Characterisation of fly-ash cenospheres from coal

Transcription

Characterisation of fly-ash cenospheres from coal-fired
power plant unit
Maciej Jacek Żyrkowski
Thesis to obtain the Master of Science Degree in
Energy Engineering and Management
Supervisor: Professor Luís Filipe da Silva dos Santos
Co-Supervisor: Doctor Rui Pedro da Costa Neto
Co-Supervisor: Professor Teresa Grzybek
Industrial Supervisor: Engineer Karol Witkowski
Examination Committee
Chairperson: Professor José Alberto Caiado Falcão de Campos
Supervisor: Professor Luís Filipe da Silva dos Santos
Member of the committee: Professor Ana Paula Vieira Soares Pereira Dias
September 2014
1
Acknowledgments
This thesis is based on work conducted within the KIC InnoEnergy Master School, in the MSc
programme Clean Coal Technologies. This programme is supported financially by the KIC
InnoEnergy. The author also received financial support from KIC InnoEnergy, which is
gratefully acknowledged.
KIC InnoEnergy is a company supported by the European Institute of Innovation and
Technology (EIT), and has the mission of delivering commercial products and services, new
businesses, innovators and entrepreneurs in the field of sustainable energy through the
integration of higher education, research, entrepreneurs and business companies.
Shareholders in KIC InnoEnergy are leading industries, research centres, universities and
business schools from across Europe.
www.kic-innoenergy.com
The MSc programme Clean Coal Technologies is a collaboration of:
AGH University of Science and Technology, Kraków, Poland,
SUT Silesian University of Technology, Gliwice, Poland
IST Instituto Superior Tecnico, Lisbon, Portugal
2
This work was done in cooperation with EDF (Electricité de France) coal fired power plant
located in Krakow (Poland), which provided samples to analyze, as well as information
regarding the power plant operation.
3
Summary
Cenospheres are one of the most desired byproducts of coal combustion process nowadays.
They are small hollow spheres with roughly 10 – 1000 µm in diameter and constitute about 1-2
% of the fly ash obtained from the coal combustion process. Because of their specific
properties, namely their low density and very high mechanical strength, cenospheres are an
important subject of coal fired power plants. The objective of this work is to characterize,
chemically and structurally, different cenospheres in order to establish the conditions that
favors their formation. For this purpose, several samples of cenospheres, obtained from fly
ash from coal-fired power plant located in Poland, were analyzed in terms of composition of
cenospheres and fly ash and its relation with glass formation principles and combustion
conditions. Moreover, detailed properties of cenospheres have been obtained and described
by techniques such as SEM, EDS, XRD, XRF and Raman spectroscopy. Results indicate that
cenospheres from coal combustion are a mixture of aluminosilicate glasses with mullite and
quartz crystalline phases. They can present different size and shell structure as well as color.
The high alumina content – roughly 25-27 wt. % - is responsible for the high mechanical
3
strength, while density of most cenospheres is lower than 1 g/cm . Regarding the formation
process, there are interesting correlations between the amount of cenospheres and the
sodium and calcium content, in different fly ash samples.
Key words: cenospheres, cenosphere yield, cenosphere formation, coal combustion, fly
ashes.
4
Resumo
Cenosferas são um dos mais desejados subprodutos da combustão de carvão hoje em dia.
São pequenas esferas ocas com ~ 8-1000 µm de diâmetro, que constituem cerca de 1-2% de
cinzas volantes num processo de combustão do carvão. Devido às suas propriedades, como
por exemplo a baixa densidade e a elevada resistência mecânica, as cenosferas são um tema
importante em muitas centrais de energia. O objectivo deste trabalho é caracterizar, química e
estruturalmente, diferentes cenosferas por forma a estabelecer as condições que favorecem a
sua formação. Assim, diversas amostras de cenosferas, obtidas a partir de cinzas volantes de
centrais de energia movidas a carvão, localizadas na Polónia foram analisadas em termos da
composição das cenosferas e das cinzas e a sua relacão com os princípios de formação
vítrea e as condições de combustão. Além disso, determinaram-se as propriedades
características das cenosferas com diferentes métodos de ensaio, como SEM, EDS, XRD,
XRF e espectroscopia Raman. Os resultados revelam que as cenosferas resultantes da
combustão de carvão são constituídas por uma mistura de vidros de aluminosilicatos com
fases cristalinas de mulite e quartzo. Elas podem apresentar diferentes tamanhos, cor e
estrutura. O elevado teor de alumina (~ 25-27 % peso) traduz-se numa elevada resistência
3
mecânica, sendo a densidade da maior parte das cenoesferas inferior a 1 g/cm .
Relativamente ao processo de formação, determinaram-se correlações interessantes entre a
quantidade de cenosferas e a composição de sódio e de cálcio em diferentes amostras de
cinzas volantes.
Palavras-chave: cenosferas, rendimento de cenosferas, formação de cenosferas, combustão
de carvão; cinzas volantes.
5
Thanks
I hereby would like to thank everyone who has helped me during writing this thesis. In
particular I would like to thank:
Prof. Luis Santos, for devoted time as being my coordinator, for his great advices and
invaluable commitment to this work,
Doc. Rui Costa Neto, for devoted time as being my co-coordinator, for his great advices and
commitment to this work,
Prof. Teresa Grzybek, for devoted time as being my co-coordinator and for great help in
dealing with formal and scientific problems,
Prof. Toste Azevedo of blessed memory, for his help and having a great concept about this
work,
Eng. Karol Witkowski, for his help and commitment during my internship, as well as during
process of writing this thesis,
For all my great family and my beloved girlfriend, who supports me all the time.
6
Table of Contents
I.
Introduction…………………………………………………………………………………… 12
1. Background and objectives…………………………………………………………………..12
2. Pulverized coal combustion………………………………………………………………… 16
3. Fuel pathway in coal fired power plant……………………………………………………..17
4. By-products from coal combustion and their application………………………………… 18
5. Cenospheres………………………………………………………………………………….. 20
6. Cenosphere formation……………..………………………………………………………... 20
7. Pulverized coal combustion boiler OP-430 review……………………………………..… 24
8. Environmental concerns…………………………………………………………………..… 28
II.
Experimental Techniques…………………………………………………………………….29
1. Cenosphere separation…..………………………………………………………………….. 30
2. XRF……………………………………………………………………………………………. 31
3. XRD…………………………………………………………………………………………… 31
4. SEM / EDS……………………………………………………………………………………. 32
5. Raman spectroscopy………………………………………………………………………... 32
III.
Results & Discussion……………………………………………………………………….. 34
1. Results from laboratory work……………………………………………………………….. 34
2. XRF……………………………………………………………………………………………. 38
3. XRD…………………………………………………………………………………………..…42
4. SEM / EDS………………………………………………………………………………….....47
5. Raman spectroscopy…………………………………………………………………………57
6. Viscosity………………………………………………………………………………………..62
7. Mineral matter transformation……………………………………………………………..…63
8. The analysis of combustion process………………….………………………………….….65
IV.
Conclusions & Future Investigation………………………………………………………... 68
1. Conclusions………………………………………………………………………………….…68
2. Future investigation……………………………………………………………………………68
References………………………………………………………………………………………………..70
7
List of Figures
Figure 1. Coal-fired power plant scheme ...............................................................................15
Figure 2. Simplified mechanism of fly ash formation .............................................................17
Figure 3. Broken cenosphere ...............................................................................................19
Figure 4. Temperature vs. Enthalpy plot for glass-forming melts ...........................................22
Figure 5. Influence of Illite, Chlorite and Montmorillonite on the yield of cenospheres . ..........23
Figure 6. Process of cenosphere formation during coal combustion in steps. ........................24
Figure 7. Variation of temperature along boiler OP-430 boiler height . ...................................26
Figure 8. Boiler OP-430 scheme and height [m] ...................................................................27
O
Figure 9. Boiler OP-430 burners layout and temperatures [ C] .............................................27
Figure 10. Temperatures distribution after burner’s outlet ......................................................27
Figure 11. Boiler OP-430 with main components and temperatures of exhaust gases ..........28
Figure 12. Stirring and sedimentation of samples. .................................................................30
Figure 13. Draining of skimmed matter..................................................................................31
Figure 14. Dried samples prepared for burnout. ....................................................................31
Figure 15. Cenospheres content In each sample of tested fly ash from ESP zone I. ..............35
Figure 16. Cenospheres content In each sample of tested fly ash from all ESP zones ..........37
Figure 17. SEM image of cenosphere sample D00Pcen........................................................38
Figure 19. SEM image of cenosphere sample D01Rcen........................................................38
Figure 27. Plots from XRF analysis for fly ash sample D00P .................................................40
Figure 28. Plots from XRF analysis for fly ash sample D03R .................................................41
Figure 29. Correlation between Na/Ca ratio in fly ash samples and content of cenospheres
obtained from given sample. .................................................................................................42
Figure 30. XRD pattern for cenosphere sample D00Pcen .....................................................43
Figure 32. XRD pattern for cenosphere sample D01Rcen .....................................................44
Figure 40. SEM picture and EDS spectrum of cenosphere sample D00Pcen.........................49
Figure 41. SEM picture and EDS spectrum of cenosphere sample D01Rcen ........................49
Figure 44. SEM picture and EDS spectrum of fly ash sample D02R ......................................50
Figure 47 SEM picture and EDS spectrum of fly ash sample D03R ......................................50
8
Figure 51 SEM picture and EDS spectrum of cenosphere sample D07Pcen..........................51
Figure 52. SEM picture and EDS spectrum of fly ash sample D07P ......................................52
Figure 53. Cenosphere SEM image relative to EDS data ......................................................53
Figure 61. Sample D00Pcen fragment A ...............................................................................59
Figure 62. Raman spectrum for sample D00Pcen fragment A ...............................................59
Figure 63. Sample D00Pcen fragment B ...............................................................................59
Figure 64. Raman spectrum for sample D00Pcen fragment B ...............................................60
Figure 65. Sample D00Pcen fragment C ...............................................................................60
Figure 66. Raman spectrum for sample D00Pcen fragment C ...............................................60
Figure 67. Sample D00Pcen fragment D ...............................................................................61
Figure 68. Raman spectrum for sample D00Pcen fragment D ...............................................61
Figure 69. Sample D03Rcen fragment A ...............................................................................61
Figure 70. Raman spectrum for sample D03Rcen fragment A ...............................................62
Figure 71. Sample D03Rcen fragment B ...............................................................................62
Figure 72. Raman spectrum for sample D03Rcen fragment B ...............................................63
Figure 73. Sample D03Rcen fragment C...............................................................................63
Figure 74. Raman spectrum for sample D03Rcen fragment C ...............................................63
Figure 75. Viscosity in function of temperature for four samples of fly ash with different
cenospheres content. ...........................................................................................................64
Figure 76. Binary phase diagram for SiO2-Al 2O3 system with relevant regions ......................65
Figure 77. System SiO2 – Al2O3 – CaO on ternary phase diagram . .......................................66
Figure 78. Exhaust gases temperature after bulkhead superheater - left side vs. Yield of
cenospheres [%]. ..................................................................................................................67
Figure 79 Exhaust gases temperature after bulkhead superheater - right side vs. Yield of
cenospheres [%]. ..................................................................................................................68
Figure 80. Graph comparing exhaust gases temperatures after bulkhead superheater (left side
and right side) and power produced by generator, in function of cenospheres yield. ..............69
9
List of Tables
Table 1. Coal composition for different types of coal .............................................................13
Table 2. Results from cenospheres separation for ESP zone I. .............................................35
Table 3. Results from laboratory work for ESP zones I, II and III. ..........................................36
Table 4. Data obtained from XRF – elementary analysis of fly ash samples in ratio of silica. .43
Table 5. SEM Pictures and elements content in single cenospheres particles .......................53
Table 6. EDS quantitative analysis of elements from fly ash samples. ...................................56
Table 7. EDS quantitative analysis of elements from cenosphere samples in [wt.%]. .............56
Table 8. EDS quantitative analysis of elements from cenosphere samples in [at.%]. .............57
Table 9. Oxide composition of cenosphere samples in [%] ....................................................57
Table 10. Oxide composition of fly ash samples in [%] ..........................................................57
Table 11. Comparison of viscosities for four fly ash samples for distinctive temperatures. .....65
10
Nomenclature
H – enthalpy [kJ/mol]
T – temperature [OC] [K]
wt. % - weight percentage
at. % - atomic percentage
η – viscosity [Pa s]
ρ – density [g/cm3]
cc – cubic centimeter
2
SSA - Specific Surface Area [cm /g]
2
A - area [mm ]
c – content of cenospheres in fly ash sample [%]
σ – standard deviation, in units of given measurement.
d – diameter [µm]
λ – Air excess coefficient
P – pressure [MPa]
W – steam load [kg/s]
m – mass [g[
ζ – tensile strength [GPa]
t – time [ms]
O
Tg – glass transformation temperature [ C]
ESP – Electrostatic precipitator
XRF – X-Ray Fluorescence
XRD – X-Ray Diffraction
SEM – Scanning Electron Microscopy
EDS - Electron Dispersive Spectroscopy
α, A’, B, B0, B1, B2, B3, – Constants
11
I.
INTRODUCTION
1. Background and objectives
Cenospheres constitute a small but important fraction of by-products from coal combustion
process. Because of some distinctive properties, like their low density and high mechanical
strength, they are applicable in many branches of industry and nowadays, cenospheres
become a desired and precious product (Petrus, 2011). This is why coal fired power plants are
strongly interested in this issue and cenosphere recovery process is now under investigation.
Cenospheres are hollow ceramic microspheres that are a by-product of coal burning power
plants. When pulverized coal is burnt at power plants, fly ash is produced. Cenospheres are
the lighter particles present in fly ash and constitute about 1-2% wt.% of the fly ash. The
objective of this work is to characterize, chemically and structurally, different cenospheres in
order to establish the conditions that favors their formation. Laboratory work performed by the
author indicated, that the content of cenospheres can vary from 0.2 wt.% to 1.9 wt.%
depending on the fly ash sample, therefore it is imperative to understand the cenosphere
formation process in order to maximize the yield of cenospheres obtained. This knowledge
would be especially important, regarding industrial cenosphere production process conducted
by coal fired power plant. The main factors can be related with the combustion process itself
as well as with the mineral composition of melting char inside the boiler. Both issues are
important to understand how cenospheres are formed and which specific factors – such as the
boiler temperature or the content of some specific elements – promote this process and allow
increasing their yield in fly ash. Therefore, despite of this work title, the author focuses not only
on the thermodynamic processes occurring in the boiler, but also on the chemical composition
of cenospheres and fly ash.
12
2. Pulverized coal combustion
Coal is an organic fossil fuel with the dominant content of carbon element that varies from
about 30% by weight in lignite coals up to 85% by weight in anthracite coals. In the middle of
this rank there are bituminous coals which are going to be a major concern of this work. In
Error! Not a valid bookmark self-reference. general differences between different coal types
are presented.
Table 1. Coal composition for different types of coal (Coal Classification, 2014).
Lignite Coals
Bituminous Coals
Anthracite Coals
Fixed Carbon
31.4 wt. %
44.9-78.2 wt. %
80.5-85.7 wt. %
Ash
4.2 wt. %
3.3-11.7 wt. %
9.7-20.2 wt. %
Sulfur
0.4 wt. %
0.7-4.0 wt. %
0.6-0.77 wt. %
Moisture
39 wt. %
2.2-15.9 wt. %
2.8-16.3 wt. %
Bulk Density
641-865 kg/m
3
673-913 kg/m
3
800-929 kg/m3
Despite of organic content in coal, there is also mineral content which is responsible for ash
formation, as well as a surface and inherent moisture. During heating, part of mineral and
organic matter in coal are being released as volatile substances – water, steam, volatile tar
and gases, such as H2,CH4, CO, CO2, C2H6 and C2H4. Because properties of coal change
along with conditions, different states of coal have been determined (Tomeczek J., 1994) :
-
Analytical state, when moisture in coal is in equilibrium with moisture in ambient air.
-
Dry state, when surface moisture is removed by drying in temperature about T = 110ºC.
-
Water and ash free state
-
Ash free state.
There are also three types of analysis to characterize coal (Tomeczek J., 1994):
-
Proximate analysis, to find out moisture, ash and volatile matter.
-
Ultimate analysis, to find out content of radicals such as C, H, S, N and O.
-
Miscellaneous analysis, to find out heating value, forms of sulfur, chlorine, trace metals,
CO2, free swelling index, dilatation, ash fusibility and ash composition.
When volatiles and all moisture are removed from coal, what remains, is called char. The total
heat of combustion depends on the amount of char, from a minimum of 40% (for strongly
volatile peat) to a maximum of 95% (for anthracite). Burning char is also the most timeconsuming process.
Combustion of pulverized coal can be divided into the following steps:
-
Heating and removing surface moisture
-
Removing inherent moisture
13
-
Releasing volatiles
-
Forming char
-
Igniting
-
Burning volatiles and char.
Combustion is basically a reaction of carbon from coal with oxygen and can be written as a set
of the following chemical reaction (Tomeczek J., 1994):
C + O2 = CO2
∆H = -393.7 kJ/mol
[1.1]
C + 0.5O2 = CO
∆H = -110.1 kJ/mol
[1.2]
C + CO2 = 2CO
∆H = +172.6 kJ/mol
[1.3]
C + H2O = CO + H2
∆H = +131.4 kJ/mol
[1.4]
CO + 0.5O2 = CO2
∆H = -283.6 kJ/mol
[1.5]
H2 + 0.5O2 = H2O
∆H = -242.4 kJ/mol
[1.6]
CO + H2O = CO2 + H2
∆H = -41.2 kJ/mol
[1.7]
When the change of enthalpy ∆H is negative, the reaction is exothermic (heat is being
released during reaction). If ∆H is positive, reaction is endothermic (heat is absorbed during
reaction). Despite of the large number of reactions, combustion is mainly based on reactions
[1.1] and [1.2]. Both products of combustion – CO2 and CO – can be named as primary
combustion products. The ratio CO/CO2 increases with temperature and decreases with
pressure. The typical residue time for coal particle in pulverized coal boiler is 1-2 second
which is usually sufficient for complete combustion (Tomeczek J., 1994) (Smoot, 1993).
As long as oxy-combustion technology is not considered, oxygen required to carry on the
combustion process is delivered with air. It is also assumed, that air contains 21% of oxygen
and 79% of nitrogen. The excess of air coefficient λ is an important parameter for combustion
process. It indicates how much air (and oxygen at the same time) is being delivered to the
combustion process with respect to the molar content derived from stoichiometric equations.
When λ<1 there is rich combustion (there is more fuel than oxygen) while λ>1 is called lean
combustion process (there is more oxygen than fuel). When λ=1, the amount of oxygen
delivered to combustion is equal to one derived from stoichiometric equations. That should be
the ideal situation (when combustion is most efficient), however in reality, oxygen cannot mix
with the fuel at a sufficient level. To enhance it, λ>1 is usually applied in pulverized coal
combustion processes. On the other hand, a high excess of air will bring about significant heat
losses of exhaust gases and decrease of combustion temperature. Typically, an optimal value
14
of λ for pulverized coal combustion is between 1.1 and 1.3 (Nocoń, Poznański, & Słupek,
1994).
Temperature of combustion is another relevant parameter and it mainly depends on the type
of coal used. For example, the temperature of combustion will be higher for bituminous coals
than for lignite coals. Besides that, temperature will rise, if air is preheated before combustion
and decrease with increasing air excess and amount of recirculated exhaust gases (Laudyn
D., 2000).
3. Fuel pathway in coal fired power plant.
Figure 1. Coal-fired power plant scheme
After transportation from a mine, coal is stored in the landfill (Fig.1-1), and then it is
transported via belt conveyor (Fig.1-A) to the coal bunkers (Fig.1-2). Next, coal is carried on
by a feeder screws (Fig.1-B) to the coal mills (Fig.1-3). There are usually a few coal mills to
avoid energy production gaps due to mill breakdown. Pulverized coal is afterwards conveyed
in the stream of primary air (Fig.1-H’) using a primary air fan (Fig.1-4). A mixture of fuel and air
is injected to the combustion chamber of the boiler (Fig.1-5) by several burners. At the same
time, secondary air (Fig.1-H’’) is being delivered to the boiler by means of a forced draft fan
(Fig.1-6). The main by-products formed during coal combustion process are slug, bottom ash
and fly ash. The heavier slug and bottom ash (Fig.1-D) are removed at the bottom of the boiler
whereas light fly ash (Fig.1-F) soars up and is transported with exhaust gases to the
electrostatic precipitator (Fig.1-7). There the separation occurs - exhaust gases are moved by
an induced draft fan (Fig.1-8) to the stack and fly ash is stored and afterwards conveyed to the
fly ash utilization unit (Fig.1-F’). Steam produced in the boiler (Fig.1-a) goes to the high
pressure part of the turbine (Fig.1-10). From there, part of steam comes back to reheat (Fig.1-
15
b) and the rest enters the low pressure part of the turbine (Fig.1-11). Next, steam with low
enthalpy flows down to the condenser (Fig.1-12) and in a liquid state returns to the boiler
(Fig.1-c).
4. By-products from coal combustion and their application.
By-products from coal combustion are an effect of the presence of inorganic mineral matter in
coal composition. As it was mentioned before, the main by-products from coal combustion are
slag, bottom ash and fly ash. If power plant involves SOx and NOx treatment, we refer to
clean-coal ash from an FGD (Flue-Gas Desulfurization) unit. Fly ash constitutes a major part
of all by-products: 70-90%, while bottom ash is usually between 10-30 %. Bottom ash is
comprised of coarse heavy particles, whereas fly ash particles are light and mostly fine, with
3
specific density from 1.6 to 2.6 g/cm (Siddique, 2010). Slag is basically a melted bottom ash.
The major mass fraction of ash (about 97%) is formed by oxides: SiO2, Al2O3, TiO2, Fe2O3,
CaO, MgO, K2O, Na2O, SO3, P2O5 and trace elements like: Zn, Pb, V, Ni, Cu, Co and Cr
(Tomeczek J., 1994).
Fly ash is the major concern of this work, because it contains a small amount of cenospheres.
According to The American Society for Testing and Materials (ASTM) there are two classes of
fly ash: Class C - from lower ranked coals - containing more than 50% of combined SiO2,
Al2O3, Fe2O3 content and class F - from higher ranked coals – containing more than 70% of
combined SiO2, Al2O3 and Fe2O3 content. This also means a higher CaO content in C class fly
ashes (Blissett R.S., 2012). Typically, a class F fly ash contains more than 85% of SiO2 and
Al2O3 and less than 2.3% of CaO (Lee S.H., 1999). Because almost all organic matter is
burned away during coal combustion, fly ash is mainly formed by incombustible mineral
matter. This in turn can form crystalline phases, such as quartz (from silica) or magnetite (from
iron), aluminosilicate minerals like mullite or amorphous glass (Ward C.R., 2006), depending
on the quenching rate and viscosity of particular melt (Shelby J.E., 1997). The main materials
that constitute fly ash are: glass, spinels, hematite, mullite, clay and mica minerals, crystalline
silicates and quartz (Hower J., 2012) (Blissett R.S., 2012). Mean fly ash particle size is usually
in the range from about 5 µm to about 35 µm. Glass content in fly ash can vary from about
66.7% to 78.5% and it turned out to be higher at full boiler load than at half load (Lee S.H.,
1999). According to Ward and French (Ward C.R., 2006), glass in fly ash from lower ranked
bituminous coals (sub-bituminous) contains lower percentage of SiO2 and higher percentage
of Fe2O3, than glass in fly ash obtained during combustion of higher ranked bituminous coals.
Figure 2 depicts the mechanism of fly ash formation. Firstly, vaporization of volatile species
like Na, Pb and Hg occurs. Those elements can undergo nucleation to form new crystallites, or
they may condense on the surface of already existing particles (Chen L., 2012). Subsequently
(or sometimes simultaneously), coal transforms to char, and inorganic elements are released
by vaporization. Char has a high temperature of burnout, and because of that, as well as
because of reducing atmosphere, even oxides with very high melting and boiling temperature
16
(like SiO2, CaO, MgO) start to volatize. Volatile forms that occur afterward can be SiO, Ca, Mg
and NaCl. What is more, the amount of vaporized content rise with increasing temperature of
char (Fenelonov V.B., 2010). In the next step, volatiles form ash particles due to condensation
and nucleation process. The described mechanism is true for fine ash particles, coarse
particles though, are formed from coalescence and fragmentation of non-volatile coal’s mineral
content (Chen L., 2012) (Blissett R.S., 2012) (Smoot, 1993). Fly ash formation during coal
combustion is also connected with the formation of deposit on the boiler walls, which is a huge
problem because of the corrosion effect and decrease of heat transfer. The most dangerous
are sodium and potassium that vaporize at low temperature and then condense forming
sulphate corrosive compounds. Those are, regarding sodium, for example NaCl, Na2SO4,
Na2SiO3 (Tomeczek J. P. H., 2002).
Figure 2. Simplified mechanism of fly ash formation, adapted from (Chen L., 2012).
Fly ashes from coal combustion have a variety of applications. Most commonly, they are used
as a component of cement (constitutes more than 70% of cement mixture). Fly ash can be
used to produce high-quality concrete pavements. High-performance concrete attains its best
durability and strength proportions with 30-40% of fly ash content (Siddique R., 2010). Cleancoal ash is used as base of road construction, as well as a concrete component. Fly ash can
be also used as a stabilizer for soil. The addition of fly ash to the soil, decrease its water
absorption capacity and therefore the swelling effect. This is beneficial, as the swelling soil can
damage pavements, roads or pipelines (Blissett R.S., 2012) (Siddique R., 2010). Fly ash can
be utilized as adsorbent of various pollutants from gases or water solutions, for example for
removing heavy metals from industrial water ash (Ahmaruzzaman, 2010). Larger size bottom
ash can be used to manufacture of masonry and as coarse aggregate. Slag is usually utilized
17
as a component for architectural concrete to improve its abrasion resistance (Blissett R.S.,
2012).
5. Cenospheres
Microspheres are hollow glass particles with low density. The term cenospheres is used for
microspheres naturally formed during pulverized bituminous coal combustion. They are a
fraction of fly ash and the major concern of this work.
There are two approaches in literature to cenospheres. The first - and more popular one assumes that they, in majority, have a density lower than the density of water, therefore they
float and can be easily collected by wet separation. This approach is justified partly because
light cenospheres are the most desirable for industries. However, there is a second approach
3
which states that in fact all particles with density below ρ = 2.2 g/cm should be treated as
cenospheres (Ghosal S., 1995). This view is based on the fact that, from all oxides present in
3
fly ash, the crystalline forms of silica have the lowest density equal to ρ = 2.2 g/cm . This
means that there cannot be an utterly solid particle with lower density than this. However, if we
follow this point of view, it must be realized that particles with density between ρ = 1g/cm3 and
3
ρ = 2.2 g/cm are going to have respectively thicker walls or contain only small voids. At the
3
same time, the wall thickness of a cenosphere with density lower than ρ = 1g/cm (called
“light”) is about 5-10% of its diameter (Barbare N., 2003), as can be observed in Figure 3.
Usually, the amount of light cenospheres in fly ash is about 1-2 wt.%, although it can go up to
3
4 wt.% (Fleming D., 2012). This is no longer valid if all particles with density ρ < 2.2 g/cm are
considered, because in this case cenospheres content can be as high as 95% (Ghosal S.,
1995). Kolay (Kolay & Bhusal, Recovery of hollow spherical particles with two different
densities from coal fly ash and their characterization, 2014) indicates, that 75% of all particles
3
with density up to 1.282 g/cm are cenospheres.
18
50 µm
Figure 3. Broken cenosphere, adapted from (Shao Y., 2009)
From the point of view of this work, the composition and structure of cenospheres are quite
important. As mentioned before, they seem to present a mixed glass and crystalline structure.
According to Shao (Shao Y., 2009), they are aluminosilicate-based particles composed of
Al6Si2O13, SiO2 and a glass phase. Vassilev (Vassilev S.T., 2004) indicates, that phasemineral composition of cenospheres is composed mainly of aluminosilicate glass, quartz,
mullite, calcite, Fe oxides, Ca silicates and sulphates while the gases inside are mainly CO2
and N2 with traces of CO, O2 and H2O. The outer shell of a cenosphere is formed from
aluminosilicate glass and its skeleton from mullite, quartz, cristobalite and feldspar. The
phase-mineral composition of cenospheres is 76% of glass, 22% of mineral matter and 2% of
char (Vassilev S.T., 2004). Cenospheres have high porosity, resulting in water absorption
capacity. According to Barbare (Barbare N., 2003), they can absorb 18-times more moisture
than sand particles due to their porous surface.
Regarding other parameters, cenospheres have low density and spherical shape, as well as
very good mechanical strength and low thermal conductivity (Wang, Liu, Du, & Guo, 2012)
(Chavez-Valdez, 2011). They are also characterized by their good resistance in highly
oxidative and corrosive environments, as well as high thermal shock resistance. This is
caused by the high mullite content in cenospheres (Chavez-Valdez, 2011). According to
Hirajima (Hirajima, 2010), 90% of the cenospheres have diameters between 100-420 µm,
while the remaining 10% have diameters lower than 100 µm. According to the same author,
3
80% of cenospheres have density lower than ρ = 1 g/cm . Vassilev (Vassilev S. T., 2004)
indicates that the diameter of cenospheres can vary from 8 µm up to 1000 µm (sometimes
3
3
cenosphere can reach even 3 mm) and density varies from 0.4 g/cm to 0.8 g/cm . Kolay
(Kolay, Recovery of hollow spherical particles with two different densities from coal fly ash and
their characterization, 2014) correlates the average diameter with density – for particles with
3
density lower than 0.857 g/cm , 80% of them have diameter between 50 - 150 µm, 15%
19
diameter between 150 – 200 µm and 5% between 200 – 250 µm. For particles with density
3
lower than 1.282 g/cm , 10% of them have diameter between 40 – 50 µm, 50% between 50 –
100 µm, 20% between 100 – 110 µm and 20% have diameter between 110 – 150 µm.
O
O
Cenospheres tend to sinter at temperature from 950 C – 1200 C and melt at temperature
from 1250
O
C to 1450
O
C (Vassilev S. T., 2004). Kolay (Kolay, Physical, chemical,
mineralogical, and thermal properties of cenospheres from an ash lagoon., 2001), indicated a
specific surface area of cenospheres of 457 cm2/g.
Cenospheres possess a bunch of distinctive properties, which makes them interesting for a
variety of applications. They can be used as a drinking water purification agent to help
removing excess of fluoride (Blissett R.S., 2012). Cenospheres can be covered by various
metals and act as a magnetic waves shield, therefore can be use in electronic and radar
applications (Blissett R.S., 2012). Cenospheres can be used in refractory materials due to
their low thermal conductivity and high temperature resistance. They are also referring as a
material for hydrogen storage. Finally, their light weight and high strength make them suitable
for the design of modern light-weight composite materials which are demanded in automotive
and aerospace industry (Blissett R.S., 2012). Because of their good water absorption capacity,
cenospheres added to concrete allow it to be desirably dry (water in concrete leads to
corrosion) and also decrease its density (Barbare N., 2003).
Cenospheres separation can be carried out by two main ways – by wet method or by dry
method. The former one was the first applied and it is still the most popular way to obtain
cenospheres. Particles with density lower than density of a given liquid will float in this liquid;
therefore, for the major fraction of cenospheres it is enough to use water. This technique
however, has several disadvantages. The most important are the environmental restrictions
and the fact that the remaining wet fly ash cannot be used for any further applications.
Regarding that, the dry separation method is now being considered. Basically, using dry
methods, the fly ash can be divided by its size, weight and density. In the first step, fly ash
must be separated into two fractions – one comprised of coarse particles and a second one
with fine particles. This process can be done using dry screening technology, such as
ultrasonic sifting (Ramme B.W., 2011). The initial screening is suggested to be from 38 to 45
micron sieve. Afterwards, the obtained fractions can undergo density separation process in
fluidized bed column without presence of water or air classification. When air classification is
concerned, fly ash is fed into a rotating turbine wheel and smaller particles are collected due to
centripetal forces while larger particles are blown away by centrifugal forces (Ramme B.W.,
2011). Apart from standard hollow cenospheres, other similar particles as plerosphers and
ferrospheres (ferrispheres) can be found in fly ash. The former one is basically a cenosphere
with another particle trapped inside, while the latter particle is an amorphous or crystalline
sphere with an increased amount of Fe (Vassilev S.T., 2004).
6. Cenospheres formation
20
The formation mechanism of cenospheres during pulverized coal combustion is complex and
strongly depends on fuel properties and combustion parameters. Ghosal (Ghosal, 1995)
points out that cenosphere formation is similar to the glass blowing process. Therefore it would
be beneficial to take a closer look into the glass formation principles. Figure 4 presents a
typical relation between enthalpy of a glass and temperature. For a normal (not very fast)
cooling rate, the melted composition, after reaching temperature Tm, would experience a rapid
drop in enthalpy and then formation of crystalline phases. However, for fast cooling rates,
there is no time for atomic diffusion to occur and structural arrangements to happen and a
super cooled liquid is obtained. As the temperature drops, the liquid viscosity increases to a
point where structural arrangements are no longer possible and an amorphous solid is
obtained. This point occurs between Tslow and Tfast, when its viscosity becomes on average
10
11.3
.
Pa s (Shelby J.E., 1997). In this case, crystals do not form and enthalpy presents a
gradual decrease. Exact temperature of glass transformation (Tg) between Tslow and Tfast
depends on the cooling rate of the melt, as depicted in Figure 4 (Shelby, 1997). Silicacontaining melts (as in the case of cenospheres), have especially high viscosity and therefore
are very agile to form glass. The addition of small amounts (up to 20%) of alkali oxides (Na2O,
K2O) or earth alkali oxides (CaO, MgO) breaks the bonds between silica and oxygen, bringing
about a rapid decrease in viscosity, although alkali oxides have a slightly higher impact here.
Those oxides are called glass modifiers. Shelby (Shelby, 1997) indicates, that the glass
O
transformation temperature is about 200 C higher for calcium (earth alkali) silicate melt than
for sodium (alkali) silicate melt, thus amorphous phase is going to be reached faster for the
first melt (Shelby, 1997). Referring to cenosphere formation, it was mentioned before, that
some of them can have amorphous outer shell and crystalline skeleton. This might be
explained by the fact, that the outer shell is subjected to faster temperature change than the
inner core (Fenelonov, 2010). According to M. Mollah (Mollah, 1999), the inner skeleton can
be either amorphous or crystalline, but usually presents a higher content of earth alkalis than
the outer shell.
21
Figure 4. Temperature vs. Enthalpy plot for glass-forming melts (Shelby J.E., 1997).
Glass constitutes about 70 % of the fly ash, which means that more than a half of all particles
undergo a glass transition process. Therefore, we need to investigate, why some of them end
up as cenospheres, as well as what could be done to increase their yield. The gass emitted
from a char particle during combustion, or from ash particle during melting, inflate inorganic
mineral matter and if quenching occurs at a proper rate, amorphous cenosphere is formed.
Gases capable to inflate spherical particle can come from the decomposition of calcium and
magnesium sulfates, kaolinite, calcium carbonate, dolomite and pyrite oxidation (Karr, 1979) all these reactions can occur at temperatures below 1000 OC. Time of residence in high
temperature zone for cenosphere-forming particle, should not exceed 200 ms (Łączny M.,
2011), and time needed to form a 50 µm cenosphere is about 0.3 ms (Karr, 1979). According
to Vassilev (Vassilev S. T., 2004), cenospheres have lower SiO2 / Al2O3 ratio than fly ash and
the melt composition favoring their formation is a mix of chlorite, illite, kaolinite, muscovite,
montmorillonite and plagioclase. They can be found in minerals such as clay, mica and
feldspars. From figure 5, chlorites and montmorillonites seem to have some catalytic
properties for cenospheres formation, while illite works as a deterring agent. Also more
cenospheres can be found in coarser grinded fly ash (Vassilev S. T., 2004).
22
Figure 5. Influence of Illite, Chlorite and Montmorillonite on the yield of cenospheres (CCC) (Vassilev S.T., 2004).
O
Regarding ash-fusion temperatures, coals with melting temperature between 1400 C – 1500
O
C turned out to be richer in cenospheres than coals with lower or higher ash-fusion
temperatures (Vassilev S.T., 2004). According to Karr (Karr C., 1979), at temperature higher
O
than 1500 C, gas evolution will be so rapid that it will escape from melting ash particle.
According to Tomeczek (Tomeczek J., 1994), sodium chloride (NaCl) liquid-to-gas phase
O
change occurs at temperature 1465 C and according to Fenelonov (Fenelonov V.B., 2010)
NaCl is one of the compounds that are more prone to volatize from char at high temperature.
This, as well as the investigation made by Vassilev (Vassilev S.T., 2004), may lead to a
conclusion that sodium chloride content in coal can be helpful for cenosphere formation. Other
known factors that can promote cenospheres formation are: bigger size of pulverized coal
grains (Vassilev S.T., 2004) (Łączny M., 2011), fast cooling rate of melt and its high viscosity
(Shelby J.E., 1997). Raask (Raask E., 1968) points out that mineral particles with diameters
below d = 10 µm are not capable of forming cenosphere.
The mechanism of cenosphere formation during pulverized coal combustion was
comprehensively described by (Fenelonov V.B., 2010). Several stages of this process are
presented in Figure 6. First step, from ”a” to “b”, represent the volatilization of species from
char particle which was explained also in the previous chapter about fly ash formation. In step
“c” char starts to burn simultaneously with further volatilization. The steps described as “e” and
“e1” represent condensation and nucleation process that vaporized elements undergo, and
fine particles formation afterwards. The transition from point “c” to “d” depicts the beginning of
coarse fly ash particles formation – char fragmentation. At this stage, as temperature
increases, bonds between carbon atoms start breaking and thus melting gets underway. The
melted substance here is comprised of unstable mineral forms and products of coal pyrolysis.
Then due to the surface capillary forces acting on melt, the spherical shape drops are being
formed in step “f”. The temperature, at which it happens, is called “spherical temperature” and
according to Itskos (Itskos G., 2010) is between 1100 OC and 1300 OC. In the next stage,
capillary forces attract single drops together, promoting their coalescence and forming one
bigger spherical particle, which is described in steps “g” and “h”. After this, the actual
cenosphere formation process begins. In step “i" gas released from the inner side of the
particle starts forming bubbles. The gas itself comes from pyrolysis of char as well as from the
23
decomposition of minerals. Pressure of this gas blows cenosphere further as depicted in the
step “j”. If the pressure is too high, thin walls will break. In such situation, a particle collapses
and shrinks, to be afterwards swollen again by gases released from its core. This process is
presented as the transition between steps “j” and “k”. These two stages can be repeated
several times, while there is gas, or until the outer shell becomes rigid due to the crystallization
or glass formation process. Depending when and how fast the cooling of the sphere occurs, it
may end up as totally hollow cenosphere “l”, plerosphers “m” or “n”, broken cenosphere “o” (if
gas pressure is too high and particle explodes) or just as small fly ash solid particles depicted
in point “p”.
Figure 6. Process of cenosphere formation during coal combustion in steps (Fenelonov V.B., 2010).
As mentioned above, viscosity is an important factor, because it is strictly related to the glassforming ability of a given melt. The higher the viscosity is, the more easily glass will form.
However, this information is only relevant when we heat above the melting temperature of a
crystal, that could possibly be formed from considering melt composition and if cooling
conditions are sufficient for glass formation - for example too slow cooling rate will promote
crystal formation (Shelby, 1997). Viscosity of fly ash as a function of temperature can be
calculated from empirical equations, such as the ones proposed by G.Urbain and latter
modified by Kalmanovitch-Frank. The calculation procedure is the following (Vergas,
Frandsen, & Dam-Johansen, 1997):
24
M = MgO + Na2O + K2O + FeO + MnO + NiO + 2TiO2 + 2ZrO2
[2.1]
α = M / (M+Al2O3)
[2.2]
.
. 2
B0 = 13.8 + 39.9355 α – 44.049 α
.
[2.3]
. 2
[2.4]
. 2
[2.5]
B1 = 30.481 - 117.1505 α + 129.9978 α
.
B2 = -40.9429 + 234.0486 α – 300.04 α
.
. 2
B3 = 60.7619 - 153.9276 α + 211.1616 α
B=
.
B0+SiO2 B1
2.
[2.6]
3.
+ (SiO2) B2 + (SiO2) B3
[2.7]
.
-lnA’ = 0.2812 B + 11.8279
. .
[2.8]
.
η = A’ T exp(1000 (B/T)
[2.9]
M – content of species (in mole fraction) which have viscosity-decreasing properties (all Fe
here is assumed to be in the form of FeO). Content of Al2O3 and SiO2 also in molar fraction.
.
η – viscosity [Pa s]
T – temperature [K]
α, A’, B, Bi - Constants
7. Pulverized coal combustion boiler OP-430 review.
All samples used in this work were obtained from waste products of coal combustion from
boiler OP-430 made by (Rafako S.A., 2014). In figure 7, variations of mass-weighted and
area-weighted average temperatures along the boiler OP-430 height are presented. The
highest temperatures in the boiler occur at heights between 8 and 12 meters – where the
temperature reaches up to 2100 OC. Burners in this boiler are installed between 12 and 17
meters height, which corresponds to the maximum of average temperature (mass-weighted
and area-weighted). At a height of about 18 meters, there is a drop in average temperature,
while maximum value of temperature does not drop at this point. Thus, if mass-weighted
O
average temperature at 18.5 m height is equal to 1450 C and maximum temperature is equal
O
O
to 1700 C at the same height, the minimum value must be equal to 1200 C. Therefore we
O
have a difference of 500 C in horizontal direction at the same level of height. This drop in
temperatures can be explained by the fact, that between 18.5 – 19.5 meters height, auxiliary
nozzles responsible for nitrogen effect reduction are installed. The author of this work
assumes, that most probably the highest temperature drop in the horizontal direction occurs
just next to the boiler walls which are cooled down by the water of temperature much below
O
500 C. The decrease in temperature in the vertical direction is also significant – temperature
O
O
falls from 1700 C to 1300 C between 12 and 26 meter of height. From Figures 8 to 11 we
can notice, that at a height level of 35 m, the average temperature in the boiler is 1210 OC
O
while at about 40 m, temperature is equal to 905 C. This consideration is important from the
cenosphere formation point of view, as we would like to know where inside the boiler particles
are subjected to the most rapid quenching process. Figures 8 and 9 present the scheme of the
OP-430 boiler, with the height of the main stages, as well as temperature and mass
distribution for three different groups of burners. The fourth group of burners is excluded until
there is emergency situation (Modlinski, 2010). It can be noticed that the burners distribute fuel
25
in such a manner that it results in lower temperature in the middle section of the combustion
O
chamber. From figure 9 we can see that this difference can be as high as 500 C.
Figure 7. Variation of temperature along boiler OP-430 boiler height (Rybak W., 2008).
The coal combustion boiler OP-430, relevant for this work, is a pulverized coal-fired boiler with
tangential burners layout. The burners’ construction allows delivering two streams (primary air
with fuel and secondary air) to the burning chamber simultaneously. Tangential layout of the
burners provides a helical shape to the whirling flame which makes a very stable flame (fig.9).
The efficiency of combustion can be controlled by the moving tilt of the burner axis. In Figure
10, an example of the temperature distribution of the burners is depicted. It can be noticed,
that the maximum value is expected at about 3 – 3.5 meters after the burner’s outlet.
(Tomeczek J. , 1994). During the operation, the boiler had the following parameters: output
steam pressure of 13.5 MPa, output steam temperature of 540ºC, maximum continuous load
of 120 kg/s and feed water temperature of 210ºC. Gross efficiency was equal to 90% (Rafako
S.A., 2014).
Temperature in the boiler can exceed 2000 ºC (fig.7) which is more than construction material
can resist. Therefore, the walls of the boiler must be cooled down by circulating water. As long
as it happens without any external work, it is called natural circulation. The heated water ends
up in the drum, from which it falls down along the boiler walls and then goes up and is heated.
After passing the walls, the water vapor is separated from liquid in a drum. Then wet steam
undergoes further heating to get rid of the existing moisture and finally becomes superheated
steam. The major components of the boiler are presented in Figure 11. Superheated steam is
then taken from the boiler and driven into a turbine. Afterwards, part of it is taken back from
the turbine and undergoes secondary superheating in the boiler (Laudyn, Pawlik, & Strzelczyk,
2000). In the boiler OP-430, air staging with utilization of OFA (Over Fire Air) nozzles is
applied. In the main burning zone (where traditional burners are installed) the combustion is
kept in rich conditions (λ<1), allowing the decrease of oxygen content, increasing the flame
26
temperature, as well as hydrocarbon radicals (CH) and therefore mitigates NOx emission.
Over the main burning zone, OFA nozzles deliver an excess of air and combustion conditions
become lean (λ>1) helping the burnout of the rest of the coal and char (Dziadula & Kosalka,
2013).
Figure 8. Boiler OP-430 scheme and height [m]
(Modlinski N., 2010)
Figure 9. Boiler OP-430 burners layout and
temperatures [OC] (Modlinski N., 2010)
Figure 10. Temperatures distribution after burner’s outlet( (Tomeczek J., 1994)
27
Figure 11. Boiler OP-430 with main components and temperatures of exhaust gases (EDF Poland)
Figure 11 represents a general scheme of the boiler OP-430 with its major components
responsible for producing superheated steam, which is used to produce electricity and hot
water for municipal heating system afterwards. The description of all elements and theoretical
temperatures for boiler OP-430 marked in figure 11, is presented below:
1 – Burners (4 burners on each level)
T1 – 1210 OC
2 – OFA nozzles
T2 – 905 C
3 – Wall heater
T3 – 815 C
4 – Ceiling superheater
T4 – 600 OC
5 – Bulkhead superheater
T5 – 375 C
6 – Final superheater
T6 – 130 C
O
O
O
O
7 – Side horizontal draft superheater
8 – Convection superheater
9 – Water heater
10 – Drum
28
8. Environmental concerns
Coal fired power plant produces significant amount of pollutants which is nowadays a global
concern. Small fly ash particles with size below 10 µm are especially dangerous for human’s
respiratory routes and lungs. They are divided into two categories – below 10 µm (index PM10)
and below 2.5 µm (index PM2.5). Fenelonov (Fenelonov, 2010) also points out, that very small
cenospheres end up in the PM10 category. To deal with particulate matter, power plants apply
utilities, such as ESPs (Electrostatic Precipitators) or fabric filters which, in principal, are very
mature technologies. However, despite the fact that ESP can reach efficiency above 99.5%,
small amounts of fine particles still goes to the atmosphere (Goodarzi, 2006). Other pollutants
released by coal fired power plant are nitrogen and sulfur compounds. Oxides of nitrogen
(NOx) are dangerous because they contribute to tropospheric ozone destruction and are
linked to acid rains, as well as photochemical smog (Zhu, Lu, Niu, Song, & Nia, 2009). To
reduce their influence several technologies have been developed, such as staged combustion,
low-NOx burners and secondary techniques like i.e. selective catalytic reduction (SCR). Sulfur
compounds are also very toxic and contribute to acid rains. This problem is addressed by
applying desulfurization units, or sorbents added to the combustion chamber in case of
fluidized-bed boilers.
European Union refers widely to environmental aspects of coal combustion (Fornea, 2011).
The document says that power plant should provide information about human health and
environmental impact of their technological processes. Moreover, it indicates the need to
further research and development to improve ash management and find new applications for
coal combustion by-products. The document also says that European Union should finance
innovative projects in this area. Cenospheres are mentioned as one of existing applications for
coal combustion by-product. Moreover, the document indicates, that more studies and
research is required to understand cenospheres’ composition, morphology and structure, as
well as new possibilities of their applications. These instructions confirm that cenospheres are
an important part of coal combustion by-products, which is especially important in the point of
view of this work.
29
II.
EXPERIMENTAL TECHNIQUES
1. Cenosphere separation
Laboratory work of cenospheres separation took place in the chemical laboratory of EDF
Krakow coal-fired power plant in August 2013. It was carried out in the frame of a student’s
internship, and the author was responsible for performing cenospheres separation from power
plant’s fly ash using the wet method. Afterwards, all necessary measurements and
calculations were carried out to determine the cenospheres content in each sample. Fly ash
was collected twice a day – at 5:00 AM (symbol R) and at 4:30 PM (symbol P) – from all three
ESP (electrostatic precipitator) zones. Samples are described using symbols like in example
D00P for fly ash samples and D00Pcen for cenosphere sample obtained from it. The ESP is
located at the end of exhaust gases pathway - just before the stack – so all produced fly ash
ends up there. Gas is ionized by means of high voltage and then it sticks to the earthed and
alternatively charged plates. Particles are being removed from plates by supplying them with a
shaking motion. In total, 60 samples were obtained from two different boilers that are installed
in EDF Krakow – 30 samples from boiler OP-380 and 30 samples from boiler OP-430.
However, for the purpose of this work, only ten samples of fly ash (all from boiler OP-430 and
ESP zone 1) are being considered (except data in Table 3). This is because comparison
between samples obtained from the same boiler and the same ESP zone would be the most
relevant. Also zone I contains about 70 wt. % of all fly ash produced by power plant (EDF
Poland). All samples were collected within a considerably short period of time - in August
2013, which means that power plant’s operation conditions, such as load and the scale of
energy production were very similar for all samples. The average mass of each collected ash
sample was about 400g.
Figure 12. Stirring and sedimentation of samples.
Cenosphere separation was carried out by wet separation (fig.12) with demineralized water;
3
therefore, all recovered particles have density lower than 1g/cm . The remaining parts were
30
poured with water, stirred for 10 minutes and left for sedimentation during one hour. Portions
of fly ash from each sample were stored for further analysis. Further on, floating particles were
skimmed and drained (Fig.13). The samples were then subjected to drying in an oven at
O
temperature of 105 C for 2 hours. Afterwards, weight of obtained matter was measured and
O
samples underwent a burnout procedure in a furnace, for 1.5 hour at temperature of 800 C
(Fig.14). Then the weight of each sample was measured again.
Figure 13. Draining of skimmed matter.
Figure 14. Dried samples prepared for burnout.
3
The ratio between mass of sample with density lower than 1 g/cm – which is supposed to be
all cenospheres – and mass of the initial fly ash indicates the cenosphere fraction present in
31
each batch. This content of cenospheres in each fly ash sample was used in this work, to
figure out if there are any factors influencing this value. Also, the content of unburned char for
each sample was determined.
2. XRF
X-Ray Fluorescence is a non-destructive technique for chemical analysis that determines the
elements present in the samples. It takes advantage of the atomic structure of an atom.
Photons delivered to the atoms convey energy capable to unbound electrons from their orbits.
When this happens, electrons from higher energy level orbital drop to the formed holes, and
this is accompanied by releasing other photons. This fluorescence light has a distinctive X-ray
energy for a given element, because the energy of the emitted photon is equal to the energy
gap between two transition orbitals and this, in turn, is constant for a specific element.
Therefore, by knowing the wavelength of emitted X-ray light, it is possible to determine the
element it comes from. XRF is widely used to determine elementary and chemical analysis in
different areas, like archeology, geochemistry, material science or engineering. Counting the
number of photons per unit of time, also allows quantitative analysis of tested material.
X-Ray Fluorescence was performed for 10 fly ash samples. The samples were in the powder
state - not grinded. The analysis was carried out with Philips PW1480 spectrometer, with three
series of measurement for each sample, with angle ranges of:
O
O
1) 50 – 140 - corresponding to the analysis of Fe, Ca and K,
2) 21O – 28.2O - corresponding to the analysis of Mg and Na, and
O
O
3) 59 – 146 - corresponding to the analysis of Si and Al.
3. X-Ray Diffraction (XRD)
X-Ray Diffraction (XRD) is a technique that yields information regarding the arrangement of
atoms and is primarily used for phase identification of a crystalline material. It was the first
method to describe the periodic atomic structure in crystals. XRD is an example of wave
interference and it is based on the Bragg’s law, which correlates distance between atomic
layers in crystals, wavelength of X-ray beams and the angle of incidence of reflected beam.
Angle Theta (Θ) indicates difference in beams reflection and therefore it allows finding out
different crystalline forms. In detail, XRD is able to measure spacing between atomic planes,
orientation of single crystals or grains, determine crystal structure of given material and
measure shape, size and internal stress of small crystalline regions. As a result of XRD tests,
patterns with specific curve shape and number of sharp peaks are obtained. The lack of
distinctive reflections and the smooth, broad band shape of the curve indicate an amorphous
structure while reflections points to the presence of crystalline phases.
32
XRD analysis was performed for 10 samples of cenospheres using a PANanalytical
diffractometer, with CuKα radiation (1.54060 Å), generated at 40 kV and 35 mA, at room
temperature, with a step of 0.05 and a scan step time of 200s.
4. Scanning Electron Microscopy (SEM) and Electron Dispersive Spectroscopy (EDS)
Scanning Electron Microscopy is a method which uses a focused beam of electrons to perform
a very precise and narrow scanning of the sample. The interaction of electrons with the
examined surface results in several different signals. The most widely used are secondary
electrons (due to inelastic scattering), which can give information about sample’s surface
topography. Other signals used are back-scattered electrons (due to elastic scattering) and
characteristic X-rays (resulting from electron ejection from the inner shell of the atom). The
former can give information about distribution of the elements in the sample, while the latter
can provide the composition and quantitative analysis of elements in a given sample. To
analyze information from characteristic X-rays, electron-dispersive X-ray spectroscopy (EDS)
analytical technique is used. In SEM, magnification can be 10 to 500.000 times, allowing a
very precise analysis of the tested material structure. SEM has got many applications in
branches such as forensic science, material science, biology, art, geology and medical
science.
SEM and EDS were performed for 10 samples of cenospheres and 4 samples of fly ash. The
analysis was carried out in a laboratory of Instituto Superior Tecnico in Lisbon, using a Bruker
Nano GmBH microscope, equipped with a XFlash 5010 detector. The measurements were
performed at 20 kV. Since the samples are ceramic, they were covered with a thin film of a
gold alloy, to make them conductive. Therefore, some gold content can be observed in EDS
spectra.
5. Raman spectroscopy
Raman spectroscopy is based on inelastic scattering of the laser light. Photons are absorbed
by targeted molecules and reemitted afterwards. Because scattering is inelastic, reemitted
photons may have different frequencies. If the targeted molecule is in excited state, photons
will have lower frequencies which is called Stokes shift. If the targeted molecule is in the less
energetic state than initial, reemitted photons will have higher frequencies which is called antiStokes shift. Finally, it may happen that frequencies of reemitted photons remain the same
and this is called Rayleigh scattering. The shift gives information about rotational and
vibrational modes in the system, which is specific for a given chemical bond or molecule. This,
in turn, allows to identify given molecules and provide reliable non-destructive chemical
composition and structure of given sample. Raman spectroscopy has plenty of applications in
variety of fields, from art to biology and chemical engineering.
Raman spectroscopy was performed on two different cenosphere samples: one with very low
yield of cenospheres and the other with the highest yield.
33
Raman spectra were collected using a LabRAM HR Evolution Confocal Microscope (Horiba
Scientific) with 532 nm excitation and a 100x objective lens (NA = 0.9). The laser power on the
samples was ~10mW. The collected Raman radiation was dispersed with a 600 lines/mm
grating and focused on a Peltier-cooled (-70º) charge-coupled device (CCD) detector allowing
-1
-1
a spectral resolution of ca. 4 cm . All spectra were recorded in the 100–4000 cm range with
an integration time of 10 s and 3 accumulations per spectrum.
34
III.
RESULTS & DISCUSSION
1. Results from cenospheres separation
In Table 2 data from laboratory work are presented. According to the particular steps of the
cenosphere separation process, Table 2 presents the mass of the tested fly ash sample, mass
3
of the skimmed matter with ρ < 1g/cm after drying and after burnout as well as percentage
3
difference between them. Finally, there is a column with a fraction with ρ <1g/cm content in
percentage of initial fly ash mass. The latter parameter is the most interesting in point of view
of this work, as it allows differentiating samples and looking for other related parameters.
3
Figure 15 illustrates that difference, where the highest content of fraction ρ < 1g/cm – which is
supposed to be comprised only of cenospheres – was obtained in sample 10.
Table 2. Results from cenospheres separation for ESP zone I.
No.
Symbol
Hour
R- 5:00,
P-16:30
Date
Mass of
ash sample
(g)
Dry sample
(<1g/cm3)
mass
Burnout
sample
(<1g/cm3)
mass
unburned
fraction
[%]
Fraction <1g/cm3
content [%]
6.26
0.2
1.35
3.7
0.3
2.31
7.02
0.4
1.49
1.44
3.22
0.4
361.04
1.63
1.55
5.16
0.4
306.13
1.45
1.4
3.59
0.5
R
526.04
3.21
3.19
0.59
0.6
R
447.23
3.91
3.57
8.67
0.8
1.09
P
454.32
4.92
4.72
4.08
0.1
28.08
R
287.08
5.77
5.52
4.3
1.9
1
D00P
25.08
P
2
D04P
29.08
P
3
D01P
26.08
P
4
D02R
27.08
R
5
D02P
27.08
6
D05R
30.08
7
D01R
8
D04R
9
10
1.47
1.38
396.8
1.4
538.23
2.48
383.15
P
R
26.08
29.08
D07P
D03R
[%]
582.37
Fraction <1g/cm3 content [%] from ESP zone I
2
1.8
1.6
1.4
1.2
1
0.8
0.6
0.4
0.2
0
1
2
3
4
5
6
7
8
9
10
Figure 15. Cenospheres content in each sample of tested fly ash from ESP zone I
Significant values are also present in samples 8 and 9. All the others are not satisfactory,
regarding hypothetical cenospheres production by the power plant. However, this particular
35
distribution can be useful for research, because it allows studying other factors related with the
samples that can promote cenosphere formation. As mentioned before, ten samples from ESP
zone I (presented in Table 2) were chosen for further investigation, nevertheless the author
briefly presents outcomes for samples collected from all ESP zones (all from boiler OP-430).
In Table 3 there are results for 30 samples including those presented in Table 2. Distribution of
3
the fraction ρ < 1g/cm can be observed in Figure 16. It is noticeable, that the percentage
content of cenospheres is almost always the highest in the first zone. At the same time, the
result for zone III in sample no. 6 (fig.16) seems to be a measurement mistake.
Table 3. Results from laboratory work for ESP zones I, II and III.
Symbol
ESP zone
Date
Mass of
ash sample
(g)
Dry sample
(<1g/cm3 )
mass
Burnout
sample
(<1g/cm3)
mass
unburned
fraction
[%]
Fraction
<1g/cm3
content
[%]
D00P
I
25.08
582.37
1.47
1.378
6.26
0.20%
D00P
II
25.08
530.21
0.48
0.434
9.58
0.10%
D00P
III
25.08
647.47
0.38
0.346
8.95
0.10%
D01R
I
26.08
526.04
3.21
3.191
0.59
0.60%
D01R
II
26.08
506.89
2.32
2.118
8.71
0.40%
D01R
III
26.08
471.79
1.32
1.269
3.86
0.30%
D01P
I
26.08
538.23
2.48
2.306
7.02
0.40%
D01P
II
26.08
458.59
1.8
1.706
5.22
0.40%
D01P
III
26.08
430.33
1.5
1.422
5.2
0.30%
D02R
I
27.08
383.15
1.49
1.442
3.22
0.40%
D02R
II
27.08
427.85
1.49
1.438
3.49
0.30%
D02R
III
27.08
489.17
1.04
0.794
23.65
0.20%
D02P
I
27.08
361.04
1.629
1.545
5.16
0.40%
D02P
II
27.08
388.38
1.534
1.456
5.08
0.40%
D02P
III
27.08
341.09
0.069
0.037
46.38
0.00%
D03R
I
28.08
287.08
5.767
5.519
4.3
1.90%
D03R
II
28.08
243.61
0.6655
0.629
5.48
0.30%
D03R
III
28.08
457.32
7.457
7.353
1.39
1.60%
D04R
I
29.08
447.23
3.9131
3.574
8.67
0.80%
D04R
II
29.08
441.3
1.5588
1.432
8.13
0.30%
D04R
III
29.08
426.29
0.8795
0.812
7.67
0.20%
D04P
I
29.08
396.8
1.3998
1.348
3.7
0.30%
D04P
II
29.08
352.51
1.7029
1.651
3.05
0.50%
D04P
III
29.08
413.24
2.148
2.103
2.09
0.50%
D05R
I
30.08
306.13
1.449
1.397
3.59
0.50%
D05R
II
30.08
290.3
1.5928
1.535
3.63
0.50%
D05R
III
30.08
300.63
0.8796
0.837
4.84
0.30%
D07P
I
1.09
454.32
4.9158
4.715
4.08
1.00%
D07P
II
1.09
462.49
4.8013
4.688
2.36
1.00%
D07P
III
1.09
342.02
2.9892
2.923
2.21
0.90%
Average
content of
fraction
<1g/cm3
0.17%
0.53%
0.39%
0.36%
0.36%
1.55%
0.64%
0.36%
0.48%
0.99%
36
Fraction <1g/cm3 content [%] from all ESP zones
2.00%
1.50%
ESP zone I
1.00%
ESP zone II
ESP zone III
0.50%
0.00%
D00P
D01R
D01P
D02R
D02P
D03R
D04R
D04P
D05R
D07P
Sample
Figure 16. Cenospheres content In each sample of tested fly ash from ESP zones I, II and III
3
In Table 3 there is also a column labeled average content of fraction ρ < 1g/cm . This is the
average percentage content calculated for all three ESP zones, in respect to the typical fly ash
distribution of the considered ESP zones – 70% of ash mass land up in the zone I, 20 % in the
zone II and 10% in the zone III. Despite the fact, that these three samples are different, they
were collected at the same time, thus this parameter can give an idea of the approximate
overall cenosphere yield for this particular power plant (and boiler OP430), if it considers their
separation. The average cenosphere content turns out to be 0.58% which is not an impressive
value, regarding that in different power plants, it can reach 1.5% (Fomenko, 2011). Figures 1726 depict SEM images of each of the ten studied cenosphere samples. As can be easily
noticed, they are quite different in terms of particle size, shape, distribution and color. In Figure
18 plenty of black particles having random shape are conditionally present. They look like
unburned char particles. If so, it would indicate that the burnout process was not efficient in
this case and thus results for this particular sample can be distorted. What can be clearly
noticed however is the difference in transparency between spheres. Some of them look like
fully transparent glassy balls; some are semi-transparent, while others are opaque. Also the
smaller spheres are usually more transparent than the larger ones. Regarding sample color,
some cenospheres are yellow, orange and brown while others are in different hue of white, but
no relevant correlation is observed with cenosphere yield. Different colors originate from trace
elements like Fe, Ti or Cu incorporated in the glass phase (Vassilev S.T., 2004). Considering
the size, the smallest particles that can be measured have the diameter of about 30 µm while
the largest are about 350 µm. Size distributions are also different in each sample. For
example, in samples depicted in fig. 20-22 there are plenty of small particles whereas in fig.
23-24 and fig. 26 they seem to be, on average, much larger. Most spheres are round-shaped;
nevertheless some opaque particles are slightly deformed. Transparent particles are clearly
hollow inside, but in case of opaque particles it is not possible to state that. Some of those
particles have other cenospheres trapped inside and then, they are called plerosphers.
37
1mm
Figure 17. SEM image of cenosphere sample D00Pcen
1mm
Figure 19. SEM image of cenosphere sample D01Rcen
1mm
1mm
1mm
1mm
1mm
1mm
38
1mm
1mm
2. XRF
The purpose of XRF analysis was to determine which elements can be found in fly ash, as
well as to find any possible correlations between them and the content of cenospheres in each
of ten samples of fly ash. Because precise quantification was not possible, results were
treated by comparison between samples and not as absolute values. Figures 27 and 28
present the XRF results for two different fly ash samples. Horizontal axis represents inclination
angle and vertical axis represents peak intensity in kcps (kilo counts per second). The results
of the analysis are summarized in Table 4, where all the elements are presented in the form of
a ratio to silica. This makes the results and comparisons between them more reliable and
accurate. Furthermore, the samples presented in this table are organized by growing content
3
of cenospheres (fraction ρ < 1g/cm ) so that any correlations between this parameter and the
content of a given element can be clearly observed. For each peak presented in Figures 27
and 28, its height was obtained after background subtraction. Then all values were compared
between each other in all feasible configurations to find out any correlations.
From the XRF results, there is a clear correlation between the content of cenospheres and the
amount of sodium. A correlation with calcium content can also be observed, but it is not so
strong. However, the best correlation observed – in terms of level of deviations and curve
slope – is the one combining the content of sodium and calcium altogether. It is given as the
last column in Table 4, as the ratio between sodium and calcium (Na/Ca). Graphical
description of this data is given in Figure 29 along with regression line and respective
equation. A general correlation seems to occur, where the amount of cenospheres increases
with the Na/Ca ratio. The samples with the most satisfactory results have higher Na/Ca ratio,
which means more sodium and less calcium content can be found in that samples. From
Figure 29 it can be noticed, that three samples with highest cenosphere content have
significantly higher Na/Ca ratio than others, but sample D04R is not well fitted to the
regression line. The same is true for sample D02P however, correlation is still quite well visible
in Figure 29.
39
Figure 27. Plots from XRF analysis for fly ash sample D00P
40
Figure 28. Plots from XRF analysis for fly ash sample D03R
41
Once the correlation of sodium and calcium content for cenosphere formation was found,
author tried to understand the nature of this phenomenon. According to Vassilev (Vassilev S.
T., 2004), the presence of chlorides can act as a catalyst for cenosphere formation. The
presence of chlorine and fluorine in an aluminosilicate peraluminous melt can increase its
viscosity - peraluminous melt occur when the fly ash molar content of Al 2O3 is higher than the
sum of Na2O and CaO (Baasner, Schmidt, & Webb, 2013). Chlorine alone can, in turn,
increase the viscosity of Na2O-Al2O3-SiO2 melts (Zimowa & Webb, 2007) and this can promote
glass formation (Shelby, 1997). This chlorine effect can give a clue, why melts containing
sodium present higher viscosity. What is more, chlorine can possibly enter the combustion
chamber and become part of the melt, as a component of sodium chloride (NaCl). We can
thus assume that sodium actually works as a chlorine carrier in fly ash. Another possibility of
chlorine effect on cenosphere formation is derived from their formation principles. Fenelonov
(Fenelonov, 2010) indicates that sodium chloride is one of possible volatizing components
which is responsible for cenosphere swelling during formation process. Tomeczek (Tomeczek
O
J. , 1994) in turn, points out, that the vaporizing temperature for NaCl is equal to 1465 C,
which is consistent with preferable cenosphere formation temperature figured out by Vassilev
(Vassilev S. T., 2004) (Karr, 1979).
0.0120
y = 0.0041x + 0.0024
0.0100
Ca / Na ratio
0.0080
0.0060
0.0040
0.0020
0.0000
0
0.2
0.4
0.6
0.8
1
1.2
1.4
Content of cenospheres in each sample [%]
1.6
1.8
2
Figure 29. Correlation between Na/Ca ratio in fly ash samples and content of cenospheres obtained from given sample.
42
Table 4. Data obtained from XRF – elementary analysis of fly ash samples in ratio of silica (Si).
Fraction <1g
Sample
Si / Al
Si / Ca
Si / Na
Si / Mg
Si / K
cm3 [%]
D00P
D04P
D01P
D02P
D02R
D05R
D01R
D04R
D07P
D03R
0.2
0.3
0.4
0.4
0.4
0.5
0.6
0.8
1
1.9
1.463
1.344
1.394
1.376
1.392
1.381
1.423
1.548
1.646
1.471
0.761
0.758
0.850
0.461
0.605
0.754
0.949
0.955
0.679
0.842
186.150
339.156
249.818
87.995
146.416
221.599
233.575
115.619
104.121
85.690
17.586
18.596
19.430
13.204
15.690
18.090
20.400
18.884
16.939
19.052
1.614
1.585
1.593
1.806
1.689
1.626
1.437
1.623
1.504
1.558
Si / Fe
0.105
0.112
0.116
0.110
0.113
0.125
0.140
0.119
0.109
0.135
Na / Ca
x 1000
4.088
2.235
3.402
5.239
4.132
3.403
4.063
8.260
6.521
9.826
3. XRD
Figures 30-39 present XRD patterns for the ten cenosphere samples The broad ovoid shape
of the background indicates the presence of an amorphous phase, while the sharp reflections
are connected with the presence of crystalline phases. From the crystalline peaks, quartz and
mullite can be identified in all samples. In the diffraction patterns of Figures 30-31 and 36-39
calcite (CaCO3) is also presented. Because carbonates, such as calcite cannot be found at
high temperature conditions – calcite is fully decomposed below 1400
O
C (Tomeczek &
Palugniok, 2002) – its occurrence could be an effect of later crystallization. The results
obtained – presence of quartz, mullite and calcite - are consistent with the composition of
cenospheres described by Vassilev (Vassilev S. T., 2004).
Figure 30. XRD pattern for cenosphere sample D00Pcen
43
Figure 32. XRD pattern for cenosphere sample D01Rcen
44
45
46
47
4. SEM / EDS
The purpose of performing SEM was to take a detailed look into the morphology, shape and
size of the examined material. In contrast to the microscope pictures presented in Chapter 1
(Results from cenospheres separation), SEM images have higher magnification and allow
detailed differences between obtained particles. EDS elementary analysis was performed to
study any significant difference between the samples, as well as between specific particles.
The results of this analysis are presented in Figures 40-60, as well as in Table 5. Moreover,
the data from EDS were used to determine the average composition of cenospheres and to
find out, if the previously found correlation between the content of sodium and calcium with the
content of cenospheres in a given sample is also observed. The data are presented in Tables
6-9. Figures 40-52 depict SEM images and the corresponding EDS spectra for the cenosphere
samples studied, and additionally, 3 samples of fly ash. The EDS spectra were taken on the
red rectangle area of the presented pictures. The corresponding elementary analysis results
are presented in Tables 6-8. The figures illustrate the differences between cenosphere
samples and fly ash samples in terms of particle shape, distribution and size. Because all the
images have the same magnification, differences in size between different particles can be
observed. From the images it may be seen, that there are predominantly small cenospheres
with 50-100 µm in the D02Pcen sample (Figure 45) while much larger particles are observed
in sample D03Rcen (Figure 46). Figures 51 and 52 disclose the difference in size between
cenospheres and average fly ash particles – the latter are generally much smaller. Among
other fly ash particles, hardly few cenospheres can be seen due to their generally low content
in fly ash.
48
0.3 mm
Figure 40. SEM picture and EDS spectrum of cenosphere sample D00Pcen
0.3 mm
Figure 41. SEM picture and EDS spectrum of cenosphere sample D01Rcen
0.3 mm
0.3 mm
49
0.3 mm
Figure 44. SEM picture and EDS spectrum of fly ash sample D02R
0.3 mm
0.3 mm
0.3 mm
Figure 47 SEM picture and EDS spectrum of fly ash sample D03R
50
0.3 mm
0.3 mm
0.3 mm
0.3 mm
Figure 51 SEM picture and EDS spectrum of cenosphere sample D07Pcen
51
0.3 mm
Figure 52. SEM picture and EDS spectrum of fly ash sample D07P
Table 5 and Figures 53-60 present SEM images and EDS spectra focused on specific
particles chosen from different samples. This comparison can give information about any
important differences in composition between particles with different sizes and shapes.
Carbon content, visible in all samples comes from the environment; however, in Figure 57, the
particle on the left noticeably contains a higher amount. This particle also has a weird shape
that allows classifying it as unburned char. Different shapes and shell structures can also be
seen for the particles in Figures 53 and 58. However, the carbon content is on the expected
level here. Cenospheres depicted in Figures 54, 56 (upper one), 59 and 60 (bottom one)
reveal a significantly content of potassium and iron. In Figure 55 there are two particles which
seem to be similar in structure but have very different shape. The bigger one contains no
calcium at all whereas it clearly exists in the smaller particle. There is also a significantly
higher content of magnesium, sodium and iron in the bigger particle, but the small one
contains additional elements like titanium and phosphorous.
Figure 60 depicts two similar particles in terms of structure and size. However, the upper one
contains some additional material attached to its shell. From elementary analysis it can be
seen that calcium has a significantly higher content in the upper particle. Therefore we can
assume that this residue might be a crystalline calcite, which would be consistent with the
results from XRD and Raman, as well as with the results presented by Vassilev (Vassilev S.
T., 2004). The same residue can be also noticed in Figure 59. Furthermore, in Figure 60 a rare
broken cenosphere can be seen. This image confirms that the wall is indeed very thin and the
particle is hollow inside. What is also worth to notice is that among all the SEM images that
were done for this work, this is the only broken cenosphere observed. This also confirms their
high mechanical strength.
52
Table 5. SEM pictures and element content in single cenospheres particles (The red rectangle in the inset images shows
the selected EDS inspection area)
Element
Carbon
Oxygen
Silicon
Aluminium
Iron
Calcium
Potassium
Sodium
Magnesium
Titanium
[norm. wt.%]
8.98
44.47
18.91
15.70
5.95
1.60
2.27
0.61
1.19
0.33
[norm. at.%]
14.74
54.84
13.28
11.48
2.10
0.79
1.14
0.52
0.96
0.14
Element
Carbon
Oxygen
Silicon
Aluminium
Iron
Calcium
Potassium
Sodium
Magnesium
[norm. wt.%]
8.47
34.22
23.20
19.24
5.11
0.50
8.10
0.23
0.93
[norm. at.%]
14.87
45.10
17.42
15.03
1.93
0.26
4.37
0.21
0.81
Figure 53. Cenosphere SEM image relative to EDS data
Element
Carbon
Oxygen
Silicon
Aluminium
Iron
Potassium
Sodium
Magnesium
Element
Carbon
Oxygen
Silicon
Aluminium
Iron
Calcium
Potassium
Sodium
Magnesium
Titanium
Phosphorus
[norm. wt.%]
6.51
46.89
22.04
14.82
4.11
2.36
1.38
1.90
[norm. wt.%]
9.52
40.44
19.07
21.77
2.28
1.00
2.40
0.63
0.52
0.91
1.47
[norm. at.%]
10.66
57.71
15.45
10.82
1.45
1.19
1.18
1.54
[norm. at.%]
15.70
50.07
13.45
15.98
0.81
0.49
1.22
0.54
0.42
0.37
0.94
53
Element
Silicon
Aluminium
Iron
Calcium
Potassium
Sodium
Magnesium
Titanium
Element
Carbon
Oxygen
Silicon
Aluminium
Iron
Calcium
Potassium
Sodium
Magnesium
Titanium
Element
Carbon
Oxygen
Silicon
Aluminium
Iron
Calcium
Potassium
Sodium
Magnesium
Element
Carbon
Oxygen
Silicon
Aluminium
Iron
Calcium
Potassium
Sodium
Magnesium
[norm. wt.%]
53.62
27.33
5.38
0.42
8.48
1.21
1.80
1.75
[norm. wt.%]
8.54
43.26
20.81
17.65
1.57
1.76
2.44
2.42
1.22
0.32
[norm. wt.%]
17.52
40.18
16.18
18.41
1.20
1.62
1.82
2.80
0.28
[norm. wt.%]
6.86
40.96
23.06
19.09
2.66
0.33
3.59
2.81
0.65
[norm. at.%]
56.00
29.71
2.83
0.31
6.36
1.54
2.17
1.08
[norm. at.%]
13.92
52.94
14.51
12.81
0.55
0.86
1.22
2.06
0.99
0.13
[norm. at.%]
26.66
45.91
10.53
12.48
0.39
0.74
0.85
2.23
0.21
[norm. at.%]
11.52
51.66
16.56
14.27
0.96
0.16
1.85
2.46
0.54
54
Element
Carbon
Oxygen
Silicon
Aluminium
Iron
Calcium
Potassium
Sodium
Magnesium
Calcium
Titanium
[norm. wt.%] [norm. at.%]
8.65
13.97
46.09
55.92
21.73
15.02
15.46
11.12
3.21
1.12
0.14
0.07
2.55
1.26
0.54
0.46
1.02
0.81
0.14
0.07
0.61
0.25
Element
Carbon
Oxygen
Silicon
Aluminium
Iron
Calcium
Potassium
Sodium
Magnesium
6.89
12.00
37.29
48.78
22.54
16.79
19.92
15.45
3.99
1.49
0.00
0.00
8.04
4.31
0.61
0.55
0.73
0.63
Element
Carbon
Oxygen
Silicon
Aluminium
Iron
Calcium
Potassium
Sodium
Magnesium
Titanium
Element
Carbon
Oxygen
Silicon
Aluminium
Iron
Calcium
Potassium
Sodium
Magnesium
11.22
17.55
48.16
56.55
16.83
11.26
13.79
9.60
1.81
0.61
3.92
1.84
2.10
1.01
0.77
0.63
1.05
0.81
0.33
0.13
8.22
13.68
41.36
51.68
21.49
15.30
18.76
13.90
1.75
0.63
0.54
0.27
6.31
3.23
0.62
0.54
0.94
0.77
55
Table 6 presents elementary analysis carried out for four fly ash samples. In particular there
are two samples which contained the highest amount of cenospheres (D03R and D07P) and
two samples with very small amount of cenospheres (D00P and D02R). For each sample,
2
EDS analysis was performed three times in different areas of 1mm size and afterwards, the
average values were calculated. Tables 7 and 8, in turn, present elementary analysis, also
obtained by EDS, for cenosphere samples, in wt.% and at.%, respectively. The oxide
composition of cenospheres presented in Table 9 was calculated from the results presented in
Table 8, while oxide composition of fly ash presented in Table 10 was calculated from the
results of Table 6 (norm. at.%). The composition regarding all cenosphere samples is between
49.49 – 60.45 % of SiO2, 23.20 – 28.30 % of Al2O3, 0.00 – 14.81 % of CaO, 1.61 – 7.83 % of
MgO, 1.77 – 5.43 % of Fe2O3, 2.04 – 3.62 % K2O, 0 – 4.83 % of Na2O and trace content of
TiO2. These results are quite interesting, because it is not common for aluminosilicate glasses
to have such a high content of alumina. This, in turn, implies very good mechanical strength. A
similar composition can be found in S-type glass (65% of SiO2 and 25% of Al 2O3) which has a
tensile strength of about 4.48 GPa and an elasticity modulus of about 85.4 GPa, making this
glass very resistant and expensive at the same time (Smith, 1995). This fact, along with
cenospheres low density, helps to understand, why cenospheres are so interesting.
Table 6. EDS quantitative analysis of elements from fly ash samples.
D00P
Element
Carbon
Oxygen
Silicon
Aluminium
Iron
Calcium
Potassium
Magnesium
Sodium
Titanium
D02R
D03R
D07P
D00P
D02R
[norm. wt.%]
39.587
36.220
8.900
6.822
4.754
1.853
0.844
0.769
0.251
0.000
40.268
38.544
7.774
6.435
3.201
1.711
0.618
0.899
0.359
0.192
D07P
D03R
[norm. at.%]
40.155
38.635
8.169
6.474
3.073
1.016
0.877
0.733
0.770
0.099
21.767
43.922
14.946
10.136
3.853
1.113
1.465
0.990
1.808
0.000
51.929
35.901
5.036
4.020
1.352
0.736
0.343
0.504
0.181
0.000
51.809
37.461
4.319
3.722
0.895
0.662
0.248
0.577
0.247
0.061
51.755
37.410
4.506
3.719
0.851
0.392
0.347
0.467
0.520
0.032
30.866
48.287
9.618
6.728
1.190
0.479
0.691
0.708
1.434
0.000
Table 7. EDS quantitative analysis of elements from cenosphere samples in [wt.%].
Sample
Al wt%
Ca wt%
C wt%
Fe wt%
Mg wt%
O2 wt%
K wt%
Si wt%
Na wt%
Ti wt%
D00Pcen
5.804
2.752
44.884
2.307
0.287
36.074
1.092
6.708
0.092
0.000
D01Rcen
12.043
0.000
23.365
5.027
0.478
40.883
2.448
14.678
0.434
0.644
D01Pcen
5.880
1.316
43.535
0.764
0.733
41.587
0.613
5.573
0.000
0.000
D02Rcen
9.140
0.287
30.503
2.021
0.746
44.455
1.407
10.607
0.459
0.373
D02Pcen
7.237
1.340
40.310
1.801
0.203
38.882
1.356
7.742
1.131
0.000
D03Rcen
8.364
0.071
34.106
1.218
0.876
43.654
0.978
9.192
1.177
0.363
D04Rcen
6.496
1.124
47.284
2.833
0.397
33.377
0.874
7.194
0.421
0.000
D04Pcen
9.518
1.511
31.634
1.551
0.254
42.772
1.686
10.591
0.483
0.000
D05Rcen
7.989
3.470
30.864
1.310
0.481
45.826
1.026
8.188
0.846
0.000
D07Pcen
8.570
0.864
32.663
1.538
0.711
44.004
0.995
9.043
1.610
0.000
56
Table 8. EDS quantitative analysis of elements from cenosphere samples in [at.%].
Sample
Al at%
Ca at%
C at%
Fe at%
Mg at%
O2 at%
K at%
Si at%
Na at%
Ti at%
D00Pcen
3.259
1.040
56.626
0.626
0.179
34.166
0.423
3.619
0.061
0.000
D01Rcen
7.866
0.000
34.283
1.586
0.347
45.034
1.103
9.211
0.332
0.237
D01Pcen
3.237
0.488
53.837
0.203
0.448
38.608
0.233
2.947
0.000
0.000
D02Rcen
5.488
0.116
41.145
0.586
0.497
45.016
0.583
6.119
0.323
0.126
D02Pcen
4.134
0.515
51.728
0.497
0.128
37.457
0.535
4.249
0.758
0.000
D03Rcen
4.883
0.028
44.727
0.344
0.568
42.976
0.394
5.155
0.807
0.119
D04Rcen
3.617
0.421
59.150
0.762
0.246
31.344
0.336
3.849
0.275
0.000
D04Pcen
5.711
0.610
42.638
0.449
0.169
43.279
0.698
6.105
0.340
0.000
D05Rcen
4.765
1.393
41.350
0.377
0.318
46.091
0.422
4.692
0.592
0.000
D07Pcen
5.054
0.343
43.280
0.438
0.466
43.772
0.404
5.124
1.115
0.000
Table 9. Oxide composition of cenosphere samples in [%]
Sample
Fraction
<1g/cm3
content [%]
Al2O3
CaO
Fe2O3
MgO
K2O
SiO2
Na2O
TiO2
D00Pcen
D04Pcen
D01Pcen
D02Rcen
D02Pcen
D05Rcen
D01Rcen
D04Rcen
D07Pcen
D03Rcen
0.2
0.3
0.4
0.4
0.4
0.5
0.6
0.8
1
1.9
23.20
27.24
28.30
26.52
26.32
25.13
25.81
25.80
25.13
26.88
14.81
5.82
8.53
1.12
6.56
14.69
0.00
6.00
14.69
0.31
4.46
2.14
1.77
2.83
3.16
1.99
5.20
5.43
1.99
1.89
2.55
1.61
7.83
4.80
1.63
3.35
2.28
3.51
3.35
6.25
3.01
3.33
2.04
2.82
3.41
2.23
3.62
2.40
2.23
2.17
51.53
58.24
51.53
59.13
54.10
49.49
60.45
54.90
49.49
56.75
0.43
1.62
0.00
1.56
4.83
3.12
1.09
1.96
3.12
4.44
0.00
0.00
0.00
1.22
0.00
0.00
1.56
0.00
0.00
1.31
MgO
K2O
SiO2
Na2O
TiO2
Table 10. Oxide composition of fly ash samples in [%]
Fraction
Sample
Al2O3
CaO
Fe2O3
<1g/cm3 [%]
D00P
D02R
D07P
D03R
0.2
0.4
1
1.9
21.79
22.76
21.26
22.91
7.98
8.10
3.03
4.83
7.33
5.47
3.76
5.24
5.46
7.06
4.47
5.75
1.86
1.52
2.18
2.14
54.60
52.83
60.77
55.52
0.98
1.51
4.53
3.20
0.00
0.75
0.00
0.39
The EDS results are in good agreement with the XRF data, since the same elements were
determined in all tested samples. These elements are Si, Al, Ca, Na, Fe, K, Mg and
occasionally Ti. Carbon, as it was mentioned before, is present as well, but it occurs as a
natural background coming from the environment and from some unburned particles. Again,
EDS data point to a clear correlation of Na and Ca content with the amount of cenospheres
obtained. From Table 6 - containing elementary analysis, made by EDS, of four fly ash
samples - some interesting conclusions can be derived. Content of sodium is higher in
samples with the higher amount of cenospheres while the content of calcium is lower. This
confirms previous assumption about sodium and calcium influence for cenospheres
57
abundance in fly ash. Similar results were also obtained by Itskos (Itskos, 2010), who reported
that bigger fly ash particles (which are mainly cenospheres) contain more sodium oxide and
less calcium oxide than the smaller ones. However, when we take a look at cenosphere
composition (Tables 7 and 8), these relations are not so clear observed. There is still clear
difference in content of sodium between edge (in terms of content of cenospheres) samples,
but there are high peaks for sodium in the middle of the stake as well. This distorts the overall
correlation. For calcium it looks similar – two edge values show very clear high difference in
the content, but the values for other samples do not maintain expected correlation. Those
results could be possibly explained by the fact, that sodium and calcium can be incorporated
in cenospheres as the amorphous phase, which is consistent with results from XRD. Another
possible explanation could be that both components act more like catalysts or inhibitors – their
existence during the combustion process is influential but they do not become a part of the
cenosphere structure.
5. Raman spectroscopy
Micro-Raman spectroscopy was performed on several particles taken from samples D00Pcen
(sample with low yield of cenospheres) and D03Rcen (sample with high yield of cenospheres).
As a result, pictures of the analyzed particles and corresponding Raman spectra are
presented in Figures 61-74. The studied particles are always spotted in the center of each
image. The analyzed area of each particle is of the order of 1 square micron. As can be
observed, different spectra were obtained for the same cenosphere samples, revealing a high
inhomogeneity of the samples. Samples are mainly amorphous, because there are few sharp
-1
peaks but plenty of broad bands instead. Broad bands at about 460 cm indicate silica glass
(Rruff Database of Raman spectra, 2014) and if they have a sharp finish, like in Figures 66
and 70, it indicates the presence of the crystalline quartz (Rruff Database of Raman spectra,
2014). The patterns presented in Figures 64 and 68 show bands due to carbon black (bands
at ~1330 and ~1590 cm-1). Both patterns are taken from sample D00Pcen, which is the
sample with very low yield of cenospheres. Raman spectroscopy confirms the occurrence of
quartz and calcite in the investigated samples (Rruff Database of Raman spectra, 2014). On
the other hand, the author was not able to confirm the presence of mullite, which was
previously indicated by XRD measurement. Instead, particles containing other minerals were
observed but could not be identified.
58
Sample D00Pcen A (figure 61)
Major peaks (figure 62):
1) Broad peak at ~467 cm
-1
– Silicate
glass
2) Peak at 1085 cm-1 - Calcite
40 µm
Figure 61. Sample D00Pcen fragment A
27000
25000
Intensity
23000
21000
19000
17000
15000
102
284
462
636
808
976
1,140
1,302
1,460
1,615
1,768
-1
Raman shift [cm ]
Figure 62. Raman spectrum for sample D00Pcen fragment A
Sample D00Pcen B (figure 63)
-1
1) Peak at 267 cm - crystalline Quartz
-1
2) Peak at 457 cm – crystalline Quartz
-1
3) Peak at 1083 cm - Calcite
-1
4) Peak at 1327 cm - Carbon black
-1
40 µm
Figure 63. Sample D00Pcen fragment B
59
20000
18000
Intensity
16000
14000
12000
10000
8000
6000
102
284
462
636
808
976
1140
1302
1460
1615
1768
Figure 64. Raman spectrum for sample D00Pcen fragment B
Sample D00Pcen C (figure 65)
1) Peak at 457 cm-1 – crystalline Quartz
-1
40 µm
Figure 65. Sample D00Pcen fragment C
8000
7000
Intensity
6000
5000
4000
3000
2000
1000
102
284
462
636
808
976
1140
1302
1460
1615
1768
Raman shift [cm-1]
Figure 66. Raman spectrum for sample D00Pcen fragment C
60
Sample D00Pcen D (figure 67)
-1
1) Peak at 267 cm - crystalline Quartz
-1
3) Peak at 1342 cm-1 - Carbon black
-1
40 µm
Figure 67. Sample D00Pcen fragment D
25000
23000
21000
Intensity
19000
17000
15000
13000
11000
9000
7000
102
284
462
636
808
976
1140
1302
1460
1615
1768
Figure 68. Raman spectrum for sample D00Pcen fragment D
Sample D03Rcen A (figure 69)
1) Broad band plus peak at ~ 458 cm-1 –
mixture of silicate glass with crystalline
Quartz.
40 µm
Figure 69. Sample D03Rcen fragment A
61
8000
7000
Intensity
6000
5000
4000
3000
2000
102
284
462
636
808
976
1,140
1,302
1,460
1,615
1,768
Figure 70. Raman spectrum for sample D03Rcen fragment A
Sample D03Rcen B (figure 71)
1) Broad band plus small peak at ~ 455
cm
-1
– mixture of silicate glass with
small amount of crystalline Quartz.
40 µm
Figure 71. Sample D03Rcen fragment B
62
12000
11000
10000
9000
Intensity
8000
7000
6000
5000
4000
3000
2000
102
284
462
636
808
976
1,140
1,302
1,460
1,615
1,768
Figure 72. Raman spectrum for sample D03Rcen fragment B
Sample D03Rcen C (figure 73)
-1
1) Broad band plus peak at ~ 458 cm –
mixture
of
silicate
glass
with
crystalline Quartz.
40 µm
Figure 73. Sample D03Rcen fragment C
25000
Intensity
20000
15000
10000
5000
0
102
284
462
636
808
976
1140
1302
1460
1615
1768
Figure 74. Raman spectrum for sample D03Rcen fragment C
63
6. Viscosity
Viscosity of fly ash was calculated using the data from Table 10 and equations [2.1]-[2.9]. The
results are presented in Table 11 (for three different temperatures) and in Figure 75. Oxide
compositions of four fly ash samples was calculated from EDS results, summarized in Table 6.
Viscosity is indeed higher for two of the samples with higher cenosphere content, as was
expected. However, the correlation is not linear because sample D07P reveals a viscosity
near twice as high as sample D03R, whereas regarding content of cenospheres, it is almost
the opposite. From Tables 10 and 11, it looks like Na2O and SiO2 content increases the melt
viscosity while CaO content, in opposite, brings about its decrease. This is expected only in
the case of CaO and SiO2, because, according to Shelby (Shelby J.E., 1997), sodium itself
causes a higher decrease in melt viscosity than calcium. Most probably the influence of
calcium is much more visible than the influence of sodium, because the former has
significantly higher content in the examined material (Kim & Sohn, 2012). Despite that, the
major component that influences viscosity is obviously silica, which is also widely used in
glass industry (Shelby, 1997). Also Vassilev (Vassilev S. T., 2004) points out, that a higher
content of quartz and cristobalite (both comprised of Si) can be found in samples with higher
cenosphere yield. The same is also visible in Table 11, however the correlation disclosed is
not linear. When looking at the figure 75, it seems that results for sample D07P are strongly
distorted by some unknown factors. It must be kept in mind, that viscosity calculations
presented below are just a rough approximation. Moreover, they are based on oxide
composition recalculated from EDS elementary analysis, which was performed on the limited
area of the sample and thus may be burdened with some inaccuracy. Summing up, there are
wide possibilities for deviations to occur in this analysis. Thus it would be beneficial to perform
precise laboratory viscosity measurements, since viscosity is an important factor regarding
cenosphere formation.
45000.00
40000.00
Viscosity [Pa s]
35000.00
30000.00
D00P
25000.00
D02R
20000.00
D03R
15000.00
D07P
10000.00
5000.00
1100
1140
1180
1220
1260
1300
1340
1380
1420
1460
1500
1540
1580
1620
1660
1700
1740
1780
1820
1860
1900
0.00
O
Temp. [ C]
Figure 75. Viscosity in function of temperature for four samples of fly ash with different cenospheres content.
64
Table 11. Comparison of viscosities for four fly ash samples for distinctive temperatures.
Sample:
Fraction <1g/cm3 [%]
Temperature [OC]
1000
1200
1500
D00P
D02R
0.2
43 132
2048
73
D07P
0.4
1
Viscosity [Pa s]
36 075
258 818
1687
7979
64
193
D03R
1.9
70 644
2866
93
7. Mineral matter transformation
To figure out, which crystalline phases can be found in the final product of transformation of
aluminosilicate material (here: fly ash), phase diagrams can be employed. The example of
binary diagram for system SiO2 – Al 2O3 is presented in Figure 76.
Figure 76. Binary phase diagram for SiO2 -Al2O3 system with relevant regions (Ernest M., 1985).
From oxide composition for cenospheres presented in Table 9, we know that SiO2 content
varies from 49.49% to 60.45% and content of Al2O3 varies from 23.2% to 28.3% depending on
the sample. Corresponding regions are depicted in Figure 76. The difference between the two
regions is related with the other oxides present in the composition. Nevertheless, as indicated,
two main minerals can occur in these compositional ranges: mullite and quartz (SiO2). This is
consistent with the results obtained in XRD.
It is also interesting to take a look at different phase diagrams, for example SiO2 – Al2O3 –
CaO system ternary diagram (Figure 77). Once again, the region corresponding to
cenosphere composition of SiO2 and Al 2O3, as given in Table 9, was depicted. According to
this graph, we could expect occurrence of Anorthite which would form at about 1600 OC
(eutectic point). However XRD analysis did not reveal the existence of this particular
crystalline form. The same is true for other possible minerals like Nepheline, Corundum,
65
Magnetite, Sillimanite, Ferrosillite, Brucite and Mayenite, which can be observed in other
ternary phase diagrams with the oxides present in the composition.
Eutectic point
for Anorthite:
CaO.Al2O3.SiO2
Figure 77. System SiO2 – Al2O3 – CaO on ternary phase diagram (Ernest M., 1985).
Regarding the temperature of crystallization, it varies from about 1000 OC to about 1600 OC for
different considered minerals, with exception of Quartz which can undergo crystallization
O
O
between 400 C –1000 C (Vassilev, Baxter, & Vassileva, 2013). XRD and Raman analysis
detected additionally occurrence of Calcite in some samples. Its formation, in turn, requires
Lime (CaO), which therefore must also exist in cenospheres. Lime crystallization temperature
O
O
can vary between 200 C – 1300 C (Vassilev, Baxter, & Vassileva, 2013). The melt during
quenching must slip over the mentioned temperatures very quickly, in order to avoid
crystallization process. According to Okada (Okada, 2008), in all glasses containing calcium,
O
glass transformation temperature Tg is always 100 – 200
C lower than crystallization
O
temperature, while for pure SiO2-Al2O3 glasses, this is only 50 C. From TGA and DTA curves
presented by Vassilev (Vassilev S. T., 2004) it may be seen that glass transformation
temperature for cenospheres is between Tg = 600 – 800
O
C. At the same time, Shelby
66
(Shelby, 1997) indicates that typical glass transformation temperature for soda-lime-silicate
O
glasses is close to 600 C. From figure 76 we see that mullite formation occurs at temperature
O
about 1600 C, but the addition of calcium (fig.77) can decrease this temperature down to
O
1400 C, which is in the expected region for cenosphere formation.
8. The analysis of combustion process
There are many parameters being measured by a power plant during its operation, such as
temperature and pressure distribution in the boiler, the amount of primary and secondary air
delivered to the combustion process, parameters of exhaust gases and many others. Figures
78-79 present temperatures of exhaust gases gauged by power plant. Figure 78 depicts
temperature measured just after bulkhead superheater on the left side, while figure 79 depicts
temperature measured after bulkhead superheater on the right side. Both temperatures are
significantly different due to the boiler asymmetry. The measurement time corresponds to the
time when cenospheres were being collected from the ESP.
As far as results are concerned, we can clearly see that the experimental points are divided
into two groups, which lay in two different temperature regions. The higher temperature in
most cases is correlated with lower cenosphere yield while the lower temperature is correlated
with a higher yield. As it can be seen in Figure 80, the change in temperature is directly
derived from the changing output power of the generator. Only the result for the sample
containing 0.8% of cenospheres is not consistent with the overall trend. Nevertheless,
obtained outcomes indicate that temperature in this region of boiler is relevant for cenosphere
formation. Moreover, this temperature region is suitable for the glass solidification process.
Thus, based on Figures 78-89 it may be concluded, that temperature after bulkhead
superheater should not exceed roughly 700 OC in order to favor cenosphere formation.
800
780
760
740
o
[ C] 720
700
680
660
0
0.5
[%]
1
1.5
2
Figure 78. Exhaust gases temperature after bulkhead superheater - left side vs. Yield of cenospheres [%].
67
850
800
750
[oC]
700
650
600
0
0.5
1
1.5
2
[%]
Figure 79 Exhaust gases temperature after bulkhead superheater - right side vs. Yield of cenospheres [%].
Temperatures presented in Figures 78-79 were measured in a significant distance from the
burning zone and, because they are correlated with the generated power, depend on the
amount of fuel delivered to the boiler. However, this parameter does not provide information
about temperature in the burning zone itself, because it depends mainly on the content of
oxygen delivered to the combustion process. Regarding that, author was not able to find any
reliable correlation between cenosphere yield and conditions in the burning zone.
Nevertheless, we can make a hypothesis that, in the burning zone, temperature favorable for
cenosphere formation should be high, in contrast with the upper part of the boiler. That is
mainly because the major component of glassy cenospheres – Silica – has a very high melting
temperature. Despite the fact, that this temperature is reduced by addition of modifiers such as
Na2O or CaO (Shelby, 1997), having a higher temperature in the burning zone should result in
more melted Silica. Finally, it should be stressed that the 10 samples investigated in this work,
is not enough to find more reliable correlations between combustion parameters and yield of
cenospheres, thus further research in this area is required.
68
[oC]
900
[MW]
140
800
120
700
100
600
500
80
400
60
300
40
200
20
100
0
Exhaust gases
temperature after
bulkhead superheater
- left side
Exhaust gases
temperature after
bulkhead superheater
- right side
Power curve [MW]
0
0.2
0.3
0.4
0.4
0.4
0.5
0.6
0.8
1
1.9
Cenospheres yield [%]
Figure 80. Exhaust gases temperatures after bulkhead superheater (left side and right side) and power
produced by the generator [MW], as a function of cenospheres yield [%]
As it was confirmed previously, the examined material has a mixed amorphous and crystalline
structure. From glass formation principles (Shelby J.E., 1997), it is known that, in order to
obtain an amorphous structure and avoid crystallization, the melt must be quenched quickly
enough. Glass transformation temperature for cenospheres should be lower than about Tg =
800
O
C (Vassilev S.T., 2004) (Shelby J.E., 1997). On the other hand, the range of
O
O
temperatures in the boiler is quite wide – from 130 C to almost 2100 C. Therefore, if we
O
could find a path where fly ash particles slip from temperature of about 1500 C - maximum
temperature for cenosphere formation (Karr C., 1979) - to temperatures below 800 OC, we
could expect high cenosphere formation in that region. From Figures 7, 8 and 11 it can be
noticed that along the boiler height, particles are subjected to very fast temperature change.
Particle residence time in a combustion chamber is lower than 2 s (Tomeczek J., 1994) and
during this time, they have to overcome the distance of about 30 m which results in quite fast
quenching indeed. From Figures 78 and 79, it may seen, that temperature of exhaust gas after
bulkhead superheater (fig.11) can be close to about 700 OC. Furthermore, the lower the
temperature in this region is, the more cenospheres are produced, which indicates that
particles can indeed be formed in this area. Another possibility is that fast quenching can also
occur very close to the boiler walls, which are comprised of pipes with water and temperatures
O
lower than 500 C. In this case we could say that most rapid change in temperature would
occur in the horizontal direction. However, this is also the main area of boiler slagging, and
thus many cenospheres can end up trapped in ash deposit that is sticked to the boiler walls.
Also, it should be kept in mind that cenospheres are not pure glass – they are also comprised
of different crystalline phases. Therefore, we cannot expect the ideal glass-forming conditions
inside the pulverized coal boiler.
69
IV.
CONCLUSIONS & FUTURE INVESTIGATION
1. Conclusions
Results disclosed a strong disproportion in the content of cenospheres among the different fly
ash samples. This gives the possibility to study them in order to find any factors, which could
promote the yield of cenospheres from coal combustion process. In terms of size and shape,
cenospheres present spherical shape with diameters from about 8 to 1000 µm. The smaller
particles are mostly transparent, while the larger particles are mostly opaque and have
different hues. The structure of cenospheres turned out to be mainly amorphous, with the
occurrence of crystalline phases such as Mullite and Quartz. This indicates the occurrence of
glass formation process inside the pulverized coal combustion boiler. High viscosity, required
in glass formation, comes from cenospheres major component – silica, while good mechanical
strength of cenospheres is a result of high alumina content. Cenospheres contained about 2527% of alumina, which is high in comparison to typical glass. From glass formation principles,
it is known, that to form an amorphous phase, fast quenching of the melt is a very important
factor. From the data obtained in work, it can be derived that cenospheres are, most probably,
formed in specific regions of the boiler, where temperature decreases rapidly from about 1500
O
O
C, to temperature below 800 C. In terms of boiler OP-430, this can presumably occur either
in horizontal position – along the furnace height, or in the vertical position – close to the boiler
walls. In the temperature zone higher than 1500 OC, gas pressure inside the cenosphereforming melt, can be too high to maintain spherical shape of the forming particle. Also,
O
temperatures lower than about 700 C in the region above bulkhead superheater, seem to be
favorable when cenosphere yield is concerned. Based on EDS and XRF data, a correlation
between cenosphere yield and the amount of sodium and calcium was observed. A higher
amount of calcium in some samples may lead to a decrease in viscosity, which is not
beneficial for the glass formation process. In these samples, the yield of cenospheres turned
out to be lower. It is more difficult to explain the positive influence of sodium. One possible
explanation is that sodium can act as a chlorine carrier (in sodium chloride NaCl) and chlorine
itself can increase melt viscosity under certain conditions. Another possibility is that chlorine
increases viscosity especially in the melts containing sodium, which in a consequence give a
correlation between sodium and the content of cenospheres in fly ash. Finally, sodium chloride
is a compound which can be released during melting process in the favorable temperature
range, and therefore it may play role in inflating cenospheres.
2. Future work
This investigation was meant to shed some light onto the cenosphere formation process
during pulverized coal combustion. However, this task is quite complex and should be an
object of further investigation. To do this, the author suggests performing laboratory
experiments of controlled coal combustion. Small amounts of fuel should be heated to different
O
temperatures from between T = 1400 – 1600 C and afterwards cooled down to temperature
70
O
below 800 C with different cooling rates. Coal chemical composition should be determined,
especially in terms of silicon, aluminum, sodium, calcium and sodium chloride content.
Chlorine alone can be added to melts with different content of sodium and calcium, to confirm
or exclude its influence. Then, from each experiment, cenospheres should be separated
3
(considering also particles with density between 1 – 2 g/cm ) and quantified. Furthermore, it
would be beneficial to perform laboratory high temperature viscosity measurements for all fly
ashes. Regarding cenospheres, it would be interesting to study their size distribution, using
dynamic light scattering and to study their porosity using BET.
For industrial cenosphere production to be considered by a power plant, the most important
parameter is cenosphere yield. In order to enhance it, power plant may employ statistical
approach like Taguchi methods. By changing one parameter and keep other fixed, this method
allows to find out the best production plan at comparatively low cost of the method itself. The
parameters that can be most interesting in case of cenospheres production, are :
-
The amount of air delivered to combustion process, because general temperature of
combustion depends on it. Furthermore, air distribution between main burners and
auxiliary nozzles (air staging) can be subjected to changes.
-
Fuel load, as it is the main factor responsible for the power output and, therefore,
temperature in the upper parts of the boiler.
-
Various types of coal with different composition. As this work stated, the content of silicon,
calcium or sodium influences cenospheres yield.
After each particular change, obviously the effect must be measured, which means
cenospheres have to be separated from the corresponding fly ash. Again, it would be
3
beneficial to make separation of all particles with density lower than 2 g/cm to obtain the real
content of produced cenospheres.
71
References
Ahmaruzzaman, M. (2010). A review on the utilization of fly ash. Progress in Energy and Combustion
Science 36 , 327-363.
Baasner, A., Schmidt, B., & Webb, S. L. (2013). The effect of chlorine, fluorine and water on the
viscosity of aluminosilicate melts. Chemical Geology 357 , 134-149.
Barbare, N., Shukla, A., & Bose, A. (2003). Uptake and loss of water in a cenosphere–concrete
composite material. Cement and Concrete Research 33 , pp. 1681 - 1686.
Birkholz, M. (2006). Principles of X-Ray Diffraction. In M.Birkholz, Thin Film Analysis by X-Ray
Scattering. Weinheim: WILEY-VCH Verlag GmbH & Co. KGaA.
Blissett R.S., (2012). A review of the multi-component utilisation of coal fly ash. Fuel 97 , pp. 1-23.
Chavez-Valdez, A. (2011). Ultra-low thermal conductivity thermal barrier coatings from recycled flyash cenospheres. Acta Materialia 59 , pp. 2556-2562.
Chen L., (2012). Oxy-fuel combustion of pulverized coal: Characterization, fundamentals, stabilization
and CFD modeling. Progress in Energy and Combustion Science 38 , pp. 156-214.
Coal Classification. (2014, 03 10). Retrieved from The Engineering ToolBox:
http://www.engineeringtoolbox.xom
Database of Raman spectroscopy. (2014, July). Retrieved July 2014, from Rruff.info: http://rruff.info
Dziadula, S., & Kosalka, J. (2013, January). Experience gained during modernization of 200 MW power
units (nr 2 and 4) in the electric power station Elektrownia Jaworzno III. Energetyka, 703 , pp. 12-14.
Ernest M. (1985). PHASE DIAGRAMS FOR CERAMISTS VOLUME I. Ohio: The American Ceramic Society.
Fenelonov V.B. (2010). The Properties of Cenospheres and the Mechanism of Their Formation During
High-Temperature Coal Combustion at Thermal Power Plans. KONA Powder and Particle Journal 28 ,
189-207.
Fleming D., (2012). Analysis on Characteristics of Fly Ash from Coal Fired Power Stations. Energy
Procedia 17 , pp. 3-9.
Fomenko, E. (2011). Fly Ash Cenospheres: Composition, Morphology, Structure, and Helium
Permeability. World of Coal Ash Conference. Denver.
Fornea, K. (2011). Opinion of the European Economic and Social Committee on ‘The processing and
exploitation, for economic and environmental purposes, of industrial and mining waste deposits in
the European Union’ (own-initiative opinion). Official Journal of the European Union C 24/11 .
Ghosal S. (1995). Particle size-density relation and cenosphere content of coal fly ash. Fuel 74 , pp.
522-529.
Goodarzi F. (2006). The rates of emissions of fine particles from some Canadian coal-fired power
plants. Fuel 85 , pp. 425-433.
72
Hirajima T. (2010). Recovery of cenospheres from coalfly ash using a dry separation process:
Separation estimation and potential application. International Jurnal of Mineral Processing 95 , pp.
18-24.
Hower J. (2012). Petrographic examination of coal-combustionfly ash. Coal geology 92 , pp. 90-97.
Itskos G. (2010). Size fraction characterization of highly-calcareous ﬂy ash. Fuel Processing
Technology 91 , 1558-1563.
Karr C. (1979). Analytical Methods for Coal and Coal Products. New York: Academic Press.
Kim, G., & Sohn, I. (2012). Effect of Al2O3on the viscosity and structure of calcium silicate-based
melts containing Na2O and CaF2. Journal of Non-Cystalline Solids 358 , 1530-1537.
Kolay P.K. (2001). Physical, chemical, mineralogical, and thermal properties of cenospheres from an
ash lagoon. Cement and Concrete Research 31 , 539-542.
Kolay, P. K., & Bhusal, S. (2014). Recovery of hollow spherical particles with two different densities
from coal fly ash and their characterization. Fuel 117 , 118-124.
Laudyn, D., Pawlik, M., & Strzelczyk, F. (2000). Elektrownie. Warszawa: Wydawnictwo NaukowoTechniczne.
Lee S.H., (1999). Characterization of fly ash directly collected from electrostatic precipitator. Cement
and Concrete Research 29 , pp. 1791-1797.
Łączny M., W. T. (2011). Modelowanie procesu powstawania cenosfer w kotłach pyłowych. Popioły z
energetyki, Zakopane, 19-21 października 2011 (pp. 191-203). Zakopane: Szczecin: Ekotech.
Mardon, S., & Hower, J. (2004). Impact of coal properties on coal combustion by-product quality:
examples from a Kentucky power plant. Coal Geology 59 , pp. 153-169.
Modlinski N. (2010). Computational modeling of a utility boiler tangentially-fired furnace retrofitted
with swirl burners. Fuel Processing Technology 91 , 1601-1608.
Mollah M.Y.A. (1999). Cristobalite formation from thermal treatment of Texas lignite fly ash. Fuel 78 ,
1277-1282.
Nocoń, J., Poznański, J., & Słupek, S. (1994). Technika cieplna. Przykłady z techniki procesów spalania.
Kraków: Wydawnictwa AGH.
Okada K. (2008). Activation energy of mullitization from various starting materials. Journal of the
European Ceramic Society 28 , 377-382.
Petrus, H. T. (2011). Performance of dry-separation processes in the recovery of cenospheres fromfly
ash and their implementation in a recovery unit. International Journal of Mineral Processing 98 , pp.
15-23.
Raask E. (1968). Cenospheres in pulverized-fuel ash. Journal of the Institute of Fuel 43 , pp. 339-344.
Rafako S.A. (2014, May 7). Retrieved from http://www.rafako.com.pl/
73
Ramme, B. W. (2011). Patent No. 8074804B2. United States of America.
Rybak W., (2008). Low sulphur coal combustion – Optimisation of boiler tunings. EDF Group & Polish
Universities R&D Cooperation Platform in Poland.
Shao Y., (2009). Characterization of porous silicon nitride ceramics by pressureless sintering using fly
ash cenosphere as a pore-forming agent. Journal of the European Ceramic Society 29 , pp. 1529-1534.
Shelby J.E. (1997). Introduction to Glass Science and Technology. Cambridge: The Royal Society of
chemistry.
Siddique R. (2010). Utilization of coal combustion by-products in sustainable construction materials.
Resources, Conservations and Recycling 54 , pp. 1060-1066.
Smith W.F. (1995). Principles of Material Science and Engineering. McGraw-Hill Companies.
Smoot, L. (1993). Fundamentals of Coal Combustion For Clean And Efficient Use. Amsterdam-LondonNew York-Tokyo: Elsevier.
Tomeczek, J. (1994). Coal Combustion. Malabar, Florida: Kriege Publishing Company.
Tomeczek, J., & Palugniok, H. (2002). Kinetics of mineral matter transformation during coal
combustion. Fuel 81 , 1252-1258.
Vassilev, S. T. (2004). Phase-mineral and chemical composition of coal fly ashes as a basis for their
multicomponent utilization. 2. Characterization of ceramic cenosphere and salt concentrates. Fuel 83
, pp. 585-603.
Vassilev, S., Baxter, D., & Vassileva, S. (2013). An overview of the behaviour of biomass during
combustion: Part I. Phase-mineral transformations of organic and inorganic matter. Fuel 112 , 391449.
Vergas, S., Frandsen, F., & Dam-Johansen, K. (1997). Performance of viscosity models for hightemperature coal ashes [project review]. Lyngby: Technical University of Denmark.
Wang, C., Liu, J., Du, H., & Guo, A. (2012). Effect of fly ash cenospheres on the microstructure and
properties of silica-based composites. Ceramics International 38 , 4395-4400.
Ward, C. R., & French, D. (2006). Determination of glass content and estimation of glass composition
in fly ash using quantitative X-ray diffractometry. Fuel 85 , 2268-2277.
Zhu, J., Lu, Q., Niu, T., Song, G., & Nia, Y. (2009). NO emission on pulverized coal combustion in high
temperature air from circulating fluidized bed–An experimental study. Fuel Processing Technology 90
, pp. 664-670.
Zimowa, M., & Webb, S. L. (2007). The combined effect of chlorine and fluorine on the viscosity of
aluminosolicate melts. Geochimica et Cosmochimica Acta 71 , 1553-1562.
74

Characterisation of fly-ash cenospheres from coal

Transcription

Similar documents

www.homark.com • • 1-218-253-2777

M Esteira Deck Freixo Termotratado (24)

WET-TO-DRY CONVERSIONS FOR COAL POWER GENERATION

Ash tree dieback – autumn and winter signs

Marshalls Seedless Ash(CMYK)

Multistage Concentration of Cenospheres in Fly Ash Using the

Arizona Electric Power Cooperative, Inc. is the generation utility

Museum Scavenger Hunt Grades 6-12

STOKER.FIRED WARM AIR Model