Versatile Interplay of Chalcogenide and Dichalcogenide Anions in the Thiovanadate Ba₇S(VS₃O)₂(S₂)₃ and Its Selenide Derivatives: Elaboration and DFT Meta-GGA Study

Abstract

Oxychalcogenides are emerging as promising alternative candidates for a variety of applications including for energy. Only few phases among them show the presence of Q–Q bonds (Q = chalcogenide anion) while they drastically alter the electronic structure and allow further structural flexibility. Four original oxy(poly)chalcogenide compounds in the system Ba–V–Q–O (Q = S, Se) were synthesized, characterized, and studied using density functional theory (DFT). The new structure type found for Ba₇V₂O₂S₁₃, which can be written as Ba₇S(VS₃O)₂(S₂)₃, was substituted to yield three selenide derivatives Ba₇V₂O₂S_9.304Se_3.696, Ba₇V₂O₂S_7.15Se_5.85, and Ba₇V₂O₂S_6.85Se_6.15. They represent original multiple-anion lattices and first members in the system Ba–V–Se–S–O. They exhibit in the first layer heteroleptic tetrahedra V⁵⁺S₃O and isolated Q^2– anions and in the second layer dichalcogenide pairs (Q₂)^2– with Q = S or Se. Selenide derivatives were attempted by targeting the selective substitution of isolated Q^2– or (Q₂)^2– (in distinct layers) or both by selenide, but it systematically led to concomitant and partial substitution of both sites. A DFT meta-GGA study showed that selective substitution yields local constraints due to rigid VO₃S and pairs. Experimentally, incorporation of selenide in both layers avoids geometrical mismatch and constraints. In such systems, we show that the interplay between the O/S anionic ratio around V⁵⁺, together with the presence/nature of the dichalcogenides (Q₂)^2– and isolated Q^2–, impacts in unique manners the band gap and provides a rich background to tune the band gap and the symmetry.

Introduction

Oxychalcogenides are, among mixed-anion compounds,¹ growing in popularity because of their great versatility and tunable properties for a large scope of applications. They open new perspectives to overcome challenges and hindrances to the development of various applications such as thermoelectrics^2,3 or photocatalytic hydrogen production,^4,5 and so on. The coexistence of oxide anions and more electronegative chalcogenide anions Q (Q = chalcogenide) leads to a highly distinctive structural chemistry.⁶

In this context, many oxychalcogenide compounds exhibit 2D layered-type structures, such as the narrow band-gap semiconductors Sr₂CuO₂Cu₂Se₂,⁷ Bi₂YO₄Cu₂Se₂,⁸ or the superconductors LaO_1–xF_xBiS₂.⁹ Besides, few complex oxychalcogenides show the presence of Q–Q bonds (Q = chalcogenide anion) leading to polychalcogenide anions, which drastically affects the electronic structure and therefore the physical properties.¹⁰ Some “simple” binaries such as cobalt persulfide CoS₂¹¹ or perselenide CoSe₂¹² are well-known dichalcogenide anion-based compounds. They are widely studied for their catalytic properties.¹³ Polychalcogenides are also crucial in lithium–sulfur batteries, which are among the most promising technologies in this field.¹⁴ Recently, new compounds could be designed by a redox metal insertion into polychalcogenide-based phases.¹⁰¹⁵ While polychalcogenide anions show great diversity, fewer complex inorganic systems containing them are found. The latter may show, for instance, very interesting features to enhance thermoelectric properties such as in BaBiTe_3–xSe_x¹⁶ or A₂BaCu₈Te₁₀ (A = K, Rb, Cs).¹⁷ Besides, oxychalcogenides with polychalcogenide anions are limited. La₂O₂S₂ and its derivatives are well-known materials for optics and are made of layers of La₂O₂ stacked with a layer of disulfide pairs.¹⁸ Recently, topochemical (de)intercalation of sulfur in the latter was demonstrated, which opened new perspectives.¹⁹ On the other hand, we recently reported the phase Ba₅(VO₂S₂)₂(S₂)₂²⁰ that contains dichalcogenide pairs occupying in a unique 1D manner channels delimited by barium cations. This compound was the first in the Ba–V–S–O system with persulfide entities, while the conventional following oxysulfides were reported: Ba₆V₄O₅S₁₁,²¹ Ba₃V₂S₄O₃,²² Ba₁₅V₁₂S₃₄O₃,²³ and Ba₁₀S(VO₃S)₆.²⁴ All of these compounds exhibit tetrahedral thiovanadates as found in the simple series A₃(VO_4–xS_x)²⁵ (A = Na, K) with disconnected VO₂S₂ in the case of Ba₅ (VO₂S₂)₂(S₂)₂.

In this work, we present four new complex polychalcogenide phases. The new phase Ba₇S (VS₃O)₂(S₂)₃ (phase 1) contains VS₃O thiovanadates and isolated S^2– anions into a layer and an original arrangement of persulfide (S₂)^2– pairs into the second layer, thus leading to a complex anionic interplay. This is amplified in the three selenide derivatives Ba₇V₂O₂S_9.304Se_3.696, Ba₇V₂O₂S_7.15Se_5.85, and Ba₇V₂O₂S_6.85Se_6.15, which exhibit five anionic species (O^2–, S^2–, Se^2–, (S₂)^2–, and (Se₂)^2–). For all phases, a comparative crystallochemical discussion based on single-crystal XRD is complemented by electronic structure analysis calculated from density functional theory together with optical properties of high-purity powder samples. We show multiple possibilities and complex anionic interplay in this system and carry out a comprehensive study.

Experimental Section

Synthesis

The powders were obtained from a mixture of 2BaO, 5BaS, 2V, and 8S pressed into pellets and heated in an evacuated sealed quartz tube. The heat treatment consisted in heating up to 750 °C for 24 h, then cooling down to 650 °C at a 5 °C/h rate, and finishing with a rapid cooling down to room temperature.

Initially, Ba₇V₂O₂S₁₃ was found in a preparation obtained at high pressure, which is described in the Supporting Information. It has a similar structure with a slightly decreased volume due to the preparation conditions. This phase could then be obtained in a sealed quartz tube as described above.

X-ray Diffraction

X-ray diffraction on single crystals for Ba₇V₂O₂S₁₃ and the selenide derivatives was performed on an X8 diffractometer equipped with a two-dimensional (2D) CCD 4K detector and an Ag Kα.

Powder XRD

The powder XRD patterns were collected on a Bruker D8 diffractometer equipped with a Lynxeye linear detector (Cu Kα) in Bragg–Brentano geometry at room temperature.

Scanning Electron Microscopy (SEM)

SEM experiments and EDX analysis were carried out on an S-3400N (Oxford instruments) and on a Hitachi S400N microscope.

DFT Calculations

DFT calculations were carried out by employing the projector augmented wave (PAW)^26,27 method encoded in the Vienna ab initio simulation package (VASP)²⁸ and the meta-GGA SCAN functional^29,30 for the exchange-correlation functionals. The full geometry optimizations were carried out using a plane-wave energy cutoff of 550 eV and 30 k points in the irreducible Brillouin zone for all models. It converged with residual Hellman–Feynman forces on the atoms smaller than 0.03 eV/Å and led to a good match with the experimental structure, i.e., within a reasonable error expected for the SCAN meta-GGA functional, which led in general to a better match with experience than simple GGA functionals. The relaxed structure was used for calculations of the electronic structure for which a plane-wave cutoff energy of 550 eV and threshold of self-consistent-field energy convergence of 10^–6 eV were used, with 72 k points in the irreducible Brillouin zone for all models.

UV–Visible Measurements

The reflectance of the sample was measured from 250 to 800 nm on a PerkinElmer Lambda 650 device.

Results and Discussion

Structure Resolution and Description

Ba₇S(VS₃O)₂(S₂)₃

A dark single crystal with a platelet shape was selected for XRD data collection. The structure could be solved with the unit cell parameters a = 8.8172(4) Å and c = 17.3428(8) Å in the space group P6₃/m (176). The data collection, refinement details, and structural parameters are given in Tables 1–4. The refinement was carried out with JANA2006³¹ software based on a structure solution obtained with the charge-flipping method implemented in JANA2006 within the SUPERFLIP module.³² EDX analysis of the single crystals led to the average atomic ratio 59.53/8.70/31.77 for S/V/Ba, respectively, in good agreement with the formula Ba₇V₂O₂S₁₃.

Table 1. Data Collection and Refinement Details of Ba₇V₂O₂S₁₃ and Its Selenide Derivatives (Phase 1 → 4).

	phase 1	phase 2	phase 3	phase 4
formula	Ba7V2O2S13	Ba7V2O2S9.304Se3.696	Ba7V2O2S7.15Se5.85	Ba7V2O2S6.85Se6.15
detailed formula	Ba7S(VS3O)2(S2)3	Ba7S1−αSeα(VS3O)2(S1−δSeδ)6	Ba7S1−αSeα(VS3O)2(S1−δSeδ)6	Ba7S1−αSeα(VS3O)2(S1−δSeδ)6
α	0	0.496	0.790	0.903
δ	0	0.533	0.843	0.875
molecular weight (g·mol–1)	1512	1694.5	1786.4	1800.5
symmetry	hexagonal	hexagonal	hexagonal	hexagonal
space group	P63/m (176)	P63/m (176)	P63/m (176)	P63/m (176)
unit cell dimensions (Å)	a = 8.8456(4)	a = 8.9290(3)	a = 8.9670(3)	a = 8.9706(2)
	c = 17.3971(8)	c = 17.6052(5)	c = 17.6747(7)	c = 17.6964(5)
volume (Å3)	1178.86(9)	1215.56(7)	1230.77(8)	1233.27 (5)
Z	2	2	2	2
Data Collection
equipment	Bruker CCD (Mo Kα)	Bruker CCD (Mo Kα)	Bruker CCD (Mo Kα)	Bruker CCD (Ag Kα)
Λ [Å]	0.71073	0.71073	0.71073	0.56087
calculated density (g·cm–3)	4.2597	4.6295	4.8204	4.8841
crystal shape	platelet	platelet	platelet	platelet
crystal dimensions (μm)	75 × 50 × 5	50 × 30 × 5	80 × 45 × 5	150 × 50 × 6
color	dark	dark	dark	dark
absorption correction	multiscan	multiscan	multiscan	multiscan
scan mode	ω, ϕ	ω, ϕ	ω, ϕ	ω, ϕ
θ (min – max) (deg)	2.34–27.88	2.31–28.7	2.3–41.16	1.82–25.26
μ (mm–1)	13.391	18.477	20.981	11.277
F(000)	1324	1464	1535	1545
reciprocal space recording	–11 ≤ h ≤ 11	–12 ≤ h ≤ 12	–16 ≤ h ≤ 16	–13 ≤ h ≤ 13
	–11 ≤ k ≤ 11	–11 ≤ k ≤ 12	–16 ≤ k ≤ 16	–13 ≤ k ≤ 13
	–22 ≤ l ≤ 22	–23 ≤ l ≤ 23	–32 ≤ l ≤ 32	–26 ≤ l ≤ 26
no. of measured reflections	18823	23554	87170	62907
no. of independent reflections	979	1084	2813	1540
I > 3σ(I) (total)	671	876	2044	1292
Refinement
number of refined parameters	41	42	43	43
refinement method	least-squares	least-squares	least-squares	least-squares
weighting scheme	sigma	sigma	sigma	sigma
R1(F) [I > 3σ(I)]/R1(F2) (all data, %)	0.0287/0.0551	0.0191/0.027	0.024/0.039	0.0404/0.0551
wR2(F2) [I > 3σ(I)]/wR2(F2) (all data, %)	0.0648/0.0720	0.0506/0.0517	0.0598/0.0627	0.0591/0.0612
goodness of fit	1.05	1.28	1.21	3.26
max/min residual electronic density (e–/Å3)	2.49/–1.56	2.47/–1.74	1.05/–1.78	3.66/–2.71
Tmin/Tmax	0.629/746	0.564/0.746	0.568/0.748	0.660/0.740

Table 4. Main Distances (Å) for Ba₇O₂S₁₃V₂ and the Selenide Derivatives.

atoms 1 and 2	d 1 and 2 [Å]	atoms 1 and 2	d 1 and 2 [Å]
Ba7S(VS3O)2(S2)3
Ba1–S2	3.280(3) × 3	Ba2–S4	3.220(3)
Ba1–S3	3.238(4) × 3	Ba2–S4	3.251(2)
Ba2–S1	3.3002(5)	Ba2–O1	2.783(2)
Ba2–S2	3.289(2)	V1–O1	1.684(2)
Ba2–S3	3.289(3)	V1–S4	2.129(2) × 3
Ba2–S4	3.3309(19)	S2–S3	2.122(4)
Ba7V2O2S9.304Se3.696
Ba1–Se1	3.3308(3)	Ba1–O1	2.7842(11)
Ba1–Se2	3.3678(8)	Ba2–Se2	3.3709(9) × 3
Ba1–Se2	3.4313(6)	Ba2–Se3	3.3261(10) × 3
Ba1–Se3	3.3611(6)	V1–O1	1.687(4)
Ba1–S4	3.2744(15)	V1–S4	2.129(2) × 3
Ba1–S4	3.257(2)	Se2–Se3	2.3253(16)
Ba7V2O2S7.15Se5.85
Ba1–Se1	3.3310(8)	Ba1–O1	2.7891(10)
Ba1–Se2	3.4075(5)	Ba2–Se2	3.3944(6) × 3
Ba1–Se3	3.3981(4)	Ba2–Se3	3.3552(5) × 3
Ba1–S4	3.3310(8)	V1–O1	1.692(4)
Ba1–S4	3.2855(8)	V1–S4	2.1364(7)
Ba1–S4	3.2664(12)	Se2–Se3	2.3832(4)
Ba7V2O2S6.85Se6.15
Ba1–Se1	3.3400(7)	Ba1–O1	2.788(3)
Ba1–Se2	3.4136(12)	Ba2–Se2	3.398(1) × 3
Ba1–Se3	3.4015(16)	Ba2–Se3	3.360(3) × 3
Ba1–S4	3.335(2)	V1–O1	1.711(9)
Ba1–S4	3.286(2)	V1–S4	2.138(3) × 3
Ba1–S4	3.274(2)	Se2–Se3	2.385(3)

Table 2. Atomic Positions and Isotropic Thermal Displacement for Ba₇O₂S₁₃V₂ and the Selenide Derivatives.

atom	Wyck.		x	y	z	Ueq
Ba7S(VS3O)2(S2)3
Ba1	2d		0.666667	0.333333	0.25	0.0180(3)
Ba2	12i		0.0865(1)	0.3307(1)	0.1148(1)	0.0159(2)
S1	2b		0	0	0	0.0225(14)
S2	6h		0.2456 (4)	0.1899(4)	0.25	0.0165(11)
S3	6h		0.3861(4)	0.4669(3)	0.25	0.0163(11)
V1	4f		0.333333	0.666667	–0.0278(1)	0.0127(6)
S4	12i		0.5628(3)	0.6745(3)	–0.0703(1)	0.0202(8)
O1	4f		0.33333	0.66667	0.0690(2)	0.014(2)
Ba7V2O2S9.304Se3.696
Ba1	12i		0.6659(1)	0.9117(1)	0.61254(1)	0.0169(1)
Ba2	2c		0.333333	0.666667	0.25	0.0191(2)
Se1	2b	0.496	0	0	0	0.0161(4)
S1	2b	0.504	0	0	0	0.0161(4)
Se2	6h	0.533	0.1794(1)	0.2365(1)	0.25	0.0148(3)
S2	6h	0.467	0.1794(1)	0.2365(1)	0.25	0.0148(3)
Se3	6h	0.533	0.4801(1)	0.3899(1)	0.25	0.0187(3)
S3	6h	0.467	0.4801(1)	0.3899(1)	0.25	0.0187(3)
V1	4f		0.33333	0.66667	0.47120(6)	0.0147(3)
S4	12i		0.1145(1)	0.6760(1)	0.4293(1)	0.0246(5)
O1	4f		0.33333	0.66667	0.5670(2)	0.0116(12)
Ba7V2O2S7.15Se5.85
Ba1	12i		0.6656(1)	0.9116(2)	0.6111(1)	0.0148(1)
Ba2	2c		0.33333	0.66667	0.25	0.0192(1)
Se1	2b	0.790	0	0	0	0.0142(1)
S1	2b	0.21	0	0	0	0.0142(1)
Se2	6h	0.843	0.0583(1)	0.8233(1)	0.25	0.0145(1)
S2	6h	0.157	0.0583(1)	0.8233(1)	0.25	0.0145(1)
Se3	6h	0.843	0.6091(1)	0.0927(1)	0.25	0.0158(1)
S3	6h	0.157	0.6091(1)	0.0927(1)	0.25	0.0158(1)
V1	4f		0.33333	0.66667	0.4708(1)	0.0134(1)
S4	12i		0.1145(1)	0.6754(1)	0.4289(1)	0.0217(2)
O1	4f		0.33333	0.66667	0.5664(1)	0.0129(6)
Ba7V2O2S6.85Se6.15
Ba1	12i		0.2472(1)	0.9124(1)	0.6113(1)	0.0114(2)
Ba2	2c		0.33333	0.66667	0.25	0.0169(3)
Se1	2b	0.9028	0	0	0	0.0151(6)
S1	2b	0.0972	0	0	0	0.0151(6)
Se2	6h	0.875	0.0580(2)	0.2347(2)	0.25	0.0109(5)
S2	6h	0.125	0.0580(2)	0.2347(2)	0.25	0.0109(5)
Se3	6h	0.875	0.4837(2)	0.0931(2)	0.25	0.0122(5)
S3	6h	0.125	0.4837(2)	0.0931(2)	0.25	0.0122(5)
V1	4f		0.33333	0.66667	0.4704(1)	0.0089(6)
S4	12i		0.3250(3)	0.8862(3)	0.4289(1)	0.0146(8)
O1	4f		0.33333	0.66667	0.5671(5)	0.010(2)

Table 3. Anisotropic Thermal Parameters U_ij (Ų) for the Sulfide and the Selenide Derivatives.

atom	U11	U22	U33	U12	U13	U23
Ba7S(VS3O)2(S2)3
Ba1	0.0151(4)	0.0151(4)	0.0237(6)	0.0076(2)	0	0
Ba2	0.0136(4)	0.0173(2)	0.0164 (2)	0.0074(2)	0.0009(2)	0.0033(2)
S1	0.0173(15)	0.0173(15)	0.033(3)	0.0087(8)	0	0
S2	0.0177(13)	0.0162(14)	0.0168(13)	0.0093(12)	0	0
S3	0.0169(14)	0.0158(14)	0.0152(12)	0.0074(11)	0	0
V1	0.0146(7)	0.0146(7)	0.0089(10)	0.0073(3)	0	0
S4	0.0179(10)	0.0198(10)	0.0231(10)	0.0095(9)	0.0049(8)	0.0019(8)
O1	0.018(3)	0.018(3)	0.004(4)	0.0092(14)	0	0
Ba7V2O2S9.304Se3.696
Ba1	0.0174(1)	0.0147(1)	0.0173(1)	0.0071(1)	–0.0032(1)	–0.0009(1)
Ba2	0.0153(2)	0.0153(2)	0.0269(3)	0.0076(1)	0	0
Se1	0.0150(5)	0.0150(5)	0.0182(8)	0.0075(3)	0	0
S1	0.0150(5)	0.0150(5)	0.0182(8)	0.0075(3)	0	0
Se2	0.0149(4)	0.0174(4)	0.0137(4)	0.0092(4)	0	0
S2	0.0149(4)	0.0174(4)	0.0137(4)	0.0092(4)	0	0
Se3	0.0185(4)	0.0181(5)	0.0181(4)	0.0081(4)	0	0
S3	0.0185(4)	0.0181(5)	0.0181(4)	0.0081(4)	0	0
V1	0.0171(4)	0.0171(4)	0.0099(5)	0.0085(2)	0	0
S4	0.0238(6)	0.0250(6)	0.0286(6)	0.0152(6)	–0.0061(4)	0.0004(5)
O1	0.0128(14)	0.0128(14)	0.009(2)	0.0064(7)	0	0
Ba7V2O2S7.15Se5.85
Ba1	0.0149(1)	0.0139(1)	0.0148(1)	0.0065(1)	–0.0013(1)	–0.0006(1)
Ba2	0.0147(1)	0.0147(1)	0.0281(2)	0.0073(1)	0	0
Se1	0.0136(2)	0.0136(2)	0.0155(3)	0.0068(1)	0	0
S1	0.0136(2)	0.0136(2)	0.0155(3)	0.0068(1)	0	0
Se2	0.0141(1)	0.0137(1)	0.0144(1)	0.0060(1)	0	0
S2	0.0141(1)	0.0137(1)	0.0144(1)	0.0060(1)	0	0
Se3	0.0160(1)	0.0174(2)	0.0155(2)	0.0094(1)	0	0
S3	0.0160(1)	0.0174(2)	0.0155(2)	0.0094(1)	0	0
V1	0.0157(2)	0.0157(2)	0.0090(2)	0.0078(1)	0	0
S4	0.0210(3)	0.0223(3)	0.0247(3)	0.0130(2)	–0.0057(2)	0.0006(2)
O1	0.0152(7)	0.0152(7)	0.0082(10)	0.0076(3)	0	0
Ba7V2O2S6.85Se6.15
Ba1	0.0114(3)	0.0099(3)	0.0119(3)	0.0046(2)	0.0007(2)	–0.0005(2)
Ba2	0.0120(4)	0.0120(4)	0.0267(7)	0.0060(2)	0	0
Se1	0.0139(7)	0.0139(7)	0.0175(10)	0.0069(3)	0	0
S1	0.0139(7)	0.0139(7)	0.0175(10)	0.0069(3)	0	0
Se2	0.0110(6)	0.0124(6)	0.0101(6)	0.0064(5)	0	0
S2	0.0110(6)	0.0124(6)	0.0101(6)	0.0064(5)	0	0
Se3	0.0118(6)	0.0145(7)	0.0104(6)	0.0067(5)	0	0
S3	0.0118(6)	0.0145(7)	0.0104(6)	0.0067(5)	0	0
V1	0.0084(7)	0.0084(7)	0.0099(10)	0.0042(3)	0	0
S4	0.0153(10)	0.0115(10)	0.0183(10)	0.0078(9)	–0.0007(9)	0.0039(8)
O1	0.013(3)	0.013(3)	0.004(4)	0.0065(15)	0	0

The original structure can be viewed as a rather open framework but can be described by two alternating layers (Figure 1d). The first layer is formed by disconnected VS₃O tetrahedra (Figure 1a) separated from each other by Ba²⁺ cations that also coordinate isolated sulfide S^2– (S1) anions into the S1Ba₆ octahedron with d_Ba1–S1= 3.282(2) Å; this distance is typical of a Ba²⁺–S^2– bond and is, for instance, comparable to that found in Ba₉V₃S₁₁(S₂)₂,³³d_Ba–S = 3.173(2) Å (see Figure 1c). In the mixed-anion tetrahedra based on V⁵⁺, we found V1–S4 = 2.121(1) Å (×3) and V1–O1 = 1.688(3) Å. These distances are comparable to those found, for instance, in thiovanadates within Ba₅V₂S₈O₄ (Figure 1h), i.e., V1–S1 = 2.178(2) Å and V1–O2 = 1.680(5) Å, or in Ba₁₅V₁₂S₃₄O₃ with comparable distances. The polar [VS₃O]^3– tetrahedral entities are arranged with their O vertices along the crystallographic unique axis c, but the space group inversion symmetry makes them alternatively point in the opposite direction along c so that no polar arrangement arises from them (Figure 1e), in contrast to K₃(VO₃S)³⁴ for instance. The second layer consists of disulfide pairs (S₂)^2– surrounded by Ba²⁺ cations (Figure 1g). These Ba²⁺ cations are distributed such as forming cavities. In the latter, the disulfide pairs are arranged pointing in three different directions. Three distinct pairs (representing the three different orientations) coordinate Ba2. Such an arrangement of disulfide pairs is original to the best of our knowledge. The two layers forming the structure are linked via weak Ba1–O1 (Figure 1b,1f) bonds. Ba₇V₂S₁₃O₂ can be written as Ba₇(S₂)₃(VS₃O)₂S to consider the structural units involved.

Figure 1.

(a) View of the heteroleptic entity VS₃O with the V1–O1 distance indicated, (b) O1 atom environments with Ba–O1 distance, and (c) S1 atom environments with S1–Ba1 distance indicated. (d) Structure of Ba₇V₂S₁₃O₂. (e) View of the tetrahedra arrangement projected along the deviated b axis, (f) Ba1 environments with S2–S3 bonds and distances indicated, and (g) S2–S3 bonds arrangement represented around Ba2; the dotted lines represent the separations S2–S2 and S3–S3 (unconnected atoms). (h) View of Ba₅(VO₂S₂)₂(S₂)₂ (ref (20)) and its (i) disulfide pair rearrangement.

Investigating Selective Substitution of Sulfur by Selenium

Starting from Ba₇S(S₂)₃(VS₃O)₂, we have attempted selective substitution by selenide anions. We targeted either the (Q₂) sites or the isolated Q^2– site or both, in Ba₇(Q^2–)(VS₃O)₂(Q₂)₃. We present here the results fulfilled for three compositions. In the high-purity powders obtained for the targeted substitutions Ba₇S(VS₃O)₂(Se₂)₃ (i.e., Ba₇V₂O₂S₇Se₆) and Ba₇Se(VS₃O)₂(Se₂)₃ (i.e., Ba₇V₂O₂S₆Se₇), single crystals were found. They correspond to similar structures as the parent phase. They show the average refined compositions of Ba₇V₂O₂S_7.15Se_5.85 and Ba₇V₂O₂S_6.85Se_6.15, i.e., close to the nominal ones, but the refinements (which will be further detailed below) show a disordered distribution of Se in both Q^2– and (Q₂) sites (VS₃O is not altered) with mixed S/Se sites. Then, Ba₇S(VS₃O)₂(SeS)₃ (i.e., Ba₇V₂O₂S₁₀Se₃) was attempted to lower the Se content and force a selective substitution in Q₂. In the corresponding preparation, a single crystal was selected and its structure resolution revealed also a composition close (Ba₇V₂O₂S_9.304Se_3.696) to the nominal one but again with the disordered incorporation of selenide in both Q/(Q₂). The disordered dichalcogenide pairs in the selenide derivatives are arranged (Figure 2a) in the same manner as the disulfide pairs in the oxysulfide parent. Considering the single crystals found in the three above-mentioned preparations, we thus have characterized, in addition to Ba₇S(VS₃O)₂(S₂)₃, the following three selenide derivatives: Ba₇S_0.504Se_0.496(VS₃O)₂((S_0.467Se_0.533)₂)₃, Ba₇S_0.21Se_0.79(VS₃O)₂((S_0.157Se_0.843)₂)₃, and Ba₇S_0.0972Se_0.9028(VS₃O)₂((S_0.125Se_0.875)₂)₃.

Figure 2.

(a) Se2–Se3 bond arrangement represented when present in their vicinity: Ba2 and its connectivity with Se2–Se3 bonds. (b) Ba1 environments. (c) Isolated Se1 atom environment. To simplify, the mixed sites Se1/S1 are designed by Se1 and so on for Se2 and Se3.

They may be named in the text, respectively, as phase 1, phase 2, phase 3, and phase 4 (with increased selenide content from phase 1 → 4) for clarity.

Structure Analysis of the Selenide Derivatives

The refinement details, structural parameters, and the main geometrical features are gathered in Tables 1–5. Figure 3 shows the regular increase of the unit cell parameters upon increasing selenide in the structure. The parameters α and δ define the substitution ratio by selenide in the Q_isolated site and in the (Q₂) pair, respectively (Table 1). Clearly, both anionic entities are impacted in more or less similar proportions upon increased selenide incorporation. For phase 2, Q_isolated is a bit more substituted. For phase 3, α and δ are very similar, and for phase 4, (Q₂) is slightly more substituted. Thus, for a lower selenide global content, Q_isolated is favored, and then, the (Q₂) sites take over progressively upon increasing selenide content for phases 3 and 4. Having said that, both entities are filled in significant amounts at all ratios investigated here. For (Q₂), the distinct crystallographic sites for each anion involved in the pair show a very close selenide ratio; thus, in the final refinement, they were constrained equal, which did not alter the quality of the refinement and is consistent with the occurrence of either S₂ or Se₂ pairs. We note that for phase 4 with the greater selenide content, the two (Q₂) sites’ occupancies show a moderate divergence when not constrained equal. This can also be related to the evolution of the refinements. For phase 1 → 3, the overall quality of the refinements is very high and the reliability factors are excellent. For phase 4, a careful analysis of the precession images from the single-crystal XRD data shows the appearance of few low-intensity extra spots (Figure S2) compared to the other phases, which we could attribute to an additional domain. A good-quality refinement could be reached after taking into account a twin domain that we found in the proportion 5.1%. Then, a careful comparison of the selenides with the parent sulfide phase shows the opposite orientation of entities in the (ab) plane within the unit cell although the structures seem identical at first sight. This is detailed in Figure S1 and is easily noticed with the Wickoff position of one of the Ba atoms, which is 2d in the sulfide and 2c in the selenides.

Table 5. Main Geometrical Information for Phases 1, 2, 3, and 4 from the Single-Crystal Refinements.

	phase 1	phase 2	phase 3	phase 4
unit cell parameters, Å	a = 8.8456(4), c = 17.3971(8)	a = 8.9290(3), c = 17.6052(5)	a = 8.9670(3), c = 17.6747(7)	a = 8.9706(2), c = 17.6964(5)
volume, Å3	1178.86(9)	1215.56(7)	1230.77(8)	1233.27 (7)
VO3S (Td) distances (Å)	V–O = 1.683(7)	V–O = 1.687(4)	V–O = 1.689(2)	V–O = 1.711(9)
	V–S = 2.129(3)	V–S = 2.129(2)	V–S = 2.1355(8)	V–S = 2.138(2)
Ch–Ch (Å)	2.122(6) (S2)	2.325(1)	2.3832(4)	2.385(3)
Ba1–Ch(isolé)	3.3002(5)	3.3308(3)	3.3323(2)	3.340(1)
Ba2–Ch(paire)	3.289(2)	3.3259(9)/3.3712(7)	3.3940(3)/3.3557(4)	3.3978(1)/3.360(2)
Ba2–S4 interlayer	3.709(2)	3.736(1)	3.74258(7)	3.748(1)
band gap (eV)	1.59	1.53	1.51	1.49

Figure 3.

Unit cell parameters (a and c) and cell volume progress (from the single crystal) while the Se percentage increases. The volume and the unit cell parameters increase smoothly with the Se content (atomic percentage of Se in the phases).

More into details of the three selenide derivatives Ba₇S_0.504Se_0.496(VS₃O)₂((S_0.467Se_0.533)₂)₃, Ba₇S_0.21Se_0.79(VS₃O)₂((S_0.157Se_0.843)₂)₃, and Ba₇S_0.0972Se_0.9028(VS₃O)₂((S_0.125Se_0.875)₂)₃, we found the distance Q2–Q3 between 2.325(2) and 2.385(3) Å lower than that found in Ba₃Ta₂Se₉ with d_Se–Se= 2.50 ų⁵ and in Sr₄Sn₂Se₉ and Sr₄Sn₂Se₁₀ with 2.459 and 2.462 Å, respectively,³⁶ also larger than that for the oxysulfide parent compound Ba₇S(VO₃S)₂(S₂)₃ for which d_S–S = 2.122 Å. The distances between the selenide derivatives are consistent with the mixing of the two types of pairs (S₂) and (Se₂) and depend on the selenide content. The large cation Ba1 with highly distorted environments is coordinated to four anionic species O^2–, (Se/S)₂^2–, (Se/S)^2–, and S^2– with one characteristic Ba–O1 distance, which is almost constant (in the range 2.783(2)–2.789(1) Å) for all compounds (Figure 2b). The isolated chalcogenide anion (Se1/S1)^2– is coordinated to six Ba²⁺ cations to form the octahedron (Se1/S1)Ba₆ with d_{Ba1–(Se1/S1)} from 3.331(1) to 3.340(1) Å in the range reported for Ba₇Sn₃Se₁₃³⁷ (d_Se–Ba ∼ 3.183 to 3.761 Å); see Figure 2c.

Polycrystalline Phase Analysis

As depicted in Figure 4, the title phases are formed with high purity. The minor impurity Ba₆(VO₂S₂)₂(VS₃O)(VS₄) was found in phase 1. Therefore, a multiphase Rietveld refinement using FullProf³⁸ based on the single-crystal structural model was carried out and led to very good reliability factors as shown in Figure 4 with a = 8.8483(1) Å and b = 17.4047(2) Å in the space group P6₃/m. The latter are very close to the one determined based on a single crystal. Rietveld quantification indicates the ratio 0.888/0.112 for the title phase Ba₇S(VOS₃)₂(S₂)₃/impurity. For the Se derivatives, the profile refinement led to the unit cell parameters a = 8.9234(1) Å and b = 17.5972(2) Å for phase 2, a = 8.9688(1) Å and b = 17.6818(2) Å for phase 3, and a = 8.9847(1) Å and b = 17.7086(2) Å for phase 4. For all phases, the unit cell parameters are consistent with the single crystal, which evidence the fact that the single-crystal compositions are close to the powder and nominal compositions. For all phases, the Rietveld refinement was then carried out using the corresponding single-crystal structure, which led to very good-quality refinements. Traces of impurity Ba₆(VO₂S₂)₂(VS₃O)(VS₄) might be present in the selenide derivatives, which are barely visible from the background and thus not taken into account for the refinements. We note that the compositions refined from single crystals for phase 2 and phase 4 have discrepancies with the nominal composition concerning the S/Se ratio, which are more significant. This might happen during the crystal growth process. Regarding the level of discrepancy and since the crystals found in the powder preparation are in a minor amount, the impact on the XRD powder pattern is difficult to observe as it would induce very small quantities of impurities.

Figure 4.

(a–d) Powder XRD Rietveld refinement of phases 1, 2, 3, and 4, respectively: the experimental (black) and calculated (red) patterns are superimposed; the difference curve and Bragg positions are represented in blue and green, respectively. The unit cell parameters refined in the space group P6₃/m are a = 8.8483(1) and b = 17.4047(2) Å for phase 1; a = 8.9234(1) Å and b = 17.5972(2) Å for phase 2; a = 8.9688(1) Å and b = 17.6818(2) Å for phase 3; and a = 8.9847(1) Å and b = 17.7086(2) Å for phase 4.

Optical Measurements

UV–visible diffuse-reflectance analysis of the polycrystalline phases (1 → 4) is represented in Figure 5. A Kubelka–Munk transformation³⁹ was applied to the measured diffuse-reflectance (R) spectra using the function F(R) = (1 – R)²/2R. Then, a Tauc plot^40,41 was used to determine the optical band gap E_g using the equation [F(R)hν]^1/n = k(hν – E_g), where hν is the photon energy, k is an energy-independent constant, E_g is the optical band gap, and n is an exponent related to the type of transition. Assuming an indirect transition (exponent n = 2), the plot of [F(R)hν]² versus hν is allowed, after drawing a tangent line at the inflection point, to determine the band gap as shown in Figure 4. This is consistent with the observed dark color of the crystals and powders. Figure S3 shows the band-gap evolution as a function of the Se content, and although in a narrow range, it decreases when the Se content increases as expected.

Figure 5.

Diffuse-reflectance spectra of (a) phase 1, (b) phase 2, (c) phase 3, and (d) phase 4 with a Tauc plot as an inset to determine the experimental band gap.

DFT Calculations Using the Meta-GGA SCAN Functional

To better understand the local geometry of selective and ordered substitution, and by consequence the origin of the experimental disorder, DFT calculations were carried out on four hypothetical phases: Ba₇S(VS₃O)₂(S₂)₃, Ba₇Se(VS₃O)₂(Se₂)₃, Ba₇S(VS₃O)₂(Se₂)-following selenide-deriv, and Ba₇Se(VS₃O)₂(S₂)₃. To allow straightforward comparison, the models were built starting from the oxysulfide unit cell in which we apply a selective substitution at the disulfide pair sites or the isolated sulfide site or both, yielding the above-mentioned models, which were fully relaxed (Tables S5–S8). The unit cells with the band structures are shown in Figure 6 and focus on the region around the Fermi level at the top of the valence band (VB) and the bottom of the conduction band (CB). The main crystallochemical information of the optimized geometries listed in Table 6 show constant distances within VOS₃ tetrahedra and Q–Q bonds considering all hypothetical phases. It highlights the rigidity of those entities that are not disturbed by the substitutions. On the other hand, we show that the Barium environment constitutes the “flexible part” and has to absorb the chemical pressure by adjusting the Ba–O and the Ba–Q distances when the complementary layer is substituted. This adaptation does not seem to be thermodynamically favorable in our experimental conditions. Instead, the systems stabilize with concomitant incorporation of selenide in both layers to avoid interlayer mismatch and constraints.

Figure 6.

Unit cell of the hypothetical phases and their band structure along high symmetry points of the Brillouin zone. Ba, V, O, S, and Se are represented in purple, black, red, yellow, and green, respectively.

Table 6. Main Geometrical Information Obtained from DFT Calculations for the Hypothetical Phases.

	Ba7S(VS3O) 2(S2)3	Ba7Se(VS3O)2(Se2)3	Ba7S(VS3O)2(Se2)3	Ba7Se(VS3O)2(S2)3
unit cell (Å)	a = 8.9088, c = 17.5491	a = 9.0677, c = 17.9483	a = 9.0391, c = 17.8805	a = 8.9359, c = 17.5970
volume (Å3)	1206.20	1278.05	1265.20	1216.88
VO3S (Td) distances (Å)	V–O = 1.6848	V–O = 1.6903	V–O = 1.6850	V–O = 1.6906
	V–S = 2.134	V–S = 2.1387	V–S = 2.1388	V–S = 2.1346
Ch–Ch (Å)	2.1141 (S2)	2.4334 (Se2)	2.4349 (Se2)	2.1138 (S2)
Ba–Ch(isolé) (Å)	3.3163	3.3798	3.3211	3.3679
Ba–Ch(paire) (Å)	3.2630/3.3025	3.4088/3.4547	3.4048/3.4423	3.2716/3.3134
Ba–Ch interlayer (Å)	3.7346	3.7938	3.7819	3.746
band gap (eV)	1.15 (indirect)	1.07 (indirect)	1.02 (indirect)	1.12 (direct)

The band structures show an indirect band gap for Ba₇S(VS₃O)₂(S₂)₃, Ba₇Se(VS₃O)₂(Se₂)₃, and Ba₇S(VS₃O)₂(Se₂)₃, and then, a transition to a direct band gap for Ba₇Se(VS₃O)₂(S₂)₃ is observed. The variation of the band-gap width is coherent (decreases with the increase of the Se/S ratio) except in the case of Ba₇S(VS₃O)₂(Se₂)₃ in comparison with Ba₇Se(VS₃O)₂(Se₂)₃, where the band gap slightly decreases by substitution of isolated Se by S. It can be explained by the positive chemical pressure induced by the smaller radius of sulfur leading to decreased Ba–Q_isolated distances as well as closer layers. These structural changes impact the band gap toward a decrease contrarily to sulfur incorporation, and the overall effects add up and lead to the slight decrease observed.

The total density of states (DOS) of each hypothetical phase is shown in Figure 7. The contribution of each type of anionic species is also represented to compare their contribution. For Ba₇S(VS₃O)₂(S₂)₃, within the VS₃O building units, the V states lie in the CB starting from ∼1.0 up to ∼3 eV for the major contribution to the CBM. In the VB, the V states’ contribution is found roughly in the range of −5 to −1 eV, where they are hybridized with the O 2p and S 3p states involved in the VS₃O building units. Owing to its more ionic bonding to vanadium and its higher electronegativity, O has a contribution of its 2p states more localized and lower in the VB (with the main large peak centered around −4.5 eV) compared to S_(VOS3) states that are found higher in the VB mainly above −4 eV. In a previous work (ref (23)), the projected DOS on distinct mixed-anion tetrahedra V(O,S)₄ as found in Ba₆(VO₂S₂)₂(VS₃O)(VS₄) allowed one to illustrate the evolution of their contributions higher in the VB when increasing the S content (from VO₃S → VS₄), thus contributing to decreasing the band gap.

Figure 7.

Projected density of states of the four hypothetical phases. The distinct contribution of each species is given: Ba; V, sulfur, and oxygen involved in a tetrahedral (VS₃O); Q (Q = S, Se) in isolated sites; and Q involved in a pair Q₂.

The disulfide pairs and the isolated S^2– contribute in a similar way as they are both found just below the Fermi level, above the broader S^2– states involved in the VOS₃ tetrahedra, with less localized states in the former due to the S–S covalent bonding. Regarding the electronic structure description and the crystallochemical situation of each anionic species, the title phase may be decomposed as a combination of the structural entities (Ba₇)¹⁴⁺(S^2–)(S₂)₃^6–((VS₃O)₂)^6–. A similar analysis can be made for all hypothetical phases.

Considering all hypothetical phases, the dichalcogenide pairs and the isolated chalcogenide states lie in a very close energy range (without interacting). However, the (Q)₂ pairs dominate except in the case of Ba₇Se(VS₃O)₂(S₂)₃, where the isolated Se is highly present at the top of the VBM.

Bader Charge Analysis

Bader charge analysis⁴² was carried out from the DFT calculated electronic structure based on the meta-GGA SCAN functional³⁰ that allows in particular a 20–50% band gap improvement⁴³ compared to the GGA functionals known to underestimate them and with no major additional computational cost compared to the costly hybrid functionals. We expect to obtain values closer to experience although here we focus on the comparison from one composition to another. Table 7 gathers the band-gap values and the net atomic Bader charges on the different types of atoms. For the Bader charges also, the comparative study is relevant, while the absolute values are not to be taken literally. We found Bader net charges consistent with those reported for comparable entities in the literature. For instance, (Se) values from a dichalcogenide pair are in the same range as reported in Sn(Se₂);⁴⁴ S^2– presents comparable values to that in FePS₃.⁴⁵ On the other hand, LiMnVO₄ exhibits for V⁵⁺ a slightly higher value of +1.80,⁴⁶ which is consistent with the fact that V⁵⁺ in our VO₃S thiovanadate involves one more covalent V–S bond contribution compared to the vanadate VO₄. The Bader charges might be another way to probe the impact of substituting one layer only, for instance, we compare Ba₇S(VS₃O)₂(S₂)₃ with Ba₇S(VS₃O)₂(Se₂)₃ and Ba₇Se(VS₃O)₂(Se₂)₃ with Ba₇Se(VS₃O)₂(S₂)₃.

Table 7. Net Atomic Bader Charges for the Four Hypothetical Phases.

net atomic Bader charges	Ba7S(VS3O)2(S2)3	Ba7Se(VS3O)2(Se2)3	Ba7S(VS3O)2(Se2)3	Ba7Se(VS3O)2(S2)3
Ba1	+1.49	+1.472	+1.476	+1.485
Ba2	+1.46	+1.442	+1.441	+1.464
V	+1.584	+1.588	+1.585	+1.583
S(VO3S)	–0.994	–1.007	–1.008	–0.994
Ch(isolated)	–1.319	–1.270	–1.338	–1.248
Ch1–Ch2(pair)	–0.752/–0.662	–0.701/–0.662	–0.700/–0.655	–0.695/–0.726
O	–1.019	–1.031	–1.014	–1.031

Overall, the changes are rather small, thus making difficult any interpretation; however, some species show a significant evolution, which may be analyzed as follows. Overall, variations of the atomic net Bader charge of ions involved in the VO₃S entities show less significant evolution in accordance with their rigid geometry through compositional change. Considering models with a fixed thiovanadate layer and containing S_(isolated)^2–, a slight augmentation (0.019) of the S_(isolated)^2– net Bader charge is observed when the nature of the complementary layer goes from (S₂) to (Se₂) pairs. This is due to the negative chemical pressure induced by the complementary layer, (S₂) → (Se₂), which enhances the ionic character of the Ba–S_(isolated) bonds. Oppositely, for Se_(isolated) in fixed thiovanadates layers, a decrease (0.022) is observed due to the positive chemical pressure from (Se₂) → (S₂) in the complementary layer yielding more covalent Ba–Se_(isolated) bonds. A similar analysis can be made for Se_(Se2), where the net Bader charge decreases through exchange of Se_isolated by S in the complementary layer, which induces a positive chemical pressure that should enhance covalent bonding. However, the counterintuitive S_(S2) net Bader charge decrease (more covalent bonding) through exchange of S_isolated by Se_isolated (negative chemical pressure) in the complementary layer indicates a more complicated scenario, which may be related to the band-gap evolution described above for this case.

Conclusions

A new series of thiovanadate polychalcogenide phases Ba₇V₂S₁₃O₂ and three selenide derivatives Ba₇V₂O₂S_9.304Se_3.696, Ba₇V₂O₂S_7.15Se_5.85, and Ba₇V₂O₂S_6.85Se_6.15 were elaborated, and their structures were solved in the space group P6₃/m (176). They represent original and complex multianionic manipulation examples. Ba₇V₂S₁₃O₂ is the sixth member in the quaternary system Ba–V–S–O (the second with disulfide pairs), while the three selenide derivatives are the first ones in the system Ba–V–Se–S–O. These phases exhibit mixed-anion building units VOS₃, isolated Q^2– (Q = S, Se) sites (present in one layer), and isolated dichalcogenide pairs Q₂ (in the second layer). We show that all attempts of a selective substitution of sulfur by selenide in our synthetic conditions led to concomitant incorporation of Se in both the isolated Q and Q₂ pairs sites with mixed S/Se occupancy. It indicates that two types of pairs S₂ and Se₂ are present in the selenide derivatives, which can be monitored by the mean Q–Q distance refined and which evolve with respect to the substitution level. Our meta-GGA DFT calculations, including Bader charge analysis, on the sulfide and three ordered selenide derivatives showed that the substitution in one layer only leads to local structural variations that are absorbed by the Ba sublattice, while VO₃S and Q₂ pairs remain rigid. This might explain the experimental disorder to avoid such constraints. Rationalization of the structure with respect to the electronic structure is interesting for designing new phases using such mixed-anion building blocks and/or dichalcogenide pairs. In particular, band-gap engineering through mastered approaches is important for a variety of properties.

Data Availability Statement

The data underlying this study are available in the published article and its Supporting Information.

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsorginorgau.3c00006.

• Data collection and refinement details of Ba₇V₂O₂S₁₃ synthesized at high pressures with a description of the synthesis (Table S1 and text below); atomic positions, isotropic thermal displacement, anisotropic thermal parameters, and main distances for Ba₇V₂O₂S₁₃ obtained with the high-pressure technique (Tables S2–S4); optimized structures information in Tables S5–S8; comparison between the oxysulfide phase and one of the selenide derivatives (Figure S1); precessions images (Figure S2); and band gap as a function of the selenide content (Figure S3) (PDF)

Author Contributions

CRediT: Batoul Almoussawi conceptualization (supporting), formal analysis (equal), investigation (lead), methodology (equal), project administration (equal), writing-original draft (lead), writing-review & editing (supporting); Hiroshi Kageyama conceptualization (supporting), funding acquisition (supporting), investigation (supporting), methodology (supporting), supervision (supporting), validation (equal), writing-review & editing (supporting); Pascal Roussel formal analysis (supporting), methodology (supporting), validation (supporting), writing-review & editing (supporting); Houria Kabbour conceptualization (lead), formal analysis (equal), funding acquisition (lead), investigation (supporting), methodology (equal), project administration (equal), supervision (lead), validation (equal), writing-original draft (supporting), writing-review & editing (lead).

The authors declare no competing financial interest.