Chalcogenide Perovskite Thin Films with Controlled Phases for Optoelectronics

Zhonghai Yu, Haolei Hui, Damien West, Han Zhang, Yiyang Sun, Sen Kong, Yin Zhang, Chenhua Deng, Sen Yang,* Shengbai Zhang, and Hao Zeng*

Chalcogenide perovskites have emerged as promising semiconductor materials due to their appealing properties, including tunable bandgaps, high absorption coefifcients, reasonable carrier lifetimes and mobilities, excellent chemical stability, and environmentally benign nature. However, beyond the well-studied $\mathtt{B a Z r S}{3}$ , reports on chalcogenide perovskite thin flims with diverse compositions are scarce. In this study, the realization of four different types of chalcogenide perovskite thin flims with controlled phases, through $\mathbf{cs}{2}$ annealing of amorphous chalcogenide precursor flims deposited by pulsed laser deposition (PLD), is reported. This achievement is guided by a thorough theoretical investigation of the phase stability of chalcogenide perovskites. Upon crystallization in the distorted perovskite phase, all materials exhibit photoluminescence (PL) with peak positions in the visible range, consistent with their expected bandgap values. However, the full-width-at-half-maximum (FWHM) of the PL spectra varies significantly across these materials, ranging from $99,\mathsf{m e V}$ for $\mathsf{s r H f S}{3}$ to 231 meV for $\mathsf{B a H f S}{3}$ . The difference is attributed to the difference in kinetic barriers between local structural motifs for the Sr and Ba compounds. The findings underscore the promise of chalcogenide perovskite thin flims as an alternative to traditional halide perovskites for optoelectronic applications, while highlighting the challenges in optimizing their synthesis and performance.

absorbers in solar cells.[1–9] However, their limited stability and toxicity have raised concerns regarding their long-term viability.[10] To address these limitations, researchers have been exploring alternative materials, and chalcogenide perovskites have emerged as promising candidates due to their attractive properties.[11–23] Chalcogenide perovskites have tunable bandgaps,[11,12,20,22] exceptionally high absorption coefficients, reasonably long carrier lifetimes,[12] and carrier mobilities comparable to those of halide perovskites,[15] making them attractive for optoelectronic applications such as solar cells, LEDs, and photodetectors. Unlike their halide counterparts, they also have excellent chemical stability and are environmentally benign,[23] which are important factors for commercial applications. Therefore, chalcogenide perovskites have attracted significant attention from academic researchers in recent years.

Hybrid organic-inorganic halide perovskites $\mathrm{ABX}_{3}$ , comprising an organic cation (A), a metal cation (B), and a halide anion (X), have demonstrated excellent efficiency when used as light

Chalcogenide perovskites are materials that have a crystal structure similar to the mineral perovskite but contain chalcogenide elements such as sulfur, selenium,or tellurium. The bandgaps of

1. Introduction

chalcogenide perovskites are in the visible regime, and can be easily tuned by varying the composition of the materials, allowing them to absorb a broad range of wavelengths of light. This tunability is crucial for developing efficient solar cells that can absorb a large portion of the solar spectrum. Among the most studied chalcogenide perovskites, $\mathrm{BaZrS}{3}$ has a bandgap of 1.8- $1.9,\mathrm{eV}$ and $\mathrm{SrHfS}{3}$ has a bandgap of $2.4–2.5\ \mathrm{eV.^{[11,19]}}$ By anion or cation alloying, the range of bandgaps can be extended to $1.5-$ $3.0\ \mathrm{eV.}^{[21,24]}$ Most recently, by stabilizing the perovskite phase of $\mathrm{BaZrSe}{3}$ using molecular beam epitaxy on $\mathrm{BaZrS}{3}$ template, bandgap of $1.4,\mathrm{eV}$ has been obtained, which is close to the optimal bandgap for a single junction solar cell.[25,26]

Another important property of chalcogenide perovskites is their exceptionally high absorption coefficient. We reported near band edge absorption coefficient $\alpha$ of $\mathrm{BaZrS}{3}$ thin films above $10^{5},\ c m^{-1}$ at ${\approx}1.9,\mathrm{\eV.^{[12]}~\ B a Z r S{3}}$ , $\mathrm{Sr}Z\mathrm{r}\mathrm{S}{3}$ , $\mathrm{BaHfS}{3}$ , and ${\mathrm{SrHfS}}{3}$ bulk polycrystalline samples all exhibit $\alpha$ values exceeding $10^{5},\mathrm{cm}^{-1}$ near the bandgaps, which is an order of magnitude higher than existing solar cell absorber materials including halide perovskites.[20,27] The high absorption coefficients can be attributed to the large density of states in the conduction band, which mainly consists of $d$ -states of the B cations such as $Z\mathbf{r}$ and Hf, as opposed to the $s\mathrm{-}$ states in conventional semiconductors. In addition, chalcogenide perovskites such as $\mathrm{BaZrS}{3}$ have reasonably long carrier lifetimes, reaching ${\approx}50$ ns as measured by time-resolved PL. The Hall mobility of $\mathrm{BaZrS}{3}$ thin films reaches $13.7,\mathrm{cm}^{2},\mathrm{V}^{1},\mathrm{s}^{.1}$ .[12] The field effect transistor (FET) mobilities in devices fabricated by low temperature processed $\mathtt{B a Z r S}{x}$ films are $16.8,\mathrm{cm}^{2}\ \mathrm{V}^{1}\ \mathrm{s}^{.1}$ and $2.6,\mathrm{cm}^{2}\ \mathrm{V}^{1}\ \mathrm{s}^{-1}$ for electrons and holes, respectively.[15] The efficient light absorption and reasonably good carrier transport in chalcogenide perovskite thin films implies high current densities in solar cells if they are used as light absorbers.[12,15,18,28]

The strong light-matter interactions in chalcogenide perovskites can further lead to strong band edge emission. Strong photoluminescence (PL) of $\beta{\cdot}{\mathrm{SrZrS}}{3}$ powder was reported with a full-width-at-half-maximum (FWMH) of $122;\mathrm{meV},$ whose intensity is comparable to single crystals of CdSe and InP.[20] $\mathrm{SrHfS}{3}$ powder also exhibits intense green PL, which is observable by naked eyes at room temperature.[19] The strong light emission and more balanced electron and hole transport are beneficial for high brightness LEDs. In addition, the presence of chalcogen elements in the crystal structure increases covalency and improves the stability of these materials in ambient conditions, making them more suitable for long-term use.

While chalcogenide perovskites have been synthesized in various forms such as powder,[19,22,20] single crystals,[13,14] and more recently nanoparticles,[29] it is crucial to realize thin films for a better understanding of carrier transport behaviors and for device applications. In 2020, our group first reported $\mathrm{BaZrS}{3}$ thin films synthesized by $\mathrm{C}\ensuremath{\mathrm{S}}{2}$ sulfurization of $\mathrm{BaZrO}{3}$ precursor layers deposited by pulsed laser deposition (PLD) at temperatures above $900,^{\circ}\mathrm{C}$ .[12] We later discovered that by changing the reaction pathways to $\mathrm{C}\mathsf{S}{2}$ annealing of PLD deposited amorphous $\mathtt{B a Z r S}{x}$ precursor films, the phase formation temperature can be lowered to $550;^{\circ}\mathrm{C}$ without inducing secondary phases.[15] Other research groups also reported the synthesis of $\mathrm{BaZrS}{3}$ thin films using various methods such as sulfurizing PLD, sputter, and solution deposited precursor films.[13,15,17,18,30,31] We note that there has been an earlier report on the sputter deposition of $\mathrm{LaYS}{3}$ thin films, a sulfide perovskite that has received less attention. However, this material contains rare-earth elements and is an indirect gap semiconductor with low optical absorption.[32,33] Epitaxial thin films of $\mathrm{BaZrS}{3}$ were realized on single crystal substrates using molecular beam epitaxy (MBE) and PLD techniques in 2021, which marked a major advance in the field.[13,14]

So far, most thin film works focus on $\mathrm{BaZrS}{3}$ , and other types of chalcogenide perovskites are rarely reported.[31,34] Furthermore, while the distorted perovskite phase exhibits promising optoelectronic properties, it is known that these materials can crystalize into competing phases with similar energies, such as the needle-like and hexagonal phases.[11] Thus, to realize phasepure chalcogenide perovskite thin films with desired phases, a deeper understanding of the thermodynamics and kinetics of the phase formation is a necessity.[35] In this work, we first theoretically investigate the phase stability of different types of chalcogenide perovskites, focusing on sulfide perovskites. Our results show that the distorted perovskite phase is stable at all temperatures for the Ba compounds $(\mathrm{BaZrS}{3}$ and $\mathrm{BaHfS}{3.}$ . However, for the $\mathrm{Sr}$ compounds $(\mathrm{Sr}Z\mathrm{r}S{3}$ and $\mathrm{SrHfS}{3}$ ), the distorted perovskite is stable at high temperatures, while at low temperatures, the stable phase transitions to the needle-like phase. Based on such understanding, we identified experimental conditions and realized four different kinds of chalcogenide perovskite thin films, namely $\mathrm{BaZrS}{3}$ , $\mathrm{BaHfS}{3}$ , $\mathrm{Sr}Z\mathrm{r}S{3}$ , and $\mathrm{SrHfS}{3}$ , in the desired distorted perovskite structure. All perovskite films exhibit PL with peak positions in the visible regime commensurate with their expected bandgap values. However, the FWHM of the PL spectra varies substantially among different materials, ranging from 99 meV for $\mathrm{SrHfS}{3}$ to more than $200\ \mathrm{meV}$ for $\mathrm{BaHfS}_{3}$ .[12] This is explained by the much smaller energy barriers between different local structural motifs of the Ba compounds compared to the Sr compounds.

2. Results and Discussion

The first-principles calculation was based on the density functional theory (DFT) as implemented in the VASP program.[36] Projector augmented wave (PAW) potentials[37] were used to describe the core-valence interaction. Plane waves with kinetic energy of $500\ \mathrm{\eV}$ were used as the basis set. The strongly constrained and appropriately normed (SCAN) metaGGA functional[38,39] was adopted for the exchange-correlation functional. In calculating the enthalpy of formation of $\mathrm{ABS}{3}$ with respect to the binary AS and ${\mathrm{BS}}{2}$ compounds, we further employed the rVV10 van der Waals functional[40] to better describe the layered compounds $Z\mathbf{r}\ensuremath{\mathsf{S}}{2}$ and $\mathrm{HfS}{2}$ . The Gibbs free energy including the zero point energy, vibration entropy and temperaturedependent enthalpy was calculated by using the phonon density of states calculated by the Phonopy package.[41]

The calculated $0,\mathrm{K}$ enthalpy of formation $(\Delta H_{\mathrm{f}})$ for Ba, Sr, and Ca compounds in the needle-like and distorted perovskite phases are shown in Table 1. Here it can be seen that Ba and Sr compounds are stable against phase separation into their respective binary compounds $(\Delta H_{\mathrm{f}}<0)$ . Interestingly, as seen from Table 1, the distorted perovskite phase is more stable for Ba compounds, while the needle-like phase is more stable for Sr compounds. However, Ca compounds are thermodynamically unstable $(\Delta H_{\mathrm{f}}>0)$ , suggesting that they would spontaneously decompose into binary compounds. This is especially true for ${\mathrm{CaHfS}}{3}$ , which has the highest enthalpy of formation, and may be difficult to form experimentally. The relative stability of these compounds is by-and-large dependent on the degree of lattice distortion (in particular the rotation and distortion of the octahedrons) and can be intuitively understood by the Goldschmidt tolerance factor,[42] $t~=(r{\mathrm{{A}}}~+r_{\mathrm{{C}}})/\sqrt{2}(r_{\mathrm{{B}}}+r_{\mathrm{{C}}})$ . In the case of chalcogenide perovskites, the difference in $t$ is almost entirely due to the difference in the size of the A cations (shown in Table S1, Supporting Information). As Ba has the largest ionic radius, the Ba series has the largest $t$ (lowest distortion), making the distorted perovskite phase more stable than the needle-like phase. The Sr series has a smaller t and hence a larger distortion, making the needle-like phase more stable. Ca has the smallest $t$ and largest distortion, making even the needle-like phase unstable compared to the binary phases.

Table 1. The 0 K enthalpy of formation for chalcogenide perovskites in the needle-like and distorted perovskite phases.

ABS3	Phase	Hf[eV/ABS3]
CaZrS3	Needle-like	0.122
CaZrS3	Perovskite	0.088
CaHfS3	Needle-like	0.139
CaHfS3	Perovskite	0.095
SrZrS3	Needle-like	-0.216
SrZrS3	Perovskite	-0.167
SrHfS3	Needle-like	-0.202
SrHfS3	Perovskite	-0.141
BaZrS3	Needle-like	-0.506
BaZrS3	Perovskite	-0.513
BaHfS3	Needle-like	-0.460
BaHfS3	Perovskite	-0.478

The tolerance factor is a crucial determinant of the stability of different types of chalcogenide perovskites. However, it is important to consider other factors that may also influence the relative stability between the distorted perovskite and the needlelike phases. Among these factors are the frequencies of phonon modes associated with the A-site cation, which affect the vibrational entropy and hence the free energy. Interestingly, our investigation reveals intriguing behavior for the Sr compounds. While the needle-like phase exhibits the lowest enthalpy of formation, the inclusion of phonon contributions leads to the distorted perovskite phase (orange curves) becoming more stable (i.e., having lower free energy) at high temperatures, as depicted in the temperature dependent free energy plots in Figure 1a,b. That is, there exists a crossover between the two phases at some critical temperature. This can be seen more clearly when the energy difference, $\Delta G\equiv G(\mathrm{perovskite})--\ G(\mathrm{need}$ le −like), is plotted in Figure 1e,f, for $\mathrm{Sr}Z\mathrm{r}\mathrm{S}{3}$ and $\mathrm{SrHfS}{3}$ , respectively (results for $\mathrm{BaZrS}{3}$ and $\mathrm{BaHfS}{3}$ are shown in Figure S1a,b, Supporting Information). For $\mathrm{Sr}Z\mathrm{r}\mathrm{S}{3}$ , $\Delta G>0,(<0)$ at $T,<,490\ K,(>490\ K)$ , which suggests that below $490~\mathrm{K}{\mathrm{i}}$ , the needle-like phase is more stable, while above that the distorted perovskite becomes more stable. $\mathrm{SrHfS}_{3}$ also shows the crossover at ${\approx}500~\mathrm{K}$ (Figure 1f). Our results suggest that to obtain phase-pure distorted perovskite phase for the $\mathrm{Sr}$ compounds, they need to be synthesized at high temperatures, with adequate cooling rate to arrest this phase at room temperature. If synthesized at lower temperatures, the Sr compounds will tend to form the needle-like phase. In contrast, for the Ba compounds, the distorted perovskite phase always exhibits lower free energy than the needle-like phase at all temperatures, as seen in Figure 1c,d, respectively.

How to understand the difference in phase stability between the Sr and Ba compounds? We note that just considering the formation enthalpy alone, needle-like phase is more stable in the Sr compound. However, the larger vibrational entropy in the distorted perovskite phase than the needle-like phase leads to the stability crossover at higher temperatures, due to the −TS term in the free energy, where S is the entropy. For the Ba compounds, on the other hand, the distorted perovskite phase is already in the ground state at $T,=,0$ . As $T$ increases, the larger entropy associated with the distorted perovskite phase just makes it even lower in energy than the needle-like phase. Thus, the distorted perovskite phase is expected to be stable at all temperatures. As one can see, symmetry and distortion play important roles in the phase stability of chalcogenide perovskites, at both low and high temperatures.

To verify our theoretical predictions, we systematically investigated the synthesis conditions for the Ba and Sr series of chalcogenide perovskite thin films, following our previously reported method of crystalizing the as-deposited chalcogenide precursor films using $\mathrm{C}\mathsf{S}{2}$ .[28] Based on our previous results, as-deposited chalcogenide precursor films were amorphous and strongly sulfur deficient, despite the PLD target being crystalline with a distorted perovskite structure. In order to crystallize the films with a stoichiometric ratio of $\mathrm{ABS}{3}$ , the as-deposited $\mathrm{ABS}{x}$ thin films were annealed in $\mathrm{C}\ensuremath{\mathrm{S}}{2}$ atmosphere at temperatures ranging from 500 to $1200\ ^{\circ}\mathrm{C}$ . We studied the structural evolution of the films as a function of annealing temperature using X-ray diffraction (XRD), Raman and energy dispersive X-ray (EDX) spectroscopies. EDX confirms the expected stoichiometry of the samples, as shown in Figure S2 (Supporting Information). We use $\mathrm{SrHfS}{3}$ as an example to illustrate the structural evolution and point out similarities and differences for other types of materials. As can be seen from the XRD patterns in Figure 2a, samples annealed at 900 and $1050,^{\circ}\mathrm{C}$ are of the distorted perovskite phase. The Miller indices of all peaks are labeled based on standard pdf-04-007- 5514 file of $\mathrm{SrHfS}{3}$ with the pnma structure, and no secondary phase can be identified within the instrument limit. However, when the annealing temperature is reduced to $800,^{\circ}\mathrm{C}$ , the needlelike phase dominates, with only a minute amount of perovskite phase present. The Rietveld refinement analysis reveals that the percentage of the needle-like phase is approximately $99.6%$ , as illustrated in Figure S3 (Supporting Information). Upon further reduction of the temperature to $700~^{\circ}\mathrm{C}$ , the sample consists entirely of the needle-like phase. This suggests that for $\mathrm{SrHfS}_{3}$ , the needle-like phase is a low-temperature phase, while the distorted perovskite phase is a high-temperature phase, consistent with our theory predictions. There is a relatively narrow temperature range between 800 and $900\ ^{\circ}\mathrm{C}$ in which two phases coexist.

A similar trend was observed for $\mathrm{Sr}Z\mathrm{r}S_{3}$ . As shown in Figure 2b, at 900 and $1050\ ^{\circ}\mathrm{C}$ , the XRD patterns match well with the distorted perovskite phase of $\mathrm{Sr}Z\mathrm{r}\mathrm{S}{3}$ . As the temperature decreases to $800\ ^{\circ}\mathrm{C}{!}$ peaks from the needle-like phase start to emerge and their intensity gets stronger at $700,^{\circ}\mathrm{C}$ , confirming that the needle-like phase is a low-temperature phase. At $700^{\circ}\mathrm{C}$ , Rietveld analysis gives $65%$ distorted perovskite and $35%$ needlelike phases, as illustrated in Figure S4 (Supporting Information).

Figure 1. The calculated Gibbs free energy as a function of temperature for a) $\mathsf{S r}Z\mathsf{r}S_{3}$ , b) $\mathsf{S r H f S}{3}$ , c) $\mathsf{B a Z r S}{3}$ and d) $\mathsf{B a H f S}{3}$ in the needle-like (blue) and distorted perovskite (orange) phases; and the energy difference between the two phases, defined as $\Delta G\equiv G$ (perovskite) −G(needle −like), as a function of temperature for e) $\mathsf{S r}Z\mathsf{r}S{3}$ and (f) $\mathsf{S r H f S}_{3}$ .

This result implies that to form $\mathrm{Sr}Z\mathrm{r}S_{3}$ thin films with pure needle-like phase, a temperature below $700,^{\circ}\mathrm{C}$ is necessary and a prolonged thermal processing is needed to improve crystallinity. The grain size of films sulfurized in this temperature range is of the order of hundreds of nm, and decreases with decreasing temperature, as evidenced by the atomic force microscope (AFM) images (Figure S5, Supporting Information). The surface roughness exhibits similar trend, decreasing from 80 to $20;\mathrm{nm}$ as the temperature decreases from 900 to $700~~^{\circ}\mathrm{C}$ At $600~~^{\circ}\mathrm{C}$ , the films are not well crystalized, with no discernible grains as seen from the AFM image.

The structure evolution as a function of sulfurization temperature for $\mathrm{SrHfS}{3}$ and $\mathrm{Sr}Z\mathrm{r}S{3}$ is further confirmed by Raman spectroscopy measurements, as shown in Figure 3. The Raman spectra show the peaks of distorted perovskite structure when the temperature is above $800\ ^{\circ}\mathrm{C}$ for both $\mathrm{SrHfS}{3}$ and $\mathrm{Sr}Z\mathrm{r}\mathrm{S}{3}$ (Figure 3). As the temperature goes below $800,^{\circ}\mathrm{C}$ , main peaks of the needle-like phase at $\approx!350,\mathrm{cm}^{-1}$ are observed. It is straightforward to understand the needle-like phases formed at lower temperatures for $\mathrm{SrHfS}{3}$ and $\mathrm{Sr}Z\mathrm{r}\mathrm{S}{3}$ thin films, as they are thermodynamically stable. However, growth at higher temperatures followed by natural cooling results in the distorted perovskite phase. This suggests that despite the needle-like phase being more stable at room temperature, once the distorted perovskite phase is formed at high temperatures, the nucleation barrier to relax to the needle-like phase is prohibitively high. Instead, the needle-like phase can only form at temperatures lower than the crossover temperature. Thus, two phases are formed in two distinct temperature regimes. The crossover temperature is found to be between 800 and $900~^{\circ}\mathrm{C}$ for $\mathrm{SrHfS}{3}$ with a rather sharp transition. For $\mathrm{Sr}Z\mathrm{r}S{3}$ , the crossover temperature is at around $700,^{\circ}\mathrm{C}.$ These results qualitatively agree with the theoretical predictions, with quantitative differences in crossover temperature arising due to the uncertainty in the calculated phonon frequencies.

In contrast, $\mathrm{BaHfS}_{3}$ shows the distorted perovskite phase for all annealing temperatures from 650 to $1050\ ^{\circ}\mathrm{C}$ , as seen in

Figure 2. a) XRD patterns of $\mathsf{S r H f S}{3}$ and b) $\mathsf{S r}Z\mathsf{r}S{3}$ thin flims sulfurized at different temperatures; the square represents the distorted perovskite phase and the triangle represents the needle-like phase.

Figure 4. In our previous publication,[28] we also found that phase-pure distorted perovskite phase was formed for $\mathrm{BaZrS}{3}$ in a wide temperature range from 550 to $1050,^{\circ}\mathrm{C}$ . With further lowering temperature to $500,^{\circ}\mathrm{C}$ , binary phases of BaS and $Z\mathbf{r}\ensuremath{\mathsf{S}}{2}$ were identified. No needle-like phase has been found in our study. This is also consistent with our theoretical prediction that the distorted perovskite phase is the ground state at all temperatures as shown in Figure 1c,d. We suggest that the needle-like phase for the Ba series might be stabilized using strain if a suitable substrate can be found.

Chalcogenide perovskites with a distorted perovskite phase possess strong light-matter interactions, differentiating this family of materials from conventional s,p semiconductors. While strong room temperature PL has been observed in $\mathrm{SrHfS}{3}$ and $\mathrm{Sr}Z\mathrm{r}\mathrm{S}{3}$ bulk powder materials, PL in thin films has only been reported in $\mathrm{BaZrS}{3}$ and $\mathrm{BaHfS}{3}$ .[30] Here we show for the first-time room temperature PL emission of other types of chalcogenide perovskite thin films. As seen in Figure 5, for all four materials with a distorted perovskite phase, strong room temperature PL has been obtained. The PL peak values are 1.8, 2.09, 2.15 and

Figure 3. Raman spectra of a) $\mathsf{S r H f S}{3}$ and b) $\mathsf{S r}Z\mathsf{r}S{3}$ thin flims sulfurized at different temperatures.

Figure 4. a) XRD patterns and b) Raman spectra of BaHfS3 thin flims sulfurized at different temperatures.

Figure 5. PL spectra of a) $\mathsf{S r H f S}{3}$ , b) $\mathsf{S r}Z\mathsf{r}S{3}$ , c) $\mathsf{B a H f S}{3}$ , and d) $\mathsf{B a Z r S}{3}$ thin flims sulfurized at $1050;^{\circ}\mathrm{C}.$ .

$2.37~\mathrm{eV}$ for $\mathrm{BaZrS}{3}$ , $\mathrm{Sr}Z\mathrm{r}S{3}$ , $\mathrm{BaHfS}{3}$ and $\mathrm{SrHfS}{3}$ , respectively. These values span the range from the red to the green spectrum, and are consistent with theoretically calculated values using HSE06 functional.[11] The optical images (Figure S6, Supporting Information) showing different colors of the films also reflect the bandgap differences in these materials. Furthermore, the FWHM values of PL spectra for $\mathrm{SrHfS}{3}$ , $\mathrm{Sr}Z\mathrm{r}\mathrm{S}{3}$ , $\mathrm{BaHfS}{3}$ and $\mathrm{BaZrS}{3}$ thin films sulfurized at $1050\ ^{\circ}\mathrm{C}$ are obtained to be 99 meV $20;\mathrm{nm})$ ), $108;\mathrm{meV}$ $\mathrm{(29;nm)}$ , $231;\mathrm{meV}$ $\mathrm{[65\nm]}$ and $200;\mathrm{meV}$ $114;\mathrm{nm})$ , respectively. Notably, ${\mathrm{SrHfS}}{3}$ and $\mathrm{Sr}Z\mathrm{r}\mathrm{S}{3}$ thin films exhibit substantially narrower PL spectra than the Ba series. This is further seen in Figure S7 (Supporting Information): while the sulfurization temperature clearly impacts the PL width, with FWHM decreasing with increasing temperature, the difference in FWHM for different materials is substantially more pronounced. These values are also narrower than previously reported values in bulk samples of $250;\mathrm{meV}$ for $\mathrm{SrHfS_{3}}^{[19]}$ and $122;\mathrm{meV}$ for $\mathrm{Sr}Z\mathrm{r}S_{3}$ ,[20] and comparable to other high-quality semiconductor films such as epitaxial GaAs and halide perovskites[43,44] with grain sizes of many microns, as well as quantum-confined nanostructures such as quantum dots and nanoplatelets, despite the polycrystalline nature of chalcogenide perovskite thin films. Such a narrow PL is beneficial for light emitting applications with high color purity, especially in the green regime, as the lower efficiency of green LEDs compared to red and blue ones is a well-known problem.

The significant difference in the PL linewidths between the Sr and Ba compounds, despite both possessing the distorted perovskite structure, presents an intriguing observation. We attribute this difference in linewidths to the tolerance factor $t,$ which is substantially larger for the Ba compounds compared to the Sr compounds. This results in a much smaller energy difference between the distorted and undistorted phases, which corresponds to the kinetic barrier between different structural motifs of local distortion, for the Ba compounds. For instance, from our calculation, the energy difference between the distorted and undistorted phases of $\mathrm{BaZrS}{3}$ is found to be $0.17,\mathrm{eV}$ per formula unit, in contrast to $0.74~\mathrm{eV}$ per formula unit for $\mathrm{Sr}Z\mathrm{r}S{3}$ . Consequently, the barriers between different structural motifs are significantly smaller in the case of the Ba compounds.

At elevated synthesis temperatures, thermal energy can facilitate the crossover of these barriers, leading to a distribution of local structural motifs with varying distortion patterns. This can be seen in molecular dynamics simulations done at $1200\mathrm{K}$ , of which snapshots of the atomic structure at regular time intervals are shown in Figure 6a,b. Molecular dynamics simulations were performed using the generalized gradient approximation of Perdew-Burke-Ernzerhof (PBE)[45] with a plane wave basis set with kinetic energy cutoff of $250\mathrm{eV}$ in a $2!\times!2!\times!2$ supercell containing 160 atoms. Simulations were performed within the NVT ensemble using a Nosé-Hoover thermostat[46] with an ionic time step of 3.0 fs and k-point sampling of a single special $\mathrm{k\Omega}$ -point at $(^{1}!/!4,,^{1}!/!4,,^{1}!/!4)$ which yielded total-energy convergence better than $0.03~\mathrm{eV}$ per supercell, as compared with supercell calculations using a Monkhorst-Pack mesh.[47] As a guide to the eye, the rotations of the $Z\mathbf{r}$ polyhedra in the distorted perovskite structure are indicated by red lines (canted to the right) and black lines (canted to the left). While $\mathrm{Sr}Z\mathrm{r}\mathrm{S}{3}$ maintains a fixed distortion network through the course of the simulation (Figure 6a), the smaller barriers of $\mathrm{BaZrS}{3}$ lead to a dynamic network in which polyhedral orientation fluctuates and locally some polyhedral orientations resemble the undistorted phase (highlighted by blue dashed lines). Upon cooling, it is expected that the dynamic network associated with the Ba compounds could lead to a distribution of local structural motifs with different distortion patterns which become locked in at room temperature.

Furthermore, as the electronic structure and bandgap are highly sensitive to the degree of distortion, with the bandgap of the undistorted phase being approximately $1,\mathrm{eV}$ smaller than the distorted perovskite phase (see Figure 6c,d), the small barrier for the Ba compounds could lead to substantial broadening of the PL spectra. Notably, local structure motifs which resemble the undistorted phase, such as those at the boundaries between two patterns of distortion (as seen in Figure 6a,b), would have a smaller bandgap and hence would preferentially contribute to the PL spectra at lower energies. This is consistent with the asymmetric broadening in our observations, as shown in Figure 5c,d. Conversely, the smaller t value for the Sr compounds and higher associated kinetic barriers, suggest the thermal energy at our synthesis conditions would be insufficient to overcome these barriers. Thus, a substantial broadening cannot occur, resulting in narrower and more symmetric PL spectra, as shown in Figure 5a,b.

3. Conclusion

In conclusion, our theoretical investigations revealed the stable phase behavior of chalcogenide perovskites, with Ba compounds maintaining the distorted perovskite phase across all temperatures, while Sr compounds transition to the needle-like phase at lower temperatures. By understanding such behavior, we successfully synthesized four different chalcogenide perovskite thin films with the desired distorted perovskite structure. These films exhibit strong PL emissions, with particularly narrow PL width of ${\approx}100\ \mathrm{meV}$ at room temperature for the $\mathrm{Sr}$ compounds. The difference in the phase stability and PL widths between the Ba and Sr compounds is closely associated with their different degrees of structural distortions and their formation energies. These findings contribute to the advancement of chalcogenide perovskite thin films for photovoltaic and optoelectronic applications.

4. Experimental Section

Preparation of the Chalcogenide Perovskite PLD Targets: Here, the $\mathsf{S r}Z\mathsf{r}S_{3}$ target was taken as an example. $\mathsf{S r}Z\mathsf{r}S_{3}$ powder was synthesized by sulfurizing $\mathsf{S r}Z\mathsf{r O}{3}$ powder in $\mathsf C\mathsf S{2}$ at $1050,^{\circ}\mathrm{C}$ for $4,\mathrm{h}$ , following the published procedure.[22] The $\mathsf{S r}Z\mathsf{r}S_{3}$ PLD target was prepared by compressing $\mathsf{S r}Z\mathsf{r}S_{3}$ powder into a $20\mathsf{m m}$ -diameter pellet, followed by sintering at $1050,^{\circ}\mathrm{C}$ for $^{2,\mathrm{h}}$ , following our published procedure.[28] Targets of $\mathsf{B a Z r S}{3}$ , $\mathsf{B a H f S}{3}$ and $\mathsf{S r H f S}_{3}$ were prepared similarly.

Fabrication of the Chalcogenide Perovskite Thin Films: A ${\mathsf{1}},{\mathsf{c m}}\times{\mathsf{1}},{\mathsf{c m}}$ (0001) sapphire substrate was loaded into the ultrahigh vacuum (UHV) deposition chamber of the PLD system equipped with a KrF pulsed excimer laser ( $!!!\lambda,=,248,\mathsf{n m}}$ , at a base pressure of $\rceil\times,70^{-8}$ torr. During the deposition, the repetition rate of the laser was $5;\mathsf{H z}$ and the fluence was 5 $|{\mathsf{c m}}^{-2}$ . The substrate and target were both rotated at $30,\mathsf{r p m}$ to ensure flim uniformity. As-deposited thin flims were moved to a tube furnace with a 2 in. diameter quartz tube for $\mathsf C\mathsf S_{2}$ sulfurization. The annealing procedure follows our published procedure,[28] with the temperature ranging from 650 to $1050,^{\circ}\mathrm{C}$ for 2 to $^{8,\mathrm{h}}$ .

Figure 6. Snapshots of the atomic structure at regular time intervals obtained from molecular dynamics simulations for a) $\mathsf{S r}Z\mathsf{r}S_{3}$ and b) $\mathsf{B a Z r S}{3}$ . The band structures and associated optical bandgaps of the distorted and undistorted c) $\mathsf{S r}Z\mathsf{r}S{3}$ (red) and d) $\mathtt{B a Z r S}_{3}$ (blue).

Thin Film Characterizations: The X-ray diffraction (XRD) $\theta{-}2\theta$ scans were performed using an X′pert Pro X-ray diffractometer operating at $7.6,\mathrm{\kW}$ $\lceil\mathsf{C u}\rceil\mathsf{K}\alpha$ radiation). The Raman spectra were obtained from a HORIBA Raman spectrometer working under $532,\mathsf{n m}$ laser excitation. The PL spectra of chalcogenide perovskite thin flims were obtained at room temperature. The laser excitation wavelengths were $325\ \mathsf{n m}$ for $\mathsf{S r H f S}{3}$ , $532;\mathsf{n m}$ for $\mathsf{S r}Z\mathsf{r}S{3}$ , $574;\mathsf{n m}$ for the $\mathsf{B a Z r S}{3}$ and $\mathsf{B a H f S}{3}$ thin flims, respectively.

Supporting Information

Supporting Information is available from the Wiley Online Library or from the author.

Acknowledgements

Z.Y. and H.H. contributed equally to this work. This work was supported by United States National Science Foundation (US NSF) ECCS-2042085, ECCS-2042126, CBET-1510121, CBET-1510948, MRI-1229208, National Natural Science Foundation of China (Grant No. 52071255, 52171190 and 52002266), the Fundamental Research Funds for the Central Universities (China), the World-Class Universities (Disciplines) and the Characteristic Development Guidance Funds for the Central Universities. H.Z. acknowledges support through the Moti Lal Rustgi Professorship.

Conflict of Interest

The authors declare no conflict of interest.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Keywords

chalcogenide perovskite, optoelectronics, phase stability, photoluminescence, PL linewidth

Received: August 11, 2023 Revised: October 10, 2023 Published online:

[1] N. J. Jeon, J. H. Noh, W. S. Yang, Y. C. Kim, S. Ryu, J. Seo, S. I. Seok, Nature 2015, 517, 476.
[2] M. Liu, M. B. Johnston, H. J. Snaith, Nature 2013, 501, 395.
[3] M. M. Lee, J. Teuscher, T. Miyasaka, T. N. Murakami, H. J. Snaith, Science 2012, 338, 643.
[4] G. Xing, N. Mathews, S. Sun, S. S. Lim, Y. M. Lam, M. Grätzel, S. Mhaisalkar, T. C. Sum, Science 2013, 342, 344.
[5] A. Kojima, K. Teshima, Y. Shirai, T. Miyasaka, J. Am. Chem. Soc. 2009, 131, 6050.
[6] M. L. Agiorgousis, Y.-Y. Sun, H. Zeng, S. Zhang, J. Am. Chem. Soc. 2014, 136, 14570. [7] J. Song, J. Li, X. Li, L. Xu, Y. Dong, H. Zeng, Adv. Mater. 2015, 27, 7162.
[8] Z.-K. Tan, R. S. Moghaddam, M. L. Lai, P. Docampo, R. Higler, F. Deschler, M. Price, A. Sadhanala, L. M. Pazos, D. Credgington, F. Hanusch, T. Bein, H. J. Snaith, R. H. Friend, Nat. Nanotechnol. 2014, 9, 687.
[9] H. Zhu, Y. Fu, F. Meng, X. Wu, Z. Gong, Q. Ding, M. V. Gustafsson, M. T. Trinh, S. Jin, X.-Y. Zhu, Nat. Mater. 2015, 14, 636.
[10] J. Berry, T. Buonassisi, D. A. Egger, G. Hodes, L. Kronik, Y.-L. Loo, I. Lubomirsky, S. R. Marder, Y. Mastai, J. S. Miller, D. B. Mitzi, Y. Paz, A. M. Rappe, I. Riess, B. Rybtchinski, O. Stafsudd, V. Stevanovic, M. F. Toney, D. Zitoun, A. Kahn, D. Ginley, D. Cahen, Adv. Mater. 2015, 27, 5102.
[11] Y.-Y. Sun, M. L. Agiorgousis, P. Zhang, S. Zhang, Nano Lett. 2015, 15, 581.
[12] X. Wei, H. Hui, C. Zhao, C. Deng, M. Han, Z. Yu, A. Sheng, P. Roy, A. Chen, J. Lin, D. F. Watson, Y.-Y. Sun, T. Thomay, S. Yang, Q. Jia, S. Zhang, H. Zeng, Nano Energy 2020, 68, 104317.
[13] I. Sadeghi, K. Ye, M. Xu, Y. Li, J. M. Lebeau, R. Jaramillo, Adv. Funct. Mater. 2021, 31, 2105563.
[14] M. Surendran, H. Chen, B. Zhao, A. S. Thind, S. Singh, T. Orvis, H. Zhao, J.-K. Han, H. Htoon, M. Kawasaki, R. Mishra, J. Ravichandran, Chem. Mater. 2021, 33, 7457.
[15] V. K. Ravi, S. H. Yu, P. K. Rajput, C. Nayak, D. Bhattacharyya, D. S. Chung, A. Nag, Nanoscale 2021, 13, 1616.
[16] T. Gupta, D. Ghoshal, A. Yoshimura, S. Basu, P. K. Chow, A. S. Lakhnot, J. Pandey, J. M. Warrender, H. Efstathiadis, A. Soni, E. Osei-Agyemang, G. Balasubramanian, S. Zhang, S.-F. Shi, T.- M. Lu, V. Meunier, N. Koratkar, Adv. Funct. Mater. 2020, 30, 2001387.
[17] C. Comparotto, A. Davydova, T. Ericson, L. Riekehr, M. V. Moro, T. Kubart, J. Scragg, ACS Appl. Energy Mater. 2020, 3, 2762.
[18] J. A. Márquez, M. Rusu, H. Hempel, I. Y. Ahmet, M. Kölbach, I. Simsek, L. Choubrac, G. Gurieva, R. Gunder, S. Schorr, T. Unold, J. Phys. Chem. Lett. 2021, 12, 2148.
[19] K. Hanzawa, S. Iimura, H. Hiramatsu, H. Hosono, J. Am. Chem. Soc. 2019, 141, 5343.
[20] S. Niu, H. Huyan, Y. Liu, M. Yeung, K. Ye, L. Blankemeier, T. Orvis, D. Sarkar, D. J. Singh, R. Kapadia, J. Ravichandran, Adv. Mater. 2017, 29, 1604733.
[21] W. Meng, B. Saparov, F. Hong, J. Wang, D. B. Mitzi, Y. Yan, Chem. Mater. 2016, 28, 821.
[22] S. Perera, H. Hui, C. Zhao, H. Xue, F. Sun, C. Deng, N. Gross, C. Milleville, X. Xu, D. F. Watson, B. Weinstein, Y.-Y. Sun, S. Zhang, H. Zeng, Nano Energy 2016, 22, 129.
[23] N. Gross, Y.-Y. Sun, S. Perera, H. Hui, X. Wei, S. Zhang, H. Zeng, B. A. Weinstein, Phys. Rev. Appl. 2017, 8, 044014.
[24] X. Wei, H. Hui, S. Perera, A. Sheng, D. F. Watson, Y.-Y. Sun, Q. Jia, S. Zhang, H. Zeng, ACS Omega 2020, 5, 18579.
[25] N. Thakur, P. Kumar, R. Neffati, P. Sharma, Phys. Scr. 2023, 98, 065921.
[26] W. Shockley, H. J. Queisser, J. Appl. Phys. 1961, 32, 510.
[27] Y. Nishigaki, T. Nagai, M. Nishiwaki, T. Aizawa, M. Kozawa, K. Hanzawa, Y. Kato, H. Sai, H. Hiramatsu, H. Hosono, H. Fujiwara, Sol. RRL 2020, 4, 1900555.
[28] Z. Yu, X. Wei, Y. Zheng, H. Hui, M. Bian, S. Dhole, J.-H. Seo, Y.-Y. Sun, Q. Jia, S. Zhang, S. Yang, H. Zeng, Nano Energy 2021, 85, 105959.
[29] R. Yang, A. D. Jess, C. Fai, C. J. Hages, J. Am. Chem. Soc. 2022, 144, 15928.
[30] J. W. Turnley, K. C. Vincent, A. A. Pradhan, I. Panicker, R. Swope, M. C. Uible, S. C. Bart, R. Agrawal, J. Am. Chem. Soc. 2022, 144, 18234.
[31] A. A. Pradhan, M. C. Uible, S. Agarwal, J. W. Turnley, S. Khandelwal, J. M. Peterson, D. D. Blach, R. N. Swope, L. Huang, S. C. Bart, R. Agrawal, Angew. Chem., Int. Ed. 2023, 62, e202301049.
[32] A. Crovetto, R. Nielsen, M. Pandey, L. Watts, J. G. Labram, M. Geisler, N. Stenger, K. W. Jacobsen, O. Hansen, B. Seger, I. Chorkendorf,f P. C. K. Vesborg, Chem. Mater. 2019, 31, 3359.
[33] K. Kuhar, A. Crovetto, M. Pandey, K. S. Thygesen, B. Seger, P. C. Vesborg, O. Hansen, I. Chorkendorf,f K. W. Jacobsen, Energy Environ. Sci. 2017, 10, 2579.
[34] Y. Liang, Y. Zhang, J. Xu, J. Ma, H. Jiang, X. Li, B. Zhang, X. Chen, Y. Tian, Y. Han, Z. Shi, Nano Res. 2023, 16, 7867.
[35] W. Feng, R. Zhao, X. Wang, B. Xing, Y. Zhang, X. He, L. Zhang, J. Energy Chem. 2022, 70, 1.
[36] G. Kresse, J. Furthmüller, Comput. Mater. Sci. 1996, 6, 15.
[37] G. Kresse, D. Joubert, Phys. Rev. B 1999, 59, 1758.
[38] J. P. Perdew, A. Ruzsinszky, G. I. Csonka, O. A. Vydrov, G. E. Scuseria, L. A. Constantin, X. Zhou, K. Burke, Phys. Rev. Lett. 2008, 100, 136406.
[39] J. Sun, A. Ruzsinszky, J. P. Perdew, Phys. Rev. Lett. 2015, 115, 036402.
[40] H. Peng, Z.-H. Yang, J. P. Perdew, J. Sun, Phys. Rev. X 2016, 6, 041005.
[41] A. Togo, L. Chaput, T. Tadano, I. Tanaka, J. Phys.: Condens. Matter 2023, 35, 353001.
[42] V. M. Goldschmidt, Naturwissenschaften 1926, 14, 477.
[43] M. I. Dar, G. Jacopin, S. Meloni, A. Mattoni, N. Arora, A. Boziki, S. M. Zakeeruddin, U. Rothlisberger, M. Grätzel, Sci. Adv. 2016, 2, e1601156.
[44] B. P. Falcão, J. P. Leitão, J. C. González, M. R. Correia, K. G. ZayasBazán, F. M. Matinaga, M. B. Moreira, C. F. Leite, A. G. De Oliveira, J. Mater. Sci. 2013, 48, 1794.
[45] J. P. Perdew, K. Burke, M. Ernzerhof, Phys. Rev. Lett. 1996, 77, 3865.
[46] S. Nosé, J. Chem. Phys. 1984, 81, 511.
[47] H. J. Monkhorst, J. D. Pack, Phys. Rev. B 1976, 13, 5188.