Der Positronenmessplatz EPOS an der Elektronenquelle ELBE

Transcription

Introduction
Positrons in 5 Minutes
Current state of EPOS
Digital Positron Lifetime
Conclusion
Der Positronenmessplatz EPOS an der
Elektronenquelle ELBE
– Aktueller Stand –
Arnold Krille (Dipl. Phys. & “Father-of-one”)
Institut für Physik, Martin-Luther-Universität Halle-Wittenberg
May 21th, 2008
EPOS an ELBE – Aktueller Stand
Arnold Krille
Introduction
Conclusion
Introduction
Arnold Krille
Introduction
Conclusion
Introduction
1
Introduction
2
3
4
The task
Different extraction methods
Simulations with gauss-pulses
Measurements on Si with BaF2 -Scintillators
Evaluation of α-Fe-data from Prague
Measuring Si with LSO-Scintillators
5
Conclusion
Arnold Krille
Introduction
Conclusion
A proof-of-concept (mini-)talk
How to explain positrons in 5 minutes (or less).
Arnold Krille
Introduction
Conclusion
Antimater and Albert Einstein
Dirac[1] found solutions with positive
charge in his electron-theory (1928)
Anderson[2] found that the particle was
not the proton and postulated the
positron, anti-particle of the electron (in
1933)
positrons annihilate with electrons, always
one e + with one e −
thanks to Einstein[3] the annihilation
results in γ-Radiation where
Eγ = (me + + me − )c 2
Figure: Einstein 1921
Arnold Krille
Introduction
Conclusion
Conservation of the pulse
Another physics law
P
pi = 0
The pulse of a system has to be conserved.
Effect on positron-electron-annihilation:
Two γ-quants are emitted in opposite
directions
Both have energy of 511keV
Figure: Feynman-diagram of
electron-positron-annihilation
Arnold Krille
Introduction
Conclusion
Conservation of the pulse
Another physics law
P
pi = 0
The pulse of a system has to be conserved.
Effect on positron-electron-annihilation:
Two γ-quants are emitted in opposite
directions
Both have energy of 511keV
Positron is in ground-state
But: Electron is in excited state
→ γ-energy changes due to p of the
electron
→ Electron-energy depends on the state,
core-electrons have higher energy than
valence-electrons
Figure: Feynman-diagram of
electron-positron-annihilation
(extended)
Arnold Krille
Introduction
Conclusion
Statistics (More or less:)
Important
Positroneneinfang durch Kristalldefekte
Electron-positron-annihilation is a highly
statistical process!
Because the diffusion of the positron in the
solid is a random-walk.
Annihilation-probability is influenced by:
Electron density
Electron energy
Atomic structure
für 22-Na:
J Quant
(1.27 MeV)
J Quant (511 keV)
• Positronen-Wellenfunktion wird im Defekt lokalisiert
• A nnihilationsparameter ändern sich, wenn Positron im Defekt zerstrahlt
• Defekte können nachgewiesen werden (Identifizierung und Quantifizieru
Figure: Positrons in the solid
Defects, voids, charge
Arnold Krille
Introduction
Conclusion
Lifetime
Lifetime of the positron is influenced mainly by
two effects:
1
2
The lower the electron density the lower
the chance to hit an electron the higher
the lifetime.
One (or many) missing atoms/cores build
a potential well the positron can’t escape
(because normal cores repulse the
positron).
→ positron lifetime increases
but this only works in neutral or negativ
traps
10000
Si; 90degree-setup; 1.5cm BaF2
1000
100
10
1
-2
-1
0
1
2
3
4
5
6
Time in samples (1 sample = 250ps)
7
8
Figure: Positron-Lifetime in
Cz-Silicon
Arnold Krille
Introduction
Conclusion
Doppler-broadening and Angular-correlation
Doppler-broadening:
measures the energy-shift in γ-direction
(usually pz )
can be measured with one detector,
→gives high background
two detectors in coincidence give low
background but longer measurement-time
energy-range is 511keV ± 10 keV
While the spectra can be calculated
theoretically, most times results are compared
to a defect-free spectrum.
22
2 Experimental Techniques
GaAs
513
514
515
Plastically deformed
Reference
Intensity [arbitrary units]
two techniques for one effect
both measure the electron-energy from the
energy-shift of the γ-quants
Aw
As
504
506
508
510
512
514
516
J–ray energy [keV]
Fig. 2.11. Doppler-broadening spectra of as-grown zinc-doped gallium
positron trapping (reference) compared with plastically deformed
1997b). The line shape parameters S and W are determined by the ind
divided by the area below the whole curve. The curves are normalized
Figure: Doppler-Spectrum[4]
The background correction is often performed as the subt
line. More sophisticated treatments use a realistic backgroun
eled by a non-linear function. This function takes into account
at a given J-ray energy is proportional to the sum of anni
higher energies. Despite such a background reduction, the Do
slightly unsymmetrical. The calculation of the W paramete
quently carried out only in the high-energy wing of the D
quality Doppler-broadening spectra are obtained by the coi
preferably with a setup using two Ge detectors. In this case, t
Arnold Krille
be taken on both sides of the curve.
Introduction
trapping,
for capture
of all positrons in defects, is
necessary. This
Digitali.e.Positron
Lifetime
Conclusion
by correlated positron lifetime measurements, which provide the fract
trons trapped in the defect. Such contour plots can also be directly com
theoretical calculations to improve the understanding of the electronic
the defect.
The first depth-resolved measurements of the two-dimensional angu
tion of annihilation radiation were reported by Peng et al. (1996). The d
ture of the SiO2/Si interface was studied by the combination of slo
beam measurements and 2D-ACAR. The positron beam system used
krypton moderator and an initial positron beam flux of about 6 107 s 1
resolution was obtained by the variation of the incident positron en
Doppler-broadening and Angular-correlation
two techniques for one effect
both measure the electron-energy from the
energy-shift of the γ-quants
Angular-correlation:
measures the energy-shift perpendicular to
the γ-direction (px and py )
has to be done in coincidence
1D- and 2D-measurements are possible
typical range is ±10 − 20mrad, resolution
is 0, 2mrad
measurement for one spectrum is typically
several days
Gives good results on the electronic structure,
but evaluating the spectra requires many
theoretical calculations.
Fig. 2.15. Combined perspective and contour plots of the measurement of two
angular correlation of annihilation radiation in gallium arsenide exhibiting no
ping in defects (Tanigawa et al. 1995). The upper panel shows the contour m
mentum distribution p in the (100) plane.
Figure: ACAR-Spectrum[4]
Arnold Krille
Introduction
Conclusion
Overview
6
2 Experimental Techniques
Birth J-ray
Annihilation J-rays
1.27 MeV
0.511 MeV
4
't
1. Positron lifetime
Sample
2. Angular correlation
px , y
4xx,,yy
4
m0 c
e+ source
22
Na
Diffusion
L+ | 100 nm
a100 µm
Thermalization
(1012 s)
3. Doppler broadening
0.511 MeV ± 'E,
'E = pzc/2
Fig. 2.1. Scheme of different positron experiments. Positrons from an isotope source (
EPOS an ELBE – Aktueller22
Stand
Arnold Krille
Introduction
Conclusion
ELBE-Overview
Arnold Krille
Introduction
Conclusion
ELBE-Overview: ELBE-Layout
Arnold Krille
Introduction
Conclusion
ELBE-Overview: ELBE-Command-Center
Arnold Krille
Introduction
Conclusion
ELBE-Overview: Buncher of the electron-beam
Arnold Krille
Introduction
Conclusion
Schema des EPOSEPOS-Systems
Arnold Krille
Introduction
Conclusion
Arnold Krille
Introduction
Conclusion
Current state of EPOS: Electron beamline in Cave 111b
Arnold Krille
Introduction
Conclusion
Current state of EPOS: RKR explaining the system
Arnold Krille
Introduction
Conclusion
Current state of EPOS: Connection to the lab
Arnold Krille
Introduction
Conclusion
Current state of EPOS: Corner-stone-ceremony
Arnold Krille
Introduction
Conclusion
Current state of EPOS: Marco at normal working speed
Arnold Krille
Introduction
Conclusion
Current state of EPOS: Lead-shield & vacuum tests
Arnold Krille
Introduction
Conclusion
Current state of EPOS: Lead-saw
Arnold Krille
Introduction
Conclusion
Current state of EPOS: Problems with leaks...
Arnold Krille
Introduction
Conclusion
Current state of EPOS: Status of the lab
Arnold Krille
Introduction
Conclusion
Current state of EPOS: Current-sources in place
Arnold Krille
Introduction
Conclusion
Current state of EPOS: A fun place to work at
Arnold Krille
Introduction
Conclusion
Current state of EPOS: Whats next?
1
2
3
Marco is finishing his diploma-thesis on the beam-system
Test of the first part of the guidance-system with an
electron-“beam”
(Its only ∼ 20 currents to adjust.)
Test of the same parts with a conventional positron-beam
6
Install the real converter and generate the EPOS-beam from
ELBE-diagnostic mode
Finish the lab. (Chopper, buncher, vacuum, conventional source,
final acceleration, sample chamber, detector setup,...)
Get external users to send in samples to measure
7
Get the nobel prize. Or at least a publication in science/nature...
4
5
Arnold Krille
Introduction
Conclusion
The Setup:
Detektor
Detektor
In1 In2
Digitizer
PC
Arnold Krille
Introduction
Conclusion
The task
Extract the timing information from a “smeared”, digitized signal
Be as accurate as possible... (FHWM should be <100ps)
...as fast as possible (should do >1000 events per second)
0.5
0.4
Input Voltage [V]
0.3
0.2
0.1
0
-0.1
-0.2
-0.3
-0.4
-0.5
-1e-08
Channel 0 (moved +0.4V)
Channel 1
-5e-09
0
5e-09
1e-08
1.5e-08
2e-08
Time [s]
Arnold Krille
Introduction
Conclusion
Different extraction methods
Interpolation is needed, different methods are
possible:
0.5
Constant-Fraction with
Polynom-Interpolation
Spline-Interpolation
Gauss-Interpolation
...all combined with “Moving average”
0.4
0.3
Input Voltage [V]
0.2
0.1
0
-0.1
-0.2
-0.3
-0.4
integral-Constant-Fraction
-0.5
-1e-08
Channel 0 (moved +0.4V)
Channel 1
-5e-09
0
5e-09
Time [s]
1e-08
1.5e-08
2e-08
Signed-integral-Constant-Fraction
Deconvolution of pulses
Polar-like plots
Figure: Pulse-pair of startand stop-signal
Resolutions from literature range from 120ps 200ps.
Arnold Krille
Introduction
Conclusion
Simulations with gauss-pulses
Arnold Krille
Introduction
Conclusion
Simulations with gauss-pulses: Noise-Levels
1.2
Extracted timing sigma
1
Identity (should-be function)
Noise 0.2
Noise 0.1
Noise 0.05
Noise 0.01
Noise 0.00
0.8
0.6
0.4
0.2
0
0.01
0.1
Artificial timing sigma
1
0.01 ≡ 51dB SNR ≡ 8bit, 0.05 ≡ 37dB SNR ∼
= 6bit
Arnold Krille
Introduction
Conclusion
Simulations with gauss-pulses: Averaging
1
0.8
Avg 1
Avg 2
Noise 0.1
Noise 0.05
0.6
0.4
0.2
0
0.01
0.1
1
Arnold Krille
Introduction
Conclusion
Simulations with gauss-pulses: Samplingrates
0.3
0.25
8GS
4GS
2GS
1GS
0.2
0.15
0.1
0.05
0
0.001
0.01
0.1
1
Arnold Krille
Introduction
Conclusion
Measurements on Si with BaF2 -Scintillators: Polynom-CF
Sample
Source
FWHM
213,4ps
203,6ps
Intensity
73,49%
21,88 %
4,61 %
0,02 %
Silizium-Probe
3 Komponenten-Fit
t1 = 0.218 ns, t2 = 0.38 ns, t3 = 1.5 ns
I1 = 97 %, I2 = 2 %, I3 = 0 %
A1 = 115275
x1 = 0.0992275 ns
e1 = 0.177226 ns
100000
10000
Counts
Error-functions:
Intensity Shift
0
70,5 %
29,5 % 192ps
Lifetime
219ps
165ps
380ps
1,5ns
Silizium
Aluminium
Salz (Bulk)
Salz (nano-Pores)
1000
Resulting Error
FWHM = ∼ 290ps
With simple
polynom-interpolation.
100
10
-1
0
1
2
3
4
Time [ns]
Arnold Krille
Introduction
Conclusion
Measurements on Si with BaF2 -Scintillators: Averaging
Running a moving-average on the raw-dataset before processing:
CF
Lifetime
Overall-FWHM
should be 219ps
Cf 0,1
214,8(6)ps
211(1)ps
Cf 0,2
235(2)ps
203(1)ps
Cf 0,3
201(4)ps
196(1)ps
Resulting Error
196-211ps
Better then just simple polynom. But still not good enough...
Arnold Krille
Introduction
Conclusion
Measurements on Si with BaF2 -Scintillators: iCF
Start
20
Counts^-1
Testing the
integral-Constant-Fraction-Method[5]:
Integrierte Energie
25
15
10
5
Integration of the pulse around the
extremum (= the real signal)
0
0
0.5
- slow
+ better energy resolution
1
Normale Energie
1.5
2
1.5
2
Stopp
20
Counts^-1
Integrierte Energie
25
+ finds “real” energy of some pulses.
+ allows to filter on risetime,
actually means
pulse-shape-discrimination
15
10
5
0
0
0.5
1
Normale Energie
Arnold Krille
Introduction
Conclusion
Measurements on Si with BaF2 -Scintillators: iCF
Lifetime (LT#3) - Si-Reference-Sample - BaF2 - iCF 0.1
1e+06
100000
10000
Counts
Resulting FWHM
248,5ps
244,8ps
233,5ps
235,8ps
227,8ps
236,8ps
255,4ps
253,1ps
243,5ps
-5
0
2
3
5
10
1000
100
10
-2
-1
0
1
4
5
Time [ns]
1e+06
without RTF
with RTF 1
with RTF 2
100000
10000
Counts
Method
CF 0,1
CF 0,1 rt1
CF 0,1 rt2
CF 0,2
CF 0,2 rt1
CF 0,2 rt2
CF 0,3
CF 0,3 rt1
CF 0,3 rt2
without RTF
with RTF 1
with RTF 2
-5
0
2
3
5
10
1000
100
10
-2
-1
0
1
4
5
Time [ns]
1e+06
Resulting Error
10000
Counts
227 - 255ps
Not that good...
without RTF
with RTF 1
with RTF 2
100000
-5
0
2
3
5
10
1000
100
Risetime-filter is good to filter pile-up at t < 0.
10
-2
-1
0
1
4
5
Time [ns]
Arnold Krille
Introduction
Conclusion
Evaluation of α-Fe-data from Prague: Polynom-CF
dataset of α-Fe used as reference-sample in Prague[5]
no dedicated start-/stop-tubes, can be evaluated in both directions
Method
pCf 0,1
pCf 0,2
pCf 0,3
FWHM
Start-Stop
193(2)ps
167(2)ps
159(2)ps
Stop-Start
200(2)ps
171(2)ps
159(2)ps
Resulting Error
FWHM of 160ps is actually as good as the analog setup in Prague.
⇒ My Polynom-CF-Method should be right and give good results.
Probably my windows are to wide-open.
Arnold Krille
Introduction
Conclusion
Evaluation of α-Fe-data from Prague: iCF
FWHM
0,1
0,1
0,1
0,2
0,2
0,2
0,3
0,3
0,3
rt
rt4
rt
rt4
rt
rt4
135(2)ps
134(2)ps
137(2)ps
12±50ps
8±55ps
8±10ps
100,1(4)ps
100,4(4)ps
99,9(5)ps
563(2)ps
549(2)ps
538(2)ps
547(2)ps
538(2)ps
535(2)ps
259(2)ps
257(2)ps
255(2)ps
Whats wrong?
Seems to indicate that something on the
implementation isn’t right.
Counts - CF 0.1
Should be 106ps
iCf
iCf
iCf
iCf
iCf
iCf
iCf
iCf
iCf
HL - Start-Stopp
L
100000
100000
10000
10000
1000
1000
100
100
10
10
-2 -1
Counts - CF 0.2
Lifetime
0
1
2
3
4
5
6
100000
100000
10000
10000
1000
1000
100
100
10
-2 -1
0
-2 -1
0
-2 -1
0
10
-2 -1
Counts - CF 0.3
Method
0
1
2
3
4
5
6
100000
100000
10000
10000
1000
1000
100
100
10
10
-2 -1
0
1 2 3
Time [ns]
4
5
6
Arnold Krille
Introduction
Conclusion
Measuring Si with LSO-Scintillators
New scintillator-material
Comparing to BaF2 :
10000
+ Better energy-resolution
+ ∼ 6× higher efficiency
- Pulse-shape un-usable for
analog electronics
1000
Counts
+ Pulse-risetime similar
LSO normal
LSO integriert, skaliert 1/24
100
10
1
Timing resolution
Current test-spectra show bad
timing-resolution. But with
wide-open energy-windows.
-5
-4.5
-4
-3.5
-3
-2.5
-2
Input Voltage [a.u.]
-1.5
-1
-0.5
0
Figure: LSO energy spectrum
Arnold Krille
Introduction
Conclusion
Conclusion: Plans for next measurements
Roadmap:
More measurements with LSO.
Test the new XP20Z8-photomultipliers.
More simulations on noise/bit-depth vs. timing-resolution,
combined with measurements using a 14bit digitizer,
maybe even combined with measurements using a 24bit digitizer?
Fix the external coincidence-unit.
Build up the system at EPOS.
Arnold Krille
Introduction
Conclusion
Conclusion: Light at the end of the tunnel
Arnold Krille
Introduction
Conclusion
Conclusion: Literature, Links, Thanks
Thanks for your attention!
Get the slides at http://positron.physik.uni-halle.de/.
P. A. M. Dirac.
The quantum theory of the electron.
Proc. R. Soc. London, Ser. A, 118:351–361, 1928.
C. D. Anderson.
The positive electron.
Phys. Rev., 43:491–494, 1933.
A. Einstein.
Ist die Trägheit eines Körpers von seinem Energieinhalt abhängig?
Annalen der Physik, 323:639–641, 1905.
R. Krause-Rehberg and H. Leipner.
Positrons in Solids.
Springer-Verlag, 2001.
F. Bečvář.
Methodology of positron lifetime spectroscopy: Present status and perspectives.
Nuclear Instruments and Methods in Physics Research B, 261:871–874, 2007.
Arnold Krille

Der Positronenmessplatz EPOS an der Elektronenquelle ELBE

Transcription

Similar documents

6. Defekte nach Implantation von hochenergetischen Ionen in Si