Zinc-Sponge Battery Electrodes that Suppress Dendrites

Brandon  J. Hopkins; Megan  B. Sassin; Joseph  F. Parker; Jeffrey  W. Long; Debra  R. Rolison

doi:10.3791/61770

JoVE Journal > Chemistry

Chemistry

Zinc-Sponge Battery Electrodes that Suppress Dendrites

Published: September 29, 2020

doi:

10.3791/61770

Brandon J. Hopkins¹, Megan B. Sassin², Joseph F. Parker², Jeffrey W. Long², Debra R. Rolison²

¹Naval Research Laboratory, National Research Postdoctoral Associate,U.S. Naval Research Laboratory, Washington, DC, 20375, United States, ²Surface Chemistry Branch,U.S. Naval Research Laboratory, Washington, DC, 20375, United States

Summary

The goal of the reported protocols is to create rechargeable zinc-sponge electrodes that suppress dendrites and shape change in zinc batteries, such as nickel–zinc or zinc–air.

Abstract

We report two methods to create zinc-sponge electrodes that suppress dendrite formation and shape change for rechargeable zinc batteries. Both methods are characterized by creating a paste made of zinc particles, organic porogen, and viscosity-enhancing agent that is heated under an inert gas and then air. During heating under the inert gas, the zinc particles anneal together, and the porogen decomposes; under air, the zinc fuses and residual organic burns out, yielding an open-cell metal foam or sponge. We tune the mechanical and electrochemical properties of the zinc sponges by varying zinc-to-porogen mass ratio, heating time under inert gas and air, and size and shape of the zinc and porogen particles. An advantage of the reported methods is their ability to finely tune zinc-sponge architecture. The selected size and shape of the zinc and porogen particles influence the morphology of the pore structure. A limitation is that resulting sponges have disordered pore structures that result in low mechanical strength at low volume fractions of zinc (<30%). Applications for these zinc-sponge electrodes include batteries for grid-storage, personal electronics, electric vehicles, and electric aviation. Users can expect zinc-sponge electrodes to cycle up to 40% depth of discharge at technologically relevant rates and areal capacities without the formation of separator-piercing dendrites.

Introduction

The purpose of the reported fabrication methods is to create zinc (Zn) sponge electrodes that suppress dendrite formation and shape change. Historically, these problems have limited the cycle life of Zn batteries. Zinc-sponge electrodes have resolved these issues, enabling Zn batteries with longer cycle lives¹^,²^,³^,⁴^,⁵^,⁶. The sponge structure suppresses dendrite formation and shape change because (1) the fused Zn framework electrically wires the entire volume of the sponge; (2) the pores hold zincate near the Zn-sponge surface; and (3) the sponge has a high surface area that decreases local current density below values identified to sprout dendrites in alkaline electrolytes⁷. However, if sponge surface area is too high, substantial corrosion occurs⁵. If the sponge pores are too large, the sponge will have a low volumetric capacity⁵. Also, if the sponge pores are too small, the Zn electrode will have insufficient electrolyte to access Zn during discharge, resulting in low power and capacity⁵^,⁶.

The rationale behind the reported fabrication methods is to create Zn sponges with appropriate sponge porosities and pore diameters. Experimentally, we find that Zn sponges with porosities from 50 to 70% and pore diameters near 10 µm cycle well in full-cell batteries and display low corrosion rates⁵. We note that existing methods to manufacture commercial metal foams fail to achieve similar morphologies on these length scales⁸, so the reported fabrication methods are needed.

The advantages of the methods reported here over alternatives are characterized by fine control of sponge features and by the ability to fabricate large, dense Zn sponges with technologically relevant areal-capacity values⁵^,⁶^,⁹^,¹⁰. Alternative methods to create Zn foams may be unable to create comparable 10 µm pores with sponge porosities near 50%. Such alternatives may, however, require less energy to fabricate because they avoid high-temperature processing steps. Alternative processes include the following strategies: cold sintering Zn particles¹¹, depositing Zn on three-dimensional host structures¹²^,¹³^,¹⁴^,¹⁵^,¹⁶^,¹⁷, cutting Zn foil into two-dimensional foams¹⁸, and creating Zn foams via spinodal decomposition¹⁹ or percolation dissolution²⁰.

The context of the reported methods in the wider body of the published literature is primarily established by work from Drillet et al.²¹. They adapted methods of fabricating porous ceramics to create one of the earliest reported three-dimensional, albeit fragile, Zn foams for batteries. These authors, however, failed to demonstrate rechargeability, likely because of the poor connectivity between the Zn particles. Prior to rechargeable Zn-sponge electrodes, the best alternative to a Zn foil electrode was a Zn-powder electrode, wherein Zn powder is mixed with a gel electrolyte. Zinc-powder electrodes are commercially used in primary alkaline batteries (Zn–MnO₂) but have poor rechargeability because Zn particles become passivated by Zn oxide (ZnO), which can increase local current density that spurs dendrite growth³^,²². We note that there are other dendrite-suppression strategies that do not involve foam or sponge architectures²³^,²⁴.

The reported Zn-sponge fabrication methods require a tube furnace, sources of air and nitrogen gas (N₂), and a fume hood. All steps can be performed at a lab desk without environmental control, but exhaust from the tube furnace during heat treatment should be piped to a fume hood. Resulting electrodes are appropriate for those interested in creating rechargeable Zn electrodes capable of high areal capacity (> 10 mAh cm_geo^–2)⁶.

The first reported fabrication method is an emulsion-based route to create Zn-sponge electrodes. The second, is an aqueous-based route. An advantage of the emulsion route is its ability to create Zn paste that, when dried, is easy to demold from a mold cavity. A disadvantage is its reliance on expensive materials. For the aqueous route, sponge preforms can be challenging to demold, but this process uses inexpensive and abundant materials.

Both methods involve mixing Zn particles with a porogen and viscosity-enhancing agent. The resulting mixture is heated under N₂ and then breathing air (not synthetic air). During heating under N₂, the Zn particles anneal and the porogen decomposes; under breathing air, the annealed Zn particles fuse and the porogen burns out. These processes yield metal foams or sponges. The mechanical and electrochemical properties of the Zn sponges can be tuned by varying Zn-to-porogen mass ratio, heating time under N₂ and air, and size and shape of the Zn and porogen particles.

Protocol

1. An emulsion-based method to create Zn-sponge electrodes

Add 2.054 mL of deionized water to a 100 mL glass beaker.
Add 4.565 mL of decane to the beaker.
Stir in 0.1000 ± 0.0003 g of sodium dodecyl sulfate (SDS) until dissolved.
Stir in 0.0050 ± 0.0003 g of water-soluble medium viscosity carboxymethyl cellulose (CMC) sodium salt by hand for 5 min or until the CMC is fully dissolved.
NOTE: Use plastic or plastic-coated stirring tools. Stirring with tools with a metallic surface can adversely affect resulting Zn sponges.
Stir in 0.844 ± 0.002 g of water-insoluble preswollen carboxymethyl cellulose resin.
NOTE: This type of water-insoluble resin is expensive (USD$420 kg^–1)⁶.
Stir this mixture at 1,000 rpm for 5 min using an overhead paddle stirrer equipped with a plastic paddle.
Pour 50 g of Zn powder (average particle size of 50 µm, containing 307 ppm of bismuth and 307 ppm of indium for corrosion suppression) into the beaker while the overhead stirrer continues to spin at 1,000 rpm.
Continue to stir the Zn paste for an additional 5 min at the same rate, 1,000 rpm.
Stop the stirrer, remove the beaker, and outgas the mixture by placing the beaker and its contents under vacuum for 5 min in a desiccator at room temperature.
Portion the Zn paste into polypropylene molds (~10 mm in diameter and ~5 mm in height) and let them dry in open air overnight. The shape of the mold dictates the form of the dried paste and resulting Zn sponges.
NOTE: Mold size and shape can vary. Past experiments⁵ successfully use cylindrical molds with diameters near 10 mm. Fill the Zn paste up to a height of 5 mm or less. The shorter the height, the shorter the required drying time. See Table of Materials for commercially available molds.
Carefully remove the dried Zn paste preforms from the molds and place them into a mesh casing that rests on a notched alumina holder⁵^,⁶.
NOTE: Fabricate mesh casing, for instance, by bending a perforated-brass sheet into a cylinder with a diameter that is slightly larger than the desired diameter of the Zn-sponge electrode. Spray the perforated-metal sheet with boron-nitride lubricant after bending into a desired shape.
Place the assembly into a tube furnace (67 mm in diameter) with ports to flow gas in and out of the tube.
NOTE: Use one port (the entrance port) to pipe gas into the furnace. Use the other (the exit port) to vent gas out of the tube furnace into a fume hood.
Pipe N₂ gas into the furnace for 30 min at a rate of 5.7 cm∙min^–1 to purge the furnace of air.
NOTE: Step 1.13 can be achieved by connecting a tank of N₂ gas with a digitally controlled flow meter to a tube connected to one of the entrance ports. Gas flow meters can be controlled manually or by a computer.
Throttle the N₂ gas to a constant rate of 2.8 cm∙min^–1 after the 30 min purge.
Program the furnace to increase temperature linearly from 20 to 369 °C over the course of 68 min, hold at 369 °C for 5 h, rise linearly from 369 to 584 °C over the course of 105 min, and then turn off.
Start the furnace program while the N₂ gas continues to flow.
Manually stop the N₂-gas flow after the 5 h temperature hold and pipe in breathing air at 2.8 cm∙min^–1.
NOTE: Step 1.17 can be achieved by connecting a tank of breathing air (not synthetic air) with a digitally controlled flow meter to a tube connected to an additional entrance port.
Once the heating program stops, let the furnace cool to room temperature without active cooling, but keep the breathing air flowing.
Remove the cooled Zn sponges and saw them and/or sand them to desired dimensions.
NOTE: A variety of sawing tools can be used such as hand-held rotary saws or vertical band saws. Abrasive or diamond blades are appropriate.

2. An aqueous-based method to create Zn-sponge electrodes

Add 10.5 mL of deionized water to a 100-mL glass beaker.
Stir in 0.120 ± 0.001 g of water-soluble high-viscosity cellulose gum, also known as carboxymethyl cellulose (CMC) sodium salt.
NOTE: Use plastic or plastic-coated stirring tools. Stirring with tools with a metallic surface can adversely affect resulting Zn sponges.
Vortex and stir this mixture by hand for 5 min or until the CMC is dissolved.
Stir in 2.400 ± 0.001 g of corn starch while vortexing for an additional 2 min.
Stir in 120.00 ± 0.01 g of Zn powder (average particle size of 50 µm, containing 307 ppm of bismuth and 307 ppm of indium for corrosion suppression) while vortexing for an additional 2 min.
Press the resulting Zn paste into desired mold cavities.
NOTE: Mold size and shape can vary. Past experiments⁶ successfully use cylindrical molds with diameters near 10 mm. Fill the Zn paste up to a height of 50 mm or less. The aqueous Zn paste is dryer than the emulsion Zn paste, so the aqueous version can be used to make larger sponges that require less drying time. The shorter the height, the shorter the required drying time. The mold needs to be able to split in half as the aqueous Zn paste minimally contracts after drying, unlike the emulsion Zn paste. Unsalted butter can be used to lubricate the molds before pressing in the aqueous Zn paste to aid in demolding. Figure 1A shows the custom-machined molds packed with Zn paste following the aqueous-based protocol. Figure 1B shows the hand-made mesh casing, notched alumina holder, and resulting Zn sponge made using the aqueous-based method.
Leave the Zn-paste-filled molds to dry overnight at 70 °C in open air in a furnace.
Follow the same handling and heat treatment steps (1.11–1.19) described for the emulsion-based method.

Representative Results

Resulting, fully heat-treated, emulsion-based Zn sponges have densities of 2.8 g∙cm^–3 while aqueous-based sponges approach 3.3 g∙cm^–3. During heating under air, a layer of ZnO forms on the Zn surfaces, which should have a thickness of 0.5–1.0 µm (observed using scanning electron microscopy)⁵. The solid in the resulting sponges should be 72% Zn (emulsion version) or 78% Zn (aqueous version) with the remainder being ZnO (measured by X-ray diffraction)⁶. Both sponges should have porosities near 50%, pore-diameter distributions centered on 10 µm, and specific surface areas of 4.0 m²∙g^–1 (measured via mercury-intrusion porosimetry)⁶. The tensile strength of both sponges should be 1.1–1.2 MPa (measured with diametral compression)⁵^,⁶. We note that the sponges should be rigid and brittle. Cross-sections of the Zn sponges should look similar to those shown in Figure 2A,B. If all the properties of the fabricated sponges fall within the provided ranges, the result is positive; if not, the result is negative.

With the stated properties, Zn sponges cycle well in properly constructed batteries. Their performance also depends on the counter electrode, electrolyte, separator, and cell construction; construction of reliable full cells is beyond the scope of this paper. To test the electrochemical validity of Zn sponges, we recommend harvesting commercial counter electrodes from nickel–metal hydride batteries⁵^,⁶. Shape a Zn-sponge to have a 10 mm diameter and 0.5-mm thickness. Cycle this sponge at 20 mA∙cm_geo^–2 (geometric area) for discharge and 10 mA∙cm_geo^–2 for charge in a nickel–zinc cell as described in the literature⁵. Assuming appropriate construction, the Zn-sponge electrode should show cycling stability at a gravimetric capacity of 328 mA∙h∙g_sponge^-1 (per gram of ZnO@Zn-sponge electrode) as shown in Figure 2C, which maps to 43% depth of discharge (the quotient of gravimetric discharge capacity with respect to every atom of Zn in the electrode divided by the theoretical gravimetric capacity of Zn). After extensive cycling, no dendrites are observed by scanning electron microscopy (Figure 3). X-ray diffraction can be used to track the state of charge of the Zn-sponge electrode by monitoring Zn and ZnO reflections¹. We note the surface of the Zn sponge undergoes restructuring during cycling. The deeper the level of discharge and the greater the cycle life, the greater the amount of restructuring⁵. These factors contribute to the difference in surface morphology shown in Figure 3A,B. If this rechargeable capacity is achieved, the result is positive; if not, the result is negative and could be caused by either the Zn sponge, poor cell construction, or failure of other cell components.

Figure 1: Zinc sponges before and after heat treatment using the aqueous-based method. (A) Photo of custom-machined molds made from Delrin or polyoxymethylene (POM) that is packed with Zn paste before heating occurs. (B) Photo of hand-made mesh casing, notched alumina holder, and resulting Zn sponge after heat treatment. Please click here to view a larger version of this figure.

Figure 2: Zinc-sponge morphology and electrochemical performance. Scanning electron micrographs of cross-sectioned (A) emulsion-based Zn sponge and (B) aqueous-based Zn sponge. (C) Voltage versus time of an emulsion-based sponge cycled in a nickel–zinc cell discharged at 20 mA∙cm_geo^–2 and charged at 10 mA∙cm_geo^–2 with a gravimetric capacity of 328 mA∙h∙g_sponge^–1. Data adapted from Hopkins et al.⁵^,⁶. Please click here to view a larger version of this figure.

Figure 3: Zinc-sponge electrodes suppress dendrite formation. Emulsion-based Zn sponge (A) before and (B) after electrochemical cycling. Data adapted from Hopkins et al.⁵. Please click here to view a larger version of this figure.

Discussion

Modifications and troubleshooting associated with these protocols include filling the freshly mixed Zn paste into a mold cavity. Care should be taken to avoid air pockets. Unwanted voids can be decreased by tapping the mold after filling or while filling. Because the aqueous Zn paste is dry, pressure can be applied directly to the Zn paste to push out air pockets while filling up the mold cavity.

A limitation of the methods is that Zn-sponge pore structure is disordered, but the Zn and porogen particle sizes can be used to alter pore morphology. A more ordered and potentially stronger and lighter Zn sponge may be fabricated using additive manufacturing. The mechanical and electrochemical properties of resulting Zn sponges, however, can be tuned by varying Zn-to-porogen mass ratio and the size and shape of the Zn and porogen particles⁵^,⁶. Another potential limitation is that the dried Zn-paste can be fragile, so transferring it into a mesh casing may be challenging and limit Zn-sponge size.

The significance of these methods with respect to existing methods is that resulting Zn sponges achieve long cycle life with high volumetric and areal capacities⁵^,⁶. Resulting Zn sponges are also mechanically robust⁵^,⁶.

Future applications of the processes could, in principle, be adapted to create other metal foams for batteries or other applications. For example, iron, magnesium, or aluminum foams may be useful as anodes for metal–air batteries²⁵^,²⁶^,²⁷. Zn-sponge electrodes, in particular, can be used to create batteries for a range of applications that include wearables, grid storage, personal electronics, electric vehicles, and electric aviation²⁸.

A critical step, which may also require modification or troubleshooting, is the heating process. Furnace temperatures can vary. The heating time under N₂, near but below the melting point of Zn, anneals the Zn particles together. The heating time under air burns out the residual porogen, fuses the Zn, and forms a ZnO layer. If the Zn particles appear to be fusing improperly, increase the heating time under N₂. If the ZnO layer is too thick, decrease the heating time under air by 10 min or more until the desired thickness of thermal oxide is achieved.

We note that a thick layer of ZnO enhances the mechanical properties of the Zn sponge but also decreases the immediately useable capacity of the Zn electrode. The Zn electrode can be charged by electrochemically converting ZnO to metallic Zn. However, stable cycling at 40% depth of discharge can be achieved without any precharge⁵. If the ZnO layer is too thin, the Zn sponge can crumble during handling⁵.

Disclosures

The authors have nothing to disclose.

Acknowledgements

This research was funded by the United States Office of Naval Research.

Materials

Corn starch	Argo	Not applicable	This acts as a porogen and viscosity-enhancing agent.
Decane	MilliporeSigma	D901
Medium viscosity water-soluble carboxymethyl cellulose (CMC) sodium salt	MilliporeSigma	C4888-500G	This CMC acts primarily as a viscosity-enhancing agent.
Overhead stirrer	Caframo Lab Solutions	BDC3030
Small cylindrical models for Zn sponges	VWR	66014-358	The caps of the vials can be used as molds.
Sodium dodecyl sulfate	MilliporeSigma	436143
Water-insoluble IonSep CMC 52 preswollen carboxymethyl cellulose resin	BIOpHORETICS	B45019.01	This CMC acts as a porogen and viscosity-enhancing agent.
Zn powder	EverZinc	Custom order

References

Parker, J. F., et al. Retaining the 3D Framework of Zinc Sponge Anodes upon Deep Discharge in Zn-Air Cells. ACS Applied Materials & Interfaces. 6 (22), 19471-19476 (2014).
Parker, J. F., Chervin, C. N., Nelson, E. S., Rolison, D. R., Long, J. W. Wiring zinc in three dimensions re-writes battery performance-dendrite-free cycling. Energy & Environmental Science. 7, 1117-1124 (2014).
Parker, J. F., et al. Rechargeable nickel-3D zinc batteries: An energy-dense, safer alternative to lithium-ion. Science. 356 (6336), 415-418 (2017).
Ko, J. S., et al. Robust 3D Zn sponges enable high-power, energy-dense alkaline batteries. ACS Applied Energy Materials. 2 (1), 212-216 (2018).
Hopkins, B. J., et al. Fabricating architected zinc electrodes with unprecedented volumetric capacity in rechargeable alkaline cells. Energy Storage Materials. 27, 370-376 (2020).
Hopkins, B. J., et al. Low-cost green synthesis of zinc sponge for rechargeable, sustainable batteries. Sustainable Energy & Fuels. 4, 3363-3369 (2020).
Yufit, V., et al. Operando Visualization and Multi-scale Tomography Studies of Dendrite Formation and Dissolution in Zinc Batteries. Joule. 3 (2), 485-502 (2019).
Ashby, M. F., et al. . Metal Foams: A Design Guide. , (2000).
Parker, J. F., Ko, J. S., Rolison, D. R., Long, J. W. Translating Materials-Level Performance into Device-Relevant Metrics for Zinc-Based Batteries. Joule. 2 (12), 2519-2527 (2018).
Hopkins, B. J., Chervin, C. N., Parker, J. F., Long, J. W., Rolison, D. R. An areal-energy standard to validate air-breathing electrodes for rechargeable zinc-air batteries. Advanced Energy Materials. 10 (30), 2001287 (2020).
Jayasayee, K., et al. Cold Sintering as a Cost-Effective Process to Manufacture Porous Zinc Electrodes for Rechargeable Zinc-Air Batteries. Processes. 8 (5), 592 (2020).
Chamoun, M., et al. . NPG Asia Materials. 7, 178 (2015).
Kang, Z., et al. 3D Porous Copper Skeleton Supported Zinc Anode toward High Capacity and Long Cycle Life Zinc Ion Batteries. ACS Sustainable Chemistry & Engineering. 7 (3), 3364-3371 (2019).
Yu, J., et al. Ag-Modified Cu Foams as Three-Dimensional Andoes for Rechargeable Zinc-Air Batteries. ACS Applied Nano Materials. 2 (5), 2679-2688 (2019).
Stumpp, M., et al. Controlled Electrodeposition of Zinc Oxide on Conductive Meshes and Foams Enabling Its Use as Secondary Anode. Journal of The Electrochemical Society. 165 (10), 461-466 (2018).
Stock, D., et al. Design Strategy for Zinc Anodes with Enhanced Utilization and Retention: Electrodeposited Zinc Oxide on Carbon Mesh Protected by Ionomeric Layers. ACS Applied Energy Materials. 1 (10), 5579-5588 (2018).
Li, C., et al. Spatially homogeneous copper foam as surface dendrite-free host for zinc metal anode. Chemical Engineering Journal. 379, 122248 (2020).
Zhou, Z., et al. Graphene oxide-modified zinc anode for rechargeable aqueous batteries. Chemical Engineering Science. 194, 142-147 (2019).
McDevitt, K. M., Mumm, D. R., Mohraz, A. Improving Cyclability of ZnO Electrodes through Microstructural Design. ACS Applied Energy Materials. 2 (11), 8107-8117 (2019).
Wang, C., Zhu, G., Liu, P., Chen, Q. Monolithic Nanoporous Zn Anode for Rechargeable Alkaline Batteries. ACS Nano. 14 (2), 2404-2411 (2020).
Drillet, J. F., et al. Development of a Novel Zinc/Air Fuel Cell with a Zn Foam Anode, a PVA/KOH Membrane and a MnO2/SiOC-based Air Cathode. ECS Transactions. 28, 13-24 (2010).
Bozzini, B., et al. Morphological evolution of Zn-sponge electrodes monitored by in situ X-ray computed microtomography. ACS Applied Energy Materials. , (2020).
Yang, Q., et al. Do Zinc Dendrites Exist in Neutral Zinc Batteries: A Developed Electrohealing Strategy to In Situ Rescue In-Service Batteries. Advanced Materials. 31, 1903778 (2019).
Yang, Q., et al. Hydrogen-Substituted Graphdiyne Ion Tunnels Directing Concentration Redistribution for Commercial-Grate Dendrite-Free Zinc Anodes. Advanced Materials. 32, 2001755 (2020).
Narayanan, S. R., et al. Materials challenges and technical approaches for realizing inexpensive and robust iron-air batteries for large-scale energy storage. Solid State Ionics. 216, 105-109 (2012).
Xiong, H., et al. Effects of Heat Treatment on the Discharge Behavior of Mg-6wt.%Al-1wt.%Sn Alloy as Anode for Magnesium-Air Batteries. Journal of Materials Engineering and Performance. 26, 2901-2911 (2017).
Hopkins, B. J., Shao-Horn, Y., Hart, D. P. Suppressing corrosion in primary aluminum-air batteries via oil displacement. Science. 362, 658-661 (2018).
Hopkins, B. J., Long, J. W., Rolison, D. R. High-Performance Structural Batteries. Joule. , (2020).

Play Video

PDF

DOI

DOWNLOAD MATERIALS LIST

Cite This Article

Hopkins, B. J., Sassin, M. B., Parker, J. F., Long, J. W., Rolison, D. R. Zinc-Sponge Battery Electrodes that Suppress Dendrites. J. Vis. Exp. (163), e61770, doi:10.3791/61770 (2020).

Automatically Generated