Nanocrystalline Lead Halide Perovskites to Boost Time‐of‐Flight Performance of Medical Imaging Detectors

Time‐of‐flight (TOF) technique, traditionally used in high energy physics (HEP) and positron emission tomography (PET), is now being explored for lower energy applications like computed tomography (CT). Regardless of the application, pushing the current boundaries in time resolution calls for novel technologies and materials exhibiting ultra‐fast time response. Semiconductor nanocrystals like cesium lead halide perovskites (CsPbBr3), benefiting from quantum confinement effects, feature ultra‐fast decay and, when combined with a suitable bulk scintillator following a heterostructure concept, can also provide the necessary stopping power. In this work, thin films of CsPbBr3 on top of BGO, LYSO:Ce, and GAGG:Ce,Mg wafers are fabricated to test their impact on the single crystal scintillator time resolution under soft X‐rays excitation (about 10 keV). It is demonstrated that the CsPbBr3 layer significantly improves the overall time resolution in all cases, achieving up to a tenfold improvement with BGO and GAGG:Ce,Mg. Under 511 keV γ‐rays, a proof‐of‐concept of the heterostructure design for TOF‐PET using CsPbBr3 thin film deposited on GAGG:Ce,Mg bulk crystal is successfully tested. Shared events depositing energy in both materials are identified, resulting in more than twofold improved coincidence time resolution: 118 ± 4 ps full‐width‐at‐half‐maximum (FWHM) compared to the 272 ± 8 ps of solely GAGG:Ce,Mg.


Introduction
Time of flight (TOF) technique is well established in high energy physics (HEP) for better pile-up rejection and particle identification [1] and in positron emission tomography (PET) to boost the detector performance, and thus leading to a significantly DOI: 10.1002/admi.202300659better quality of the reconstructed image. [2]In HEP, time resolution values below 30 ps (sigma) have already been reached, [1] while the state-of-the-art TOF-PET scanner exhibits a coincidence time resolution (CTR) of about 200 ps full-width-at-half-maximum (FWHM). [3] challenge has been set to achieve 10 ps FWHM CTR in TOF-PET, as such time resolution corresponds to the best desirable spatial resolution along the line of response of 1.5 mm, then is intrinsically limited by the physics of  + decay. [4]he reason for the different time resolution range for these applications is explained by the timing capability scaling with the inverse square root of the deposited energy, limiting the applications that can benefit from the TOF technique.However, improvements in the whole detector chain over the last decade make it feasible to tackle also low energy (10-120 keV) medical imaging applications instead of only TOF-PET (511 keV) or prompt gamma imaging (a few MeV) for particle therapy. [5,6]Computed tomography (CT), and more generally, X-ray radiography, are morphological medical imaging techniques that, by measuring the attenuation of X-ray photons through a body, provide an image of its internal structure.One of the main sources of noise in the reconstructed X-ray images is due to scattered photons.As the unscattered X-rays follow a known path and consequently have a known TOF (they travel at the speed of light in a vacuum), by applying the TOF technique to X-ray imaging, it is possible to discard scattered photons. [7,8]Because of the lower energy range of X-ray imaging, the aimed time resolution cannot be the same as that of PET.However, simulation work [9] has shown that a time resolution of 100 ps can potentially remove 75% of the scattered X-rays, with a clear improvement in the reconstructed image.Reaching 100 ps time resolution at 20-120 keV is not straightforward: the state-of-the-art inorganic scintillator LYSO:Ce coupled to optimized silicon photomultipliers (SiPMs) and readout electronics [10] shows a time resolution of about 700 ps at 15 keV (corresponding to 250 ps at 120 keV). [11]ead halide perovskite nanocrystals (CsPbX 3 ; X = Cl, Br, and I) have captured significant attention of the scientific community in the last decade, especially in the field of optoelectronics, due to their very fast and bright luminescence with narrow and easily tunable emission in the visible region. [12,13]Especially bromides have also been extensively studied as scintillators, with a focus on their application as scintillating screens for low-energy imaging techniques. [14]ecently, given the very fast emission with sub-nanosecond decay time components, their use to improve the timing performance of radiation detectors is been investigated for various applications, including both medical imaging and high energy physics experiments. [15,16]In this case, the main drawback is the need for macroscopic (cm 3 -scale) large detectors in order to achieve a reasonable stopping power.For convenience and protection from detrimental environmental influence, perovskite nanocrystals are usually embedded in a solid matrix (polymer or glass). [17,18]Another requirement for some applications such as high energy physics is radiation hardness.First investigations on cesium lead bromides showed significant radiation stability of the material and potential for high energy applications, [19,20] and a suitable material for embedding could help in this regard.Moreover, semiconductor nanocrystals such as lead halide perovskites and also metal chalcogenide quantum dots (e.g., CdSe) suffer from the small Stokes shift. [12,21]Therefore, to guarantee good transparency and efficient light transport, such nanocomposites can neither be thick enough to provide sufficient stopping power nor dense enough in terms of nanocrystal loading, resulting in a detector with poor statistical collection or poor scintillation efficiency.To give an idea, in our previous work, [16] 10% weight filling factor (which is, to the best of our knowledge, among the highest studied in the literature for lead halide perovskite nanocrystals) in the 100 μm thick sample already resulted in deteriorated transmittance.Because not being capable of stopping a detectable amount of 511 keV gamma photons, only soft X-rays (10 keV) could be used to characterize its scintillation performance.
In this work, we show the potential of CsPbBr 3 (CPB) nanocrystals for the TOF technique in medical imaging, focusing on X-ray but also providing some perspective for TOF-PET.To overcome the problem of low stopping power, we follow a heterostructure concept, [22][23][24][25] where thin films of free nanocrystals in solution are deposited on a dense single crystal scintillator with superior stopping power and energy resolution.Based on the target application, it may be necessary to utilize either a single heterostructure unit, comprising only one layer of nanocrystalline thin film on one dense single crystal wafer (e.g., in TOF X-ray), or stack multiple units together in a pixel (e.g., in TOF-PET), in order to achieve the stopping power required.
We tested different single-crystal scintillators: BGO, LYSO, and GAGG.The former two were chosen as well-established scintillators currently used in commercial detectors in medicine (e.g., BGO for traditional PET systems and LYSO for the firstgeneration of TOF-PET scanners), [2] while GAGG has recently been extensively studied as a fast bulk single crystal scintillator attractive for high-energy physics experiments. [26,27]We demonstrate that the presence of CsPbBr 3 significantly improves the overall detector time resolution under soft X-ray excitation compared to the solely single crystals.Similar results were also obtained under 511 keV for the heterostructure unit made of CPB on GAGG.

Scintillation Decay of Solely CPB
The first step to assess the timing properties of the synthesized CPB nanocrystals consisted of measuring the scintillation decay kinetics of the sample, specifically the decay time.The scintillation time profile obtained following X-ray excitation and acquired in time correlated single photon counting mode is shown in Figure 1.The intrinsic scintillation rate was modeled as the sum of three exponential decay components ( d, 1/2/3 ) and the Dirac-delta function () to describe the ultra-fast emission of the sample. [16,28]For more details about the experimental and modeling methods, please refer to the Experimental Section.One can observe that, considering the prompt component and the fastest exponential component, about 45% of the photons are emitted within the first nanosecond.
Overall, the results demonstrate high initial photon time density (i.e., the number of photons emitted in the first few nanoseconds), a crucial parameter giving the material the best potential to achieve the ultrafast time resolution we seek.To assess this potential, samples for time resolution measurements under both the X-ray and gamma-ray excitations were prepared following the heterostructure concept outlined in Section 1 (and already assessed in previous works [22,24,25] ), as described in Experimental Section 4.

Time Resolution upon X-Ray Excitation
To evaluate the time resolution performance of CPB for TOF Xray, we coupled the bulk, dense scintillator with the SiPM and deposited a layer of CPB nanocrystals (about 50 μm thick) on top of it, directly facing the incoming X-rays (schematically shown in the inset of Figure 2c).
The measured time delay distributions are shown in Figure 2a, together with the FWHM values of all samples summarized in the table below (Figure 2b).The effect of the CPB layer on timing was significant in all cases, achieving a detector time resolution (DTR) of about 240 ps regardless of the type of bulk scintillator.That is more than a tenfold improvement in timing capabilities compared to solely GAGG and BGO crystals (2.4 and 2.9 ns, respectively) and more than a twofold improvement compared to the state-of-the-art TOF-PET scintillator LYSO (590 ps).We also measured the time resolution of the stand-alone perovskite solution by drop-casting it directly on the SiPM, obtaining 196 ps FWHM.
To investigate the influence of the CPB layer in detail, we dropcasted different volumes of the nanocrystal solution (namely 0.5, 2, and 5 μL corresponding to approximately 1, 8, and 20 μm thick layers) on a GAGG plate.We chose to focus on GAGG for this experiment because, together with BGO, it showed the greatest improvement in timing while having comparable light output and energy resolution to those of LYSO.These features make GAGG the best candidate to observe any variation due to the different CPB layer thicknesses.The results are shown in Figure 2c and summarized in the table below (Figure 2d).One can observe that already with only 1 μm thick CPB layer the time resolution (FWHM) decreases from 2.4 ns to 414 ps.Then it improves quickly to 273 ps with 8 μm, 258 ps with 20 μm, and considering the aforementioned results, 229 ps with 50 μm.
These results suggest that the overall timing of the heterostructure is dominated by the CPB layer and that a very thin film (about 1 μm thick) is sufficient to lead to a sizeable improvement in time resolution.This is explained by the time resolution being inversely proportional to the initial photon time density: [29] DTR ∝ The improvement of the time resolution with increasing thickness is due to increasing stopping power of the CPB layer, thereby due to more fast photons contributing to the time response.A rough estimate allows us to determine that layer thicknesses of 1, 8, 20, and 50 μm of CPB can stop about 5%, 33%, 64%, and 92% of X-rays, respectively.These values are overestimated as we took into account the density of CPB bulk crystal while that of nanocrystal is generally lower (for more details about the assumptions and calculations made, please refer to the Experimental Section).However, this estimation suggests that a 50 μm thick layer of CPB solution stops most of the X-rays.The reason why the DTR of the heterostructure units do not go down to the value of the only CPB is due to the contribution from the bulk crystal.The influence of the bulk crystal can be observed, especially in the tail of the time delay distribution.With only a 1 μm thick CPB layer, less than 5% of the incoming X-rays are stopped, and the time delay distribution shows a long tail due to GAGG contribution.To quantify this effect, we evaluated the full-width-at-tenthmaximum (FWTM) which decreased by a factor of two (from 9.3 to 4.5 ns) with a 1 μm thick CPB layer.With CPB layer thicknesses greater than about 10 μm the FWTM stabilized around 1-1.5 ns.We want to emphasize that increasing the CPB thickness further is not the best approach, as the transparency of the solution is not optimal, and this could cause the loss of part of the ultra-fast photons produced in the outer layer.
It is interesting to note that in heterostructures, where a fraction of X-rays interacts with the CPB layer, we do not observe a loss in the collected light.Figure 3 shows the energy spectrum of GAGG, LYSO, and BGO without and with the CPB layer and that of stand-alone CPB nanocrystals: the latter has higher light output than solely BGO but lower than solely GAGG and LYSO, and the heterostructure units showed comparable (LYSO) or higher (BGO and GAGG) light output compared to the corresponding solely single crystal plates.This effect was already observed by Děcká et al. [30] While it cannot be explained by the simple sum of the two contributing materials, it can be due to surface effects.The scintillation light produced by the bulk crystal and emitted in the opposite direction from the photodetector is probably scattered back from the nanocrystals.The effect could be less visible for LYSO as its whole emission spectrum is positioned below the absorption edge of CPB nanocrystals (for the absorption spectra of CPB nanocrystals, please refer to our previous work [31] ).
These results lead us to conclude that a thin film of CPB nanocrystals deposited on a given scintillator can significantly improve its time resolution under X-ray irradiation.With GAGG, an improvement of factor 6 was obtained with only a 1 μm thick film and up to a factor 10 when increasing the thickness to about 10 μm.At the same time, we do not observe a loss in the collected light compared to the sole crystals.Overall, this study constitutes a successful proof of concept for a TOF X-ray detector: the CPB layer improves the time response while keeping both the high stopping power and energy resolution of the "host" crystal, with the ultimate result of enabling the scattering rejection.

Coincidence Time Resolution under 511 keV 𝜸-Ray Excitation
The coincidence time resolution (CTR) of CPB deposited on a GAGG plate was also measured under 511 keV -ray excitation.
We chose to focus on GAGG as CPB is mostly transparent to its emission wavelengths, while this is not the case for BGO and LYSO.In X-ray measurements, this was not a problem, as the CPB was not in between the crystal and the photodetector, unlike in the case of 511 keV -ray excitation (compare the inset in Figure 2c with Figure 4a).The reason is to exploit the energy sharing mechanism and maximize the energy deposited in CPB, thereby the number of fast photons produced.Placing a heavy, thin scintillator between the source and the CPB layer allows the former to stop the incoming -ray via the photoelectric effect, while the recoil photoelectron can escape from it and travel into the CPB where it deposits the remaining energy generating ultra-fast photons.This phenomenon (depicted in Figure 4a) is the idea behind the heterostructure concept for TOF-PET, as the shared events (i.e., events depositing energy in both materials, see Figure 4a) improve the CTR. [22,24,25]ow to identify shared events depends on the materials that are combined.In our case, GAGG and CPB show different light output and decay kinetics.This is reflected by a different pulse shape.The amplitude and the integrated charge of the pulse are two features allowing us to perform event classification, as the correlation plot between these two quantities shows in Figure 4b.Also, the signal rise time (defined as the time interval between when the signal crosses two fixed thresholds) allows to distinguish events based on the material where the energy was deposited.In Figure 4c, the correlation between the signal rise time and the integrated charge is shown.We can observe two peaks in the signal rise time distribution: one centered around 200 ps, which also corresponds to the majority of photopeak events (peak at about 18 nWb in the integrated charge distribution).The second and sharper peak is centered around 120 ps.Most of these events have low integrated charge (see the sharp peak at about 1 nWb), but we observe a long tail extending along the charge distribution.These are events shared or depositing energy only in CPB.
We evaluated the CTR of the photopeak events and that of the CPB/shared events (identified as shown by the red boxes in Figure 4c) obtaining 272 and 118 ps, respectively.It is important to note that the twofold improvement would be even greater if we were selecting only shared photopeak events.However, their identification is not straightforward for this specific case, and we decided to select events depositing at least part of the energy in CPB according to the signal rise time, without imposing any constraint on the amount of total energy deposited.
This measurement constitutes a successful proof of the concept of how CPB can be coupled to bulk scintillators to enhance their timing performance at high (511 keV) energy.Further optimization from the material standpoint is needed to incorporate CPB in a complete (multi-layer) heterostructure.It is not feasible to obtain such a structure on a large scale with CPB in solution, and its embedding in a host matrix is therefore necessary.However, the present embedding techniques do not allow for reaching high concentrations of CPB nanocrystals while keeping the sample transparent. [16]To pursue this R&D line, new candidates as host matrix (e.g., glass instead of polystyrene because of its higher density and better radiation hardness) and embedding techniques will be considered.

Conclusion and Outlook
We fabricated CsPbBr 3 (CPB) nanocrystalline thin films on BGO, LYSO:Ce, and GAGG:Ce,Mg bulk scintillators to study their influence on the timing capabilities of the detector.Our work was focused on time resolution under soft X-ray (10 keV) excitation, but in the case of GAGG:Ce,Mg, the samples were also tested under 511 keV -ray excitation to provide some insight into their possible application in time-of-flight positron emission tomography (TOF-PET).
We demonstrated that in all cases, the CPB layer significantly improved the performance of the respective bulk scintillators, reaching values around 240 ps regardless of the type of the bulk scintillator.Already the thinnest tested layer, about 1 μm thick, resulted in an acceleration of time resolution from 2.4 ns of solely GAGG:Ce,Mg to about 400 ps.The layer of stand-alone CPB nanocrystals, directly deposited on the SiPM, featured a time resolution below 200 ps, without causing any loss in collected light after the deposition of CsPbBr 3 nanocrystals in all our samples.
The test under 511 keV -ray excitation of the heterostructure unit comprising of GAGG:Ce,Mg and CPB revealed that the identification of shared events leads to more than twofold improvement in time resolution compared to that of solely GAGG:Ce,Mg, almost reaching 100 ps.This result paves the way for the investigation of this material for a wider range of applications using energies above a few hundred keV, from TOF-PET to HEP.
In conclusion, our results confirm the capability of CPB to boost the timing performance of the detector.Despite the typical energy range of X-ray imaging is higher than the one available in our laboratory (20-120 keV), more than half of this en-ergy is already covered by our test sample, and, by extrapolation of our results (DTR of 240 ps at 10 keV), the required time resolution below 100 ps should already be achieved at 60 keV (240 ps × √ 10 keV∕60 keV).Future work should be focused not only on testing the present samples under hard X-rays to validate the extrapolation, but also on embedding the nanocrystals in a suitable matrix.This step will not only protect them from detrimental environmental effects, but it will also enable them to obtain a large-scale multi-layer heterostructure, as more elements will be needed to fully stop higher energy radiation.We therefore expect that proper optimization of our samples should allow us to reach coverage of all TOF X-ray energies and, as the ultimate goal, TOF-PET and eventually HEP experiments.
Sample Preparation: CsPbBr 3 nanocrytals were synthesized by the standard hot-injection method following the procedure described by Protesescu et al. [12] The preparation of cesium oleate was modified by using a high Cs:OA ratio of 1:5 to yield Cs-oleate soluble at room temperature. [32]bBr 2 (0.752 mmoL) was dissolved and dried in 20 mL of octadecene (ODE) using 1.78 mL of oleic acid (OA) and 2 mL of oleylamine (OAm) under vacuum at 110 °C.Then, the temperature was increased in an argon atmosphere to 170 ˜○C and 0.5 mL of cesium oleate (0.4 m) was injected.The reaction was quenched after 10 s in an ice bath.
Hot-injection synthesis was followed by the ligand exchange reaction using DDAB. [33]Crude reaction mixture was added to the DDAB solution in toluene and mixed for 2 min.A ratio of 3 mL of reaction mixture and 2 mL of 55 mm DDAB solution was used.Nanocrystals were precipitated by ethyl acetate and isolated by centrifugation.To obtain a final solution with a concentration of around 35 mg mL −1 , CsPbBr 3 nanocrystals were redispersed in toluene.Detailed characterization of the luminescence and other properties of as-prepared CsPbBr 3 nanocrystals can be found in the work of Děcká et al. [30] CsPbBr 3 thin films on various substrates were prepared by drop-casting of a colloidal solution onto a wafer.In standard procedure, the solution was added by 1 μL drops, and the drop-casting procedure was repeated as many times as needed to deposit the desired amount of nanocrystals.
Decay Kinetics: The scintillation kinetics of CPB were measured in time-correlated-single-photon-counting (TCSPC) mode [34] under X-ray excitation.A few layers of CPB solution were drop-casted on a glass plate and irradiated with a pulsed X-ray tube (XRT N5084 from Hamamatsu) operating at 40 kV.The scintillation light was collected in reflection mode with a hybrid photomultiplier tube (HPM 100-07 from Becker&Hickl) optimized for TCSPC measurements.For a more detailed description of the experimental setup, please refer to Pagano et al. [11] The data were fitted with the convolution between the instrumental response function of the system IRF (having 160 ps FWHM) [11,28] and the intrinsic scintillation rate.The latter was modeled as the sum of three bi-exponential functions and the Dirac-delta function.This model was suggested by Děcká et al. [16] to properly take into account the ultra-fast emission of CPB nanocrystals, on the trail of the one used by Gundacker et al. [28] to describe the scintillation time profile of BGO, which comprised both scintillation and prompt Cherenkov photons.
The sample showed an instantaneous rise time that could not be resolved by the IRF of the system.As it was found that it did not improve the quality of the fit, in order to have more stability on the decay part (as in the bi-exponential function, the rise and decay components are correlated), it was fixed at 0 ps.

Time Resolution under X-Ray Excitation:
The experimental setup and method used to measure the time resolution under X-ray excitation is thoroughly discussed by Pagano et al. [11] In brief, the samples were coupled to a SiPM (S13360-3050CS from Hamamatsu, 53 V breakdown voltage, 61 V bias voltage) collecting the scintillation light produced following X-ray excitation.The SiPM signal was read out by optimized high-frequency electronics, which separately process the time and energy information. [10,29]As the authors used a pulsed X-ray tube (XRT N5084 from Hamamatsu), the time resolution was obtained as the FWHM of the time delay distribution between the SiPM signal and the external trigger of the X-ray tube.By integrating the energy signal, information about the light output of the sample can also be obtained. [11]he time resolution of the stand-alone GAGG, BGO, and LYSO plates was measured first, optically coupling them with the SiPM using Meltmount glue.Later, 12 μL of CPB, which considering the surface of the plates, corresponded to a thickness of about 50 μm, were drop-casted on each of them.This allowed comparing the performance of the samples with and without CPB while keeping the same coupling conditions with the SiPM.
It is worth emphasizing that the positioning of the CPB layer to directly face the X-ray beam (see Figure 5a) was important for this application as, given the low energy of the radiation, it would otherwise be entirely absorbed by the dense scintillator used.
To study the influence of different thicknesses of the CPB layer, a single drop of 0.5 μL of CPB solution was deposited on a GAGG plate, and the amount was subsequently increased by depositing up to 2 and 5 μL.For the same considerations as made above, these three volumes corresponded to about 1, 8, and 20 μm thickness, respectively.To assess the timing performance of solely CPB nanocrystals, 12 μL of their solution was drop-casted directly on the SiPM.
Calculation of the X-Ray Attenuation in CPB Layer: The peak energy of the used X-rays was 10 keV (characteristic X-rays of Tungsten).The corresponding mass attenuation coefficient was obtained from the NIST XCOM database [35] μ m = 104.9cm 2 g −1 .The linear attenuation coefficient was calculated using the CPB orthorhombic bulk crystal density 4.83 g cm −3 obtained from the ICDD PDF-2 database (version 2013), card no.01-072-9729, which was the crystallographic phase identified in the samples here. [30]t should be emphasized that the value used for the CPB density was one of the bulk crystal.Nanocrystals generally have a lower density due to the significant proportion of surface atoms in their structure.Moreover, they were surrounded by organic ligands, further reducing the overall stopping power of such layers.Another factor to take into account was the nonuniformity in the thicknesses of the layers due to the drop-casting process.Finally, the X-ray energy distribution had a tail extending up to 40 keV due to bremsstrahlung radiation.The obtained values should therefore be considered as rough estimates, however useful to get an idea of the variation in stopping power with the increasing thickness.
Coincidence Time Resolution under 511 keV -Ray Excitation: For this study, 20 μL of CPB were drop-casted on a 6 × 6 mm 2 (corresponding to a thickness of about 20 μm) SiPM (S13360-6050PE from Hamamatsu, 53 V breakdown voltage, 61 V bias voltage) and placed a 6 × 6 × 0.2 mm 3 GAGG plate on top of it (see Figure 5b).This geometry allowed for exploiting the energy sharing mechanism outlined in Section 2.3.
The experimental setup used to measure the coincidence time resolution under 511 keV -ray excitation was the one described by Gundacker et al. [36] The samples were placed on the opposite sides of a 22 Na source and measured in coincidence.For this study, the samples were measured in coincidence with a reference crystal (2 × 2 × 3 mm 3 LSO:Ce:Ca0.4%,61 ps CTR FWHM).
The SiPM signal was read out by a high-frequency readout circuit, [10,29] like the one used for the time resolution under X-ray, and finally digitized by a LeCroy DDA735Zi oscilloscope (3.5 GHz bandwidth, 20 Gs s −1 sample rate).
Figure 6 shows the diagram of the digitized SiPM signal and the measurement extrapolated from the waveform for the data analysis.Both the amplitude and integrated charge of the energy signal were recorded: for materials with sufficiently different decay kinetics, the correlation between these two features enabled a clear pulse shape discrimination, allowing to distinguish the events according to the material where the energy was deposited. [25]Figure 4b presents the resulting 2D histogram showing the correlation between the integrated charge and the amplitude of the energy signal: two regions of accumulation of events could be distinguished, attributed to events with the energy deposition solely in either GAGG or CPB.The events lying in between were the shared events.
The rise time of the time signal was measured as the time difference between the two fixed thresholds through which the signal passes (see Figure 6).This quantity depended on the scintillation kinetics of the material (but must not be confused with the scintillation rise time) and, as shown in Figure 4c, also allowed to discriminate the events according to the material where the energy was deposited.
Finally, the time delay between the time signals of the test sample and the reference detector was measured.The FWHM of the resulting distribution was extrapolated to estimate the CTR.It should be noted that the measured FWHM does not directly represent the CTR of the sample, as two identical samples were not measured in coincidence.This effect was corrected by subtracting the contribution of the reference detector in quadrature,that is, CTR = √ 2 × FWHM 2 meas.− CTR 2 ref .

Figure 1 .
Figure 1.Scintillation time profile of solely the CsPbBr 3 layer on the nonscintillating glass wafer.Blue dots are the measured data, the green line is their average, and the red curve is the fit function.The intrinsic scintillation rate was modeled as the sum of three exponential decay components ( d, 1/2/3 ) and the Dirac-delta function () to describe the ultra-fast emission of the sample.

Figure 2 .
Figure 2. a) Time delay distribution of the different samples measured: GAGG, LYSO, and BGO (yellow, green, and blue, respectively) without and with the CPB layer (dotted and solid lines, respectively) and b) corresponding measured DTR (FWHM) values.c) Time delay distribution of bulk GAGG without deposited CPB solution (yellow), with 0.5 μL (blue), 2 μL (dark green), and 5 μL (light green) of CPB solution (inset: geometry for time resolution measurements under X-ray excitation), and d) corresponding measured DTR values (FWHM and FWTM).

Figure 3 .
Figure 3. Energy spectrum under X-ray irradiation of the different samples measured: GAGG, LYSO, and BGO (yellow, green, and blue, respectively) without and with CPB layer (dotted and solid lines, respectively) and only CPB (red).

Figure 4 .
Figure 4. a) Schematically depicted different interactions of -ray with the sample.b) 2D histogram showing the correlation between the integrated charge and amplitude of the signal.c) 2D histogram showing the correlation between the signal rise time and the integrated charge, together with the projection of the individual distribution.d) Time delay histogram of the photopeak and CPB/shared events and their associated FWHM values.

Figure 5 .
Figure 5. Geometry of the sample for time resolution measurements a) under X-ray and b) under 511 keV photon excitation.

Figure 6 .
Figure 6.Diagram of the SiPM signal outputs of the two detector in coincidence, digitized by the oscilloscope and the extrapolated measurements.