HAL Id: jpa-00226152
https://hal.archives-ouvertes.fr/jpa-00226152
Submitted on 1 Jan 1986
HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are pub- lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers.
L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.
DETECTION SYSTEMS FOR FLUORESCENCE EXAFS MEASUREMENT
J. Baines, C. Garner, S. Hasnain, C. Morrell
To cite this version:
J. Baines, C. Garner, S. Hasnain, C. Morrell. DETECTION SYSTEMS FOR FLUORESCENCE EXAFS MEASUREMENT. Journal de Physique Colloques, 1986, 47 (C8), pp.C8-163-C8-166.
�10.1051/jphyscol:1986830�. �jpa-00226152�
JOURNAL DE PHYSIQUE
Colloque C8, supplhment au n o 12, Tome 47, dhcembre 1986
DETECTION SYSTEMS FOR FLUORESCENCE EXAFS MEASUREMENT
J.
BAINES, C.D. GARNER*, S.S. HASNAIN and C. MORRELL
SERC, Daresbury Laboratory, GB-Warrington WA4 4AD, Great-Britain
"~epartment of Chemistry, University of Manchester, GB-Manchester,
M 1 3 9 P L ,Great-Britain
Abstract: The fluorescence detection systems currently in use for the measurement of EXAFS at the Daresbury Synchrotron Radiation Source are described. Each system consists of a n array of NaI(T1) scintillation detectors. Methods of reducing the proportion of counts due to scatter are discussed. The advantages and problems associated with alternative detection systems are reviewed. Results are presented from measurements of energy resolution and count rate capability obtained with a prototype MWPC with a 10 cm thickness of gas.
Introduction:
The time needed to obtain a n EXAFS spectrum is governed by the requirement for the 'signal-to-noise' ratio to be better than some minimum over a given energy range above the absorption edge. The amplitude of the EXAFS 'signal' is proportional to the increase in count rate over the edge, F (the fluorescence count rate); the 'noise' due t o statistical fluctuations is proportional t o the square root of the total count rate above the edge, F+B (where B is the background count rate). Thus a figure of merit, X , can be defined as follows :
where f is the fluorescence fraction, F/<F+B). The requirements for a fluorescence detection system are, therefore, high detection efficiency and good background rejection. In practice there will be a trade-off between these two factors. Since the amplitude of the EXAFS oscillations is callibrated by reference to the edge height, a n additional requirement is that the detection efficiency of the system be constant over the range of count rates encountered in moving from before the edge t o after the edge.
Background reiection:
Contributions t o the background arise from elastic and inelastic scatter of the monochromatic beam and its 'harmonics' (higher order Bragg reflections from the monochromator). A further source is 'background fluorescence' from constituents of
the sample other than that of interest.
As the radiation from the SRS is polarized, the scatter count rate will be a minimum for a detector placed in the horizontal plane at 90' to the beam (detector 1). The count rate obtained fr4m the detection system can be increased by increasing the solid angle coverage of this detector. However, if no method of background rejection is applied, the proportion of background counts will be increased and the
Article published online by EDP Sciences and available at http://dx.doi.org/10.1051/jphyscol:1986830
JOURNAL
DE PHYSIQUE
net effect may be to decrease the signal-to-noise ratio. This can be prevented in a multi-element detection system by using the fluorescence fraction to weight the counts from the different elements before summation.
The method used to reduce the proportion of background in the data is dependant on the energy resolution of the detector. Whichever method is used to discriminate against scatter will also discriminate against the K R fluorescence of the sample (which contributes-20% of fluorescence counts). The resolution required to resolve the KKfluorescence from the elastic scatter (at half maximum) is 11% at the Cu K- edge, and 13% at the Mo K-edge. Better resolutions of about 8% and 6% at the respective edges are required to discriminate against inelastic scatter. If the energy resolution of the detector does not meet these requirements, filters can be used to preferentially absorb the scatter! For studies at the K-edge of a 3d metal of atomic number 2 , a (2-1) metal foil is used as the filter; a (2-2) foil is used for 4d metals. The improvement obtained with a filter is limited by the fluorescence of the filter itself. The amount of filter fluorescence reaching the detector can be reduced by a collimator assembly placed between the foil and the detector.
Concen- Relative data collection time tration NaI(Tl)+filters solid
(a) (b) (c> state
1 OmM 1 0.97 0.97 0.43 %
5mM 3.0 2.8 2.1 0.86 /'
7 i Em kmft E n
0
of paperImM 48 40 14 4.3
Table 1: The relative data collection -Collunamr
times calculated for different concentra- Sclnttiiator tion samples and different detection Photom~lt~plner systems. The beam intensity and the solid
angle coverage of the detection system
are the same in each case. @ s a ~ F O 1 l PLAN VIEW END VIEW
Figure 1: The arrangement of detectors within the detection systems at the SRS.
Detection svstems at the SRS:
Each of the three fluorescence detection systems currently in use at the SRS consists of four or five NaI(T1) scintillation detectors positioned around the sample at a distance of 8 cm. Each detector covers a solid angle of about 4% of 4iT (see figure 1). The detection system is described in more detail elsewhere? The outputs from the detectors are combined and input to a fast multi-channel analyser (MCA).
The energy resolution of the detectors is used to discriminate against the scatter of harmonics and in many cases it can also be used to discriminate against background fluorescence. Metal filters are used to reduce the scatter. A collimator assembly is being constructed which is designed to absorb 90% of filter fluorescence with a loss of only about 15% of sample fluorescence.
The count rate of the present detection system must be limited to avoid pile-up in the readout electronics. Pile-up reaches the 10% level at a total incident count rate of 4.5x10~s~'. The count rates observed at the SRS vary between different stations and depend on the nature of the sample and the X-ray energy. At the EXAFS station on the Wiggler beam line incident count rates of greater than
lo6
s-' per detector would be encountered at the higher energies if the beam were not collimated to reduce intensity. With foils in place the maximum count rate incident upon any detector on this station would be about 5x10' s-' at full intensity.A readout system providing a separate amplifier, discriminator and scaler for each detector is currently being tested at the EXAFS station on the Wiggler line. With this readout system the count rate will be limited only by the dead time of the detectors which lcads to a 10% loss of detection efficiency at an incident count rate of 4.4~10' s-' per detector. In addition, with this system it is possible to weight the contributions from the different detectors before summing the counts. In the near future this readout system will be commissioned for routine use on the Wiggler line station and on the high brightness EXAFS station where focussing monochromators have been installed. The approximate relative times required to collect data with the same signal-to-noise ratio from a lOmM, 5mM and 1mM sample using the existing detection systems are given as case (a) in table 1. The improvement expected when the counts from the different detectors are weighted (case (b)) and when in addition the collimator assemblies are installed (case (c)) are also given.
Other detection systems:
A detection system capable of resolving the sample fluorescence from scatter would give a significant reduction in the data collection time, particularly for very dilute samples, as shown in the last column of table 1. These results are based on the assumptions that the solid angle coverage and beam intensity are the same in each case. The number of detectors and readout channels required will depend on the dead time of the detectors and the pulse length of the signals produced.
The energy resolutions of several different detectors are listed in table 2. The count r-ate limits resulting both from detector dead time (or space charge effects in the case of the wire chambers) and from pile-up in the readout electronics are also given. Only solid state detectors are capable of fully resolving fluorescence from scatter. However a large number of these detectors would be required in order to obtain the desired count rate capability? With such an array the dead area between detectors would account for a significant fraction of solid angle. This is also a problem with NaI(T1) scintillators. The'necessity of cooling solid state detectors to liquid Nitrogen temperatures would make a detection system covering a large solid angle extremely bulky and expensive. Room temperate solid state detectors may provide a solution In the future: Although the Multi-Wire Proportional Chamber (MWPC) is not able to resolve fluorescence from scatter, it is capable of much higher count rates.
Large area MWPCs can be constructed with the counts split amoung several readout channels. In this way count rate capabilities of 10' s-' per detector can be achieved?
Plastic scintillators are capable of count rates of lo7 s-' with a single readout channel. However they have extremely poor energy resolution. It may be possible to at least partially resolve fluorescence from scatter using a wire chamber operated either as a gas scintillation proportional counter (GSPC)~ or a parallel plate proportional counter (PPPC~." The latter type of detector is also capable of higher count rates than the MIJPC.
DETECTOR ENERGY MAX. COUNT RESOLUTION RATE (s-')
( X ) (a) (b)
NAI(T~) 80 4.4~10' 4 . 5 ~ 1 0 ~
Plastic
- lo7 lo7
Solld state 6 - lo4 f-d,
* C d 2 Cathale-
MWPC 18 104 mm-2 lo6
PPPC 12 10'mm-~ lo7
GSPC 9 -Anode wires
12pm slummnlred zollm
+
Table 2: The energy resolution for 6 keV %,%w-.
photons and the count rate llmit (a) per detector and (b) per readout channel given for varlous different detectors.
Figure 2: The electrode configuration of the MWPC.
JOURNAL
DE PHYSIQUE
A MWPC for EXAFS:
To obtain good detection efficiency, a MWPC must have a large thickness of gas, or be operated at high pressure. A prototype wire chamber has been constructed which has removable wire planes, the spacing of which can be altered. This detector can be operated either as a MWPC or a PPPC. Preliminary results are presented for the operation of this detector as a MWPC with a total thickness of 96mm. The chamber was divided into a conversion region 88mm thick, and an amplification region 8 mm thick (see figure 2). The detector was filled with a mixture of 90% Xe and 10% CO, at atmospheric pressure. The anode wire was held at a potential of 3.2 kV and a drift field of 100 ~ c m - ' was applied. An energy resolution of 20% was measured for 7.5 keV photons absorbed in the amplification region. The energy resolution was much poorer for photons absorbed in the conversion region. Better energy resolutions may be possible at higher drift fields. Resolutions of 18% at 6 keV have been reported for photons absorbed in a conversion region 10 cm
thick'
The energy resolution was degraded at count rates above about 10' s-' cm-,.
When the incident radiation was localised in a region 0.8mm x 2mm, the output count rate reached a plateau at about 2x10~s-'. However, such high local ~ntensities are not encountered in fluorescence EXAFS measurements.Conclusion:
Only a detection system composed of solid state detectors would be capable of fully resolving fluorescence from scatter. With such a detection system rhe time taken to collect data for a 1mM sample would be reduced by about a factor of 3 compared to the time required by a detection system with the same solid angle coverage but using foils and collimator slit assemblies to reduce the amount of scatter. However such a system would be bulky and expensive with a significant fraction of dead area. The energy resolution of NaI(T1) scintillators and MWPCs is sufficiently good to be used to discriminate against some of the sources of backkround. Large area detectors capabli of count rates of 10' s-' per detector are possible with MWPCs and plastic scintillators. The latter requires less readout channels, but has an extremely poor energy resolution.
We would like to thank the Science and Engineering Research Council for funding this work and for the provision of the facilities of the Daresbury Laboratory. The prototype MWPC was built at NIKHEF in the Netherlands as part of a collaborative project
.
References:
1. E.A. Stern and S.M. Heald, Nucl. Instr. and Meth. 172 (1980) 397
2. J.T.M. Baines, C.D. Garner. S.S. Hasnain and C. Morrell, Nucl. Instr. and Meth.
A246 (1986) 565
3. J.H. Howes and F.L. Allsworth, IEEE Trans. Nucl. Sci. 33 (1986) 283 4. J.S. Iwanczyk et al.,-IEEE Trans. Nucl. Sci. 33 (1986) 355
5 . G.C. Smith, Nucl. Instr. and Meth. 222 (1984) 230
6. W.J.C. Okx et al., IEEE Trans. Nucl. Sci. 33 (1986) 391 7. H.E. Schwarz and I.M. Mason, Nature 309 (1984) 532
8. A . Peisert. Nucl. Instr. and Meth. 217 (1983) 229
9. G. Charpak, C. Demierre, R. Kahn, J.C. Santiard and F. Sauli, Nucl. Instr. and Meth. 141 (1977) 449
