The optoelectronic properties improvement of double perovskites Cs2SnI6 by anionic doping (F−)

Tin-based perovskite material is the best choice to replace heavy metal element lead during the last several years. Cs2SnI6 with Sn4+ is a fascinating optoelectronic material, which is a more air-stable composite cesium tin halide peroxide variant from CsSnI3. However, the optoelectronic performance between N and P type of Cs2SnI6 varies considerably. Herein, we synthesized uniform Cs2SnI6 by modified two-step method, which thermal evaporated CsI firstly, and followed annealing in the SnI4 and I2 vapor at 150 °C resulted in uniform Cs2SnI6 films. SnF4 is used as a dopant source to improve the optoelectronic properties of Cs2SnI6 films. Results indicate that good crystallinity was obtained for all films and the doped films underwent a crystalline plane meritocracy transition. The doped films had a flat, non-porous morphology with large grains. The high transmittance of the doped films in the infrared region led to the avoidance of self-generated thermal decomposition. With the help of F−, the films became more conductive and had higher carrier mobility. DFT calculations showed that doping with F reduced the surface energy of (004), resulted in a preferred orientation transition in the crystal of Cs2SnI6. Fluorine doped double layer perovskite materials would have a broader application prospect.

www.nature.com/scientificreports/ Herein, we synthesized Cs 2 SnI 6 film by evaporating CsI firstly and annealing @150 °C for 1 h in SnI 4 (SnF 4 ) and I 2 environment in a quartz tube furnace, which modified from Mitzi's group 17 . This method enables the preparation of perovskites at low temperatures under atmospheric conditions, providing a new idea for the synthesis of stable and efficient perovskites. The structure and properties of Cs 2 SnI 6 thin films with the amount of doping were observed, measured, and analyzed the influence of F − doping. Meanwhile, to understand the underlying mechanism of performance change, we studied crystalline Cs 2 SnI 6 over the amount of doped via density functional theory (DFT).

Results and discussion
The phase structures of Cs 2 SnI 6 films with F doped were characterized by XRD, and the results are shown in Fig. 1. Specifically, according to PDF #051-0466, the three strongest two-theta peaks located at 26.4731°, 30.6832°, and 43.9581°, corresponding to the (222), (004), and (044) diffraction planes of of Cs 2 SnI 6 , respectively 17 . Meanwhile, the well-matched peak of 27.5982°, assigned to CsI phase (110) plane (PDF #006-0311), was observed in Cs 2 SnI 6 and Cs 2 SnI 9/2 F 3/2 films. As predicted, the peak at 23.9748° was attributed to Cs 2 SnF 6 phase (011) plane (PDF #070-0141) in F doped films. However, from the intensity of diffraction peaks, the main phase composition of the films was Cs 2 SnI 6 .
It is worth mentioning that the preferred orientation plane of Cs 2 SnI 6 phase changed from (222) to (004) after doping, as shown in Fig. 1a. It can be seen from the local enlarged patterns (Fig. 1b, c) that as the doping amount increased, the intensity of the (222) diffraction peak decreased, and the FWHM increased. However, the intensity and FWHM of the (004) diffraction peak both gradually increased. In order to explain the differences of Cs 2 SnI 6 (222) and (004) peaks on different F − contents in this study, the diffraction peak offsets, FWHM, interplanar spacing offsets of the (222) and (004) peaks, the lattice constant and the volume of unit crystal on different F − contents are compared and listed in Table 1. It can be seen from the periodic table that the ionic radius of F − is smaller than that of I − . When F − is doped into the Cs 2 SnI 6 unit cell, the lattice constant is compressed. It can confirm this view by calculating the results of the volume. These changes are in full compliance with Prague's law. It further proved that F − was successfully doped into Cs 2 SnI 6 . It is undeniable that after a large amount of F − doping, a new phase was formed, which Cs 2 SnI 6 was alloyed. Regarding the preferred orientation transition, it may be related to the change of the crystal plane energy and surface energy by the doping of F − . Surface/interfacial energy is the main factor influencing the formation and evolution of the film weave 18 . The higher the crystal planes' energy, the quicker the atomic accumulation rate, and then the faster the growth rate in the direction perpendicular to the crystal plane. There are two consequences of this: (1) The crystal grows rapidly along the direction perpendicular to the crystal plane; (2) The crystal plane disappears during the growth process.  www.nature.com/scientificreports/ Therefore, we have performed first-principles calculations on the (222) and (004) crystal planes of Cs 2 SnI 6 , and we will discuss this issue in more detail in the part of calculation.
To explore the surface chemistry, the pristine and doped films were studied by X-ray photoelectron spectroscopy (XPS), which were estimated by curve fitting of the Cs 3d, Sn 3d, I 3d and F 1s spectra. The XPS spectra of these species are shown in Fig S1. These peaks were assigned according to the reference database 19 . All the curve-fitting analysis of XPS peaks for Cs, Sn, I and F demonstrate very identical spectra indicating the absence of multiple states of constituent elements. The peaks with binding energy of 723.22 and 737.14 eV can be attributed to the Cs 3d 5/2 and Cs 3d 3/2 , respectively. It can be clearly seen that the spectra can be divided into distinct two peaks which can be attributed to I 3d 5/2 (617.85 eV) and I 3d 3/2 (629.33 eV), respectively. The XPS spectra for Sn showed a well symmetric characteristic peak at ~ 485.83 and 494.25 eV for Sn 3d that correspond to Sn 4+ species. This result is consistent with earlier reports 20 . For doped films, the characteristic peaks at 684.42 eV ( Fig  S1d) were assigned for F 1 s that suggests the Fstate. The results of XPS indicate that these elements exist in the doped films as Cs + , Sn 4+ , I − and F − .
The surface morphological structures of the Cs 2 SnI 6 films with F doped are displayed in Fig. 2 and inset showed the cross-sectional image of respective films. The morphology of the Cs 2 SnI 6 films on glass substrates were relatively flattening. However, holes were present in the films and decreased with increasing doping F − , eventually becoming almost invisible in Cs 2 SnI 3 F 3 . The appearance of the holes is due to the expansion of the crystal lattice during the chemical reaction between cesium iodide (CsI) and tin iodide (SnI 4 ), releasing a lot of stress. This phenomenon is consistent with the results of Byungho Lee's experiments 21 . The grain size of the Cs 2 SnI 6 films also became larger as doped F − increased, see 50KX, most likely due to the aggregation of small grains. In previous works, the surface coverage of FASnI 3 can be improved by adding F − to reduce the pinhole and void forms present in pure perovskites films 22,23 and even to obtain larger crystal sizes 24 . Furthermore, to confirm the elemental distribution, we also performed the elemental mapping for pristine and the doped films ( Fig S2). It is found that Cs + , Sn 4+ , Iand Fcations are evenly distributed in all the doped films. This further shows that the doping of F helps to obtain a film with a flat surface, no holes, and larger grains.
To investigate the optical properties of these Cs 2 SnI 6 films, we performed the UV-VIS-NIR transmittance spectrum measurements within the range of 500-1100 nm, as shown in Fig. 3a. Obviously, there is a significant difference between the doped films and the undoped film. The intrinsic film absorbed almost all within the test range. However, Fdoped films had higher transmittance in the infrared band. This effectively avoids the problem of self-decomposition of the perovskite material during application due to thermal effect of infrared radiation. The excellent optical properties of Cs 2 SnI 6 films made up for this issue. The bandgaps of Cs 2 SnI 6 films with doped F − were also calculated by the UV-VIS-NIR transmittance spectrum and the plot of (αhv) versus photon energy (hv) 25 , as shown in Fig. 3b. The bandgaps of all the Cs 2 SnI 6 films over doped Fwere 1.36 eV, 1.56 eV and 1.47 eV, separately. It is reported that Cs 2 SnI 6 had a direct optical gap of 1.25-1.62 eV 7,17,26,27 . The results of the optical bandgap are similar to those of previous studies. It can be known from the DFT calculations of predecessors that the bandgap of Cs 2 SnF 6 was 5.337 eV 28 , which is much larger than 1.3 eV of Cs 2 SnI 6 29 . The results showed that the metal-halogen bond angle has an important influence on the band gap, which the bandgap decreases with the increase of the bond angle 30 . The minimum bond angle of F-Sn-F measured from the unit crystal of Cs 2 SnF 6 was 87.591°, while that of I-Sn-I in Cs 2 SnI 6 was 90°. F − doping may induce increase of bandgap. What is surprising is that the band gap of Cs 2 SnI 3 F 3 had become smaller. For semiconductors with the same crystal structure, small  www.nature.com/scientificreports/ lattice constant means small interatomic distance, and hence, a strong electrostatic attraction. The band-gap represents the energy needed for bond-breaking, which reflects the strength of the attractive force. So, a direct consequence of decreasing the lattice constant is the increase in the energy gap. This is supported by the lattice constant in XRD (Table 1), which the lattice constant of Cs 2 SnI 9/2 F 3/2 was smaller than that of Cs 2 SnI 3 F 3 . It is thus clear that F − can adjust the bandgap of Cs 2 SnI 6 . In the future, by controlling the doping ratio of I-F, continuous modulation of the absorption of perovskite materials can be achieved. The electrical properties and the thickness of the Cs 2 SnI 6 films with doping F − were shown in Table 2. All films were P-type semiconductor properties. In Table 2, it's also worth noting that the resistivities of Cs 2 SnI 6 films with doped were an order of magnitude smaller than that of undoped film. And then the undoped Cs 2 SnI 6 film exhibited the carrier densities of ~ 10 15 cm −3 , and the carrier density of Cs 2 SnI 6 films remained almost the same order of magnitude with the increase of SnF 4 content. At the same time, the carrier mobility of doped films is an order of magnitude greater than that of the undoped. The carrier concentrations were not significantly varied, while the carrier mobility of Cs 2 SnI 6 for doped F exhibited the large varied, compared with Byungho Lee's experiments about ≈5 × 10 16 cm −3 and 4-8 cm 2 v −1 s −121 . These changes can be analyzed from the micro and macro perspectives, respectively. On one hand, F ions have a distortion effect on the crystal structure. The electronegativity of F is the largest in the periodic table. Therefore, more electrons were taken from Sn via F than I, the bond length of F-Sn was shorter, and the bond energy was higher 6 . On the other hand, F ions may promote crystal growth, and close to fewer defects. The lower porosity without visible crack formatted in the doped films from the SEM images (Fig. 2). In the crystal, the fewer defects (such as voids, grain boundaries, impurities, etc.), the less hindrance to the carrier, So the higher mobility is displayed in the films. In a word, the integrity of the lattice structure is poor, and there are many defects. The scattering and trapping of carriers are greatly enhanced. The carrier concentration and mobility are low, and the resistivity of the film is increased. The beneficial increase brought by F ions is the improvement of electrical properties.
The crystal structure was optimized to calculate the (222) and (004) crystal plane formation energy for Cs 2 SnI 6 . The crystalline surface formation energy 31 is calculated as follows: where E f is the formation energy of a crystal plane, E slab and E bluk are the total energies of the surface slab and the bulk unit cell, respectively, N is the number of atoms in a unit cell, 2 is forming two news along the Z-axis in the surface slab, and A is the surface area in the surface slab.
The calculated results are illustrated in Table 3. The lower the formation energy, the easier and more stable it is to form. There is a large variation in the formation energy after doping, especially in the (004) crystal plane. It is noteworthy that the formation energy increment of (004) is much larger than that of (222), which is 0.057 eV  www.nature.com/scientificreports/ and − 0.178 eV, respectively. It can be seen that the growth rate of (004) crystal plane increased rapidly when doped with 1 F − , while the growth rate of (222) crystal plane slowed down. Then, at 25% or 50% doping, a change in crystal orientation is a natural consequence. The calculation results also prove the changing trend of the morphology of the thin films. After doping F, the pores of the films decreased, and the grains became larger.

Conclusion
In conclusion, we reported the double perovskite Cs 2 SnI 6 film was successful prepared through a modified two-step method. Meanwhile, the effect of doped F − on the optoelectronic properties of Cs 2 SnI 6 films was also investigated. All the films revealed good crystallization as revealed by XRD. In particular, the (004) crystal plane is used as the preferred orientation of the F-doped films. In terms of film morphology, as the amount of doping increased, the surface tended to be flat and free of holes, and larger grain sizes were obtained. It is observed that the film analysis by XPS and EDX mapping indicates uniform composition in the film surface. Moreover, all films exhibited excellent light absorption in the visible range and the films with doped F − increased transmission in the infrared region, which effectively avoids thermal-effect of infrared radiation. And then, the electrical properties of Cs 2 SnI 6 have also been optimized. The resistivity of the films was reduced, while the carrier mobility was enhanced. These results may provide useful assistance for the application of Cs 2 SnI 6 on HTM or LAM in solar cells. First-principles calculations confirm that doping with F changed the formation energy of the crystal plane. In the future, by controlling the doping ratio of I-F, continuous modulation of the absorption of perovskite materials, optimizing electrical performance, structure and morphology can be achieved. Therefore, we submit that doping F − will be a sustainable study for future tin-based perovskite studies.

Experimental details
Materials. 99 Synthesis of Cs 2 SnI 6 . The process of sample preparation was shown in Fig. 4. The glass slides were washed in acetone, ethanol, and deionized water for 15 min. The cleaning glass slide was transferred into a DM-450C vacuum system for CsI layer evaporation at a pressure of 3 × 10 −3 Pa at room temperature. CsI powder was placed onto a tungsten (W) boat of size of 50 mm × 15 mm × 2 mm, which was 160 mm below the substrate. And then CsI films were placed in a quartz tube furnace with SnI 4 and I 2 powder in an alumina boat and annealed at 150 °C for 60 min (ramp rate: 4 °C/min) with Ar flow. After the reaction, the white translucent CsI films turned to mirror-like black Cs 2 SnI 6 films. The experimental doping amount refers to the ratio of cell atoms, such as, F at% = 0, 25% and 50%, refer to Table 4.
Characterizations of Cs 2 SnI 6 . Structure and morphology characterization. X-ray diffraction (XRD) pattern were obtained by a X'Pert powder X-ray diffractometer in glancing angle scanning mode at 0.5° incident angle with Cu Ka X-ray (λ = 1.5418 Å) from 10° to 90° and analyzed by High-Score software. The top-view FESEM images of Cs 2 SnI 6 films were observed by SIGMA HD Field Emission Scanning Electron Microscopy (FESEM), which was operated at 15 kV. www.nature.com/scientificreports/ Optoelectronic characterization. The thicknesses of the films were measured by using a KLA Tencor Alpha-step D-100 type profilometer on a step created on the films by masking the glass substrates. The electrical properties of the films were measured by a HALL 8800 Hall Effect Measurement device. The samples were cut into 10 × 10 mm 2 samples in a square shape. The measurement was carried out in four-point mode with gold electrodes by DC voltage in a 4000-gauss magnetic field at room temperature. A CARY 5000 UV-Vis-NIR spectrometer was used to measure the transmittance of the films over the wavelength range of 300 to 1500 nm.
First-principles calculations. Calculations by the CASTEP module in Materials Studio were used to optimize the structure of Cs 2 SnI 6 and Cs 2 SnI 6 with F doped. After the structure was optimized, the (222) and (004) crystal plane formation energy of Cs 2 SnI 6 phase were calculated. Then the effect of doped F − on the change of preferred orientation is analyzed. The Cs 2 SnI 6 phase belongs to the room temperature phase stable structure, and its ground state threedimensional model is a cubic crystal system. The structure of crystalline Cs 2 SnI 6 (6 × 6 × 6) was obtained from the Materials Project database, with code number 27636 32 ; the crystalline Cs 2 SnI 6 had a FM-3 M space group symmetry. Combined with the doping ratio, a super cell with 18 atoms is established, and the valence electron configurations were Cs 6s 1 , Sn 5s 2 5p 2 , I 5s 2 5p 5 , and F 2p 5 . The super cell model is shown in Fig. 5a, and the F atom substitution doping model is shown in Fig. 5b. The (222) and (004) crystal plane models of Cs 2 SnI 6 are shown in Fig. 5c,d, respectively.