High-Capacity, Microporous Cu 6 Sn 5 Sn Anodes for Li-Ion Batteries

Transcription

1 /2009/156 5 /A385/5/$23.00 The Electrochemical Society High-Capacity, Microporous Cu 6 Sn 5 Sn Anodes for Li-Ion Batteries Lynn Trahey, a,b, *,z John T. Vaughey, a, ** Harold H. Kung, b and Michael M. Thackeray a, ** a Argonne National Laboratory, Electrochemical Energy Storage Department, Chemical Sciences and Engineering Division, Argonne, Illinois 60439, USA b Department of Chemical and Biological Engineering, Northwestern University, Evanston, Illinois , USA A385 Three-dimensional, microporous Cu 6 Sn 5 Sn architectures were created by electrodeposition of copper and tin onto sintered copper foam substrates and evaluated as anodes for lithium-ion batteries. The electrodes were characterized before and after cycling by X-ray diffraction, scanning electron microscopy, and energy-dispersive X-ray spectrometry. Before cycling, the electrochemically deposited films consisted of a combination of crystalline Cu 6 Sn 5 and Sn, whereas after cycling, the films appeared amorphous to X-rays. When evaluated in coin cells against metallic lithium, the composite Cu 6 Sn 5 Sn electrodes delivered a reversible capacity of 670 mah/g, which is significantly greater than the capacity achieved previously from powdered ballmilled and thin-film sputtered Cu 6 Sn 5 electrodes, typically mah/g The Electrochemical Society. DOI: / All rights reserved. Manuscript submitted December 8, 2008; revised manuscript received January 21, Published March 16, Rechargeable lithium-ion batteries are extremely attractive, portable energy storage devices. The world s increasing dependency on natural oil reserves for transportation has intensified international efforts to use lithium-ion batteries in hybrid-, plug-in hybrid-, and all-electric vehicles to constrain the growth of oil imports. For those goals to be realized, and particularly to increase the electric range of vehicles, the batteries need to store more energy, last longer, and most importantly, be safer than those currently manufactured. Safety limitations of state-of-the-art lithium-ion batteries arise from a number of factors, such as manufacturing defects, metal oxide cathode decomposition, overcharging a fully lithiated graphitic carbon LiC 6 anode at or below the potential of metallic lithium, and the high reactivity between charged electrodes and flammable electrolyte solvents, particularly during thermal excursions. Research on alternative anodes for lithium-ion batteries has been conducted for many years. 1-4 Various classes of materials have been investigated, notably: i metals e.g., Al, Sn 5, ii metalloids e.g., Si 6,7, iii intermetallic compounds e.g., CoSn, 8,9 Cu 6 Sn 5, 10,11 Cu 2 Sb, 12,13, and iv metal oxides e.g., Li 4 Ti 5 O 12 14,15. Metals, metalloids, and intermetallic compounds are of particular interest because they offer significantly higher theoretical volumetric and gravimetric capacities compared to graphite 372 mah/g and 818 mah/ml, and because they react with lithium several hundred millivolts above the potential of metallic lithium. However, metal, metalloid, and intermetallic electrodes have densely packed structures and therefore expand considerably during reaction with lithium, whether by an insertion- or a displacement-type mechanism. The repeated expansion and contraction during electrochemical cycling pulverizes the electrode particles; they can lose electrical contact with the current collector, causing capacity fade. One approach to counter the problem of volume expansion is to engineer threedimensional 3D electrode architectures with sufficient porosity to accommodate the expansion. 11,16-19 Intermetallic compounds that have structural similarities before and after lithiation, as observed in the Cu 6 Sn 5 to Li 2 CuSn transition, 10,20 and systems containing active and inactive components, such as CoSn, in which Sn is electrochemically reactive with lithium and Co is not, have been shown to help mitigate the effects of structural changes and metal displacement reactions on reversibility. 8,21 The major objective of our work is to design a 3D, microporous copper architecture that can act as a current collector and substrate foracu 6 Sn 5 electrode and with sufficient void volume that can * Electrochemical Society Student Member. ** Electrochemical Society Active Member. z trahey@anl.gov accommodate the volumetric expansion associated with the reaction of Cu 6 Sn 5 with lithium. Of particular relevance to this study is that such a design has already been successfully employed in hightemperature Na/NiCl 2 zebra cells; in this case, a porous, sintered nickel current collector is used as the substrate for the electrochemically deposited nickel particles generated during the discharge of cells. 22 The displacement reaction 2Na + NiCl 2 2NaCl + Ni which is accompanied by a 64% increase in volume at the cathode, is fully reversible at 300 C. In principle, therefore, it seems that it should be possible to design an analogous copper-current-collecting architecture and substrate for a Cu 6 Sn 5 electrode and to engineer and improve the reversibility of copper displacement reactions in room-temperature Li/Cu 6 Sn 5 cells. If the maximum uptake of lithium by tin is Li 17 Sn 4, alternatively Li 4.25 Sn, 23 the full reaction of lithium with Cu 6 Sn 5 can be represented by the two-stage process 10Li + Cu 6 Sn 5 5Li 2 CuSn + Cu 11.25Li + 5Li 2 CuSn 1.25Li 17 Sn 4 +5Cu The theoretical capacity of the Cu 6 Sn 5 electrode for Reactions 2 and 3 is 584 mah/g. The electrode swells by 189% upon complete reaction with lithium. In this paper, we report the fabrication and characterization of a 3D, composite Cu 6 Sn 5 Sn architecture and its evaluation as an electrode in lithium cells, in which a sintered, microporous copper foam substrate acts as the current collector. A similar investigation of Cu 6 Sn 5 has also been reported very recently. 11 Composite electrodes containing both Cu 6 Sn 5 and Sn offer a higher capacity than Cu 6 Sn 5 alone, because the theoretical value of Sn is 960 mah/g and that of Cu 6 Sn 5, 584 mah/g assuming that fully lithiated tin is Li 17 Sn Moreover, it was conjectured that composite electrodes in which Sn particles were stabilized by a copper tin matrix might yield improved capacity retention to pure Sn electrodes 24 that are not prone to good electrochemical cycling behavior. 25 Pulsed electrodeposition was used to coat the copper foam current collector with electrochemically active Cu 6 Sn 5 Sn particles, thus foregoing the use and added mass of binders and conductive carbon. Electrodeposition seems a good choice for the synthesis of electrodes because i the deposited electrode material has inherent electrical contact to the substrate, ii there are many ways to tune the composition and physical properties of the electrode, and iii the technique is commercially viable

2 A386 Experimental Copper foam current collectors were fabricated on copper foil substrates by electrodeposition following a literature procedure. 29 The deposition solution consisted of 0.2 M CuSO 4 5H 2 O and 1.5 M H 2 SO 4. The counter electrode was copper gauze and the applied current was 1.33 A. Depositions were performed for 70 s at room temperature without stirring. The resulting copper foam was sintered at 500 C for 45 h in an argon atmosphere to create a robust, 3D, current-collecting copper substrate. The substrate was subsequently weighed on a microbalance Cahn C-33 prior to copper and tin deposition. The reduction potentials of divalent copper and tin were first determined at room temperature by cyclic voltammetry CV in a solution of 0.05 M CuCl 2 2H 2 O and 0.05 M SnCl 2 with 10 vol % HCl to stabilize the solvated tin cations. The CV experiments were performed using an electrochemical cell with a Pt disk working electrode, a Pt gauze counter electrode, and a saturated calomel reference electrode SCE ; a sweep rate of 50 mv/s was employed. The solutions for the deposition of Cu 6 Sn 5 Sn films on copper foam substrates consisted of M CuCl 2 2H 2 O, 0.2 M SnCl 2, and 12 vol % HCl. The films were deposited at 600 mv vs SCE using the following potential square wave within a scanner-loop program: Step 1: 0 V vs open circuit potential OCP, 50 ms; Step 2: 600 mv vs SCE, 100 ms; Step 3: 0 V vs OCP, 50 ms; Step 4: OCP monitoring, 3 s. This procedure was carried out 400 times for each film deposition. All depositions were carried out using a PAR 273A potentiostat/galvanostat with Corrware software. Depositions were made, while stirring the solutions, on stainless steel, copper foil, and sintered copper foam substrates at room temperature. The coated substrates were subsequently annealed at 150 C for 100 h and weighed to determine the mass of the deposited copper tin material. For the coin cell data presented, the mass of the active material was determined to be 120 g. Copper tin electrodes were analyzed by scanning electron microscopy SEM, JEOL 6400, energy-dispersive X-ray spectrometry EDS, Oxford INCA, and X-ray diffraction XRD, Siemens D5000, Cu K radiation. The electrodes were assembled and evaluated in coin cells size 2032 containing a metallic lithium counter electrode, a Celgard separator, and 1.2 M LiPF 6 in ethylene carbonate/ ethyl methyl carbonate 3:7 wt % electrolyte. Coin cells were cycled galvanostatically at 0.08 ma from 0 to 2 V at room temperature. Figure 1. Cyclic voltammogram image of a 1:1 deposition solution 0.05 M CuCl 2 2H 2 O, 0.05 M SnCl 2, 10 vol % HCl, SCE, Pt working and counter electrodes, 50 mv scan rate, room temperature. Results and Discussion A typical cyclic voltammogram of a 0.05 M CuCl 2 2H 2 O, 0.05 M SnCl 2, 10 vol % HCl deposition solution in a cell with Pt working and counter electrodes, cycled at room temperature between 1.0 and 1.0 V vs a SCE, is shown in Fig. 1. The Cu 2+ ions reduce at more positive potentials than Sn 2+ ions, as expected, the former process having a peak potential of 310 mv vs SCE and the latter process around 660 mv under the sweep rate used 50 mv/s. XRD data of deposited films indicated that a Cu 6 Sn 5 phase was formed together with Sn when the potential reached approximately 600 mv, i.e., well after the Sn 2+ ions had started to reduce 500 mv. When the deposition potential was more negative 680 mv the deposits showed higher variances in regional composition by EDS and more dendritic growth. Depositions performed at more positive potentials than 600 mv primarily produced Cu 4 Sn. Further experimentation showed that Cu 6 Sn 5 formation occurred more readily when a large Sn:Cu ratio, typically 4:1 or higher, was employed. In the absence of excess tin, i.e., Cu:Sn = 1:1, XRD and EDS data showed that Cu 4 Sn was the dominant phase in the electrodeposited film. In order to obtain electrodeposited Cu 6 Sn 5 Sn products with appreciable concentrations of Cu 6 Sn 5 and Sn and without the production of inactive Cu Sn phases Cu 4 Sn, Cu 3 Sn, we found it necessary to use solutions of CuCl 2 and SnCl 2 that were heavily rich in Sn. For our experiments, a 100:1 ratio of Sn:Cu was selected. Electrodeposition was carried out using a potential square wave in a scanner loop program as described in the Experimental section. The XRD pattern of a product, electrodeposited on copper foil from a M CuCl 2 2H 2 O, 0.2 M SnCl 2, 12 vol % HCl solution, confirmed that a tin-rich composite of Sn and Cu 6 Sn 5 had been formed Fig. 2a. The copper peaks in the XRD patterns in Fig. 2a and b were due solely to the copper foil, not electrodeposited copper, because the corresponding XRD patterns of Cu 6 Sn 5 Sn products obtained by electrodeposition from the same solution on a stainless steel substrate not provided were essentially identical to those deposited on copper foil in terms of both overall composition and crystallinity, and showed no evidence of any copper metal. Annealing the product at 150 C under argon increased the crystallization of the Cu 6 Sn 5 component significantly, as evident from the growth and improved resolution of the Cu 6 Sn 5 XRD peaks in Fig. 2b. In this respect, annealing is known to enhance the Cu 6 Sn 5 content at interfaces of Cu and Sn. 24,28,30 Deposits formed on stainless steel were annealed for comparison and also showed improved resolution of Cu 6 Sn 5 XRD peaks. This indicates that the Cu substrate/deposit interaction is not solely responsible for the enhanced Cu 6 Sn 5 crystallinity upon annealing. The typical morphology of the as-grown, crystalline Cu 6 Sn 5 /Sn product is shown in an SEM image in Fig. 2c. EDS analyses of the Cu 6 Sn 5 Sn product shown in Fig. 2c indicated that the crystals consisted of, on average, 90 atom % Sn and 10 atom % Cu. Regions with higher Cu content 20 atom % were present at the rougher edges of the tin-rich crystals, possibly indicating nucleation points for the Cu 6 Sn 5 phase. An SEM image of a copper foam substrate produced by the electrodeposition of copper onto copper foil is shown in Fig. 3a. The as-grown foam is brittle and powdery. Therefore, in order to strengthen the foam and its contact to the underlying copper foil, the substrate was sintered at 500 C under argon for 45 h. The resulting morphology change is notable Fig. 3b. The thickness of the foam is roughly 90 m, as determined from micrometer and SEM crosssectional measurements. Although the foam lost some porosity, the intrinsic strength of the 3D copper architecture increased significantly and provided a sufficiently robust substrate for depositing the active Cu 6 Sn 5 Sn electrode material. An SEM micrograph of the final Cu 6 Sn 5 Sn on Cu electrode architecture, after annealing at 150 C, is shown in Fig. 3c. The crystallite size and morphology of the particles in Fig. 3c are similar to those deposited on copper foil substrates Fig. 2c. Furthermore, the XRD patterns showed that the relative amounts of deposited Cu 6 Sn 5 and Sn on the foam were essentially the same as those on copper foil substrates. EDS analyses

3 A387 Figure 2. a XRD pattern of an electrodeposited Cu 6 Sn 5 Sn product asgrown on Cu foil; b XRD pattern of a after annealing at 150 C; and c SEM micrograph of the Cu 6 Sn 5 Sn product in a. Scale bar represents 20 m. of the sample in Fig. 3c typically detected a 20 atom % minimum of Sn on the copper foil at the base of the electrode as well as on the walls of the microporous electrode architecture and 90 atom % Sn on the outermost surfaces. The use of pulse deposition conditions is believed to be the key contributor to relatively uniform, although by no means perfect, Sn Cu 6 Sn 5 coverage of the high-surface-area Cu electrode. During each deposition cycle, the applied potential square wave was followed by a 3 s OCP monitoring step. The open-circuit step allowed sufficient time for the Cu 2+ and Sn 2+ ions in the bulk solution to replenish the depleted diffuse double layer and for the OCP in the potential square wave to be continuously adjusted throughout the deposition process. Basic electrochemical principles can be used to examine the current distribution in the electrode; however, with two electrode deposition reactions, a pulse plating protocol, and complex porous electrode geometry, a detailed analysis of the deposition process is beyond the scope of this paper; these electrochemical principles and analyses will be presented and discussed in a subsequent paper. The voltage profiles of the first through sixth charge/discharge cycles of a lithium coin cell containing a Cu 6 Sn 5 Sn composite electrode on a copper foam current collector, cycled between 2 and 0 V at 0.08 ma, is shown in Fig. 4a; a capacity vs cycle number plot for the first 54 cycles is provided in Fig. 4b. Open-circuit interrupts, as seen in Fig. 4a, were part of the cycling program used and can be ignored for the purpose of this paper. A small plateau at Figure 3. SEM micrographs of a copper foam, as-grown, b sintered copper foam 500 C, and c electrodeposited Cu 6 Sn 5 Sn film on sintered copper foam. All scale bars represent 100 m. around 1.2 V in the first cycle signals the presence of some oxides likely SnO in the deposit. Figure 4b shows that four break-in cycles, during which the capacity dropped steadily from an initial value of 1020 mah/g, were necessary before the capacity steadied

4 A388 Figure 5. SEM micrograph of the Cu 6 Sn 5 Sn electrode after 54 cycles. Scale bar represents 80 m. Figure 4. a Voltage vs time profile for the first to sixth cycles of Li/Cu 6 Sn 5 Sn cell for the voltage range 2 0 V, and b charge and discharge capacity vs cycle number of coin cell tested. at 670 mah/g for the next 26 cycles. The significant irreversible capacity loss that occurs during the break-in cycles is typical for metal and intermetallic electrodes and is attributed largely to irreversible reactions with the electrolyte that form a solid electrolyte interphase SEI layer on the active copper tin particles. The discharge capacity delivered by the reaction of lithium with the Cu 6 Sn 5 Sn electrode of the Li/Cu 6 Sn 5 Sn cell is slightly larger than the charge capacity; this behavior is attributed to the use of lithium as the counter electrode and that our cycling conditions may not allow the Cu 6 Sn 5 Sn electrode to come to equilibrium and fully utilize all of the available capacity of the electrode during the electrochemical discharge reactions. The experimental capacity of 670 mah/g obtained from the composite Cu 6 Sn 5 Sn electrode is significantly higher than that achieved from Cu 6 Sn 5 electrodes alone, as expected, because Cu 6 Sn 5 has a theoretical capacity of 584 mah/g. For example, ballmilled Cu 6 Sn 5 electrodes have been reported to yield a rechargeable capacity of 200 mah/g, 10 whereas sputtered thin-film Cu 6 Sn 5 electrodes have been reported to yield a rechargeable capacity of 350 mah/g. 31 A recent report has indicated that Cu 6 Sn 5, when fabricated by a similar method to that described in this paper, yields a capacity of between 500 and 450 mah/g for 20 cycles, thereby providing supporting evidence for the advantages of using electrodeposition techniques and copper foam current-collecting substrates to synthesize high-capacity copper tin electrodes. 11 Although the Cu 6 Sn 5 Sn composite electrode delivered an impressive capacity during the early cycles, there was an abrupt onset of capacity fade at cycle 30 Fig. 4b, the exact reasons for which have not yet been determined. After 54 cycles, the cell was stopped in the charged state, i.e., at 2.0 V after lithium had been extracted from the Cu 6 Sn 5 Sn electrode. The electrode was removed from the coin cell, dried by evaporation in air, and analyzed by SEM, EDS, and XRD. The SEM image shown, for example, in Fig. 5 indicates that the copper foam backbone of the electrode had remained intact during electrochemical cycling; the rounded morphology of the electrode architecture compared to its parent state Fig. 3c was attributed to the electrochemical sintering of Cu 6 Sn 5 Sn particles and the possible reaction of the Sn component with the underlying copper foam, at least at the surface of the foam, during electrochemical cycling. Of particular interest were the XRD patterns of cycled electrodes that showed only the presence of crystalline copper, presumably from the current collector, suggesting that after 54 cycles the copper tin and tin particles were too small to be detected by X-rays and therefore appeared amorphous. The EDS analyses, however, showed that both copper and tin appeared throughout the electrode, the average atomic composition of the particles at any one point ranging 20 30% Sn and 70 80% Cu. The lower Sn content on the outer electrode surface after cycling can be attributed to the pulverized particles being more finely dispersed on the current collector, becoming detached from the current collector, or migration of Sn into the current collector. The loss in electrochemical capacity after 30 cycles is tentatively attributed to the loss in electronic contact between some of the copper tin or tin particles from the copper foam substrate during prolonged electrochemical cycling due to i the formation of an electrically resistive SEI layer that would block the active Cu 6 Sn 5 Sn particles from contact with the Cu current collector, ii active particles falling off the current collector as a result of the repetitive changes in crystallographic volume, and iii the formation inactive amorphous phases such as Cu 3 Sn and Cu 4 Sn. Work is in progress, first, to determine the precise reasons for, and overcome the capacity fade limitations, and thereafter to evaluate these copper tin on copper foam electrodes in a full Li-ion cell configuration. Conclusions Significant progress has been made in increasing the reversible capacity of Cu 6 Sn 5 electrodes by using a composite Cu 6 Sn 5 Sn system, fabricated by electrochemical deposition on a copper foam current collector. The results indicate that with improved electrode design it may be possible to engineer a Cu 6 Sn 5 Sn anode that will provide a rechargeable capacity of 600 mah/g for at least several

5 A389 hundred cycles; the results provide the motivation to continue with this research endeavor in our efforts to develop safe, high-energy, and high-power Li-ion batteries. Acknowledgments The authors thank Dennis W. Dees for helpful discussions. Financial support from the Office of Vehicle Technologies of the U.S. Department of Energy under contract DE-AC02-06CH11357 and from the Center for Energy Efficient Transportation at Northwestern University is gratefully acknowledged. Argonne National Laboratory assisted in meeting the publication costs of this article. References 1. R. A. Huggins, Solid State Ionics, 152, B. Scrosati, Electrochim. Acta, 45, M. M. Thackeray, J. T. Vaughey, C. S. Johnson, A. J. Kropf, R. Benedek, L. M. L. Fransson, and K. Edstrom, J. Power Sources, 113, M. Winter and J. O. Besenhard, Electrochim. Acta, 45, D. Larcher, S. Beattie, M. Morcrette, K. Edstroem, J. C. Jumas, and J. M. Tarascon, J. Mater. Chem., 17, U. Kasavajjula, C. S. Wang, and A. J. Appleby, J. Power Sources, 163, C. K. Chan, H. L. Peng, G. Liu, K. McIlwrath, X. F. Zhang, R. A. Huggins, and Y. Cui, Nat. Nanotechnol., 3, J. J. Zhang and Y. Y. Xia, J. Electrochem. Soc., 153, A Q. Fan, P. J. Chupas, and M. S. Whittingham, Electrochem. Solid-State Lett., 10, A K. D. Kepler, J. T. Vaughey, and M. M. Thackeray, Electrochem. Solid-State Lett., 2, X. Y. Fan, F. S. Ke, G. Z. Wei, L. Huang, and S. G. Sun, Electrochem. Solid-State Lett., 11, A L. M. L. Fransson, J. T. Vaughey, R. Benedek, K. Edstrom, J. O. Thomas, and M. M. Thackeray, Electrochem. Commun., 3, M. Morcrette, D. Larcher, J. M. Tarascon, K. Edstrom, J. T. Vaughey, and M. M. Thackeray, Electrochim. Acta, 52, E. Ferg, R. J. Gummow, A. Dekock, and M. M. Thackeray, J. Electrochem. Soc., 141, L K. Ariyoshi, R. Yamato, and T. Ohzuku, Electrochim. Acta, 51, H. C. Shin and M. L. Liu, Adv. Funct. Mater., 15, T. Jiang, S. C. Zhang, X. P. Qiu, W. T. Zhu, and L. Q. Chen, J. Power Sources, 166, F. S. Ke, L. Huang, J. S. Cai, and S. G. Sun, Electrochim. Acta, 52, M. Wachtler, M. R. Wagner, M. Schmied, M. Winter, and J. O. Besenhard, J. Electroanal. Chem., 510, L. Fransson, E. Nordstrom, K. Edstrom, L. Haggstrom, J. T. Vaughey, and M. M. Thackeray, J. Electrochem. Soc., 149, A A. D. W. Todd, R. A. Dunlap, and J. R. Dahn, J. Alloys Compd., 443, J. L. Sudworth, J. Power Sources, 100, C. Lupu, J. G. Mao, J. W. Rabalais, A. M. Guloy, and J. W. Richardson, Inorg. Chem., 42, N. Tamura, R. Ohshita, M. Fujimoto, S. Fujitani, M. Kamino, and I. Yonezu, J. Power Sources, 107, S. F. Yang, P. Y. Zavalij, and M. S. Whittingham, Electrochem. Commun., 5, S. D. Beattie and J. R. Dahn, J. Electrochem. Soc., 150, A A. Finke, P. Poizot, C. Guery, and J. M. Tarascon, J. Electrochem. Soc., 152, A W. H. Pu, X. M. He, J. G. Ren, C. R. Wan, and C. Y. Jiang, Electrochim. Acta, 50, H. C. Shin, J. Dong, and M. L. Liu, Adv. Mater. (Weinheim, Ger.), 15, N. Tamura, R. Ohshita, M. Fujimoto, M. Kamino, and S. Fujitani, J. Electrochem. Soc., 150, A I. S. Kim, J. T. Vaughey, and O. Auciello, J. Electrochem. Soc., 155, A