WO2016153886A1

WO2016153886A1 - Biphasic electrode suspension for a semi-solid flow cell

Info

Publication number: WO2016153886A1
Application number: PCT/US2016/022740
Authority: WO
Inventors: Jennifer A. Lewis; Teng-Sing WEI
Original assignee: Harvard University
Current assignee: Harvard University
Priority date: 2015-03-25
Filing date: 2016-03-17
Publication date: 2016-09-29
Anticipated expiration: 2017-09-25

Abstract

A biphasic electrode suspension for a semi-solid flow cell comprises a polar solvent, a nonionic dispersant, a first plurality of particles having repulsive interactions and comprising an electrochemically active material, and a second plurality of particles having attractive interactions and comprising an electronically conductive material. The nonionic dispersant is present in an amount sufficient to sterically stabilize the first plurality of particles, but insufficient to sterically stabilize the second plurality of particles. The second plurality of particles forms a percolating conductive network through the first plurality of particles.

Description

BIPHASIC ELECTRODE SUSPENSION FOR A SEMI-SOLID FLOW CELL

RELATED APPLICATIONS

[0001] The present patent document claims the benefit of priority under 35 U.S.C. §119(e) to U.S. Provisional Patent Application No. 62/137,973, filed March 25, 2015, which is hereby incorporated by reference in its entirety.

TECHNICAL FIELD

[0002] The present disclosure is related generally to rechargeable battery technology and more particularly to electrode suspensions for semi-solid flow cells.

FEDERALLY SPONSORED RESEARCH OR DEVELOPMENT

[0003] The invention described in this disclosure was made with

government support under Contract No. DE-AC02-06CH11357 from the U.S. Department of Energy. The government has certain rights in the invention.

BACKGROUND

[0004] The rapidly increasing deployment of wind and solar energy has resulted in an urgent need for a smarter, more efficient and reliable electric grid for load balancing. Currently, only a small fraction of the total electric production is tied to grid storage, with the vast majority being pumped hydro installations. While the latter technology is mature, cost effective, and efficient, it is geographically limited. Alternate modes of energy storage that can be deployed in a distributed manner include batteries, compressed air, thermochemical energy, and flywheels.

[0005] Flow batteries are particularly attractive due to their decoupled energy and power, providing design flexibility especially at large scales. A typical flow battery includes an anolyte and a catholyte containing one or more electroactive materials that flows through an electrochemical cell, reversibly converting chemical energy directly to electricity. Additional electrolyte may be stored externally, generally in tanks, and is usually pumped through the cell (or cells) of the reactor. Flow batteries can be rapidly recharged by replacing the electrolyte while simultaneously recovering the spent material for re-energization.

[0006] Many flow battery chemistries, however, suffer from limited complex solubility and low nominal voltage, resulting in low energy densities. To overcome this, a semi-solid flow cell (SSFC) in which traditional liquid catholytes and anolytes are replaced by attractive colloidal suspensions composed of Li-ion compounds has been developed. In SSFCs, traditional stationary current collectors are replaced by conductive networks within the flowing suspensions. Similar concepts have recently emerged for

electrochemical flow capacitors and polysulfide flow batteries.

BRIEF SUMMARY

[0007] A biphasic electrode suspension for a semi-solid flow cell comprises a polar solvent, a nonionic dispersant, a first plurality of particles having repulsive interactions and comprising an electrochemically active material, and a second plurality of particles having attractive interactions and

comprising an electronically conductive material. The nonionic dispersant is present in an amount sufficient to sterically stabilize the first plurality of particles, but insufficient to sterically stabilize the second plurality of particles. The second plurality of particles forms a percolating conductive network through the first plurality of particles.

[0008] A method of making a biphasic electrode suspension for a semisolid flow cell comprises mixing together a polar solvent, a salt, a first plurality of particles comprising an electrochemically active material, and a non-ionic dispersant to form a first suspension. A second plurality of particles

comprising an electronically conductive material is mixed into the first suspension, thereby forming a biphasic electrode suspension.

[0009] The terms "comprising," "including," and "having" are used

interchangeably throughout this disclosure as open-ended terms to refer to the recited elements (or steps) without excluding unrecited elements (or steps). BRI EF DESCRI PTION OF TH E DRAWINGS

[0010] FIG. 1 A is a 3D reconstruction (top) and 2D x-y slice (bottom) of a nano-CT scan acquired on a biphasic suspension containing 20 vol.% LFP, 1 .25 vol.% KB and 0.3 wt.% PVP.

[0011] FIG. 1 B is a 3D reconstruction (top) and 2D x-y slice (bottom) of a nano-CT scan acquired on a purely attractive suspension containing 20 vol.% LFP, 1 .25 vol.% KB and 0 wt.% PVP.

[0012] FIGs. 1 C-1 F show electron micrographs of exemplary biphasic suspensions including increasing amounts (0 vol.%, 10 vol.%, 20 vol.% and 30 vol.%, respectively) of repulsive particles for a constant volume fraction (2 vol.%) of attractive particles.

[0013] FIG. 1 G shows the influence of repulsive particle volume fraction on the electronic conductivity of an exemplary biphasic suspension.

[0014] FIG. 2A is a plot of shear elastic modulus and electronic conductivity as a function of varying KB content (at 20 vol.% LFP, 0.3 wt.% PVP).

[0015] FIG. 2B is a plot of shear elastic modulus and electronic conductivity as a function of varying LFP content (at 1 .5 vol.% KB, 0.3 wt.% PVP).

[0016] FIG. 2C is a plot of shear elastic modulus and electronic conductivity as a function of varying PVP content (at 20 vol.% LFP, 1 .5 vol.% KB).

[0017] FIG. 3A is a log-log plot of shear stress as a function of shear rate for biphasic (0.3 wt.% PVP) and purely attractive (0 wt.% PVP) electrode suspensions including 20 vol.% LFP and 1 .25 vol.% KB. The lines represent fits of the Herschel-Bulkley model to the experimental data.

[0018] FIG. 3B is a log-log plot of shear elastic storage (G') and loss (G") moduli for biphasic (0.3 wt.% PVP) and purely attractive (0 wt.% PVP) electrode suspensions composed of 20 vol.% LFP and 1 .25 vol.% KB.

[0019] FIG. 4A shows flow curves for a biphasic LFP suspension

(20LFP/1 .25KB/ 0.3PVP) measured at 25 °C using different rheometer plate gaps (H = 0.8, 0.6 and 0.4 mm) against a smooth stainless steel surface. The overlapping results indicate no-slip boundary conditions at the surface.

[0020] FIG. 4B shows flow curves for a biphasic LTO suspension

(20LTO/1 .5KB/0.3PVP) measured at 25 °C using different rheometer plate gaps (H = 0.8, 0.6 and 0.4 mm) against a smooth stainless steel surface. The overlapping results indicate no-slip boundary conditions at the surface.

[0021] FIG. 5A shows apparent viscosity as a function of shear rate for biphasic (20LFP/1 .25KB/ 0.3PVP) and purely attractive (20LFP/1 .25KB/0PVP) LFP suspensions. The flow curves for the biphasic suspensions are slip- corrected, while those reported for the purely attractive suspensions are acquired at a rheometer plate gap of 0.8 mm. The latter data provide a lower bound on the true flow curves for these suspensions.

[0022] FIG. 5B shows apparent viscosity as a function of shear rate for biphasic (20LTO/1 .5KB/0.3PVP) and purely attractive (20LTO/1 .5KB/0PVP) LTO suspensions. The flow curves for the biphasic suspensions are slip- corrected, while those reported for the purely attractive suspensions are acquired at a rheometer plate gap of 0.8 mm. The latter data provide a lower bound on the true flow curves for these suspensions.

[0023] FIG. 6A shows a log-log plot of shear stress as a function of shear rate for biphasic (0.3 wt.% PVP) and purely attractive (0 wt.% PVP) electrode suspensions composed of 20 vol.% LTO and 1 .5 vol.% KB. The lines represent fits of the Herschel-Bulkley model to the experimental data.

[0024] FIG. 6B shows a log-log plot of shear elastic storage (G') and loss (G") moduli for biphasic (0.3 wt.% PVP) and purely attractive (0 wt.% PVP) electrode suspensions composed of 20 vol.% LTO and 1 .5 vol.% KB.

[0025] FIG. 7A shows analytical predictions of the electrode thickness, C- rate, and mean velocity as a function of electronic conductivity for biphasic and purely attractive LFP suspensions of varying composition, under the constraint of constant 90% voltage efficiency between charge and discharge.

[0026] FIG. 7B shows predicted pressure drop contours for suspensions with 20 vol.% (solid lines) and 5 vol.% LFP (dotted lines). The effective ionic conductivity is calculated for 1 mol L^"1 LiTFSI in a PC solvent, which has a viscosity of 8 mPa-s. Contours of electronic transference number, defined as T = a_eff/(a_eff + K_eff , where a_eff and K_eff are effective electronic and ionic conductivity of the suspension, are also shown. [0027] FIG. 7C shows surfaces of constant pressure-drop in the three- dimensional space of active-material loading, shear yield-stress, and electronic conductivity. The six data points represent a biphasic LFP suspension (20LFP/1 .25KB/0.3PVP), a purely attractive LFP suspension (20LFP/ 1 .25KB/0PVP), a biphasic LTO suspension (20LTO/1 .5 KB/0.3 P VP), and a purely attractive LTO suspension (20LTO/1 .5KB/0PVP) along with two reference samples (22.4LCO/0.6KB) and (7.9LTO/2.2KB) reported by other groups.

[0028] FIGs. 8A-8C show 3D CAD images of the electrochemical measurement cells used for electronic conductivity testing, including a static galvanostatic cycling cell and a flow galvanostatic cycling cell.

[0029] FIG. 9A shows capacity and Coulombic efficiency as a function of cycle number for a biphasic LFP suspension (20LFP/1 .5KB/0.3PVP) galvanostatically cycled vs. Li metal anode in a non-flowing Swagelok cell.

[0030] FIG. 9B shows selected cycles obtained for the same LFP. Cycles 1 and 90 are performed at C/8; Cycle 6 is performed at C/4.

[0031] FIG. 9C shows intermittent-flow cycling of the biphasic LFP suspension (20LFP/1 .25KB/0.3PVP) in a lab-scale flow cell. Two consecutive aliquots are first charged, then the second one is discharged, and finally the first aliquot is discharged. The flow channel is 20 mm long and has a 1 .5 mm x 1 .5 mm square cross-section.

[0032] FIG. 10A shows charge capacity and Coulombic efficiency for a biphasic LTO suspension (25LTO/2KB/0.8PVP) cycled galvanostatically between 2.5 V and 1 .0 V, with the first cycle at C/5 and subsequent cycles at C/8 rate.

[0033] FIG. 10B shows selected cycles for galvanostatic cycling of the same suspension as in FIG. 10A: cycle 1 is performed at C/5, while the other cycles are performed at C/8.

[0034] FIG. 1 1 and 12 show the results of cyclic voltammetry experiments to compare the cycling stability of two nonionic dispersants, Triton X-100 and PVP, respectively. DETAILED DESCRIPTION

[0035] Described herein are biphasic electrode suspensions for semi-solid flow cells (SSFCs) that may exhibit high energy density, fast charge transport and low-dissipation flow. A key element of optimizing semi-solid flow cells is maximizing the active material content of the electrode suspensions while retaining satisfactory flowability and electrical conductivity. One challenge is that an increased solids loading of the electrode suspensions can lead to dramatic changes in rheological properties, which can inhibit flow, and the electrochemically active materials may be inherently resistive. The inventors have recognized that it is possible to achieve a high active material content without sacrificing flowability or conductivity by tailoring the interactions among the particles present within the electrode suspensions.

[0036] Accordingly, a biphasic electrode suspension for a semi-solid flow cell includes a polar solvent, a nonionic dispersant, a first plurality of particles having repulsive interactions and comprising an electrochemically active material, and a second plurality of particles having attractive interactions and comprising an electronically conductive material. The dispersant is present in an amount sufficient to sterically stabilize the first plurality of particles, but insufficient to sterically stabilize the second plurality of particles. Thus, the particles comprising the electrochemically active material are dispersed in the suspension, and the particles comprising the electrically conductive material form a percolating network through the first plurality of particles. The biphasic nature of the electrode suspension is believed to enable advantageous rheological properties and excellent electrical conductivity in conjunction with a high active material content.

[0037] The polar solvent may comprise a nonaqueous polar solvent such as propylene carbonate (PC). Other suitable nonaqueous polar solvents may include ethylene carbonate (EC), a linear ester or carbonate, a fluorinated ester, a fluorinated carbonate, a fluorinated ether, a cyclic carbonate, a sulfone, a sulfonamide, tetrahydrofuran (THF), 2-methyltetrahydrofuran (2-Me- THF), polymethoxy ether, dimethoxy propane, diethyl ether, diethyoxyethane (DEE), dimethoxyethane (DME), dimethyl carbonate (DMC), diethyl carbonate (DEC), ethylmethyl carbonate (EMC), propylmethyl carbonate (PMC)), chloroethylene carbonate (CIEC), fluoroethylene carbonate (FEC),

trifluoropropylene carbonate (TFPC), γ-butyrolactone (yBL), glycol borate ester (BEG), ethyl propionate (EP), and/or ethyl butyrate (EB). In one example, the polar solvent may be a mixture of two or more nonaqueous polvent solvents. Alternatively, the polar solvent may comprise an aqueous polar solvent, such as water.

[0038] The biphasic suspension may further comprise a salt, such as a lithium salt, at a suitable concentration to achieve a high ionic strength.

Lithium ions from the salt are a crucial part of the lithiation (discharge) and delithiation (charge) electrochemical reactions that occur during use of an electrochemical cell. For example, the salt may be incorporated at a concentration of up to about 5 M, although in some cases an ionic strength (or concentration) of up to about 1 M may be preferred. It is more typical for aqueous solvents to incorporate salts at higher ionic strengths (e.g., up to about 5 M) due to the high ionic conductivity of water, whereas nonaqueous solvents typically include one or more salts at an ionic strength of up to about 1 M. A minimum concentration for the salt may be about 0.1 M. The salt may comprise one or more of the following: lithium bis(trifluoromethane)sulfon- amide (LiTFSI), a derivative of LiTFSI , lithium nitrate (LiNO₃), lithium perchlor- ate (LiCIO₄), lithium hexafluoroarsenate (LiAsF₆), lithium tetrafluoroborate (LiBF₄), lithium hexafluorophosphate (LiPF₆), lithium 4,5-dicyano-1 ,2,3-triazo- late, lithium bis(trifluoroborane)imidazolide (Lild), lithium tris(trifluoromethane- sulfonyl)methide (LiMe), lithium borate with aromatic ligands, and/or lithium fluoroalkyl phosphate (LiFAP).

[0039] Since colloidal particles can rapidly aggregate due to van der Waals interactions when suspended in polar solvents under high ionic strength conditions, a nonionic dispersant (or nonionic surfactant) is used for steric stabilization. As would be recognized by one of ordinary skill in the art, the term "nonionic surfactant" may be used interchangeably with "nonionic dispersant." Particles that are sterically stabilized have primarily or

exclusively repulsive interactions and tend to remain dispersed within the suspension. Particles having attractive interactions have a tendency to flocculate into particle aggregates. As discussed further below, it has been found that a suitable amount of the nonionic dispersant to achieve selective stabilization may be at least about 0.1 wt.%, and is preferably at least about 0.3 wt.%. Typically, the nonionic dispersant may be present in the biphasic suspension in an amount no greater than about 3 wt.%, e.g., from about 0.3 wt.% to about 3 wt.%. The nonionic dispersant may be polyvinylpyrrolidone (PVP), polyvinyl alcohol (PVA), polyvinyl pyridine or another nonionic polymer, such as those based on acrylics, esters, ethers, vinyl acetates, vinyl esters, carbonates, and ketones, or block copolymers, such as PPO-PEO-PPO (e.g., 25R4 Pluronic copolymer) with a combination of hydrophobic and hydrophilic groups. PVP is believed to be especially useful for SSFC applications since it can confer stability even in systems with high salt concentrations that undergo electrochemical charging and discharging. Another suitable nonionic

dispersant is octylphenol ethoxylate, or Triton™ X-100, which is commercially available from Dow Chemical Company. Octylphenol ethoxylate, which may be referred to as Triton X-100 in the present disclosure, may have some advantages over other nonionic dispersants in terms of electrochemical cycling stability. A nonionic dispersant having good electrochemical cycling stability does not break down within the operating voltage window during cycling.

[0040] The electrochemically active material may be selected from among: lithium titanate (Li₄Ti₅Oi2; LTO), lithium manganese nickel oxide

(LiNio_.5Mn-i .5O4; LNMO), lithium cobalt oxide (LiCoO₂), lithium iron phosphate (LiFePO₄; LFP), LiTi₂(PO₄)₃, vanadium oxide (e.g., V₂O₅, LiV₃O₈), manganese oxide (MnO₂), Sn-based oxides, composite alloys, and other materials known in the art for semi-solid flow cells. The electronically conductive material may comprise carbon. Examples of suitable carbon particles may include carbon black, carbon nanotubes, carbon nano- or microfibers, graphene particles, graphite flakes, or carbon-coated particles and flakes.

[0041] In the biphasic electrode suspension described herein, which contains a suitable amount of the nonionic dispersant, the interactions among the particles comprising the electrochemically active material are primarily or exclusively repulsive, and interactions among the particles comprising the electronically conductive material are primarily or exclusively attractive.

Consequently, as indicated above, the particles comprising the

electrochemically active material are dispersed in the suspension, and the particles comprising the electronically conductive material form a percolating conductive network through the first plurality of particles.

[0042] It should be noted that particles having repulsive interactions may be referred to as repulsive particles, and particles having attractive

interactions may be referred to as attractive particles. Similarly, suspensions containing only or predominantly attractive particles (or only or predominantly repulsive particles) may be referred to as attractive suspensions (or repulsive suspensions). The term "biphasic" is used in reference to a suspension comprising two populations of particles, where one population makes up a repulsive phase and the other population makes up an attractive phase. Cross-interactions between the two populations of particles tend to be repulsive. Also, the terms "biphasic suspension" and "biphasic electrode suspension" are used interchangeably.

[0043] The method used to prepare the biphasic suspension may influence the selective stabilization of the two populations of particles. In addition, the selective stabilization may be aided by a substantial difference in the particle size, and consequently the specific surface area, of the particles in the first and second pluralities. More specifically, the particles in the second plurality (which comprise the electronically conductive material) may have a smaller mean particle size and a higher specific surface area compared to the particles in the first plurality (which comprise the electrochemically active material). Given the higher specific surface area of the second plurality of the particles and the order in which they are added to the mixture during synthesis, as discussed further below, it is possible to preferentially coat the first plurality of particles with the dispersant while the particles in the second plurality remain substantially uncoated, or are insufficiently coated to effect stabilization.

[0044] The specific surface area of the second plurality of particles may be an order of magnitude or more greater than the specific surface area of the first plurality of particles. The particles of the first plurality may have a microscale mean particle size, while the particles of the second plurality may have a nanoscale mean particle size. In one example, the first plurality of particles may comprise a mean particle size of from about 0.1 micron to about 20 microns, or from about 1 microns to about 10 microns, and they may have a specific surface area of from about 1 m²/g to about 100 m²/g, or from about 10 m²/g to about 100 m²/g. The second plurality of particles may have a mean particle size of from about 1 nm to about 1000 nm, from about 1 nm to about 500 nm, or from about 10 nm to about 200 nm, and they may have a specific surface area of from about 100 m²/g to about 2000 m²/g, or from about 500 m²/g to about 2000 m²/g.

[0045] In one example, the first plurality or population of particles may comprise LiFePO₄ (LFP) having a mean particle size in the range of 0.1 -5 microns, and the second plurality or population of particles may comprise carbon black (e.g., Ketjenblack EC-600JD; "KB") having a mean particle size in the range of 30-100 nm (0.03-0.1 micron). These two particle populations may be suspended and mixed sequentially in a nonaqueous polar solvent such as propylene carbonate (PC) with up to 1 M of LiTFSI .

Polyvinylpyrrolidone (PVP) is employed in a suitable amount to selectively stabilize the first plurality of particles, but not the second plurality.

Alternatively, Triton X-100 may be employed. The carbon black may be added to the mixture after the PVP or Triton X-100 is mixed with the LFP particles.

[0046] In another example, the first plurality or population of particles may comprise Li₄Ti₅Oi2 (LTO) having a mean particle size in the range of 0.1 -5 microns, and the second plurality or population of particles may comprise carbon black (e.g., KB) having a mean particle size in the range of 30-100 nm (0.03-0.1 micron). These two particle populations may be suspended and mixed sequentially in a nonaqueous polar solvent such as propylene carbonate (PC) with up to 1 M of LiTFSI . Polyvinylpyrrolidone (PVP) is employed in a suitable amount to selectively stabilize the first plurality of particles, but not the second plurality. Alternatively, Triton X-100 may be employed. The carbon black may be added to the mixture after the PVP or Triton X-100 is mixed with the LTO particles.

[0047] In a third example, the first plurality or population of particles may comprise LiTi₂(PO₄)3 having a mean particle size in the range of 0.1 -5 microns, and the second plurality or population of particles may comprise carbon black (e.g., KB) having a mean particle size in the range of 30-100 nm (0.03-0.1 micron). These two particle populations may be suspended and mixed sequentially in an aqueous polar solvent such as water with up to 5 M of LiNO₃. PVP is employed in a suitable amount to selectively stabilize the first plurality of particles, but not the second plurality. Alternatively, Triton X- 100 may be employed. The carbon black may be added to the mixture after the PVP or Triton X-100 is mixed with the LiTi₂(PO₄)3 particles.

[0048] In a fourth example, the first plurality or population of particles may comprise LFP having a mean particle size in the range of 0.1-5 microns, and the second plurality or population of particles may comprise carbon black (e.g., KB) having a mean particle size in the range of 30-100 nm (0.03-0.1 micron). These two particle populations may be suspended and mixed sequentially in an aqueous polar solvent such as water with up to 5 M of UNO3. PVP is employed in a suitable amount to selectively stabilize the first plurality of particles, but not the second plurality. Alternatively, Triton X-100 may be employed. The carbon black may be added to the mixture after the PVP or Triton X-100 is mixed with the LFP particles.

[0049] Due to the effective dispersion of the first plurality of particles and the presence of the percolating network of conductive particles, a high concentration of the electrochemically active material may be used. For example, the electrochemically active material may be present at

concentration of at least about 15 vol.%, at least about 20 vol.%, or at least about 25 vol.%, and the amount may be as high as 50 vol.%. The conductive material may be present at a concentration of at least about 0.1 vol.%, at least about 0.5 vol.%, at least about 1 vol.%, or at least about 1.5 vol.%, and is typically no higher than 5 vol.%.

[0050] To broadly explore the effects of suspension composition, electronic conductivity and oscillatory measurements are carried out on biphasic electrode suspensions of varying amounts of the electrochemically active material, electronically conductive material, and dispersant. For these examples, the electrochemically active material comprises LFP, the

conductive material comprises KB, and the dispersant comprises PVP. LFP may be advantageous as it has a low volume expansion (Cnnear ~ 2.2%) when charged or discharged. For comparison, purely attractive electrode

suspensions that do not include a dispersant are also prepared and

evaluated.

[0051] First, the microstructures of exemplary biphasic and purely attractive electrode suspensions are characterized using nanoscale-computed tomography (nano-CT). The 3D reconstructed images and 2D slices obtained from each system are provided FIGs 1A and 1 B. The exemplary compositions are 20LFP/1.25KB/0.3PVP (FIG. 1A) and 20LFP/1.25KB/0PVP (FIG. 1 B), where the numbers denote the volume percent of LFP and KB particles and weight percent of PVP (in solution) in each suspension. The biphasic mixtures, which contain PVP-stabilized LFP particles, can be observed to be more homogeneous than their purely attractive counterparts, which do not contain PVP.

[0052] This is further illustrated in FIGs. 1 C-1 F, which show the

homogenizing effect of increasing amounts of repulsive particles on the microstructure of a biphasic suspension. In this example, repulsive S1O2 particles (S1O2 particles stabilized by PVP) are used to mimic the effect of PVP-stabilized electrochemically active particles in a biphasic suspension containing attractive carbon black particles (Ketjenblack EC-600JD or "KB") at a concentration of 2 vol.%. Repulsive SiO₂ particles are used due to their optical transparency in the index-matched liquid medium, and thus it is possible to observe the de-agglomeration of the attractive KB particles. The micrographs of FIGs. 1 C-1 F show the biphasic microstructure at 0 vol.% S1O2, 10 vol.% SiO₂, 20 vol.% SiO₂ and 30 vol.% SiO₂, respectively. FIG. 1 G shows that the conductivity of the biphasic suspension rises exponentially as a function of SiO₂ volume fraction for a constant volume (2 vol.%) of the KB particles.

[0053] Referring to FIGs. 2A-2C, the effect of varying the amount of each component (electrochemically active material, conductive material or dispersant) is investigated. Biphasic suspensions comprising a varying amount of conductive material (KB) and a fixed amount of electrochemically active material and dispersant (20 vol.% LFP and 0.3 wt.% PVP, respectively) are investigated first. A positive correlation between the electronic

conductivity, shear elastic modulus, and KB content of the biphasic

suspensions is observed. Even at the lowest KB content explored (1 vol.%), the particles form a percolating network that gives rise to a measurable electronic conductivity and shear elastic modulus. However, upon increasing to 2 vol.% KB, there is a power law growth in both electronic conductivity and shear elastic modulus (FIG. 2A), indicating that that the interparticle bonds within this percolating particle network have dramatically increased.

[0054] Next, the amounts of KB (1.5 vol.%) and PVP (0.3 wt.%) are fixed and the LFP content is varied (from 0 vol.% to 25 vol.%) to explore the effect on the electronic conductivity and shear elastic modulus of the biphasic suspensions (FIG. 2B). The presence of LFP particles having repulsive interactions alters both the long-range and local structure of the percolating network, which includes KB particles having attractive interactions. In the absence of the LFP particles, the percolating KB network comprises large, dense clusters that surround open regions filled with solvent and salt species. When repulsive LFP particles are included in the biphasic suspension, the percolating network of attractive KB particles becomes more homogenous, favoring the formation of more tenuous, linear chains with fewer bonds between the KB particles. The repulsive particles may have a significantly slower mobility than do solvent molecules or ionic species. When randomly distributed amongst a population of attractive particles, these species can frustrate the formation of attractive particle bonds, thereby yielding aggregated systems that are kinetically trapped in a more structurally uniform state. The concomitant rise in electronic conductivity and shear elastic modulus with increasing LFP content at a fixed number density of attractive KP particles reflects the microstructural evolution within these biphasic suspensions.

Referring to FIG. 2B, as the conductive network becomes more homogeneous due to the increased repulsive interactions at higher LFP concentrations, there are more pathways for transporting electronic current within the electrode suspensions and the shear elastic modulus increases. In the absence of attractive KB particles, the shear elastic modulus is indeed quite small (1.2 Pa), indicative of a structureless liquid state expected for a well-dispersed LFP suspension (20 vol.% LFP, 0.3 wt.% PVP).

[0055] In an additional set of experiments, both the LFP and KB contents are fixed at 20 vol.% and 1.5 vol.%, respectively, and the amount of PVP is varied from 0 to 0.5 wt.% to determine its effects on performance. Results show that once a critical amount of PVP (0.3 wt.%) is introduced to the suspension to stabilize the particles, both electronic conductivity and shear elastic modulus vary minimally with further addition of PVP, as can be seen in FIG. 2C. The above data reveal that optimizing biphasic electrode

suspensions for use in semi-solid flow cell batteries is not straightforward. Both their ability to charge/discharge efficiently and flow readily through the flow cell is important. These criteria are often in tension, since the percolating conductive network gives rise to both the desired electronic conductivity (which facilitates charge transfer) and a yield stress (which increases pumping pressure). [0056] As indicated by the above data, the biphasic suspension may be designed with advantageous flow properties in conjunction with an electronic conductivity of at least about 1 mS/cm, at least about 5 mS/cm, or at least about 10 mS/cm. In general, it is desirable to have the electronic conductivity match the magnitude of the ionic conductivity of the suspension, which may be determined by the electrolyte.

[0057] To explore dispersant effects on flow behavior, both shear viscometry and oscillatory measurements are carried out on biphasic and purely attractive electrode suspensions with the same compositions described above. The measured flow curves and shear elastic and loss moduli are shown in FIGs. 3A and 3B, respectively. Concentrated colloidal suspensions are prone to wall slip effects during the measurements. Here the flow curves for biphasic LFP suspensions are slip-corrected and exhibit nearly identical behavior over multiple gap heights, as can be seen in FIGs. 4A-4B. The inhomogeneity of the purely attractive electrode suspensions makes slip correction difficult. The flow curve for these suspensions is acquired at the gap height of 0.8 mm and represents a lower bound. Both the biphasic and purely attractive suspensions exhibit a shear thinning response (FIGs. 5A and 5B) and a shear yield stress (FIG. 3A). However, the apparent viscosity of the biphasic system is nearly an order of magnitude lower than its purely attractive counterpart at a given shear rate. The same trend is observed for shear yield stress, which is an order of magnitude lower for the biphasic LFP suspension. Fits to the Herschel-Bulkley model are shown in FIG. 3A and regression to the experimental data gives an extrapolated yield stress of 661 Pa and 5,488 Pa for the biphasic and purely attractive LFP suspensions, respectively.

[0058] The corresponding viscoelastic storage ( G') and loss moduli ( G") are provided in FIG. 3B for the same biphasic and purely attractive LFP electrode suspensions. Both suspensions behave predominantly as elastic solids (G'»G"), indicating that a percolating network of conductive particles exists within each system. However, once again, the biphasic suspensions exhibit an order of magnitude lower stiffness (0.145 MPa) than the purely attractive suspensions (2.46 MPa). It is noted that the design of electrolyte suspensions that exhibit a shear thinning response, a shear yield stress, and viscoelastic behavior is advantageous for SSFC applications, because it reduces particle sedimentation, phase segregation, and unwanted mixing of charged and discharged regions.

[0059] This strategy can be readily applied to other active materials, such as lithium titanate (LTO). To demonstrate this, both biphasic and purely attractive electrode suspensions with compositions of 20LTO/1 .5KB/0.3PVP and 20LTO/1 .5KB/0PVP, respectively, are prepared. The slip-corrected flow curves for the biphasic LTO suspension exhibits nearly identical behavior over multiple gap heights, as can be seen in FIG. 4B. Both LTO suspensions exhibit a shear thinning response, a shear yield stress and viscoelastic behavior (see FIG. 5B). Their measured flow curves and viscoelastic moduli are shown in FIGs. 6A and 6B, respectively. However, akin to the biphasic LFP suspensions, the apparent viscosity (at a given shear rate), shear yield stress and shear elastic modulus of the biphasic LTO suspensions are nearly an order of magnitude lower than their purely attractive counterparts. For example, Herschel-Bulkley fits give extrapolated shear yield stress values of 355 Pa and 4,085 Pa for the biphasic and purely attractive LTO suspensions, respectively.

[0060] As evidenced above, the biphasic suspension may exhibit flow properties that far surpass conventional electrode suspensions. For example, the apparent viscosity may be no greater than about 10⁴ Pa s at a shear rate of 10^~1 s^~1 , and the shear yield stress may be no greater than about 700 Pa. The biphasic suspension may also exhibit desirable values of shear elastic modulus, such as a shear elastic storage modulus (G') of no greater than about 6 x 10⁴ Pa, and a shear elastic storage modulus (G') of no greater than about 1 x 10⁴ Pa.

[0061] Also described herein is a method of making a biphasic electrode suspension for a semi-solid flow cell. The method may entail mixing together a polar solvent, a salt, a first plurality of particles comprising an electrochemically active material, and a nonionic dispersant to form a first suspension. A second plurality of particles comprising an electronically conductive material may be mixed into the first suspension, thereby forming the biphasic suspension.

[0062] Nonionic polymers can physically or chemically absorb onto particle surfaces to effect stabilization. PVP, PVA, and Triton X-100 are examples of phys-absorbing nonionic dispersants, which may be effective when added to the first suspension either before or after the salt is added. For chem- absorbing nonionic dispersants (e.g., comb polymers with anionic or cationic backbones with non-ionic teeth), it may be beneficial to add the dispersant to the first suspension prior to adding the salt to promote coverage of the particle surfaces.

[0063] In some cases, the polar solvent, the first plurality of particles and the non-ionic dispersant may be mixed together before adding the salt. In other cases, the polar solvent, the non-ionic dispersant and the salt may be mixed together before adding the first plurality of particles. Typically, the mixing is carried out in a controlled environment comprising a moisture and oxygen content of less than 0.5 ppm.

[0064] The polar solvent may be propylene carbonate (PC) or another suitable nonaqueous or aqueous polar solvent, as described above. The biphasic suspension may comprise an ionic strength or salt concentration of typically up to about 1 M, and the salt may be a lithium salt. The nonionic dispersant may be polyvinylpyrrolidone (PVP), Triton X-100 or another nonionic polymer such as one of those named above. To achieve the desired characteristics of the biphasic suspension, the dispersant may be present at a concentration of at least about 0.3 wt.%. The biphasic suspension may have any of the characteristics set forth in this disclosure. For example, the first plurality of particles may have a mean particle size of from about 0.1 to about 20 microns and/or a specific surface area of from about 1 m²/g to about 100 m²/g. The second plurality of particles may have a mean particle size of from about 1 nm to about 1000 nm and/or a specific surface area of from about 100 m²/g to about 2000 m²/g. In another example, the electrochemically active material employed in the above-described method may be present at concentration of at least about 20 vol.%, and may be selected from among: Li₄Ti₅Oi2 (LTO), LiNio.5Mn1.5O4 (LNMO), LiCoO₂, LiFePO₄ (LFP), V₂O₅, LiV₃O₈, MnO₂, Sn-based oxides and composite alloys. The conductive material may be present at a concentration of at least about 0.5 vol.% and may comprise carbon.

Electrode Scale Model

[0065] To identify suitable SSFC suspension compositions, an electrode- scale model that incorporates efficiency and flowability criteria has been developed. In order to impose a constant electrochemical performance criterion, all suspensions have the same round-trip voltage efficiency between charge and discharge. In all cases, the modeled cell is subjected to a current density / of 10 mA cm^"2 with a current-collector length L_cc of 20 mm. For each suspension, the electrode thickness w is treated as an adjustable parameter that produces a voltage efficiency η_ν of 90%, where η_ν is defined as:

[0067] where ^~<j)_eq is the average open-circuit voltage of the cell, and ΔΦ is the average polarization. It is assumed that the electrochemical reactions propagate via a planar front through the thickness of the electrode, producing an average polarization that decreases as the effective electronic conductivity a_eff and effective ionic conductivity K_eff increase:

[0068] ΔΦ = iw (K_eff + a_eff)/(2K_effa_eff) (2)

[0069] This trend is shown in FIG. 7A for suspensions with various loading levels of LFP.

[0070] The model couples charge-transfer and rheology properties by accounting for the high flow velocities required to cycle thin electrodes at a given current density. The mean flow-velocity required to maintain

stoichiometric conditions (where one full charge or discharge is completed in one flow pass) increases as the electrode thickness decreases:

[0071] u = iL {qw) (3)

[0072] Here, q is the volumetric charge-capacity of the suspension that depends on the type and loading of electroactive material used. Because electrode thickness decreases with decreasing electronic conductivity, the mean velocity increases as electronic conductivity decreases. A Bingham- plastic rheology (where shear stress increases linearly with shear rate¹³⁴¹) is assumed to estimate the corresponding pressure drop ΛΡ , which increases with the flow's dimensionless Bin ham number:

[0074] The Bingham number (a characteristic ratio of elastic-to-viscous stresses in the flow) is defined in terms of the fluid's yield stress r₀ and plastic viscosity μ_ρ

[0075] Bn = T₀W j (ΐμ_ρ w) (5)

[0076] With the variables defined, the model provides a contour plot as shown in FIG. 7B, where the pressure drops can be estimated for

suspensions according to their charge-transfer and rheology properties. By inserting in the measured electronic conductivity and shear yield stress values for the biphasic and attractive suspensions, it is found that the attractive suspensions require roughly ten-times the pressure drop. The predicted pressure drops for the biphasic suspensions are from about 1 psi to about 10 psi. It is worth noting that the biphasic LFP and LTO electrode suspensions exhibit electronic conductivities that are nearly two orders of magnitude higher than prior art electrode suspensions, which allows for roughly 25 times thicker electrodes (see FIG. 7A), greatly reducing the time-average pumping rate. The biphasic suspensions also lead to higher areal capacities compared to suspensions with lower electronic conductivity or active material content, enabling longer discharge time at the same current density. Furthermore, the biphasic suspensions possess nearly optimal electronic transference numbers, where electronic and ionic conductivities are of similar magnitude, as indicated in FIG. 7B. These properties are ideal for maximizing cell cycling rates, while minimizing the shunt currents between multiple cells in a stack. By contrast, the low electronic conductivities measured for prior art cells would give rise to dramatic ionic shunt currents owing to their low transference numbers as shown in FIG. 7B, if electrodes of moderate thickness are employed (e.g., from 100 μηι to about 1 mm). FIG. 7C shows the surfaces of constant pressure-drop in the three-dimensional material-selection space defined active-material loading, shear yield-stress, and electronic conductivity; data points on the plot show the properties of materials developed in the present work and in previous work. Because volumetric charge-storage capacity is proportional to the loading of active material in semi-solid

suspensions, the vertical axis is a surrogate scale for energy-density. Based on the plot, it appears that only biphasic suspensions can produce

simultaneously high energy-density (i.e., high active-material loading), fast charge transfer (i.e., high electronic conductivity), and low-dissipation flow (i.e., low pressure drop).

[0077] Accordingly, also described herein is an electrode for a semi-solid flow cell that comprises a biphasic suspension as set forth herein, where the biphasic suspension may have any of the characteristics and/or compositions described above. Advantageously, due to the excellent electronic conductivity of the biphasic suspension, the electrode may comprise a thickness of at least about 100 μηι. As shown by the test data below, the electrode may exhibit a charge capacity of at least about 120 mAh/g at a rate of C/8 and/or a

Coulombic efficiency of at least about 99%.

Electrochemical Tests

[0078] FI Gs . 8A and 8B show 3D CAD images of electrochemical measurement cells used for electronic conductivity testing, including an electronic conductivity cell 800 and a static cycling cell 802, respectively. The static cell 800 includes current collectors 810, a compression spring 808, a contact plate 806, a containment ring 804 and Swagelock fittings 812, and the flow cell 802 includes current collectors 810, a compression spring 808, a sample well 812, a separator (e.g., Celgard) 814 and a metal foil (e.g., lithium) 816. To test its electrochemical performance, a biphasic LFP suspension (20LFP/1 .5KB/0.3PVP) is first cycled in non-flowing configuration in a modified Swagelok cell. The initial specific capacity of 129 mAh g^"1 is obtained at a rate of C/8. The capacity is stable for over 90 cycles, with a capacity of 123 mAh g^"1 on the 90^th cycle, or a loss of 0.05% per cycle, as shown in FIG. 9A. Capacity is roughly the same at a rate of C/4, but drops significantly at C/2. Notably, the biphasic suspensions have a two-fold higher LFP content compared to prior reported systems. Hence, a C-rate of C/4 for a 20 vol.% suspension represents a current equivalent to C/2 with a 10 vol.% suspension, or 1 C with a 5 vol.% suspension.

[0079] FIG. 9B reveals that polarization increases with cycle number, as given by the voltage differences between the galvanostatic charge and discharge curves, indicating that the capacity loss is due to impedance growth, rather than true capacity fading. Coulombic efficiency is consistently over 99% for the biphasic LFP suspension (20LFP/1 .5KB/0.3PVP). Referring to FIGs. 10A and 10B, similar measurements were carried out for a biphasic LTO electrode suspension (25LTO/2KB/0.8PVP) and a capacity up to 170 mAh g^"1 at a rate of C/8 was obtained, with Coulombic efficiencies exceeding 99%.

[0080] To further investigate their performance, the optimized biphasic LFP suspensions are tested in a lab-scale half-flow cycling cell (FIG. 8C) against a Li metal negative electrode, using the intermittent flow mode. The half-flow cell 820 includes cathode and anode contacts 822,824, a separator 826, two inlets 828, two outlets 830, and flow channels 830. In this approach, the material inside the electroactive region is fully charged or discharged under non-flowing condition, and then the aliquot is quickly replaced with a fresh one using computer-controlled syringe pumps. This protocol is known to reduce inefficiency due to pumping and electrochemical losses. In this test, an amount of suspension equal to twice the channel volume is charged and discharged, as shown in FIG. 9C. An overall Coulombic efficiency of 72.3% and energetic efficiency of 63.1 % are obtained, as compared to first cycle Coulombic efficiency of 91 % in the static cell. This is believed to be the highest LFP concentration used in a semi-solid flow battery electrodes, and additional performance improvements are expected upon further optimizing flow conditions. The overall capacity of LFP in the suspension is 131 mAh g^"1 (equivalent to 93 Wh L^"1), at a current density of 1 .67 mA cm^"2, or a rate of C/9 per aliquot. During the charge and discharge processes, the capacity of the first aliquot is larger than that of the second one, which may be due to electroactive zone extension, in which the catholyte's high electronic conductivity allows material slightly outside the channel to be cycled.

[0081] FIG. 10A shows charge capacity and Coulombic efficiency for a biphasic LTO suspension (25LTO/2KB/0.8PVP) cycled galvanostatically between 2.5 V and 1 .0 V, with the first cycle at C/5 and subsequent cycles at C/8 rate, and FIG. 10B shows selected cycles for galvanostatic cycling of the same suspension as in FIG. 10A. Cycle 1 is performed at C/5, while the other cycles are performed at C/8.

[0082] The cycling stability of two nonionic dispersants, PVP and Triton X- 100, are compared by dissolving each dispersant into an electrolyte (1 M LiTFSI in PC) and performing cyclic voltammetry experiments. The three- electrode experimental set-up includes lithium reference and counter electrodes and a glassy carbon working electrode, with the experiments conducted at a scan rate of 20 mV/s. The results are shown in FIG. 1 1 for Triton X-100 (in comparison with data from the electrolyte alone, upper curve) and in FIG. 12 for PVP. The cycling data for the pure electrolyte and the electrolyte containing the dissolved Triton X-100 are also shown in FIG. 12, but no details can be discerned due to the difference in the y-axis scale. If the dispersant is stable within the cycling window, no extra electrochemical activity (exhibited by extra peaks) should be observed. The cycling data of FIG. 12 show an extra peak for the electrolyte including dissolved PVP, which suggests that PVP may be consumed during cycling. In contrast, the data for the electrolyte containing dissolved Triton X-100 in FIG. 11 suggest that the anionic dispersant is stable within the cycling window.

[0083] As evidenced above, a new class of biphasic electrode suspensions has been developed for semi- solid flow cell batteries. The electrode suspensions may comprise a high active material content while exhibiting improved flow behavior and enhanced electronic conductivity. The ability to independently tune the stability of two (or more) particle populations enables one to engineer concentrated suspensions that exhibit flow behavior akin to that observed for purely attractive electrode systems, while achieving far higher electronic conductivities. Given their enhanced performance, thicker electrodes can be used, allowing for more desirable transference numbers and higher theoretical areal energy densities. This approach has been demonstrated for both LFP and LTO suspensions including carbon-based percolating networks but is more broadly applicable to other electrochemically active and electronically conductive materials.

Experimental Section

[0084] Suspension preparation: The active materials consist of a carbon- coated LiFePO₄ (LFP) powder (M121 , Advanced Lithium Electrochemistry Co., Ltd., Taoyuan, Taiwan) with a mean particle size of 4 m, a specific surface area of 13 m² g^"1 , and a density of 3.551 g cm^"3 and carbon-coated Li₄Ti₅Oi2 (LTO) powder (LTO-1 , BTR NanoTech Co., Shenzhen, China) with a mean particle size of 1.1 μηι, a specific surface area of 10.68 m² g^"1 , and a density of 3.539 g cm^"3. The conductive material consists of a Ketjenblack (KB) powder (EC-600JD, Azko Nobel Polymer Chemicals LLC (Chicago, USA) with a mean particle size ranging from 30-100 nm, a specific surface area of 1400 m² g^"1 , and a density of 2.479 g cm^"3. Lithium

bis(trifluoromethane)sulfonamide (LiTFSI), propylene carbonate (PC) and polyvinylpyrrolidone (MW = 40,000 g mol^"1) are acquired from Sigma Aldrich.

[0085] Electrode suspensions are prepared in an argon-filled glovebox with moisture and oxygen content maintained under 0.5 ppm. All dry materials are heated at 120°C overnight under vacuum to remove moisture. First, 250 ml HDPE bottles are filled with 200 g of 5 mm and 100 g of 0.5 mm yttrium stabilized zirconia (YSZ) milling beads. Next, 50 g of PC, 0.3 g of PVP, and 10 g of LFP or LTO powder are added. The bottles are sealed and the

suspensions are ball- milled under ambient conditions for 24 h. The

suspensions are then filtered through 20 μηι stainless steel sieve in the argon- filled glovebox. The filtered suspensions are then sealed and centrifuged at 12,500 g in the glove box for approximately 1 hour to collect the dispersed particles. After removing the supernatant, the dense sediment (typically 70 wt.% solids) is collected and homogenized using a planetary mixer (Thinky AR-100). Additional PC and 1 % PVP/PC solution are added, followed by ultra- sonication and homogenization. LiTFSI is then added to achieve a 1 M electrolyte concentration. Finally, KB powder is added and homogenized. Suspensions containing either 0 or 0.1 PVP% may be too flocculated to pass through a 20 μηι sieve. Hence, those samples are prepared by planetary mixing of PC with PVP, followed by adding LiTFSI, then active material, and, finally KB.

[0086] Rheological characterization: Rheological measurements are carried out on electrode suspensions of varying composition using a torsional rheometer (Malvern Kinexus Pro) enclosed in an argon-filled glove box. Both steady shear viscometry and oscillatory shear tests are performed using the smooth parallel plate geometry (diameter of 20 mm; mean roughness Rq of

[0087] 0.36 Mm). All tests are performed at 25 °C, as regulated by a Peltier plate system. All samples are pre-sheared at 100 s^"1 for 5 s prior to

measurement and left to equilibrate until the normal force relaxes (-15 min). Steady shear viscometry tests are performed with decreasing applied shear rates, as described by Ovarlez et al. (Journal of Non-Newtonian Fluid

Mechanics, 2013, DOI 10.1016/j.jnnfm.2012.06.009) to ensure the existence of a simple yield stress for the material and to avoid possible transient shear banding. Following Yoshimura and Prud'Homme (Journal of Rheology (1978- present), 1988, DOI 10.1 122/1 .549963), the same sample is tested at three different gap heights to probe and correct for slip effects. If the flow curves at different gap distances superimpose, the material does not slip. If gap- dependent rheology is observed, a correction is applied to extract the true shear rate applied on the sample at each value of the applied stress.

Oscillatory strain amplitude sweep tests are performed at a fixed frequency ω = 1 rad s^"1 and used to extract their plateau shear elastic modulus.

[0088] Electronic conductivity characterization: Electronic conductivity is measured by the DC method, where the voltage is swept from 0 V to 0.15 V (Biologic VMP-3). The test cell used is a modified Swagelok cell with a cylindrical test geometry (6.35 mm wide, 200 μηι thick) sandwiched by two stainless steel electrodes. Contact resistance between suspension and current collector is neglected in these measurements.

[0089] Galvanostatic cycling characterization: Static measurements are performed in two-electrode Swagelok-type cells, using lithium metal foil (Alfa Aesar) as a counter electrode. Electrode suspensions are placed in a stainless steel rod with a 0.5 mm deep well, which is sputter-coated with gold. A porous polymer separator (Celgard) soaked with electrolyte is sandwiched between the electrodes. All electrochemical tests are performed using a Biologic VMP-3 potentiostat.

[0090] Flow cell characterization: The electrode suspensions are tested in a lab-scale half flow cell, with both the positive and negative sides consisting of a 1 .5 mm x 1 .5 mm x 20 mm electroactive region machined into a PVDF body. This region is metallized by sputter-coating with gold on the positive side. A lithium metal negative electrode is inserted into the region on the negative side, and the two halves are bolted together with a Celgard separator wetted with electrolyte in between. Pumping is performed using syringe pumps (Cetoni) with glass syringes (Hamilton Co.), at a flow rate of 30 μΙ_ s^"1. A syringe is connected to each end of the flow channel; during flow, the suspension is pushed from one syringe, while simultaneously pulled into the other. Flow cell tests are performed in "intermittent flow" mode in which the material in the electroactive region is fully charged or discharged, before another suspension aliquot in pumped in.

[0091] Although the present invention has been described in considerable detail with reference to certain embodiments thereof, other embodiments are possible without departing from the present invention. The spirit and scope of the appended claims should not be limited, therefore, to the description of the preferred embodiments contained herein. All embodiments that come within the meaning of the claims, either literally or by equivalence, are intended to be embraced therein.

[0092] Furthermore, the advantages described above are not necessarily the only advantages of the invention, and it is not necessarily expected that all of the described advantages will be achieved with every embodiment of the invention.

Claims

1. A biphasic electrode suspension for a semi-solid flow cell, the biphasic electrode suspension comprising:

a polar solvent;

a nonionic dispersant;

a first plurality of particles having repulsive interactions and comprising an electrochemically active material; and

a second plurality of particles having attractive interactions and comprising an electronically conductive material;

wherein the nonionic dispersant is present in an amount sufficient to sterically stabilize the first plurality of particles, but insufficient to sterically stabilize the second plurality of particles, the second plurality of particles forming a percolating conductive network through the first plurality of particles.

2. The biphasic electrode suspension of claim 1 wherein the polar solvent comprises a nonaqueous polar solvent.

3. The biphasic electrode suspension of claim 1 or 2, wherein the polar solvent comprises a salt present at an ionic strength of up to about 1 M.

4. The biphasic electrode suspension of claim 3, wherein the salt comprises a lithium salt.

5. The biphasic electrode suspension of any one of claims 1-4, wherein the nonionic dispersant is selected from the group consisting of: polyvinylpyrrolidone (PVP), polyvinyl alcohol (PVA), polyvinyl pyridine, octylphenol ethoxylate, a polymer based on an acrylic, ester, ether, vinyl acetate, vinyl ester, carbonate, or ketone, and a block copolymer with a combination of hydrophobic and hydrophilic groups.

6. The biphasic electrode suspension of any one of claims 1-5, wherein the amount of the nonionic dispersant is at least about 0.1 wt.%.

7. The biphasic electrode suspension of any one of claims 1-6, wherein the electrochemically active material is selected from the group consisting of: Li₄Ti₅O₁₂ (LTO), LiNio.5Mn 1.5O4 (LNMO), LiCoO₂, LiFePO₄ (LFP), LiTi₂(PO₄)₃, V₂O₅, LiV₃O₈, MnO₂, Sn-based oxides and composite alloys.

8. The biphasic electrode suspension of any one of claims 1-7, wherein the electronically conductive material comprises carbon.

9. The biphasic electrode suspension of any one of claims 1-8, wherein the electrochemically active material is present at concentration of at least about 20 vol.%.

10. The biphasic electrode suspension of claim any one of claims 1- 9, wherein the conductive material is present at a concentration of at least about 0.1 vol.%.

11. The biphasic electrode suspension of any one of claims 1-10, wherein the first plurality of particles comprises a mean particle size of from about 0.1 to about 10 microns.

12. The biphasic electrode suspension of any one of claims 1-11 , wherein the first plurality of particles comprises a specific surface area of from about 1 m²/g to about 100 m²/g.

13. The biphasic electrode suspension of any one of claims 1-12, wherein the second plurality of particles comprises a mean particle size of from about 1 nm to about 1000 nm.

14. The biphasic electrode suspension of any one of claims 1-13, wherein the second plurality of particles comprises a specific surface area of from about 100 m²/g to about 2000 m²/g.

15. The biphasic electrode suspension of any one of claims 1-14 comprising an electronic conductivity of at least about 1 mS/cm.

16. The biphasic electrode suspension of any one of claims 1-15 comprising an apparent viscosity of no greater than about 10⁴ Pa s at a shear rate of 10^"1 s^"1.

17. The biphasic electrode suspension of any one of claims 1-16 comprising a shear yield stress of no greater than about 700 Pa.

18. The biphasic electrode suspension of any one of claims 1-17, wherein the shear elastic storage modulus (G') is no greater than about 6 x 10⁴ Pa.

19. The biphasic electrode suspension of claim 18, wherein the shear elastic storage modulus (G') is no greater than about 1 x 10⁴ Pa.

20. An electrode for a semi-solid flow cell comprising the biphasic electrode suspension of any one of claims 1-15, wherein the electrode comprises a thickness of at least about 100 microns.

21. The electrode of claim 20 comprising a charge capacity of at least about 120 mAh/g at a rate of C/8 and a Coulombic efficiency of at least about 99%.

22. A method of making a biphasic electrode suspension for a semisolid flow cell, the method comprising:

mixing together a polar solvent, a salt, a first plurality of particles comprising an electrochemically active material, and a nonionic dispersant to form a first suspension; and

mixing into the first suspension a second plurality of particles

comprising an electronically conductive material, thereby forming a biphasic electrode suspension.

23. The method of claim 22, wherein the polar solvent, the first plurality of particles and the non-ionic dispersant are mixed together before adding the salt.

24. The method of claim 22, wherein the polar solvent, the nonionic dispersant and the salt are mixed together before adding the first plurality of particles.

25. The method of any one of claims 22-24, wherein the mixing is carried out in a controlled environment comprising a moisture and oxygen content of less than 0.5 ppm.

26. The method of any one of claims 22-25, wherein the salt is present at an ionic strength of up to about 1 M.

27. The method of any one of claims 22-26, wherein the nonionic dispersant is selected from the group consisting of: polyvinylpyrrolidone (PVP), polyvinyl alcohol (PVA), polyvinyl pyridine, octylphenol ethoxylate, a polymer based on an acrylic, ester, ether, vinyl acetate, vinyl ester, carbonate, or ketone, and a block copolymer with a combination of hydrophobic and hydrophilic groups.

28. The method of any one of claims 22-27, wherein the amount of the nonionic dispersant is at least about 0.1 wt.%.

29. The method of any one of claims 22-28, wherein the first plurality of particles comprises a mean particle size of from about 0.1 to about 10 microns.

30. The method of any one of claims 22-29, wherein the first plurality of particles comprises a specific surface area of from about 1 m²/g to about 100 m²/g.

31. The method of any one of claims 22-30, wherein the second plurality of particles comprises a mean particle size of from about 1 nm to about 1000 nm.

32. The method of any one of claims 22-31 , wherein the second plurality of particles comprises a specific surface area of from about 100 m²/g to about 2000 m²/g.

33. The method of any one of claims 22-32, wherein the

electrochemically active material is present at concentration of at least about 20 vol.%.

34. The method of any one of claims 22-33, wherein the conductive material is present at a concentration of at least about 0.1 vol.%

35. The method of any one of claims 22-34, wherein the

electrochemically active material is selected from the group consisting of: Li₄Ti₅Oi2 (LTO), LiNio.5Mn1.5O4 (LNMO), LiCoO₂, LiFePO₄ (LFP), LiTi₂(PO₄)₃, V2O5, L1V3O8, MnO₂, Sn-based oxides and composite alloys.

36. The method of any one of claims 22-35, wherein the

electronically conductive material comprises carbon.

37. The method of any one of claims 22-36, wherein the polar solvent comprises one or more of the following: propylene carbonate (PC), ethylene carbonate (EC), a linear ester or carbonate, a fluorinated ester, a fluorinated carbonate, a fluorinated ether, a cyclic carbonate, a sulfone, a sulfonamide, tetrahydrofuran (THF), 2-methyltetrahydrofuran (2-Me-THF), polymethoxy ether, dimethoxy propane, diethyl ether, diethyoxyethane (DEE), dimethoxyethane (DME), dimethyl carbonate (DMC), diethyl carbonate (DEC), ethylmethyl carbonate (EMC), propylmethyl carbonate (PMC)), chloroethylene carbonate (CIEC), fluoroethylene carbonate (FEC), trifluoropropylene carbonate (TFPC), γ-butyrolactone (yBL), glycol borate ester (BEG), ethyl propionate (EP), ethyl butyrate (EB), and water.