ON THE ENERGY DEPENDENCE OF THE DETECTOR EFFICIENCY OF A Si-PIN DIODE

Size: px

Start display at page:

Download "ON THE ENERGY DEPENDENCE OF THE DETECTOR EFFICIENCY OF A Si-PIN DIODE"

Tobias Barnett
6 years ago
Views:

$Copyright(C)JCPDS-International Centre for Diffraction Data 2, Advances in X-ray Analysis, Vol.42 36 Copyright(C)JCPDS-International Centre for Diffraction Data 2, Advances in X-ray Analysis, Vol.$

1 Copyright(C)JCPDS-International Centre for Diffraction Data 2, Advances in X-ray Analysis, Vol Copyright(C)JCPDS-International Centre for Diffraction Data 2, Advances in X-ray Analysis, Vol ON THE ENERGY DEPENDENCE OF THE DETECTOR EFFICIENCY OF A Si-PIN DIODE H.Ebel, M.Mantler, S.Saxinger, RSvagera, B.Wernsperger, P.Wobrauschek and A.C.Huber3 Institut fiir Angewandte und Technische Physik, Technische UniversitLit Wien, Wiedner flauptstrafle 8-1, A-14 Wien (Austria) Atominstitut der iisterreichischen Universitiiten, SchiittelstraJe 11.5, A-12 Wien (Austria) 3 AMPTEK, Inc.,6 De Angelo Dr., Bedford, MA 173 (USA) ABSTRACT Characteristic Q-radiations of 15 elements from Z=21 to 44 were excited by monochromatic Ag K-radiations from Cd-19 (27mCi). The characteristic radiations in the photon energy range from 4.9keV to 19.28keV were detected by a Si-PIN diode with a silicon thickness of 3pm. The measured signals were corrected for the influence of the finite size of the plane specimens on the measured countrate/cm2 and also for the influence of the distances between specimen, source and detector on the signal strength. Both, the geometry and the activity of the source were well defined and allowed the computation of absolute values of the fluorescence signals. Our algorithms ) employed photoabsorption coefficients from McMaster et a12, fluorescence yields from Hubbell et a13 and photon energies and transition probabilities from Johnson and White4. The fluorescence yield of silver and the decrease of source activity from its last calibration were considered in our calculations. For the detector efficiency we used the expression described by Fiori et a?. The comparison of theoretical and experimental results was performed by minimizing the standard deviation between corresponding values of the elements in dependence on the true activity of the source and the effective active thickness of the Si-PIN diode. The evaluations allowed the following conclusions: I. the activity of the source is 26% smaller when compared to the expected value II. the,,effective active thickness of the Si-PIN diode is 17Opm and thus, smaller when compared to the expected thickness of 3~. The reduced active thickness is caused by rise-time discrimination (RTD) in the course of electronic pulse processing. RTD reduces spectrum background and improves line shapes by gating off events occurring in the back of the detector, where slow charge collection leads to low amplitude pulses. Therefore, a calibration is recommended when absolute countrates are to be measured with Si-PIN diodes. EXPERIMENTAL We developed a remote controlled vehicle for an application in quantitative soil analysis. Fig.1 depicts a cross section of the setup with a pair of wheels, the detector, the ring source and the specimen. A Cd-19 annular source from DuPont Pharma with an activity of 27mCi has been used and the time after calibration was approximately 9 days. For correction of this influence on the activity we used a half-life period of 464 days. From DuPont Pharma we got the following detailed information on the source: Inner diameter 28.7mm, outer diameter 35.5mm, active annular area 3.17mm2, a computed plated activity thickness of 3.12mg/cm2, the active volume is electroplated onto a.25mm thick commercial pure (>99.9%) silver target, the capsule has a.15mm to.18mm thick aluminium window, the 88keV photon

$X-ray Analysis. Sponsored by the International Centre for Diffraction Data (ICDD).$ This document is provided by ICDD in cooperation with the authors and presenters of

This document is provided by ICDD in cooperation with the authors and presenters of

All copyrights for the document are retained by ICDD.

2 This document was presented at the Denver X-ray Conference (DXC) on Applications of X-ray Analysis. Sponsored by the International Centre for Diffraction Data (ICDD). This document is provided by ICDD in cooperation with the authors and presenters of the DXC for the express purpose of educating the scientific community. All copyrights for the document are retained by ICDD. Usage is restricted for the purposes of education and scientific research. DXC Website ICDD Website -

3 Copyright(C)JCPDS-International Centre for Diffraction Data 2, Advances in X-ray Analysis, Vol Copyright(C)JCPDS-International Centre for Diffraction Data 2, Advances in X-ray Analysis, Vol emission is 2.9 * lo6 photons per second per steradian and the computed total 22keV photon emission (Ag Ku-radiation) is lo7 photons per second per steradian. Fig.1 Cross-section through the vehicle with geometry between source, specimen and detector. Normal distance from specimen to the plane of source is 1.22cm and normal distance from specimen to detector crystal is 2.82cm. An XR-1CR Si-PIN diode from AMPTEK has been used for detection of fluorescence signals. AMPTEK gave us the following detailed information on the detector: Detector size 2.4 *2.8rnm2, silicon thickness 3OOpm, detector window (Be) 12.5pm. From these data the detector efficiency of Fig.2 has been calculated by the algorithms described by Fiori et a T 6 7.z 6 ; 5 t5 4 H z Photon Energy (kev) Fig.2 Computed detector efficiency of the Si-PIN diode. The thickness of the active volume has been assumed to be 3OOpm.

4 Copyright(C)JCPDS-International Centre for Diffraction Data 2, Advances in X-ray Analysis, Vol Copyright(C)JCPDS-International Centre for Diffraction Data 2, Advances in X-ray Analysis, Vol First experiments have been performed on the influence of the specimen area on measured signals. For this pu.rpose thin quadratic and rectangular Zn-platelets of.5mm thickness and 1 exp * Isf - %z 9 E ;i 5.s z ul 2 a 3 rs Area (cma2) Fig.3 Measured Zn Ku-signals of Zn platelets with different specimen area Area (cma2) Fig.4 Reduced Zn Ka-signal. The linear response describes the measured fraction of the characteristic signal oer cm2 when compared to a specimen with an infinitesimal small area.

5 Copyright(C)JCPDS-International Centre for Diffraction Data 2, Advances in X-ray Analysis, Vol Copyright(C)JCPDS-International Centre for Diffraction Data 2, Advances in X-ray Analysis, Vol areas from.3cm2 to 4.2cm2 have been measured. The result can be seen from Fig.3. We observed no significant difference between square and rectangular platelets of comparable area. The decrease of the characteristic Zn Ku-signal per cm2 with increasing area is due to the systematic change of the geometry of incident source radiation and take-off of Zn fluorescence radiation. To describe the result in a general way we extrapolated the linear response towards area=o. In Fig.4 the ratio of measured signal and extrapolated value versus area is plotted. This response can be used to determine the cleared countrate of platelets from any chemical element in the following way. The measured countrate of element x is divided by the area of the platelet and by the reduced Zn Ku-signal of Fig.4 corresponding to the area of the platelet. Thus, the influence of the specimen area on the characteristic countrate per cm2 is eliminated. Another essential influence on measured countrates comes from the distance between source, detector and the specimen. We investigated again Zn-platelets and varied the distance by mounting the platelets on thin foils of different thicknesses. The result of the measurements can be seen from Fig.5. An increasing distance from the reference point (compare with the numerical values of Fig.1) means a decreasing distance between specimen and source. A variation of approximately 3mm causes a variation of the Zn Ka-countrate from 87 cts.crne2 s-l to 112cts.cmT2s. Thus, we observed an increase of the measured countrate of nearly 1% per mm. The full line through the upper point at reference position has been computed from our theoretical description of the countrate considering a variation of the specimen position with regard to the reference point. Similar to the treatment of the area dependence of the countrate it is possible to normalize the response with regard to the reference point (Fig.(i) and to correct measured countrates for distance variations by dividing the measured value by the numerical value of the reduced response corresponding to the distance from the reference point. 14 I I 12 exp. lin - g 1 i 3i E 8 z.p ci, 6 2 Q 52 4 rs Distance from Reference Point (cm) Fig.5 Variation of the reduced Zn Ka-countrate by an increasing distance of the platelet surface from the reference point.

6 Copyright(C)JCPDS-International Centre for Diffraction Data 2, Advances in X-ray Analysis, Vol.42 4 Copyright(C)JCPDS-International Centre for Diffraction Data 2, Advances in X-ray Analysis, Vol f! 9.F s Q z rs $ z 5 6 a I I I I I Distance from Reference Point (cm) Fig.6 Normalized response of the Zn Ka-countrate versus distance from the reference point. The reference point is defined by the numerical values given together with Fig.1. An increasing distance from the reference point means a decreasing distance between specimen and source (and detector). Thus, the measured countrate increases and the increase is approximately 1% per mm. The sequence of corrections is I. the measured countrate is divided by the area of the specimen II. the numerical value of the countrate per cm2 is reduced to a specimen of,,small area III. this numerical value is reduced to reference position of the specimen The validity of the procedure of data reduction is restricted to the given geometry and distance variations in the mm-range. For quantitative applications of the analytical setup these two influences on measured countrates ask for plane specimen surfaces, defined specimen areas and defined specimen distances. The next step of our program has been the measurement of the detector efficiency. This quantity depends on the photon energy. A well defined variation of the photon energy can be performed by the measurement of characteristic fluorescence radiations of different chemical elements. The following pure elements have been investigated (K&radiation): SC, Ti, V, Cr, Mn, Fe, Co, Ni, Cu, Zn, Ge, Y, Nb, MO, Ru An extension to higher Z-values is restricted by the possibility to excite the fluorescence radiation by Ag K-radiation and an extension to lower Z-values by the poor excitation probability, the decreasing fluorescence yield and the absorption of the fluorescence radiation in the ambient atmosphere. With the above mentioned chemical elements from SC to Ru we cover an energy interval from 4keV to 19keV.

7 Copyright(C)JCPDS-International Centre for Diffraction Data 2, Advances in X-ray Analysis, Vol Copyright(C)JCPDS-International Centre for Diffraction Data 2, Advances in X-ray Analysis, Vol THEORY The excitation by monochromatic Ag Ka- and Ag KP-radiations of the ring source enables a simple computation of the fluorescence countrates. We used Sherman s algorithm ) for primary excitation, photoelectric absorption coefficients from McMaster s tables2), fluorescence yields from Hubbell s tables3), transition probabilities from Johnson s tables4 and calculated the detector effiency by Fiori s algorithm5 with the above given characteristic data. The expected detector efficiency can be seen from Fig.2. Fig.7 depicts the comparison of computed and measured countrates. The theoretical response shows a certain similarity with the experimental energy dependence but, on a first glance differs from the abolute values of the signals. A first attempt to find an agreement between the two responses has been made by a systematic variation of the primary photon flux from the ring source. The comparison of the curves gave an evidence for an incorrect description of the energy dependence of the detector efficiency. Consequently, the complete set of computed fluorescence countrates has been treated by a least squares fit procedure where the primary photon flux and the thickness of the Si-crystal of the Si-PIN diode have been used as variables. Fig.8 gives the result of the lsfcalculations. The primary photon flux has to be reduced by a factor.74 and the,,effective thickness of the Si-crystal is 17pm. A discussion with the supplier of our detector lead to the following explanation for the discrepancy between geometrical and effective thickness. For background reduction and improved line shapes rise time discrimination (RTD) is used in the course of electronic pulse processing. As a consequence of RTD events ocurring in the back of the detector are suppressed. In other words, the detector crystal appears to be thinner when compared to its geometric thickness. Consequently, a calibration of the detector becomes unavoidable as a quantitative analysis with absolute signals has to be performed z 16 P E 14 Y 8 12 u, g 1.9 cn t W 2 fb n' Q 2 - * t" i-q t t t t Photon Energy (kev) Fig.7 Comparison of theoretical and measured fluorescence countrates. Theoretical values have been obtained with original data of primary photon flux and thickness of Si-crystal. I exp 8 theor. $

8 Copyright(C)JCPDS-International Centre for Diffraction Data 2, Advances in X-ray Analysis, Vol Copyright(C)JCPDS-International Centre for Diffraction Data 2, Advances in X-ray Analysis, Vol Fig.9 represents the comparison of the expected and the true energy dependence of the detector efficiency of the Si-PIN diode. P 1-t 'exp' - 'theor.' t IO Photon Energy (kev) Fig.8 Comparison of theoretical and measured fluorescence countrates. Theoretical values have been obtained with.74 times the expected primary photon flux and a thickness of 17~ instead of 3pm of the Si-crystal Photon Energy (kev) Fig.9 Comparison of the true (dashed curve) detector efficiency with the expected (full curve).

9 Copyright(C)JCPDS-International Centre for Diffraction Data 2, Advances in X-ray Analysis, Vol Copyright(C)JCPDS-International Centre for Diffraction Data 2, Advances in X-ray Analysis, Vol REFERENCES 1 J.Sherman, Spectrochim.Acta 7 (1955) W.H.McMaster, N.K.del Grande, J.H.Mallett and J.Hubbell, Compilation of X-Ray Cross Sections, UCRL-5174, Sect.11, Rev.1 Lawrence Radiation Laboratory, University of California, Livermore, CA (1969) 3 J.H.Hubbell, P.N.Trehan, Nirmal Singh, B.Chand, D.Mehta, M.L.Garg, R.R.Garg, Surinder Singh and SPuri, J.Phys.Chem.Ref.Data, 23 (1994) G.G.Johnson Jr. and E.W.White, X-Ray Emission and kev Tables for Nondiffractive Analysis, ASTM Data series DS 46, Philadelphia (197) 5 C.E.Fiori, R.L.Myklebust and K.F.J.Heinrich, Anal.Chem. 48 (1976) 172

WIDE ANGLE GEOMETRY EDXRF SPECTROMETERS WITH SECONDARY TARGET AND DIRECT EXCITATION MODES

WIDE ANGLE GEOMETRY EDXRF SPECTROMETERS WITH SECONDARY TARGET AND DIRECT EXCITATION MODES Copyright(C)JCPDS-International Centre for Diffraction Data 2000, Advances in X-ray Analysis, Vol.42 11 Copyright(C)JCPDS-International Centre for Diffraction Data 2000, Advances in X-ray Analysis, Vol.42