Scholarly article on topic 'A varied shaping time noise analysis of Al0.8Ga0.2As and GaAs soft X-ray photodiodes coupled to a low-noise charge sensitive preamplifier'

A varied shaping time noise analysis of Al0.8Ga0.2As and GaAs soft X-ray photodiodes coupled to a low-noise charge sensitive preamplifier Academic research paper on "Physical sciences"

CC BY
0
0
Share paper
Keywords
{AlGaAs / Detector / Diode / GaAs / Photodiode / Preamplifier / X-ray}

Abstract of research paper on Physical sciences, author of scientific article — A.M. Barnett, J.E. Lees, D.J. Bassford, J.S. Ng

Abstract The noise sources affecting Al0.8Ga0.2As and GaAs spectroscopic X-ray photon counting p+–i–n+ photodiodes connected to a custom low-noise charge sensitive preamplifier are quantified by analysing the system's response to pulses from a signal generator and varying the system's shaping amplifier's shaping time (from 0.5μs to 10μs). The system is investigated at three temperatures (−10°C, +20°C and +50°C) in order to characterise the variation of the component noise sources and optimum shaping time with temperature for Al0.8Ga0.2As and GaAs diodes. The analysis shows that the system is primarily limited by dielectric noise, hypothesised to be mainly from the packaging surrounding the detector, for both types of diode and at each temperature.

Academic research paper on topic "A varied shaping time noise analysis of Al0.8Ga0.2As and GaAs soft X-ray photodiodes coupled to a low-noise charge sensitive preamplifier"

vi.

ELSEVIER

Contents lists available at SciVerse ScienceDirect

Nuclear Instruments and Methods in Physics Research A

journal homepage: www.elsevier.com/locate/nima

NUCLEAR INSTRUMENTS A METHODS IN

PHYSICS HE5EARCH

A varied shaping time noise analysis of Alo.8Gao.2As and GaAs soft X-ray photodiodes coupled to a low-noise charge sensitive preamplifier

A.M. Barnetta*, J.E. Leesa, D.J. Bassforda, J.S. Ngb

a Space Research Centre, Department of Physics and Astronomy, Michael Atiyah Building, University of Leicester, Leicester LEI 7RH, UK b Department of Electronic and Electrical Engineering, University of Sheffield, Mappin Street, Sheffield SI 3JD, UK

ARTICLE INFO ABSTRACT

The noise sources affecting Al08Ga02As and GaAs spectroscopic X-ray photon counting p+-i-n+ photodiodes connected to a custom low-noise charge sensitive preamplifier are quantified by analysing the system's response to pulses from a signal generator and varying the system's shaping amplifier's shaping time (from 0.5 ms to 10 ms). The system is investigated at three temperatures (-10 °C, + 20 °C and + 50 °C) in order to characterise the variation of the component noise sources and optimum shaping time with temperature for Al0.8Ga0.2As and GaAs diodes. The analysis shows that the system is primarily limited by dielectric noise, hypothesised to be mainly from the packaging surrounding the detector, for both types of diode and at each temperature.

© 2012 Elsevier B.V. All rights reserved.

Article history:

Received 22 September 2011 Received in revised form 4 January 2012 Accepted 6 January 2012 Available online 17 January 2012

Keywords:

AlGaAs

Detector

Photodiode

Preamplifier

1. Introduction

Wide band gap compound semiconductor photodiodes for photon counting X-ray spectroscopy in high temperature and intense radiation environments have attracted increased attention in recent years, with Al08Ga02As [1-4], GaAs [5] and SiC [6-12] among the materials which have had soft X-ray spectroscopic results reported at temperatures b 20 °C. High temperature and radiation tolerant soft X-ray detectors are likely to have terrestrial and space applications, including real time oil condition monitoring [13] in high value mechanical machinery (including railway locomotives, aircraft, ships, Formula 1 racing cars and military vehicles), in situ analysis of geological materials around active hydrothermal vents, planetary X-ray fluorescence spectroscopy missions to hot extraterrestrial environments such as the surface of Mercury and Venus and in extreme radiation environments such as those that would be encountered in missions to study the Jovian [14] or Saturnian [15] aurorae, or to study X-ray emissions from Jupiter's Galilean moons [16].

Al08Ga02As and GaAs photon counting soft X-ray p+-i-n+ photodiodes operating at temperatures from -30 to + 90 °C and from - 30 to + 80 °C have been previously reported by Barnett et al. [1,5]. In both cases the detectors were coupled to the same charge sensitive preamplifier electronics. The X-ray spectral

* Corresponding author. Tel.: +44 116 223 1 042; fax: +44 116 252 2464. E-mail address: amb67@le.ac.uk (A.M. Barnett).

0168-9002/$-see front matter © 2012 Elsevier B.V. All rights reserved. doi:10.1016/j.nima.2012.01.009

performance, as measured by the FWHM of the Mn Ka (5.9 keV) peak from an 55Fe radioisotope source, was reported as 0.9-2.5 keV and 0.8-1.6 keV for the Al0.8Ga0.2As and GaAs diodes, respectively. The initial noise analyses presented in Refs. [1,5] indicated that the majority of the noise (spreading of the detected Mn Ka peak beyond that expected if the resolution was Fano limited) came from a source other than the parallel white noise from the leakage current of the diodes. It was reported that it was suspected that a significant portion of the noise was dielectric noise from the packaging of the diodes. However, this contribution was not separated from other sources, such as the series white noise, so the hypothesis was not tested.

Using a pulse generator and collecting spectra at various shaping amplifier shaping times (0.5 ms r t r 10 ms) and temperatures (-10, + 20 and + 50 °C), we calculate the individual contributions of the parallel white, series white, 1/f and dielectric noises [17] for the Al08Ga02As and GaAs diodes reported in Refs. [1,5].

2. Noise sources in X-ray photodiodes

The fundamental (statistically limited) X-ray spectral resolution (FWHM in eV) of a photodiode is

DE [eV] = 2.35o/ — (1)

where o is the energy required to create an electron-hole pair in the diode material, E is the energy of the incident X-ray and F is the Fano factor [18], which quantifies the observed deviation in number of electron-hole pairs created from the absorption of a photon of given energy from that predicted by Poissonian statistics.

This fundamental spectral resolution is degraded by terms R and A, defined below, causing the practical spectral resolution of a semiconductor X-ray detector to become

DE [eV] =

The factor R is the equivalent noise charge (in r.m.s. e_) introduced by the detector during the movement of the charge to the contacts (e.g. by charge trapping), and A is the equivalent noise charge (in r.m.s. e") introduced by the detector's leakage current, capacitance and the properties of preamplifier [17]. Assuming that peak broadening due to partial charge collection is negligible (the 5.9 keV peaks observed with the Al0.8Ga0.2As and GaAs detectors are Gaussian (see Fig. 4 of Ref. [5]), which would be unexpected if partial charge collection or trapping were dominant factors [19-21]), the measured noise beyond that predicted by Eq. (1) can be attributed solely to A, which is a combination of the parallel white (NPW), series white (NSW), induced gate current, 1/f (Nf and dielectric noise (ND) contributions, defined below. A comprehensive introduction to the various electronics noise source contributions in photon counting X-ray photodiodes coupled to charge sensitive preamplifiers can be found in Ref. [17], the salient points of which are summarised in Sections 2.1-2.5 to give equations for the contributions of the different noise components. Further discussion and results regarding the disentangling of the noise components affecting semiconductor radiation detectors and their preamplifiers can also be found in Ref. [22].

2.1. Parallel white noise

The parallel white noise is from the shot noise of the currents flowing through the input node of the preamplifier. It is primarily dependent on the leakage currents of the detector, ILD, and the preamplifier input field effect transistor (FET), ILT [17].

The parallel white noise power spectral density, SPW, can be expressed as

Spw = 2q(/iD + Ilt )+ — (3)

where q is the charge on an electron, k is the Boltzmann constant, T is the temperature (in K) and r is the resistance of the preamplifier feedback resistor (if the preamplifier has one) [17]. The preamplifier used in this work does not have a feedback resistor so the 4kT/r term is omitted. An example design of a charge sensitive preamplifier without a feedback resistor can be found in Ref. [23]. The contribution (measured in r.m.s. e~) of SPW to the equivalent noise charge A (Eq. (2)) is

Npw = — V (A3/2)Spw t

where A3 is a constant dependent on signal shaping function [24], and t is the shaping time [17].

2.2. Series white noise

drain current [17]: c 4kT

Ssw = g — (5)

where 0.7 r g r 1 depending on FET characteristics, and gm is the transconductance of the FET [17]. The contribution (measured in r.m.s. e_) of Sws to the equivalent noise charge, A, (Eq. (2)) is

Nsw = -yj (Ai/2)SSWC2(1/t)

where A1 is a constant depending on signal shaping function [24] and CT is the total capacitance at the preamplifier input (= Cd+Ci+Cf + Cs, where Cd is the detector capacitance, Ci is the input transistor capacitance, Cf is the feedback capacitance and Cs is the stray capacitance) [17]. The equivalent noise charge contribution from NSW becomes increasingly significant at shorter shaping times because of the 1/t dependence.

2.3. 1/f noise

The noise from the drain current of the preamplifier input FET is also the main constituent of 1/f noise. The contribution to the equivalent noise charge A (Eq. (2)) is

ENCyf = 1v/A2pA/C? (7)

where Af is a characteristic constant dependent on the FET and A2 is a constant ranging from 0.64 to 2 depending on signal shaping function [17,24,25].

2.4. Dielectric noise

Dielectrics in close proximity to the preamplifier, such as the packaging of the FET and detector contribute

Nd = A22kTDCdie (8)

to the total electronics equivalent noise charge, A, where Cdie is the capacitance of the dielectrics, D is the dielectric dissipation factor [17] and q, A2, k and T have all been previously defined. It is therefore desirable to design the input FET and detector packaging to minimise exposure to dielectrics, for example by reducing the capacitance of the FET and detector assembly by integrating the FET onto the detector.

2.5. Induced gate current noise

Drain current noise (Section 2.2) causes charge fluctuations in the FET gate current, this gives rise to the induced gate current noise. The contribution from this to A is dependent on Ssw (Eq. (5)). Like the series white noise, the induced gate current noise becomes important at short shaping times because of the 1/t dependence. Bertuccio et al. state that Eq. (5) can be modified by a factor, Gc, (OGc e 0.8) to take account of this noise in the FET's gate [17]:

SSWC = SSWGc ■ (9)

Consequently the contribution to A, becomes

Nswc = (A!/2)SswCcC2(1/t) = Nsw pGc (10)

The series white noise primarily arises from the effect of thermal noise on the drain current of the input FET [17]. When secondary sources (e.g. stray resistance in series with the input FET's gate) are negligible, the series white noise power spectral density, SSw, can be approximated to the thermal noise of the FET

2.6. Electronic noise sources in combination

When considered together, the parallel white (Section 2.1) and series white (Section 2.2) noise contributions' dependences on t

and 1/t, respectively, can give rise to an optimum shaping time, which minimises the combined noise from these sources [26]. In Section 4, it is shown how the parallel white, series white including induced gate current, 1/f and dielectric noise components combine to produce the overall FWHMs reported in Refs. [1,5] and how these are affected by varying the shaping time, t.

3. Method

The previously reported [1,5] circular (100 mm radii) mesa Al0.8Ga0.2As D1 (Table 1) and GaAs D9 (Table 2) diodes, which were mounted in TO-5 packages (gold plated, 9 mm diameter, 12 pins, see Fig. 1), were each in turn connected to a charge sensitive preamplifier, and individually reversed biased at 10 V by a Thurlby Thander PL330QMD stabilised power supply. The preamplifier used a Si JFET (Vishay Siliconix 2N4416, capacitance=2 pF) as the input transistor [27]. The preamplifier was connected to an EG&G Ortec 571 shaping amplifier [28] whose output signal was connected to an Ortec multichannel analyser (MCA) [29]. A pulse generator was connected to the system. A block diagram of the electronics chain is given in Fig. 2. The diodes and preamplifier were placed in a Design Environmental FS55-65 Temperature Test Chamber to control their operating temperature. The diodes and preamplifier were assumed to be in thermal equilibrium with each other, and their temperatures were monitored by thermocouple and Fluke 50D digital thermometer. The MCA scale was calibrated in energy terms for each diode using the zero energy noise peak and the Mn Ka line of an 55Fe radioisotope X-ray source with which the diodes were illuminated as per Refs. [1,5].

To assess the shaping time and temperature dependence of the electronics' noise, spectra were accumulated with shaping amplifier shaping times, t, of 0.5,1, 2, 3, 6 and 10 ms at temperatures of + 50, + 20 and -10 °C. By measuring the FWHM of the peak produced by the pulse generator, the performance of the electronics coupled to each diode can be characterised. Since the electronics' performance depends on the properties of the detector to which it is connected, the detectors remained connected to the system and reverse biased (10 V) during these measurements.

Unlike the parallel white (NPW), series white including induced gate current (NSWC) and the 1/f (N1/f) noise contributions, the

Table 1

Layer details of the Al0.sGa0.2As diodes [1].

Layer Material Thickness Dopant Type Doping density

(mm) (cm-3)

1 GaAs 0.01 Be p 2.5 x 101s

2 Al0.sGa0.2As 1 Be p 2.0 x 101s

3 Al0.sGa0.2As 1 Undoped

4 Al0.8Ga0.2As 1 Si n 2.5 x 101s

5 GaAs 0.25 Si n 2.5 x 101s

Substrate n+ GaAs

Table 2

Layer details of the GaAs diodes [5].

Layer Material Thickness (mm) Dopant Type Doping density (cm-3)

1 GaAs 0.01 Be p++ 1.0 x 1019

2 GaAs 0.2 Be p+ 2.0 x 101s

3 GaAs 2 Undoped < 1015

4 GaAs 0.1 Si n+ 2.0 x 101s

5 GaAs 0.2 Si n+ 2.0 x 101s

Substrate n+ GaAs

Fig. 1. Al0.sGa0.2As chip mounted on a TO-5 package.

Fig. 2. Block diagram of the electronics chain.

dielectric noise (ND) is not readily calculable because the stray capacitances are not easily measured. However, by calculating the noise contributions from NPW, NSwc and N1/f, and subtracting these in quadrature from the total noise, NT, shown by the FWHM of the pulser peak, it is possible to obtain an estimate for the dielectric noise, ND, assuming any noise sources other than those mentioned are insignificant.

The parallel white noise contribution for each shaping time and temperature was calculated from Eqs. (3) and (4), assuming signal shaping function constant A3 = 2 [24,28]. The series white noise contribution was calculated from Eqs. (5) and (6) assuming g=0.85 [17], transconductance of the FET gm = 5 mS [27] and that the total capacitance at the preamplifier input, CT, is dominated by the capacitances of the detector (Al08Ga02As diode: 3.1 pF, GaAs diode: 1.94 pF) and FET (2 pF). The capacitances of the detectors were measured before packaging by directly probing the devices using a Hewlett Packard 4275 LCR metre with the AC test voltage signal magnitude and frequency set at 50 mV r.m.s. and 1 MHz, respectively. The series white noise contribution was adjusted for induced gate current noise (Eqs. (9) and (10)) by assuming OGc e 0.8 [17]. The 1/f noise contribution was calculated from Eq. (7) with the above assumptions and also that A2 = 1.2 and Af = 3 x 10-15 V2 [24].

4. Shaping time results

The calculated parallel white, series white (including induced gate current noise) and 1/f noise components together with the measured pulser peak FWHM and estimated dielectric noise

1.6n 1.4 « 1.2 H

£ 0.«

S 06 H

0.4 0.2 H 0

_ -kr -

X-X---X---X----------X-

---------x

Shaping time (|s)

1.6 1.4

> 1.2 \ ^

5 0.6 as

j? 0.4

0.2 H 0

1.61 1.4 1.21 -

0 0.6 H 0.4 0.2 0

-X---- ------- ----X

6 8 10

Shaping time (|s)

----------

---------x

Shaping time (|is)

Fig. 3. FWHM versus shaping time data for the preamplifier and Al0 8Ga0 2As diode at (a) 50 °C, (b) 20 °C and (c) -10 °C. Pulser FWHM (circles), dielectric noise (squares), parallel white noise (triangles), series white noise (diamonds), 1/f noise ( x symbols). Lines are guides for the eyes only.

contribution (ND=N2 - NpW- N|WC- N2/f) at each shaping time for temperatures + 50, + 20 and -10 °C are shown for Al0 8Ga0.2As in Fig. 3(a)-(c), and for GaAs in Fig. 4(a)-(c), respectively. Even though the preamplifier electronics used for the Al0 8Ga0.2As and GaAs devices are the same, their measured performances are somewhat different because they are dependent on the diodes' properties.

With both types of diode, the contribution from the parallel white noise decreases and the series white contribution increases as the shaping time is lengthened at all three temperatures. The increase in parallel white noise with longer shaping time is more significant at higher temperature because of the temperature dependences of the leakage currents in the detectors (see Fig. 3 of Ref. [5]) and FET. The series white noise contribution can be seen decreasing with increasing shaping time, as expected from its 1/t dependence (Eq. (6)). The 1/f noise contribution is not shaping time dependent and is, therefore, constant (Eq. (7)).

For both diode types, the shaping time at which the functions representing the parallel white and series white noise contributions intersect lengthens with decreasing temperature (t=2, 5.25, 8.45 ms at 50, 20 and -10 °C for Al0.8Ga0.2As; t=2, 5.4, 7.8 ms at 50, 20 and -10 °C for GaAs). For the Al0 8Ga0.2As detector, there is a clear lengthening of optimum shaping time with decreasing temperature (t=2, 3, 6 ms at 50, 20 and -10 °C, respectively), but the system's optimum shaping time dependence on temperature is less clear with the GaAs detector because the minima representing the optimum shaping time at 20 and - 10 °C are not so well defined. However, a general trend of lengthening optimum shaping time with decreasing temperature is still observable.

5. Discussion

The data shown in Figs. 3 and 4 suggest that dielectric noise is the most significant noise source for the temperature range investigated.

The estimated dielectric noise is broadly similar across the measured shaping times as would be expected from a shaping time independent contribution (Eq. (8)). However, slight deviations from constant are apparent.

Apparent increases in dielectric noise at short shaping times may be attributable to underestimating NSwc (Eq. (10)), possibly as a consequence of underestimating the total capacitance at the preamplifier input i.e. the stray capacitance may not be insignificant compared with those of each of the diodes, 3.1 pF (Al0.8Ga0.2As) and 1.94 pF (GaAs), and FET (2 pF). However, if the apparent increases in dielectric noise are attributable to underestimating the significance of the stray capacitance, it is unclear why only some measurements show rises at short shaping times. The slight increase shown in calculated dielectric noise at long shaping times (particularly Al0.8Ga0.2As at 20 and - 10 °C) could be attributed to an underestimate of the parallel white noise as a consequence of underestimating the leakage current of the FET (the FET's parameters were taken from its data sheet rather than being measured directly in the lab). The leakage currents of the FET and detector are greater and more significant at higher temperature, explaining why little apparent increase is observed in the parallel white noise at long shaping times in the -10 °C data.

A computer model of the noise, which varies parameters such as the additional stray capacitances (capacitances in addition to those

H 0.8-s

---X----------X---

----♦

Shaping time (^s)

JL 0.8 H

__—1

*---X----------X----

______A

------»

------X

Shaping time (^s)

..V-V-v*-----— -------x-----

Shaping time (^.s)

Fig. 4. FWHM versus shaping time data for the preamplifier and GaAs diode at (a) 50 °C, (b) 20 °C and (c) -10 °C. Pulser FWHM (circles), dielectric noise (squares), parallel white noise (triangles), series white noise (diamonds), 1/f noise ( x symbols). Lines are guides for the eyes only.

2.5 2 1.5

-,—e-

o Al0.8 Ga0 2As D1 A GaAs D9 i Al08Ga02As D1 without ND * GaAs D9 without N

-40 -30 -20 -10 0 10 20 30 40 50 60 70 80 90 Temperature (C)

Fig. 5. Measured FWHM at 5.9 keV of the system with Al08Ga02As (open circles, [1]) and GaAs (open triangles, [5]) diodes and predicted FWHM for each if the dielectric noise was completely eliminated (Al0sGa0 2As: filled circles, GaAs: filled triangles).

of the detector and FET), to fit the experimental data, may produce better estimates for the individual noise contributions of the systems, but the analysis presented here is sufficient to show that the dielectrics around the detector and FET may be a very significant source of noise in the system and that, if the dielectric noise could be reduced, the performance of the system is likely to improve.

To demonstrate the importance of reducing the dielectric noise contribution, Fig. 5 shows the measured spectral resolutions (FWHM at 5.9 keV) at a shaping time of 3 ms for the Al0.8Ga0.2As and GaAs diodes [1,5] and the expected spectral resolutions if the

dielectric noise contributions were to be eliminated. Under this circumstance, spectral resolutions of 380 and 370 eV FWHM at 5.9 keV are predicted, compared to 1.07 keV and 800 eV with the current dielectric noise contributions, for Al0.8Ga0.2As and GaAs, respectively, at t=3 ms.

6. Conclusions and future work

New investigations of the noise sources contributing to the spectral resolutions of Al0.8Ga0.2As and GaAs diodes reported in Refs. [1,5] have shown that dielectric noise, thought to arise primarily from the diode and FET packaging, is the most significant factor in limiting the spectral resolutions. If the dielectric noise could be eliminated through improvements to the diode and FET packaging (e.g. custom designed packaging, perhaps using ceramic or Teflon mountings), spectral resolutions (FWHM at 5.9 keV) of 380 and 370 eV at 20 °C at a shaping amplifier shaping time constant of 3 ms are predicted for the Al0 8Ga0 2As and GaAs diodes, respectively. Future investigations of the noise contributions of the system with low-noise diode packaging and heating/cooling the diode only, while holding the preamplifier electronics at constant temperature will further elucidate the temperature dependence of the noise components in Al08Ga02As and GaAs X-ray photodiodes.

Acknowledgements

A.M. Barnett acknowledges the individual financial support received from Science and Technologies Facilities Council during the course of this study. The authors jointly acknowledge the support of

Science and Technologies Facilities Council through IPS awards (ST/ H000127/1 and ST/H000143/1). J.S. Ng acknowledges support from the Royal Society through her University Research Fellowship.

References

[1] A.M. Barnett, et al., Nuclear Instruments and Methods A 621 (2010) 453.

[2] J.E. Lees, et al., Nuclear Instruments and Methods A 594 (2008) 202.

[3] A.M. Barnett, et al., Nuclear Instruments and Methods A 626 (2011) 25.

[4] A.M. Barnett, et al., Nuclear Instruments and Methods A 629 (2011) 154.

[5] A.M. Barnett, et al., Nuclear Instruments and Methods A 654 (2011) 336.

[6] J.E. Lees, et al., Journal of Instrumentation 6 (2011) C01032.

[7] J.E. Lees, et al., Nuclear Instruments and Methods A 578 (2007) 226.

[8] J.E. Lees, et al., Nuclear Instruments and Methods A 604 (2009) 174.

[9] G. Bertuccio, R. Gasiraghi, IEEE Transactions on Electron Devices ED 50 (2003) 175.

[10] G. Bertuccio, et al., Nuclear Instruments and Methods A 518 (2004) 433.

[11] G. Bertuccio, et al., Nuclear Instruments and Methods A 522 (2004) 413.

[12] G. Bertuccio, et al., Nuclear Instruments and Methods A 652 (2011) 193.

[13] Z. Yang, et al., Talanta 59 (2003) 673.

[14] G.R. Gladstone, et al., Nature 415 (2002) 1000.

[15] J.-U. Ness, et al., Astronomy and Astrophysics 41 (2004) 337.

[16] R.F. Elsner, et al., Astrophysical Journal 572 (2002) 1077.

[17] G. Bertuccio, et al., Nuclear Instruments and Methods A 380 (1996) 301.

[18] U. Fano, Physical Review 72 (1947) 26.

[19] T.A. McMath, M. Martini, Nuclear Instruments and Methods 86 (1970) 245.

[20] A.D. Short, A. Keay, M.J.L. Turner, SPIE 3445 (1998) 13.

[21] A.D. Short, R.M. Ambrosi, M.J.L. Tuner, Nuclear Instruments and Methods A 484 (2002) 211.

[22] G. Bertuccio, A. Pullia, Review of Scientific Instruments 64 (1993) 3294.

[23] G. Bertuccio, P. Rehak, D. Xi, Nuclear Instruments and Methods A 326 (1993) 71.

[24] E. Gatti, et al., Nuclear Instruments and Methods A 297 (1990) 467.

[25] A. Pullia, Nuclear Instruments and Methods A 405 (1998) 121.

[26] A. Pullia, G. Bertuccio, Nuclear Instruments and Methods A 380 (1996) 1.

[27] Siliconix, 2N4416/2N4416A/SST4416 N-Channel JFETs, Data Sheet, S-52424, Rev. E, Vishay Electronic GmbH, Selb, Germany, 14 April 1997.

[28] EG&G Ortec, Model 571 Spectroscopy Amplifier Operating and Service Manual, Rev. A, Part no. 733490, Oak Ridge, Tennessee, USA.

[29] Advanced Measurement Technology Inc., Ortec Easy-MCA-8k Digital Gamma-Ray Spectrometer Hardware User's Manual, Manual Revision A, Ortec Part no. 931044, Oak Ridge, TN, USA, 2008.