Bright and fast scintillation of organolead perovskite MAPbBr 3 at low temperatures

We report the excellent scintillation properties of MAPbBr 3 , an organic– inorganic trihalide perovskite (OTP). The characteristic scintillation time constants were determined using pulsed monochromatic 14 keV X-rays from a synchrotron. We find that between 50 and 130 K the MAPbBr 3 crystal exhibitsavery fastandintensescintillationresponse,withthefast ( s f ) and slow ( s s ) decay components reaching 0.1 and 1 ns, respectively. The light yield of MAPbBr 3 is estimated to be 90000 (cid:2) 18000 ph MeV (cid:3) 1 at 77 K and 116000 (cid:2) 23000 ph MeV (cid:3) 1 at 8 K.

Scintillators detect ionising radiation by converting energy deposited in them to a proportional number of photons.They are omnipresent in large-scale technical and commercial applications around us.For example, they are found in many branches of physics, security scanners, and medical applications such as nuclear imaging for cancer diagnostics. 1,2An ideal scintillator emits a maximum number of scintillation photons per unit energy deposited, has a high absorption coefficient for gamma quanta, and exhibits a narrow timing profile for its scintillation photons.Brighter and faster scintillators facilitate better timing resolution, which is crucial for measuring the time of the initial particle or radiation interaction with high precision.At present, the dominant limitation of modern scintillators is their timing resolution.The state-of-the-art resolution in coincidence timing has just broken the 100 ps barrier, with the lowest value of 73 AE 2 ps reported for LSO-Ce scintillators 3 and 80 AE 4 ps in LGSO-Ce. 4 The main limitation towards the ambitious goal of 10 ps resolution is the low light yield and the long decay time of scintillators. 5,6Currently, the best scintillator for this application is deemed to be LaBr 3 -Ce, exhibiting 16 ns decay time and 70 000 ph MeV À1 . 7However, this is still far from ideal.To achieve a timing resolution below 10 ps, a scintillator with a light yield of at least 140 000 ph MeV À1 and a decay time of r1 ns is required. 5,8espite the decades of intensive effort, it has been found that meeting all requirements in one material is very difficult; hence nowadays the strategy is to develop and optimize scintillators for specific applications. 9Recently, hybrid metal-halide perovskites have attracted considerable attention.These materials have had a long history, with the first studies of photoconductivity in CsPbX 3 (X = Cl, Br or I) carried out in 1958, 10 revealing semiconducting behaviour.A shift towards organic-inorganic materials occurred in 1978 when small molecules were included. 11A number of investigations on structure, optical and luminescence properties were published in the following decades, albeit in studies that were not necessarily aimed at application.A drastic change occurred a few years ago with a significant increase in interest due to the rapid evolution of solid-state photovoltaics based on organic-inorganic trihalide perovskites (OTPs) -materials with the general formula MAPbX 3 , where MA = methylammonium and X = Cl, Br and I. 12 It was quickly realised that the remarkable physical properties of OTPs, in particular the exceptional tolerance of their optical and electronic characteristics to structural defects, made them highly attractive for various optoelectronic applications.In particular, the outstanding high photoluminescence quantum yield of OTPs enabled bright light-emitting devices and lasers, 13,14 whereas their high current conversion efficiency upon light exposure underpins their application as photodetectors. 15Photovoltaic OTPs have also stimulated solar cell research.However, due to the Shockley-Queisser limit, photovoltaic perovskites are typically tuned to a very narrow band gap range, thus excluding the majority of highquality perovskites, especially those with larger band gaps.
Less attention was paid to higher-mass elements with a correspondingly high atomic number (Z) used in OTPs, i.e.Pb, Br and I atoms, which make OTPs inherently suitable for applications in which good X-ray absorption capability is required since the cross section increases with BZ 4 .Furthermore, OTPs exhibit a high mobility of charge carriers, which makes them optimal for radiation detection through direct conversion of X-ray photons into current.Detection of soft X-rays (o10 keV), using the photoelectric effect in polycrystalline MAPbI 3 films, was first demonstrated by Yakunin et al. 16 Improving the detection probability for hard X-rays (4100 keV) called for a decrease in absorption length.This prompted the development of X-ray detectors based on OTP single crystals [17][18][19] or thick films. 20The energy spectra measured with OTP detectors demonstrated energy resolutions of 35% for 59.6 keV of 241 Am and 6.5% for 662 keV of 137 Cs. 17,19 This shows that OTPs are promising materials for the detection of ionising radiation, motivating a systematic exploration for their application in scintillation detectors.An important aim is avoiding limitations that arise from extracting charged particles.This inherent feature of photodetectors with direct photon-to-current conversion imposes two basic constraints.First, it eventually limits the thickness of the absorber and hence the conversion efficiency for high-energy photons.Second, the transit time of charge carriers in the material dictates the relatively slow (B100 ms) response time of OTP photodetectors. 17The advantage of a scintillation detector is its non-reliance on extracting charged particles from the material.Light can be detected from the bulk of the crystal absorber with a response time governed by the probability of radiative decay of excited particles, and this can be very fast as is the case for exciton emission.
Fully inorganic perovskites have the right properties to be fast scintillators.An intense nanosecond X-ray luminescence of free excitons in CsPbX 3 (X = Cl, Br, I) at 77 K was reported by Voloshinovskii et al. already in 1993. 21The light yield at room temperature, however, was too low (o500 ph MeV À1 ) 22 for their application in conventional scintillation detectors.In 2004, interest increased again when sub-nanosecond scintillation decay at room temperature was found in the layered hybrid metal-halide compound (C 6 H 13 NH 3 ) 2 PbI 4 , 23 however, with a light yield of only B6000 ph MeV À1 .Recently, light yields of 9000 and 14 000 ph MeV À1 were reported for the layered perovskites (EDBE)PbCl 4 (EDBE = 2,2 0 -ethylenedioxy)-bis(ethylamine) 24 and (C 6 H 5 C 2 H 4 NH 3 ) 2 PbBr 3 , 25 highlighting the potential of these materials for scintillation detectors.Other researchers compared theoretically the scintillation properties of (EDBE)PbCl 4 and MAPbX 3 (X = Br and I) at low temperatures and posited, without measurement and assuming no thermal quenching, that the X-ray luminescence yield of MAPbX 3 could possibly exceed 150 000 ph MeV À1 at T = 10 K. 24 Furthermore, the characteristic decay time of luminescence observed at room temperature in OTPs is below 100 ns. 24,26Hence, there is the potential for a representative of the OTP family to outperform modern scintillators due to its higher light yield and faster decay time.
Therefore, in this study we investigated the scintillation light yield and decay time of relatively easy-to-synthesize MAPbBr 3 crystals over the temperature range 8-295 K.We found that at cryogenic temperatures the perovskite crystals exhibit exceptionally high light yields (4100 000 ph MeV À1 ) and sub-nanosecond decay times.This finding underpins the potential of OTPs for novel detector applications that rely on the fast timing of scintillation detectors at cryogenic temperatures.In this study, we used synchrotron radiation for measurements of timing characteristics and a multi-photon counting technique for measuring the scintillation light yield at cryogenic temperatures. 27The Materials and methods section contains detailed information on techniques for precisely recording temperature-dependent measurement data and the experimental setup.
When excited with X-rays, MAPbBr 3 exhibits narrow, nearedge emission bands peaking at 560 nm with a very pronounced temperature dependence. 24We studied the scintillation kinetics of MAPbBr 3 crystals over a wide temperature range, from 8 to 295 K, using pulsed X-rays.Fig. 1 displays the corresponding scintillation decay curves.The decay curves exhibit very fast, non-exponential kinetics, which is indicative of bimolecular recombination of the charge carriers as reported recently. 28he main feature of the measured scintillation decay curves of the MAPbBr 3 crystal, and common for the majority of scintillation materials, is an increase in the decay time constant with a decrease of temperature.We also clearly observe that cooling the crystal results in an increase of the background.Furthermore, the amplitude of the scintillation pulse initially increases with cooling, but when the temperature decreases below 50 K, it starts to decrease.Careful inspection of the plots reveals that the scintillation pulse in MAPbBr 3 also undergoes significant changes in shape at low temperatures: the fractional contribution of the background rapidly increases and the long component of the decay curves becomes more pronounced.These features are indicative of a slowing down in the recombination dynamics.
For a more quantitative comparison of these properties and trends, we fitted the measured decay curves with a sum of exponential functions: , where A i is the amplitude, t i the decay time constant and y 0 the background.][31] The quality of the fit was only marginally different between twoand three exponential fits.Two exponentials and a constant background are sufficient for an adequate representation of the measured decay curves.The fitting parameters for MAPbBr 3 as functions of temperature are displayed in Fig. 2.
An analysis of the plots reveals further details in the temperature evolution of the luminescence kinetics of the crystal.As can be derived from the decay time versus temperature dependence, the fast and slow decay time constants in the crystal are about 0.1 and 1 ns at T 4 50 K.This correlates well with the results from photoluminescence decay studies of MAPbBr 3 down to 77 K. 32 With cooling to lower temperatures, the decay rate of the luminescence kinetics in MAPbBr 3 exhibits steep changes, resulting in a significant increase of the decay time constants, so that at T = 8 K, t f = 2 ns and t s = 50 ns.The amplitudes of the fast and slow components initially increase with cooling, while below 40 K they start to decrease; in particular, the amplitude of the fast component decreases by about a factor of five.At the same time, the amplitude of the background y 0 exhibits a steady increase with cooling, becoming comparable to the amplitude of the fast component at T = 8 K.This shows that at this temperature the radiative dynamics is dominated by the slow recombination processes due to trapping and release of charge carriers.This effect causes an afterglow, which has a detrimental impact on the temporal response of the scintillator.Importantly at T 4 60 K the fast and slow scintillation components dominate in the radiative decay, while the fractional contribution of the background does not exceed 1%.This implies that at the higher temperature the major fraction of the scintillation response from the crystal is released over a nanosecond time interval, following an excitation pulse.This is further supported through the measurements of the scintillation light yield discussed later.The evidence of the very fast scintillation response of MAPbBr 3 is shown in Fig. 3a by comparing it with an LYSO-Ce scintillator.Another example is given in Fig. 3b, which displays the sequence of X-ray pulses from the synchrotron (interval Dt = 2 ns) as detected by MAPbBr 3 and LYSO-Ce.It is clear from these figures that the timing performance of the MAPbBr 3 crystal, exhibiting a sub-nanosecond decay time, is superior to that of LYSO-Ce.The latter exhibits a decay time constant of 33 ns and offers an example of one of the best results in coincidence timing resolution, which relies on fast timing. 53][34][35] Upon high-energy excitation, the thermalized electrons and holes form free excitons, which in turn can interact with defects or impurities.The narrow  luminescence bands with a small Stokes shift observed in OTPs at low temperature are attributed to free and bound excitons.The luminescence is very bright at low temperature but exhibits significant thermal quenching.This is due to the exciton binding energies being tens of meV 36 and the increase in temperature causing dissociation.Yet the other characteristic features of exciton emission are the fast decay kinetics.The free excitons emit promptly, while the excitons captured at defects or impurity sites recombine more slowly through de-trapping.Consequently, the scintillation mechanism in the crystals at low temperatures is controlled by two main processes that give rise to the fast and slow emission components.The fast decay component corresponds to the radiative decay of free excitons, while the slow component of the emission is attributed to the radiative decay of electrons and holes released from the traps.
The observed temperature dependence of the luminescence decay in both channels can be explained in the framework of a simple quantitative model by considering the dynamics of radiative and non-radiative transitions between the excited and ground states of the emission center.In terms of this model the measured transition rate (the inverse of the luminescence decay constant t) can be determined as the sum of the radiative (k r ) and non-radiative (k nr ) rates: The changes in decay time with temperature are attributed to the process of depopulation of the excited state through the thermally promoted transfer of excited particles over the energy barrier that leads to non-radiative decay.The rate associated with the non-radiative process exhibits a strong temperature dependence, thus controlling the variation of the non-radiative decay with temperature: where K is the probability of non-radiative decay, DE is the activation energy for the non-radiative transitions, and k is the Boltzmann constant.Substituting (2) into (1) brings about the classical formula: 37 Using this formula, we fitted the experimental results (see Fig. 2) and found that this theoretical model successfully describes the t = f (T) dependence over a range of low temperatures from 8 to 80 K.This indicates that at these temperatures the radiative decay of free and bound excitons is controlled by the thermal activation processes.However, at higher temperature there is a disparity between the model and experimental results, evidencing that the model based on the assumption of isolated emission channels is not valid anymore.We posit that this is rather an expected observation that can be explained as follows: at higher temperature, when the excitons start to dissociate and electronhole pairs can escape trapping sites without recombination, there is a probability for particles exchanging between different radiative decay channels.In other words, the recombination of free and bound excitons may contribute to both emission components of the luminescence decay. 38,39This effect manifests itself by an increase of the decay time constant with heating, which is also observed upon photoexcitation. 32It is worth pointing out that OTP crystals possess exceptionally low trap densities. 26Consequently, radiative decay is a dominant channel for the relaxation of excited states.This is one main cause for the very high luminosity at low temperatures when thermal quenching is suppressed.Further to assess the performance of MAPbBr 3 as a scintillator, we studied a series of energy spectra induced by a-particles in MAPbBr 3 as a function of temperature.Fig. 4 shows the pulse height spectrum measured at 50 K that features a peak with a Gaussian shape attributed to the detection of 5.5 MeV a-particles emitted using an 241 Am source by the MAPbBr 3 crystal.
The position of the peak centre is proportional to the amplitude of the scintillation response of the crystal so that it can be used as a measure of scintillation light output at different temperatures.In Fig. 5, we show the variation of the scintillation light output of the MAPbBr 3 crystal with temperature (together with CsI and LYSO-Ce).A clearly measurable scintillation response can be detected when the crystal is cooled below 180 K.The scintillation efficiency of MAPbBr 3 increases gradually as the temperature is decreased until a plateau is reached at around 70 K.At T 4 60 K the individual scintillation event recorded using a data acquisition system exhibits a very short, intense peak that decays within hundred nanoseconds.This peak is caused by the overlap of many scintillation photons arriving over a short initial time interval after excitation by alpha particles.Hence, at these temperatures only the fast emission contributes to the scintillation signal.An increase of the light output by about 20% is observed as the temperature decreases below 30 K.This increase correlates with the rapid increase of the fractional contribution of the afterglow observed at very low temperatures.We also noted that as the temperature decreases below 50 K, a delayed signal appears, distributed over the entire time window of the 1.6 ms used to record the individual scintillation events.This signal is responsible for the additional emission enhancement, observed over this temperature range; it is due to the process of radiative recombination of charge carriers released by shallow traps with activation energies between 10 and 90 meV, as established using thermoluminescence data. 24aving demonstrated that over the 50-150 K temperature range MAPbBr 3 crystals exhibit fast scintillations, and taking into consideration the theoretical estimates, we then evaluated the scintillation light yield using as references the commercial scintillators CsI and LYSO-Ce.In such evaluations, it is preferable to use a reference scintillator with characteristics not too dissimilar to the crystals under study.Undoped CsI has a very high light output of B100 000 ph MeV À1 at 77 K 40 and exhibits a strong temperature dependence, 41 although the decay time is relatively long (B1 ms at 77 K).LYSO-Ce is renowned for its high light yield (34 000 ph MeV À1 ) and fast decay time, both changing only insignificantly with cooling. 42,43The light collection efficiency of the experimental setup used in this study is determined predominantly by the geometrical factors that are constant parameters.Because of their low penetration depth, the energy of the alpha particles is fully absorbed by the thin samples; hence the scintillation light yield can be evaluated by comparing the measured light outputs of the reference scintillator and the perovskite crystals corrected for the difference in the emissionweighted spectral sensitivity, e l . 44The emission spectra of MAPbBr 3 , CsI and LYSO-Ce crystals as well as the quantum sensitivity of the photomultiplier used for the calculation of the emission-weighted sensitivity, e l , are shown in the inset of Fig. 5.
Taking the light yield of CsI as 100 000 ph MeV À1 at 77 K, we determined that the light yield of MAPbBr 3 is equal to Fig. 4 Pulse height spectra of scintillations excited through a-particle interaction from 241 Am in MAPbBr 3 at 50 K and CsI at 50 K before correction for the spectral response of the photomultiplier.The pulse height distributions that signify scintillation response due to a-particles are fitted by Gaussians (red lines).90 000 ph MeV À1 at 77 K and 116 000 ph MeV À1 at T = 8 K. On the other hand, measuring LYSO-Ce, we found that the scintillation light yield increases to 40 500 ph MeV À1 upon cooling to T = 8 K, giving a light yield of MAPbBr 3 equal to 110 000 ph MeV À1 at this temperature.The estimated values correlate very well despite the relatively large error AE20%, which stems from the uncertainty of e l and the determination of the centroid in the pulse height spectra.The significance of these values can be appreciated in full when compared with the characteristics of the best modern scintillators (see Table 1).
A comparison of the MAPbBr 3 parameters with commercial scintillators shows that OTPs are very promising scintillation materials.Of particular interest is the excellent initial photon density calculated as the ratio of light yield to decay time -the most important parameter that determines the timing precision of the scintillator detector.The higher density of photons near the peak enables a higher precision in determining the time of interaction.A conservative evaluation shows that this parameter is higher by a factor 20 in MAPbBr 3 compared to the best modern scintillator LaBr 3 -Ce.It should be noted that there are a few other materials with fast scintillations at cryogenic temperatures discussed in the literature (ZnO, PbI 2 , HgI 2 51) but a low value of the light yield is a major limitation.The stopping power of MAPbBr 3 , which is defined by the photoelectric fraction of the absorption coefficient, is also very competitive in comparison with other scintillators; only two materials exhibit a higher value.For illustration, the energy dependence of the photoelectric absorption of the three crystals discussed in this study is displayed in Fig. 6.
In summary, in this study we measured the decay time and light output of MAPbBr 3 crystals down to a temperature of 8 K, using X-ray and particle excitation.We found an exceptionally fast and intense scintillation response -the key characteristics for a scintillation detector.At 77 K the fast and slow components of the decay are found to be B0.1 ns and 1 ns, respectively.The light yield of MAPbBr 3 is estimated as 90 000 AE 18 000 ph MeV À1 at 77 K and 116 000 AE 23 000 ph MeV À1 at 8 K.
The observation of intense sub-nanosecond scintillations in OTPs at low temperatures indicates that they are materials with extremely promising scintillation properties that may offer major advantages over conventional scintillation materials.The potential benefits might outweigh additional effort needed for the development of equipment that provide a low-temperature environment for operation of such scintillation detectors.It should be highlighted that the advanced scintillation characteristics of OTP crystals are attained even upon moderate cooling to a temperature just below 100 K, which can be achieved at relative ease through liquid nitrogen-based refrigeration systems.Modern developments in cryogenics made these temperatures also accessible through using dry cryogenic systems, 52 while advances in CMOS silicon photodetectors allow reliable detection of single photons at these temperatures. 53The moderate cooling requirement and the flexibility of the production technology make this approach worth pursuing.There are still challenges in tuning the optical and scintillation parameters, developing the technology to produce crystals with desirable quality and size as well as integration with photodetectors in scintillation arrays.Nonetheless, the compositional flexibility of OTP crystals offers freedom in engineering their optoelectronic properties, whereas solution processing facilitates easy integration with industrial processes.

Methods
The crystals of MAPbBr 3 studied in this work were produced by a solution method as described in ref. 54.In this study, commercially available, undoped CsI (Hilger Crystals (UK)) and Ce-doped LYSO (Saint-Gobain (France)) were used as stateof-the-art reference scintillators.
The scintillation decay curves of the crystals were measured at beamline B16 of the Diamond Light Source using a 14 keV monochromatic X-ray beam from the synchrotron.The measurements  were carried out under special beam conditions by triggering on a single X-ray pulse with an FWHM of Dt = 60 ps, separated from the following pulse by a 300 ns gap.The single crystalline sample with dimensions 4 Â 3 Â 1 mm 3 was glued with silver adhesive to the holder of a continuous-flow, He-cryostat (Oxford Instruments).The temperature in the cryostat was stabilized using a controller using a PID loop with a Si-diode as a temperature sensor and a resistive heater, allowing controlled cooling.The cryostat was attached to an XYZ-translation stage to facilitate swift alignment.One port of the cryostat is equipped with a 0.2 mm thick aluminised Mylar window allowing X-rays to penetrate.Through this window the X-ray beam with a flux of approx.10 9 ph (s mm 2 ) À1 irradiated the sample holder placed at 451 to the incoming radiation, while the luminescence was collected in reflection mode at 451 through a quartz window.The emission was detected using an ID100 single photon counting detector sensitive over a 400-900 nm spectral range and a PicoHarp 300 time-correlated single photon counting module.
To measure scintillation light output as a function of temperature, we used the multi-photon counting technique, developed specifically for measurements of the scintillation characteristics of materials at very low temperatures. 29In this method, the sequence of scintillation events is recorded and analyzed.As the measured charge per scintillation event is proportional to the energy released in the crystal, it is used to derive the value of the scintillation light yield.The sample was mounted on the copper sample holder with a 0.9 mCi 241 Am source placed just behind the sample.The assembly was then placed inside a He-flow cryostat.The scintillation light emitted by the crystal was detected using a multi-alkali photomultiplier 9124A (Electron Tube Enterprises, Ruislip, UK) connected to the optical window of the cryostat.The photomultiplier detects photons in the 280-650 nm wavelength range, thus allowing measurements of the light yield of the MAPbBr 3 crystal.To eliminate the contribution from thermoluminescence due to charge release during heating, the measurements were performed while cooling the crystals.

Fig. 1 X
Fig. 1 X-ray data.Decay curves of X-ray luminescence measured in the MAPbBr 3 crystal at different temperatures in the range 8-132 K (where the emission is most pronounced).The luminescence is excited with 60 ps pulses of synchrotron radiation (E = 14 keV).

Fig. 3
Fig. 3 Scintillation decay.(a) Normalised scintillation decay curves observed upon excitation with 14 keV X-ray pulses in MAPbBr 3 (T = 77 K, red curve) in comparison with LYSO-Ce (T = 292 K, black curve).(b) The sequence of X-ray pulses from electron bunches in the synchrotron ring (time interval 2 ns, FWHM = 60 ps) as recorded with an ID100 photon counter using MAPbBr 3 (red) and LYSO-Ce (black) at T = 77 K.The picture demonstrates the exceptionally good timing resolution of the scintillation response for OTP crystals as opposed to the featureless signal from LYSO-Ce, which is due to the presence of a strong slow component.

Fig. 5
Fig. 5 Light output as a function of temperature.The scintillation light yield as a function of temperature for the MAPbBr 3 crystal (black squares) measured for excitation with 5.5 MeV alpha particles from 241 Am.The plot also displays the comparison with measurements of the commercial scintillators CsI (green triangles) and LYSO-Ce (red circles) with known light yields.The inset shows the normalised emission spectra of CsI (T = 77 K) in green, LYSO-Ce (T = 295 K) in red, MAPbBr 3 (T = 10 K) in black.The dotted line indicates the normalised quantum sensitivity of the photomultiplier 9124A used in the measurements of the scintillation light yield.

Fig. 6
Fig.6Photoelectric fractions of absorption of gamma-rays in CsI, LYSO-Ce and MAPbBr 3 .The data are calculated using an XCOM web-tool.50

Table 1
Properties of modern scintillation materials at room temperature.Data for CsI and MAPbBr 3 are shown at T = 77 K Calculated using an XCOM web-tool.50 a