Subscriber access provided by CORNELL UNIVERSITY LIBRARY
Article 2+
+
Systematic Investigation into Mg /Li Dual-Cation Transport in Chevrel Phases Using Computational and Experimental Approaches Jae-Hyun Cho, Jung-Hoon Ha, June Gunn Lee, Chang-Sam kim, Byung Won Cho, Kwang-Bum Kim, and Kyung Yoon Chung J. Phys. Chem. C, Just Accepted Manuscript • Publication Date (Web): 17 May 2017 Downloaded from http://pubs.acs.org on May 17, 2017
Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
The Journal of Physical Chemistry C is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.
Page 1 of 15
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
The Journal of Physical Chemistry
Systematic Investigation into Mg2+/Li+ Dual-Cation Transport in Chevrel Phases Using Computational and Experimental Approaches Jae-Hyun Cho,†,‡,¶ Jung Hoon Ha,‡,¶ June Gunn Lee,‡,¶ Chang-Sam Kim,‡ Byung Won Cho,‡ KwangBum Kim,**,† and Kyung Yoon Chung*,‡ †
Department of Materials Science and Engineering, Yonsei University, 50 Yonsei-ro, Seodaemun-gu, Seoul, 03722, Republic of Korea. ‡ Center for Energy Convergence Research, Korea Institute of Science and Technology, Hwarang-ro 14-gil 5, Seongbuk-gu, Seoul, 02792, Republic of Korea. ABSTRACT: Computational and experimental investigations into the Li+, Mg2+, and Mg2+/Li+ dual-cation transport properties within the Chevrel phase Mo6S8 have been performed. Five representative paths were selected for 3D diffusion, and their corresponding energy barriers were determined. Based on density functional theory calculation results, we reveal phenomena of the cation trapping, sluggishness of Mg2+ ion transport, and synchronized movement of inserted cations induced by repulsive interactions. The computational results were further validated by cyclic voltammetry carried out at ambient to high temperatures, from which apparent diffusion constants and activation energies for each case were determined. We found broad agreement between the theoretical and experimental results and suggest an optimum scenario for charge-discharge processes within the dual-cation hybrid system. issue, and it became clear that the Mg2+/Li+ dual-salt hybrid concept, which introduces fast Li+ ion insertion at the cathode by simply dissolving Li salts in the electrolyte, appears to be the most promising strategy. After the first report by Yagi et al. who adopted a LiFePO4 cathode,8 several materials such as TiS2, FeS2, Li4Ti5O12, MgCo2O4, S, Se, FeFe(CN)6, and LiMn2O4 were consecutively used with this new approach,9-16 including Mo6S8.17-18 While these studies demonstrate that the dual-salt concept is an appealing and highly feasible strategy for exploiting Mg anodes, most of the studies to date have simply focused on the enhancement of battery performance (i.e., the attainment of high energy densities). As a result, we still do not have a profound understanding of the underlying mechanisms of Mg2+/Li+ co-insertion kinetics and their transport properties, which is essential for the design of dualsalt hybrid systems.
1. INTRODUCTION As there are large demands for safe and cost-effective energy storage systems (ESSs), the development of battery types that employ polyvalent non-noble metal anodes (e.g., Mg, Al, Ca, and Zn) is of great interest. Since a prototype was first proposed by Aurbach et al. in the early 2000s, a rechargeable Mg battery that uses a Mg anode is regarded as one of the most promising candidates for ESSs because of several attractive advantages.1-2 First of all, compared with other polyvalent metal anodes, Mg has a relatively large theoretical capacity (gravimetric capacity: 2205 mAh g-1; volumetric capacity: 3833 mAh cm-3) and a negative standard reduction potential (-2.36 V, vs. standard hydrogen electrode, SHE), affording high energy density. Also, metallic Mg is environmentally benign, cost-competitive on account of its abundance in the Earth’s crust, and has guaranteed safety because it is possible to restrain dendrite formation by tuning electrolyte composition and concentration.2-4
The dual-salt hybrid battery can be classified into two major groups: the first group is the Daniel cell type, which only allows Li+ ion insertions at the cathode, due to the structural and electronic characteristics of the host material (e.g., LiFePO4). In this case, the battery uses two different kinds of cations at each electrode (i.e., Li+ at the cathode, and Mg2+ at the anode) as is the case in the classic Daniel cell (i.e., Cu2+ at the cathode, and Zn2+ at the anode) that has a similar working mechanism. However, this hybrid battery system does not involve a septum, unlike the Daniel cell, since the cathode material already acts as the “Li+ ion pass filter,” and hence, the battery fails to operate when Li salts in the electrolyte become depleted.8 The second group exploits both Li+ and Mg2+ ions as charge carriers (e.g., Mo6S8), and at this point in time, the activities of these two cations are understood to play salient roles in battery performance such as operating potential,
The Chevrel phase (CP) Mo6S8 system, with Mg2+ ion insertions, has been one of the primary focuses in the development of rechargeable Mg batteries.1 This system has two kinds of insertion sites for guest ions: “inner-” and “outer-ring” sites. These sites provide two distinct potential plateaus with discharge processes around 1.3 and 1.1 V (vs. Mg/Mg2+). Unfortunately, the inner-ring sites that occupy half of the entire capacity of the system are only partially available at room temperature (RT), and thus merely ~80 mAh g-1 capacities are operative out of a total theoretical capacity of 122 mAh g-1.5-6 Researchers immediately realized that the trapping of Mg2+ ions and their sluggish transport act as major factors that cause this problem.7 Various experimental and computational approaches have been developed in an attempt to resolve this
1
ACS Paragon Plus Environment
The Journal of Physical Chemistry
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
capacity, and rate capability.17 We previously demonstrated that thermodynamic changes occur with variations in the Mg2+/Li+ dual-cation activity, resulting in remarkable electrochemical performance induced by prevailing Li+ ion insertions.18
Page 2 of 15
The all-phenyl complex (APC) solution used as the electrolyte in this study was prepared by reacting 2 M PhMgCl in THF (Sigma-Aldrich) and anhydrous aluminum chloride (AlCl3, 99.999%, Sigma-Aldrich), with at a 2:1 molar ratio. After the reaction was completed, the APC solution was diluted to 0.4 M, calculated on a basis of the concentration of Mg2+ ions, by adding additional THF.31 The electrolytes for Mg hybrid batteries were composed of 0.2 M APC solution containing 0.8 M anhydrous lithium chloride (LiCl, 99%, SigmaAldrich) as the Li+ ion source.
Herein, we present a systematic investigation, using a combined computational and electrochemical approach, on the dual-cation transport behavior within the Mo6S8 structure, which is still the best framework because of its exceptional ability to accommodate Mg2+ insertions at RT. By carefully considering five possible representative diffusion pathways for each of the guest ion sites in Mo6S8, we quantified energy barriers for Li+, Mg2+, and Mg2+/Li+ dual-cation diffusion. Furthermore, these computational results were subsequently supported by cyclic voltammetry (CV) experiments performed at ambient to high-temperatures.
2.3. Cell Configuration and Electrochemical Measurements. For electrochemical experiments, 2032-type coin cells (Wellcos) were assembled in an Ar-filled glove box (MB200B, MBRAUN, less than 0.1 ppm of O2, and H2O). Mg alloy foil (3% Al and 1% Zn, Alfa Aesar) served as the counter electrode, which had an area twice that of the working electrode, to minimize polarization of the Mg depositionstripping processes, and to more accurately measure the potential at the working electrode. CV experiments were performed using a multi-channel potentiostat (VMP3, BioLogic) at various scan rates from 0.02 to 2 mV s-1 at ambient to high-temperatures, and every fourth CV profile was collected for the stabilized cell data set.
2. EXPERIMENTAL SECTION 2.1. DFT Computational Methods. Calculations were carried out within the density functional theory (DFT) framework using the Vienna ab initio Simulation Package (VASP) and MedeA-VASP.19-21 We relaxed electrons using projector augmented wave (PAW) potentials with the Perdew-BurkeErnzerhof (PBE) exchange-correlation energy parameterization,22-23 within the generalized gradient approximation (GGA).24-26 We adopted Mo_sv (4p6 4d5 5s1), S (3s2 3p4), Li_sv (1s1 2s1 2p1), and Mg_sv (2p6 3s2) as valence-electron potentials. An energy cutoff of 520 eV was used as well as the gamma-centered scheme that generated a 4 × 4 × 4 k-point mesh (k-spacing = 0.2 Å-1). We confirmed total energy convergences of less than 10−5 eV and optimized the geometric structure using the conjugate gradient method until the forces fell below 0.01 eV Å-1. We did not consider spin polarization, dipole correction, or strong-correlation since they do not affect energy difference calculations in this case. For the energy barrier determinations, we used the nudged elastic band (NEB) method using VASP transition state theory (VTST) with six or ten images, including the initial and final configurations, within structural convergence to -0.05 eV Å-1.27 The illustration related to these computational calculations were created by using the Visualization for Electronic and Structural Analysis (VESTA) program.28
3. RESULTS AND DISCUSSION 3.1. Model Systems. Figure 1 depicts the model systems selected in this study; the M36(Mo6S8)3 supercell, where M is the insertion site, is shown in Figure 1a, while Figure 1b shows the five representative pathways selected for the diffusion study involving foreign cations, such as Li+ and Mg2+ in (Mo6S8)3, through the insertion sites. In Figure 1a, we can see that each block consists of a cube of eight S atoms (yellow balls) that enclose an octahedron of six Mo atoms (violet balls). Note that, in Figure 1b, the green balls denote the inner-ring sites, and the red balls denote the outer-ring sites for guest ion insertions. Here, Path 1 represents in-innerring diffusion; Path 2 is an inner-ring to the next inner-ring diffusion that passes through two outer-ring sites; Path 3 is outer-ring diffusion between two facing outer-ring sites across from each other; Path 3i is a detour of Path 3; and Path 4 is outer-ring diffusion between two neighboring outer-rings via a short bridge (~1.2 Å). Note that among these pathways, only 2, 3, and 3i make 3D diffusion possible.
2.2. Preparation of the Cathode and Electrolyte. The host material CP Mo6S8 was synthesized through the molten salts method described earlier.29 More specific experimental sequences are outlined in the Supporting Information. The cathode was prepared by casting onto Ni foil (99.5%, Alfa Aesar) a homogenized slurry of Mo6S8 active material, acetylene black conductor (Denka Black, Denki Kagaku), sodium carboxymethyl cellulose thickening agent (CMC, average molecular weight: 90,000, Sigma-Aldrich), and styrene-butadiene rubber binder (SBR, 40 wt% in water, Zeon) with a weight ratio of 92:5:1:2. The electrode was then dried and roll-pressed. The amount of working electrode loaded was ~3 mg cm-2. We adopted the CMC/SBR binder system to achieve high coulombic efficiency, which could not be obtained with the polyvinylidene fluoride (PVDF, Solvay) binder. In the former study, tetrahydrofuran (THF) was reported to cause the swelling of PVDF binders as it permeates into the polymer.30 Therefore, it is recommended that PVDF binders be avoided with any electrolyte that uses THF as the solvent (for further details, see Figure S1 in the Supporting Information).
3.2. DFT Calculations of Li+ and Mg2+ Ion Transport Energy Barriers. Firstly, we examined the kinetic aspects of the LixMgy(Mo6S8)3 system (where x = 0-12, y = 0-6, and x + 2y ≤ 12) through DFT calculations. During the NEB study, all the initial structures were fully relaxed until the forces on each atom fell below 0.01 eV Å-1. We confirmed that the calculated lattice parameters were slightly larger, by up to 0.15 Å, than the corresponding X-ray diffraction (XRD) results from previous experiments;5, 7, 32-33 this is attributed to the overestimating nature of PBE potentials used in this study. In order to trace the trajectories of ions transported between the initial and final stages, the structural force convergence was eased to 0.05 eV Å-1 in order to reduce the computational cost. Preliminary calculations showed that this criterion is sufficient to provide an energy barrier with the accuracy of less than 0.01 eV. The energy barriers for these pathways are summarized in Table 1, which lists barriers that are always smaller for Li+, compared to those for Mg2+, for every
2
ACS Paragon Plus Environment
Page 3 of 15
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
The Journal of Physical Chemistry
pathway. In addition, Paths 2, 3, and 3i play crucial roles as controlling
Figure 1. The computational model. (a) The M36(Mo6S8)3 supercell, where M is the insertion site. Color-coding: Mo, violet; S, yellow. (b) Five representative pathways for 3D diffusion within M36(Mo6S8)3. Color-coding: inner-ring sites, green; outer-ring sites, red.
steps for 3D diffusion of the foreign cations. In contrast, for Path 1 (in-inner-ring diffusion), energy barriers are only 0.006-0.009 eV for a Li+ ion in the Li3(Mo6S8)3 system, and 0.04-0.13 eV for a Mg2+ ion in the Mg3(Mo6S8)3 system, when all three inner-ring sites are occupied. This suggests that any inner-ring diffusion is almost barrierless at RT, except that a Mg2+ ion may require a relatively small amount of agitation. In fact, either cation can readily move to new positions in the inner-ring, if necessary, to achieve a more stable configuration.
Table 1. Summary of calculated energy barriers (eV) for Li+ and Mg2+ ion transport. Path
Lix(Mo6S8)3
Mgx(Mo6S8)3
1 2 3 3i 4
0.006-0.009 0.43 0.21 0.25 0.08
0.04-0.13 0.51 0.77 0.32 N/A
We often noticed that the movement of a cation along Paths 2 or 3 is accompanied by spontaneous movement of another cation residing in the inner-ring to ease the electrostatic repulsion between inserted cations, especially when the depth of discharge (DOD) exceeds 1/3. In addition, the occupation of inner-ring sites by more than one cation is energetically unfavorable; any additional cation is readily ejected to an outerring site. Path 4 represents the diffusion between two outerring sites (two red balls connected by Path 4 in Figure 1b). We found an energy barrier of only 0.08 eV for this process within the Li5(Mo6S8)3 system as shown in Figure S2. In this case, the Li+ ions in the inner-rings do not result in ring expansion; consequently, the two outer-ring sites that lie across Path 4 remain as two individual local minima. However, when the Mg4(Mo6S8)3 system finds itself in the analogous situation, the
Mg2+ ions in the inner-rings promote ring expansion; as a consequence, the two outer-ring sites that lie across Path 4 become closer (from ~1.2 Å to less than ~0.5 Å) resulting in a single local minimum. Therefore, it is not possible to identify Path 4 in the case of Mg2+, and the diffusion of a Mg2+ ion between two neighboring outer-rings cannot follow this path. Figures 2a-c depict the DFT-calculated energy profiles for Li+ and Mg2+ ion transport, while Figures 2d-f show images of the corresponding trajectories (for further details, see Figures S3-S5 in the Supporting Information). At the very early stages of discharge, when all inner-ring sites are empty, we found that the inner-ring sites are the only minima available for insertion. This indicates that there are no local minima at outerring sites, meaning that any foreign atom will fall spontaneously into the empty inner-ring. Once an ion occupies an inner-ring site, it requires considerable energy to escape (0.43 eV for Li+, and 0.51 eV for Mg2+) due to relatively deep potential minima at the inner-ring sites, as shown in Figure 2a. This explains the trapping problem (i.e., the cyclical motion of a cation within the same inner-ring) often encountered experimentally upon the last stage of charging processes at RT, especially for the MgxMo6S8 system.34 Path 2 in Figure 2d is the trajectory adopted by an inserted cation as it overcomes these large energy barriers, and involves movement to another inner-ring site. In the Mo6S8 system, there are six inner-ring and six outerring sites as shown in Figure 1b, but in order to maintain local electroneutrality within the host material after insertion of foreign ions, the maximum number of electrons is limited to four per formula unit (fu) (i.e., four Li or two Mg fu-1, which can be formulated to be twelve Li or six Mg per M36(Mo6S8)3 supercell).35 For Path 3, based on the above discussion, we assume that all inner-rings are initially occupied by three cations, with subsequent movement of one cation from an outer-ring site to another across from it. Figures 2b and 2e display the energy profiles of Path 3 for Li4(Mo6S8)3 and Mg4(Mo6S8)3 systems, and their trajectories, respectively. Owing to the double charges of the Mg2+ ion, it has to overcome a higher energy barrier (0.77 eV) because of interactions
3 Environment ACS Paragon Plus
The Journal of Physical Chemistry
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 4 of 15
from one side, and an existing ion exits from the other side, completing net transport. We applied this observation to each model Li3(Mo6S8)3 and LiMg2(Mo6S8)3 system, as shown in Figures S5 and 2f, respectively. In these cases, Path 3i was extended all the way to the next inner-ring in order to observe the barriers around the outer-ring sites through NEB calculations. Furthermore, for calculational convenience, we only
Figure 2. Calculated energy profiles and barriers: (a) Path 2 for G1(Mo6S8)3 (G ≡ Li or Mg); (b) Path 3 for G4(Mo6S8)3; and (c) Path 3i for Li3(Mo6S8)3 and Li1Mg2(Mo6S8)3 systems. Guest ion trajectories (from position i to f): (d) Path 2 (scarlet ball, Li or Mg) for G1(Mo6S8)3; (e) Path 3 for G4(Mo6S8)3 (for further details of the differences between the Li+ and Mg2+ ion trajectories, see Figures S3 and S4); and (f) Path 3i for the Li1Mg2(Mo6S8)3 system (green ball, Li; orange ball, Mg). The trajectory within the Li3(Mo6S8)3 system is provided in Figure S5 in the Supporting Information.
considered cation movements out of the inner-ring, without accounting for another cation entering simultaneously. This means that the calculated energy barriers are the upper limits since cations entering simultaneously assist the departure of an existing inner-ring ion. The calculated energy barriers are 0.25 and 0.32 eV for Li+ and Mg2+ ions, respectively, as shown in Figure 2c. These results imply both Paths 3 and 3i are available for Li+ ions, while only Path 3i is feasible as the Mg2+ ion diffusion path. 3.3. Determination of Dapp and Ea via CV. Based on the computational results presented above, we performed electrochemical analyses to verify the correspondence between theoretical and experimental approaches. CV was employed for this study as it is a facile and accurate way of determining apparent diffusivity (Dapp) and activation energy (Ea). In Figures 3a and 3b, voltammograms at different scan rates for Mg2+ insertion/deinsertion (Mg battery) and Mg2+/Li+ coinsertion/deinsertion (Mg hybrid battery) are displayed. CV Peaks a/a’ match those of guest ion insertion/deinsertion for inner-ring sites, while Peaks b/b’ correspond to outer-ring sites. In the Mg hybrid battery system, where Li+ ions predominantly insert/deinsert into Mo6S8, there are two distinct peaks (Peaks b/b’ and c/c’) for the outer-ring sites for
each charge and discharge processes, which signifies that the electrochemical reactions associated with the outer-ring sites are divided into two steps in this case.7 In Figure 3a, the Mg2+ insertion/deinsertion trapping problem near RT is observed as addressed by Path 2; hence, only one first-order phase transition (Peaks b/b’) is found. In order to reveal two distinct first-order phase transitions, the energy barrier for Mg2+ entrapment needs to be overcome; consequently, CV measurements need to be carried out at elevated temperatures (above 42.5 oC in this paper; for details see Figures S6 and S7 in the Supporting Information). In the case of the hybrid system, however, three distinct first-order phase transitions already exist (Peaks a/a’, b/b’, and c/c’) at RT, implying that the introduction of Li+ ions results in significant kinetic improvement (Figures 3b and S8).7, 17-18 Figure 3c shows the temperature dependency of the insertion/deinsertion potentials of the Mg hybrid battery. Clearly, Peaks a’ and b/b’ are shifted toward negative potentials as the temperature rises, and both guest ions find more energetically favorable compositions in the Mo6S8 host. We regard these results to originate from increases in the proportion of co-inserting/deinserting Mg2+ ions at elevated temperatures, as their activities are raised dramatically. In our
4 Environment ACS Paragon Plus
Page 5 of 15
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
The Journal of Physical Chemistry
prior report, we unraveled the relationship between the activities of dual-inserting cations and potential variations in Mo6S8, which may well explain this phenomenon.18 To obtain Dapp values through the Randles-Sevcik equation, we linearly fitted plots of normalized peak current density against the square root of scan rate (Figures 3d and 3e; more detail is provided in the Supporting Information). It can be observed that only Peaks b/b’ of the Mg battery (Figure 3a), and Peaks a/a’ and b/b’ of the hybrid battery systems (Figure 3b) have good symmetric relationships between cathodic and anodic current density among all of the redox couples. Based on these results, we can ascertain the reversibility of reactions related to these peaks.36 However, the peak current densities corresponding to Peaks c/c’ in Figure 3b do not proportionally
Figure 3. Normalized CV profiles at RT at various scan rates: (a) Mg2+ insertion/deinsertion (Mg battery); (b) Mg2+/Li+ coinsertion/deinsertion (Mg hybrid battery); and (c) the temperature dependency of the Mg hybrid battery’s peak potential at a scan rate of 0.05 mV s-1. Linear fitting of normalized peak current density against square root of scan rate: (d) Mg2+ insertion/deinsertion; (e) Mg2+/Li+ co-insertion/deinsertion corresponding to Peaks a/a’ and b/b’; and (f) Mg2+/Li+ co-insertion/deinsertion corresponding to Peaks c/c’, in which quasi-reversible behavior is observed.
Table 2. Dapp (cm2 s-1) for Mg2+ insertion/deinsertion into CP Mo6S8 at different temperatures. Temp. (oC)
Peak a
Peak a’
Peak b
Peak b’
25 42.5 60
2.3 × 10-16 8.4 × 10-16 1.8 × 10-15
1.7 × 10-17 9.4 × 10-17 3.4 × 10-16
1.4 × 10-15 3.6 × 10-15 5.8 × 10-15
1.5 × 10-15 2.9 × 10-15 5.1 × 10-15
Table 3. Dapp (cm2 s-1) for Mg2+/Li+ co-insertion/deinsertion into CP Mo6S8 at different temperatures. Temp. (oC) 25 42.5 60
Peak a -14
1.3 × 10 2.7 × 10-14 3.2 × 10-14
Peak a’ -15
9.0 × 10 2.2 × 10-14 3.4 × 10-14
Peak b -14
6.0 × 10 1.1 × 10-13 1.2 × 10-13
Peak b’ -14
6.9 × 10 1.3 × 10-13 1.6 × 10-13
5 Environment ACS Paragon Plus
Peak c -15
3.1 × 10 8.4 × 10-15 2.9 × 10-14
Peak c’ 2.5 × 10-16 1.9 × 10-15 5.2 × 10-15
The Journal of Physical Chemistry
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
increase with the square root of scan rate, but increase in direct proportion to the first order of scan rate (Figure 3f), which indicates that the reaction associated with Peaks c/c’ is quasireversible.37 The quasi-reversible nature of this reaction implies that it is not only limited by diffusion rate but is also controlled by charge transfer rate. This can be explained by the loss of electrical conductivity as the concentration of guest ions increases within the Mo6S8 host, which originally had metallic properties. DFT calculations from prior research proved that the Fermi level (EF) shifts toward the band gap as the insertion level increases within Mo6S8, which eventually exhibits semiconductor behavior when it becomes fully inserted.35, 38 As we were unable to find an accurate way of calculating diffusivities for the quasi-reversible reaction, therefore, the Dapp related to Peaks c/c’ were determined based on the results obtained by plotting the normalized peak current density against the square root of scan rate. It should be noted that we excluded some peak current densities at high scan rates (at 0.2 and 0.5 mV s-1) because they show serious deviations from the linear line of best fit (Figure S9). The Dapp values obtained by CV are listed in Tables 2 and 3. Compared to the conventional Mg battery, Dapp values are enhanced by about two to three orders or magnitude higher at the inner-ring site, and one to two orders improved at the outer-ring site at around RT by introducing the hybrid battery system. However, as mentioned above, the reactions associated with Peaks c/c’ of the Mg hybrid battery show rather low Dapp values, similar to those of Mg2+ ions. Note that the values suggested in this work vary with experimental conditions (e.g., electrode loading, electrode composition, electrolyte viscosity, and electrolyte conductivity).36 We determined the Ea of each peak by plotting Dapp versus inverse operating temperature (Arrhenius plots, Figures 4 and S10); the calculated Ea values are listed in Table 4. Ea is substantially improved by replacing Mg2+ insertion/deinsertion with those of Li+ ions, especially at the inner-ring sites, in good agreement with the guidelines obtained computationally. However, the reactions related to Peaks c/c’ still have high Ea values resulting from wide variations in Dapp over the experimental temperature range. The substantial improvement in the Dapp values of Peaks c/c’ at mid- to high-temperatures might be the result of the increased insertion ratio of Mg2+ ions at outer-ring sites as their activities are considerably raised. Therefore, with this in mind, we suggest that the Mg2+/Li+ co-insertion/deinsertion at this region facilitate guest ion transport within the Mo6S8 structure.
Page 6 of 15
Figure 4. Arrhenius plots of Dapp for discharge processes. Ea values for Mg2+ insertions (blue) and Mg2+/Li+ co-insertions (red) are derived from the slopes of these plots.
Table 4. Summary of calculated Ea (eV mol-1) for Mg2+ and Mg2+/Li+ diffusions. Peak a/a’ b/b’ c/c’
Mg2+ diffusion
Mg2+/Li+ diffusion
Cathodic
Anodic
Cathodic
Anodic
0.60 0.34
0.73 0.31
0.24 0.18 0.55
0.32 0.21 0.75
N/A
3.4. Idealistic Mg2+/Li+ Dual-Cation Transport Scenario. Founded on the theoretical and practical aspects presented above, we recommend the ideal Mg2+/Li+ dual-cation insertion scenario for the Mo6S8 system. (i) Initially, Li+ ions predominantly diffuse into the system and fall into the inner-rings if the proper concentration of Li+ ions in the electrolyte is provided. (ii) Thereafter, as the concentration of Li+ ions in the electrolyte decreases, both Li+ and Mg2+ ions diffuse into the host and occupy outer-ring sites. (iii) Finally, as the Li+ ions in the electrolyte are almost completely consumed, Mg2+ ions alone occupy the remaining outer-ring sites. This scheme discourages any insertion of Mg2+ cations into the inner-ring site, and as a consequence resolves the trapping issue. We expect faster transport of Mg2+ ions in steps (ii) and (iii) since Li+ ions reside in inner-rings and, hence, Mg2+ ions are less tightly bound to surrounding atoms. In addition, it should be stressed that the reverse process (i.e., charging) does not involve any trapping problem since all inner-ring sites are filled by fastmoving Li+ ions.
4. CONCLUSIONS We have combined computational and electrochemical methods to investigate guest ion diffusion behavior within Mo6S8. To calculate the energy barriers for Li+ and Mg2+ ion transport, we considered five representative pathways. The calculated energy barriers for each pathway was always lower for Li+ than that for Mg2+, implying that the sluggish transport properties, as well as the trapping problem ascribed to the insertion/deinsertion of Mg2+ cations, can be resolved by introducing the dual-salt concept. Although we did not find a
6 Environment ACS Paragon Plus
Page 7 of 15
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
The Journal of Physical Chemistry
strict match between the results from theoretical and empirical approaches, because of a number of unavoidable experimental factors, the cyclic voltammetry experiments broadly show similar diffusion properties to those obtained computationally. Our theory-aided investigation provides a profound understanding of the dual-salt hybrid battery system, especially its diffusion kinetics, and further suggests guidelines for achieving optimal compositions for Mg2+/Li+ co-insertion/deinsertion chemistry.
ASSOCIATED CONTENT Supporting Information The Supporting Information is available free of charge via the Internet at http://pubs.acs.org. Experimental procedures; details of measurements and calculations; Supplementary Figures S1–S10; Supplementary references 1–3.
AUTHOR INFORMATION Corresponding Authors *E-mail:
[email protected]. Phone: +82-2-958-5225. Fax: +822-958-5229. **E-mail:
[email protected]. Phone: +82-2-365-7745. Fax: +82-2-312-5375.
Author Contributions ¶ These authors contributed equally. Notes The authors declare no competing financial interest.
ACKNOWLEDGMENTS This work was supported by the Energy Efficiency & Resources Program of the Korea Institute of Energy Technology Evaluation and Planning (Project No. 20112010100140) grant funded by the Korean government Ministry of Trade, Industry & Energy and the KIST institutional program (Project No. 2E26292). We also would like to acknowledge the support of Materials Design (www.materialsdesign.com/medea) and Kyungwon Enc. (www.kwenc.kr) for granting us the use of the MedeA-VASP program.
7
ACS Paragon Plus Environment
The Journal of Physical Chemistry
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 8 of 15
8266. (16) Pan, B.; Feng, Z.; Sa, N.; Han, S. D.; Ma, Q.; Fenter, P.; Vaughey, J. T.; Zhang, Z.; Liao, C. Advanced Hybrid Battery with a Magnesium Metal Anode and a Spinel LiMn2O4 Cathode. Chem. Commun. 2016, 52, 9961-9964. (17) Cheng, Y.; Shao, Y.; Zhang, J.-G.; Sprenkle, V. L.; Liu, J.; Li, G. High Performance Batteries Based on Hybrid Magnesium and Lithium Chemistry. Chem. Commun. 2014, 50, 9644-9646. (18) Cho, J.-H.; Aykol, M.; Kim, S.; Ha, J.-H.; Wolverton, C.; Chung, K. Y.; Kim, K.-B.; Cho, B. W. Controlling the Intercalation Chemistry to Design High-Performance DualSalt Hybrid Rechargeable Batteries. J. Am. Chem. Soc. 2014, 136, 16116-16119. (19) Kohn, W.; Sham, L. J. Self-Consistent Equations Including Exchange and Correlation Effects. Phys. Rev. 1965, 140, A1133-A1138. (20) Kresse, G.; Furthmüller, J. Efficient Iterative Schemes for ab initio Total-Energy Calculations Using a Plane-Wave Basis Set. Phys. Rev. B 1996, 54, 11169-11186. (21) Materials Design/MedeA® VASP Program. http://www.materialsdesign.com/medea (accessed May 10, 2017). (22) Blöchl, P. E. Projector Augmented-Wave Method. Phys. Rev. B 1994, 50, 17935-17979. (23) Ernzerhof, M.; Scuseria, G. E. Assessment of the Perdew–Burke–Ernzerhof Exchange-Correlation Functional. J. Chem. Phys. 1999, 110, 5029-5036. (24) Perdew, J. P.; Chevary, J. A.; Vosko, S. H.; Jackson, K. A.; Pederson, M. R.; Singh, D. J.; Fiolhais, C. Atoms, Molecules, Solids, and Surfaces: Applications of the Generalized Gradient Approximation for Exchange and Correlation. Phys. Rev. B 1992, 46, 6671-6687. (25) Perdew, J. P.; Burke, K.; Ernzerhof, M. Generalized Gradient Approximation Made Simple. Phys. Rev. Lett. 1996, 77, 3865-3868. (26) Becke, A. D. Density-Functional Thermochemistry. IV. A New Dynamical Correlation Functional and Implications for Exact-Exchange Mixing. J. Chem. Phys. 1996, 104, 1040-1046. (27) VTST. http://theory.cm.utexas.edu/vtsttools (accessed May 10, 2017). (28) Momma, K.; Izumi, F. VESTA: A Three-Dimensional Visualization System for Electronic and Structural Analysis. J. Appl. Cryst. 2008, 41, 653-658. (29) Lancry, E.; Levi, E.; Mitelman, A.; Malovany, S.; Aurbach, D. Molten Salt Synthesis (MSS) of Cu2Mo6S8— New Way for Large-Scale Production of Chevrel Phases. J. Solid State Chem. 2006, 179, 1879-1882. (30) Munari, S.; Bottino, A.; Capannelli, G. Casting and Performance of Polyvinylidene Fluoride Based Membranes. J. Membr. Sci. 1983, 16, 181-193. (31) Mizrahi, O.; Amir, N.; Pollak, E.; Chusid, O.; Marks, V.; Gottlieb, H.; Larush, L.; Zinigrad, E.; Aurbach, D. Electrolyte Solutions with a Wide Electrochemical Window for Rechargeable Magnesium Batteries. J. Electrochem. Soc. 2008, 155, A103-A109. (32) Ritter, C.; Gocke, E.; Fischer, C.; Schöllhorn, R. Neutron Diffraction Study on the Crystal Structure of Lithium Intercalated Chevrel Phases. Mat. Res. Bull. 1992, 27, 1217-1225.
REFERENCES (1) Aurbach, D.; Lu, Z.; Schechter, A.; Gofer, Y.; Gizbar, H.;
Turgeman, R.; Cohen, Y.; Moshkovich, M.; Levi, E. Prototype Systems for Rechargeable Magnesium Batteries. Nature 2000, 407, 724-727. (2) Yoo, H. D.; Shterenberg, I.; Gofer, Y.; Gershinsky, G.; Pour, N.; Aurbach, D. Mg Rechargeable Batteries: An onGoing Challenge. Energy Environ. Sci. 2013, 6, 2265-2279. (3) Muldoon, J.; Bucur, C. B.; Gregory, T. Quest for Nonaqueous Multivalent Secondary Batteries: Magnesium and Beyond. Chem. Rev. 2014, 114, 11683-11720. (4) Gregory, T. D.; Hoffman, R. J.; Winterton, R. C. Nonaqueous Electrochemistry of Magnesium Applications to Energy Storage. J. Electrochem. Soc. 1990, 137, 775-780. (5) Levi, E.; Lancry, E.; Mitelman, A.; Aurbach, D.; Ceder, G.; Morgan, D.; Isnard, O. Phase Diagram of Mg Insertion into Chevrel Phases, MgxMo6T8 (T = S, Se). 1. Crystal Structure of the Sulfides. Chem. Mater. 2006, 18, 54925503. (6) Levi, E.; Lancry, E.; Mitelman, A.; Aurbach, D.; Isnard, O.; Djurado, D. Phase Diagram of Mg Insertion into Chevrel Phases, MgxMo6T8 (T = S, Se). 2. The Crystal Structure of Triclinic MgMo6Se8. Chem. Mater. 2006, 18, 3705-3714. (7) Levi, M. D.; Lancry, E.; Gizbar, H.; Lu, Z.; Levi, E.; Gofer, Y.; Aurbach, D. Kinetic and Thermodynamic Studies of Mg2+ and Li+ Ion Insertion into the Mo6S8 Chevrel Phase. J. Electrochem. Soc. 2004, 151, A1044-A1051. (8) Yagi, S.; Ichitsubo, T.; Shirai, Y.; Yanai, S.; Doi, T.; Murase, K.; Matsubara, E. A Concept of Dual-Salt Polyvalent-Metal Storage Battery. J. Mater. Chem. A 2014, 2, 1144-1149. (9) Gao, T.; Han, F.; Zhu, Y.; Suo, L.; Luo, C.; Xu, K.; Wang, C. Hybrid Mg2+/Li+ Battery with Long Cycle Life and High Rate Capability. Adv. Energy Mater. 2014, 5, 1401507. (10) Zhang, Y.; Xie, J.; Han, Y.; Li, C. Dual-Salt Mg-Based Batteries with Conversion Cathodes. Adv. Funct. Mater. 2015, 25, 7300. (11) Wu, N.; Yang, Z.-Z.; Yao, H.-R.; Yin, Y.-X.; Gu, L.; Guo, Y.-G. Improving the Electrochemical Performance of the Li4Ti5O12 Electrode in a Rechargeable Magnesium Battery by Lithium-Magnesium Co-Intercalation. Angew. Chem. 2015, 127, 5849-5853. (12) Ichitsubo, T.; Okamoto, S.; Kawaguchi, T.; Kumagai, Y.; Oba, F.; Yagi, S.; Goto, N.; Doi, T.; Matsubara, E. Toward “Rocking-Chair Type” Mg–Li Dual-Salt Batteries. J. Mater. Chem. A 2015, 3, 10188-10194. (13) Gao, T.; Noked, M.; Pearse, A. J.; Gillette, E.; Fan, X.; Zhu, Y.; Luo, C.; Suo, L.; Schroeder, M. A.; Xu, K., et al. Enhancing the Reversibility of Mg/S Battery Chemistry through Li+ Mediation. J. Am. Chem. Soc. 2015, 137, 12388-12393. (14) Zhao-Karger, Z.; Lin, X.-M.; Bonatto Minella, C.; Wang, D.; Diemant, T.; Behm, R. J.; Fichtner, M. Selenium and Selenium-Sulfur Cathode Materials for High-Energy Rechargeable Magnesium Batteries. J. Power Sources 2016, 323, 213-219. (15) Dong, H.; Li, Y.; Liang, Y.; Li, G.; Sun, C. J.; Ren, Y.; Lu, Y.; Yao, Y. A Magnesium-Sodium Hybrid Battery with High Operating Voltage. Chem. Commun. 2016, 52, 8263-
8
ACS Paragon Plus Environment
Page 9 of 15
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
The Journal of Physical Chemistry
(33) Aurbach, D.; Suresh, G. S.; Levi, E.; Mitelman, A.; Mizrahi, O.; Chusid, O.; Brunelli, M. Progress in Rechargeable Magnesium Battery Technology. Adv. Mater. 2007, 19, 4260-4267. (34) Levi, E.; Levi, M. D.; Chasid, O.; Aurbach, D. New Insight on the Unusually High Ionic Mobility in Chevrel Phases. Chem. Mater. 2009, 21, 1390-1399. (35) Kaewmaraya, T.; Ramzan, M.; Osorio-Guillén, J. M.; Ahuja, R. Electronic Structure and Ionic Diffusion of Green Battery Cathode Material: Mg2Mo6S8. Solid State Ionics 2014, 261, 17-20. (36) Yu, D. Y. W.; Fietzek, C.; Weydanz, W.; Donoue, K.; Inoue, T.; Kurokawa, H.; Fujitani, S. Study of LiFePO4 by Cyclic Voltammetry. J. Electrochem. Soc. 2007, 154, A253A257. (37) Takahashi, M.; Tobishima, S.-i.; Takei, K.; Sakurai, Y. Reaction Behavior of LiFePO4 as a Cathode Material for Rechargeable Lithium Batteries. Solid State Ionics 2002, 148, 283-289. (38) Saha, P.; Datta, M. K.; Velikokhatnyi, O. I.; Manivannan, A.; Alman, D.; Kumta, P. N. Rechargeable Magnesium Battery: Current Status and Key Challenges for the Future. Prog. Mater. Sci. 2014, 66, 1-86.
9
ACS Paragon Plus Environment
The Journal of Physical Chemistry
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
TOC Graphic
10 Environment ACS Paragon Plus
Page 10 of 15
Page 11 of 15
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
The Journal of Physical Chemistry
Figure 1. The computational model. (a) The M36(Mo6S8)3 supercell, where M is the insertion site. Colorcoding: Mo, violet; S, yellow. (b) Five representative pathways for 3D diffusion within M36(Mo6S8)3. Colorcoding: inner-ring sites, green; outer-ring sites, red. 211x75mm (300 x 300 DPI)
ACS Paragon Plus Environment
The Journal of Physical Chemistry
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Figure 2. Calculated energy profiles and barriers: (a) Path 2 for G1(Mo6S8)3 (G ≡ Li or Mg); (b) Path 3 for G4(Mo6S8)3; and (c) Path 3i for Li3(Mo6S8)3 and Li1Mg2(Mo6S8)3 systems. Guest ion trajectories (from position i to f): (d) Path 2 (scarlet ball, Li or Mg) for G1(Mo6S8)3; (e) Path 3 for G4(Mo6S8)3 (for further details of the differences between the Li+ and Mg2+ ion trajectories, see Figures S3 and S4); and (f) Path 3i for the Li1Mg2(Mo6S8)3 system (green ball, Li; orange ball, Mg). The trajectory within the Li3(Mo6S8)3 system is provided in Figure S5 in the Supporting Information. 199x109mm (300 x 300 DPI)
ACS Paragon Plus Environment
Page 12 of 15
Page 13 of 15
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
The Journal of Physical Chemistry
Figure 3. Normalized CV profiles at RT at various scan rates: (a) Mg2+ insertion/deinsertion (Mg battery); (b) Mg2+/Li+ co-insertion/deinsertion (Mg hybrid battery); and (c) the temperature dependency of the Mg hybrid battery’s peak potential at a scan rate of 0.05 mV s-1. Linear fitting of normalized peak current density against square root of scan rate: (d) Mg2+ insertion/deinsertion; (e) Mg2+/Li+ co-insertion/deinsertion corresponding to Peaks a/a’ and b/b’; and (f) Mg2+/Li+ co-insertion/deinsertion corresponding to Peaks c/c’, in which quasi-reversible behavior is observed. 199x103mm (300 x 300 DPI)
ACS Paragon Plus Environment
The Journal of Physical Chemistry
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Figure 4. Arrhenius plots of Dapp for discharge processes. Ea values for Mg2+ insertions (blue) and Mg2+/Li+ co-insertions (red) are derived from the slopes of these plots. 90x74mm (300 x 300 DPI)
ACS Paragon Plus Environment
Page 14 of 15
Page 15 of 15
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
The Journal of Physical Chemistry
table of contents 199x114mm (300 x 300 DPI)
ACS Paragon Plus Environment
