Communication pubs.acs.org/JACS
Amorphous Metal Polysulfides: Electrode Materials with Unique Insertion/Extraction Reactions Atsushi Sakuda,*,†,# Koji Ohara,‡,§ Katsutoshi Fukuda,‡ Koji Nakanishi,‡ Tomoya Kawaguchi,‡ Hajime Arai,‡ Yoshiharu Uchimoto,∥ Toshiaki Ohta,⊥ Eiichiro Matsubara,‡ Zempachi Ogumi,‡ Toyoki Okumura,† Hironori Kobayashi,† Hiroyuki Kageyama,† Masahiro Shikano,† Hikari Sakaebe,† and Tomonari Takeuchi*,† †
Research Institute of Electrochemical Energy, Department of Energy and Environment, National Institute of Advanced Industrial Science and Technology (AIST), 1-8-31 Midorigaoka, Ikeda, Osaka 563-8577, Japan ‡ Office of Society-Academia Collaboration for Innovation, Kyoto University, Gokasho, Uji, Kyoto 611-0011, Japan § The Research & Utilization Division, Japan Synchrotron Radiation Research Institute (JASRI) 1-1-1 Kouto, Sayo, Hyogo 679-5198, Japan ∥ Graduate School of Human and Environmental Studies, Kyoto University, Nihonmatsu-cho, Yoshida, Sakyo-ku, Kyoto 606-8317, Japan ⊥ SR Center, Ritsumeikan University, 1-1-1 Noji-Higashi, Kusatsu, Shiga 525-8577, Japan S Supporting Information *
of these materials and, in particular, their structural changes, as doing so could lead to novel concepts for designing electrode materials with high energy densities. However, it has been difficult to determine their charge/discharge mechanism because of the difficulty in analyzing the unusual structure of amorphous metal polysulfides. Here, we report the charge/discharge mechanism of amorphous TiS4 as determined by combining advanced techniques such as X-ray pair distribution function (PDF) analysis and first-principles molecular dynamics calculations. We found that amorphous transition metal polysulfide electrodes exhibit an anomalous (dis)charge mechanism that is neither the typical intercalation/deintercalation mechanism nor the conversion-type mechanism, but a mixture of the two. Figure 1a shows the voltage versus capacity curves for various oxide and polysulfide electrode materials.1−8 The oxide-based electrodes show high voltages from 3.5 to 5.0 V. In contrast, the voltages of the metal polysulfide electrodes are almost half of those of the oxide electrodes. On the other hand, the capacities of some of the transition metal polysulfides are more than three times larger than those of the oxides. Among these materials, aTiS4 shows a large capacity.1,2 Figure 1b,c shows the initial charge/discharge curves of a-TiS4. In this study, a-TiS4 was discharged and charged for compositions between Li0TiS4 and Li4TiS4 by a 1-electron process. As shown, a-TiS4 showed a smooth and long plateau in the 2 V region, and the capacity corresponding to discharging to Li4TiS4 is 609 mAh g−1. The local structures of the samples labeled A−I in Figure 1b,c were characterized by X-ray PDF analysis, which is a powerful technique for investigating the structures of amorphous materials. Figure 2 shows the experimentally obtained X-ray PDF data (g(r)). The PDF of the as-prepared a-TiS4 (sample A) shows the characteristic shoulder and peaks
ABSTRACT: A unique charge/discharge mechanism of amorphous TiS4 is reported. Amorphous transition metal polysulfide electrodes exhibit anomalous charge/discharge performance and should have a unique charge/discharge mechanism: neither the typical intercalation/deintercalation mechanism nor the conversion-type one, but a mixture of the two. Analyzing the mechanism of such electrodes has been a challenge because fewer tools are available to examine the “amorphous” structure. It is revealed that the electrode undergoes two distinct structural changes: (i) the deformation and formation of S−S disulfide bonds and (ii) changes in the coordination number of titanium. These structural changes proceed continuously and concertedly for Li insertion/extraction. The results of this study provide a novel and unique model of amorphous electrode materials with significantly larger capacities.
C
urrently available lithium-ion battery technology is reaching its limit, while demands for increases in battery energy density remain strong. Thus, the development of new concepts for producing high-energy electrode materials is essential. We have recently reported novel materials such as amorphous titanium polysulfides (a-TiSx).1−4 These materials are expected to show high performances, exhibiting the advantages of both metal sulfide batteries and lithium−sulfur batteries. For example, a-TiS4 exhibits a high capacity of ca. 700 mA h g−1,2 and a-TiSx is semiconducting in nature. This is a significant advantage when designing batteries with high gravimetric and volumetric energy densities. Furthermore, polysulfide dissolution is significantly suppressed by the formation of metal−sulfur bonds in batteries based on amorphous polysulfides.2 These unique characteristics highlight the importance of elucidating the charge/discharge mechanism © 2017 American Chemical Society
Received: April 18, 2017 Published: June 16, 2017 8796
DOI: 10.1021/jacs.7b03909 J. Am. Chem. Soc. 2017, 139, 8796−8799
Communication
Journal of the American Chemical Society
any of the samples, suggesting that the amorphous structure is maintained during the charge/discharge reaction. Figure 2b shows the changes in the local structure during the initial discharging process. The as-prepared a-TiS4 sample contains a number of homopolar S−S bonds in its structure. The intensity of the peak of the S−S bonds decreases with discharging, suggesting that the covalent S−S bonds dissociate during the discharging process. The correlations attributable to the Ti−S and the second-neighbor S−S become shorter and longer, respectively, with discharging, suggesting that the coordination number of S around Ti decreases with discharging. This is reasonable because the ratio of positively charged atoms (Ti and Li) and negatively charged atoms (S) changes significantly, from 1:4 to 5:4, when a-TiS4 is discharged to a-Li4TiS4. The structural change that occurs during the discharging process is largely reversible as shown in Figure 2c. To elucidate the structure of a-TiS4 before and after the discharge/charge process, we developed an amorphous structure model for titanium polysulfides (TiS4, Li2TiS4, and Li4TiS4) by melt quenching using computer simulations, which were performed based on density functional theory-based molecular dynamics (DFT-MD). Figure 3a−c shows the structural models derived from the DFT-MD calculations. The structure factor values, S(Q), of the experimental and calculated models of a-TiS4 are compared in Supporting Information (Figure S1). The calculated amorphous structure model agreed well with the experimental one. Figure 3d−f shows the distribution of disulfides (S−S bonds with bond lengths smaller than 2.1 Å). The number of disulfides decreases with discharging. Note that the long-chain S−S bond is not observed even in a-TiS4. Figure 3g−i shows the relationship between the number of pairs and the distance, r, for the S−S, Ti−S, and Ti−Ti correlations. The average coordination numbers of Ti for aTiS4, a-Li2TiS4, and a-Li4TiS4 are ca. 6.9, 6.4, and 5.6, respectively. We found that, in the charged state, the amorphous TiS4 electrode contains the disulfide unit and that the coordination number of Ti is higher than 6 (it is close to 7 in the DFT-MD simulations). During the discharging process, the coordination number decreases to less than 6. This leads to an increase in the average length of the second-neighbor S−S correlation. The specific g(r) values of Ti−S, Ti−Ti, and S−S as well as the calculated g(r) values are shown in Figures S2 and S3. The peak shifts were consistent with the experimental results (Figure S3) although the distances and widths of the peaks in the simulated ones are somewhat different from those of the experimental ones in a strict sense. Figure 3j,k shows the local structure models of a-TiS4 and a-Li4TiS4. Discharging (i.e., an increase in the lithium content and changes of electronic state of Ti and S) induced structural changes, namely, the dissociation of the S−S bonds and a decrease in the coordination number of Ti. The Ti K-edge and S K-edge XANES spectra of a-TiS4 are presented in Figure S4. The electronic states for both Ti and S changed dramatically with discharging; and, after full charging, the electronic states were basically recovered reversibly. As shown above, the charge/discharge mechanism of a-TiS4 involves (i) the deformation and formation of S−S disulfide bonds and (ii) sharp changes in the coordination number of Ti. Although these changes are unique, the continuous change in the coordination number of Ti is a particularly distinct charge/ discharge mechanism. Furthermore, these two structural changes occur simultaneously. In the discharged state, the
Figure 1. (a) Voltage versus capacity for oxide and sulfide electrode materials. The dashed line in the figure shows the contour line of energy density calculated by the total weight of respective positive electrodes and lithium negative electrode. (b) Discharge/charge curves of a-TiS4 for investigation of structural change. (c) Discharge/charge curves with the number of lithium atoms per formula unit shown on the x-axis.
Figure 2. (a) Experimental X-ray radial pair distribution function (PDF) data for a-TiS4 with different state of charge, g(r) basically shows the probability of finding a pair of atoms, weighted by their scatter power, at a distance of r. (b) Change of PDF data during discharging. (c) Change of PDF data during charging.
at approximately 2.0, 2.4, and 3.4 Å. The shoulder observed at 2.0 Å is attributable to highly covalent homopolar S−S bonds. Further, the peak observed at ca. 2.4 Å is mainly related to the Ti−S bond, while the peak observed at 3.4 Å is mainly attributable to the second-neighbor S−S correlation. A significant middle-range order (>5 Å) was not observed in 8797
DOI: 10.1021/jacs.7b03909 J. Am. Chem. Soc. 2017, 139, 8796−8799
Communication
Journal of the American Chemical Society
materials with high capacities involving multielectron processes can be explained by a simple conversion mechanism.11 In this study, we directly observed the gradual and distinct structural change in an amorphous host; the formation of the expected conversion products such as Ti and Li2S was not observed in the XANES spectra and the XRD pattern in the discharged state (a-Li4TiS4) for the investigated voltage range in this study (Figures S4 and S5). These results indicate clearly that the charge/discharge mechanism is related to both a conversion reaction and Li insertion/extraction in the structure of the amorphous host in the composition range between TiS4 and Li4TiS4. It should be noted that, both the discharging and charging curves of a-TiS4 exhibited a single voltage plateau with a gradual slope for 4.0-electron processes. Further, the charge and discharge curves corresponded to a gradual change in the structure of the amorphous host. We analyzed the connectivity of the -Ti-S- network. The ratios of Ti atoms with more than 10 continuous -Ti-S- or -Ti-S-S- chains in a-TiS4, a-Li2TiS4, and aLi4TiS4 were 88, 82, and 62%, respectively. In a-LixTiS4 compounds with low Li contents, the structure is networklike, while in compounds with high Li contents, it is chain-like. This behavior was also confirmed by a Qn analysis, as shown in Figure S6. Although the coordination number of Ti decreased with an increase in the Li content, many Ti atoms in a-Li4TiS4 still formed -Ti-S- chains. This behavior arisen from the amorphous structure suppressed the formation of Li2S crystals. Thus, the reaction equation of a-TiS4 can be written as shown below: a‐Li 0TiS4 (with network‐like structure) + x Li+ + x e− ⇔ a‐LixTiS4 (with chain‐like structure)
(0 ≤ x ≤ ca. 4) (1)
Figure 3. (a−c) Structure models of (a) a-TiS4, (b) a-Li2TiS4, and (c) a-Li4TiS4 designed using the DFT-MD simulation. (d−f) Distribution of S−S bonding in (d) a-TiS4, (e) a-Li2TiS4, and (f) a-Li4TiS4 models. (g−i) Relationship between number of pairs and distance, r, for the S− S, Ti−S, and Ti−Ti correlations derived from (g) a-TiS4, (h) aLi2TiS4, and (i) a-Li4TiS4 models. The number of first-neighbor Ti−S represents the average coordination number of Ti. (j,k) Local structure models of (j) a-TiS4 and (k) a-Li4TiS4.
The discharge/charge curves of the initial 20 cycles of the cell using a-TiS4 are shown in Figure S7. The cell shows reversible discharging and charging although the capacity loss in some degree is observed. Batteries based on some transition metal polysulfides show long cycling lives in the all-solid-state batteries,7,8 suggesting that the reversibility of the metal polysulfide electrodes thereof is intrinsically high. The cell using a-TiS4 shows a large capacity of more than 500 mAh g−1 at the current density of 4 mA cm−2 (400 mA g−1, a-TiS4) (Figure S8). As described above, the (dis)charge mechanism involves two unique structural changes: (i) the dissociation/formation of S− S bonds and (ii) sharp changes in the coordination number of Ti. These two mechanisms can occur simultaneously, resulting in high-capacity charging and discharging. The related structural changes can occur continuously and concertedly. Designing electrodes based on the concept of a continuous change in the coordination number will be useful in the case of electrode materials with an amorphous phase. Metal polysulfide electrodes, which can be considered as representative of electrodes based on this concept, should be highly suited for use in next-generation batteries with significantly larger capacities. Lithium−sulfur batteries suffer from a few drawbacks, such as a low volumetric energy density, because they require a large amount of carbon and because the choice of potential electrolytes for these batteries is limited, owing to the adverse side reactions involved. On the other hand, metal polysulfides have several advantages compared to sulfur electrodes, such as (i) a high energy density because of their higher density and electronic conductivity and (ii) a large range
height of the pre-edge peak in the Ti K-edge XANES spectrum increased (Figure S4a,d). It has been reported that the symmetry around Ti strongly affects the pre-edge transition; the pre-edge peak is very weak when the coordination number of Ti is 6 and increases in intensity with a decrease in the Ti coordination number.9 Thus, the increase in the intensity of the pre-edge peak in the discharged state suggested that the coordination number of Ti became lower than 6. This result was consistent with the results of the PDF analysis and DFTMD simulations. The dramatic change in the discharge voltage in this composition range was likely owing to the large structural change. Titanium sulfide (IV), TiS2, is well known as the most transitional insertion electrode material.10 The TiS2 shows layered structure and Li+ ion inserts into and extracts from its host structure with 1-electron process at 2 V region. The charge compensation is usually explained by the change of the oxidation state of Ti: namely, Ti changes between Ti3+ and Ti4+. On the other hand, the contribution of the redox of S is very important in the metal polysulfide electrodes. The Mulliken charge analyses show that both Ti and S atoms become reduced and oxidized during discharging and charging (Table S1). The charge/discharge processes of many electrode 8798
DOI: 10.1021/jacs.7b03909 J. Am. Chem. Soc. 2017, 139, 8796−8799
Communication
Journal of the American Chemical Society Present Address
of potential electrolytes, because the polysulfides show low solubility in many organic solvents and because the undesirable side reactions are suppressed.2,4 The latter advantage is based on the formation of M−S bonds (M = metal) and the elimination of long S−S chains. By fully exploiting these advantages, it should be possible to develop novel rechargeable batteries. There are some pioneering works in amorphous metal polysulfide electrodes such as a-MoS3, including materials for primary batteries.12 Moreover, there are some reports on amorphous metal polysulfides with high S/M ratio, such as aMoSx (x = 5−6).13 Development of the novel high-capacity electrode materials combining these previous works is required. We anticipate that the findings of this study will help design novel electrode materials for innovative batteries. An overview of the methods is given below; details are provided in the Supporting Information. Materials and Electrochemical Cells. The a-TiS4 sample was prepared by a mechanochemical process.2 To prepare the a-TiS4 working electrode, acetylene black (AB) was kneaded into a-TiS4 using ball milling, and the obtained powder and polytetrafluoroethylene (PTFE) were mixed in an agate mortar.2 The weight ratio of a-TiS4, AB, and PTFE was 9:1:1. A 1 M solution of LiPF6 in a 50:50 (by volume) mixture of ethylene carbonate and dimethyl carbonate was used as the electrolyte. The electrochemical measurements were performed at 30 °C at a current density of 20 mA g−1 (a-TiS4). High-Energy X-ray Diffraction Measurements. The Xray total scattering measurements for the PDF analysis were carried out at SPring-8 BL28XU.14 The incident X-ray energy was 38.0 keV. The analyzed Q-range was 0.7−14 Å−1. The collected data sets were corrected for the absorption, background, and polarization effects. Details of the correction and normalization procedures have been reported elsewhere.15 DFT-MD Simulations. The first-principles-based simulations were performed using OpenMX.16 For the MD calculations, the O(N) Krylov-subspace method was used.17 The FOCUS supercomputer system was used. The model cells for a-TiS4, a-Li2TiS4, and a-Li4TiS4 contained 360, 448, and 576 atoms, respectively. The atoms in the cells were melted at 2000 K for 7000 MD steps and quenched from 2000 to 300 K for 7000 MD steps. The atomic models were described using VESTA18 or VMD19 software.
■
#
A.S.: Department of Applied Chemistry, Graduate School of Engineering, Osaka Prefecture University, 1-1, Gakuen-cho, Naka-ku, Sakai, Osaka 599-8531, Japan Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS This work was supported by the RISING and RISINGII projects of the New Energy and Industrial Technology Development Organization (NEDO), Japan.
■
ASSOCIATED CONTENT
S Supporting Information *
The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/jacs.7b03909.
■
REFERENCES
(1) Matsuyama, T.; Sakuda, A.; Hayashi, A.; Togawa, Y.; Mori, S.; Tatsumisago, M. J. Mater. Sci. 2012, 47, 6601−6606. (2) Sakuda, A.; Taguchi, N.; Takeuchi, T.; Kobayashi, H.; Sakaebe, H.; Tatsumi, K.; Ogumi, Z. Electrochem. Commun. 2013, 31, 71−75. (3) Hayashi, A.; Matsuyama, T.; Sakuda, A.; Tatsumisago, M. Chem. Lett. 2012, 41, 886−889. (4) Sakuda, A.; Taguchi, N.; Takeuchi, T.; Kobayashi, H.; Sakaebe, H.; Tatsumi, K.; Ogumi, Z. ECS Electrochem. Lett. 2014, 3, A79−A81. (5) Rout, C. S.; Kim, B.-H.; Xu, X.; Yang, J.; Jeong, H. Y.; Odkhuu, D.; Park, N.; Cho, J.; Shin, H. S. J. Am. Chem. Soc. 2013, 135, 8720− 8725. (6) Sakuda, A.; Takeuchi, T.; Okamura, K.; Kobayashi, H.; Sakaebe, H.; Tatsumi, K.; Ogumi, Z. Sci. Rep. 2015, 4, 4883. (7) Matsuyama, T.; Hayashi, A.; Ozaki, T.; Mori, S.; Tatsumisago, M. J. Mater. Chem. A 2015, 3, 14142−14147. (8) Sakuda, A.; Takeuchi, T.; Shikano, M.; Sakaebe, H.; Kobayashi, H. Front. Energy Res. 2016, 4, 4. (9) Farges, F.; Brown, G. E.; Rehr, J. J. Geochim. Cosmochim. Acta 1996, 60, 3023−3038. (10) Whittingham, M. S. Prog. Solid State Chem. 1978, 12, 41−99. (11) Poizot, P.; Laruelle, S.; Grugeon, S.; Dupont, L.; Tarascon, J.-M. Nature 2000, 407, 496−499. (12) Jacobson, A. J.; Chianelli, R. R.; Rich, S. M.; Whittingham, M. S. Mater. Res. Bull. 1979, 14, 1437−1448. (13) Afanasiev, P.; Bezverkhy, I. Chem. Mater. 2002, 14, 2826−2830. (14) Tanida, H.; Fukuda, K.; Murayama, H.; Orikasa, Y.; Arai, H.; Uchimoto, Y.; Matsubara, E.; Uruga, T.; Takeshita, K.; Takahashi, S.; Sano, M.; Aoyagi, H.; Watanabe, A.; Nariyama, N.; Ohashi, H.; Yumoto, H.; Koyama, T.; Senba, Y.; Takeuchi, T.; Furukawa, Y.; Ohata, T.; Matsushita, T.; Ishizawa, Y.; Kudo, T.; Kimura, H.; Yamazaki, H.; Tanaka, T.; Bizen, T.; Seike, T.; Goto, S.; Ohno, H.; Takata, M.; Kitamura, H.; Ishikawa, T.; Ohta, T.; Ogumi, Z. J. Synchrotron Radiat. 2014, 21, 268−272. (15) Kohara, S.; Itou, M.; Suzuya, K.; Inamura, Y.; Sakurai, Y.; Ohishi, Y.; Takata, M. J. Phys.: Condens. Matter 2007, 19, 506101. (16) Ozaki, T. OpenMX, http://www.openmx-square.org/, accessed March 2016. (17) Ozaki, T. Phys. Rev. B: Condens. Matter Mater. Phys. 2006, 74, 245101. (18) Momma, K.; Izumi, F. J. Appl. Crystallogr. 2011, 44, 1272−1276. (19) Humphrey, W.; Dalke, A.; Schulten, K. J. Mol. Graphics 1996, 14, 33−38.
Experimental details and anaylses, including Figures S1− S9 and Table S1 (PDF)
AUTHOR INFORMATION
Corresponding Authors
*
[email protected] *
[email protected] ORCID
Atsushi Sakuda: 0000-0002-9214-0347 Katsutoshi Fukuda: 0000-0002-7895-650X Hajime Arai: 0000-0001-6695-637X 8799
DOI: 10.1021/jacs.7b03909 J. Am. Chem. Soc. 2017, 139, 8796−8799
