WP1_TlBr_Stability_FINAL.pdf

White Paper #1 on TlBr (External):

Stability (long term, “room temperature” operation)

Note: See separate White Paper on temperature stability.

(Version: 200825b)

RMD is developing two configurations of TlBr crystal arrays for Radio‐Isotope Identification (RIID) instruments. One is the pixelated TlBr arrays, where the core detector module (CDM) consists of two 2

TlBr crystals, each with an 11×11 pixelated area with approximate dimensions of 2cm × 2cm × 1cm. RMD is also developing crystals for Position Sensitive ‐ Capacitive Frisch Grid (PS‐CFG) arrays for RIID instruments, where the core detector module (CDM) includes 16 TlBr crystals, each with dimensions of

5mm x 5mm x 12mm.

As part of the development process, RMD is studying the stability of these detector configurations. In the meanwhile, however, RMD has a considerable amount of information on earlier, smaller detector configurations. The purpose of this document is to summarize RMD’s data on the stability of a variety of

TlBr device configurations.

The stability of devices depends strongly on the electrode metal contacts; namely, what metal materials are used, how the surface of the TlBr is prepared before application of the electrodes, and how the metals are applied.

Hence, the following discussion is organized by not only crystal form‐factor, but also by surface preparation (process) and electrode material. Below, data from the following detector configurations is presented: planar, pixelated arrays, and Position Sensitive ‐ Capacitive Frisch Grid (PS‐CFG) detectors.

In the descriptions below, RMD shows the data on 1‐mm thick planar devices with 3‐mm‐diameter electrodes. Data is also presented on two designs of pixel arrays: (i) 3x3 pixels with 1 mm pixel size, 5‐ mm thickness and (ii) 3x3 pixels with 1.74 mm pixel size, 10‐mm thickness. Finally, data is provided for the PS‐CFG detectors that are 5 mm × 5 mm × 12 mm in size.

Current Surface Processing and Choice of Electrode Metal:

Pixelated:

The pixelated array devices are fabricated using two main processes described below:

Process #1: etching with etchant#1, and a specific combination (#1) of anode and cathode electrodes

Process #2: etching with etchant#2, and a specific combination (#2) of anode and cathode electrodes

CFG:

Currently, the PS‐CFG crystals are prepared using process #1.

Process #1 is a modified version of RMD’s legacy process. The devices fabricated with RMD’s legacy process would typically last for ~100 days. By modifying the electrodes in the legacy process, Process #1 has been developed to improve the stability.

Process #2 was developed as a parallel process to achieve long‐term stability.

a) Results on 1‐mm thick planar TlBr detectors

i. Process #1

The 1‐mm thick planar TlBr samples are the earliest samples that had been monitored for stability. For these, stability is defined as a pass/fail test wherein it passes if the amplitude of the photo‐peak for the

60‐keV peak from a 241‐Am source is within 10% of its nominal value after an initial conditioning period

[see Note 1]. A continuous bias is applied to the detectors at room temperature and spectra are periodically recorded. Table 1 shows the time for which the detectors operated stably after which they were removed to utilize the test equipment for other newer devices.

Table 1. Stability summary for 1‐mm thick Planar TlBr sensors under continuous bias:

Sample Config.

Thickness mm Process

Bias (V) Days Years

1 Planar 1 1 100, 140 2282 6.2

2 Planar 1 1 100, 140 1343 3.7

3 Planar 1 1 100, 140 1343 3.7

4 Planar 1 1 100, 140 1343 3.7

5 Planar 1 1 100, 140 1343 3.7

6 Planar 1.57 2 200 600 1.6

The table also includes the operating bias. The following figures show examples of the stability data as a function of time for devices fabricated using process #1. Figure 1 shows the relative 60‐keV peak position as a function of time for sample #1, where RMD changed the operating bias from 100 V to 140 V after

~1400 days.

Figure 1. 60‐keV photo‐peak channel plotted versus time under bias for a 1‐mm thick planar TlBr detector #1 fabricated with process #1.

Figure 2 shows a plot of the stability of planar 1‐mm thick TlBr devices fabricated with process #1 (referred to in plot as “new electrode”) as a function of time. Results are also included for a control sample, labeled as "control electrode" in the plot, with legacy process. This control device with the legacy process polarized rapidly as expected, while the four devices fabricated with process #1 survived much longer.

This data demonstrates that modifying the legacy process by choosing appropriate electrodes (to create process #1) mitigated the polarization of the devices, extending the stability to beyond 1000 days.

ii. Process #2

Figure 3 plots the location of the 60‐keV photo‐peak as a function of time for device #6, which was prepared using process #2. As seen in the plot, the detector demonstrates stable performance for ~600 days. In this case, and many other cases, such experiments were terminated due to demands on the setup and not due to device failure.

Figure 2. 60‐keV photo‐peak channel plotted versus time under bias for four 1‐mm thick planar TlBr detectors (#2, #3, #4, and #5) with process #1, and the device labeled "control electrode", which is a one 1‐mm thick planar TlBr detector with RMD’s legacy process (blue data).

Figure 3. 60 keV photo‐peak channel as a function of time under bias for a 1.57‐mm thick planar TlBr detector (#6) fabricated using process #2. This detector was operated with an applied electric field of 1.27 kV/cm.

b) 5‐mm thick (1‐mm & 1.25‐mm pitch) TlBr arrays

Table 2 summarizes the results for selected 5‐mm thick TlBr array samples, showing the sample identification number, the operating bias, and number of days that the device had been operating. For the array samples, RMD defines the stability in terms of the nominal position of the 662‐keV photo‐peak from 137‐Cs irradiation after an initial conditioning period [see Note 1]. As with the 60‐keV peak, the stability is a pass/fail test where the device passes when variations in the photo‐peak position are less than 10% of its nominal value after an initial conditioning period. The first two samples in the table (#7 and #8) have a 1‐mm pitch between the pixels, while the third sample (#9) has a 1.25‐mm pitch between the pixels.

Table 2. Stability summary for 5‐mm thick TlBr arrays:

Sample Config.

Thickness mm Process

# Vbias Days

7 Array 5 1 1000 248

8 Array 5 2 1000 251

9 Array 5 2 1500 386

Figure 4, left, shows the peak position for the average of the 9 pixels in a 5‐mm thick 3 × 3 TlBr array

(sample #7), fabricated with process #1, versus the number of days of operation. The plot on the right shows the estimated FWHM from measurements of the uncorrected anode spectra as a percentage of the 662‐keV photo‐peak for the pixels versus the days of operation. This experiment was terminated to use the equipment and lab space for other experiments.

Figure 4. Left: Photo‐peak position and Right: energy resolution as a function of time under bias for a 5‐mm thick TlBr array (sample #7) fabricated with process #1.

0%

20%

40%

60%

80%

100%

120%

140%

0 50 100 150 200 250

N o rm al iz e d A ve ra ge P e ak P o si ti o n

Days of Operation

0.0

2.0

4.0

6.0

8.0

10.0

0 50 100 150 200 250

FW

H M a t ke V

Days of Operation

i. Process #2

Figure 5 (left) shows the position of the 662‐keV photo‐peak, averaged over all of the 9 elements in the array (sample #8), as a function of the number of days of operation. This 5‐mm thick array was fabricated with process #2. Figure 4 (right) shows the energy resolution value (in % FWHM of the 662 keV peak) for the array elements, estimated from the uncorrected (raw) energy spectra.

Figure 5. Left: Photo‐peak position and Right: energy resolution as a function of time under bias for a 5‐ mm thick TlBr array (sample #8) fabricated with process #2.

Figure 6 shows the depth corrected spectrum of sample #9 for the “best pixel” (2 × 2 pixels, 1.25‐mm pitch, 5‐mm thick) after 386 days of operation. The FWHM was 1.1% after 386 days, and the applied field was 3 kV/cm. The sample was fabricated using process #2.

c) 10‐mm thick (1.72‐mm pitch) TlBr arrays

Table 3 summarizes the results for selected 10‐mm thick

TlBr pixelated array samples, showing the sample identification, the fabrication process, the operating bias and the number of successful operating days.

Table 3. Stability summary for 10‐mm thick TlBr arrays:

Sample Config.

Thickness mm

Process # V bias

Days

10 Array 10 1 1000, 1500 212

11 Array 10 2 1000 240

12 Array 10 2 1000 237

13 Array 10 2 1000 200

14 Array 10 2 1000 150

15 Array 10 2 1000 180

Figure 6. Depth corrected 137Cs spectrum recorded after 386 days of operation from a pixel of a 5‐mm thick array (#9) fabricated with process #2.

Figure 7 shows the anode spectra for a 3 x 3 pixelated array (sample #10) after 171 days of operation.

This is a 10‐mm thick array fabricated with process #1, where the best energy resolution for the 137Cs photo‐peak is estimated to be 1.8% for an uncorrected (no depth‐of‐interaction correction applied) spectrum.

Figure 7. Raw 137Cs (not depth of interaction corrected) spectra from a 10‐mm thick TlBr array (sample #10) fabricated with process #1. These spectra were recorded after 171 days under bias. The applied electric field was 1.5 kV/cm.

ii. Process #2

Figure 8 shows the depth‐corrected 137Cs spectra, accumulated over all 9 pixels, for the following samples of 10‐mm thick TlBr sensors: Left: #15, Center: #14 Right: #13. The depth‐corrected spectra exhibit an energy resolution better than 2%.

500 1 2 3

500 4 ~1.8%

( ) y

5 6

0 500 1000

N um be r of E ve nt s

Pulse Amplitude (Channel) 0 500 1000

0 500 1000

Figure 8. Depth corrected 137Cs spectra from single‐pixel events for three 10‐mm thick full arrays fabricated using process #2. Left: #15, Center: #14 Right: #13. The arrays (from left to right) operated for more than 180, 150 and 200 days, respectively.

d) 5‐mm × 5‐mm × 12 mm Position‐Sensitive Capacitive Frisch Grid (PS‐CFG) TlBr sensor

Figure 9 plots the data from an on‐going measurement campaign for monitoring the stability of a PS‐CFG detector, sample #16, fabricated using process #1 under a continuous bias of 1500 V at room temperature.

The electric field is 1.25 kV/cm. After the initial spectrum, labeled 03Apr, the 662‐keV photo‐peak separates from the shoulder associated with the escape peak. This initial change represents the conditioning [see Note 1] of the sample.

Figure 9. Left: Raw anode spectra from a PS‐CFG detector (#16), continuously biased at 1500 V and irradiated with 137‐Cs for several dates in 2020, in chronological order, where the top curve is the earliest.

Right: Plot of the peak position and width for the 662‐keV photo‐peak and the tail‐pulse generator (TPG) reference as a function of the number of days under continuous bias. The photo‐peak was fitted to a Gaussian peak convolved with a “reversed” exponential, i.e., a pulse function with a “low side” tail. The dashed lines in the plot are a guide for the eye.

While the plot shows subtle changes in the spectrum over time, the plot on the right indicates that the photo‐peak position and width from a peak‐fitting analysis are relatively stable after the initial conditioning, where the subtle changes in the raw spectra can be attributed to changes in the tailing of the peak. The energy resolution depends on the applied bias, and we anticipate improvements when increasing the bias; however, we have selected 1500 V as a “relatively safe” bias to study the longevity.

Future experiments will explore the operation of the PS‐CFG detectors at higher bias and higher energy resolution.

Notes:

1. The “Conditioning Period” is a period in the factory wherein the crystals are initially brought up to nominal operating conditions in a controlled fashion in order to achieve long‐term stability

0 1000 2000 keV

C ou nt R at e

(H z)

Pulse Amplitude (MCA chan.)

03Apr

06Apr

09Apr

13Apr

20Apr

28Apr

04May

11May

22May

28May

02Jun

TPG

137-Cs irradiation

Date

0.0k

0.5k

1.0k

1.5k

0 10 20 30 40 50 60

TPG

P o si tio n (M

C A c ha n keV

W id th

F

W H

M

Days