HAL Id: jpa-00227330
https://hal.archives-ouvertes.fr/jpa-00227330
Submitted on 1 Jan 1987
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.
MONTE CARLO SIMULATION OF THE RESPONSE OF Si(Li) X-RAY DETECTORS TO PROTON
INDUCED K X-RAYS OF LIGHT ELEMENTS
M. Geretschläger, O. Benka
To cite this version:
M. Geretschläger, O. Benka. MONTE CARLO SIMULATION OF THE RESPONSE OF Si(Li)
X-RAY DETECTORS TO PROTON INDUCED K X-RAYS OF LIGHT ELEMENTS. Journal de
Physique Colloques, 1987, 48 (C9), pp.C9-127-C9-130. �10.1051/jphyscol:1987921�. �jpa-00227330�
MONTE CARL0 SIMULATION OF THE RESPONSE OF Si(Li) X-RAY DETECTORS TO PROTON INDUCED K X-RAYS OF LIGHT ELEMENTS
M. GERETSCHLAGER and 0. BENKA
J o h a m e s Kepler Universitat Linz, Institut fiir Experimentalphysik, A-4040 Linz, Austria
A new detector model is deduced from measured ratios of low energy background counts to photopeak counts. This detector model assumes both the existence of a surface layer with reduced charge carrier collection efficiency and the existence of a low concentration of small regions of detector defects which have enhanced charge carrier recombination probability within the volume of the Si(Li) detector.
Monte Carlo simulations of the response of two different Si(Li) detectors to K X-rays have been performed in order to quantify possible contributions of detector front contact and dead layer, respectively, to the measured X-ray spectra. Upper limits of those contributions and upper limits of dead layer thickness are given.
The consequence of these results for detector efficiency calibration measurements are discussed.
Spectra of monoenergetic X-rays measured by means of semiconductor detectors show a continuous background below the photopeak. Our interest in the nature of the low energy background (LEB) arises from efficiency calibration measurements of X-ray detectors. We have shown [1,2], that one can use commercially available surface barrier detectors (SBDs) to measure X-ray spectra for X-ray energies E x > l keV. For SBDs one can determine the thickness of the entrance window tw, the dead layer tD and the sensitive layer tS with high accuracy. Consequently the X-ray absorption efficiency for the sensitive layer, EA=exp(-pwtw) exp(-ptD) [1-exp(-ptS) 1, where p, pw are the absorption
Article published online by EDP Sciences and available at http://dx.doi.org/10.1051/jphyscol:1987921
C9-128 JOURNAL DE PHYSIQUE
coefficients, can be determined with an error as low as 2 % for X-ray energies 1 keVsExs5 keV. However, there could also be s o m e contributions t o the measured X-rays which arise from photoabsorption events in t h e window and the dead layer of the SBD.
Therefore one cannot take advantage of the high accuracy of c A , as long as t h e origin of the LEB is not well understood.
S X < L ~ ) U Z P PHOTO EL.
A AUGER EL.
:
"'"II . E 3 1.L.
X - R A Y ENERGY Lev3
Fig.1. Calculated layer thickness T o f incomplete charge collection and mean projected range [ 4 ] of Si K-photoelectrons and Si KLL- Auger electrons.
Fig.2. Layer thickness T of in- complete charge collection a s a function o f the mean range 1/p o f X-rays. The solid lines are fit results. 2 0 % errors have been assumed.
In an earlier work [3] w e have measured proton induced K X-ray spectra of light elements (1252532) using two different Si(Li) detectors. From those measurements w e determined t h e number of LEB counts NB and the number of fully collected counts NF (photopeak counts and escape peak counts). I f w e assume that all effects responsible for t h e LEB can be represented by a Si surface layer of constant thickness T with charge collection efficiency 0 < ~ < 1 , absorption in layer T leads t o counts NB whereas absorption in the region behind leads t o counts NF. T is then given by T=ln(l+NB/NF)/p. Figure 1 shows the evaluated T and the results for electrons escaping from the region having V=1. The energy dependence of the ratio NB/NF indicates the presence o f two components in NB, one component NBC depending on energy in a manner s o that it could be well described by the existence of a constant layer of (energy independent) thickness Tc and the other component NBE being constant and independent o f X-ray energy. This constant component N B E can be tentatively explained if w e assume that t h e total volume of the detector has a low concentration of clusters o f crystal imperfections where the recombination probability is enhanced and where
v
therefore is V<1. In this model the calculated thickness Tplotted versus 1/p. 2 0 % errors have been assumed for the values T.
The linear dependence found is remarkable. W e do not assume any dead S i layer.
CHANNEL CHANNEL
CHANNEL CHANNEL
Fig.3. Examples o f Monte Carlo simulations for K X-ray spectga compared t o experimentally measured spectra. The values tW=40pg/cm
,
t D = O were used. The contributions of the gold window t o the LEB continuum is also shown.
In this work w e have further improved a computerprogram described earlier [ 3 ] , which simulates the experimental spectra by a Monte Carlo calculation. The simulation program now takes fully into account our detector model. The functions P(4) and 4(Z) (see ref. [ 3 ] ) were determined from proton induced K X-ray spectra of germanium and phosphorus, respectively. To calculate the energy loss of photoelectrons and Auger electrons w e use the extrapolated range 141. This procedure assumes a straight electron path during the slowing down process. Examples o f our Monte Carlo simulations for K X-ray spectra of Mg t o G e are plotted in fig. 3 together with the experimental spectra. For X-ray energies E x > 5 kev ( V t o Ge) there
C9-130 JOURNAL DE PHYSIQUE
is a significant difference between simulated and measured spectra in the region between the K, and Kg lines and immediately below the K
, line. This difference occures because we did not take radiative Auger contributions into account in our simulations. From our Monte Carlo simulations of Si(Li) spectra we deduced the following results:
1) The combined contribution of photoabsorption in the window and dead layer is always less than 1.5 % of the total spectrum (less than 0.5 % for E x > 6 keV). The absorption efficiency for the sensitive layer (7?>0), & corresponds therefore to more than
A
98.5 % of the number of total counts (extrapolated to channel 0) of the measured spectrum.
2) Any dead Si layer will produce a Si K fluorescence line. If a Si K line is not visible in a proton induced X-ray spectrum of chlorine or scandium having NTt10 counts the upper limit of a 6
dead Si layer is tD'5pg/cm 2
.
We find therefore that in efficiency calibration measurements one should determine both, photopeak efficiency e p and absorption efficiency E A , although in practice E A is only useful for K X-ray spectra of monoelemental targets. The measurement of E p involves nonlinear least squares fit procedures which introduce uncertain systematical errors, which depend upon the chosen fit interval and upon the chosen background function. These systematical fit errors can be appreciable for X-ray energies just above the Si K absorption edge. The extrapolation and interpolation of E p depends critically on the assumed model for those parts of the detector which produces background counts (regions with 0<??<1). The use of E A avoid all those drawbacks mentioned above and therefore one should take advantage of the inherently higher accuracy of E A whenever this is feasible. E A should be used as an "internal standard" for the extrapolation and interpolation of the photopeak efficiency E p .
[l] Geretschlager, M., Nucl. Instr. and Meth. (1982) 117 [2] Geretschlager, M., Nucl. Instr. and Meth.
A204
(1983) 479[3] Geretschlager, M., Nucl. Instr. and Meth. B (1983), to be published
[ 4 ] Iskef, J., Cunningham, J.W., Watt, D.E., Phys. Med. Biol. 28
(1983) 535
