Cathodoluminescence spectral and lifetime mapping of Cs4PbBr6: fast lifetime and its scintillator application

Highly efficient green emission of Cs4PbBr6 has been attributed to intermediate states formed by embedded CsPbBr3 nanocrystals or defects. However, direct experimental confirmation of the presence of such nano-emitters is not straightforward and the emission mechanism remains elusive. By using cathodoluminescence (CL) imaging with a high spatial resolution, we demonstrate that CsPbBr3 nanocrystals within the Cs4PbBr6 matrix contribute to the green emission, exhibiting optical behavior distinct from the matrix. Additionally, we explore its potential as an electron beam scintillator, given its high CL intensity and exceptionally short lifetime.


M
etal halide perovskites (MHPs) are recognized as next-generation materials for photovoltaic and lightemitting devices through their outstanding optical properties, driving a wide range of research areas. 1,2)For the light emitter applications, Cs 4 PbBr 6 stands out among these MHPs due to its highly efficient green emission and easy solution-based synthesis process, making it a potential material for applications in LEDs and laser diodes. 3,4)The highly luminescent characteristics at RT, short emission lifetime, and mass-production possibility of this material also extend the application to scintillators, used in imaging and detection of high energy particles/rays. 5,6)Despite all these diverse applications of Cs 4 PbBr 6 , the mechanism of its efficient green emission still remains unclear.Initially, the green emission had been attributed to the recombination of excitons that are trapped in the local structures. 7)][13][14][15] Indeed, samples coexisting with CsPbBr 3 and Cs 4 PbBr 6 exhibited intense green emission, and crystal structure analysis through transmission electron microscopy (TEM) measurements has revealed the existence of CsPbBr 3 , providing support for this hypothesis. 12,13)ntrinsic Cs 4 PbBr 6 defects, such as bromine vacancies, have also been proposed as candidates for the green emission source. 16,17)hile the origin of the green emission has been debated, experimental results supporting luminescence arising from microscopic structures such as CsPbBr 3 nanocrystals or defects have not been conclusively reported due to the spatial resolution limitations in photoluminescence (PL) measurements.
Cathodoluminescence (CL) imaging, which offers higher spatial resolution than PL measurements, revealed the presence of bright spots in the Cs 4 PbBr 6 particles and that only these regions exhibited green emission. 18)However, the CsPbBr 3 nanocrystals in the study are too sparse to exhibit such high luminescence intensity as the entire material.Therefore, further confirmation and understanding of the local optical properties, such as spectrum or emission lifetime of these nanocrystals, is needed.
In this study, we utilized CL spectral imaging as well as CL lifetime evaluation using a Hanbury-Brown-Twiss (HBT) interferometer to investigate the optical properties of Cs 4 PbBr 6 samples.The obtained results support that the green emission of Cs 4 PbBr 6 powder samples originates from CsPbBr 3 nanocrystals, which were observed to have a high CL intensity and extremely short lifetime.][21] We also demonstrated the application of the Cs 4 PbBr 6 scintillator as a part of our scanning transmission electron microscopy (STEM) imaging system.
The synthesis of the Cs 4 PbBr 6 powder sample was conducted via a solution method. 7)CsBr and PbBr 2 precursors (1:1 molar ratio) were dissolved in dimethyl sulfoxide (DMSO) and stirred for one hour.The yellow precipitate was obtained by adding antisolvent (methanol) and washed with DMSO to remove CsPbBr 3 which co-precipitated with Cs 4 PbBr 6 .Crystalline structure analysis by X-ray diffraction confirmed the presence of Cs 4 PbBr 6 without detectable CsPbBr 3 .
CL measurements were performed using a STEM (JEOL JEM-2000FX), as illustrated in Fig. 1(a).The emitted light from the sample was collimated by a parabolic mirror and transferred to a spectrometer for the CL spectrum measurement and an HBT interferometer for the lifetime measurement.The HBT method allows for lifetime measurement without pulsing the electron beam by evaluating the obtained second-order correlation function ( ) ( ) t g .
2 [22][23][24][25][26][27] By synchronizing the light detection and the electron beam scan, CL mappings were performed. W used the acceleration voltage of 80 kV with the beam current of 11 pA in order to avoid sample damage.All measurements were conducted at RT.
The secondary electron (SE) image and the panchromatic CL image of the measured powder are presented in Figs.2(a) and 2(b), respectively.The CL image revealed a high emission intensity at the central part of the particle, particularly highlighting the presence of bright spots with high emission intensities.The green emission that was observed only from the bright spots can be attributed to CsPbBr 3 nanocrystals, according to the previous CL study. 18)The matrix regions other than the bright spots without strong CL signal were considered as the non-luminescent host phase of Cs 4 PbBr 6 .Considering the similarities in the CL image appearance, the bright spots observed in our study are likely CsPbBr 3 nanocrystals.A region with densely distributed bright spots [red square in Figs.2(a) and 2(b)] was selected for higher magnification CL imaging and spectral measurements.The magnified CL image [Fig.2(c)] revealed the presence of many bright spots including a large (approximately 100 nm-sized) one at the upper central region of the image and a number of smaller ones.The area-integrated spectrum of the measured region [Fig.2(d)] displayed emission peaks at 375 nm and 520 nm (referred to as Peak1 and Peak2 in this paper respectively), along with a broad peak in the longer wavelength range.In the previous research, 10) Peak1 is attributed to the optical transition of Pb 2+ ions, indicating emission from the defect of the Cs 4 PbBr 6 matrix.Peak2 is the typical green emission of Cs 4 PbBr 6 , possibly the CsPbBr 3 nanocrystals.The broad peak observed in the longer wavelength range seems to correspond to Br-related defects reported for radioluminescence with X-ray excitation. 6,28)o separately analyze the contribution of the multiple peaks, spectral fitting was performed using three Gaussian functions corresponding to the three peaks, [red line in Fig. 2(d)].Through this fitting, the peak intensity, wavelengths, and FWHM of each peak can be extracted.We map Peak1 and Peak2 to see the contribution of the host phase  The intensity mapping of Peak2 showed higher intensity in the region of the bright spots [Fig.2(h)].For comparison, representative spectra were extracted from regions including a single bright spot (named "Spot1," "Spot2"), many bright spots ("Spots"), and no bright spots ("Matrix"), respectively.Spectra of "Spot1," "Spot2" and "Spots" areas exhibited strong green emission (Peak2 intensity), while the spectrum of "Matrix" exhibited a very small intensity of this Peak2.These results suggest that the CsPbBr 3 nanocrystals (corresponding to Peak2) are distributed in a homogeneous Cs 4 PbBr 6 host phase (Peak1).This is also consistent with the previous report. 18)Thanks to the relatively long penetration depth 29) (estimated around 40 μm) of the accelerated electrons in STEM, we also access the emitters buried in depth, which however has only weak intensities and gives blurred and broadened spot images compared to those on the surface (such as "Spot1" and "Spot2").This is different from the measurement by scanning electron microscopy, 18) which has lower electron acceleration energy (its penetration depth estimated around 0.5 μm at 5 keV) and is more sensitive to the surface.At the clearly recognizable bright spots ("Spot1," "Spot2"), Peak2 redshifts in wavelength and decreases in FWHM, compared to the surroundings, as seen in Figs.2(i) and 2(j).The larger spot (Spot1) gave more redshifts compared to Spot2, Spots and the surroundings, which presumably included smaller nanocrystals.Such size-dependent behavior is consistent with CsPbBr 3 nanoparticles. 5,30)o further investigate the local optical properties, lifetime measurements (HBT measurement) were conducted in the same region as the spectral analysis, as shown in Figs.3(a), 3(b).A very short lifetime (approximately 0.4 ns) was observed across the entire measured region [Fig.3(c)], which is close to the instrument time resolution limit of 0.3 ns.Lifetime mapping revealed a shorter lifetime in the region where more nanocrystals exist (higher CL intensity area on the right of the image) compared to the region with fewer nanocrystals (on the left of the image with lower CL intensity) [Fig.3(e)], but no significant lifetime change was observed at the large bright spot area [Fig.3(c)].Due to the limited spatial resolution of the lifetime mapping, which is constrained by the acquisition time of the ( ) ( ) t g 2 curve per electron beam position, it was not possible to resolve the spatial variation of the emission lifetime originating from small nanocrystals in the region on the right side of the image, where smaller nanocrystals seem densely buried.
In Fig. 3(d),  ( ) value, corresponding to the peak top of the second-order correlation function, is significantly higher at the bright spot (Spot1) than the surrounding.From the result of the spectrum [Fig.2(k)], it is evident that the green emission (Peak2) of the bright spots region exhibits higher intensity compared to the broad peak.Considering the negligible detection efficiency of the photon counter for the UV region, i.e.Peak1, the correlation function obtained from the HBT measurement suggests that only Peak2 contributes to bunching, for comparison.In the inset, the ( ) ( ) t g 2 curve of "Spot1" with a larger time window is shown.015005-3 © 2024 The Author(s).Published on behalf of The Japan Society of Applied Physics by IOP Publishing Ltd while the broad peak becomes the constant background in the ( ) ( ) t g 2 curve. 27)This implies that Peak2 has an exceptionally short lifetime, and the broad peak indicates a long lifetime beyond the measurable range in the HBT measurement, similar to the bulk radioluminescence. 28)In the ( ) ( ) t g 2 curve with the time range of 50 ns, as shown in the inset of Fig. 3(e), only a sub-nanosecond decay component is found.
Considering the fact that this Cs 4 PbBr 6 powder exhibits high CL intensity and a short lifetime of sub-nanosecond, we applied it to an electron beam scintillator for the STEM bright field (BF) imaging.This emission lifetime is significantly shorter than that of a conventional scintillator (Y 2 SiO 5 :Ce) for electron detection and imaging, which is approximately 50 ns. 31,32)To realize STEM imaging with this material, we built a STEM detector consisting of a scintillation part filled with the Cs 4 PbBr 6 powder as shown in Fig. 4(a).To detect light from the scintillator, light optics consisting of two lenses are accommodated in the detector system, as depicted in Fig. 4(b).This electron detector was installed in the camera chamber for the STEM mode.
For the STEM imaging, nanodiamond (ND) particles were used as the demonstration sample.The STEM-BF image of the ND sample obtained using the Cs 4 PbBr 6 scintillator is shown in Fig. 4(c).The STEM image shows clear bright and dark contrasts reflecting the particle shape of the sample, comparable to the conventional electron detector.Although we cannot demonstrate fast-scan due to limited instrumentation (scan system, electronics, etc), this success of realizing a fast electron detector using the Cs 4 PbBr 6 powder scintillator decisively demonstrates its potential to break the current bottleneck of the fast STEM imaging.
In summary, we synthesized Cs 4 PbBr 6 powder samples using a solution-based method and investigated their nanoscale optical properties using CL.CL spectral imaging revealed the presence of the CsPbBr 3 nanocrystals showing bright spots.We observed the CL emission around 375 nm from the matrix Cs 4 PbBr 6 phase, and green emission at around 520 nm of the CsPbBr 3 nanocrystals embedded in the host phase, which showed a size-dependent spectral shift.The lifetime measurement revealed the material's green emission has a very short lifetime.Furthermore, we successfully realized an electron detector using the Cs 4 PbBr 6 powder scintillator and demonstrated STEM-BF imaging.This highlights Cs 4 PbBr 6 as a promising candidate for ultrafast electron beam scintillators, showing its potential for the application of ultrafast STEM imaging.The mass-fabricable nature of this perovskite material is also an advantage of such applications.

Fig. 1 .
Fig. 1.(a) Schematic illustration of the STEM-CL measurement setup.CL emitted from the sample is collimated by the parabolic mirror situated above the sample and is forwarded to the spectrometer or the HBT interferometer system.(b) Photograph of the Cs 4 PbBr 6 powder sample.

Fig. 2 . 2 ©
Fig. 2. CL spectral analysis results.(a) SE image and (b) panchromatic CL image of the measured particle.(c) Panchromatic CL image magnified in the area surrounded by red in panels (a) and (b).(d) Integrated CL spectrum of the entire image area in panel c.The fitted curve is superposed.(e)-(j) Intensity, wavelength, and FWHM mappings of (e)-(g) Peak1 and (h)-(j) Peak2 as indicated in panel d.(k) CL spectra extracted from the four square regions as indicated by "Spot1," "Spot2," "Spots," and "Matrix" in panels c and h.015005-2

Fig. 4 .
Fig. 4. (a) Illustration and photograph of the prepared scintillator.The Cs 4 PbBr 6 powders are filled in a 2 mm hole in a stainless-steel plate.(b) Illustration and photograph of the detector system.(c) STEM-BF image acquired using the Cs 4 PbBr 6 scintillator.