tu b e a re added 5 m l. of b a sa l p a n to th e n ic a cid -fre e m edium plus in gredients noted below . A fte r th e 10-m inute h e a t tre a tm e n t 1 ml. of y e a st
suspension (0.3 m g. of m o ist y e ast) is a d d ed to each. T h e tu b e s a re th e n shaken a t 30° for 16 hours, th e tu rb id ity is m easured, an d th e tu b e s a re re tu rn e d
to th e in c u b a to r fo r 2 h o u rs a n d m easu red again.
16 H ours
18 H ours
No.
1
2
3
4
5
6
7
8
9
I I jO
ML
ML
err
/O
4 .0
3 .5
3 .0
2 .5
0
6
2.0
1.0
0
3
10
11
2
1
0
12
3
13
14
15
16
17
18
19
20
21
22
23
--------------- A dded-------------------------- . A bsorption
2
1
0
3
2
1
0
3
2
1
0
0 .5 o
50 M y of calcium p a n to th e n a te 0
o 100 Sam e
1
1 .5 o 150 Sam e
o 200 Sam e
2
o 300 Sam e
3
o 400 Sam e
4
1
2
o
0 .0 2 5 m l. w hole urinc*>
3
4
1
2
o 0 . 1 m g. y e a s t e x tra c t ( d r y )c
3
4
1
2
M y /tu b e
P a n to th e n a te
y /o - or
y /m l.
21
P a n to ­
th e n a te
%
M y /tu b e
P a n to ­
th e n a te
y/g- or
y /m l.
A verage
y /m l. or
y/g.
7
21
16
29
4 1 .5
4 8 .5
5 9 .5
6 6 .5
31
48
5 4 .5
A bsorption
37
48
55
66
73
70
130
195
250
57
2.8
2.6
2.6
2 .5
2 8 .5
3 6 .5
5 5 .5
0 1 .5
72
130
203
250
2 2 .5
40
53
57
55
112
165
212
550
560
550
530
550
17.5
3 1 .5
4 3 .0
5 0 .5
165
215
570
550
550
538
110
2 .9
2.0
2.8
2 .7
2 .5
o
0.5 m g. d ry 200-B yeast«*
2 2 .5
4 1 .5
56
63
75
155
265
355
150
155
177
173
29
4 8 .5
6 1 .5
6 7 .5
75
155
250
317
150
155
166
159
163
o
10 m g. w h e at-h ard w in te r N o. 2 e
2 5 .5
4 1 .5
5 4 .5
6 2 .5
87
157
250
350
8 .7
7 .9
8 .3
3 2 .5
4 9 .5
60
6 7 .5
160
235
320
86
8.6
8.0
8 .3
3
4
1
2
P a n to th e n a te
3
4
8.8
7 .8
8.0
u £ alcium p a n to th e n a te , p re p a re d fro m 1 m g. p e r m l. of re frig e ra te d solution.
& N o tre a tm e n t, p reserv ed a t p H 5.6 for 24 ho u rs, 10 m l. d ilu te d to 400 ml.
j J jg e s tc d w ith clarasc for 2 d ay s a t 45° C ., 100 m g. d ilu te d to 1 lite r.
« D igested w ith c larase for 2 d ay s a t 45° C., 100 mg. d ilu te d to 200 m l., solution centrifuged u n til clear.
• W a te r ex tra c tio n , au to clav cd 15 m in u tes a t 16 p ounds p ressu re, 1 g ra m dilu ted to 100 ml.
an am ount containing only 10 to 20 microgram s: hence 100 mg.
are weighed o u t a n d after digestion a g reater dilution is made.
T he solutions are centrifuged, if necessary, to o btain a clear ex­
tra c t.
W a t e r E x t r a c t i o n ( H i g h T e m p e r a t u r e ) . Some cereals do
n o t ap p ear to require enzyme digestion (9, 10)— e.g., w heat and
w heat products—and for these w ater extraction m ay be used.
Suspend an am ount of sample estim ated to contain 5 to 10 micro­
gram s of p an to th en ate—e.g., 1 gram of w heat—in 80 ml. of water,
add 1 ml. of buffer, and a d ju st th e pH to 5.6 to 5.7, using dilute
sodium hydroxide or sulfuric acid. H eat th e suspensions in an
autoclave a t 7 kg. (15 pounds) for 15 m inutes, cool, and dilute to
100 ml. T his treatm en t does n o t always yield a clear extract
even a fter centrifuging. A sh o rt incubation a t 45° for 15 m inutes
a fte r th e addition of a knife point of clarase will usually produce
flocculation and a clear su p ern atan t fluid.
E n z y m e D i g e s t i o n F o l l o w e d b y W a t e r E x t r a c t i o n . Pro­
ceed exactly as in th e simple enzyme digestion b u t rinse th e con­
ten ts of th e te s t tube into a flask w ith 60 ml. of w ater, check the
p H and adju st, if necessary, to 5.6 to 5.7, and th e n h e at a t 15
pounds for 15 m inutes. Cool and dilute as usual.
METHOD
MILLIMICROGRAMS- PANTOTHENATE
Figure 1.
A. 16 hours
Reference Curve
B. 18 hours
T he th ird m ethod has been found necessary w ith certain m ate­
rials—e.g., fresh green peas—because of th e presence of a sub­
stance inhibitory to yeast grow th which, however, is inactivated
by a short treatm en t in th e autoclave.
E n zy m e D ig e s tio n .
Insoluble substances should be pow­
dered or dispersed in w ater w ith a W aring Blendor or its equiva­
lent. W eigh or m easure a portion of th e unknow n estim ated to
contain 10 to 20 m icrogram s of p a n to th en ate into a 40-ml. te st
tu b e graduated a t 10 and 20 ml. A dd 1.0 ml. of th e buffer and
sufficient w ater to m ake th e volume to 10 ml. H e a t in flowing
steam for 5 m inutes, cool, and add a weighed am ount of clarase
roughly equal to th e dry w eight of th e sam ple. Dissolve by
gentle shaking and a fter adju stin g th e volum e to 20 ml. ad d 0.5
ml. of benzene. C ork securely an d incubate a t 45° for 2 days
or 37.5° for 3 days. M ake th e volum e to 200 ml. w ith w ater.
W ith m aterials of high potency it is n o t convenient to weigh out
F ive milliliters of basal pantothenate-free medium plus a solu­
tion of th e unknow n o r an aliquot of th e p an to th en ate standard
solution are placed in a series of te st tubes together w ith sufficient
w ater to m ake th e volum e in each tu b e 9 ml. T h e tubes are
plugged w ith cotton, steam ed for 10 m inutes, cooled, and inocu­
lated w ith 1 ml. each of th e yeast inoculum. T h e tubes are then
shaken a t 30° C. and th e y east grow th is estim ated a t 16 and 18
hours by turbidim etric m easurem ents m ade directly on th e tubes
w ith th e photoelectric colorim eter. E ach assay run includes a
series of tubes which are used to construct th e reference curve.
T his series is m ade w ith th e following levels: 0, 5 0 ,100,150, 200,
300, an d 400 millimicrograms p er tube. F or assay runs contain­
ing more th a n 25 tubes two reference series are included, one a t
th e beginning and one a t the end of the run, and th e results of the
tw o are averaged to construct th e reference curve.
T he basal m edium for 20 assay tubes is prepared by mixing the
stock solutions in th e following proportions: sugar and salts
solution, 50 m l.; potassium citrate buffer, 10 m l.; inositol solu­
tion, 5 m l.; amm onium sulfate solution, 5 m l.; thiam ine solution,
5 m l.; pyridoxine solution, 5 m l.; biotin solution, 5 m l.; aspara­
gine solution, 12.5 m l.; an d w ater to 100 ml.
I t is n o t essential to prepare this m edium fresh for each run.
Larger batches m ay be prepared and stored a t a tem p erature a
few degrees below 0° for as long as 3 m onths w ith no observed ill
effects upon th e assay run.
ANALYTICAL
January 15, 1944
T h e protocol of a typical assay ru n in w hich representative
m aterials were assayed is shown in T able I and th e reference
curve is given in Figure 1. In practice th is curve is plo tted on
ordinary graph paper and values for th e unknow n are obtained
by interpolation. T he estim ated potency is an average of the
value a t each assay level and a t bo th 16 an d 18 hours. T he aver­
age deviation from th e m ean is a b o u t 4 per cent. If m ore th an
tw o of th e eight values obtained for each assay deviate from the
m ean by more th a n 10 per cent, th e assay is usually repeated.
R esults are reported as p a n to th en ate co n ten t an d are based
on d-calcium p an to th en ate as th e prim ary stan d ard w ith o u t con­
version to th e equivalent w eight of free acid. T here seems to be
good precedent for this in the use of thiam ine hydrochloride and
pyridoxine hydrochloride equivalents w ith o u t conversion to th e
equivalent w eight of free base.
Table II.
Determination of Pantothenate in Presence of Its
H y d ro ly tic Products
C om position of M ix tu re
C om pletely
h ydrolyzed
calcium
C alcium
p a n to th e n a te
p a n to th e n a te
M g.
M g.
0
1.0
0 .2 5
0 .5 0
0 .7 0
0 .8 0
0 .9 0
0 .9 5
1.00
0 .7 5
0 .5 0
0 .3 0
0 .20
0 .10
0 .0 5
0 .0 0
A ssay
C alcium
p a n to th e n a te
(by assay)
Mg.
R eco v ery
of a d d ca
calcium
p a n to th e n a te
%
1.0 2
102
0 .7 0
. 0 .5 0
0 .2 9
0 .1 9
10 0
0 .10
100
110
0 .0 5 5
< 0 .0 1
93
97
95
RESULTS
P a n t o t h e n a t e E s t i m a t i o n in t h e P r e s e n c e o f /3 -A la n in e .
Since /3-alanine is a p o tential interfering substance in this assay
it is desirable to determ ine the lim its of the interference. T here
is no evidence th a t /3-alanine occurs as such in n atu re except as a
degradation product of pan to th en ate. C onsequently, th e in­
fluence of /3-alanine on the assay w as studied by analyzing a se­
ries of m ixtures representing p an to th en ate in various stages of
hydrolysis.
A solution of com pletely hydrolyzed p an to th en ate was pre­
pared by heating 10 mg. of th e calcium salt dissolved in 10 ml. of
12V and 42V sodium hydroxide, respectively, a t 15 pounds pres­
sure for 2 hours and th en cooling and neutralizing. T h e panto­
th en ate rem aining was determ ined after adding a known am ount
of pantothenate to the solution. T he assay showed 1.5 per cent
of the p an to th en ate left in the 12V tre a te d solution and less th an
1 per cent in the 42V treated solution.
In order to determ ine the am ount of /3-alanine produced b y th e
hydrolysis the basal m edium was modified b y th e omission of
asparagine, th e inclusion of Z-leucine, and th e su bstitution of
pure biotin for the crude w hich contains /3-alanine. A reference
series of grow th tests w ith increasing am ounts of /3-alanine then
supplied d a ta for a reference curve, which was used to determ ine
the /3-alanine content of th e hydrolyzates. T heoretically each
100 parts of calcium p an to th en ate should yield 37.3 of /3-alanine.
T he 12V solution assayed 32.5 p arts or 87 per cent and the 42V
solution 30.25 parts or 81 per cent. A pparently a portion of th e
/3-alanine is destroyed in th e course of th e hydrolysis.
Assuming th a t th e 42V solution represented “ com pletely” hy­
drolyzed pantothenate, the experim ent described in T able I I was
made. T he hydrolyzate was mixed in various proportions w ith
freshly dissolved p an to th en ate and th e m ixtures were assayed as
unknowns. T he results show th a t little interference from 23alanine is to be expected u n til 95 p er cent or more of th e p an to ­
th en ate has been destroyed.
T he ease w ith which the present m ethod m ay be converted to a
m ethod for /3-alanine determ ination suggests th a t it m ay be use­
ful in determ ining the nature of th e loss of p an to th en ate activ ity
in th e processing of vitam in concentrates.
I n f l u e n c e o f pH o n t h e S t a b i l i t y o f P a n t o t h e n a t e .
Since pan to th en ate is readily hydrolyzed by either acid or alkali,
it is desirable to establish th e p H range of m axim um stability, so
th a t extracts for assay m ay be m ade w ith a m inim um of loss.
EDITION
69
Solutions of p an to th en ate were heated a t 7 kg. (15 pounds)
pressure for 15 m inutes and a t 9 kg. (20 pounds) for 1 hour.
T h e heating a t 20 pounds was used to accentuate th e destruction;
th e m ilder heating is used in th e assay. T he solutions contained
in 1 liter: 1 mg. of calcium pantothenate, 10 ml. of potassium
citrate buffer, and enough citric acid to a d ju st the p H to 5.0.
Potassium hydroxide (30 per cent) was added to a d ju st th e pH to
various levels an d aliquots of th e solution were rem oved a t each
level. A fter heating and cooling, th e solutions were diluted to
correspond to a concentration of 100 millimicrograms p er ml. and
assayed by th e usual procedure.
As can be seen from Figure 2, the m ost stable region is between
p H 5.5 and 6.0. In practice th e p H is ad justed to betw een 5.6
and 5.7 in th e extraction procedures. A lthough a suspension of
whole w heat had a p H of 6.4, extraction w ith o u t lowering th e pH
showed no significant loss. R elatively large q u antities of organic
m a tte r m ay have a protective action on p an to th en ate. I t is de­
sirable, however, to a d ju st th e p H to th e m ost stable range in
routine analysis, unless investigation shows it to be unnecessary.
Clarification of aqueous ex tracts of starch y substances by the
short clarase m ethod proceeds best a t p H 5.6 to 5.7.
E n z y m a tic D ig e s tio n .
E arly p an to th en ate assays of certain
substances by the microbiological m ethods did n o t agree w ith
chick assays (9). T h e discrepancy is now believed to be due to
th e existence of a bound form of p anto th en ate available to the
chick b u t n o t to th e microorganism . D igestion w ith various en­
zyme preparations liberates th e bound pan to th en ate an d tends
to bring th e assay results by th e tw o m ethods into som ewhat
b etter agreem ent. W aism an and Elvehjem (22) found pancreatin satisfactory, Cheldelin el al. (3) prefer takadiastase,
whereas Strong, Feeney, and E arle (9) recom mend clarase. T he
au th o rs have com pared clarase w ith a few other preparations and
find it satisfactory an d furtherm ore relatively low in color and in
p an to th en ate content. T he p roduct is labeled “ D iastase, Vera,
highly concentrated ‘C larase’ ” , and m ay be obtained from Eim er
and Amend, New Y ork. T he p an to th en ate content of this
preparation varies betw een 2 and 4 m icrogram s per gram . W hen
clarase tre a tm e n t is used, th e results of th e assay m u st be cor­
rected for th e p an to th en ate content of th e enzyme.
pH
Figure 2. Influence of p H on Stability on Pantothenate
A. 15 pounds for 15 minutes
B. 20 pounds for 1 hour
T h e enzyme is active in th e p H range of m axim um stab ility of
p an to th en ate. M axim um liberation of bound p a n to th en ate was
obtained by digestion for 3 days a t 37° C. or 2 days a t 45° C.
H igher tem peratures did n o t ap p ear to offer an y advantage.
I t is essential to m ain tain an excess of benzene or toluene in the
digestion m ixture during incubation.
INDUSTRIAL
70
Table III.
AND
Y e ast e x tra c t
D ried y e a st
U rine
C alcium p a n to th e n a te
Table IV .
O riginal
P a n to th e n a te
C o n te n t
t / (7509
509
104
3 .2 7 /m l.
1
mg.
R esidual
P a n to th e n a te
7/(7-
2 .5
5 .0
10.0
0 . 0 0 7 /m l.
1 0 . 0 7 /m g .
D estru ctio n
%
9 9 .5
9 9 .0
9 4 .0
98 .4
99
Recovery of Pantothenate
T o tal
Added
F ound
R ecovered
7/(7-
7/(7-
7/(7-
7/(7-
R ecovery
%
8 .3
8 .3
8.0
8.0
17.4
10.4
9 .1
8.1
110
101
Y east e x tra c t
550
550
500
500
1053
1 0 S0
503
530
101
10G
D rie d y e a st
104
104
200
200
371
357
207
193
104
97
Av. 103.0
Pantothenate Content of V arious Substances
D escription
P a n to th e n a te
D eterm ined
W ater
C larase
e x tra c tio n
(autoclave) digestion
C ereais
W hole w heat A
W hole w heat B
W hite flour (p a te n t)
W hole w h eat b read (air-dry)
W hite b read (air-dry)
Y ellow corn
C itru s fru its
G ra p e fru it (fresh)
O ranges (fresh)
C larase A
B
C
M eats
P ork muscle (fresh)
P ork liver (fresh)
B eef m uscle (fresh)
B eef liv e r (fresh)
Liver c o n ce n trate No. 20 (dry)
7/(7-
7/(7.
8 .9
8 .3
4 .0
8 .9
3 .8
4 .2
8 .4
2 .7
1.9
3 .7 6
3.3 6
1.9
4 .0
3 .3
T he relative ease w ith which
p antothenate activity is destroyed by hydrolysis suggests an
added te st of the specificity of th e assay m ethod. Q u antitative
destruction of the pantothen ate activ ity by alkaline hydrolysis
would indicate th e absence of an alkali-stable, nonspecific grow th
factor in the preparation under stu d y . Sim ilarly, acid hydroly­
sis would indicate th e absence of an acid-stable factor.
Y east extract (100 mg.) and d ry y east (100 mg.) were subjected
to clarase digestion and then 10iV sodium hydroxide was added
to m ake the final concentration IN . Urine and calcium panto­
th en ate were prepared in sim ilar fashion b u t w ithout digestion.
All were heated a t 20 pounds for 2 hours, cooled an d neutralized,
and then assayed for pantoth en ate after superim posing a known
q u an tity of p an to th en ate on th e test. T he results are shown
in T able I I I . I t is ap p aren t th a t essentially all th e activ ity is
destroyed by alkaline hydrolysis. Sim ilar results were obtained
by acid hydrolysis.
R e c o v e ry o f A dded P a n to th e n a te .
T he recovery of p anto­
th en ate which has been added to th e unknow n is a necessary test
of the specificity of an assay m ethod of th is sort. T able IV gives
th e results obtained w ith w heat, y east extract, and dried yeast,
when th e p an to th en ate w as added a t th e beginning of th e extrac­
tion. T he calculated recovery is based upon th e added p an to ­
th en ate and the average recovery of 103.6 p er cent is considered
satisfactory, since each estim ate is based upon two assays.
A s s a y o p M i s c e l l a n e o u s S u b s t a n c e s . T able V gives the
results of assays on a num ber of representative m aterials. These
assays are in essential agreem ent w ith those appearing in the
literature if allowance is m ade for th e fa c t th a t m any early micro­
biological assays were m ade w ith o u t enzymic digestion. W heat
and w heat derivatives appear to give m axim um values w ith simple
aqueous extraction, b u t yellow corn requires enzymic digestion.
I t would th u s app ear difficult to generalize a b o u t the best ex­
tractio n m ethod for cereals.
C itrus fruits contain some of th e bound p an to th en ate. M eats,
anim al tissues and extracts, and y east products have a high pro­
portion of bound p an to th en ate and m u st alw ays be subjected to
enzymic digestion. M ilk an d urine do n o t ap p ear to contain any
significant proportion of bound p an to th en ate. Enzym e diges­
tion of fresh w-hole m ilk does n o t yield results as high as aqueous
extraction. T here are indications th a t fresh m ilk m ay have an
inhibitory substance w hich is destroyed by heat. M ost fresh
vegetables contain bound pan to th en ate. F resh green peas con­
tain a substance w hich is m arkedly inhibitory to y east growth
b u t which appears to be completely destroyed w hen th e autoclave
m ethod of extraction is em ployed. Simple enzymic digestion is
therefore inadequate and th e authors follow th e enzyme treatm en t
w ith autoclaving, when assaying fresh vegetables and fruits.
M ilk
P asteurized A
P a steu riz ed B
D ried skim milk
M ilk (w hey)c
3 .9
9 .3
8 .5
6 0 .0
8.2
5 9 .0
472
7 / ml.
or y /y .
H y d ro ly s is o p P a n to th e n a te .
Vol. 16, No. 1
CHEMISTRY
Table V .
A lk a lin e H yd ro lysis of Pantothenate
Sam ple
W h ea t
ENGINEERING
3 .2
3 .3 .
4 3 .0
3 .2
7 /m l.
or y /y .
1.8.’ 2*4 6
4 0 .0
P a n to th e n a te
L ite ra tu re V alue
y /o 1 2 ( 4 ). 1 2 . 8 ( 10 )
8 .3 (9), 1 1 .2 “ (S)
3 .5 (4), 5 .7 (10)
8 .S (4)
6 .9 (4)
9 .0 (9)
2 .9 (4)
3 .4 (4)
4 .7 , 5 .8 (4)
50 (11)
10 «?), 9 , 8 “ (5)
70 (4). 0 1 .5 (9)
7 /m l. or 7 / y.
4 .0 (9)
' 44, 47 (9)
y /d a y
7 /d a y
7 /d a y
U rine, norm al 24-hour excretion**
3000-5000 (9)
3410
3250
S u b je c t A
4400
S u b je c t A
4400
S u b je c t B
V egetables (frefch)
4 .5 6
3 .7 (4)
2 .0
T o m ato es
2 .5 6
2 .1
C abbage
1.8 (4)
1.06
1.3
B eets
1 .1 U)
3 .1 6
2 .3
2 .5 (4)
C arro ts
3 .9 6
G reen peas
3 .8 (4>
3 .8
2 .5 6
2 .1
3 .2 (4)
P o ta to e s
Y east
100
22*. 80* (9)
B rew ers’ (dry)
104
"¿8
200-B (dry)
240 * ’({?),’ 260 * (5)
319
500
Y e a s t e x tra c t (dry)
48
A utoclaved (dry)
“ C hick assay.
b C larase digestion followed b y auto clav in g w ith w ater.
c C asein co ag u lated w ith m in eral acid an d s u p e rn a ta n t fluid n e u tra lised
an d a u to c lav e d for assay.
4 p H of u rin e a d ju sted and assayed w ith o u t fu rth e r tre a tm e n t.
* N o enzym ic digestion.
T h e specificity of th e present m ethod is supported by a num ber
of observations: T h e values estim ated a t th e various testing
levels and a t 16 an d 18 hours show no significant drift, recovery of
added p an to th e n ate is v irtu ally q u an titativ e, th e poten tial in ter­
ference of /S-alanine has been elim inated, th e p an to th en ate ac­
tiv ity of extracts prepared for assay m ay be destroyed by alkaline
or acid hydrolysis, and the results obtained are in su bstantial
agreem ent w ith reported values obtained b y tests w hich employ
o th er microorganism s, and also in a lim ited num ber of cases w ith
th e results of th e chick assay m ethod.
In order to o b tain a m easure of th e reproducibility of the
m ethod, a carefully refrigerated sam ple of dried y east was as­
sayed te n tim es over a period of 3 m onths. T h e complete assay
including clarase digestion was perform ed each tim e. T h e mean
of th e ten assays was 166.3 m icrogram s of p a n to th en ate p er gram
w ith an average deviation of 3.2 p er cent and a sta n d a rd devia­
tion of 3.8 per cent.
SUMMARY
A y east microbiological m ethod for th e determ ination of pan­
to th en a te is described. Specificity of response to p an to th en ate
in th e presence of /3-alanine is obtained by th e inclusion of a rela­
tively large proportion of asparagine in th e m edium in addition
to am m onium sulfate. T he y e ast is grown in te s t tu b es which are
shaken a t 30° C. for 16 to 18 hours. Y east grow th is estim ated
w ith th e aid of a photoelectric colorim eter. M ethods of ex trac­
tio n of th e vitam in have been studied an d recovery experim ents
are described. E x tra c ts of substances to be assayed are prepared
by aqueous extraction under pressure (16 pounds for 15 m inutes)
ANALYTICAL
January 15, 1944
a t pH 5.6 to 5.7, by enzym e digestion a t th e same pH , o r by en­
zym e digestion followed by aqueous extraction (15 pounds for
15 m inutes).
(5)
(6)
(7)
(8)
LITERATURE CITED
(9)
(1) A tkin, L ., Schultz, A. S., and Frey, C. N ., Arch. Biochem., 1, 9
(1942).
(2) A tkin, L., Schultz, A . S., "Williams, W . L ., and F rey, C . N ., I n d .
E n g . C h e m ., 1 5 , 141 (1 9 4 3 ).
(3) C heldelin, V . H ., Eppright, M . A ., Snell, E . E ., and Guirard,
B . M „ U niv. Texas P ubl. 4 2 3 7 ,1 5 (1942).
(4) C heldelin, V . H „ and W illiam s, R . J., Ib id .. 4237, 105 (1942).
71
EDITION
(10)
(11)
(12)
(13)
Jukes, T . H ., J . B iol. Chem., 117, 11 (1937).
Jukes, T . H „ and L epk ovsk y, S., Ib id ., 114, 117 (1936).
K uhn, R ., and W endt, G ., B er., 71, 780 (1938).
Pennington, D „ Snell, E . E ., and W illiam s, R . J., J . B iol. Chem.,
135, 213 (1940).
Strong, F . M ., F een ey, R . E ., and Earle, A nn., I n d . E n g . C h e m .,
A n a l . E n ., 13, 566 (1941).
T cply, L . J., Strong, F . M ., and E lvehjem , C. A ., J . N u trition ,
24, 167 (1942).
W aism an, H . A ., and E lveh jem , C . A ., “ V itam in C on ten t of
M eat" , M inneapolis, M in n ., Burgess P ublishing C o., 1941.
W illiam s, R . J., L ym an, C. M ., G oodyear, G . H ., T ruesdail, J.
H ., and H olad ay, D ., J . A m . Chem. Soc., 55, 2912 (1933).
W illiam s, R . J., and R ohrm an, E ., Ib id ., 58, 695 (1936).
Polarographic Determination of Copper, Lead, and
Cadmium in High-Purity Z in c A llo y s
R. C. H A W K I N G S a n d H . G . T H O D E
M cM aster University, Ham ilton, O ntario, Canada
A study has been made on the application of the polaro­
graphic method of analysis in determining trace elements
(down to 1 X 1 0 " ‘ per cent) found In zinc-base die casting
alloys. Trace amounts of lead, cadmium, and tin cause
intergranular corrosion which results in a serious weakening
of the allo y. A polarographic procedure has been de­
veloped for the direct determination of copper, cadmium,
and lead in these alloys. The samples are dissolved in
hydrochloric and nitric acids, evaporated to near dryness,
redissolved, treated with hydroxylam ine hydrochloride,
and finally diluted to volume. The solution is then electrolyzed cathodically over a range of approxim ately 0.8
L L U R G IS T S have found in recent years th a t traces
M EofT Alead,
cadm ium , and tin in zinc-base die-casting alloys
tend to cause intergranular corrosion which results in a serious
w eakening of th e alloy. F o r th is reason, specifications for the
m anufacture of these alloys are very rigid, often requiring th a t
lead shall n o t exceed 0.003 per cent, cadm ium 0.003 p er cent, and
tin 0.001 per cent. T he purpose of th is investigation was to de­
velop a system of analysis for zinc die-casting alloys of the
M azak ty p e in w hich th e polarograph could be used w ith a d v an ­
tage to determ ine trace am ounts down to 10 per cent w ith high
precision and accuracy.
T he determ ination of trace quan tities b y w et m ethods of
analysis is exceedingly difficult. I t is usually necessary to use
large sam ples (100 gram s or m ore), an d to m ake repeated tim econsum ing separations. T h e polarographic m ethod, on the
o th er hand, is particularly suited to trace am ounts, th e very
n ature of which greatly reduces th e necessity for m aking separa­
tions, an d w hich com pares very favorably w ith th e spectrochemical m ethod w ith respect to th e lim its of determ ination
possible. A polarographic determ ination, where applicable, is
considerably cheaper th a n a spectrochem ical determ ination, and
can often be carried out m ore rapidly.
H eyrovsky (/,) in a review of th e applications of polarograpby
noted th a t it is possible to determ ine lead and cadm ium in zinc,
a n d gave curves for a 0.5-gram sam ple in 5 ml. of hydrochloric
acid, in w hich th e concentration of lead is 0.0050 per cent and
cadm ium is 0.0037 per cent. T erui {11) determ ined lead and
cadm ium to th e nearest 1 X 10“ * p er cent by dissolving 8 grams
of zinc in 70 ml. of 5N hydrochloric acid w ith a few drops of ni­
tric acid and evaporating to 50 ml. Ensslin (3) reported a po­
larographic m ethod for lead an d cadm ium in pure zinc, in which
th e zinc w as dissolved in n itric acid and th e resulting solution
com bined w ith different base solutions. T h e lead an d cadm ium
were determ ined separately; th e lead w ith an accuracy of 20 per
volt to obtain waves for copper, lead, and cadmium. Using
an 8-gram sample in 50 ml. of solution, these elements can
be determined with a precision of * 1 X 10~* per cent
of the sample weight. National Bureau of Standards zinc
samples have been analyzed using the above procedure
and the results found to agree very well with the certificate
value. Samples of high-purity zinc and zinc alloys have
been analyzed without difficulty. Nineteen elements
have been considered from the standpoint of possible
interference. The results indicate that trace amounts of
copper, cadmium, and lead can be determined polarographically with high precision and accuracy.
cent of th e to ta l am ount present, from 3 X 10 "3 to 5 X 10 "1 per
cent on w h at would correspond to a 100-gram sample in 1 liter of
solution. ICrossin (8) applied th e polarograph to th e analysis of
copper- an d alum inum -bearing zinc alloys for lead, an d bism uth
by m eans of precipitation w ith sodium sulfide. Seith an d Esche
(9) determ ined lead, cadmium, bism uth, thallium , an d tin in
zinc by th e polarographic m ethod. The lead, cadm ium , and
bism uth were determ ined sim ultaneously by treatin g a 5-gram
sample w ith hydrochloric acid and diluting to 25 ml. before elec­
trolysis a t 28° C. T h e thallium and tin were determ ined by
difference from th e sum of cadmium and thallium and lead and
tin, respectively. R esults are reported to th e nearest 1 X 10“ 3
per cent except for tin, w hich is lim ited to 1.5 X 10-3 per cent.
H ohn (5) in a review of polarographic m ethods of analysis o u t­
lined a m ethod for copper, lead, an d cadm ium in zinc, b u t m ade
no m ention of its accuracy or precision.
In th e work reviewed above, only one p aper deals with th e de­
term ination of im purities in zinc alloys, an d th is involves an ob­
jectionable sulfide separation w ith its a tte n d a n t errors. T he
present investigation was u n dertaken in an effort to develop a
m ethod for th e direct determ ination of copper, lead, an d cadmium
in high-purity zinc-base die-casting alloys of th e M azak ty p e w ith
special em phasis on accuracy and precision in th e region of 10
p er cent of th e sam ple weight.
APPARATUS A N D REAGENTS
T he prelim inary studies were m ade w ith a Leeds & N o rth rup
Electro-C hem ograph, and th e w ork w as concluded w ith a H ey­
rovsky polarograph M odel X I (E. H . Sargent a n d Co.). The
same capillary w as used thro u g h o u t th e investigation. The
capillary constant in 2.5.V zinc chloride w as found to be 1.37
m g.*/3 sec.- 1 / 3 when h = 36.5 cm., t = 3 .3 seconds, and tem pera­
tu re = 25 ± 0.5° C. W hen th e curves showed irregularities
traceable to fluctuations in th e drop tim e, th e capillary was
cleaned w ith concentrated nitric acid as directed b y K olthoff and
Lingane (7). T he pressure on th e dropping electrode was m ain­
tained b y using th e Leeds & N o rth ru p electrode assem bly in con-
72
junction w ith a large flask to serve as a pres­
sure regulator. T his kept the pressure constant
to w ithin ± 0 .5 mm. of m ercury over periods
of not less th an 45 m inutes. An ordinary
electrolysis cell w ith an internal anode was
used throughout these experiments. T he step
heights were m easured by th e slope intercept
m ethod—i.e., straight lines were draw n along
the principal slopes of the curve and th e v er­
tical distances between the points of intersec­
tion of the extensions of these lines were m eas­
ured w ith a m illim eter scale (/) (see Figure 1).
All work was carried o u t a t 25 =*= 0.5° C.
M ost commercial reagents contain traces of
various elem ents, and it was found necessary
to check the p u rity of the reagents used under
th e conditions existing in th e procedure out­
lined below. I t was n o t possible to procure
zinc m etal of such pu rity th a t no steps were
obtained under the operating conditions.
(The word “step ” is intended to describe th e increase in current
caused by the discharge of an ion, 7.)
Typical Polarogram and M ethod of M easuring W ave Heights
A sample of zinc was m ade uniform by reducing the sam ple
to shavings and mixing. A portion of this sample was analyzed
polarographically, using the procedure outlined, w ith o u t addi­
tion of m etal ions. T his gave th e residual cu rren t for the en­
suing determ inations.
C o n c e n t r a t e d H y d r o c h l o r i c A c i d (sp. gr. 1.19). One
hundred m illiliters of concentrated hydrochloric acid were evap­
orated alm ost to dryness in a 125-ml. Pyrex beaker. An 8-gram
sam ple of zinc was then added to this residual liquid and the
whole carried through the procedure for analysis. A fter sub­
tracting the residual current duo to the zinc, it was found th a t
for the 32 ml. of hydrochloric acid required, a correction of 5 X
10-5 per cent of cadm ium and 5 X 10-6 per cent for copper would
have to be applied. No trace of lead was found.
C o n c e n t r a t e d N i t r i c A c i d (sp. gr. 1.42). One hundred
milliliters of concentrated nitric acid were treated in a m anner
sim ilar to th a t used for the hydrochloric acid. No traces of
copper or lead were found, b u t cadmium corresponding to
0.00015 per cent in an 8-gram zinc sam ple was detected. In a s­
m uch as th e am ounts of nitric acid used were seldom in excess of
10 m l., th is am ount of cadm ium was considered negligible for the
present purpose.
H y d r o x y l a m i n e H y d r o c h l o r i d e (2Ar). T en m illiliters of
2iV hydroxylam ine hydrochloride were evaporated alm ost to
dryness and treated as was th e hydrochloric acid. No traces of
copper, lead, or cadm ium w ere found.
G e l a t i n S o l u t i o n (0.2 per cent aqueous). One gram of
gelatin was ashed in a porcelain crucible and th e residue taken
up in a few milliliters of concentrated hydrochloric acid. The
contents of th e crucible were then added to a zinc sam ple as for
hydrochloric acid. C opper corresponding to 3 X 10“ 4 per cent
and cadmium to 1 X 10“ 4 per cent in an 8-gram zinc sample were
found. Since only 2.5 ml. of th e 0.2 per cent solution are used
in an analysis, these im purities were considered negligible.
D is tille d W a te r.
F ive hundred m illiliters of w ater were
evaporated to dryness and tre ated as for hydrochloric acid.
N o detectable am ounts of copper, lead, or cadm ium were found.
To 8 gram s of turnings in a 125-ml. Pyrex beaker add slowly
25 ml. of concentrated hydrochloric acid. A fter th e first violent
reaction has subsided, add cautiously a few milliliters of concen­
tra te d nitric acid, and warm to effect solution. W hen solution
is complete, add sufficient nitric acid to m ake the volum e of
added nitric acid 5 ml. E v ap o rate on a h o t sand b a th until
salts begin to crystallize out, an d th e m ixture boils like thick
sirup. If desired, th e evaporation m ay be hastened som ewhat
by careful heating on a wire gauze. Allow the m ixture to cool
for a short time, or un til solid. W ash down w ith distilled w ater
until a b o u t 10 ml. of w ater have been added.
Add 7 ml. of concentrated hydrochloric acid and h eat on a sand
b a th until any hydrolyzed alum inum is redissolved. T his may
require 10 to 15 m inutes. T ransfer to a 50-ml. volum etric flask
w ith the m inim um of distilled w ater. Add 2.5 ml. of 0.2 per cent
gelatin solution an d 0.1 ml. of 2 N hydroxylam ine hydrochloride.
Shake and h eat un til th e solution becomes colorless. If neces­
sary, add a second portion of hydroxylam ine hydrochloride.
(Occasionally a solution m ay rem ain colored in spite of a large
excess of hydroxylam ine. If th e excess corresponds to 100 to 200
tim es th e am o u n t of iron present, th is color can usually be ig­
nored.) D ilute w ith freshly boiled distilled w ater, cool, and di­
lute of volume. Bubble w ith nitrogen in th e electrolysis cell for
15 to 20 m inutes to rem ove dissolved oxygen, and eleetrolyze
from —0.04 volt to th e discharge potential of th e supporting
electrolyte (approxim ately —0.8 volt) using a bridge poten tial of 1
volt. T h e sensitivity should be ad justed to give th e largest pos­
sible steps in the curves.
Using the procedure described below, it was found th a t the
over-all effect of im purities in the reagents am ounts to copper
0.00005 per cent, lead 0.00000 per cent, and cadm ium 0.00005 per
cent. These am ounts m ay be neglected for m ost purposes.
T he stan d ard solutions of copper, lead, and cadm ium , required
for th e calibration of th e capillary were prepared by diluting stock
solutions.
S t a n d a r d S tock S o l u t i o n f o r C o p p e r .
A 0 .2 3 / solution
of cupric n itrate was prepared by dissolving 12.714 gram s of
electrolytic copper in dilute nitric acid a n d diluting to 1 liter.
T his solution was analyzed by slow deposition, and found to be
0.1982 =*= 0.00023/ (average of four determ inations).
S t a n d a r d S t o c k S o l u t i o n f o r C a d m iu m . A 0 .2 3 / solution
of cadm ium n itrate was prepared by dissolving 22.496 gram s of
c.P. cadm ium in dilute nitric acid an d diluting 1 liter. T his solu­
tion was analyzed by electrical deposition and found to be 0.1996 ±
0.00043/ (average of three determ inations).
S ta n d a r d S to c k S o lu tio n f o r L e a d .
A 0.23/ solution of
lead n itrate was prepared by dissolving 41.458 gram s of c .P : lead
in dilute nitric acid and diluting to 1 liter. T his solution was
analyzed by th e lead acid m ethod and found to be 0.1981 ±
0.00023/ (average of four determ inations).
PROCEDURE
T he ferric iron is readily reduced to the ferrous sta te by tre a t­
ing th e warm hydrochloric acid solution w ith hydroxylam ine (2).
In th is state, th e iron will n o t interfere w ith th e determ ination of
th e copper, lead, an d cadm ium . S trubl {10) also m ade use of
this reagent in th e analysis of zinc blende which was high in iron.
T he tim e required for a determ ination of copper, lead, and
cadm ium in a zinc-base alloy using th e above procedure is about
3 hours. However, a large num ber of sam ples m ay be run a t the
sam e tim e, as only 10 m inutes are required for a polarogram after
th e sam ple is prepared.
I f copper is present in th e alloy in excess of 0.1 p er cent, it is
advisable to rem ove th e copper by electrodeposition from nitric
acid solution as follows:
T o 8 gram s of zinc in a tail-form 250-ml. beaker, add 50 ml. of
w ater and th en add 23 ml. of nitric acid in sm all portions. W hen
all th e acid has been added, boil to complete solution, dilute to
100 ml. (or sufficient volum e to cover th e electrodes), and electrolyze a t 4 am peres an d 3 to 4 volts for 1 hour, w ith a ro tatin g
gauze anode and a gauze cathode. A t the end of this tim e, wash
down the cover glass and beaker and continue for a n o th er 5
January 15, 1944
ANALYTICAL
m inutes. Carefully rem ove th e electrodes, while washing w ith a
heavy stream of w ater. U nder no circum stances should the
circuit be broken before th e electrode is completely free of acid.
Rinse th e e le c tro d e several tim es in 95 per cent alcohol, shake
free of excess alcohol, and d ry by revolving rapidly over a B un­
sen flame after igniting the film of alcohol. Weigh as pure
copper. Replace the anode, which m ay have lead oxide deposited,
in th e electrolyte and h eat the whole to boiling. T hen wash the
electrode and remove it and boil th e solution down to incipient
crystallization. Add hydrochloric acid and carry o u t the proce­
dure as for zinc w hich is low in copper. Care m u st be taken to
remove all excess nitric acid. T o do this, an extra evaporation
w ith 10 ml. of hydrochloric acid is recommended before addition
of the 7 ml. of hydrochloric acid and 10 m l. of w ater to redissolve
th e hydrolyzed alum inum .
T his modification increases th e to ta l tim e required for an
analysis, b u t when a large n um ber of sam ples are to be run, this
increase is considerably lessened. T h e results obtained for the
copper by this m ethod are usually slightly high (0.03 per cent
high for 3 per cent copper). N o traces of cadm ium or zinc were
found in th e deposit when th e deposit h ad been redissolved and
deposited, and th e electrolyte examined polarographically.
Table I.
E lem e n t
Cu
Pb
Cd
Calibration Constants for Copper, Le ad , and Cadmium
Leeds & N o rth ru p
2
A verage
% / micro­ % / m icro­ % / micro­
ampere
ampere
ampere
1
0 .0 1 8 1 5
0 .0 2905
0.018G9
0 .0 1864
0 .0 3009
0 .0 1873
0.0184
0 .0 2 9 6
0.0 1 8 7
S a rg e n t
2
A verage
% /m ic ro ­ % /m icro ­ % / micro­
ampere
ampere
ampere
1
0 .0 2040
0.03334
0 .0 1 8 7 5
0 .0 2 0 7 8
0 .0 3413
0 .0 1 8 8 8
0 .0 2 0 6
0 .0 3 3 8
0.0 1 8 8
RESULTS A N D DISCUSSION
T he capillary was calibrated by m aking additions of copper,
lead, and cadm ium from th e diluted stock solutions. T h e cali­
b ratio n was carried out over a concentration range of 1 0 ~ W to
3 X 10 ~*M on both the Leeds & N o rth ru p and the Sargent
polarographs. T he step heights were obtained by difference
from the residual current of th e zinc used as a supporting electro­
lyte, and those produced by th e zinc plus added ions. In the
lower concentration range (1 X 10"5 to 5 X 10-6ilf), th is residual
cu rrent am ounted to from te n to tw en ty tim es th e increase in
current due to the added ions, and, accordingly a large error was
introduced. T his error was m ore significant when using th e
Leeds & N o rth ru p instrum en t th a n when em ploying th e Sargent
a p p aratu s. T h e fact th a t th e la tte r has over twice th e m axi­
m um sensitivity of the form er would account in p a rt for this dif­
ference in th e results.
T h e calibration constants for tw o successive calibrations using
tw o different zinc sam ples as a supporting electrolyte are given in
T able I.
T he factors were obtained in term s of per cent p er microam pere
for each m etal for the sake of convenience in calculating th e
values from the step heights. T he basis for th e calculation is an
8-gram sample in 50 ml. of solution.
In order to determ ine the precision of th e m ethod, and to dis­
cover the g reatest source of erro r in th e polarographic proce­
dure, five' series of determ inations were carried out, using the
S argent instrum ent (Table II).
1. Precision of repeated determ inations on th e sam e cell.
2. Precision of repeated determ inations on different aliquots
of the same solution.
3. Precision of repeated determ inations on different sam ples
of th e sam e alloy.
4 and 5. Precision of repeated m easurem ents of th e sam e curve
by different individuals and by the sam e individual.
These results indicate th a t th e mean deviation of m easurem ent
in all cases is approxim ately one half of the to tal m ean deviation
of th e procedure. T he deviation is n o t significant, however, in
th a t it barely affects th e fourth place of decimals. T he over-all
EDITION
73
Table II.
Precision of Polarographie Procedure
C opper
S cries“ A verage D e v iatio n &
%
%
1
O.OOI81
0 .00006
2
O.OOI81
0.00003
3
0.0018s
0.00011
4
O.OOlOt
0.00003
5
0 .0 0 1 0 2
0.00003
L ead
A verage Deviation*»
%
%
0.00579
0.00008
0 .0 0 5 8 t
0.00003
0.0056o
0 .00008
0.0017!
0.00004
O.OOI67
0.00005
C adm ium
A verage D eviation^
%
%
0 . 0005«
0 .00004
O.OOO62
0 .00005
O.OOO62
0 .00005
0.00002
0.0009s
0.0009a
0.00003
° E a c h of series 1, 2, and 3, is resu lt of seven d e te rm in a tio n s. In series
3, each d e te rm in a tio n w as m ade in du p lic a te. Series 4 is re su lt of d u p lic a te
m easurem ents by nine d ifferen t individuals. Series 5 is re s u lt of te n m easure­
m ents of sam e curve.
&A verage dev iatio n from a rith m e tic a l m ean.
precision would indicate th a t it is possible to determ ine copper,
lead, and cadm ium to w ithin 1 X 10-< per cent for th e range of
concentrations encountered in high-purity alloys of th e M azak
type.
A t th e tim e this investigation was carried out, it was im pos­
sible to procure a N ational B ureau of S tandards zinc die-casting
alloy of th e ty p e desired (M azak 3). In lieu of this, an analysis
was m ade on an alloy high in copper. Analyses are also pre­
sented for several stan d ard zinc spelters (Table II I).
I t has been found, as a result of th e analysis of a large num ber
of commercially analyzed zinc sam ples, th a t the polarographic
results are usually high. N o explanation is available for this
phenom enon. T he calibrations have been checked an d re­
checked repeatedly in th e a u th o rs’ labo rato ry by different m eth­
ods. T here is a possibility th a t because of th e sm all am ount of
handling, the polarographic results represent a closer approxim a­
tion to th e tru e value th a n analyses which are the result of m any
successive m anipulations. T he degree of precision of th e polaro­
graphic procedure is high, as illustrated, and th e deviations from
o th er analyses do n o t show signs of a constant error. T his is
evidenced by exam ination of th e above results for th e B ureau of
S tan d ard s sam ples.
Table III.
A ccu ra cy of M ethod
(S arg en t in stru m e n t)
C o p p er
L ead
%
N a tio n a l B u re au of S ta n d a rd s.
E x p erim en tal
Precision
C ertificate v a lu e
2 .8 3 “
± 0 .0 1
2 .8 2
N a tio n a l B u re au of S ta n d a rd s.
E x p e rim en ta l
P recision
C ertificate v a lu e
0 .0007
± 0 .0 0 0 1
0 .0 0 0 5
N a tio n a l B u re au of S ta n d a rd s.
E x p e rim en ta l
Precision
C ertificate valu e
0 .0 0 0 4
± 0 .0 0 0 1
0.0004
%
C adm ium
%
Sam ple 94
0 .0318
± 0 .0 0 0 2
0.031
0.0025-
=*=0.0001
0.0 0 4
S am ple 109
0.0025
± 0 .0 0 0 1
0.0020
0.001 9
*=0.0001
0.001 8
S am ple 108
0.0505
± 0 .0 0 0 2
0 .0 4 7
0.096 0
*=0.0007
0.0 9 2
“ B y electrodeposition.
INTERFERING ELEMENTS
T here are some nineteen elem ents w hich m ay be found in zinc,
eith er as im purities, or as alloying elem ents: Cu, P b, Cd, N i, Co,
M n, Ag, As, H g, T l, Bi, Sb, Al, Fe, Ge, Ga, In , Mg, and Sn. (“ In ­
terference” is intended to describe th e prelim inary or alm ost co­
incidental discharge of some undesired ion which either results
in a m asking of th e step desired or m akes impossible th e record­
ing of the desired ion a t m axim um sensitivity.)
E xperim ents w ith 0.05 p er cent each of nickel, cobalt, m anga­
nese, silver, arsenic, m ercury, and indium show th a t th ere is no
detectable effect on the steps for copper, lead, and cadm ium
within ±0.0001 p er cent. M agnesium was tried up to 0.5 p e r
cent an d no interference was detected. T his is to be expected,
since th e discharge p o ten tial of m agnesium is well above th a t of
74
INDUSTRIAL
AND
ENGINEERING
zinc. Copper up to 0.1 per cent has been determ ined polarographically w ithout reducing th e sam ple size while retaining a
high degree of precision for lead an d cadm ium . R esults of th e
electrodeposition of copper in conjunction w ith th e polarographic
determ ination of lead and cadm ium indicate th a t there is no loss
of lead an d cadmium during this operation an d th a t less th an
0.01 per cent of th e copper rem ains after deposition. B ism uth
was found to give a step which precedes th a t of copper and for
this reason interferes w ith th e ensuing determ ination of copper,
lead, and cadmium when present in excessive am ounts. I t is
not close enough, however, to m ask th e copper step. T he factor
for bism uth was found to be approxim ately 0.024 per cent per A.
for an 8-gram zinc sam ple in 50 ml. A ntim ony a t 0.05 p er cent
gives a poorly defined step w hich interferes w ith th e steps for
copper, lead, and cadm ium ; however, a t 0.01 p e r cent and less,
no interference w as found. T h u s antim ony concentrations of
0.01 per cent can be tolerated. T h e n atu re of th e interference
would seem to indicate a sm all diffusion coefficient for Sb+++++
in this particular m edium . T hallium gives a well-defined wave
which comes between lead and cadm ium in 2.5 zinc chloride.
F or trace am ounts of copper, lead, an d cadmium, only 0.002
per cent of thallium can be present. T his corresponds to th e
findings of Seith and Esche (9) w ith regard to th e lim it of detec­
tion of thallium in zinc. G erm anium was n o t tested for in te r­
ference, because of th e extrem e volatility of its chloride. Gallium
was not tested because of th e difficulty in obtaining a salt of th is
m etal. N o interference is to be expected from gallium because
of its high discharge potential.
Stannic tin when present in am ounts greater th a n 0.0015 per
cen t will rive a m easurable increase in th e step height for lead.
Occasionally, am ounts greater th a n th is m ay be tolerated due to
volatilization of stannic chloride, b u t such am ounts of tin are n o t
volatilized appreciably by this p articu lar procedure. Kalovsek
(5) indicates th a t the electroreauction of stannic tin is n o t re­
versible except in hydrochloric acid solutions of high concentra­
tion (above 0.1 A7), where, however, th e reduction process a p ­
pears inhibited. T his would explain w hy such a large am ount of
stannic tin would cause no interference. These results confirm
those of Seith and Esche for th e lim it of detection of tin in zinc.
Aluminum up to 6 per cent has been found to be w ith o u t d etec t­
able effect on the height of th e steps for copper, lead, and cad­
mium. H igher concentrations of alum inum m ight have an effect
only in so far as they affected th e concentration of the supporting
electrolyte. Iron, a fter reduction w ith hydroxylam ine hydro­
chloride is w ithout significant effect in th e determ ination of
copper, lead, and cadm ium a t 2.5 per cent. I t is necessary, of
course, to ad ju st the am ount of hydroxylam ine hydrochloride
used in accordance w ith the iron content for satisfactory results.
CHEMISTRY
Vol. 16, No. 1
A stu d y of the polarographic determ ination of trace quantities
of tin , alum inum , and m agnesium in high-purity zinc alloys is in
progress.
CONCLUSION
T race am ounts of copper, lead, and cadm ium can be rapidly de­
term ined by th e polarographic m ethod in high-purity zinc and
zinc die-casting alloys w ith a high degree of precision. Such a
m ethod should prove of value in industrial laboratories where tim e
is a t a prem ium .
In an effort to m ake possible a complete polarographic analysis
of high-purity zinc and zinc die-casting alloys, procedures are
being developed for th e determ ination of other elem ents present.
ACKNOWLEDGMENTS
T h e authors wish to th a n k th e D ep artm en t of Physics, U ni­
v ersity of T oronto, for th e use of its Leeds & N o rth ru p E lectroChem ograph. T hanks are due also to D . H . Sim pson for his
valuable assistance in th e prelim inary investigations, and to F. E.
Beamish, U niversity of T oronto, who kindly provided a sam ple
of indium chloride.
LITERATURE CITED
(1) Borcherdt, G . T., M eloche, V. W ., and A dkins, H., J . A m .
Chem. Soc., 59, 2171 (1937).
(2) Bray, W . C., Sim pson, M. E ., and M acK enzie, A. A., Ibid., 41,
1366 (1919).
(3) E nsslin, F ., M elall u. E rz, 37, 1 71-2 (1940).
(4) H eyrovskjb J., Chim ie an d In du strie, Special N o ., 1933, 20 4 -1 0 .
(5) H ohn, H ., Ch.em.-Ztg., 62, 7 7 -8 1 (1938).
(6) K alovsek , M., Collection Czechoslov. Chem. Com mun., 11, 5 9 3 613 (1937).
(7) K olthoff, I. M., and Lingane, J. J., "Polarography” , p. 243,
N ew Y ork, Interscienco P ublishers, 1941.
(8) K rossin, E ., MetaU u. E rz., 38, 1 0-12 (1941).
(9) S eith, W ., and E sche, W ., Z. M elallkunde, 33, 8 1 -3 (1941).
(10) Strubl, R ., Collection Czechoslov. Chem. Commun., 10, 466
(1938).
(11) Terui, Y ., B u ll. In st. P h ys. Chem. Res. ( Tokyo), 17, 6 4 4 -8
(1938).
r e s e n t e d before th e D ivision of A n a ly tic al and M loro C h em istry a t th e
105th M eeting of th e A m e r i c a n C h e m i c a l S o c i e t y , D e tro it, M ich.
P
Determination of Pectin in Biological Materials
M odification of Pentose-Furfural M ethod
E D W IN F. B R Y A N T , G R A N T H . P A L M E R , A N D G L E N N H . J O S E P H
California Fruit Grower» Exchange, Research Department, Corona, Calif.
Pectin is converted to a pentose which produces the furfural
with which this method is concerned. Data are presented
to show the normal levels of furfural-yielding substances in
various organs and fluids from rabbits. The analytical pro­
cedures described make it possible to recover 95 to 100
per cent of pectin which has been added to animal tissues
and fluids.
N T w artim e conditions, w hich have increased in ter­
P RestE SinEpectin
sols for intravenous use in treatin g shock, have
m ade it necessary to devise a sem im icrom ethod for th e d eter­
m ination of pectin in biological m aterials.
Pectinum N . F. V II which is suitable for intravenous use is es­
sentially a pure polygalacturonic acid ester. W hen such a pectin
is refluxed a t elevated tem p eratu res w ith 12.5 per cent hydro­
chloric acid the polygalacturonic anhydride u n its are decarboxylated, producing a mole of carbon dioxide for each carboxyl, and
forming furfural from th e newly formed pentose. A ccurate
qu an titativ e procedures based upon th e determ ination of th e
carbon dioxide evolved an d upon the furfural produced have been
developed during th e forty years w hich have elapsed since the
general reactions were first described by Tollens (5).
T he m ethods which have been developed for q u an titativ e esti­
m ations of furfural are sensitive to extrem ely sm all am ounts and
adaptable to colorim etric procedures. T h e carbon dioxide m eth­
ods are useful only when relatively large am ounts of m aterial are
available; hence for purposes of determ ining pectin in biological
system s th e furfural scheme is preferred.
F u rfu ral m ethods and th eir applications to pectin analyses
were discussed by Browne and Zerban in 1941 (2). Y oungburg
(7) in 1927 described a particu lar ad ap ta tio n of furfural estim a­
tion useful for biological m aterials. B ry an t, Palm er, an d Joseph
(4) used a modification of Y oungburg’s m ethod in these labora­
tories early in 1941 for an exam ination of th e liver and other
organs of th e rab b it. T h e Y oungburg scheme involved steam
distillation from 85 per cent phosphoric acid an d colorim etric
determ ination of furfural in the distillate b y th e furfural-aniline
acetate reaction. D etails for the use of the Y oungburg m ethod,
January 15, 1944
ANALYTIC
L E DI TIO N
75
sta n d a rd (1.00 ml. equivalent to 0.01 mg. of furfural) should be
made fresh the day of use.
Freshly distilled aniline, free from furfural. T e st b y adding
0.25 ml. to 2.0 ml. of glacial acetic acid. If a pink color develops
w ithin one m inute, furfural is present.
G lacial acetic acid, 85 per cent orthophosphoric acid, 10 per
cent by w eight solution of sodium tu n g sta te (NasW OiYHjO),
0.5Ar sulfuric acid, and isopropyl or eth y l alcohol, a t least 95 per
cent.
PREPARATION OF SAM PLES
M e th o d .
T he u tility of th is
m ethod in determ ining th e fate of injected pectin depends upon
having a n accu rate figure for th e furfural equivalent p er m illiliter
o f the p ectin sol being used in th e anim al experim ents. Each
lo t of Exchange P ectinum N . F . V II will have a definite value for
the furfural equivalent, usually varying from 190 to 215 mg. of
furfural p er gram of pectin. These furfural values, obtained by
the m ethod described below, have been checked by using th e
sta n d a rd gravim etric phloroglucin m ethod (6). T he values by
the two m ethods agree alm ost perfectly.
Animal work w ith pectin usually involves 1.0 to 2.0 per cent
sols w hich have been sterilized by autoclaving and filtered to
sparkling brilliance. S tandardization analyses on such sols
should be run on sam ples prepared as follows: 5.0 ml. of 1.0 to
2.0 per c e n t pectin sol, diluted to 200.0 ml. w ith distilled w ater;
2.0 ml. of this d ilu ted sol should be used for each analysis.
U rin e .
N orm al urine from rab b its usually contains so little
furfural-yielding m aterial th a t 4.0-ml. sam ples are required for
a n analysis. W hen pectin has been injected into an anim al, i t
is necessary to use only a 2.0-ml. sam ple of urine to g et good re­
sults, while pectin is being excreted.
P ip et duplicate sam ples of urine into 25 X 150 ml. Pyrex te s t
tubes an d add 10 volumes of a t least 95 per cent alcohol. A fter
stan d in g for an hour or two centrifuge the tubes an d pour off the
su p e rn ata n t liquor. Place th e tubes containing the residue in a
drying oven a t 100° C. an d d ry for ab o u t 30 m inutes, or until no
odor of alcohol can be detected. T hese tubes containing th e
dried residue m ay th en be covered an d stored until ready for
analysis.
U rin e a n d A d d ed P e c tin , a s C o n t r o l o n M e th o d .
P rep are
a check solution by diluting 5.0 ml. of a 1.0 to 2.0 per cent pectin
solution w ith urine u p to 200.0 ml. in a volum etric flask, and
tre a t 2.0-ml. sam ples w ith alcohol as described above.
B lo o d .
P repare a Folin-W u blood filtrate b y th e following
m ethod: W eigh 2.0 ml. of oxalated blood (10 mg. of potassium
oxalate per 5 ml. of blood) accurately in a 30-ml. centrifuge tube.
Add to this b y p ip et 13.0 ml. of distilled w ater, 2.0 ml. of 10
per cen t sodium tu n g state solution, and 3.0 ml. of 0.5IV sulfuric
acid. M ix th e contents of th e tubes well, allow the samples to stand
15 m inutes, then centrifuge a t ab o u t 1000 r.p.m . for 30 m inutes.
Rem ove th e clear centrifugate and save for analyses, using 2.0 ml.
for each sample, as described below.
A n i m a l T i s s u e s . A nalyses m ay be m ade upon th e organs
im m ediately a fter th e y have been rem oved from the anim al and
weighed, or th e m aterial m ay be frozen w ith d ry ice and stored
for la te r use. In either case it is necessary to m acerate th e organ
(or a portion of it in cases of large organs) in a m ortar. T ransfer
as m uch as possible from th e m o rtar (n o t to exceed 10 gram s of
the larger organs), into a tared beaker and th en ad d a b o u t 10
tim es as m uch distilled w ater as th e w eight of th e organ used.
S tir th is weighed m ixture thoroughly an d transfer to a mixer
such as the W aring B lendor where it is reduced to a homogeneous
slurry. P ip et a volum e of this slurry equivalent to a b o u t 0.3
to 0.5 gram of th e original organ into a 30-ml. centrifuge tube,
and weigh. A dd distilled w ater to bring th e volume to 15.0
ml., then add 2.0 ml. of 10 p er cent sodium tu n g state solution
and 3.0 ml. of 0.51V sulfuric acid, mixing th e contents of th e tube
by careful swirling after th e addition of each reagent. M ix th e con­
tents of th e tubes well and allow the sam ples to sta n d 15 m inutes.
T hen centrifuge a t ab o u t 1000 r.p.m . for 30 m inutes. Remove
the clear centrifugate an d save for analyses, using 2.0 ml. for each
distillation.
P e c tin S o ls f o r C h e c k o n
particularly th e step involving trichloroacetic acid tre a tm e n t of
sam ples, were given by Andersch an d Gibson (1).
M odifications of the Y oungburg m ethod previously published
failed to allow reasonable recovery of pectin w hich had been added
to urine, and also failed to provide sam ples from blood an d tissue
m aterials w hich could be distilled w ith o u t excessive foaming.
T h e m ethod described here provides for an alcohol tre a tm en t of
urine sam ples w hich perm its complete recovery of pectin added
to urine. I t su b stitu tes sodium tu n g sta te for trichloroacetic
acid an d centrifuging for filtration, in preparing o th er sam ples,
and continues w ith th e regular Y oungburg distillation. F oam ­
ing is elim inated and recoveries from control sam ples are excellent.
T h e technique for the determ ination of furfural in th e distillate by
a photoelectric colorim eter is described. T h e procedures given
below are the result of several hundred analyses of ra b b it blood,
urine, an d organs, and of m any m ixtures of pectin w ith these ani­
m al m aterials.
APPARATUS
T he distillation apparatus shown in Figure 1 is composed of
a steam generator, A , m ade from a large eth er can w ith burner
underneath regulated by screw clam p for control of steam flow
into B , the distillation unit, a Pyrex 25 X 150 mm . te s t tube, w ith
therm om eter covering th e range 0° to 200° C .: C, tin cylinder
shield for m icrobum er; and D, a w ater-cooled condenser, 250mm . jacket, w ith inside tube preferably 6 to 7 mm . in diam eter
and tu rn ed down a t end for delivery into E . a 50-ml. g raduated
cylinder. An ordinary steam generator m ade from glass m ay be
used instead of th e m etal one described.
A num ber of the Pyrex 25 X 150 mm. te s t tubes should be
available because urine sam ples are prepared in them and then
m ay be stored until tim e is available for distillation. I t is con­
venient to have 2.0-, 3.0-, 4.0-, and 5.0-ml. volum etric pipets, as
well as M ohr pipets, 1.0-ml. size graduated in 0.10 divisions. A
centrifuge a t least as large as th e In tern atio n al N o. 1 and several
30-ml. centrifuge tubes are required. A photoelectric colorim­
eter is recom mended. T he F isher Electropnotom eter w ith a blue
filter (N o. 425B) and th e K lett-Sum m erson photoelectric colorim­
eter w ith a green filter (No. 54) have been used successfully w ith
this m ethod.
REAGENTS
METHOD
Freshly - prepared furfural, vacuum -distilled a t a pressure of
20 to 30 mm.
Standard furfural stock solution prepared from th e freshly dis­
tilled furfural; 1.000 gram is diluted to 500 ml. w ith distilled
w ater. T his stock solution m ay be k ep t for several weeks in a
refrigerator.
S tandard dilute solution of furfural for calibration of photo­
electric colorim eter or as a com parison solution w ith th e D uboscq-type colorim eter. T his solution is m ade by diluting 5.0
ml. of th e stock to 1000 ml. w ith distilled w ater. T his dilute
T he sample for analysis should be in a 25 X 150 mm . Pyrex te s t
tube, prepared as discussed under “ P reparation of Sam ples” .
Add to th e sam ple in the te s t tu b e 5.0 ml. of 85 per cen t phos­
phoric acid and connect w ith th e condenser and th e steam gen­
erator. L ight the m icroburner and raise th e tem perature rapidly
to 170° to 175° C. T he tem perature should be m aintained a t
this po in t and never allowed to go higher during the 20- to 30m inute distillation period.
Collect 40 ml. of distillate in th e 50-ml. g raduated cylinder used
as a receiver, and an additional 10 ml. of distillate in a te s t tu b e
INDUSTRIAL
76
AND
ENGINEERING
previously m arked a t the 10-mi. level. T e st this last distillate
for th e presence of furfural as follows: Mix 2.0 ml. of th e distil­
late, 0.25 ml. of aniline, and 2.0 ml. of glacial acetic acid in a test
tube. If no color develops in one m inute, stop the distillation and
discard th e final 10 ml. of distillate. If color does ap p ear in the
te s t sam ple, distill an additional 10 ml. and te s t a portion for fur­
fural as above described. W hen analyzing blood or urine of un­
tre a te d anim als, 30 ml. of d istillate are usually sufficient to con­
ta in all th e furfural.
Table I.
Recovery of Pectin A d d e d to Rabbit Urine
M ate rial A nalyzed
2.0 ml. of 1.75% p ectin sol B-9221
48.0 m l. of u rin e + 2.0 ml. of p ectin sol B-9221
48.0 m l. of u rin e 4- 2.0 ml. of distilled w a ter
D ifference, d u e to a d d ed pectin
R ecovery of ad d ed p ectin , %
Tabic II.
F u rfu ra l F o u n d
Y oun g b u rg P rese n t
m eth o d
m eth od
M g.
Mg.
7 .0 8
11 .4 6
7 .3 3
4 .1 3
5 8 .3
7 .0 8
9 .2 2
2 .0 8
7 .1 4
100.9
Furfural Equivalent of 1 .0 0 Gram of Pectinum N . F. V II
(S am ple 444-H -3.
CHEMISTRY
Vol. 16, No. 1
I t is desirable to know th e actual pectin concentration of th e pec­
tin sol used or else th e furfural equivalent per gram of th e pectin
used to m ake th e sol. T he typical examples of these d a ta shown
in T ables I I and I I I illustrate th e ranges for these values as well
as show th e precision and accuracy of the m ethod.
T h e d a ta in T able I I I show th a t 100 ml. of solution B-9221
would be equivalent to 355 mg. of furfural; hence th e solution
m ust be 355/201 or 1.76 p er cent pectin. P ectin sol B-9221, when
analyzed by a modified Lefevre-Tollens m ethod (described by
B ry an t and Joseph, 3), showed 0.2315 gram of galacturonic acid
p er 15.0-ml. sample. Pectin 444-H-3 had previously been found
by th e sam e m ethod to contain 87.6 per cent of galacturonic acid;
hence th e 15.0-ml. sample of B-9221 contained 100 (0.2315)/15.0
X 0.876 o r 1.76 p er cent of pectin.
T he exten t to which furfural-producing substances occur in
ra b b it organs is indicated in T able IV, along w ith an alytical
values showing how th e present m ethod elim inates interference
from th e various anim al m aterials an d perm its practically com­
plete recovery of any added pectin.
A nalyses m ade on w a ter solu tio n c o n tain in g 750 m g. of
p e ctin p e r liter)
O p erato r
F u rfu ra l F o u n d
M g J g ra m
444-H -3
201
E .F .B .
G .H .P .
G .H .P .
G .H .P .
A verage v a lu e used
198
205
200
201
Combine and mix 20.0 ml. of th e original 40-ml. distillate and
5.0 ml. of each 10-ml. distillate showing a positive te st for fur­
fural, Place 5.0 ml. of this m ixture (or 5.0 ml. of the original
40-ml. distillate when the furfural test in th e next 10 ml. was
negative) in a tube and mix w ith 0.5 ml. of aniline an d 4.5 ml. of
glacial acetic acid, carefully m easured w ith a pipet.
Set this m ixture aside in the d ark a t 20° to 25° C. for exactly
15 m inutes, a t which tim e the color in ten sity is determ ined w ith a
photoelectric colorimeter. T h is 15-minute period has been deter­
mined experim entally as giving th e m ost exact value. T he col­
orim eter should be calibrated by using th e stan d ard dilute fur­
fural solution described above, spacing sam ples over the range
from 0.001 to 0.050 mg. of furfural per 10 ml. of solution contain­
ing 0.5 ml. of aniline, 4.5 ml. of glacial acetic acid, and th e diluted
stan d ard furfural solution.
W hen using the D uboscq type of in stru m en t a stan d ard for
colorim etric comparison m u st be prepared for eacli sample. T his
is done by mixing 2.0 ml. of th e dilute stan d ard furfural solution
m ade tiia t day, w ith 3.0 ml. of distilled w ater, 0.5 ml. of aniline,
a n d 4.5 ml. of glacial acetic acid. T h e tube containing this mix­
tu re and also th e one w ith the unknow n are se t aside in th e dark
for exactly 15 m inutes, for color developm ent. T he unknown
a n d the stan d ard should have a b o u t th e sam e color intensity.
If they do not, the solutions should be rem ade, reducing the
volum e of either the unknow n or stan d ard , adjusting th e to tal
volum es w ith distilled w ater to keep them th e sam e as outlined
above.
DISCUSSION
A n illustration of the accuracy of th e present m ethod com­
p ared w ith th a t of the original Y oungburg scheme, in the case of
ra b b it urine, is given by T able I.
I t was found th a t urea added to pectin sols decreased th e re­
covery of pectin by th e Y oungburg procedure. M any analyses,
a s illustrated by T able I, show th a t even though th e alcohol pre­
c ip itatio n scheme does rem ove some of th e natu rally occurring
aldehyde-producing substances of urine, it also removes th e urea
a n d thereby allows complete recovery of th e added pectin. The
increase in furfural due to added pectin in urine (hum an as well
as rab b it), using th e alcohol precipitation m ethod, has alw ays
been found equivalent to th e furfural obtained from th e pectin
w hen analyzed alone.
W hen th is m ethod is used to follow pectin excretion from ani­
m als to which pectin sol transfusions have been given, i t is neces­
sa ry to know th e furfural equivalent for th e pectin sol used and
th e norm al furfural values for th e organs or fluids being examined.
Table 111.
Furfural Equivalent of 1 .0 0 M l. of A u to clav e d 1.75 Per
Cent Isotonic Pectin Sol
(Sol B-9221 m ade from P e c tin 444-H -3. A nalyses m ade on d ilu tio n of 5.0
m l. of B-9221 to 200 m l. to ta l)
O p erato r
D a te
G .H .J.
G .H .P .
E .F .B .
G .H .J.
E .F .B .
G .H .J.
E .F .B .
A verage value
Table IV .
F u rfu ra l
M g ./m l. B-9221
Sept. 11, 1942
Sept. 14, 1942
O ct. 5, 1942
O ct. 7, 1942
M ay 27, 1943
Ju ly 21, 1943
J u ly 21, 1943
3 .5 9
3 .4 6
3 .5 4
3 .5 7
3 .5 8
3 .5 4
3 .5 6
3 .5 5
Furfural O b tained from Rabbit Tissues and Recovery of
Pectin A d d e d to These Tissues
(1.00 ml. of p e ctin sol B-9221 a dded is eq u iv a len t to 3.55 m g. of fu rfu ra l,
T a b le II I)
M a te ria l
Blood
B ile
B one
H e a rt
K idney
L iv e r“
Spleen
Sam ple
W eight
Grams
1.903
1.539
0 .3 2 0
0 .3 2 0
0 .5 0 0
0 .5 0 0
0 .3 3 0
0 .3 3 0
0.5 3 5
0 .5 3 5
0.4 9 2
0 .4 9 6
0.3 2 0
0 .3 2 0
F u rfu ra l O btained
T issue plus
1.0 m l. of
p e ctin sol
B-9221
M g.
O riginal
tissue
sam ple
M g.
F u rfu ra l
D u e to
A dded
P e c tin
M g.
R ecovery
of A dded
P e c tin
%
101.7
0 .2 3
3 .8 0
3 .6 1
3 .7 1
3 .4 2
9 6 .3
3 .6 2
3 .5 2
9 9 .2
3 .6 2
3 .4 8
9 8 .0
3 .7 4
3 .5 2
9 9 .2
5 .6 0
3 .6 5
102.8
3 .6 2
3 .4 8
9 8 .0
0 .2 9
0 .1 0
0 .1 4
0 .2 2
1 .9 5
0 .1 4
° B e tte r resu lts a re ob tain ed w hen sam ple w eight of liv e r is a b o u t 0.2 to
0.3 gram .
LITERATURE CITED
(1) Andersch, M arie, and G ibson, R . B ., J . P harm acol., 52, 390-407
(1934).
(2) B row ne, C. A ., and Zerban, F . W ., “P hysical and C hem ical M eth ­
ods of Sugar A n alysis” , 3rd od., pp. 914-18, N ew Y ork, John
W iley & Sons, 1941.
(3) B ryant, E . F ., and Joseph, G . H ., paper presented before the
D ivision of Agricultural and F ood C hem istry at th e D alla s
M eetin g of the A m erican C hemical S ociety , 1938.
(4) B ryant, E . F ., Palm er, G. H ., and Joseph, G. H ., P roc. Soc.
E x p tl. B iol. M ed., 49, 2 7 9 -8 2 (1942).
(5) L efevre, K . U., and T ollens, B ., B cr., 40, 4 5 1 3 -2 3 (1907).
(6) Norris, F . W ., and R esell, C . E ., Biochem. J ., 29, 1590-6 (1935);
also A ngell, S tan ley, Norris, F . W ., and R esoh, C ., Ib id . 30,
2 1 4 6 -5 4 (1936).
(7) Y oungburg, G . E „ J . B iol. Chem., 73, 5 99-606 (1927).
P h e s e n te d
of th e
before th e D ivision of B iological C h em istry a t th e 105th M eeting
D e tro it, M ich.
A m e k ic a n C h e m ic a l S o c ie ty ,
January 15, 1944
ANALYTICAL
15
EDITION
M ETALLURGICAL PRODUCTS A N A LYZED
SPEEDILY A N D ACCURATELY
Slomin High Sp e e d Electrolytic A n a lyze rs
T h e rap id a cc ep ta n ce o f this in str u m e n t fo r
m e ta llu r g ic a l an alysis is o u tsta n d in g en d orsem en t o f its p ro v en a b ility and con sisten t re lia b ility . O v er 80 0 Slom in A n a ly zer s are n o w in
u se in m e ta llu r g ic a l lab oratories.
E lectrod e d esign , c u r r e n t efficien cy and im p ro v ed p roced u res re d u ce d ep osition tim e fo r m e r ly req u ired b y oth er system s as m u ch as 25
to 4 0 % . U n d er th ese h ig h speed co n d ition s h a rd ,
sm o o th , b r ig h t and clo sely g rain ed deposits th a t
firm ly adhere to th e electrod es are p ro d u ced ,
th u s a ssu rin g good r e p r o d u c ib ility o f re su lts.
U sers rep ort an a c c u r a c y o f 0.01 to 0.04% fo r
r o u tin e d eterm in a tio n s.
E ach m od el is p o rta b le and en closed in a -welded steel case finished in acid resista n t baked
w h ite en am el. T h e b rushless m otor is v a p o r tig h t
and is th e re fo r e u n a ffe cted by co rro siv e fu m es,
B o th m odels h a v e an e le c tr ic a lly h e a te d , rh eos ta t co n tro lled b eak er p la tfo r m fo r a d ju stin g
so lu tio n tem p era tu res, an d v o ltm e te r s an d a m m eters so th a t d eta iled stu d ies ca n be m ade,
E ach p o sitio n o f th e t w o p la ce a n a ly ze r is a
co m p lete c ir c u it th a t op erates in d e p e n d en tly o f
th e o th er. C o n seq u en tly th is u n it ca n be u sed
fo r th e sim u lta n eo u s d eterm in a tio n o f tw o sam p le s h a v in g w id e ly d iv e r g e n t ch a ra c teristics,
N o a ccesso ry g en era to rs o r rh eo sta ts are
req u ired — a ll m odels are se lf-c o n ta in e d ,
p o rta b le and op erate fro m sta n d a rd e lec­
tr ic c ir c u its.
A la b o ra to ry m a n u a l o f h ig h speed e lec­
t r o ly tic m ethod s o f an alysis w r itte n b y
G . W . Slom in is su pp lied w it h each a n a ly ­
zer. In d iv id u a l copies are a v a ila b le at
.5 0 each.
• S -2 9 4 6 0 S lom in E le c tr o ly t ic A n a ­
ly zer. O n e p o s itio n , 5 A m p e r e M o d el,
w ith H e a tin g P la t e . For op eration from
115 v o lt, 60 cy cle circu its. Each $ 1 7 0 .0 0
• S -2 9 4 6 2 D itlo . B u t for op eration from
23 0 v o lts, 60 cy cle circu its.E a ch $ 1 7 5 .0 0
• S -2 9 4 6 5 S lom in E le c tr o ly t ic A n a ­
ly zer, T w o p o s itio n s , 5 A m p e r e M o d el,
w ith H e a tin g P la t e . For op eration from
115 v o lt, 60 c y c le circu its. Each $ 3 1 0 .0 0
• S -2 9 4 6 7 D itto . B u t fo r op eration from
23 0 v o lts, 60 cy c le circu its. Each $ 3 2 0 .0 0
High Speed Electrodes for Use
with Slomin Electrolytic
Analyzers
• S -2 9 6 3 2 C o r r u g a t e d P latin u m A n o d e
(P a t. p e n d in g ). P rice su b ject to m a rk et.
• S-29& 72 C o r r u g a t e d P latin u m C a t h ­
o d e (P a t. p e n d in g ). P rice su b ject to
m ark et.
Lite ra tu re on Request
E. H. SARGENT & CO., 155-165 E. Superior St., Chicago 11, III.
M ichigan D ivision:
1959 East Jefferson, Detroit 7, Michigan
16
INDUSTRIAL
AND
ENGINEERING
CHEMISTRY
Vol. 16, No. 1
5 7 In d ic a t o r s fo r D e t e r m in in g
H y d r o g e n - I onC o n c e n t r a t i o n
E
astm an O rganic Chem icals include 57 indicators covering the entire
pH range. Careful laboratory tests assure the dependability and
uniformity of these compounds. In addition, each indicator of the
sulfonphthalein type is tested spectrophotometrically to make certain
that it exhibits maximum color change and sensitivity.
A table giving the solvents employed and the concentrations in which
these Eastman indicators are generally used, as well as a chart showing
their pH ranges and color changes, will be supplied free upon request. . ..
Eastman Kodak Company, C h e m ic a l S a le s D iv is io n , Rochester, N. Y .
ER
The r e ar e mor e fhan 3 4 0 0 '
CO DAK.
M odem E F F IC IE N C Y
COUPLED
W ITH
50 Y E A R S O F Q U A L I T Y
E DIAL OF A
ß h ja b sw jd siA .
MOISTURE
TESTER
The[M OISTUR£ TESTER does
away with the old, laborious
method of determinins mois­
ture content. O n e instrument
takes the place of the oven,
the balance and the desicca­
tor. Moisture content can be
determined w hile samples are
in the oven.
W h e re v e r m o istu re c o n te n t Is a a c ­
to r— in th e la b o r a to r y o r In th e p la n t
— th e r e th e M O IS T U R E TESTER w ill
m ake Its p la c e .
• TEN TESTS A T ONCE
• SENSITIVE TO .002s
• READINGS TO ,05%
W R IT S FOR F U L L IN FO RM ATIO N
SCHAAR
& COMPANY
Complete Laboratory Equipment
754 WEST LEXINGTON ST., CHICAGO
INDUSTRIAL
LA B O R A TO R Y TABLE 6610
For hall a century Peterson engineers Lave constantly
studied the requirem ents of laboratory workers an d the
various techniques w hich em anate from the classroom
on into commercial practice. Every bit of Peterson
laboratory furniture is designed with full cognizance of
m odem improvements necessary to the current technique
. . . plus the fact that it is backed by the reputation for
quality w hich stems back through half a century.
W RITE FOR PE T E R SO N 'S CATALOG, G RATIS OF COURSE
LEONARD PETERSON & CO., Inc.
1222-34 F U L L E R T O N
AVE.
C H IC A G O , U. S . A.
ANALYTICAL
January 15, 1944
17
EDITION
m
a
The mos t i mp o r t a n t
bot tl e of c h e m i c a l s
in the world. . Jo
You
It s the next bottle of reagent chemicals you
take from your laboratory shelf, for upon the
quality, purity and uniformity of the product
it contains may depend the outcome of your
careful analyses . .. perhaps the success or fail­
ure of vital research that “can’t wait!’
B &A recognizes your need for dependable
chemicals. That’s why all B&A Reagents are
produced only under the strictest manufacturing standards.
Into every B&A product—whether it is a
reagent you use every day or one you need
infrequently—goes the same careful attention
to minute details and constant laboratory con­
trol that assure you chemicals of unusually
high uniformity and purity.
When you reach for a bottle of B&A Reagent
Chemicals you reach for the best. Be sure to
specify Baker and Adamson Reagents!
S E T T I N G
THE
PACE
IN
T*
B a k e r
C H E M I C A L
PURITY
SINCE
1882
7Ï
A
d a m
s o n
D iv is io n o f G E N E R A L C H E M IC A L C O M P A N Y ,4 0 R e cto r S t . , N e w Y o r k 6 , N . Y .
C .'R A c û / s
T echn ical S e n k te O ffices: A t l a n t a •
B a ltim o re • B o s to n • B rid g e p o rt ( C o n n .) * B u ffa lo • C h a r lo tte ( N . C .)
C h ic a g o
•
C le v e la n d
•
D enver
•
D e tr o it
*
H o u s to n
•
K a n s a s C ity
•
M ilw a u k e e
•
M in n e a p o lis
N ew Y o rk • P h ila d e lp h ia • P i tts b u r g h • P ro v id e n c e ( R . I . ) • S t. L o u is • U tic a ( N . Y .)
P a cific C oast T ech n ical S e rv ic e O ffic e s : S a n F ra n c is c o • L o s Angeles
P a cific N o rth w est T ech n ical S ervice O ffic e s : W e n a tc h e e ( W a s h .) • Y a k im a ( W a s h .)
In C a n a d a : T h e N ic h o ls C h e m ic a l C o m p a n y , L im ite d • M o n tr e a l • T o r o n to • V a n c o u v e r
18
INDUSTRIAL
AND
Liters
Micron
Liters
Micron
CHEMISTRY
Vol. 16, No.
For its capacity—The Speediest
Pump at a n y
Pressure
PUMPING SPEED
14.9
1
24.2
1
ENGINEERING
Per Minute at
at 304 R.P.M.
Per Minute at
at 528 R.P.M.
W elch DUO-SEAL
VACUUM PU M P
Greater free-air capacity and higher
ultimate vacuum assure fastest pumping
at all pressures.
In addition to superior performance
it is
QUIET RUNNING
No drumming in the ears even where
banks of them are pumping.
G UARANTEED V A C U UM
.35 M icro n — .00005 m m H g
.
FREE A IR C A PA C ITY
3 3 .4 L iters P e r M in u te
O PTIM U M O PERATING SPEE D
3 00 R ev olutions P e r M in.
• OIL REQUIRED
6 5 0 m l D uo*Seal O il
• ELECTRONICS
No. 140S-H
• FREEZING
• DRYING
D U O -SEA L V A C U U M PUM P, M otor D riven.
V acu u m .05 m ic ro n — f re e a ir c a p a c ity of 33.4
lite rs p e r m in u te ................................................... $140.00
W ith la r g e r m otor g iv in g 57 lite rs fre e a ir c a p a c ity
p e r m in u te a n d v a c u u m of 0.1 m i c r o n .. . .$ 1 5 5 .0 0
T H IS BOOKLET
E X PL A IN S ALL. “
O rd er to d a y for p ro m p t d e liv e ry .
W. M. Welch ESstacb lish
i eedn1880
tific C o m pan y
1518 Sedgwick Street
Chicago, Illinois, U.S.A.
W rite for it to d a y . C on­
ta in s fu ll d e ta ils a n d d ia ­
g r a m s o n W e lc h P u m p s .
LaMOTTE OUTFIT FOR
DETERMINATION OF VITAMIN A
CHEM ICAL AND SCIENTIFIC PO RCELA IN W ARE
djaiic equipment
OF A L L M O D E R N
TESTING
LABORATORIES
E V A P O R A T IK C D IS U
A Goleiade Product
fijj ^ W e x ld m id e A a n te
The im portance oi grading fish oils and similar m aterials
according to the vitamin A content has dem anded a
simple means of determ ining vitamin A. This LaMotte
outfit employing the method of Dr. B. E. O ser provides a
reliable chem ical procedure to m eet this need. The
unit comes complete with color standard special com­
parator, reagents and full instructions. Price $35.00
f.o.b. Towson-4, Baltimore, Md.
L a M D T T E C H E M I C A L P M D J C T S CO.
D ept. F.
Tow son-4
B A L T IM O R E , M D .
» T a C oors P o r c e l a in C o m p a n y
B e W 8 ^
. GOLDEN, COLORA DO .
^*masaJM
January 15, 1944
ANALYTICAL
EDITION
19
It’s UfOt a Fancy Furnace —
But It’s Built to Do the Job
T
H IS H o s k in s Furnace, T y p e FD,
is an o ld stan d -b y to m o st la b o ra ­
to ries. F or a lm o st 3 0 years, it has
b een b u ild in g that rep u tation by its
g o o d p erform an ce. C h em ists lik e its
relatively c o o l surface tem perature.
T h e y lik e the g o o d -siz e d w o r k sh elf
and th e ea sily o p erated s lid in g d o o r.
A ls o , repairs are s o easy to m ake
b eca u se every th in g is s o easy to g et
at. F in ally, th e C h ro m el u n its are s o
d u ra b le that w h e n th ey at la st w ea r
o u t, it o n ly re m in d s you o f h o w very
lo n g they h ave lasted . O p erates on
45/s'' of Insulation
means efficient operation,
and a cooler laboratory
lin e v o lta g e. F ull sto ry in C atalog5 8 , fro m you r d ea le r o r ou rselves.
H o s k in s
M an u factu rin g
C om pany,
D e tr o it 8 , M ich ig a n .
Sturdy
Long-Lasting
Chromel Elements—
EASY TO APPLY
H O S K IN S PR O D U CTS
E L E C T R IC H EA T T R E A T IN G FU R N A C E S • • H E A T IN G ELEM ENT A L L O Y S • • TH ER M O C O U P LE A N D
LEAD WIRE • • PYROMETERS • • WELDING WIRE • • HEAT RESISTANT CASTINGS • • ENAMELING
FIXTURES
SPARK PLUG ELECTRODE WIRE • • -SPECIAL A LLO YS OF N ICKEL • • PROTECTION TUBES
INDUSTRIAL
20
AND
ENGINEERING
Vol. 16, No. 1
CHEMISTRY
NO W . . .
FASTER D ELIVERIES O N
K
l e
B u rre ll
T em p .
TYPE K F U R N A C E
t t .
.
.
.
P h o to m e te rs
K U tt-£ u * tu n e sU & ri
P h o to e le c tr ic
Q laM . C e ll
Golo>U*net*si
For the varied needs of industrial, clinical
and agricultural laboratories. Complete
and self contained — no accessory equip­
m en t necessary. Fused cells for readings
on solu tion depths of 2.S, 10, 20, and 40 m m .
Selective ligh t filters available.
fio *
up
ta m p a ta tW c a i
to
2550°F
This furnace accommodates laboratory crucibles
and dishes of various sizes and may be used for
heating articles which can be laid directly on
the hearth plate.
• 1 U K b it
fy lu o sU m e i& i
• Heating chamber measures 4" wide x 4"
high x 9” deep. Throat is 41/ 2" deep.
• Floor is a removable, refractory plate with
moulded sides to prevent spillage from
reaching isulation.
• Heating elements give long life and can
be replaced in a few minutes without
cooling furnace.
• Overall outside dimensions without table
are 26" wide x 30” high x 24" deep.
No. Î0 T 1
. Designed for the
rapid and accurate
d e t e r m i n a t i o n of
thiam in , riboflavin, and other substances which
fluoresce in solu tion. The sensitivity and sta ­
b ility are su ch th a t it has been found particu­
larly u seful in d eterm ining very sm all am oun ts
of these substances.
To further increase its adaptability a separate
circuit and scale is provided for colorim etric
determ inations.
For complete information on all standard models
of box, muffle, combustion-tube and pit types,
write for Burrell Catalog F-241.
LITERATURE SE N T U PO N REQUEST
B u r r ell
TECHNICAL SU PPLY C O M P A N Y
K
1 9 3 6 - 4 2 Fifth A v e n u e
179 E A S T 8 7 T H S T R E E T
Pittsburgh (1 9), Pa.
le
tt
M anufacturing Co.
NEW YORK, N.T.
ANALYTICAL
January 15, 1944
EDITION
M A K ER
Platinum Laboratory Ware
T e s t e d b y U se
\
Tests of any piece of apparatus under conditions that simulate those under which it
jT
\
will have to operate are valuable, hut notli\
ing can equal workaday use really to show
gA
the quality of a piece of equipment.
Baker Laboratory W are is under the conV tinual trial of test by use, because we em­
ploy it freely all the time in our own
\ laboratories. In this way we discovered S'
how the steins of stationary type electrodes
I
\tended to break at their juncture with the
i
cylinder and corrected the fault by reinfbrcing the joint there.
Among our other improvements are design >.
v
'y changes like our reinforced rim crucibles /
CSl
ar>d 'dishes, the Baker Low Form Crucible /
and Our platinum -rhodium alloy, now so,
widely, used for laboratory ware.
SPECIAL APPABATUS — When you are
■«B>y
in need of special apparatus, o u r plant is
the logical place to have it produced. We
are the largest refiners and worker^ of
s Mjjpfr
platinum , gold and silver in the world.; Our
experience covers a period of m ore than
seventy-five years and our research staff is
always ready to work with you.
BAK ER & CO
S M E L T E R S , R E F I N E R S A N D W O R K E R S O F P L A T IN U M . G O L D A ND S I L V E R
113 A s t o r S t ., N e w a r k . 5 , N . «V.
SA IV I 'l I A N C I S t 'O
21
22
INDUSTRIAL
AND
E N G I N E E R I N G ¡ " . C H E MI S T R Y
Vol. 16, No. 1
: liaae auailahle i&me
ICAL MICROSCOPES
N ew a n d R e c o n d itio n e d
H eating stages
P olarizers and analyzers
A lso
FLUORESCENT
U T I L I T Y LI GHT
This fluorescent Utility Light is an efficient, flexible, attrac­
tively designed lighting unit for laboratory or office use.
BIOLOGICAL and
M etallurgical M icroscopes
Photom icrographic Cameras
types o f m icroscope lenses,
new and used
The reflector housing is of Bakelite molded in one
piece—-overall length 18 inches. The ballast is
built into and concealed inside the housing.
A manually-operated switch with separate stopand-start buttons assures dependable operation
and fluorescent lamp efficiency. Light output is
boosted by an aluminized metal reflector inside
the housing.
A universal ball-joint on the mounting hub per­
mits adjustment of the lamp to any angle. The
unit is completely adjustable and, when separated
from the support base, may be used atop a
balance.
10303—F luorescent U tility L ight. As illus­
trated, with 8" diameter stand and 15" support rod
and clamp. For 110 volts, 60 cycle,
W e are experienced and successful in the design
and manufacture of high temperature furnaces particu­
larly for research laboratories requiring a range from
2 0 0 0 ° to 3 0 0 0 ° F. In this field we have served some
of A m e ric a ’s leading Chem ical and Industrial plants.
If your furnace requirements are unusual, our d e ­
signs as w ell as our quotations w ill interest you.
A d d re ss Harper.
H L -7 6 1 0 -F High
Temperature Electric
Laboratory Furnace
$15.75
Write fo r d e sc rip tiv e b u lletin
W ILL
CORPORATION
R O C H E S T E R , N. Y.
Office and Warehouses
W ill C o r p o r a t io n , 596 B r o a d w a y , N ew Y o r k C i t y
B u f f a lo A p p a r a t u s C o r p ., B u f f a lo , N . Y .
HARPER
ELECTRI C
FURNACE
ANALYTICAL
January 15, 1944
EDITION
23
''Spectrophotelom eter"
A typ ical assembly of the "Spectrophotelom eter“ and
accessories for work in the visible wave lengths.
fyecduneA. oJ, £tcceile*ice.
•
E A S E O F O P E R A T IO N
•
ACCURACY
AND
R E P R O D U C I­
B IL IT Y O F M E A S U R E M E N T S
•
F A C IL IT Y
B R A T IO N
•
9
9
9
FO R
C H E C K IN G
T h e Cenco-Sheard “Spectrophotelom eter” is unsurpassable for
accurate and economical spectral transm ission m easurem ents
in research or control. M ore th a n a hundred o f these in stru ­
m ents are being used daily in educational and industrial labo­
C A L I­
R A P I D IT Y IN M A K I N G R E A D I N G S
‘
ratories. F o r work in th e visible range o f 380 to 700 milli­
microns, th e 1 cm. cell u n it, com plete w ith light source and
all necessary accessories, m ay be purchased for $572.90. The
5 cm cell unit, covering th e sam e range, com plete w ith ac­
N E G L IG IB L E S T R A Y L IG H T EFFECT
cessories, sells for $577.90. I f th e range 325 to 750 millimicrons
is desired, th e addition of a special high am perage light source
A D A P T A B IL IT Y
FO R
V A R IO U S
S I Z E S A N D K IN D S O F A B S O R P ­
T IO N C E L L S
and suitable transform er, rheostats, and other accessories is
necessary. T he 1 cm u n it and accessories covering th is range
to ta l $720.25. T he 5 cm u n it and accessories for th is range
N O M I N A L B A N D W ID T H S F R O M
2l/2 T O 2 0 mu.
to ta l $725.25. T he I cm outfits are equipped to handle 1/ t inch
te st tubes. W rite for o th er detailed inform ation.
CENTRAL SCIENTIFIC COMPANY
S C IE N T IFIC
IN S T R U M E N T S
ffN (()
LABORATORY
A PPA R A TU S
RbSrtpLr otf.
NEW Y O R K
TO RO NTO
CHICAGO
BO STO N
S A N F R A N C IS C O
INDUSTRIAL
AND
ENGINEERING
CHEMISTRY
Vol. 16, No. 1
GLASS ABSORPTION CELLS
oj p,He Q uality,
■
■
■
made. Ly K . I c t t
Optical Flat
Wa l l s . M a n y
stock sizes are
available. Special
s i z e s m a d e to
order.
Fused in an elec­
tric furnace with
cem ent t hat is
acid, alkali and
solvent resistant.
Sole manufacturer in the United States
of fused Electrophoresis Cells
■
Makers of complete Electrophoresis Apparatus
Klett
Manufacturing Co.
177
ea st
87TH
str e e t,
new
York,
n.
y.
I t ’s VERSATILE!
IT’S NEW!
R E C O R D IN G
Viscosimeter
PERM AN EN T
I Records — available only
with the Recording Vis|cosimeter
] C H AN G ES
• Mills, washers and calenders in a
range of sizes for laboratory work
in natural and synthetic rubber and
plastics. Heavily constructed and
precision built. For full details write
NATIONAL RUBBER MACHINERY CO.
in viscosity can now be
] studied easily
j ADAPTABLE
I to many m a t e r i a l s . In
j many industries
V A R IA T IO N
in speed, paddles, springs
I easily accomplished
SCHAAR
For paints, pigments, varnishes,
printing inks, adhesives, oils,
greases, chocolate, ice cream,
starches, and other viscous ma­
terials, from approxim ately 1
poise up. Write for information.
H a v e y o u r e c e iv e d y o u r co p y
of C A T A L O G 43, P a r t O n e?
& COMPANY
Complete Laboratory Equipment
G e n e r a l O ffic e s : A k r o n , O hio
754
WEST LEXINGTON ST., CHICAGO
January 15, 1944
A N A L YTI C A L E D I TI O N
25
LU M ETR O N
N EW
American Chemical Society
Monographs
PHOTOELECTRIC COLORIMETER
VEGETABLE FATS
AND OILS
t
The Chem istry, Production and U tilizatio n of
V e g e ta b le Fats and O ils for E d ib le , M e d icin a l and
Technical Purposes
Secon d Edition
by George S. Jamieson
Revised and re -w ritte n to include n ew in fo rm atio n from
S o u th A m erica th is m o n o g rap h is packed w ith v ita l d etails
on th e sources o f these p roducts, th e ir ch aracteristics, com ­
p o sitio n , p ro p erties and uses; com m ercial and la b o ra to ry
processes fo r th e p rep aratio n and e x tra c tio n o f fats and oils
from oleag in o u s seeds, clarification, b leach in g , d eo d o rizatio n ,
h y d ro g en atio n , refining and o th e r treatm en ts in special cases
are discussed. G rad in g is described. M e th o d s are given for
sam p lin g , fo r ex am in atio n o f seeds, o ils, fats, press cake and
m eals as w ell as tests for e v a lu a tio n purposes and for the
detectio n o f a d u lte ra n ts . . . all ex h au stiv ely docum ented.
;o S Pages U -71
A highly sensitive instrument covering an un­
usually wide field of application.
•
•
•
•
A b rid g e d spectrophotometry with 14 monochromatic
filters, trichromatic filters.
G reat variety of sample holders, microcells, test tubes,
absorption cells up to 150 mm light path.
Line-operated, highest reproducibility and stability due
to use of balanced circuit, no batteries or voltage
stabilizers required.
A p p lic a b le (with accessories) to faint turbidities, fluores­
cence, ultraviolet absorption, reflection of opaque
liquids, powders, pastes, solids.
W rite for new 15-page bulletin on L U M E T R O N Mod. 402-E.
TUNGSTEN
PHOTOVOLT CORPORATION
Its H isto ry, G e o lo g y , O re -D re ssin g, M etallu rgy,
Chem istry, A n a ly s is , A p p lic a tio n s ,
and Econom ics
95 M a d iso n A v e .
4 BOOK
by K . C. L i and Chung Y u W ang
T h e increasing im p o rtance o f tu n g sten in the m etallurgical
and electronic in d u stries h as evoked a need for a com pre­
hensive tre a tm e n t o f its g eology, processing and uses, w h ich
th is b o o k ex p ertly m eets. W ritten by tw o o u tsta n d in g
au th o ritie s in th e field, it presents detailed discussions o f the
occurrence, co m p o sitio n and p re p a ra tio n o f tu n g sten ores in
all p a rts o f th e w o rld . Includes th ree strik in g co lo r rep ro ­
d uctio n s o f tu n g sten allo y s, m aps and o th e r illu stra tio n s.
A valu ab le reference w o rk for a ll those concerned w ith the
uses o r pro p erties o f th is im p o rta n t m etal, ¡ i ; Pages Illu s ­
tra ted
$7.00
BIOCHEMISTRY OF THE
FATTY ACIDS AND THEIR
COMPOUNDS THE LIPIDS
by W . R. Bloor
A m uch-needed stu d y o f th e fats — the o rganic com pounds
so closely associated in th e hum an body w ith carb o h y d rates
and p ro tein s — o f g re a te st value to physicians, n u tritio n is ts ,
o rg an ic chem ists, b io ch em ists, and to th o se in th e food and
iharm aceutical fields. Discusses th e p a rt played by fats and
¡{fids in d ig estio n and n u tritio n , and th e re la tio n o f fa tty
acid m etabolism to such p e rtin e n t subjects as v itam in s, en­
zym es, th e rep ro d u c tiv e cycle, em bolism , anaem ia, cancer,
diab etes, sy p h ilis, a rth ritis , and o th e rs. }$y Pages Illustrated
{
NEW YORK CITY
YOU WILL FIND HELPFUL
Ju st w rite for it
I 11 tills booklet th e '. pH ,
chlorine, and phosphate con­
trol problems o f 34 IN D I­
V ID U A L IN D U S T R IE S are
discussed and brief solutions are given . . . 84 pages o f in­
form ation you will find helpful in m aintenance of precise
process control. IT IS YOURS FO R T H E A SK IN G .
TAYLOR
COM PARATORS
offer m any advantages for pH ,
chlorine, and phosphate control in
practically every in dustry. M olded
o f durable plastic, th ey will la st a
lifetime. A ccuracy o f determ ina­
tion is assured a t all tim es because T aylor Liquid Color
S tandards C A R R Y A N U N L IM IT E D G U A R A N T E E
A G A IN ST F A D IN G . E xtrem ely easy to use — p erm it
quickest m ethod o f determ ination. T h eir cost is insignifi­
c a n t com pared to the savings they m ake possible.
$6.00
S e e y o u r d e a le r or utrite d ir e c t.
Senel for our N ew 1944 Book Catalog
" L e t’s Look It U p "
REINHOLD PUBLISHING CORPORATION
3 3 0 W . 4 2n d Street
N e w Y o rk 1 8 , N .
Y.
\
W. A. TAYLOR co?
7 3 0 2 Y O R K RD. • BALTIM ORE-4, MD.
26
INDUSTRIAL
AND
ENGINEERING
Vol. 16, No.
CHEMISTRY
%T
I
' - Xfe'
/*
I
spe
\ .
o
CSV-'": '
V.r.: •' uD,W
■
X
ß
T hese things fight ! They fight 24 hours a
day! Striking hard blows at the enemy by
c o n sta n tly giving o u r so ld ie rs d e ad lier
weapons, faster planes, b etter arm or. The
few sym bols and words above are a small
p art o f technology, th at know how o f produc­
tion which has become as im portant as
logistics or strategy. These are the weapons
C o l e m a n
i n s t r u m e n t s
a r e
so Id
b y
o f m odern men, the realists o f research
and industry, always seeking the practical,
newer, b etter way o f doing their work.
W e offer these weapons in their m ost
efficient form, proper im plem ents for m en
who have no m ore roots in the past th an
are useful to roll back the future — and
win a war
l e a d i n g
d e a l e r s
in
l a b o r a t o r y
s u p p l i e s
pH E L E C T R O M E T E R S
P H O TO FL UO RO M E TE R S
SPECTROPHOTOMETERS
C O L E M A N
E L E C T R I C
C O .
•
M A Y W O O D ,
ILL.
•
NEW
Y O R K ,
N.
Y.