A novel Li⁺-doped CsCu₂I₃ single crystal for dual gamma–neutron detection

Abstract

Halide perovskites are among the most attractive research hotspots in optoelectronic applications due to their excellent luminescence properties. In this paper, Li⁺ was successfully incorporated into a one-dimensional perovskite CsCu₂I₃ host and constructed as a novel neutron/gamma detection scintillator via the Bridgman method. XRD and ICP-MS were carried out, and the results were discussed. The Li-doped CsCu₂I₃ crystal exhibits a main emission peak at 575 nm originating from strongly localized 1D exciton emission and a high photoluminescence quantum efficiency of 18.7%. The CsCu₂I₃:Li crystal has a light yield of 10 900 photons per MeV and an energy resolution of 11.5% under ¹³⁷Cs irradiation. The figure-of-merit (FoM) of the CsCu₂I₃:Li crystal has been estimated (about 0.8) under ²⁵²Cf radiation source. The crystal has a potential application for dual gamma–neutron detection.

---

Efficient neutron detectors have been used widely in many fields, such as homeland security, crystallography and medicine.¹ Because of the increasing application of neutron detection, scintillators that can simultaneously detect or distinguish neutrons and gamma rays have attracted wide attention. A number of inorganic scintillating crystals containing ⁶Li have been discovered and developed for neutron detection, including LiI:Eu (ref. 2) and elpasolite-structured crystals like Cs₂LiYCl₆:Ce (CLYC:Ce),³ Cs₂LiLaBr₆:Ce (CLLB:Ce),⁴ and LiCaAlF₆:Eu.⁵ However, these crystals have some disadvantages, such as extremely high hygroscopicity and incongruent melting.

In recent years, because of the advantages of confinement exciton emission, a large Stokes shift, good air stability and high photoluminescence quantum yield (PLQY), perovskite-structured metal halide materials have attracted more and more attention for radiation detectors.^6–8 At present, many promising Cu-based low-dimensional perovskite scintillators with excellent scintillation properties have been discovered, including Cs₃Cu₂I₅ (ref. 9) and CsCu₂I₃.¹⁰ They can be used as general scintillators to detect a wide energy range from soft X-ray to hard gamma radiation. Recently, Tl⁺ doped Cs₃Cu₂I₅ single crystal was developed and has a high light output of 87 000 photons per MeV under ¹³⁷Cs source. Meanwhile, it presents an extremely low X-ray induced afterglow of 0.03% at 10 ms and an excellent energy resolution of 3.4% at 662 keV.¹¹ In our previous work, a ⁶Li⁺ doped Cs₃Cu₂I₅ crystal was also prepared and presented excellent neutron and gamma discrimination (FoM > 2.0).¹² However, the Cs₃Cu₂I₅ crystal still undergoes incongruent melting. Fortunately, the CsCu₂I₃ crystal exhibits congruent melting and has a lower melting point (371 °C) and good air stability. It has the great advantage of the preparation of large-size single crystals. Moreover, as previously reported,¹³ the pure CsCu₂I₃ single crystal has a high light yield (16 000 photons per MeV) and a good energy resolution of 7.8% under ¹³⁷Cs source at 662 keV. Similarly, the CsCu₂I₃ crystal has potential application in the neutron detection field by doping with Li⁺. In this paper, an environmentally friendly scintillator with good air-stability for dual gamma-ray and neutron detection is developed by enriched ⁶Li⁺ doping. The single crystal of ⁶Li-doped CsCu₂I₃ has a scintillation yield of 10 900 photons per MeV under ¹³⁷Cs source at 662 keV. The FoM value is estimated to be about 0.8 under ²⁵²Cf source.

Anhydrous high-purity CuI (99.999%, ALDRICH), CsI and ⁶LiI (99.99%, ⁶Li enrichment above 95%, APL Engineered Materials, Inc.) were used as starting materials. High-quality Li⁺-doped CsCu₂I₃ single crystals with a size of Φ 12 mm × 30 mm were grown successfully via a self-seeding vertical Bridgman technique with a capillary tube, as shown in Fig. S1a (ESI†). The temperature field for CsCu₂I₃ growth is shown in Fig. S1b.† The experimental results show that the CsCu₂I₃ crystal is easy to crack during the cooling and machining processes, as shown in Fig. S1c.† In order to avoid structural defects and crystal cracks, a suitable temperature gradient (20 °C cm⁻¹), slow growth rate (<1 mm h⁻¹), and cooling rate (<5 °C h⁻¹) are crucial for obtaining a high-quality CsCu₂I₃ crystal. The incorporation of Li ions was confirmed by XRD and inductively coupled plasma mass spectrometry (ICP-MS). Furthermore, the PLQY, X-ray excited emission luminescence (XEL) and decay time were also measured. Finally, the application potential of CsCu₂I₃:Li crystals in the field of dual gamma–neutron detection is explored. The detail of the growth process and characterization methods can be seen in the ESI.†

The transparent and free of impurities crystal sample with a size of Φ 12 mm × 3 mm was machined, as shown in Fig. 1a. Bright yellow light can be observed under ultraviolet light. The powder XRD patterns of the as-grown CsCu₂I₃ and CsCu₂I₃:⁶Li are shown in Fig. 1b. The patterns were matched well with that of the PDF card #45-0076,¹⁴ and no other phase can be identified. Therefore, the Li⁺ doped CsCu₂I₃ crystal still belongs to the orthogonal space group Cmcm. Two spatially oriented [CuI₄]³⁻ tetrahedral units are alternately connected to form a 1D [Cu₂I₆]⁴⁻ tetrahedral unit chain separated by Cs, thus forming the 1D electronic structure of CsCu₂I₃,¹⁵ as exhibited in Fig. S2a.† This unique localized structure can result in the formation of self-trapping excitons. From Fig. S2b,† it can be found that the position of diffraction peaks shifts to a higher angle after Li⁺ doping. The radius of Li⁺ (0.59 Å) is closer to Cu⁺ (0.60 Å), while it is far less than the radius of Cs⁺ (1.74 Å). Therefore, the crystal lattice shrinks when Cu⁺ is replaced by Li⁺. It is rational that Li⁺ tends to substitute the site of Cu⁺.

Fig. 1 (a) As-grown CsCu₂I₃:5 mol%Li crystal sample and the 3 mm thick machined CsCu₂I₃:5 mol%Li crystal sample under sunlight and UV irradiation. (b) Powder XRD patterns of pure and different Li⁺-doped CsCu₂I₃ crystals. (c) Sampling location for testing the Li concentration by ICP-MS. (d) The fitted curve of the segregation coefficient. (e) Actual measured value of the Li concentration at the sampling location.

In order to further prove the successful incorporation of Li ions and calculate the segregation coefficient of Li⁺ in the CsCu₂I₃ crystal, the actual Li concentration was measured by an ICP-MS method. First, five different positions were determined along the direction of the CsCu₂I₃:5mol%Li crystal growth axis for a small amount of sampling, as shown in Fig. 1c. The Li⁺ doping concentration was confirmed and increased from 2.92 mol% to 5.13 mol% along the direction of the crystal growth, as listed in Table S1.† Because of the slow growth rate of the CsCu₂I₃ crystal, the growth process can be regarded as quasi-static. Therefore, the segregation coefficient K_Li of the CsCu₂I₃:5mol%Li crystal can be calculated using the following formula:¹⁶

where C_L is the nominal doping concentration, L represents the total length of the CsCu₂I₃:Li crystal, and the relative position and doping concentration of any two points on the crystal are (Z₁, ) and (Z₂, ), respectively. A segregation coefficient K_Li of 0.45 was obtained by linear fitting in the CsCu₂I₃ crystal, as shown in Fig. 1d. The calculated curve is in good agreement with the actual doping concentration, as shown in Fig. 1e. The segregation coefficient is low (<1), indicating that a large number of Li ions will be enriched in the tail of the CsCu₂I₃ crystal.

The optical absorption spectra of pure CsCu₂I₃ and CsCu₂I₃:5 mol%Li crystals are shown in Fig. 2(a and b), and the inset is the fitted optical band gap. Based on the onset of the light absorption edge, the optical band gaps of pure CsCu₂I₃ and ⁶Li⁺-doped CsCu₂I₃ crystals are calculated to be 3.31 eV and 3.23 eV, respectively. The band gap was narrowed after Li⁺ doping. In order to discuss the effect of Li⁺ doping on the luminescence properties, the PL, PLE, and RL spectra of pure and Li⁺-doped CsCu₂I₃ were measured, as shown in Fig. 2(c and d). The Li⁺-doped CsCu₂I₃ crystal has a broad excitation and emission bands peaking at 334 and 575 nm, respectively, which are consistent with those of the pure CsCu₂I₃ crystal. The Li⁺-doped CsCu₂I₃ crystal still has a large Stokes shift of 1.54 eV and no self-absorption. The RL spectra under X-ray excitation are similar to the PL spectra with a slight redshift. The red shift maybe related with the ionization luminescence mechanism, which is more efficient in capturing charge carriers and can be more observed in RL spectra, as discussed in ref. 17. Obviously, the strategy of Li⁺ doping does not introduce a new luminescent center. The PL decay curves of pure CsCu₂I₃, as well as CsCu₂I₃:5mol%Li, were measured at room temperature (λ_ex = 334 nm; λ_em = 575 nm), as shown in Fig. S3.† The decay curves of pure CsCu₂I₃ and CsCu₂I₃:5mol%Li crystal can be well fitted using two-components. The decay time of the CsCu₂I₃ crystal is 82 ns (56%) and 1332 ns (44%). Compared to the pure CsCu₂I₃ crystal, the proportion of the fast component increased slightly from 56% to 63%, while the decay time of the slow component decreased from 1332 ns to 1204 ns after Li⁺ doping. The fast decay component is attributed to trap-assisted recombination of the photo-excited electrons and holes at the CsCu₂I₃ surface, while the slow decay component is related with the recombination inside CsCu₂I₃, as reported in a previous study.¹⁴ Obviously, Li⁺-doping was beneficial to improving the fast decay component. The PLQY of the pure and Li⁺-doped CsCu₂I₃ crystal was measured at room temperature. A blank quartz plate was also measured in an integrating sphere under excitation as a reference, using an absolute photoluminescence measurement system with an integrating sphere, as shown in Fig. 2(e and f). It is found that the PLQY improved from 17.4% to 18.7% when Li⁺ was introduced. This indicates that the PLQY has a slight improvement after Li⁺ doping.

Fig. 2 (a and b) The optical absorption spectra of the CsCu₂I₃ and CsCu₂I₃:5mol%Li crystals, and the inset is the fitted optical band gap. (c and d) The PLE, PL, and X-ray excited RL spectra of CsCu₂I₃ and CsCu₂I₃:5 mol%Li. (e and f) The PLQY spectra of CsCu₂I₃ and CsCu₂I₃:5mol%Li. The reference curve was measured by placing a blank quartz plate in the integrating sphere.

The energy resolution, light yield and decay time of the pure and CsCu₂I₃:5 mol%Li crystals were measured under ¹³⁷Cs gamma-ray source at 662 keV with a Hamamatsu R6231-100 PMT. To estimate the absolute light yield of the crystal samples, the standard sample of the CsI:Tl crystal was also chosen and measured as a reference. The pulse height spectra are plotted in Fig. 3a. Under ¹³⁷Cs excitation source, the absolute light yield of the CsI:Tl standard sample is about 56 000 photons per MeV. The absolute light yield of the pure CsCu₂I₃ and CsCu₂I₃:5 mol%Li crystal samples was estimated to be 10 248 photons per MeV and 10 900 photons per MeV, respectively. The light yield is improved by Li⁺ doping, which is consistent with the results of PLYQ. But, the energy resolution of the CsCu₂I₃ crystal is deteriorated from 10.5% to 11.5% after Li⁺ doping at 662 keV, which is worse than that of the CsI:Tl crystal (E.R. = 6.8%). The deterioration of the energy resolution is related to the transparency of the crystal and its degree of coloring. The scintillation decay profiles of both crystals were measured at room temperature under gamma ray. All the decay curves can be fitted well using a three-exponential function, as shown in Fig. 3b. The decay time of the CsCu₂I₃ crystal is 4.7 ns (2.5%), 94.1 ns (67.1%) and 316 ns (30.4%), which should be ascribed to the self-trapped exciton recombination. The fast component was shortened from 94.1 ns to 70 ns, while the slow component was reduced to 233 ns after Li⁺ doping. It is possible that the charge carrier mobility will be improved by Li⁺ doping, as observed in ref. 18. However, the mechanism is still unclear. More experiments need to be done.

Fig. 3 (a) Pulse height spectra of pure CsCu₂I₃ and CsCu₂I₃:5 mol%Li and CsI:Tl under ¹³⁷Cs γ-ray irradiation. (b) Scintillation decay curves of pure CsCu₂I₃ and CsCu₂I₃:5 mol%Li. (c) PSD scatter plot and (d) PSD spectrum of the 5 mol%Li⁺-doped CsCu₂I₃ crystal under the excitation of a ²⁵²Cf source.

Finally, the performance of the Li⁺-doped CsCu₂I₃ crystal was examined further for dual gamma ray and neutron detection application. A two-dimensional PSD scatter plot was generated for the neutron and gamma-ray events under a ²⁵²Cf radioactive source, as shown in Fig. 3c. It can be recognized that the gamma signal and the neutron signal have a relatively clear separation. Fig. 3d shows one-dimensional projection of the neutron and gamma-ray events to the Y axis. Gaussian fitting of the neutron and gamma ray peaks is performed to calculate the number of merits (FoM), which is used to evaluate the discrimination performance of the neutrons and gamma rays, which can be defined by the following equation.¹⁹

D is the distance between the centroids of the neutron and gamma peaks in the PSD ratio histogram, and nFWHM and γFWHM are the corresponding FWHMs. An FoM value of 0.8 is estimated. At present, the FoM value is still low, and future work should focus on improving the Li ion doping concentration and improving the quality of the crystal to further improve the neutron detection efficiency.

In summary, Li⁺-doped CsCu₂I₃ bulk single crystals were prepared by a Bridgman method for the first time. The Li⁺ effective segregation coefficient (k_eff) of the CsCu₂I₃:Li crystal was calculated to be about 0.45 using the deformed equilibrium segregation equation. The PLQY value increased slightly from 17.4% to 18.7% after Li⁺ doping. The absolute light yield of Li⁺-doped CsCu₂I₃ is about 10 900 photons per MeV under ¹³⁷Cs radiation, and Li⁺ doping also leads to a faster decay time under gamma-ray. Due to the successful combination of the ⁶Li isotope, the CsCu₂I₃ crystal material can be used in the field of dual gamma–neutron detection. The crystal quality and scintillation performance need to be improved in the future.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

The authors gratefully acknowledge the financial support provided by the “Pioneer” and “Leading Goose” R&D Program of Zhejiang (Grant No. 2022C01046), the National Natural Science Foundation of China (NSFC) (No. 11975220, 51972291), the Natural Science Foundation of Zhejiang (No. LGG22E020001), and the Fundamental Research Funds for the Provincial Universities of Zhejiang (No. 2021YW14).

References

1. K. Yang, P. R. Menge and V. Ouspenski, IEEE Trans. Nucl. Sci., 2017, 64, 2406–2413.
2. A. Syntfeld, M. Moszynski, R. Arlt, M. Balcerzyk, M. Kapusta, M. Majorov, R. Marcinkowski, P. Schotanus, M. Swoboda and D. Wolski, IEEE Trans. Nucl. Sci., 2005, 52, 3151–3156.
3. W. M. Higgins, J. Glodo, U. Shirwadkar, A. Churilov, E. V. Loef, R. Hawrami, G. Ciampi, C. Hines and K. S. Shah, J. Cryst. Growth, 2010, 312, 1216–1220.
4. J. Lin, Q. H. Wei, D. Zhang, G. Tang, L. S. Qin and H. S. Shi, Cryst. Res. Technol., 2019, 54, 1900047.
5. T. Yanagida, N. Kawaguchi, Y. Fujimoto, K. Fukuda, Y. Yokota, A. Yamazaki, K. Watanabe, J. Pejchal, A. Uritani, T. Iguchi and A. Yoshikawa, Opt. Mater., 2011, 33, 1243–1247.
6. R. C. Lin, Q. Zhu, Q. L. Guo, Y. M. Zhu, W. Zheng and F. Huang, J. Phys. Chem. C, 2020, 124, 20469–20476.
7. F. Maddalena, L. Tjahjana, A. Xie, A. Arramel, S. Zeng, H. Wang, P. Coquet, W. Drozdowski, C. Dujardin, C. Dang and M. D. Birowosuto, Crystals, 2019, 9, 88.
8. B. Yang, L. X. Yin, G. D. Niu, J. H. Yuan, K. H. Xue, Z. F. Tan, X. S. Miao, M. Niu, X. Y. Du, H. S. Song, E. Lifshitz and J. Tang, Adv. Mater., 2019, 31, 1904711.
9. X. M. Li, J. X. Chen, D. D. Yang, X. Chen, D. L. Geng, L. F. Jiang, Y. Wu, C. F. Meng and H. B. Zeng, Nat. Commun., 2021, 12, 1–9.
10. R. C. Lin, Q. L. Guo, Q. Zhu, Y. M. Zhu, W. Zheng and F. Huang, Adv. Mater., 2019, 31, 1905079.
11. S. L. Cheng, M. Nikl, A. Beitlerova, R. Kucerkova, X. Y. Du, G. D. Niu, Y. C. Jia, J. Tang, G. H. Ren and Y. T. Wu, Adv. Opt. Mater., 2021, 9, 2100460.
12. P. Xiang, Q. H. Wei, C. E. Wang, P. Q. Cai, Y. F. Tong, G. Tang, F. Yang, H. S. Shi, Z. G. Liu and L. S. Qin, J. Mater. Chem. C, 2022, 10, 15400–15407.
13. S. L. Cheng, A. Beitlerova, R. Kucerkova, E. Mihokova, M. Nikl, Z. Y. Zhou, G. H. Ren and Y. T. Wu, ACS Appl. Mater. Interfaces, 2021, 13, 12198–12202.
14. X. M. Mo, T. Li, F. C. Huang, Z. X. Li, Y. L. Zhou, T. Lin, Y. Ouyang, X. M. Tao and C. F. Pan, Nano Energy, 2021, 81, 105570.
15. Y. Hui, S. Y. Chen, R. C. Lin, W. Zheng and F. Huang, Mater. Chem. Front., 2021, 5, 7088–7107.
16. X. G. Zhang, Z. Kang, Z. C. Cai, H. W. Zhai, Z. Yin, W. Q. Jie and T. Wang, J. Cryst. Growth, 2021, 573, 126308.
17. Q. Wang, C. E. Wang, Z. H. Wang, X. L. Sun, M. Nikl, X. P. Ouyang and Y. T. Wu, J. Phys. Chem. Lett., 2022, 13, 9066–9071.
18. F. Maddalena, A. Xie, A. Arramel, M. E. Witkowski, M. Makowski, B. Mahler, W. Drozdowski, T. Mariyappan, S. V. Springham, P. Coquet, C. Dujardin, M. D. Birowosuto and C. Dang, J. Mater. Chem. C, 2021, 9, 2504–2512.
19. K. N. Li, X. P. Zhang, Q. Gui, P. Jin and G. Tian, Nucl. Sci. Tech., 2018, 29, 1–6.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d2ce01263d

---