ReportsChemistry

Rechargeable nickel–3D zinc batteries: An energy-dense, safer alternative to lithium-ion

See allHide authors and affiliations

Science  28 Apr 2017:
Vol. 356, Issue 6336, pp. 415-418
DOI: 10.1126/science.aak9991

Zinc can compete with lithium

Although lithium-based batteries are ubiquitous, there are still challenges related to their longevity and safety, as well as concerns about material availability. Aqueous rechargeable batteries based on zinc might provide an alternative, but they have been plagued by the formation of dendrites during cycling. Parker et al. show that when zinc is formed into three-dimensional sponges, it can be used with nickel to form primary batteries that allow for deep discharge. Alternatively, the sponges can be used to produce secondary batteries that can be cycled thousands of times and can compete with lithium ion cells.

Science, this issue p. 415

Abstract

The next generation of high-performance batteries should include alternative chemistries that are inherently safer to operate than nonaqueous lithium-based batteries. Aqueous zinc-based batteries can answer that challenge because monolithic zinc sponge anodes can be cycled in nickel–zinc alkaline cells hundreds to thousands of times without undergoing passivation or macroscale dendrite formation. We demonstrate that the three-dimensional (3D) zinc form-factor elevates the performance of nickel–zinc alkaline cells in three fields of use: (i) >90% theoretical depth of discharge (DODZn) in primary (single-use) cells, (ii) >100 high-rate cycles at 40% DODZn at lithium-ion–commensurate specific energy, and (iii) the tens of thousands of power-demanding duty cycles required for start-stop microhybrid vehicles.

The present energy-storage landscape continues to be dominated by lithium-ion batteries despite numerous safety incidents (1, 2) and obstacles, including transportation restrictions (3), constrained resource supply (lithium and cobalt) (4), high cost (5), limited recycling infrastructure (6, 7), and balance-of-plant requirements (8)—the last of which constrains the energy density of Li-ion stacks. Despite these disadvantages, Li-ion batteries are widely used because they provide high energy density, high specific power, and long cycle life—attributes that must also be met by any alternative battery system in order to compete for market share.

The family of zinc-based alkaline batteries (Zn anode versus a silver oxide, nickel oxyhydroxide, or air cathode) is expected to emerge as the front-runner to replace not only Li-ion but also lead-acid and nickel–metal hydride batteries (9, 10). This projection arises because Zn is globally available and inexpensive, with two-electron redox (Zn0/2+) and low polarizability that respectively confer high specific capacity and power. The long-standing limitation that has prevented implementing Zn in next-generation batteries lies in its poor rechargeability due to dendrite formation (1113).

We bypass this obstacle to cycling durability by redesigning the Zn electrode as a monolithic, porous, aperiodic architecture in which an inner core of electron-conductive metallic Zn persists even to deep levels of discharge, schematically depicted in Fig. 1A (14, 15). In primary 3D Zn–air cells, this “sponge” form factor (3D Zn) discharges >90% of the Zn (16), a 50% improvement over conventional powder-bed composites (17). When cycling Zn sponges at the demanding current densities that otherwise induce dendrite formation in alkaline electrolyte—typically greater than 10 mA cm–2 (18)—the 3D Zn restructures uniformly without generating separator-piercing dendrites (14).

Fig. 1 Possibilities with rechargeable Ni–Zn.

(A) Schematic of the effect of recharging Ni–Zn (conventional powder zinc anodes) versus Ni–3D Zn in which the anode is redesigned as a monolithic aperiodic sponge ensuring persistent 3D wiring of the metallic Zn core. Dendrites that form at powder-composite Zn anodes can reach hundreds of micrometers in length (30, 31). (B) The calculated specific energy of a fully packaged Ni–Zn cell as a function of increasing Zn depth of discharge versus a capacity-matched NiOOH electrode. The shaded areas highlight the specific energy range of common battery chemistries. For example, at ≥40% DODZn (percentage of theoretical utilization), Ni–Zn becomes competitive with Li-ion at the single-cell level.

The performance of the Zn anode enables us to explore the secondary Ni–Zn system. This battery chemistry uses a rechargeable cathode (NiOOH) that is further developed than the air cathode of rechargeable Zn–air and is more economically feasible than Ag–Zn. Nickel-zinc batteries discharge via the oxidation of Zn metal coupled with the reduction of nickel oxyhydroxide according to the anodic (Eqs. 1 to 3; Zn) and cathodic (Eq. 4; NiOOH) reactions.

Electrooxidation: Zn(s) → Zn2+(aq) + 2e(1)Complexation: Zn2+(aq) + 4OH(aq) → Zn(OH)42–(aq)(2)Dehydration/Precipitation: Zn(OH)42–(aq) → ZnO(s) + H2O(l) + 2OH(aq)(3)Electroreduction: 2β-NiOOH(s) + 2H2O(l) + 2e → 2β-Ni(OH)2(s) + 2OH(aq)(4)

The theoretical specific energy for Ni–Zn is 372 Whkg–1, whereas a practical Ni–Zn battery delivers up to 135 Whkg–1 (~300 WhL–1 on a volumetric basis) depending on battery-design considerations and Zn depth of discharge (DODZn). Contrasting the specific energy for a fully packaged Ni–Zn cell as a function of increasing DODZn with that for lead-acid, nickel-cadmium, and nickel–metal hydride shows that the performance of Ni–Zn is comparable or superior (Fig. 1B), even at modest utilization of the Zn (10 to 20% DODZn). Deeper depths are required (≥40% DODZn) to bring Ni–Zn to a specific energy that becomes competitive with common Li-ion batteries at the single-cell level. These calculations assume that the Zn and Ni electrodes are present at 39% of the total packaged weight (19)—a conservative assumption because the percentage of packaging weight (casing) is expected to decrease when scaling Ni–3D Zn cells up to vehicle-relevant stacks.

We previously reported on electrolyte formulations and electrode additives that minimize shape change of Zn sponge electrodes cycled 20 times to 20% DODZn in a Ni–3D Zn configuration (20). The proper electrolyte formulation should include additives that force dehydration of soluble zincate [Zn(OH)42–(aq) to ZnO(s)], (Eqs. 2 and 3) at lower concentrations than occur in unadulterated 6 M KOH. For the deep-discharge and long-term cycling conditions of this study, we used an electrolyte formulation of 6 M KOH + 1 M LiOH in conjunction with a Ca(OH)2-infused Zn sponge electrode. This combination of additives provides superior round-trip cycling efficiency—a cell-based metric that convolves performance from the cathode and anode—because (i) Li+ augments NiOOH rechargeability by suppressing O2 evolution (21); (ii) Ca(OH)2 induces zincate supersaturation (22); and (iii) 300 parts per million (ppm) of In and 300 ppm of Bi predoped into the Zn suppress H2 evolution.

For potential application in consumer electronics, the higher cell voltage of Ni–Zn over traditional, single-use alkaline batteries (MnO2–Zn) is a compelling feature if it can be coupled to essentially complete use of the Zn anode. The ability of Zn sponge anodes to discharge to high-Zn mass-normalized capacity and be recharged without inducing dendritic shorts was probed by exhaustively discharging Ni–3D Zn cells (Fig. 2A) at a current density of ~10 mA cm–2 (C/9; i.e., the entire capacity of the battery is discharged in 9 hours) and then recharging at the same rate. These cells reached an average 91% DODZn (743 mAhgZn–1; 1202 WhkgZn–1) and could be recharged to >95% capacity from these extreme depths (Fig. 2B). Similar Zn depths of discharge were obtained in our previous 3D Zn–air studies, but we could not probe capacity recovery in that configuration because of the lack of a mature recharge-capable air cathode (14). The emulsion-based route to Zn sponges also affords great flexibility in application-specific x-y-z size and form factors because the mold defines the anode size and shape (fig. S1).

Fig. 2 Cycling performance of nickel–3D zinc cells.

(A) Schematic design of the nickel–3D zinc coin cell used in this study. (B) Nickel–3D zinc cells tap >90% of the theoretical Zn capacity upon discharge (black circles, at 10 mA cm–2) and >95% of that discharged capacity can be recovered upon subsequent recharge (red squares, at 10 mA cm–2) with a half-cycle voltage hysteresis of <300 mV. (C and D) The voltage-time curves for cells discharged at 25 mA cm–2 to 40% DODZn and recharged at either (C) 5 mA cm–2 or (D) 10 mA cm–2. The constant voltage at 1.93 V indicates the potentiostatic region of the charge profile.

To address the feasibility of Ni–3D Zn in fields of use that demand multicell stacks, high cycle life, and power performance, we cycled Ni–3D Zn cells to a DODZn (40%) that translates to a specific energy competitive with Li-ion (Fig. 1B). The long-term experiments began with a 5-mA cm–2 break-in cycle consisting of a ~50 mAh discharge (50% DODZn) and a recharge of ~40 mAh. This first-cycle capacity mismatch was chosen to saturate the electrolyte with zincate and to introduce a buffering amount of ZnO and Ni(OH)2 into the respective electrodes to minimize gas evolution upon charging (23). In subsequent cycles, the cells were discharged at 25 mA cm–2 (a C/1.5 rate with respect to a nominal capacity of 328 mAhgZn–1) and recharged at either 5 mA cm–2 (Fig. 2C) or 10 mA cm–2 (Fig. 2D). A ~3 mAh potentiostatic hold at 1.93 V was added to the end of each charge to ensure exhaustive oxidation of the NiOOH electrode while avoiding O2 evolution (6).

The cells ran for 111 and 141 cycles for the 5 mA cm–2 and 10 mA cm–2 charging cases, respectively, before falling below 50% of nominal cycling capacity. Upon >20% capacity fade, typically at >80 cycles, injection of electrolyte or water into the cathode compartment revives the nonhermetically sealed plastic cells back to nominal capacity, demonstrating that the fade arises from dehydration rather than irreversible passivation of either the cycled 3D Zn or Ni electrodes.

These Ni–3D Zn cells maintained 100% of the required discharge capacity for 85 and 65 cycles, respectively, with an average energy efficiency of 84% before capacity fading [comparable to the 85% energy efficiency found in Li-ion batteries (24)]. The cycling stability achieved in electrolyte-limited Ni–3D Zn cells stands in contrast to the capacity fade in commercial products (6) and in 3D-inspired designs comprising electrodeposited Zn coatings on Ni mesh (25). Scanning electron micrographic analysis of the postcycled Zn sponges (Fig. 3, D to I) reveals uniform restructuring of the surfaces, the absence of dendrites, and maintenance of the porosity and interconnectivity of the monolithic sponge; some densification is noted, however, relative to the precycled microstructure, Fig. 3, A to C.

Fig. 3 Postcycling microstructural analysis of 3D Zn sponges.

Scanning electron micrographic analysis of (A to C) precycled and (D to I) postcycled Zn sponges after >100 cycles, verifying that minimal shape change occurs and no dendrites are formed when the Ni–3D Zn cell is discharged at 25 mA cm–2 to 40% DODZn and recharged at either [(D) to (F)] 5 mA cm–2 or [(G) to (I)] 10 mA cm–2.

With a demonstration of pulse-power capability, Ni–3D Zn could compete in a third field of use—replacing lead-acid batteries within microhybrid vehicles. The duty cycles for “start-stop” operation involve pulses for engine start and restart as well as auxiliary constant-use loads such as air conditioning and entertainment systems. State-of-the-art start-stop batteries for microhybrid vehicles currently use lead-acid cells with absorbed glass mat (AGM) technology. Lead-acid AGM has the advantage of low cost and excellent shelf life in the charged state but suffers from such disadvantages as low specific and volumetric energy, life-cycle concerns due to toxic active materials (Pb and PbO2) (26), electrolyte instability in the discharged state (27), and poorer cycle life and price point compared with standard SLI (starting-lighting-ignition) lead-acid batteries.

To validate the applicability of Ni–3D Zn as a start-stop battery, we approximated the current-versus-time duty cycle of the BMW microhybrid battery (28) as scaled to our typical single-cell dimensions (Fig. 4A). We used the following assumptions: (i) the specific power of individual Ni–3D Zn cells will match that of individual Pb-acid cells within the AGM battery commonly used in BMW’s microhybrid systems (e.g., Exide EK900); (ii) a scaled-up Ni–3D Zn battery requires eight cells to achieve the necessary voltage (~12 V) and would therefore deliver 33% more power than its six-cell Pb-acid counterpart; and (iii) Zn will occupy 19% of the packaged weight (19). The through-connected void structure of the sponge serves to ameliorate transport limitations under high-rate demands (fig. S2), such as those required during the acceleration phase of a start-stop duty cycle.

Fig. 4 Long-term performance of Ni–3D Zn single cell as cycled under start-stop conditions.

(A) The current-time duty cycle modeled from a BMW AGM start-stop drive cycle (28) scaled to our 1-cm2 Ni–3D Zn coin cells. (B) The measured current-time curves for Ni–3D Zn coin cells at early (solid line, 4000 cycles) and late (dashed line, 54,000 cycles) points in the 4.5-month-long, nonstop cycling. (C) Micrographic analysis of a postcycled Zn sponge after ~54,000 cycles, which verifies that minimal shape change occurs and no dendrites are formed.

For start-stop batteries to achieve >104 cycles, it is customary to keep the percentage of capacity used intentionally low. Per 4-min duty cycle, the capacity tapped of the Ni–3D Zn coin cells was kept to <1% DODZn. More than 50,000 cycles (Fig. 4B) were achieved, with cycling stopped only when the high load pulse (~65 mA cm–2) reached a preset voltage limit of 0.8 V. With a nominal 20 start-stop cycles in a round-trip commute, Ni–3D Zn would provide ~2500 days of start-stop performance (>6.8 years of daily use), approaching the average 11.4-year age of U.S. cars (29). The cumulative discharge capacity for ~54,000 cycles is ~3 times that achieved in the 40% DODZn/100+ cycles discussed above. Postmortem analysis of the months-long-cycled (still nonhermetically sealed) cells revealed a dry cell concomitant with an increased cell resistance. The postcycled Zn sponge remains visibly monolithic; scanning electron microscopy reveals that the pore–solid architecture of the Zn sponge is retained and that no anomalous macroscale dendrites are electrogenerated (Fig. 4C).

We then assessed the effect of a 3D Zn anode-based battery on the energy-storage requirements of various electric vehicles (EVs). The quantitative assessment fixed the energy capacity for each EV application using the current state of the art for (i) an electric bicycle (versus standard lead-acid), (ii) a start-stop microhybrid (versus lead-acid AGM), and (iii) an all-electric battery vehicle (versus Li-ion). Weight and volume savings result for all three applications by using Ni–3D Zn (Table 1).

Table 1 Summary of the projected effect of the nickel–3D zinc–based battery on various weight and normalized capacity metrics of relevance to electric vehicles (EVs).

SLA, sealed lead-acid; AGM, absorbed glass mat.

View this table:

A projected Ni–3D Zn battery pegged to the specific capacity of the Nissan Leaf (24 kWh) saves 100 kg of weight. Much of the weight and potential cost savings with Ni–3D Zn over Li-based EV batteries come from the reduction or elimination of subsystems that are required for Li-ion battery packs, which include thermal management, sophisticated electronic controls, and structural protection to manage any catastrophic events. The 3D Zn–based batteries will not require comparably complex subsystems. The advantages of Ni–3D Zn–based batteries—not just the projected range and cost improvements in EV applications—are augmented by eliminating dangers associated with fire risk from incidents of Li-ion thermal runaway, all while using a nonstrategic, globally available, recyclable natural resource.

Supplementary Materials

www.sciencemag.org/content/356/6336/415/suppl/DC1

Materials and Methods

Figs. S1 to S3

References

References and Notes

  1. J. E. Oxley, C. W. Fleischmann, Improvement of zinc electrodes for electrochemical cells, First, second, and third quarterly reports N66-13568, N66-16956, N66-26870 (Leesona Moos Labs, Great Neck NY, 1965–1966).
Acknowledgments: The U.S. Naval Research Laboratory and EnZinc, Inc., teams wish to acknowledge the Advanced Research Projects Agency–Energy (ARPA-E) for financial support (award DE-AR-0000391) and for the guidance of the ARPA-E staff of the Robust Affordable Next Generation Energy Storage (RANGE) program. This work was also supported by the U.S. Office of Naval Research. M.F.B. and M.M. thank B. Dussia, R. Tarr, K. Dias, and S. Mohanta for their essential input, as well as the staff of the Venture Greenhouse green technology incubator for helping assess the commercial potential of Zn-based batteries equipped with Zn sponge anodes. I.R.P. was a Naval Research Laboratory–National Research Council Postdoctoral Associate. D.R.R. served as technical lead for the project. J.F.P., J.W.L., and D.R.R. oversaw experimental design. J.F.P. and I.R.P. prepared Zn sponges for use in NiOOH–3D Zn coin cells. C.N.C. harvested the NiOOH cathode and multilayered the charged cathode tape into multi-ply, Ni foam–held NiOOH cathodes. J.F.P. and I.R.P. assembled NiOOH–3D Zn coin cells and ran charge-discharge cycling experiments. J.F.P., C.N.C., and J.W.L. analyzed electrochemical data. J.F.P. and C.N.C. collected electron micrographs of the postcycled Zn sponge anodes. M.F.B. served as project and systems engineering lead for the project. M.M. served as lead for battery design engineering. D.R.R., J.F.P., and J.W.L. are inventors on patent applications US 2014/0147757 and US 2016/0093890, submitted by the U.S. Department of the Navy, which covers the fabrication of 3D interconnected Zn sponges and their use as electrodes in electrochemical cells.
View Abstract

Navigate This Article