First Principles Study on the Features of Ca_xSr_2−xTa₂O₇ (x = 0, 1) as a Photocatalytic Material ^†

Abstract

With hydrogen as one of the energetic vectors craved for use in the future, the successful de-carbonization of the energy sector will require an increase in hydrogen production from renewable resources. Materials that are able to catalyze the water-splitting reaction through sunlight absorption have been widely studied as an adequate solution for green hydrogen generation. Among the proposed tantalum-based oxide materials, Sr₂Ta₂O₇ displays moderate photocatalytic activity. Aiming to improve the photocatalytic properties by means of compositional modifications, this work presents a DFT study of the Sr substitution with Ca. The structural, energetic, and electronic features of the phases of Ca_xSr_2−xTa₂O₇ (0 < x < 1) have been examined. The computational results utilizing the SCAN functional show that there is a slight decrement in the band gap value (from 3.65 eV for x = 0 to 3.50 eV for x = 1) concomitant to a minor distortion of the crystal structure.

1. Introduction

Climate change might be one of the biggest challenges that we face nowadays as a society. Currently, most of the produced energy comes from fossil fuels, which contributes to greenhouse gas emissions. Energy production through renewable sources is being explored as a solution to this major issue. In that respect, hydrogen (H₂) seems to be the energetic vector of the future [1,2]; it presents a high combustion efficiency and the highest energy content per unit mass of all fossil fuels normally used [3]. However, hydrogen production rarely comes from zero-emission sources. In this sense, hydrogen production through water splitting using sunlight (i.e., photocatalysis) presents a key to green hydrogen generation [4] (p. 3).

The photocatalytic dissociation of water can be achieved using semiconductor materials that are able to absorb sunlight, leading to the promotion of electrons from the valence band (VB) to the conduction band (CB). There are, however, some vital concerns that must be considered: the VB maximum should be below the O₂/H₂O oxidation potential (1.23 eV) and the CB minimum should be above the H₂O/H₂ reduction potential (0.0 eV). Likewise, the band gap should be smaller than 3.0 eV to use visible light and larger than 1.23 eV to fulfill band position requirements [5]. Several compounds have been studied for this purpose, for example, TiO₂, WO₃, CdSe, CdS, and Ta₂O₅ [6]. Among the investigated d⁰-TM oxides, Sr₂Ta₂O₇ [7,8] shows some exciting characteristics, particularly in its CB position relative to H₂O/H₂ reduction potential. However, its large band gap (4.6 eV) limits the light absorption to the UV region of sunlight.

The Sr₂Ta₂O₇ compound presents a perovskite-derived structure [9], where four perovskite layers form 2D nets perpendicular to the c axis, as shown in Figure 1. In the ABO₃ perovskite blocks, Ta occupies the B position, being coordinated to six O atoms. Sr occupies two different positions, Sr(1) and Sr(2), that are coordinated to 8 and 12 O atoms, respectively. Sr(1) can be described as an atom between perovskite block layers, while Sr(2) can be seen as an atom inside those blocks.

Previous computational studies [10,11] suggest that anion modifications along with cation doping in the Ta site tune the band gap value and the relative positions of the VB/CB, thereby influencing the photocatalytic activity As for cation doping in the Sr site, some authors [12] have experimentally demonstrated that the partial substitution of Sr for Ba improves the photocatalytic activity. Specifically, the solid solution Sr_2−xBa_xTa₂O₇ was successfully synthesized in the compositional range 0 < x < 0.4, with the term x = 0.4 displaying the best catalytic performance. This makes the investigation of the substitution of Sr with other cations such as Ca, appealing in that it could potentially enhance the photocatalytic activity of Sr₂Ta₂O₇. For this purpose, this work focuses on the computational study through DFT calculations of the structural and electronic changes that the substitution of Sr with Ca can provoke in the Ca_xSr_2−xTa₂O₇ family. In particular, the x = 0 and 1 phases are studied.

2. Materials and Methods

2.1. Crystallographic Models

Crystallographic models were built from the Sr₂Ta₂O₇ structure (ICSD file 601) [9]. The structure was used to model the hypothetical Sr_2−xCa_xTa₂O₇ compound with, x = 1. However, since there are two possible crystallographic sites for the Sr substitution (Figure 1), two models of CaSrTa₂O₇ were constructed. The substitution of Sr(1) for Ca led to the model named CaSrTa₂O₇(1), and the substitution of Sr(2) for Ca generated the model named CaSrTa₂O₇(2).

2.2. Computational Methods

DFT calculations were performed using the VASP package (Vienna ab-initio simulation package) developed at the Universität Wien [13,14]. The Projector Augmented Wave (PAW) [15] method was used to describe the interaction of core electrons with nuclei, specifically 4s²4p⁶5s² for Sr, 5p⁶6s²5d³ for Ta, 2s²2p⁴ for O, and 3s²3p⁶4s² for Ca were treated as valence electrons. A meta-GGA exchange-correlation functional was utilized; in particular, the recently developed strongly constrained and appropriately normed (SCAN) functional [16]. The energy cutoff was set at 600 eV throughout all calculations. Integration of the first Brillouin zone was carried out under the determination of k-points by the Monkhorts–Pack scheme. The k-point meshes were set at 6 × 2 × 6 in all cases, along with a Gaussian smearing parameter of 0.05 eV. For the density of states (DOS) calculations, the tetrahedron method with Bloch corrections [17] was employed. The tolerance threshold in total energy to achieve self-consistency was set at 1 × 10⁻⁴ eV. Structural relaxation was performed over atoms’ position, as well as cell shape and volume. No symmetry constraints were imposed during a relaxation for the Ca models calculations.

3. Results and Discussion

3.1. Crystal Structure

The structural optimization results for Sr₂Ta₂O₇ within the SCAN functional are shown in

Table 1. Lattice parameters and atomic distances in the Sr₂Ta₂O₇ (S.G. Cmcm) structure and CaSrTa₂O₇ models (S.G. Cmcm).

Lattice Parameters (Å)
	Sr2Ta2O7			CaSrTa2O7(1)	CaSrTa2O7(2)
	Experimental	Calculated	% Error	Calculated	Calculated
a	3.9376	3.9537	0.4	3.9084	3.9273
b	27.1986	27.1894	0.03	27.0999	27.0855
c	5.6927	5.6951	0.04	5.6119	5.6646
Volume (Å3)	609.50	612.22	0.4	594.40	602.56
Distances (Å)
	Sr2Ta2O7			CaSrTa2O7(1)	CaSrTa2O7(2)
	Experimental	Calculated	% Error	Calculated	Calculated
Sr1/Ca1-O	2.47 (x4)	2.480	0.4	2.416	2.469
Sr1/Ca1-O	2.60 (x2)	2.575	1.0	2.435	2.572
Sr1/Ca1-O	3.27 (x2)	3.313	1.3	3.338	3.347
Sr2/Ca2-O	2.71 (x4)	2.719	0.3	2.697	2.653
Sr2/Ca2-O	2.827 (x4)	2.837	0.4	2.809	2.848
Sr2/Ca2-O	2.85 (x2)	2.852	0.1	2.811	2.849
Sr2/Ca2-O	2.70	2.629	2.6	2.650	2.476
Sr2/Ca2-O	2.74	2.743	0.1	2.731	2.721
Ta1-O	1.87 (x2)	1.855	0.8	1.871	1.859
Ta1-O	1.979 (x2)	1.989	0.5	1.976	1.990
Ta1-O	2.16 (x2)	2.071	4.1	2.141	2.128
Ta2-O	1.89 (x2)	1.855	1.9	1.832	1.907
Ta2-O	1.98 (x2)	1.997	0.9	1.969	1.978
Ta2-O	2.07 (x2)	2.157	4.2	2.067	2.032

. The lattice parameters obtained are in good agreement with experimental measurements [9], with errors below 0.5%. Likewise, distances results are, in general, in line with experimental values, showing deviations below 4.5%. Thus, the SCAN functional correctly reproduces the crystal structure of Sr₂Ta₂O₇.

Regarding the relative stability of the two crystallographic models proposed for CaSrTa₂O₇, DFT calculations show that the CaSrTa₂O₇(1) model has lower energy than the CaSrTa₂O₇(2) model (energy difference of 332 meV). The relative stability can be rationalized considering the ionic radii of Ca²⁺ (r = 1.12 Å for C.N. = 8 and 1.34 Å for C.N. = 12 [18]) and Sr²⁺ (r = 1.26 Å for C.N. = 8 and 1.44 Å for C.N. = 12 [18]). This calculated energy indicates that a more stable structure is formed if the larger Sr²⁺ occupies the position inside perovskite blocks (C.N. = 12), which is to say, the smaller Ca²⁺ cation will tend to occupy the site between perovskite layers position (C.N. = 8).

Lattice parameters and atomic distances for the most stable CaSrTa₂O₇ model are listed in Table 1. As expected, the introduction of the smaller Ca²⁺ ion produces a contraction of the unit cell and a reduction in all lattice parameters. Overall, the introduction of Ca ions preserves the initial crystal structure of Sr₂Ta₂O₇.

3.2. Electronic Structure

Figure 2 and Figure 3 show the calculated density of states and band structure of the Sr₂Ta₂O₇ and DOS of the two CaSrTa₂O₇ models, respectively.

The DOS for Sr₂Ta₂O₇ is represented in Figure 2a. The Fermi level is set at the 0 of energy. The VB is mainly formed by the 2p states of O while CB is composed of the 5d Ta empty states (in blue). Yet, a good hybridization, signature of Ta-O covalence, exists between O and Ta states in both bands. Sr barely contributes to forming the valence and conduction bands. As shown in Figure 2a,b, band gap of 3.65 eV is obtained using the SCAN functional. This result deviates from the experimental one (4.6 eV), given that band gap underestimation is a known failure of DFT methods [19].

Despite the limitations of DFT to reproduce accurate band-gap values, the trends in band-gaps due to chemical composition modifications are fully reliable [11].

Figure 3 shows the calculated DOS for the two models used to simulate the CaSrTa₂O₇ compounds. CaSrTa₂O₇(1) has a band gap of 3.55 eV while the CaSrTa₂O₇(2) band gap is 3.50 eV. Therefore, calculations suggest that in the Sr_2−xCa_xTa₂O₇ family, the trend is to reduce the band gap as x increases. Importantly, band gap variations are of only about 0.15 eV when going from x = 0 to x = 1. It is well documented that in d⁰ transition metals oxides crystallizing in the ABO₃ perovskite structure, the A substitution with other ions can provoke a distortion in the crystal structure, which could ultimately affect the electronic structure producing significant band-gap values modifications [20]. Since, in the present case, the partial substitution of Sr with Ca causes minor crystal structure modifications, the variations in the electronic structure are subtle. Additionally, valence and conduction bands in CaSrTa₂O₇ models have essentially the same features as in the Sr₂Ta₂O₇ phase.

Considering the exposed results, the substitution of Sr with Ca in Sr₂Ta₂O₇ only leads to a slight reduction in the band gap value that is independent of the Ca position. This seems insufficient to cause an effective change in the photocatalytic activity of the material. Taking into account the experimentally observed improvement of photocatalytic activity when substituting Sr with Ba in some Sr_2−xBa_xTa₂O₇ phases [12], more work is needed to assess the impact of the Sr substitution in Sr₂Ca₂O₇ with Ca/Ba and other cations in the electronic structure. It should be highlighted that other critical features could account for an improved catalytic activity in Sr_2−xBa_xTa₂O₇, such as the decrement of the electron-hole recombination.

4. Conclusions

Computational investigations at the level of Density Functional Theory (DFT) permit the prediction of some basic and critical features in view of the photocatalytic activity of materials. In this work, DFT methods are used to predict the crystal and electronic structures of the potential photocatalyst Sr_2−xCa_xTa₂O₇. We found that the x = 1 phases preserve the perovskite-related structure of the parent Sr₂Ta₂O₇. The calculated DOS shows a band narrowing of 0.10–0.15 eV because of the Sr substitution with Ca. With band gaps of ~4.5 eV, sunlight absorption by the SrCaTa₂O₇ photocatalyst is limited to the UV region. Work is in progress to investigate other substitutions that may enhance the photocatalytic activity of Sr₂Ta₂O₇.