Attachment WP3_Updated_TlBr_Temp_Depend_FINAL_(CLEANED).pdf

White Paper #3 on TlBr:

Updated Temperature Dependence of TlBr Gamma-Ray Crystals & Associated Electronics

Note: See separate White Paper on long-term room-temperature stability.

(Version: 211206)

This work presents a summary to date of the temperature dependence of TlBr when used to measure gamma ray spectra for applications such as Radio-Isotope Identification Detectors (RIIDs) and represents a snapshot of our evolving information data base.

The performance of TlBr detectors, as with all semiconductor detectors, depends on operating temperature. The following two factors affect the energy resolution performance of the TlBr itself (before electronics): the generation of the signals at the electrodes, and the fluctuations (noise) introduced by leakage current.

The generation of the pulse signals at the electrodes depends on geometric factors, including “weighting potentials”, as well as the transport of the transient signal charge to the electrodes. The transport of the signal charge depends on the electric field and the charge mobility and lifetime. The temperature dependence of the leakage current depends on the concentration and temperature dependence of the mobility of charge carriers (including both electronic and ionic components).

The effect of temperature cycling on the spectral performance of a TlBr detector was measured for two temperature cycles using an Environmental Test Chamber (ETC) to control the temperature. Figure 1 (left) shows 137Cs pulse height raw (uncorrected for depth of interaction) spectra from the anode as a function of temperature for a 5mm × 5mm × 12mm TlBr CFG* (Capacitive Frisch Grid) device as the temperature cycles through the following values for the first temperature cycle: cycle #1: 22, 10, 22, 33, and 22 °C. Figure 1 (right) shows measurements for a repeat of the same temperature cycle (cycle #2).

The figures show vertically offset spectra to allow easy visualization of the spectral shapes.

Figure 1. 137Cs raw (uncorrected for depth of interaction) spectra recorded with a 5mm ⨯ 5mm ⨯ 12mm TlBr CFG detector as the temperature is cycled. First cycle (#1, left) and second cycle (#2, right). TPG refers to a reference signal from a Tail Pulse Generator of 18 mV pulse injected through a 1pF capacitor.

After achieving each temperature setting, the device is held at that temperature for 15 minutes before the first spectrum reading is taken. The spectrum at that temperature setting is then measured 2 more times, at 15-minute intervals. Between temperature settings, the temperature changes at a rate of

0.2 °C per minute until reaching the next temperature setting, at which time the same procedure is followed. The legend shows the nominal temperature (Tnom), which is the chamber temperature, the leakage current in nA, and the temperature from a thermistor in the aluminum box containing the CFG detector within the chamber. The bias voltage (1500 V) and shaping time (12 µs) were held constant for all these measurements. Both the detector and the readout electronics were inside the chamber, and the 137Cs source irradiated the detector from outside the chamber, a setup which produces a relatively large number of events at low energy (due to scattering) below channel number 300 in the MCA (Multi- Channel-Analyzer) spectrum compared to the number of events in the 662-keV photo-peak.

Regarding results, one first sees that the measurements in Figure 1 demonstrate that the spectrum of the TlBr CFG device changes very little with the temperatures tested (note that the spectra are not depth corrected). Also, Figure 1 demonstrates that the spectra essentially “recover” from the small changes that do occur after changing the temperature upon returning to the original temperature, confirming the integrity of the contacts over this temperature range.

The position of the photo-peak moves to slightly higher amplitudes as the temperature decreases, and the width of the photopeak narrows. These small changes in the position and resolution of the photo-peak might be attributed be an improved charge collection at the lower temperatures during the integration period.

0 500 1000 1500

C ou nt s in s

Pulse Amplitude (MCA Chan.)

Tnom, I(nA), T(°C) 22°C, 2.6, 20.7 22°C, 2.6, 20.8 22°C, 2.7, 20.9 33°C, 5.9, 32.2 33°C, 5.9, 32.3 33°C, 5.7, 32.2 22°C, 2.6, 20.7 22°C, 2.6, 20.7 22°C, 2.5, 20.8 10°C, 1.3, 9.3 10°C, 1.3, 9.3 10°C, 1.3, 9.4 22°C, 2.5, 20.7 22°C, 2.5, 20.8 22°C, 2.6, 20.7 keV

TlBr PS-CFG 171D2-5R

1.5 kV, 137-Cs Cycle #1TPG

18mV 1pF

0 500 1000 1500

C ou nt s in s

Pulse Amplitude (MCA Chan.)

Tnom, I(nA), T(°C) 22°C, 2.8, 20.7 22°C, 2.8, 20.7 22°C, 2.9, 20.9 33°C, 6.0, 32.2 33°C, 6.0, 32.3 33°C, 5.8, 32.2 22°C, 2.9, 20.8 22°C, 2.9, 20.7 22°C, 2.9, 20.8 10°C, 2.0, 9.4 10°C, 2.0, 9.3 10°C, 2.0, 9.5 22°C, 2.9, 20.7 22°C, 2.9, 20.7 22°C, 2.9, 20.8 keV

TlBr PS-CFG 171D2-5R

1.5 kV, 137-Cs Cycle #2TPG

18mV 1pF

The tail-pulse generator reference, labeled “TPG” in the plots, demonstrates constancy of the preamplifier gain with temperature and the relatively small contribution of leakage current and readout noise to the energy resolution of the spectra. The 23 °C temperature range of the experiment produces a negligible change, ~0.07%, in the 1-pF value of the coupling capacitor for the TPG reference, which is in the environmental chamber with the pre-amplifiers as its temperature coefficient is ±30 ppm/°C.

Processing the data for 3D correction factors would help correct for geometric effects, thus improving the energy resolution and possibly reducing the variations seen in the spectra at different temperatures, but 3D correction was not applied to the spectra shown above. Note that the 2.7 eV bandgap of TlBr makes the intrinsic carrier concentration and its contribution to the leakage current negligible at the measured temperatures.

Figure 2 shows the offset 137Cs raw (uncorrected for depth of interaction) pulse height spectra from the anode as a function of temperature for a different 5mm × 5mm × 12mm TlBr CFG device (denoted as 186B2-4R) as the temperature cycles through the following values sequentially in time: 20, 10, 0, -10, -20, 20, 30, 40, 50, 60, and 20 °C, nominal. Like the measurements illustrated in Figure 1, three spectra were recorded at each temperature, however, the plot in Figure 2 shows the sum of the three spectra. The temperature and leakage current illustrated in Figure 2 represent an average of the three values measured with each spectrum, and the Temperature represents the value measured from a sensor near the TlBr detector. The temperature rate, dwell times, and other conditions are the same as described for Figure 1. Also, as with the previous measurements, the 662-keV photopeak becomes more pronounced at lower temperatures and shifts slightly to the left at higher temperatures. The shape of the spectra at 20 °C is similar when comparing the start of the cycle to the middle and end of the cycle. The spectra demonstrate the capability of operating the TlBr PS-CFG detector over a temperature ranging from -20 to +56 °C.

Figure 3 shows a plot of the leakage current as a function of temperature on a log scale, where the linear increase illustrates an exponential dependence, as would be expected.

Figure 2. Offset 137Cs spectra (uncorrected), in reverse chronological order (bottom 1st), recorded with a 5×5×12 mm3 TlBr PS-CFG detector as the temperature is cycled using a bias of

1.5 kV. TPG refers to an 18-mV reference signal injected through a 1 pF capacitor.

0 500 1000 1500

5k

10k

15k

C o u n ts in s

Row Numbers

T(°C), I(nA) 19.3, 1.4 19.3, 1.4 56.5, 44 48.4, 19 40.6, 8.8 30.7, 3.5 19.1, 1.4 -20.4, 0.17 -11.1, 0.25 -1.5, 0.29 7.6, 0.56 19.3, 1.4

TPG

keV

137Cs

186B2-4R

1.5 kV 18Jan2021

The lowest temperature data points deviate slightly from the anticipated trend and are plotted in grey because condensation on the TlBr surface, which would be prevented in the future, may have contributed to the measured leakage. The change in the leakage current with temperature reflects the effect of temperature on the material resistivity (the inverse of conductivity). The conductivity depends on the mobility, which includes contributions from lattice scattering, impurity scattering, ionic conduction, and carrier concentrations (intrinsic and dopants). Ionic conduction increases the conductivity (decreases the resistivity) with temperature and thus increases the leakage current. Also, the effect of impurity scattering is reduced at increased temperatures, which increases the mobility and leakage current. The effect of lattice scattering, however, exhibits the opposite effect with temperature.

The photopeaks (uncorrected for depth of interaction) and TPG peak in the spectra of a similar detector, denoted 198B1-13, which exhibits similar spectra as those seen in Figure 2, were analyzed with a peak fitting routine in the commercial graphing program, Origin. Figure 4 plots the offset spectra. The fitting analysis of the photopeaks consisted of an asymmetric component, with a complementary error function to quantify the “tailing”, and a symmetric, Gaussian component. The purpose of the analysis was to quantify the effect of the temperature on the photopeak position, Gaussian width, and the relative width of the “tail”. Figure 5 plots the results.

Figure 5 (left) plots the centroid of the photopeak and TPG peak as a function of temperature, with the TPG position references to the left axis and the photopeak centroid referenced to the right axis, both with a range of 150 channels. The constant position of the TPG peak demonstrates the stability of the gain of the preamplifiers and coupling capacitors, which are also in the environmental chamber with the TlBr detector. The position of the photopeak decreases by ~5% over the 60 °C temperature range.

Figure 3. Temperature dependence of the leakage current (log y-axis) vs temperature. The curves are a visual guide, and the data points, below -5 °C, are grey due to possible error from additional leakage in some of the equipment presently being used.

Figure 4. Offset 137Cs spectra (uncorrected), in chronological order starting at bottom, recorded with a 5×5×12 mm3 TlBr PS-CFG detector, as the temperature is cycled using a bias of

1.5 kV. TPG refers to an 18-mV reference signal injected through a 1 pF capacitor.

0 500 1000 1500

5k

10k

C o u n ts in s)

Row Numbers

T(°C), I(nA)

19.7, 3.5

19.6, 3.5

58.0, 40

49.6, 24

41.4, 14

31.5, 7.5

19.5, 3.6

19.7, 3.8

-2.1, 0.73

7.8, 1.5

19.6, 3.5

198B1-13, Br/MeOH, Sputtered, 1.5 kV

TPG

keV

Figure 5. Left: Temperature dependence of the centroid from the TPG and from the photopeak. The curves are a visual guide. Center: Temperature dependence of the leakage current (dark current) and width of the TPG pulser. Right: the width and “tail” of the photopeak from the peak fitting analysis. The curves are a visual guide.

Figure 5 (center) plots the leakage current as a function of temperature referenced to the left axis, where the exponential behavior is apparent on the linear scale and is consistent with the trend seen in Figure 3. The plot also shows the width of the TPG peak referenced to the right axis. The width of the TPG depends on the readout noise as well as the fluctuations caused by the leakage current. The fluctuations caused by the leakage current, and the TPG width, are expected to increase with the temperature and leakage current, as illustrated in the plot.

Figure 5 (right) plots the width of the Gaussian component, referenced to the left axis, and the relative photopeak tail, referenced to the right axis, as a function of temperature. The relative photopeak tail corresponds to a relative width of the asymmetric component, or low energy “tail”, of the photopeak. As seen in the plot, the Gaussian component changes by ~40% over the 60 °C temperature range. Correcting the data for the interaction depth often reduces the amount of asymmetry, or “tailing,” of the photopeak in the spectrum. Again, the plotted data is not depth corrected.

Note: There are several ways, both passive and active, to minimize changes in TlBr detector performance with varying operating temperature. These methods could either reduce the magnitude of changes in temperature or maintain a constant (e.g., room-temperature) temperature of the detector, much as is done with CZT detectors. Temperature control would also reduce effects from the 56 ppm/°C thermal expansion coefficient of TlBr. Again, in the experiments described herein, no passive or active temperature controls were implemented since the aim of these experiments was to extract the explicit temperature dependencies of the TlBr crystals and associated electronics, were those temperatures to be experienced in actual fielded operation.

* The devices were configured as PS-CFG (Position Sensitive – Capacitive Frisch Grid) detectors but the position-sensitive functionality was not utilized, and the 3D depth correction was not performed on this data. They were operated as simple CFG configurations.

0 10 20 30 40 50 60

TPG

Photopeak

Temperature (°C)

T P

G C en tr o id

C ha n ne l)

P h ot o pe ak

C en tr oi d (C h an ne l)

0 10 20 30 40 50 60

Temperature (°C)

Leakage TPG width

L ea ka ge nA

T P

G W id th

C h an n el

R M

S

0 10 20 30 40 50 60

5 Gaussian Width Relative Tail

Temperature (°C)

G au ss ia n P ho to p ea k W id th

F W

H M

R el at iv e P h ot o pe ak T a il