Activation of anionic redox in d0 transition metal chalcogenides by anion doping

Abstract

Expanding the chemical space for designing novel anionic redox materials from oxides to sulfides has enabled to better apprehend fundamental aspects dealing with cationic-anionic relative band positioning. Pursuing with chalcogenides, but deviating from cationic substitution, we here present another twist to our band positioning strategy that relies on mixed ligands with the synthesis of the Li₂TiS_3-xSe_x solid solution series. Through the series the electrochemical activity displays a bell shape variation that peaks at 260 mAh/g for the composition x = 0.6 with barely no capacity for the x = 0 and x = 3 end members. We show that this capacity results from cumulated anionic (Se^2-/Se^n-) and (S^2-/S^n-) and cationic Ti³⁺/Ti⁴⁺ redox processes and provide evidence for a metal-ligand charge transfer by temperature-driven electron localization. Moreover, DFT calculations reveal that an anionic redox process cannot take place without the dynamic involvement of the transition metal electronic states. These insights can guide the rational synthesis of other Li-rich chalcogenides that are of interest for the development of solid-state batteries.

Fig. 1. Structural and electrochemical characterization of Li₂TiS_3 − xSe_x.

a PXRD patterns of synthesized materials in Li₂TiS_3 − xSe_x: a shift towards lower 2θ angles (higher d-spacing) is observed upon Se introduction. b Lattice parameters refined in $R\overline{3}m$ plotted against the Se content x in Li₂TiS_3 − xSe_x. c Voltage traces of the first two galvanostatic cycles. d Change of the differential capacity dQ/dE of the first oxidation process with Se doping. The lowering of the activation potential indicates successful manipulation of the anionic band. The decreased voltage is consistent with increased covalency of Ti–Se bonds compared to Ti–S bonds and with the lower chemical hardness of Se compared to S. e Gravimetric discharge capacity of the first discharge process (black) and the electrochemical activity (amount of removable Li upon the first oxidation) (red).

Fig. 2. Electrochemistry of Li₂TiS_2.4Se_0.6 and Li₂TiS_1.5Se_1.5.

a Voltage trace of Li₂TiS_2.4Se_0.6 during 25 cycles (discharge-voltage drops from 2.33 to 2.25 V), inset: dQ/dE vs E for the 1st, 2nd and 25th cycle. b GITT experiment of Li₂TiS_2.4Se_0.6. c GITT experiment of Li₂TiS_2.4Se_0.6, inset: comparison of discharge energy of Li₂TiS_2.4Se_0.6 with Li₂TiS_1.5Se_1.5 during the first 25 cycles.

Fig. 3. Crystal structure of Li_yTiS_2.4Se_0.6.

a Structure of Li₂TiS_2.4Se_0.6 shown along the stacking axis: Li and Ti occupy all octahedral interstices present in the cubic close packing of Ch²⁻ anions. Partial ordering of the cations leads to the formation of [Li]Ch₂ and [Li_1/3Ti_2/3]Ch₂ layers. b Representation of the mixed [Li_1/3Ti_2/3]Ch₂ slab: the partial ordering of Li vs Ti. c Evolution of the PXRD patterns of Li_yTiS_2.4Se_0.6 during the 1st cycle of an operando PXRD experiment. d Corresponding evolution of the lattice parameters of Li_yTiS_2.4Se_0.6 (space group: $R\overline{3}m$). e, f Local bonding environments around Ti1 of pristine Li₂TiS_2.4Se_0.6 and fully discharged Li₂TiS_2.4Se_0.6 (color code: blue: Ti, gray: Li, green: S, pink: Se).

Fig. 4. XPS/HAXPES experiments on Li_yTiS_3 − xSe_x.

a Binding energy difference ΔE(Ti–Ch) and the amount of partially Seⁿ⁻ as a function of x in pristine Li₂TiS_3 − xSe_x as deduced from XPS data. b HAXPES S 1s, S 2p/Se 3p and Ti 2p spectra of Li_yTiS_2.4Se_0.6 at different states of charge. c Proportion of reduced Ch²⁻ and partially oxidized Chⁿ⁻ in Li_yTiS_2.4Se_0.6 at different states of charge as deduced from HAXPES spectra.

Fig. 5. EPR results on Li₂TiS_3 − xSe_x.

a Temperature-dependent EPR spectra of Li₂TiSe₃ (b) evolution of the g-factor in Li₂TiS_3 − xSe_x at 110 K. c Main: temperature dependence of the g-factor of relithiated Li_yTiS_2.4Se_0.6. Inset: EPR spectra of pristine Li₂TiS_2.4Se_0.6 (black), delithiated Li_yTiS_2.4Se_0.6 (red) and relithiated Li_yTiS_2.4Se_0.6 (blue) at 5 and 110 K, respectively.

Fig. 6. Solid-state NMR on Li₂TiSe₃.

a Static ⁷Li spectra as a function of temperature. b ⁷Li longitudinal relaxation components R_1a and R_1b plotted against temperature.

Fig. 7. Theoretical calculations on Li_yTiS_3 − xSe_x.

a Projected density of states (pDOS) of Li₂TiS₃, Li₂TiSe₃, and Li₂TiS_1.5Se_1.5. b Fukui functions computed to probe the electronic states involved in the oxidation process for the Li₂TiS₃, Li₂TiSe₃, and mixed Li₂TiS_1.5Se_1.5 electrodes: these results clearly show that the holes generated upon oxidation (yellow volume) are mainly localized on the Ch (S and Se) anions for the pure-S and pure-Se electrodes, while the transition metal (Ti) is also involved for the mixed S/Se phases. Note that the Se selectivity in the mixed S/Se electrode is also clearly seen with a larger hole density around Se (pink atoms) much larger than around S (green atoms), therefore highlighting the Se selectivity in the mixed phases. c Atomic Bader net populations of the most oxidized S and Se atoms in the structure as a function of the Li content showing that Se is much more oxidized and S is much less oxidized in mixed S/Se compared to pure-S and pure-Se phases for an equivalent number of extracted lithium. All calculations were performed with the metaGGA SCAN functional (see “Methods” section).