当前位置:首页 >> 能源/化工 >>

Semi-Solid Lithium Recharageable Flow Battery


www.advenergymat.de www.MaterialsViews.com

COMMUNICATION

Semi-Solid Lithium Rechargeable Flow Battery
Mihai Duduta, Bryan Ho, Vanessa C. Wood, Pimpa Limthongkul, Victor E. Brunini, W. Craig Carter, Yet-Ming Chiang*
Global energy and climate-change concerns have accelerated the electri?cation of vehicles, aided by advances in battery technology. It is now recognized that low-cost, scalable energy storage will also be key to continued growth of renewable energy technologies (wind and solar) and improved ef?ciency of the electric grid. While electrochemical energy storage remains attractive for its high energy density, simplicity, and reliability, existing battery technologies remain limited in their ability to meet many future storage needs. Here we propose and demonstrate a new storage concept, the semi-solid ?ow cell (SSFC), which combines the high energy density of rechargeable batteries with the ?exible and scalable architecture of fuel cells and ?ow batteries. In contrast to previous ?ow batteries, energy is stored in suspensions of solid storage compounds to and from which charge transfer is accomplished via dilute yet percolating networks of nanoscale conductors. These novel electrochemical composites constitute ?owable semi-solid ‘fuels’ that are here charged and discharged in prototype ?ow cells. Potential advantages of the SSFC approach include projected system-level energy densities that are more than ten times those of aqueous ?ow batteries, and the simpli?ed low-cost manufacturing of large-scale storage systems compared to conventional lithiumion batteries. Demand for batteries of higher energy and power has driven several decades of research in electrochemical storage materials, resulting recently in signi?cant improvements in the stored energy of cathodes and anodes.[1,2] However, most batteries have designs that have not departed substantially from Volta’s galvanic cell of 1800, and which accept an inherently poor utilization of the active materials.[3] Even the highest energy density lithium ion cells currently available, e.g., 2.8–2.9 Ah 18650 cells having >600 Wh L?1, have less than 50 vol% active material. The reduced energy density, along with higher cost, result because the high-energy-storage compounds are diluted by inactive and costly components necessary to extract power (e.g., currentcollector foils, tabs, separator ?lm, liquid electrolyte, electrode binders and conductive additives, and external packaging). Further dilution of energy density, by about a factor of two, occurs between the cell and system level.[4] Electrode designs that minimize inactive material, bio- and self-assembly, and 3D architectures are new approaches that promise improved design ef?ciency but have yet to be fully realized.[5–9] Decoupling power components from energy-storage components so that stored energy can be scaled independently of power is a strategy for improving system-level energy density. Redox ?ow batteries have such a design, in which active materials are stored within external reservoirs and pumped into an ion-exchange/electron-extraction power stack.[10] As the system increases in capacity, its energy density may asymptotically approach that of the redox active solutions. Aqueous-chemistry ?ow batteries are of much current interest for stationary applications due to their scalability, relative safety, and potentially low cost. However, they currently use low energy density chemistries limited by electrolysis to ≈1.5 V cell voltage and have low ion concentrations (typically 1–2 M), yielding ≈40 Wh L?1 energy density for the ?uids alone.[10] Furthermore, the large ?uid volumes that must be pumped produce parasitic mechanical losses that detract signi?cantly from round-trip ef?ciency. The ?ow-cell’s design advantages are therefore offset by the use of low-energy-density active materials. In a new system we call a semi-solid ?ow cell (SSFC, Figure 1), the inherent advantages of a ?ow architecture are retained while dramatically increasing energy density by using suspensions of energy-dense active materials in a liquid electrolyte.[11] This approach to ?owable electrodes can produce more than 10 times the charge storage density of typical ?ow-battery solutions, due to the much greater energy density inherent to solidstorage compounds. For example, in molarity units the concentration of reversibly-stored lithium in lithium-ion cathodes such as LiCoO2, LiFePO4, LiNi0.5Mn1.5O4, and 0.3 Li2MnO3–0.7 LiMO2 (M = Mn, Co, Ni), and anodes such as Li4Ti5O12, graphite, and Si (assuming ≈1000 mAh g?1 reversible capacity), is 51.2, 22.8, 24.1, 39.2, 22.6, 21.4, and 87 M, respectively.[2] Assuming a solids content of 50% (up to 70 vol% solids have been achieved in ?owable suspensions of other materials), the volumetric capacity of the semi-solid suspensions is 5–20 times greater (e.g., 10 to 40 M) than that of aqueous redox solutions (≈2 M). The semi-solid approach may be applied to aqueous chemistries, in which case the volumetric energy density is also 5–20 times greater since cell voltages remain limited by electrolyte hydrolysis to ≈1.5 V.[12,13] When applied to nonaqueous Li-ion chemistries, however, energy density is further increased by another factor of 1.5–3, in direct proportion to cell voltage. (Energy density is the product of volumetric charge capacity, e.g., in molarity or Ah l?1 units, and the cell voltage. Speci?c values for the systems studied are given later.) In this work, we demonstrate working prototype SSFCs using ?owable suspensions having up to ≈12 M concentration. We show that in addition to energy density advantages, the SSFCs can operate at low ?ow rates with very low mechanical energy dissipation. The design ?exibility inherent in the SSFC approach may enable new use-models for electrical storage, such as rapid refueling

M. Duduta, B. Ho, Dr. V. C. Wood, Dr. P. Limthongkul, V. E. Brunini, Prof. W. C. Carter, Prof. Y.-M. Chiang Massachusetts Institute of Technology Cambridge, MA 02139, USA E-mail: ychiang@mit.edu

DOI: 10.1002/aenm.201100152

Adv. Energy Mater. 2011, XX, 1–6

? 2011 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

wileyonlinelibrary.com

1

www.advenergymat.de www.MaterialsViews.com

COMMUNICATION
Figure 1. a) Scheme: semi-solid ?ow cell (SSFC) system using ?owing lithium-ion cathode and anode suspensions could enable new models such as transportation ‘fuels’ tuned for power versus range, or cold versus warm climates, with ?exible refueling and recycling options. b) Fluid semi-solid suspension containing LiCoO2 powder as the active material and Ketjen black as the dispersed conductive phase, dispersed in alkyl carbonate electrolyte. c) Laboratory cell using monolithic copper and aluminum current collectors, with lithium metal reference electrode, and fed by tubing using a peristaltic pump. d–f) Galvanostatic charge/discharge curves for semi-solid suspensions having 26 vol% LiCoO2 (LCO) dispersed in 1.3 M LiPF6 in an alkyl carbonate blend, 20 vol% LiNi0.5Mn1.5O4 (LNMO) dispersed in 1 M LiPF6 in ethylene carbonate:dimethyl carbonate (3:7), and 25 vol% Li4Ti5O12 (LTO) dispersed in 1 M LiPF6 in dimethyl carbonate.

of vehicles by fuel or fuel tank exchange, tuning of suspensions as needed for power, energy, and operating temperature, and extension of service life by renewing suspension chemistry or incorporating serviceable system components. Energy-Dense, Electrochemically Active Semi-Solid Suspensions: The proposed concept cannot work without effective charge transfer from the active material particles to the current collectors of the cell. Our approach to accomplishing this objective, illustrated in Figure 2, takes advantage of two limiting cases of particle aggregation behavior to produce novel, electrochemically active composites: 1) diffusion-limited cluster aggregation (DLCA) of conductive nanoparticles at low volume fraction (≈1 vol%) to form percolating conductor networks; and 2) volumetrically dense packing of micrometer-scale storage particles to maximize storage energy density. Interactions between nanoscale particles are typically dominated by colloidal rather than gravitational forces.[14] In solutions of high salt concentration such as ionically conductive electrolytes, surface charges are screened and attractive London–van der Waals attractions dominate, resulting in ‘hit and stick’ behavior that forms fractal particle networks by DLCA.[15] Testing numerous nanoscale carbons including various carbon blacks, vapor-grown carbon ?bers (VGCF), and multiwall carbon nanotubes (MWNTs), we found that percolating networks form in typical nonaqueous lithium-conducting electrolytes at particle concentrations less than 1 vol%. Figure 2c shows an example for Ketjen black in

alkyl carbonate electrolyte (1.3 M LiPF6 salt), viewed by wet scanning electron microscopy (SEM). Into this structure we added micrometer-scale particles of electrode-active cathodes (e.g., LiCoO2, LiFePO4, LiNi0.5Mn1.5O4) and anodes (e.g., graphite, Li4Ti5O12). We found that the nanoscale conductor network serves a second critical function beyond ‘wiring’ the active material for charge transfer: it stabilizes the larger particles from settling out of suspension. Figure 2d shows a wet SEM image of the LiCoO2 particle distribution in such a suspension, while Figure 2e shows an X-ray tomography showing short-range clustering of the LiCoO2 particles. Evidence for physically and electronically percolating networks was found through rheological and transport measurements. Figure 2a shows that 0.3–0.6 vol% Ketjen black produces strong shear-thinning behavior indicative of network formation, whether the suspension contains Ketjen black alone or LiCoO2 at 22.4 or 40 vol%. In fact, the suspensions containing both solid phases have several-fold higher viscosity than either suspension alone, indicating complex interactions (to be elucidated elsewhere). A conductivity cell designed to operate with stationary or ?owing liquids was used to obtain Nyquist plots such as in Figure 2b. Since ?ow rates up to 10 mL min?1 (1.6 mm cylindrical channel) produced less than a factor of 2 change in conductivity in most cases (see Supporting Information (SI)), we use results obtained under non?owing conditions to illustrate composition effects for several suspensions.

2

wileyonlinelibrary.com

? 2011 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Adv. Energy Mater. 2011, XX, 1–6

www.advenergymat.de www.MaterialsViews.com

COMMUNICATION
3

The electrochemical activity of the semisolid suspensions provides a more critical functional test. Figure 1d–f show galvanostatic charge/discharge curves for ?owable suspensions containing LiCoO2, LiNi0.5Mn1.5O4, and Li4Ti5O12 as active materials, measured in non?owing half-cell con?guration (each versus a Li-metal electrode). The LiCoO2 semi-solid suspension (26 vol%), when galvanostatically cycled at C/3.2 rate (room temperature), exhibited ≈115 mAh g?1 capacity; at C/20 rate, the capacity reached 140 mAh g?1 (not shown). A 20 vol% suspension of the high-voltage spinel LiNi0.5Mn1.5O4 exhibited the expected capacity (120 mAh g?1) and voltage (4.7 V average) and showed stable behavior over 30 cycles at C/5 rate. Similarly, the Li4Ti5O12 spinel exhibited the expected insertion voltage (1.55 V) and speci?c capacity (≈170 mAh g?1). Note the modest polarization (voltage hysteresis) in Figure 1d–f. Thus each semi-solid electrode exhibits the electrochemical activity of a conFigure 2. a) Viscosity versus shear rate for suspensions of nanoparticulate carbon (Ketjen ventional solid lithium ion battery electrode. black) and LiCoO2 (LCO) in alkyl carbonate electrolyte. The suspensions show shear-thinning The generality of this behavior is further behavior consistent with the presence of Ketjen networks partially disrupted by shear stress. shown in the SI for a non?owing lithium ion b) Nyquist plots for the different suspensions and their components. The high frequency intercept on the real axis provides the ionic conductivity, and is the same for the pure alkyl car- cell using a nanoscale olivine as the cathode bonate electrolyte and the 0.6% Ketjen suspension in the same electrolyte (Z = 55 Ω intercept and Li4Ti5O12 as the anode. For semi-solid anodes, a major concern is corresponds to 22 mS cm?1 ionic conductivity since cell con?guration factor = 1.2 cm?1). At higher solids fractions the ionic conductivity decreases, e.g., the slurry containing 22.4% LCO the likely deleterious effects of solid-electrolyte and 0.6% Ketjen has 6 mS cm?1 ionic conductivity (Z = 200 Ω). The 22.4% Li4Ti5O12 (LTO) + interface (SEI) formation on electron trans0.6% Ketjen suspension uses dioxolane solvent (1 M LiPF6) and has lower ionic conductivity of port. The SEI is formed upon reduction of ?1 (Z = 750 Ω). The electronic conductivity, extrapolated from low frequency data, is 1 mS cm alkyl carbonate solvents at potentials (relative about 102 lower than the ionic conductivity, being 0.06 and 0.01 mS cm?1, for the LCO + Ketjen + and LTO + Ketjen suspensions, respectively. c) Wet-cell SEM images of Ketjen black in alkyl to Li/Li ) of ≈0.8 V or less, and is not a concarbonate electrolyte show extended percolating networks formed by diffusion-limited cluster cern for Li4Ti5O12 due to its higher potential aggregation, whereas in (d) a suspension of 22.4 vol% LCO and 0.6 vol% Ketjen in the same (1.55 V versus Li/Li+). For graphite anodes electrolyte shows uniform distribution of LCO particles. e) A 3D reconstruction of a 10 vol% (e.g., MCMB) having ≈0.15 V potential versus LCO and 0.6 vol% Ketjen suspension obtained using X-ray tomography shows clusters of LCO Li/Li+, decreasing electronic conductivity due particles without apparent long-range percolation. to SEI formation could be readily detected in our tests. We found that by decorating the In Figure 2b, the high frequency intercept on the real axis proMCMB graphite with metallic Cu using electroless deposition, vides the ionic conductivity, and is the same for the pure alkyl a substantial degree of electronic percolation could be maincarbonate electrolyte and its mixture with 0.6% Ketjen black; tained, and we therefore do not rule out the use of low-voltage the Z = 55 Ω intercept yields 22 mS cm?1 ionic conductivity anodes in our approach. Nonetheless, shifting the potential since the cell con?guration factor is 1.2 cm?1. At higher solids window upwards by using a higher-voltage cathode and anode fraction the ionic conductivity decreases: the slurry containing together, such as LiNi0.5Mn1.5O4 with Li4Ti5O12 (producing a 22.4% LiCoO2 and 0.6% Ketjen in alkyl carbonate electrolyte 3.2 V cell) is effective for maintaining high energy density while exhibited 6 mS cm?1 ionic conductivity (Z = 200 Ω). The 22.4% minimizing SEI. Li4Ti5O12, 0.6% Ketjen suspension was prepared in a lower conContinuous and Intermittent Flow Cell Tests versus Li/Li+: ductivity electrolyte consisting of 1 M LiPF6 in dioxolane; the Semi-solid suspensions that were con?rmed to be electrochemintercept at Z = 750 Ω indicates 1 mS cm?1 ionic conductivity. ically active in non?owing cells were subsequently tested under For these particular suspensions, the electronic conductivity of ?owing conditions. Half-cells in which cathode suspensions the suspension, extrapolated from the low frequency data, is were cycled against a ?xed Li-metal negative electrode were about 102 lower than the ionic conductivity, and is expected to be tested under two pumping modes: 1) a continuously circurate-limiting during electrochemical cycling. For the LiCoO2 + lating mode in which the suspension is only partially charged/ Ketjen (in alkyl carbonate electrolyte) and Li4Ti5O12 + Ketjen (in discharged during its residence time in the ?ow channel; and dioxolane electrolyte) suspensions, the electronic conductivity 2) an intermittent mode where a single cell volume of semiis 0.06 and 0.01 mS cm?1, respectively. These variations can be solid is pumped into the cell, completely charged or discharged attributed to differences in aggregation behavior when either under static conditions, then displaced by a new volume of electrolytes or active materials differ. fresh semi-sold. (Conventional ?ow cells operate in the ?rst

Adv. Energy Mater. 2011, XX, 1–6

? 2011 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

wileyonlinelibrary.com

www.advenergymat.de www.MaterialsViews.com

COMMUNICATION
Figure 3. Semi-solid half-?ow-cell test: multistep galvanostatic charge/ discharge of LiCoO2 suspension (22.4 vol% solids (11.5 M), 0.7 vol% Ketjen, 1.3 M LiPF6 in a blend of alkyl carbonates) ?owing at 20.3 mL min?1, separated from stationary Li metal negative electrode by microporous separator ?lm.

mode: circulating using one anode and one cathode reservoir, the state-of-charge (SOC) of the ?uids’ entirety is continuously increased/decreased.) A typical single ?ow channel had 0.85 cm2 active membrane area, being 80 mm long and 1.6 mm wide and having depths of 1.4 to 3.2 mm (i.e., distance between the separator and the bottom of the channel). C-rates that are reported below refer to the capacity in a single cell volume, not the system-level C-rate, which depends on the relative capacities of the stack and the storage tanks. We used microporous polymer separators (Celgard 2500 and Tonen) developed for lithium-ion batteries that have a mean pore size of <0.1 μm, more than ten times smaller than the mean diameter of the cathode and anode particles used. We saw no evidence of particle crossover or shorting through the separator within the scope of the experiments conducted. Our ?rst successful demonstration of a SSFC appears in Figure 3. Here, a LiCoO2 suspension (22.4 vol% active (11.5 M), 0.7 vol% Ketjen black, 1.3 M LiPF6 in alkyl carbonate blend, ≈7 cP viscosity) was continuously circulated at 20.3 mL min?1 through a single channel half-?ow-cell while conducting multistep galvanostatic charging/discharging between 2 and 4.4 V (rate varied between C/8.8 and D/4.4). The charge capacity at 4.4 V (rest voltage 4.2 V) corresponds to a LiCoO2 speci?c capacity of 146 mAh g?1 (system value), while the discharge capacity corresponds to 127 mAh g?1. Compared to the expected reversible capacity of ≈140 mAh g?1, these values demonstrate high utilization of the system’s LiCoO2. The 12.5% lower discharge capacity was obtained at a higher average discharge than charge rate and does not represent the maximum achievable coulombic ef?ciency. However, the continuous mode’s pumpingdissipation/total-discharge-power ratio was calculated to be 22% (see SI); this is about 75× greater than in the intermittent mode described next.

Figure 4. Intermittent-?ow SSFC tests: a) ?ve iterations of semi-solid injection followed by constant-current constant-voltage charging for a half-?ow-cell using LiCoO2 suspension (10 vol% with 1.5 vol% Ketjen black, 1.3 M LiPF6 in alkyl carbonate blend). b) Two iterations of injection and galvanostatic cycling for a full lithium-ion ?ow cell operating between 0.5 and 2.6 V at C/8 rate. Suspensions are 20 vol% LiCoO2, 1.5 vol% Ketjen black and 10 vol% Li4Ti5O12, 2 vol% Ketjen black, both in 1 M LiPF6 in dimethyl carbonate.

Results from two intermittent-mode experiments are shown in Figure 4. The intermittent-mode permits straightforward electrochemical analysis of the suspensions while still demonstrating ?owability, and allows the relationship between C-rate, time-averaged pumping rate, and active material utilization to be determined. Figure 4a shows results characterizing a charging process: ?ve intermittent pumping and charging iterations are shown, during each of which a slug of LiCoO2 suspension was injected into the cell, charged at 1 C rate to 4.5 V, then held at 4.4 V until the current dropped to C/10 rate. The current decay during each interval shows that complete charging of a single slug occurs in ≈2 h. The charge capacities, which ranged from 118–145 mAh g?1, show generally high utilization of the LiCoO2 although there is likely some variability in the pumped volume. The ability to completely charge discrete slugs of suspension is signi?cant as it demonstrates that the electroactive zone of the continuously conductive semi-solid stream is largely con?ned to the region between the current collectors. Upon discharge of such suspensions, the percentage of capacity delivered varied from nearly 100% at 1 C rate to ≈75% at 2.5 C rate; at 2.5 C the power per unit separator area is 900 W m?2, similar to values for aqueous ?ow battery stacks.[10] Higher power should be achievable with increased loading since power density scales in direct proportion to solids concentration at a given C-rate. Moreover,

4

wileyonlinelibrary.com

? 2011 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Adv. Energy Mater. 2011, XX, 1–6

www.advenergymat.de www.MaterialsViews.com

COMMUNICATION

electronic conductance should increase with loading and enable higher C-rates. Intermittent ?ow shows signi?cantly lower parasitic loss than continuous ?ow, with the pumping-dissipation/ total-discharge-energy ratio in this experiment being less than 1% of the discharge energy (see SI). Full-Cell Tests in Intermittent Flow Mode: Figure 4b shows two charge/discharge cycles for a full lithium-ion ?ow cell that uses LiCoO2 (20 vol% (10.2 M) and 1.5% Ketjen black) as the cathode and Li4Ti5O12 (10 vol% (2.3 M) and 2% Ketjen black) as the anode. At each iteration, an anode and a cathode slug were simultaneously injected into their respective channels, displacing the previous electrode volume. Galvanostatic C/8 charge/discharge cycles were performed on each slug pair. Each electrode’s working voltage was measured with a lithium reference electrode located between two abutting separator layers. The SSFC’s capacity, limited here by the anode because of its lower loading (the cathode is incompletely charged/discharged), approach the low-rate capacity of Li4Ti5O12 (170 mAh g?1), and the coulombic ef?ciencies are 73% and 80% in the ?rst and second cycles. Since non?owing full lithium-ion cells using similar suspensions have shown up to 98% coulombic ef?ciency and 88% energy ef?ciency (SI, Figure S3), we believe that much higher round-trip ef?ciencies under continuous ?ow are possible with improvements such as closer matching of cathode and anode suspension capacities and tuning ?ow rates to match capacities. The results in Figure 4b represent the ?rst demonstration of a fully operational lithium-ion SSFC. A ?rst-order systems analysis for SSFCs assumes similar scaling as in aqueous ?ow batteries from reactant to systemlevel. For two reservoirs of 2 M aqueous redox solutions at 1.5 V discharge voltage, the reactants-only theoretical energy density is 40 Wh L?1 (≈50 Wh kg?1). Realized systems values include the effects of voltage and coulombic inef?ciencies, parasitic losses, the mass and volume of system hardware, and are 20–35 Wh L?1.[10] By comparison, a semi-solid system with 40 vol% solids in each suspension has the following theoretical energy densities for the two semi-solids combined: LiCoO2–Li4Ti5O12 (2.35 V average discharge voltage) has 397 Wh L?1 (168 Wh kg?1); LiNi0.5Mn1.5O4–Li4Ti5O12 (3.2 V average discharge voltage) has 353 Wh L?1 (150 Wh kg?1); and LiCoO2–graphite (3.8 V average discharge voltage) has 615 Wh L?1 (309 Wh kg?1). With reasonable allowances for the stack, storage vessels, and balanceof-plant, we estimate that optimized SSFC systems using established lithium intercalation compounds could have energy densities of 300–500 Wh L?1 (speci?c energy 130–250 Wh kg?1), which would satisfy metrics considered necessary for widespread adoption of all-electric vehicles.[4] Further improvements would be possible by ‘dropping in’ higher-energy-density or lower-cost storage compounds in the SSFC platform as they are developed. For large-scale applications, the SSFC design should also provide lower materials and manufacturing cost than conventional lithium ion battery technology. At near-term costs of $10–15 kg?1 for active materials and $14 kg?1 for nonaqueous electrolyte, the semi-solid suspensions alone have an energyspeci?c cost of $40–80 kWh?1 depending on the speci?c chemistry, which leaves substantial room to achieve system-level cost targets of $250 kWh?1 and $100 kWh?1 for transportation and grid level storage, respectively.[16] A simpli?ed and streamlined manufacturing process would use prepared semi-solids to ?ll

assembled systems, bypassing many of the unit operations (e.g., electrode coating, calendering, slitting and tabbing, cell assembly, module assembly) and associated capital equipment costs of current battery manufacturing.

Experimental Section
The following active materials were used to prepare the semi-solid suspensions: LiCoO2 from AGC Seimi Chemical Co., Ltd. (Kanagawa, Japan), jet-milled to D50 of 2.9 μm, LiFePO4 from Advanced Lithium Electrochemistry Co., Ltd. (Tayouan City, Taiwan), a proprietary iron-containing nanoscale olivine from A123 Systems (Watertown, Massachusetts, USA), LiNi0.5Mn1.5O4 and Li4Ti5O12 from NEI Corporation (Somerset, NJ, USA), Li4Ti5O12 of 2.7 μm average particle size from AltairNano (Reno, Nevada, USA), and MCMB graphite, grade 6–28 (6–28 μm particle size) from Osaka Gas Co. (Osaka, Japan). The graphite was copper-decorated following the method of Caturala et al.[17] prior to use in suspensions. Ketjen ECP600JD from Azko Nobel Polymer Chemicals LLC (Chicago, Illinois, USA) was used as a conductive additive (speci?c surface area = 800–1200 m2 g?1). Electrolytes were obtained from Novolyte Technologies (Independence, Ohio, USA). Compositions used included: 1.3 M LiPF6 in a proprietary blend of alkyl carbonates, 1.0 M LiPF6 in dimethyl carbonate, 1.0 M LiPF6 in 3:7 mixture of ethylene carbonate:dimethyl carbonate, and 1.0 M LiPF6 in dioxolane. Microporous separators used in electrochemical testing were Celgard 2500 (Celgard LLC, Charlotte, North Carolina, USA) and Tonen (Tonen Chemical Corporation, city, Japan). Handling and storage of suspension components, suspension formulation and rheological testing, and electrochemical testing and conductivity measurements were all conducted in argon-?lled glove boxes (MBRAUN, Newburyport, MA, USA) with oxygen and water levels maintained below 5 and 0.1 ppm, respectively. The active materials and carbon were weighed and mixed in a 20 mL glass vial and electrolyte was added using a variable volume pipette. The resulting suspensions were sealed in glass vials, removed from the glove box, and sonicated in a Branson 1510 ultrasonic bath for a period of time of 60 min prior to use. Single-channel ?ow cells as shown in Figure 1c were machined from 101 copper alloy (negative pole) and 6061 aluminum alloy (positive pole). SI, Figure S1 shows a design drawing of the electrochemical ?ow cell. The working surface of the aluminum was sputtered with gold in a Pelco SC-7 instrument to reduce interfacial impedance. Electrochemical testing was performed using a Solartron potentiostat operating a 1400 Cell Test System (AMETEK Inc., Paioli, Pennsylvania, USA). Continuous ?ow experiments were performed using a Master?ex peristaltic pump (Master?ex, Vernon Hills, Illinois, USA); Chem-Sure tubing (W. L. Gore and Associates, Elktron, Maryland, USA) was used inside the pump and was connected to the cell using Master?ex Chem-Durance tubing. Intermittent ?ow experiments were conducted using manually pumped, calibrated, gas-tight syringes. Viscosity measurements were conducted in an argon-?lled glove box using a Brook?eld controlled shear-rate viscometer (Model RVDV-II + Pro, Brook?eld Engineering, Middleboro, Massachusetts, USA). Suspensions were sheared under Couette ?ow in a concentric cylinder measurement system (Brook?eld spindle model SC4–15 and chamber model 7R). Data were recorded at 4 s intervals over 2 min at each preprogrammed shear rate. Conductivity measurements of the semi-solids were performed in an apparatus with a ?ow channel of the same geometry as used in the ?ow cell tests, ensuring that the semisolid suspensions under test are presented with the same shear conditions for a given ?ow rate. The conductivity cell was assembled from interlocking PTFE pieces and stainless steel electrodes (SI, Figure S2). The area of the electrodes was chosen so that the cell con?guration factor was 1.2 cm?1. Prior to measuring the conductivity of the semisolid suspensions measurement, the cells were tested and calibrated using Oakton conductivity standards. SEM of the semisolid suspensions in their dispersed, wet state was carried out using a QuantomiX WETSEM capsule. The images were

Adv. Energy Mater. 2011, XX, 1–6

? 2011 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

wileyonlinelibrary.com

5

www.advenergymat.de www.MaterialsViews.com

COMMUNICATION

obtained with a 15 kV electron beam and a 9 mm working distance in a FEI/Philips XL30 FEG ESEM. X-ray tomography experiments providing 3D reconstruction of the semisolid suspensions were conducted at the TOMCAT beamline of the Swiss Light Source (Paul Scherrer Institute, Villigen, Switzerland). The semisolid suspension was injected into a hollowed cylindrical column of aluminum with an inner diameter of 1.0 mm and a side-wall thickness of 200 μm to enable transmission of the X-rays through the sample. The column was scanned at an X-ray energy of 20 keV with a theoretical pixel size of 0.74 micrometers. 1501 projections (115 ms exposure each) were recorded to reconstruct the 3D density distribution information. Gray values representing the density distribution were used to determine the threshold value separating the LiCoO2 from the Ketjen black and alkyl carbonate electrolyte.

Supporting Information
Supporting Information is available from the Wiley Online Library or from the author.

Acknowledgements
We thank Dr. Federica Marone and Professor Dr. Marco Stampanoni for performing the tomography measurements at the Swiss Light Source (Paul Scherrer Institute), and Katherine M. Blancha for assistance with ?gures. This work was initially funded by Defense Advanced Research Projects Agency (DARPA) Contract No. FA8650–09-D-5037), and subsequently funded by the Advanced Research Projects Agency–Energy (ARPA-E)), US Department of Energy, under Award Number DE-AR0000065. The information, data, or work presented herein was funded in part by an agency of the United States Government. Neither the United States Government nor any agency thereof, nor any of their employees, makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Reference herein to any speci?c commercial product, process, or service by trade name, trademark, manufacturer, or otherwise does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States Government or any agency thereof. The views and opinions of authors expressed herein do not necessarily state or re?ect those of the United States Government or any agency thereof. Received: March 22, 2011 Revised: April 18, 2011 Published online:

[1] a) Q. Zhong, A. Bonakdarpour, M. Zhang, Y. Gao, J. R. Dahn, J. Electrochem. Soc. 1997, 144, 205; b) Z. Lu, L. Y. Beaulieu, R. A. Donaberger, C. L. Thomas, J. R. Dahn, Electrochem. Soc. 2002, 149, A778; c) M. M. Thackeray, S. Kang, C. Johnson, J. Vaughey, R. Benedek, S. A. Hackney, J. Mater. Chem. 2007, 17, 3112; d) F. Badway, F. Cosandey, N. Pereira, G. G. Amatucci, J. Electrochem. Soc. 2003, 150, A1318. [2] a) M. Winter, J. O. Besenhard, Electrochim. Acta 1999, 45, 31; b) L. Y. Beaulieu, R. L. Turner, A. Bonakdarpour, A. A. Abdo, L. Christensen, K. W. Eberman, L. J. Krause, J. R. Dahn, J. Electrochem. Soc. 2003, 150, A149; c) M. N. Obrovac, L. Christensen, Electrochem. Solid-State Lett. 2004, 7, A93; d) P. Poizot, S. Laruelle, S. Grugeon, L. Dupont, J. M. Tarascon, Nature 2000, 407, 496; e) C. K. Chan, H. Peng, G. Liu, K. McIlwrath, X. F. Zhang, R. A. Huggins, Y. Cui, Nature Nanotechnol. 2008, 3, 31. [3] A. Volta, Philos. Trans. R. Soc. London 1800, 90, 403. [4] ARPA-E BEEST FOA, https://arpa-e-foa.energy.gov/FoaDetailsView. aspx?foaId = 85e239bb-8908--4d2c-ab10-dd02d85e7d78 (accessed May, 2011). [5] W. Lai, C. K. Erdonmez, T. F. Marinis, C. K. Bjune, N. J. Dudney, F. Xu, R. Wartena, Y.-M. Chiang, Adv. Mater. 2010, 22, E139. [6] K. T. Nam, D. Kim, P. J. Yoo, C. Chiang, N. Meethong, P. Hammond, Y.-M. Chiang, A. M. Belcher, Science 2006, 312, 885. [7] Y. K. Cho, R. Wartena, S. M. Tobias, Y.-M. Chiang, Adv. Funct. Mater. 2007, 17, 379. [8] J. W. Long, B. Dunn, D. R. Rolison, H. S. White, Chem. Rev. 2004, 104, 4463. [9] P. H. L. Notten, F. Roozeboom, R. A. H. Niessen, L. Baggetto, Adv. Mater. 2007, 19, 4564. [10] C. Ponce de León, A. Frías-Ferrer, J. González-García, D. A. Szánto, F. C. Walsh, J. Power Sources 2006, 160, 716. [11] Y.-M. Chiang, W. C. Carter, B. Y. Ho, M. Duduta, Int’l Patent Application No. WO2009151639A1, 2009. [12] W. Li, J. R. Dahn, D. S. Wainwright, Science 1994, 264, 5162. [13] R. Ruffo, C. Wessels, R. A. Huggins, Y. Cui, Electrochem. Comm. 2009, 11, 247. [14] J. Israelachvili, Intermolecular and Surface Forces, Academic Press, London 1992. [15] M. Y. Lin, H. M. Lindsay, D. A. Weitz, R. C. Ball, R. Klein, P. Meakin, Nature 1989, 360, 339. [16] Arpa-e GRIDS FOA https://arpa-e-foa.energy.gov/Default. aspx?Search=GRIDS&SearchType= (accessed May, 2011). [17] F. Caturla, F. Molina, M. Molina-Sabio, F. Rodríguez-Reinoso, A. Esteban, J. Electrochem. Soc. 1995, 142, 4084.

6

wileyonlinelibrary.com

? 2011 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Adv. Energy Mater. 2011, XX, 1–6


相关文章:
锂离子液流电池的研究进展.pdf
[ 9 ] Duduta M ,Ho B ,Wood V C ,et al. Semi-Solid Lithium recharageable flow battery [J]. Advanced Energy Materials ,2011 ,1 ( 4 ): 511516...
Organosilicon Electrolytes for Li-ion Batteries(On ....pdf
(On 26th international battery seminar)_能源/化工...largely by SEI (Solid-electrolyte interface) ...and recharageable lithium-ion batteries with very...
lithium-battery-guidance-document-2017-en_图文.pdf
lithium-battery-guidance-document-2017-en_电子/...able to charge consumer devices such as mobile ...Classification Flowcharts The following (2) ...
All-solid-state lithium battery with LiBH4 solid el....pdf
All-solid-state lithium battery with LiBH4 solid...ammable organic liquid electrolyte, it has been ...The diameter of the semicircle corresponds to the...
更多相关标签: