Improvement of carrier transport properties of CsPbBr3 thin films by moisture absorption and TbCl3 doping technique

Moisture absorption and TbCl3 doping of CsPbBr3 thin films were investigated to improve the carrier transport properties. We found that post-deposition moisture-absorbing treatment improved the carrier diffusion length of CsPbBr3 thin films. The moisture-absorbing treatments under a relative humidity of about 20%–40% were effective to improve the carrier diffusion length. TbCl3 doping during the thermal evaporation of CsPbBr3 affected the structure of the deposited films. An excessive amount of TbCl3 doping leads to the formation of CsPb2Br5 additional phase, but a small amount of TbCl3 doping (1%) can improve the carrier diffusion length. The moisture-absorbing treatment and TbCl3 doping are promising techniques to improve the optoelectronic properties of CsPbBr3.


Introduction
Cesium lead bromide (CsPbBr 3 ) is one of the widegap perovskite materials which has been investigated as the light-absorbing layer of solar cells, [1][2][3][4] the light-absorbing layer of photovoltaic power converters in optical wireless power transmission, [5][6][7] and the light emitting layer of LEDs. 8,9) This material has a bandgap of about 2.3 eV 10,11) and its stability against humidity is superior to the other halide perovskite materials. 12) Therefore, this material is very suitable for the light-absorbing material of the top cell of a four-terminal tandem solar cell or the light-absorbing material of the photovoltaic power converter for blue light. The deposition of CsPbBr 3 can be performed by solution-based techniques, 12,13) thermal evaporation, 14,15) sputtering, 16) and close space sublimation. 17) Thermal evaporation is a promising technique to deposit CsPbBr 3 thin films because it allows precise control of the film thickness. In addition, thermal evaporation does not require any organic solvent. The evaporation of CsPbBr 3 from a single evaporation source is very promising due to the simple deposition process. Relatively good quality CsPbBr 3 thin films were obtained by thermal evaporation using a single evaporation source (mixture of CsBr and PbBr 2 ), but the deposition rate (0.06 nm s −1 ) still needs to be improved. 18) In our previous publication, we reported that high quality CsPbBr 3 thin films were fabricated by the combination of sequential deposition of PbBr 2 and CsBr and thermal annealing. 6) However, this process requires a very long evaporation time and high temperature annealing (500°C) to obtain high quality CsPbBr 3 thin films. This very low deposition rate is a disadvantage in reducing the cost of the devices such as solar cells. The high temperature annealing also limits the choice of the material underneath of the CsPbBr 3 . Therefore, it is important to develop a technique to deposit high quality CsPbBr 3 thin films with a high deposition rate without a high temperature annealing process.
In this study, we focus on one-step thermal evaporation of CsPbBr 3 by using CsPbBr 3 powder. Thermal evaporation of CsPbBr 3 is possible because the evaporation temperature of CsPbBr 3 is about 600°C. 12) Close space sublimation of CsPbBr 3 powder was recently reported, 17) but thermal evaporation of CsPbBr 3 by using CsPbBr 3 powder has not been well investigated and the quality of the deposited CsPbBr 3 thin films is not sufficient. The effect of humidity on the structural properties of CsPbBr 3 thin films was investigated and the positive effect of a certain amount of humidity exposure was reported. 19,20) However, the effect of the humidity exposure of CsPbBr 3 thin films on the optoelectronic properties has not been discussed in detail. Tb doping in CsPbBr 3 is also a promising technique to improve the opt-electronic properties of CsPbBr 3 thin films. The effect of Tb doping was investigated by using CsPbBr 3 thin films deposited by a solution-based technique. 13) However, there is no report on the Tb doping during the thermal evaporation of CsPbBr 3 . Therefore, we investigate a post-deposition moisture-absorbing treatment and Tb doping using TbCl 3 during thermal evaporation in order to improve the optoelectronic properties of thermally evaporated CsPbBr 3 thin films.

Experimental methods
2.1. Thermal evaporation of CsPbBr 3 CsPbBr 3 thin films were deposited by thermal evaporation of CsPbBr 3 powder (purchased from TCI) on an Eagle XG glass substrate. The base pressure of the evaporation chamber is about 3 × 10 −3 Pa. The CsPbBr 3 powder with a weight of about 200 mg was thermally evaporated by using an aluminacoated tungsten basket (Nilaco B-3). The diameter of the tungsten wire of the basket was 1 mm. The evaporation was controlled by an electric current (14.5 A). Approximately 250 nm-thick CsPbBr 3 thin films were obtained by evaporating all the CsPbBr 3 powder in the basket. Figure 1 shows the deposition rate during thermal evaporation. We started to apply electric current to the tungsten basket at 0 min. After 1.5 min, the electric current reaches the setting value of 14.5 A. The evaporation started at about 7 min and the deposition rate increased up to about 2 nm s −1 . In our study, this high rate of evaporation was employed for all samples. After the deposition of CsPbBr 3 thin films, the samples were annealed at 300°C for 3 min by using a hot plate under an N 2 atmosphere. Finally, Au coplanar electrodes with a gap of 0.2 mm were thermally evaporated on the surface of the CsPbBr 3 thin films.

Post-deposition moisture-absorbing treatment
The procedure for the post-deposition moisture-absorbing treatment is shown below. As shown in Fig. 2, an annealed CsPbBr 3 thin film on a glass substrate was dipped into 2propanol for about 1 s, and then the 2-propanol on the surface of the CsPbBr 3 thin film was removed by N 2 blow for about 5 s. The dipping and N 2 blow processes were repeated about 300-500 times. This process was carried out in a humiditycontrolled (relative humidity up to 60%) grove box. During the N 2 blow, the surface of the CsPbBr 3 thin film was cooled by the vaporization heat, therefore, a small amount of moisture was absorbed on the surface. In our study, this process is referred to as post-deposition moisture-absorbing treatment.

Thermal evaporation of TbCl 3 -doped CsPbBr 3
TbCl 3 -doped CsPbBr 3 thin films were deposited by thermal evaporation of Tb-doped CsPbBr 3 powder. The TbCl 3 -doped CsPbBr 3 powder was prepared by mixing and grinding a mixture of TbCl 3 and CsPbBr 3 powders using a mortar and a pestle in an N 2 -filled grove box. We prepared 1.0 and 2.5 mol% TbCl 3 -doped CsPbBr 3 powders. For comparison, we also prepared ground CsPbBr 3 powder as a reference. We deposited TbCl 3 -doped CsPbBr 3 thin films by using these TbCl 3 -doped CsPbBr 3 powders. The sample fabrication procedure was the same as the procedure explained in Sect. 2.1.

Characterization of the CsPbBr 3 thin films
We mainly focus on the optoelectronic properties of the CsPbBr 3 thin films. One of the most important semiconductor parameters for solar cells or photovoltaic power converters is the carrier diffusion length. The carrier diffusion length is proportional to the square root of the product of mobility and carrier lifetime (μτ product) and the photo-conductivity of a semiconductor material is proportional to the μτ product. The photo-conductivity is a good measure of the carrier diffusion length. The measurements of photo-conductivity were performed under the illumination of a solar simulator with AM1.5 spectrum and 100 mW cm −2 light intensity by using the Au coplanar electrodes. The diffusion length of the selected samples was also evaluated by Steady-State Photoconductivity (SSPC) 21,22) measurements using monochromatic light with a wavelength of 400 nm. The crystal structure and grain size were evaluated by X-ray diffraction (XRD) and atomic force microscope (AFM), respectively. All measurements were performed under a normal atmosphere.

Fabrication of CsPbBr 3 photovoltaic power converters (solar cells)
We also fabricated CsPbBr 3 photovoltaic power converters. The structure of the devices is anti-reflection (AR) film/glass/ FTO/nanocrystalline GaN (nc-GaN)/CsPbBr 3 /P3HT/Au. After ultrasonic cleaning using acetone and 2-propanol and ozone cleaning of FTO glasses, an nc-GaN layer 23) was deposited on FTO. CsPbBr 3 was thermally evaporated on the nc-GaN layer. After the thermal evaporation, the samples were annealed at 300°C on a hot plate in an N 2 atmosphere for 3 min. After the post-deposition moisture-absorbing treatment, the P3HT layer was prepared on the CsPbBr 3  The Japan Society of Applied Physics by IOP Publishing Ltd layer by spin coating of a mixture of P3HT (10 mg)/chloroform (1 ml) and subsequently annealed at 180°C for 10 min in N 2 ambient. The Au back electrode with a size of 3 mm square was thermally evaporated on the P3HT layer. Finally, the AR film (Geomatec g.moth 5021) was attached to the glass surface. The fabricated CsPbBr 3 photovoltaic power converters were characterized by I-V measurements under a solar simulator with AM1.5 spectrum with a light intensity of 100 mW cm −2 .

Results and discussion
3.1. Effect of post-deposition moisture-absorbing treatment on the properties of CsPbBr 3 thin films Figure 3 shows the influence of the number of post-deposition moisture-absorbing treatments on the diffusion length and dark conductivity of the CsPbBr 3 thin films. In these experiments, the post-deposition moisture-absorbing treatments were performed in a humidity-controlled grove box, but the conductivity measurements were performed outside of the grove box (relative humidity about 50%). In addition, the number of treatments was the cumulative number of treatments. After the treatment, the sample was taken out from the grove box and the conductivity measurements were performed. After the conductivity measurements, the sample was received the treatment in the grove box. This process was repeated.
The diffusion length for the sample without the postdeposition moisture-absorbing treatment was about 0.1 μm. For the post-deposition moisture-absorbing treatment under a relative humidity of 20% and 40%, we observed a clear improvement of carrier diffusion length from 0.1 μm to over 0.2 μm. The rapid improvement was observed from the treatment under a relative humidity of 40%, but the final diffusion length is higher for the treatment under a relative humidity of 20%. On the other hand, the diffusion length decreased by the post-deposition moisture-absorbing treatment under relative humidity of 60%. In addition, the color of the film changed from yellow to almost transparent after 20times treatment under the relative humidity of 60%, indicating that the CsPbBr 3 thin film was degraded. We also performed the post-deposition moisture-absorbing treatment under a relative humidity of less than 5%. The value of 5% is the lower detection limit of the relative humidity meter. The diffusion length was also improved by the treatment under a relative humidity of less than 5%, but the improvement rate was smaller than the treatments under a relative humidity of 20% and 40%. We also measured the carrier diffusion length of a sample treated under relative humidity less than 5% after 600 continuous treatments (without taking out the sample outside of the grove box) because there is a possibility that humidity during the conductivity measurements outside of the grove box affects the diffusion length. In this case, the improvement of the diffusion length was very small. These results clearly indicate that the humidity during the treatments is the key factor for the improvement of the carrier diffusion length and the best carrier diffusion length can be obtained by the treatment under appropriate humidity (30%-40%). The dark conductivity of the samples was also shown in Fig. 3. The dark conductivity was not influenced by the moistureabsorbing treatment after deposition under a relative humidity of less than 40%. The increase of the dark conductivity of the sample with the treatment under a relative humidity of 60% is due to the degradation of the CsPbBr 3 thin film as mentioned above. These results suggest that the improvement of the carrier diffusion length is due to the improvement of the carrier lifetime. Figure 4 shows the surface SEM images of the CsPbBr 3 thin films after the post-deposition moisture-absorbing treatments under different relative humidities. According to the SEM images, it is clear that the treatment does not affect the grain size of the CsPbBr 3 thin films. The grain size of all the samples was about 0.5 to 1 μm. This indicates that the postdeposition moisture-absorbing treatment does not affect the structure of the CsPbBr 3 thin films. Figure 5 shows the XRD patterns of the CsPbBr 3 thin films after the post-deposition moisture-absorbing treatments. From all the samples, we observed the peaks located at 11.6°, 15.1°, and 30.7°. The peak located at 44.4°was observed only from the sample after the treatment under 20% and 40% relative humidity. The peak located at 11.6°is assigned to CsPb 2 Br 5 24,25) and the others are assigned to CsPbBr 3 . [26][27][28] The ratio of the peak intensity of the peak located at 15.1°to the peak intensity of the peak located at 11.6°decreased by the treatment regardless of the relative humidity. This result indicates that the crystallinity of the CsPbBr 3 thin film slightly deteriorated by dipping the sample into 2-propanol. The presence of the peak located at 44.4°after the treatment under a relative humidity of 20% and 40% suggests that the crystalline orientation of CsPbBr 3 is slightly changed by the treatment. According to the results shown in Fig. 3, the carrier diffusion length is increased by the moisture-absorbing treatment, but the dark conductivity is almost the same. Therefore, the improvement of the carrier diffusion length is probably due to the improvement of the carrier lifetime. The surface SEM shows that the grain size does not change after the moisture-absorbing treatments. These results suggest that the moisture-absorbing treatment improves the grain boundary properties of CsPbBr3 thin films. More detailed analysis is needed to clarify the improvement mechanism.
We also investigated the aging effect of the sample with the post-deposition moisture-absorbing treatment. After the treatments and conductivity measurements, we kept the samples in a vacuum desiccator for 1 month. As shown in Fig. 6, the sample with the treatment under w relative humidity of less than 5% does not show a clear improvement of the carrier diffusion length, but the samples with the treatment under a relative humidity of 20% and 40% show the improvement of carrier diffusion length after 1 month of aging. This suggests the gradual reaction of the moisture absorbed in the CsPbBr 3 thin films or the gradual release of the excessive amount of absorbed moisture from the   CsPbBr 3 thin films. Although the post-deposition moistureabsorbing treatment and subsequent aging can improve the carrier diffusion length of thermally evaporated CsPbBr 3 thin films, this technique requires many treatments and long-term aging. Therefore, it is important to explore a way to reduce the number of treatments and the aging time in the future study.
Finally, we evaluated the CsPbBr 3 photovoltaic power converter (solar cell) with the CsPbBr 3 layer which received an 100-times post-deposition moisture-absorbing treatment under the relative humidity of 30%. As shown in Fig. 7, the CsPbBr 3 photovoltaic power converter exhibited a good I-V curve. The conversion efficiency for AM1.5 100 mW cm −2 illumination was 5.56% (Voc = 1.26 V, Jsc = 6.68 mA cm −2 , FF = 0.66). This result confirmed that device-quality CsPbBr 3 thin films can be prepared by applying the post-deposition moisture-absorbing treatment. The device performance of a device without moisture-absorbing treatment is not shown, but it is easy to understand that very poor device performance is expected according to our previous report. 6) In our previous work, we reported that the device using CsPbBr 3 thin film with a carrier diffusion length of about 0.1 μm showed very poor performance, especially with a very small Jsc less than 0.5 mA cm −2 . As shown in Fig. 3, the carrier diffusion length of the CsPbBr3 thin films without moisture-absorbing treatment is about 0.1 μm, suggesting that the device performance will be very poor if the device is fabricated using CsPbBr 3 thin films without moisture-absorbing treatment. Figure 8 shows the powder XRD patterns of evaporation sources. Both the ground CsPbBr 3 powder and the 2.5 mol% TbCl 3 -doped CsPbBr 3 powder show almost the same as that of the raw CsPbBr 3 powder. The grinding and TbCl 3 doping do not affect the crystal structure of the evaporation source. Figure 9 shows the XRD patterns of CsPbBr 3 thin films thermally evaporated using TbCl 3 -doped CsPbBr 3 powders. As shown in Fig. 9, it is clear that TbCl 3 -doping reduced the crystallinity of the CsPbBr 3 phase. We did not observe any additional phases from the CsPbBr 3 thin film deposited using 1 mol% TbCl 3 -doped CsPbBr 3 powder, but the presence of CsPb 2 Br 5 phase was clearly confirmed from the CsPbBr 3 thin film deposited using 2.5 mol% TbCl 3-doped CsPbBr 3 powder. Surface SEM and EDS mapping results shown in Fig. 10 indicate that rod-shaped CsPb 2 Br 5 phases segregated on the surface of the CsPbBr 3 thin film. The size of the rodshaped CsPb 2 Br 5 was about 2-5 μm. The small amount of TbCl 3 added to CsPbBr 3 powder does not affect the structure of CsPbBr 3 powder, but there is a strong impact on the CsPbBr 3 thin film growth. Figure 11 shows the AFM images of the samples. The grain size of the CsPbBr 3 phase was significantly increased by 300°C annealing for the CsPbBr 3 thin film deposited from the ground CsPbBr 3 powder. The grain size of the as-deposited TbCl 3 -doped CsPbBr 3 thin films was larger than that of the undoped CsPbBr 3 thin film, however, the improvement of the grain size after annealing was not observed. As a result, the grain size of the TbCl 3 -doped CsPbBr 3 thin films after the annealing was smaller than that of the annealed undoped CsPbBr 3 thin film. TbCl 3 -doping of CsPbBr 3 thin films deposited by solutionbased technique demonstrated the improvement of the grain size, but TbCl 3 -doping is not effective to increase the grain size of deposited CsPbBr 3 thin films.      Figure 12 shows the absorption and photoluminescence spectra of the TbCl 3 -doped CsPbBr 3 thin film and photographs of the doped and undoped thin films. As shown in the photographs of the deposited CsPbBr 3 thin films, the TbCl 3 -doping does not affect the optical absorption of the films, however, the photoluminescence properties were strongly affected by the TbCl 3 -doping. Strong green photoluminescence was observed only from the TbCl 3 -doped film. The emission wavelength is very similar to the bandgap of CsPbBr 3 and the wavelength is different from the emission of Tb. Therefore, the strong green photoluminescence indicates the improvement of the film quality. Figure 13 shows the aging effect of TbCl 3 -doped CsPbBr 3 thin films on the photo-and dark-conductivity. The TbCl 3 -doped CsPbBr 3 thin film deposited using 2.5 mol% TbCl 3 -doped CsPbBr 3 powder showed high dark conductivity and significant reduction of both photo-and dark-conductivity after aging. This high dark conductivity before aging suggests the existence of a thin TbCl 3 layer on the surface. Our dark conductivity measurements of TbCl 3 thin films demonstrated that thermally evaporated TbCl 3 has a high dark-conductivity of about 10 −3 S cm −1 . If a 3 nm-thick TbCl 3 exists on the surface of the CsPbBr 3 thin film, the apparent conductivity of the structure of 3 nm-thick TbCl 3 /250 nm-thick CsPbBr 3 is expected to be about 10 −5 S cm −1 . This value is very similar to the dark conductivity of the TbCl 3 -doped CsPbBr 3 thin film deposited from 2.5 mol% TbCl 3 -doped CsPbBr 3 powder before aging. The mp of TbCl 3 is slightly higher than that of CsPbBr 3 , therefore, TbCl 3 evaporation mainly occurs at the final stage of the evaporation of a mixture of CsPbBr 3 and TbCl 3 . This produces the thin surface TbCl 3 layer. The TbCl 3 -doped CsPbBr 3 thin film deposited using 1 mol%  TbCl 3 -doped CsPbBr 3 powder showed a completely different behavior. Before the aging, the photo-conductivity was slightly smaller than that of the undoped CsPbBr 3 thin film, but the aging significantly increased the photo-conductivity. After 14 d of aging, the photo-conductivity was higher than that of the undoped CsPbBr 3 thin film. After 56 d of aging, the best CsPbBr 3 thin film deposited from 1 mol% TbCl 3 -doped CsPbBr 3 powder showed high photo-conductivity above 10 −4 S cm −1 and a large carrier diffusion length of 2.3 μm. This value is very promising because the required thickness of CsPbBr 3 solar cells and photovoltaic power converters is about 600 nm. The mechanism of the improvement of photo-conductivity and carrier diffusion length by TbCl 3 -doping is not clear at present. A possible explanation is that TbCl 3 in or on the CsPbBr 3 thin film enhances the absorption of moisture in the air before the samples are placed in the vacuum desiccator since TbCl 3 has a strong hygroscopic nature. As explained in the previous section, moisture absorption in the CsPbBr 3 thin film improves the carrier diffusion length. Therefore, it is highly probable that the enhancement of moisture absorption by TbCl 3 addition leads to the improvement of carrier diffusion length of CsPbBr 3 thin films. In addition, there is a possibility that Cl in TbCl 3 affects the improvement of the electrical properties. It is well known that the electronic properties of CH 3 NH 3 PbI 3 are improved by a small amount of Cl-doping. 29) The presence of Cl on the surface of CH 3 NH 3 PbI 3 and the increase of the bandgap at the surface have also been reported. These reports suggest that the grain boundary is improved by Cl-doping. In order to clarify the reason for the improvement of CsPbBr 3 thin film by TbCl 3 doping, it is necessary to perform a more detailed analysis of the grain boundary of our samples. 30) Figure 14 shows the I-V curves of the CsPbBr 3 photovoltaic power converter (solar cell) with the TbCl 3 -doped CsPbBr 3 layer deposited from 1 mol% TbCl 3 -doped CsPbBr 3 powder. The device with the TbCl 3 -doped CsPbBr 3 without aging shows very poor performance, but aging for 14 d can improve the device performance. This indicates that the improvement of the carrier diffusion length of the CsPbBr 3 thin film leads to the improvement of the CsPbBr 3 photovoltaic power converter. However, the device performance, especially Jsc, is very small. We measured the external quantum efficiency of this device and found that the external quantum efficiency is low for all wavelength regions. This implied that the quality of the CsPbBr 3 bulk layer is sufficient, but there is a carrier transport barrier at the interface of the nc-GaN/CsPbBr 3 interface or CsPbBr 3 /P3HT interface. The interface of nc-GaN/CsPbBr 3 is the same as that of the device explained in the previous section (Fig. 7). Therefore, the poor Jsc and external quantum efficiency are due to the CsPbBr 3 /P3HT interface. As mentioned in the previous paragraph, we pointed out that the existence of a thin TbCl 3 layer on the CsPbBr 3 layer is highly probable. The TbCl 3 layer is probably one of the limiting factors of the Jsc of the devices.

Conclusions
We investigated the post-deposition moisture-absorbing treatment and TbCl 3 -doping into CsPbBr 3 thin films in order to improve the optoelectronic properties of CsPbBr 3 thin films. Both techniques showed an improvement in the carrier diffusion length of CsPbBr 3 thin films. The TbCl 3 -doping is more promising since a carrier diffusion length of 2.3 μm was achieved for the CsPbBr3 thin film deposited from 1 mol% TbCl 3 -doped CsPbBr 3 powder. However, we need to apply long-term aging (56 d) to obtain such a large carrier diffusion length. Further investigation is required to obtain CsPbBr 3 thin films with large carrier diffusion lengths by using short-term processes.