Uncontrolled ion migration has been well-known in perovskite-based semiconductor devices. Here, we show that instead of being detrimental, ion migration can be used to enhance the performance of perovskite CsPbBr3 semiconductor gamma-ray detectors. Through deliberate application of electrical biasing, we actively control ion migration to modify the metal-CsPbBr3 interface barrier height in devices with asymmetric electrodes. Ion migration plays a pivotal role in reducing bulk defects, as evidenced by the contact potential difference measurement, thermally stimulated current spectroscopy, and photoluminescence measurements. The evidence suggests that biasing-induced ion migration in CsPbBr3 results in a reduction in electron traps. As a result, record-breaking performance of 57Co gamma ray spectrum for CsPbBr3 detector was achieved by intentionally biasing the detector. As biasing at elevated temperatures expedites ion migration, preconditioning the CsPbBr3 crystals through reverse biasing is a promising strategy for enhancing their performance.
Article
Ion migration as a tool to enhance the performance of perovskite CsPbBr3 γ-ray detectors
https://doi.org/10.21203/rs.3.rs-3932916/v1
This work is licensed under a CC BY 4.0 License
Version 1
posted
You are reading this latest preprint version
Semiconductor detectors operating at room temperature have broad applications in security inspection, medical imaging, and scientific research. 1–3 Efficient interception of hard X-rays and gamma rays necessitates the use of semiconductors with high atomic number (Z). While CdZnTe has been successful as a high-Z semiconductor detector, 4 its wide-scale adoption is limited due to the high cost associated with crystal growth challenges. 5 TlBr has attracted attention for many years, but the uncontrollable migration of ions has emerged as a significant problem, leading to notable performance degradation in TlBr detectors. 6,7 On the other hand, lead halide perovskites have recently emerged as promising candidates for high-performance semiconductor radiation detectors. These perovskites exhibit high Z, high mobility lifetime products, and the potential for cost-effectiveness. 8,9 CsPbBr3, an all-inorganic perovskite compound, has garnered significant interest among various types of perovskites. 10,11 This attention stems from its exceptional thermal and moisture stability, surpassing hybrid organic-inorganic counterparts. CsPbBr3 single crystal detectors have achieved remarkable milestones in the detection of gamma rays, 12–14 high flux X-ray photon counting, 3 and ultra-high flux synchrotron X-ray detection. 15 The superior properties of CsPbBr3 make it a compelling candidate for wide applications in radiation detection.
Despite the record-breaking achievements of perovskite-based semiconductor devices in various fields, uncontrolled ion migration has been known to alter or degrade device performance, especially in photovoltaics16–18. The phenomenon of device current hysteresis is often attributed to the migration of ions.16–18 This migration process is typically described as the accumulation of migrated ions from the interior of the perovskite to the electrode-perovskite interface region, creating an opposing electric field in the bulk that counteracts the external electric field, as evidenced by report on a MAPbBr3 detector19. Consequently, the charge collection efficiency of perovskite devices would decrease. Moreover, ion migration has been observed to alter interface energy barriers in perovskite solar cells20–22 and photodetectors23, thereby influencing the current-voltage (I-V) behavior of these devices. These intricate ion migration mechanisms are not fully understood and can significantly influence the operational characteristics of perovskite devices. Furthermore, the impact of ion migration extends beyond modifying the electric field within the device. In perovskite materials, the mobile ions behave as crystal defects, which encompass vacancies, interstitials, and antisite defects16,20,24. The correlations between these ionic defects and charge carriers transport (i.e., electrons and holes) have been demonstrated16,25,26. Consequently, ion migration can induce changes in the material's defect characteristics, leading to further alterations in the operational properties of perovskite radiation detectors. A recent report has highlighted defect proliferation in a solution-grown CsPbBr3 detector and attributed it to ion migration27. Intriguing observations have been made regarding the role of ion migration in the reversible disappearance and re-appearance of Schottky diode current-voltage (I-V) behavior in melt-grown CsPbBr3 detectors14. This migration of ionic defects often results in detrimental effects on the device's performance and stability, prompting the exploration of various methods to suppress ion migration20,24,26,28. Although ion migration is extensively researched in various thin film perovskite devices (such as solar cell), there remains a gap in the comprehensive understanding of its impact on the electronic and defect properties in CsPbBr3 single crystals. Specifically, with melt grown CsPbBr3 crystals12,29, we found that some CsPbBr3 detectors delivered optimal performance right after device fabrication, while some required electrical biasing to reduce the initial noise and achieve the expected performance. We hypothesize that the difference between these two types of devices regarding the necessity of electrical biasing arises from the different number of defects associated with the mobile ions within the CsPbBr3 crystal.
Driven by such hypothesis, in this study, we investigated the role of ion migration in melt-grown CsPbBr3 gamma-ray detectors with asymmetric electrodes, i.e., a low work function metal like bismuth on one side and a high work function metal like gold on the other. With experimentally measured electronic energy band diagrams, we show biasing-induced ion migration could modify the metal-CsPbBr3 interface barrier, which explains the seemingly different electronic I-V behaviors of the CsPbBr3 detector, such as I-V hysteresis, current drifting, and reversible Schottky diode I-V behavior. The bulk of CsPbBr3 becomes more intrinsic following electrical biasing, which is further corroborated by Thermally Stimulated Current (TSC) spectroscopy and photoluminescence (PL) measurements. The Bi/CsPbBr3/Au detectors exhibited an enhanced gamma-ray energy spectra because of the decreased bulk defects caused by the biasing-induced ion migration. This spectrum improvement is more pronounced in electron collection mode compared to hole collection mode, suggesting that the ion migration causes a reduction of electron traps in CsPbBr3. Furthermore, the rate of biasing-induced ion migration in CsPbBr3 increases when reverse biasing is applied at slightly elevated temperatures. Consequently, preconditioning CsPbBr3 crystals with a large bias voltage at elevated temperatures can effectively and swiftly reduce crystal defects, thereby enhancing the overall performance of the detectors.
CsPbBr3 detector I-V behavior
The CsPbBr3 single crystals were synthesized using the high-temperature Bridgman method, as detailed in our previous work.12,29. We fabricated CsPbBr3 detectors from cut and polished crystals with thickness in the range of 2.1–3.2 mm with an asymmetric electrode structure, i.e., Bi/CsPbBr3/Au (Fig. 1a). The Bi and Au form a Schottky and Ohmic contact with the CsPbBr3, respectively 30,31. Schottky contact creates an energy barrier at the metal-semiconductor interface, leading to current rectifying behavior with lower current at reverse voltage direction. Hence, reverse bias voltage with Bi as anode and Au as cathode is used to maintain a small leakage current for normal detector operation. To further minimize the surface leakage current, we used the guard ring structure where the center Au electrode (diameter 2 mm) is surrounded by a guard Au electrode, while the opposite side is the Bi planar electrode that covers the entire crystal surface (Fig. 1a). The center and guard electrodes were held at the same bias voltage under normal operation, and signal was read out only from the center electrode. As shown in Fig. 1b, when the guard electrode was enabled at the same bias voltage as the center electrode, both the signal current amplitude and its decaying were much smaller than that when the guard electrode was disabled (i.e., floating). With minimal surface leakage current, we can observe the true CsPbBr3 detector I-V behavior.
We first observed I-V hysteresis (Fig. 1c), the shape of which depends on the voltage sweep starting point, direction, and speed. Here, we started the voltage scan from 0 V to 300 V (referred to as the forward direction) where the current became larger as voltage increased (90 nA at 300 V). Then the voltage was swept from 300 V to -300 V. The current reached a maximum value at some voltage between 300 V and 0 V (490 nA at 170 V) and then during the reverse scan direction it increased slowly from 0 V to -300 V (-17 nA at -300 V). The current under reverse bias, i.e., negative voltage, is much smaller than that under forward bias, i.e., positive voltage, due to the Schottky diode structure. Upon voltage sweeping from − 300 V to 0 V, the current was slightly smaller than that from 0 V to -300 V (see Fig. 1c inset, e.g., current is -9.6 nA at -200 V for sweeping from − 300 V to 0 V and current is -12.4 nA at -200 V for sweeping from 0 V to -300 V). The observed I-V hysteresis can be attributed to the consistent increase and decrease of current under continuous forward and reverse bias, respectively, in the CsPbBr3 detector (Fig. 1d). For example, the maximum in current in Fig. 1c (490 nA at 170 V) when sweeping from 300 V to 0 V is a combined result of the forward biasing that leads to current increase over time and the decrease of the forward bias voltage that leads to smaller forward current. Likewise, the reverse biasing process, which involves sweeping from 0 to -300 V, results in a gradual decrease in reverse current over time, leading to a slightly smaller current when the scan is switched back from − 300 V to 0 V (as seen in Fig. 1c).
Another interesting phenomenon is the observed reversible change of CsPbBr3 detector Schottky diode I-V behavior upon the high voltage biasing. As shown in Fig. 1e where voltage swept from 0 V to 300 V and then 0 V to -300 V separately to avoid current maximum like in Fig. 1c, without intentional biasing ("initial state” in Fig. 1e), the I-V curve showed a rectifying Schottky diode behavior. After reverse biasing at -1000 V for some time (e.g., 10 mins in this case), however, the I-V curve became more linear and nearly symmetric. Then I-V curves measured after forward biasing at 1000 V show a recovery of the asymmetric rectifying behavior and the degree of recovery depends on the forward biasing time. We further observed that the reduced reverse current after reverse biasing could persist for a long time (e.g., more than several days to several weeks). As shown in Fig. 1f, under intermittent reverse biasing where reverse − 1000 V and 0 V were applied alternately for 60 s duration, the current kept decreasing following the previous path and the reduced dark current can remain low for an extended time. After the intermittent reverse biasing, forward biasing of the device resulted in a significant increasing in the forward dark current (see Fig. 1f inset). When the device was subjected to a new cycle of continuous reverse biasing at -1000 V, the current exhibited a similar decreasing pattern as observed under previous intermittent biasing. Additionally, applying a higher reverse bias voltage led to a faster decline in the reverse current (see Supplementary Fig. 1). While the reversible alteration of the diode's I-V behavior and the rate of current change during forward and reverse biasing are likely linked to ion migration, the precise underlying interface mechanisms remain unclear.
Effect of biasing-induced ion migration on electronic and defect behavior in CsPbBr3
The device I-V behavior is a direct manifestation of its underlying electronic energy band configuration. Kelvin Probe Force Microscopy (KPFM) is a useful tool to obtain the energy band diagram of semiconductor devices and has been used to study the perovskite-electrode interface dynamics19. To gain insights into the relationship between ion migration and CsPbBr3 detector I-V behavior, we used a Kelvin Probe setup to scan the Contact Potential Difference (CPD) of the Bi/CsPbBr3/Au structure in the dark after different biasing conditions (Fig. 2a). Typical vertical device structure, as shown in Fig. 1a, is not suitable for CPD measurement as the metal electrodes are too thin (~ 100–300 nm) to produce any meaningful results. Hence, rectangular-shaped Bi and Au electrodes were deposited in parallel on the same side of the CsPbBr3 crystal with a separation of 2.6 mm (see Supplementary Fig. 2 for device picture). Initially, a CPD scan was conducted on the CsPbBr3 detector without any deliberate biasing, referred to as the "initial state" in Fig. 2a. Subsequently, the detector was subjected to reverse biasing at -1000 V for a duration of 10 minutes. Following the reverse biasing, the CPD measurement taken immediately showed a lower value compared to the initial state. The CPD change of the Au electrode (i.e., ΔCPD-Au ~ -400 mV) is larger than that of the Bi electrode (i.e., ΔCPD-Bi ~ -200 mV). The CsPbBr3 detector was then forward biased at 1000 V for 5 mins. We note that we limit the forward biasing time to be shorter than the reverse biasing time, because the heat generated by the large forward current at 1000 V may damage the detector. The CPD measured immediately after forward biasing shows a recovery, which agrees with the previous reversible I-V behavior.
With the known work function of the Kelvin Probe tip (5 eV) through pre-calibration, we constructed the thermal equilibrium energy band diagram of the Bi/CsPbBr3/Au structure at the initial state and after reverse biasing (Fig. 2b). The diagram is drawn to scale proportional to the relevant work function, electron affinity, and bandgap values (see Supplementary Fig. 3 for these values at initial state). There are several notable changes after reverse biasing. First, the work function of Bi and Au electrode both became smaller, while the electron affinity and bandgap of CsPbBr3 should remain constant. However, when we examine the interface energy barrier for hole injection from metal electrode to CsPbBr3 (referred to as ΦB−Bi and ΦB−Au), we observe an increase in the Schottky barrier height for both Bi and Au. The barrier height for Bi increased from 0.8 eV to 1 eV after reverse biasing, while the barrier height for Au increased from 0.4 eV to 0.8 eV after reverse biasing. Additionally, the bulk electric field, which originates from the asymmetric electrode work function, decreased from 0.4 V to 0.2 V. The increased ΦB−Bi and decreased bulk electric field are directly responsible for the observed decrease in current during reverse biasing, (as shown in Fig. 1d). Furthermore, the increased ΦB−Au after reverse biasing leads to a reduction in forward current, which explains the symmetric I-V behavior seen in Fig. 1e. During forward biasing, the energy band diagram can return to its initial state. In this process, the ΦB−Au continues to decrease, resulting in an increase in current under forward biasing, as depicted in Fig. 1d. The metal electrode work function and interface barrier change are attributed to mobile ion accumulation (likely bromide ions) at the metal-CsPbBr3 interface due to electrical biasing. It is worth to note that although the ion accumulation at the metal-CsPbBr3 interface is reversible during the time scale of experiments (e.g., several days), the long-term effect regarding interfacial electrode reaction with CsPbBr3 under continuous reverse bias is possible and remains to be investigated in the future. In addition to affecting device electronic behavior, the ion migration within CsPbBr3 could also modify crystal defect behavior. Indeed, the fermi level of the CsPbBr3 moves ~ 0.4 eV closer to the middle of the bandgap after reverse biasing, which indicates that the CsPbBr3 bulk become more intrinsic with less defects.
To further observe the effect of biasing-induced ion migration on CsPbBr3 defects, we conducted Thermally Stimulated Current (TSC) spectroscopy to directly measure the defect density of a non-biased and a biased CsPbBr3 device. TSC has been successfully used for quantitative defect characterization on various semiconductors, such as GaAs32, CdZnTe33, and CsPbBr334,35. Here, two CsPbBr3 crystals adjacent in the Bridgman ingot with essentially the same crystal quality were fabricated into two devices (Fig. 3a). Planar Au and Ga/In metal alloy were used as opposite electrodes. The control device, named as non-biased device, was not biased before any testing, while the biased device was subject to reverse biasing at 600 V for 36 hours. Gamma ray energy spectra were first measured for these two devices before the TSC measurements. As shown in Fig. 3b, the non-biased device cannot resolve the 59.5 keV and 122 keV photopeak of the 241Am and 57Co gamma ray sources, while the biased device clearly resolves the photopeak with higher channel number. The TSC results showed that the non-biased device produced a larger thermally stimulated current due to a higher defect density than the biased device (Fig. 3c), which is consistent with the improved charge collection of the biased device. Further quantitative estimation shows a total defect density of 5×1012 cm− 3 in the temperature range of 79 K − 300 K for the biased device, which is two orders of magnitude smaller than the total defect density of 2×1014 cm− 3 for the non-biased device (see Supplementary Info for calculation procedure of the defect density).
Previously, we reported the photoluminescence (PL) spectroscopy as an effective method to investigate the defects of CsPbBr3 crystals (see Supplementary Fig. 4) 15. Specifically, a shorter PL decay time of the as-grown CsPbBr3 crystal is associated with excessive defects facilitating carrier trapping and the inability to collect the generated charges. In contrast, a longer as-grown PL decay time suggests fewer defects and longer carrier drifting lengths under electric field. Here, to study the PL change due to electrical biasing, a large CsPbBr3 crystal with average PL decay time of ~ 43 ns was broken into four pieces. One served as the non-biased control sample, while the remaining three were exposed to reverse biasing at 300 V. After different biasing treatments, the PL was measured from freshly cleaved crystal surfaces to observe the true PL behavior of the bulk unaffected by surface deterioration. As shown in Fig. 3d, the PL emission spectra of these crystals with different biasing times exhibited a slightly red-shifted maxima by 1–5 nm (see Table S1 in Supplementary Info). Then, the PL decay times of these four crystals were measured at the same wavelength of 529 nm. Figure 3e shows the fitted PL decay curves. For the three biased crystals, the average decay times were 70 ns, 66 ns, and 62 ns, corresponding to biasing durations of 8, 12, and 15 hours, respectively (see Table S1 in Supplementary Info). These decay times are on average notably longer than the 43 ns decay time of the control non-biased crystal. The same behavior of the elongated PL decay time after reverse biasing was observed for another set of four crystals (see Table S2 in Supplementary Info). Defects in a crystal can act as non-radiative recombination centers, which can shorten the PL decay time because they provide alternative and faster pathways for electrons to lose their energy without emitting light. The observed extension of the PL decay time following biasing suggests a reduction in the number of defects, aligning with the expectation that longer PL decay time is indicative of fewer non-radiative defects and thus, superior crystal quality, as shown in Supplementary Fig. 4. Consequently, the biasing-induced defect reduction, as indicated by the PL measurements, agree with the CPD and TSC results discussed above. This further substantiates the impact of biasing-induced ion migration on the electronic and defect properties of CsPbBr3.
Effect of biasing-induced ion migration on gamma ray spectral performance
The defect reduction effect in CsPbBr3 due to biasing-induced ion migration can have a positive impact on its gamma ray detection performance by reducing charge trapping. The increased interface barrier after reverse biasing leads to a decrease in dark current, which is desirable for gamma spectrum acquisition as it reduces detector noises. Biasing-induced ion migration also leads to a decrease of the built-in bulk electric field for the tested CsPbBr3 detector from 0.4 V to 0.2 V, as shown in Fig. 2b. The 0.2 V reduction of built-in electric field is negligible compared to the externally applied electric field that is usually several hundred volts or higher. Hence, the detector charge collection efficiency and spectral performance is expected to benefit from the bias-induced ion migration.
To evaluate the impact of biasing-induced ion migration, we conducted measurements on the CsPbBr3 detector gamma spectrum during the biasing process. Initially, we measured a gamma ray spectrum after the CsPbBr3 detector had been subjected to reverse biasing for a significant duration (e.g., 30 minutes or longer) until the detector reached a steady state regarding electronic and defect behavior. This steady state was indicated by the asymptotic dark current (Fig. 4a). Deduced from the behavior of reversible I-V and energy band diagram of the CsPbBr3 detector, the detector ion migration state is reversible by applying electric field with opposite direction (Fig. 1f). Hence, we purposely applied forward biasing until the forward current become significantly large (Fig. 4a), under which condition the detector behavior is reversed back to that close to the initial state. Then immediately following the forward biasing, we measured gamma ray spectra consecutively under continuous reverse biasing (Fig. 4a). Due to the repeatability of detector state, the gamma ray spectra obtained after applying forward biasing encompass the period during which ion migration occurs within the bulk. This period corresponds to the duration in which the reverse dark current decreases and approaches a steady value. Moreover, to investigate the influence of ionic defect on electron and hole transport, we conducted gamma spectra acquisition under specific setups where either the hole-induced or electron-induced signal dominates. We used the 59.5 keV gamma-ray emitted by 241Am, which has a limited penetration depth in CsPbBr3. 1 mm of CsPbBr3 can attenuate ~ 91% of the 59.5 keV gamma rays (refer to Supplementary Fig. 5). Based on the Shockley-Ramo theorem,36 It can be approximated that the total induced signal predominantly originates from holes or electrons when using the setup of irradiating the anode or cathode of CsPbBr3 detectors with thickness from 2.1 mm to 3.2 mm in this study. This approximation holds valid for the 59.5 keV gamma rays. However, for the more penetrating 122 keV gamma rays emitted by 57Co, the approximation is less accurate.
When using 241Am to irradiate the anode (Bi electrode) of a CsPbBr3 detector, with an acquisition time of 30 seconds for each spectrum and no interval between consecutive spectra after forward biasing (as shown in Fig. 4b), we observed that the gamma ray spectra obtained before and after forward biasing remained nearly identical when the signal was acquired in hole collection mode. Similarly, when the anode was irradiated by the more penetrating 57Co gamma-rays, with hole collection mode (as shown in Fig. 4c), there was no noticeable change in the spectra before and after forward biasing. These findings suggest that the mobile ions do not have a noticeable impact on hole transport in CsPbBr3. In contrast, when the cathode (Au electrode) was irradiated by 241Am with the same procedure for spectra acquisition (Fig. 4d), there was an abrupt degradation of the first spectrum acquired after forward biasing compared to that before forward biasing. Following the acquisition of the first spectrum after forward biasing, the gamma spectrum gradually improved over time as the CsPbBr3 detector was continuously subjected to reverse biasing. Eventually, the gamma spectrum returned to its initial state prior to forward biasing. This degradation and subsequent recovery of the spectrum, under condition where electron-induced signal dominates (electron collection mode), indicate that the ionic defects in CsPbBr3 act as electron traps. During reverse biasing, the mobile ions in the bulk of CsPbBr3 migrate towards the surface region. This migration process repairs the ionic defects that can trap electrons within the bulk, resulting in better electron collection efficiency. When the cathode was irradiated by 57Co, same spectrum degradation and recovery were observed (Fig. 4e). However, the contrast of spectrum change appears less prominent than for 241Am, because of the higher penetration depth of 122 keV gamma ray than the 59.5 keV gamma ray and the consequent less electron transport contribution to the 57Co gamma spectrum. Additionally, the fact that the gamma ray spectrum does not degrade over time under continuous reverse biasing suggests that the reduction of the built-in electric field due to ion migration has negligible adverse effect on CsPbBr3 detector performance. Overall, these results demonstrate that biasing-induced ion migration can reduce bulk defects of CsPbBr3, which could improve CsPbBr3 detector gamma ray spectral performance.
With such insights gained regarding electron collection and trapping in CsPbBr3, we were able to break the performance record of the 57Co gamma spectrum through deliberately biasing the detector. As shown in Fig. 5, the 57Co spectrum of the non-biased CsPbBr3 detector was the best achieved performance in our previous work12 without deliberate biasing. In comparison, the 57Co spectrum of the biased CsPbBr3 detector is the current best performance with intentional biasing applied to the detector. Notably, the biased detector produced a more symmetric photopeak with less low energy tailing as opposed to that of the non-biased detector. Since the low energy tailing is resulted from poor electron transport in CsPbBr3 13, less low energy tailing indicates improved electron collection of the biased detector, which is consistent with the conclusion that biasing-induced ion migration reduces electron traps in CsPbBr3.
Ion migration acceleration at elevated temperature
Since biasing-induced ion migration can reduce ionic defects in the bulk of the CsPbBr3 crystals, intentional biasing can be used as an effective preconditioning step to improve crystal quality and device performance. To reduce the biasing preconditioning time, a higher ion migration rate is desired. As shown in Supplementary Fig. 1, a large reverse bias voltage leads to faster current decay, which means a stronger applied electric field can accelerate the ion migration process. Since ion migration rate is temperature dependent in perovskites28, increasing the temperature could also help with a faster ion migration of CsPbBr3. We then investigated the effect of temperature on biasing-induced ion migration in CsPbBr3 by examining the current drifting under constant forward or reverse biasing. We note that the migrated ions after biasing cannot be reversed during a time scale of hours or days without applying an opposite electric field (see Fig. 1f). Hence, to observe the forward current increasing at different temperatures for the same device, each time after forward current testing at a given temperature, we need to reverse bias the detector to reset the large forward current to the initial small value before the next forward current testing can begin at another temperature. As shown in Fig. 6a, the forward current increasing under constant forward 1000 V biasing became faster when temperature was increased from 26 °C to 60 °C and became slower when temperature was decreased from 60 °C to 26 °C (Fig. 6b). We note that we restricted the temperature to be lower than the phase transition point of CsPbBr3 (i.e., 88 °C37) to avoid potential crystal damage. Since the forward current increasing is due to ion migration induced Au-CsPbBr3 barrier change (see Fig. 2b), faster current increasing means faster ion migration at higher temperature, which is consistent with thermally activated ion migration. Similarly, to observe the reverse current decreasing at different temperatures, we need to apply forward bias to reset the small reverse current to a large initial value before the next reverse current testing can begin at another temperature. As shown in Fig. 6c and Fig. 6d, the dark current value after decreasing to steady state became larger when temperature was increased from 21 °C to 60 °C, and became lower when temperature was decreased from 60 °C to 21 °C. The dark current steady value is affected by thermionic emission through Schottky barrier, so the stronger thermionic emission at higher temperature leads to larger dark current. To evaluate the rate at which the reverse current decreases at different temperatures, we normalized the dark currents depicted in Fig. 6d to their respective initial values, as presented in Fig. 6e. This analysis reveals that the reverse current diminishes more rapidly at higher temperatures with biasing time, suggesting an increased rate of biasing-induced ion migration at elevated temperatures.
We observed that at the onset of biasing (initial tens of seconds), the reverse currents of the device shown in Fig. 6c exhibited a significant peak before commencing the overall trend of current decreasing. Theoretically, such reverse current peak could be due to ionic current or trapped carrier release (carrier de-trapping) because the device was initially forward biased before commencing reverse biasing and thus significant charge injection and trapping could occur during forward biasing. To further find out the mechanism of the reverse current peak, we did intermittent reverse biasing of the device in Fig. 6c (i.e., reverse − 1000 V and 0 V were applied alternately for 60 s duration), and the reverse current is shown in Fig. 6f. The peak exists for every reverse biasing of -1000 V. Given the nature of ion migration, it is expected that the ionic current will produce a monotonic decreasing curve during intermittent biasing, instead of showing a peak. Hence, the large current peak in the beginning of reverse biasing should be due to electronic carrier de-trapping which is a result of charge injection and trapping during the prior forward biasing. As shown in Supplementary Fig. 6, the longer time the forward biasing lasts, the larger the reverse current carrier de-trapping peak is. The reverse current peak is related to the trap characteristics (e.g., trap density and trapping/de-trapping time constant) of the CsPbBr3, so different devices could show different peak behavior. The reverse currents of device in Fig. 6d also showed a peak feature, but less prominent compared with that of Fig. 6c. In comparison, the reverse currents of the device in Supplementary Fig. 1 did not show observable peak feature. Lastly, carrier de-trapping could also be a contributing factor for the reverse current decreasing in addition to the Schottky barrier height increasing as shown in Fig. 2b. The number of carriers generated from de-trapping decreases as function of time, which contributes to lower reverse current.
Atomic-level mechanism of ion migration in CsPbBr3
One open question so far is the atomic-level mechanism of biasing-induced ion migration, such as the specific types of ionic defects and how the migration of these ionic defects leads to reduction of bulk defects in CsPbBr3. We have seen that forward biasing can cause gamma spectrum degradation (Fig. 4), which indicates that the reduced bulk defects following reverse biasing re-occurred after forward biasing. Since PL measurement can probe the bulk defect change, it would be interesting to see how the PL decay time change after forward biasing. We observed that the forward biased samples showed PL decay times (58.5 ns, 86.7 ns and 78.2 ns) longer than that of the non-biased sample (54 ns) (see Supplementary Table S3 for details). Such PL results indicate that some defects are reduced after forward biasing. The fact that forward biasing reduces some bulk defects but also results in gamma spectrum degradation leads to the speculation that there may be more than one type of defects that are subject to the effect of biasing. Some type of defects can be healed while some other type of defects cannot. As reported, the interstitial Br- vacancies in CsPbBr3 process low formation energy11,38,39 and occur during crystal growth process. Such vacancies generate localized charged regions within the crystal lattice, characterized by the undercoordinated Pb2+ atoms (as illustrated in Supplementary Fig. 7). We propose that as the Br- ions migrate through the lattice, they may encounter and re-bind to the strongly positively charged undercoordinated Pb2+ atoms. Such phenomenon effectively repairs this type of vacancy defects. We note that the discussions here regarding the atomic-level mechanism of ion migration are speculations, which is intended to stimulate further research in this area instead of being considered as conclusive. Such open question highlights the need for better understanding of the complex defect dynamics in perovskite materials.
In this work, we demonstrated that biasing-induced ion migration could induce change of metal electrode work function and interface energy barrier, which is the reason of CsPbBr3 detector current drifting under constant biasing. The fermi level of CsPbBr3 shifts towards the middle of the bandgap following reverse biasing. This shift signifies that the CsPbBr3 bulk becomes more intrinsic, indicating a decrease of defects. The fewer bulk defects are also supported by thermally stimulated current and photoluminescence measurements. We further identified that the ionic defects in CsPbBr3 act as electron traps because they affect electron transport within the detector. With less defects after biasing-induced ion migration, the collection of gamma ray generated free electrons gets improved, which leads to record-breaking performance of the 57Co gamma ray energy spectrum of the CsPbBr3 detector. Beneficial from the effect of biasing-induced ion migration, preconditioning the CsPbBr3 crystal under large electric field and high temperature can be an effective method to improve CsPbBr3 crystal quality and device performance. With the important role of ionic defects revealed, tuning the crystal growth process to reduce ionic defects may be an effective approach to further improve CsPbBr3 crystal quality. Besides, interfacial electrode reaction may be possible, and it could affect the long-term operational stability of the CsPbBr3 detector, which needs to be investigated in future work. The use of different electrodes in the CsPbBr3 detector may also result in different ion migration and biasing characteristics. Our findings contribute to the understanding of the role of mobile ions in CsPbBr3 and their impact on the overall performance of gamma detectors based on this material.
Crystal growth and device fabrication
The CsPbBr3 ingots were grown from high temperature Bridgman method with growth process parameters described in detail in our previous work12,29. CsPbBr3 Ingots were cut into pieces of appropriate dimensions for device fabrication. The CsPbBr3 crystal surfaces were finely polished using sandpaper and then cleaned with toluene. Gold electrode (~ 100 nm thickness) and Bi electrode (~ 300 nm thickness) were thermally evaporated onto the crystal surface. Gallium indium metal alloy electrode was brushed onto crystal surface. The guard ring electrode pattern was formed using a shadow mask and the gap between the two separate guard electrodes were filled in with gold paint. Copper wires were connected to the electrodes using silver paint. The CsPbBr3 devices were placed on a glass holder and encapsulated in paraffin wax.
Current, Kelvin probe, photoluminescence measurement
Gamma energy spectrum acquisition
The gamma spectroscopy system includes a high voltage power supply (ORTEC 556), a preamplifier (eV-550), a shaping amplifier (ORTEC 572A), and a multi-channel analyzer (ORTEC 927). Alternatively, a digital pulse processing system (Amptek PX5) can be used to replace the shaping amplifier and the multi-channel analyzer.
Thermal stimulated current (TSC) spectroscopy
CsPbBr3 device was mounted in a cryostat in the dark and cooled down to 79 K with liquid nitrogen. Subsequently, at 79 K, the device was photoexcited with a Xe arc lamp (150 W broadband wavelengths) for ~ 15 minutes. About 10 minutes after the excitation was turned off, the device was heated up at a constant heating rate from 79 K to 310 K with a 0.5 K step. During heating, the current after photoexcitation was obtained by a Keithley 2636 electrometer under a 20 V applied reverse bias. The above procedure was repeated but without photoexcitation to obtain a dark current. Finally, the thermally stimulated current was obtained by subtracting the dark current from the current after photoexcitation.
Acknowledgments
This research was supported by the Defense Threat Reduction Agency (DTRA) under the Interaction of Ionizing Radiation with the Matter University Research Alliance (IIRM-URA) under contract number HDTRA1-20-2-0002. The work at Argonne National Laboratory was supported by the Department of Energy, National Nuclear Security Administration, Office of Defense Nuclear Nonproliferation Research and Development under contract no. DE-AC02-06CH11357 and the Department of Energy, Basic Energy Sciences (for exploration of Hi-Z detector materials for synchrotron detector applications).
Competing interests
Mercouri G. Kanatzidis and Duck Young Chung are cofounders of Actinia Inc, a company that applies perovskite materials to radiation detection.
Data availability
All data are available in the main text and Supplementary Information.
- Johns, P. M. & Nino, J. C. Room temperature semiconductor detectors for nuclear security. J Appl Phys 126, 040902 (2019).
- Veale, M. C. et al. Cadmium zinc telluride pixel detectors for high-intensity x-ray imaging at free electron lasers. Journal of Physics D: Applied Physics 52, 085106 (2018).
- Pan, L., He, Y., Klepov, V. V., Michael, C. & Kanatzidis, M. G. Perovskite CsPbBr3 single crystal detector for high flux X-ray photon counting. IEEE Transactions on Medical Imaging (2022).
- Kim, Y., Lee, T. & Lee, W. Radiation measurement and imaging using 3D position sensitive pixelated CZT detector. Nuclear Engineering and Technology 51, 1417-1427 (2019).
- Bolotnikov, A. E. et al. Effects of Te Inclusions on the Performance of CdZnTe Radiation Detectors. IEEE transactions on nuclear science 55, 2757-2764 (2008).
- Kozlov, V., Kemell, M., Vehkamäki, M. & Leskelä, M. Degradation effects in TlBr single crystals under prolonged bias voltage. Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment 576, 10-14 (2007).
- Vaitkus, J. et al. Investigation of degradation of electrical and photoelectrical properties in TlBr crystals. Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment 531, 192-196 (2004).
- He, Y., Hadar, I. & Kanatzidis, M. G. Detecting ionizing radiation using halide perovskite semiconductors processed through solution and alternative methods. Nature Photonics 16, 14-26 (2022).
- Wei, H. & Huang, J. Halide lead perovskites for ionizing radiation detection. Nat Commun 10, 1066 (2019).
- Zhang, C. et al. Thermal stability of CsPbBr3 perovskite as revealed by in situ transmission electron microscopy. APL Materials 7, 071110 (2019).
- Kang, J. & Wang, L.-W. High Defect Tolerance in Lead Halide Perovskite CsPbBr3. The journal of physical chemistry letters 8, 489-493 (2017).
- He, Y. et al. CsPbBr3 perovskite detectors with 1.4% energy resolution for high-energy γ-rays. Nature photonics 15 (2020).
- Pan, L., Feng, Y., Kandlakunta, P., Huang, J. & Cao, L. R. Performance of Perovskite CsPbBr3 Single Crystal Detector for Gamma-Ray Detection. IEEE transactions on nuclear science 67, 443-449 (2020).
- Toufanian, R., Swain, S., Becla, P., Motakef, S. & Datta, A. Cesium lead bromide semiconductor radiation detectors: crystal growth, detector performance and polarization. Journal of Materials Chemistry C 10, 12708-12714 (2022).
- Pan, L. et al. Perovskite CsPbBr3 single-crystal detector operating at 1010 photons/s/mm2 for ultra-high flux X-ray detection. (2022).
- Calado, P. et al. Evidence for ion migration in hybrid perovskite solar cells with minimal hysteresis. Nature communications 7, 1-10 (2016).
- Lan, D. The physics of ion migration in perovskite solar cells: Insights into hysteresis, device performance, and characterization. Progress in Photovoltaics: Research and Applications 28, 533-537 (2020).
- Tress, W. Metal Halide Perovskites as Mixed Electronic–Ionic Conductors: Challenges and Opportunities From Hysteresis to Memristivity. The journal of physical chemistry letters 8, 3106-3114 (2017).
- Ahmadi, M. et al. Spatially resolved carrier dynamics at MAPbBr3 single crystal–electrode interface. ACS applied materials & interfaces 11, 41551-41560 (2019).
- Bi, E., Song, Z., Li, C., Wu, Z. & Yan, Y. Mitigating ion migration in perovskite solar cells. Trends in Chemistry 3, 575-588 (2021).
- Chen, B. et al. Impact of capacitive effect and ion migration on the hysteretic behavior of perovskite solar cells. The journal of physical chemistry letters 6, 4693-4700 (2015).
- Chen, Y. et al. In situ management of ions migration to control hysteresis effect for planar heterojunction perovskite solar cells. Advanced Functional Materials 32, 2108417 (2022).
- Wu, T., Ahmadi, M. & Hu, B. Giant current amplification induced by ion migration in perovskite single crystal photodetectors. Journal of Materials Chemistry C 6, 8042-8050 (2018).
- Lee, J.-W., Kim, S.-G., Yang, J.-M., Yang, Y. & Park, N.-G. Verification and mitigation of ion migration in perovskite solar cells. APL materials 7, 041111 (2019).
- Zheng, X. et al. Correlations of Ionic Migration and Deep-Level Traps Leads to Surface Defect Formation in Perovskite Solar Cells. The Journal of Physical Chemistry C 125, 19551-19559 (2021).
- Hou, Y. et al. Suppression of hysteresis effects in organohalide perovskite solar cells. Advanced Materials Interfaces 4, 1700007 (2017).
- Zhang, B.-B. et al. Defect proliferation in CsPbBr3 crystal induced by ion migration. Appl Phys Lett 116, 063505 (2020).
- Lin, Y. et al. Suppressed ion migration in low-dimensional perovskites. ACS Energy Letters 2, 1571-1572 (2017).
- He, Y. et al. High spectral resolution of gamma-rays at room temperature by perovskite CsPbBr3 single crystals. Nat Commun 9, 1609 (2018).
- Pan, L., Feng, Y., Huang, J. & Cao, L. R. Comparison of Zr, Bi, Ti, and Ga as Metal Contacts in Inorganic Perovskite CsPbBr₃ Gamma-Ray Detector. IEEE transactions on nuclear science 67, 2255-2262 (2020).
- He, Y. et al. Sensitivity and Detection Limit of Spectroscopic‐Grade Perovskite CsPbBr3 Crystal for Hard X‐Ray Detection. Advanced Functional Materials, 2112925 (2022).
- Pavlović, M. & Desnica, U. Precise determination of deep trap signatures and their relative and absolute concentrations in semi-insulating GaAs. J Appl Phys 84, 2018-2024 (1998).
- Pavlović, M., Jakšić, M., Zorc, H. & Medunić, Z. Identification of deep trap levels from thermally stimulated current spectra of semi-insulating CdZnTe detector material. J Appl Phys 104, 023525 (2008).
- Zhang, M. et al. Determination of defect levels in melt-grown all-inorganic perovskite CsPbBr3 crystals by thermally stimulated current spectra. The Journal of Physical Chemistry C 122, 10309-10315 (2018).
- Wang, F. et al. Precursor Engineering for Solution Method-Grown Spectroscopy-Grade CsPbBr3 Crystals with High Energy Resolution. Chemistry of Materials 34, 3993-4000 (2022).
- He, Z. Review of the Shockley–Ramo theorem and its application in semiconductor gamma-ray detectors. Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment 463, 250-267 (2001).
- Stoumpos, C. C. et al. Crystal growth of the perovskite semiconductor CsPbBr3: a new material for high-energy radiation detection. Crystal growth & design 13, 2722-2727 (2013).
- Biega, R.-I. & Leppert, L. Halogen vacancy migration at surfaces of CsPbBr3 perovskites: insights from density functional theory. Journal of Physics: Energy 3, 034017 (2021).
- Cohen, A. V., Egger, D. A., Rappe, A. M. & Kronik, L. Breakdown of the static picture of defect energetics in halide perovskites: the case of the Br vacancy in CsPbBr3. The Journal of Physical Chemistry Letters 10, 4490-4498 (2019).
Yes there is potential Competing Interest.
Mercouri G. Kanatzidis and Duck Young Chung are cofounders of Actinia Inc, a company that applies perovskite materials to radiation detection.
- SICLBIonmigration252024.docx
Supplementary Information