Increasing the Electrolyte Capacity of Alkaline Zn–Air FuelCells by Scavenging Zincate with Ca(OH)2

Aaron L. Zhu,*[a] Darek Duch,[d] Gregory A. Roberts,[e] Sasha X. X. Li,[f] Haijiang Wang,[f]

Konrad Duch,[g] Eunjoo Bae,[h] Kwang S. Jung,[h] David Wilkinson,[a] andSergei A. Kulinich *[b, c]

1. Introduction

After decades of development, lithium-ion batteries are notonly being mass produced, but are also gaining acceptance asa viable energy source in electric and hybrid-electric vehicles.Yet, substantial publicly and privately funded research and de-

velopment efforts are still underway to address concernsabout their cost, safety, and longevity.[1] Although polymer ex-change membrane fuel cells (PEMFCs) do show certain advan-tages over Li-ion batteries, especially when applied in electricvehicles,[2] their state-of-the-technology is still quite far frommass manufacturing and commercialization. Thus, there are nonear-term alternatives that can outperform the Li-ion batterieswith their combination of high energy and power densities.Hence, alternative chemistries that are inherently safer and lessexpensive attract increasingly more attention.[3, 4] Among them,metal–air batteries and fuel cells represent a new class ofpromising alternative power sources, owing to their remarka-bly high theoretical energy output.[5–8]

Of all metal–air batteries/fuel cells that are currently underdevelopment, Zn–air batteries/fuel cells have the most promis-ing potential, as they offer an appealing combination of highenergy density, addressable technical challenges, low cost, andinherent safety.[8–11] A renowned example is the low-power pri-mary zinc–air hearing-aid battery.[12] With a design combiningaspects of conventional batteries and modern fuel cell designs,the current remodeled Zn–air fuel cell (ZAFC) offers evenhigher energy densities than its traditional Zn–air batterycounterpart.[13–15] Raw material costs for its components (zincmetal, aqueous alkaline electrolytes, inexpensive separator ma-terials, and non-precious metal catalysts for cathode) are low.However, three major barriers keep hindering the ZAFC com-mercialization: 1) degradation of the air cathode, 2) low elec-trolyte capacity, and 3) difficulties with recharging zinc anodesbecause of zinc oxide formation.

Unlike Nafion in PEMFCs, mature alkaline polymer mem-branes have not been fully developed so far. Therefore, potas-

The use of calcium hydroxide for scavenging zincate species isdemonstrated to be a highly effective approach for increasingthe electrolyte capacity and improving the performance of thezinc–air fuel cell system. A fundamental approach is estab-lished in this study to quantify the formation of calcium zin-cate as the product of scavenging and the amount of watercompensation necessary for optimal performance. The goodagreement between predicted and experimental results provesthe validity of the proposed theoretical approach. By applyingthe results of theoretical predictions, both the electrolyte ca-

pacity and the cell longevity have been increased by morethan 40 %. It is also found that, using Ca(OH)2 to scavenge zin-cate species in concentrated KOH solutions, affects mostly theremoval of zincate, rather than ZnO, from the electrolyte,whereas the presence of excess, free, mobile H2O plays a keyrole in dissolving ZnO and forming zincate. The results ob-tained in this study demonstrate that the proposed approachcan widely and effectively be applied to all zinc–air cell sys-tems during their discharge cycle.

[a] Dr. A. L. Zhu, Prof. D. WilkinsonDepartment of Chemical and Biological EngineeringUniversity of British Columbia, 2360 East MallVancouver, BC, V6T 1Z4 (Canada)E-mail : researchman12001@yahoo.com

[b] Dr. S. A. KulinichInstitute of Innovative Science and TechnologyTokai University, Hiratsuka, Kanagawa, 259-1292 (Japan)E-mail : skulinich@tokai-u.jp

[c] Dr. S. A. KulinichSchool of Engineering and Applied ScienceAston University, Birmingham, B4 7ET (UK)

[d] D. DuchChemistry Division, Power Air Canada Corp.Suite 310, 4475 Wayburne DriveBurnaby, BC, V5G 4X4 (Canada)

[e] Dr. G. A. RobertsAmprius Inc. , 1430 O’Brien DriveSuite C, Menlo Park, CA, 94025 (USA)

[f] Dr. S. X. X. Li, Dr. H. WangEnergy, Mining and EnvironmentNational Research Council of Canada4250 Wesbrook Mall, VancouverBC, V6T 1W5 (Canada)

[g] K. DuchFaculty of Mathematics, Waterloo UniversityWaterloo, ON, N2L 3G1 (Canada)

[h] Dr. E. Bae, Dr. K. S. JungH-plus Holding Ltd, 6F J-Tower, 158-3, Seokchon-dongSongpa-gu, Seoul, 138-844 (Korea)

ChemElectroChem 2015, 2, 134 – 142 � 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim134

ArticlesDOI: 10.1002/celc.201402251

sium hydroxide is still the first choice electrolyte for ZAFCs. Ingeneral, Zn electrochemistry in alkaline electrolytes seemsrather simple. The zinc–air chemistry is generally described byReactions (1)–(3):[14, 15]

Anodic reaction : Znþ 2 OH� ! ZnOþ H2Oþ 2 e� ð1Þ

Cathodic reaction : 1=2 O2 þ H2Oþ 2 e� ! 2 OH� ð2Þ

Overall reaction : Znþ 1=2 O2 ! ZnO ð3Þ

Accordingly, ZnO precipitation [given by Reaction (3)] mayseem favorable. If it were the case, the electrolyte compositionshould not change during battery operation. However, in reali-ty, the above simple equations hide all of the complexity of re-actions involving numerous intermediates.[16] Zinc oxidationproduces soluble species [Reaction (4)] , that is, the tetrahy-droxyzincate anion (denoted as “zincate” thereafter), which canundergo a series of progressive reactions until ZnO eventuallyprecipitates [Reaction (5)] . Even though ZnO precipitation isthermodynamically favored, the zincate anion is metastable,and this is known to hinder ZnO precipitation. In fact, supersa-turated zincate solutions with concentrations two to threetimes higher than its equilibrium solubility limit have beenreported:[17]

Znþ 4 OH� ! ZnðOHÞ42� þ 2 e� ð4Þ

ZnðOHÞ42� ! ZnOþ H2Oþ 2 OH� ð5Þ

Also, zincate is always present (at least up to its saturationlevel) as the first intermediate product of the ZnO formation.The latter is known to form through a series of intermediates,resulting from different steps that lead to changes in the elec-trolyte balance among zinc ions, OH� , and water.[15, 16] ZnO pre-cipitation might be welcomed as it decreases the electrolyteviscosity and, therefore, should improve the performance ofthe battery, which is technically attractive. However, as ZnOprecipitates as an insulating paste and tends to coat the cellcomponents, it can reduce zinc particle-to-particle conductivityand lead to certain mechanical problems. Moreover, the KOHconcentration is lowest at the surface of the zinc particles(where it is being depleted), and low pH favors ZnO precipita-tion. So, it is inevitable that ZnO precipitation will preferentiallyoccur on the anode and anodic current collector. Hence, im-proving the electrolyte capacity and minimizing the negativeeffect of ZnO deposition become important issues in ZAFC en-gineering and development.

It is well known that electrolyte additives, such as potassiumsilicate, inhibit ZnO precipitation. When an additive is used, allelectrochemically dissolved zincate species are often assumedto remain in the solution. At least, additive species inhibit anyheterogeneous precipitation of ZnO on surfaces, although it ispossible that some homogeneous nucleation may occur andthen lead to the formation of fine particles. This avoids manyproblems caused by ZnO precipitation, but it is clear that Reac-tion (4) leads to a certain depletion of hydroxide in the electro-lyte, owing to different reaction rates of Reactions (4) and (5)

and to the accumulation of zincate ions. This decreases theelectrolyte conductivity, increases its viscosity, and causes anenergy efficiency loss through concentration polarization.

The loss in performance associated with the accumulation ofzincate ions and depletion of hydroxide ions has major impli-cations. First, the overall ZAFC performance will decay. Second,the electrolyte has an effective capacity, which contributessubstantially to the weight and volume requirements of theZAFC. If the dissolved zincate could be stored more efficiently,the weight and volume of required electrolyte would be dra-matically reduced. To investigate whether the storage efficien-cy of zincate could be improved, we report herein on a newapproach that could possibly provide a simple and efficient so-lution to the problem. The approach is based on the use ofcalcium hydroxide as a compound for trapping and storing thedissolved zincate as a solid.[18–20] The chemistry described inthis paper is relevant to any Zn battery/fuel cell with an alka-line electrolyte. Our preferred approach is to provide the calci-um hydroxide in the electrolyte storage tank to minimize inter-ference with the zinc anode, but some battery designs may in-corporate calcium hydroxide as a component in the zincanode. It is also supported by recent findings, which show thatcalcium zincate is beneficial for the zinc electrode, outperform-ing the mixture of Ca(OH)2 with ZnO in the Ni–Zn battery[21]

and acting as a good catalyst for biodiesel production throughethanolysis.[22] According to Reaction (6),[19, 20] the product thatforms in the system upon adding calcium hydroxide is calciumzincate:

CaðOHÞ2 þ 2 ZnðOHÞ42� þ 2 H2O

! CaðOHÞ2 � 2 ZnðOHÞ2 � 2 H2O þ 4 OH�ð6Þ

Calcium hydroxide is an attractive material for zincate re-moval for several reasons. First, it is only sparingly soluble inconcentrated KOH. The precipitation of any soluble material isnot encouraged during cell running, as it may block the aircathode pores, resulting in a loss of the cell performance.Second, the product, calcium zincate, is also poorly soluble inconcentrated KOH. Therefore, the starting material and the re-action product can be confined in a desired area of the ZAFC.Third, based on some preliminary results obtained in the trialand error tests with mechanically rechargeable zinc fuel cells,calcium hydroxide is expected to have reasonably fast reactionkinetics when reacted with zincate species in a concentratedalkaline electrolyte. Fourth, from the chemical point of view,one Ca atom traps two Zn atoms, which makes Ca(OH)2 an effi-cient and effective material for removing zincate anions. How-ever, three important issues still need to be addressed:

1) Reports regarding the composition of calcium zincate thatforms during chemical reactions between Ca(OH)2 and zin-cate ions are quite controversial. There are two causes pos-sibly generating the controversy. One is the lack of a preciseinterpretation of the crystallographic patterns of calciumzincate products, whereas the other is the uncertaintyabout calcium zincate obtained by using the two different

ChemElectroChem 2015, 2, 134 – 142 www.chemelectrochem.org � 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim135

Articles

prevailing preparative methods, that is, electrochemicalanodic dissolution of zinc in concentrated KOH and chemi-cal dissolution of ZnO in water-diluted KOH. It is imperativeto clarify the exact stoichiometric form of the calcium zin-cate product, as this is pertinent to the working efficiencyof Ca(OH)2 in ZAFCs.

2) It is not yet clear how much Ca(OH)2 should be added intothe KOH electrolyte containing zincate, as well as howmuch extra electrolyte capacity can be acquired by scav-enging zincate.

3) The effect of water compensation on maintaining the elec-trolyte capacity is also generally unknown.

This study presents the results of work that uses Ca(OH)2 asa scavenger of zincate species, with the aim to significantly in-crease the electrolyte capacity and the longevity of the cell.The aforementioned three key issues are specifically addressedin this report.

Experimental Section

Electrochemical Tests

A five-cell ZAFC bipolar stack from Power Air Corporation, whichuses inexpensive Zn pellets as the anode material, was used forthe electrochemical experiments (an expanded view of an individu-al cell is shown in Figure 1). There are many creative designs forZn–air systems, but this particular study was performed for a ZAFCthat uses loose Zn pellets in the anode, an alkaline electrolyte thatis pumped through the cell area from a reservoir, and a non-pre-cious metal oxygen-reduction cathode in which oxygen is obtainedfrom the air like a fuel cell. Fresh zinc can be fed into the cell bya gravity-fed hopper as the zinc dissolves, so the ZAFC used in thisreport could also be classified as a mechanically rechargeable bat-tery. The O-cathode (from eVionyx) and Zn pellet (<1 mm cut wireshot, Pellets LLC) were used as the cathode and the anode, respec-tively. KOH electrolyte (original source: 45 wt %, in liquid form with98 wt % purity, Anachemia) was circulated by using a pump con-nected to the electrolyte tank. A custom-configured Arbin BT2000

battery-testing station (Arbin) was employed for data acquisitionand control. Cell polarization and chronopotentiometry curveswere used to determine the electrolyte capacity.

Analysis of Zn Content in KOH Electrolyte

The total amount of zinc dissolved in the electrolyte was measuredwith atomic absorption spectrometry (AAS, AAnalyst 200, Perkin-Elmer). AAS was selected as the analytical technique for measuringZn content because of its high instrumental sensitivity. Comparedto conventional titration, this technique permits to avert the matrixinterferences assigned to high KOH concentrations by applyinghigh dilution factors (250- to 1000-fold dilution). An air–acetylenemixture was used as the flame source. A set of five standards rang-ing from 0.5 (0.0 was used as the blank) to 16 ppm of Zn was usedto attain a linear absorbance–concentration curve, with a correla-tion coefficient greater than or equal to 0.995.[23–25] To control thestability of the system (reading drift), a standard check was appliedafter every ten samples (starting from the standard sample withthe lowest concentration).

Titrations were mainly used to monitor the OH� species in solu-tions and were performed by using a R24906-52 type Schott Uni-versal Titrator (Cole-Parmer). The detailed measurement proceduresare reported elsewhere.[26]

Calcium Zincate Formation and Crystal StructureIdentification

The nepheloid layer, containing suspended and/or precipitatedwhite powders that could be calcium zincate products, was firstpumped into a volumetric collecting flask, filtered by a MillivacMini filtering system (with pore size �0.45 mm, Millipore), and final-ly dried in a vacuum oven (Cole-Parmer) at 325 K for 12 h. Thedried powder was then identified by X-ray diffraction (XRD) byusing a Bruker D8 advanced diffractometer (CuKa1 irradiation, l=1.5406 �) with the scattering angle (2q) scanned from 58 to 908 atthe rate of 0.028 s�1. Prior to the XRD examinations, the diffractom-eter was calibrated by using an Al2O3 NIST-1976 standard (Vantec).

ZnO Detection

The presence of ZnO in the precipitates was examined by usingXRD and energy-dispersive X-ray spectroscopy (EDS), with thelatter unit being attached to the scanning electron microscopy(SEM) instrument (FEI Tecnai G2).

Light Scattering Measurements

To determine whether colloidal species were formed, as well as todetect and track the occurrence and evolution of ZnO, light scat-tering spectroscopy (Zetasizer Nano-S, Melvern Instruments) wasutilized. Prior to each measurement, the instrument was calibratedby setting the curve of double-distilled water or fresh KOH electro-lyte in a quartz cuvette as the baseline.

Electrolyte Preparation and Ca(OH)2 Addition

Additive-free KOH electrolytes were prepared by diluting a com-mercial (45 wt %) KOH solution (Anachemia, Fe level below20 ppm) with triple-distilled water (with a resistance of18.2 MW cm) to concentrations of 6–10 m. Ca(OH)2 (Anachemia,

Figure 1. Expanded view of the half-cell assembly used in this study.1) Frame, 2) air cathode, 3) copper bar used as current collector, 4) polymerseparator, 5) copper mesh, and 6) T-shaped copper bus bar bonded to thecopper mesh.

ChemElectroChem 2015, 2, 134 – 142 www.chemelectrochem.org � 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim136

Articles

95 % purity) was mixed with 60 wt % polytetrafluoroethylene (PTFE,Dupont, the weight ratio of PTFE to KOH being 1:10). The latterwas used as an adhesive agent and powder stabilizer. The mixturewas put into the electrolyte tank at the beginning of the cell run.Alternatively, Ca(OH)2 was directly added to the electrolyte at theend of cell discharge to form calcium zincate. The solution sampleswere stirred for 12 h, after which the transparent liquid above thefine powdered precipitation was decanted and the precipitate wasdried in a vacuum oven at 325 K for 12 h.

In order to realize the highest possible scavenging efficiency of theZn species dissolved in the cell electrolyte, the work loop used inthe present study was designed as follows:

1) It was necessary to clarify the products of the scavenging reac-tions and then to establish the corresponding reaction equa-tions, so as to accurately obtain the molecular weight ratio ofreactive Ca(OH)2 to Zn dissolved in the electrolyte.

2) It was then necessary to precisely measure the content of Zndissolved in electrolyte.

3) As Zn dissolved in highly concentrated alkaline electrolyte un-dergoes a series of complex transitions, all possible side reac-tions throughout the entire scavenging process had to be con-sidered, especially those related to and involving water. Thiswas dictated by the necessity to control the scavenging reac-tions and increase the scavenging efficiency.

4) Based on the abovementioned reactions, the calculation ap-proach could then be established to quantify the amount ofCa(OH)2 that needed to be added to the electrolyte for scav-enging the dissolved zinc, as well as the amount of water to beadded for compensation.

5) Finally, the scavenging effect was evaluated experimentally,both with and without adding free water, in an industriallyused cell presented in Figure 1, after which the work would goback to processes (1)–(4) to refine the calculation approach.

2. Results and Discussion

2.1. XRD Examination of Zinc–Calcium Product

Ca(OH)2 was put into the electrolyte tank prior to cell opera-tion. The electrolyte samples taken from the cell outlet atvaried cell discharge times were examined by using AAS tomeasure the content of dissolved zinc and evaluate the zincateconcentration during cell operation. The precipitated productswere dried in a vacuum oven and then examined by means ofXRD.

In the database, four calcium zincates, Ca[Zn(OH)3]2·2 H2O,Ca[Zn(OH)3]2·2 H2O, CaZn2(OH)6·2 H2O, and Ca(OH)2·2 Zn(OH)2

·2 H2O are described,[27, 28] and the type of calcium zincate thatemerges during the ZAFC run depends on the preparationmethod, water loss, and pH value. The most compact productof Ca zincate self-polymerization found so far is Ca(OH)2·2 Zn(OH)2·2 H2O.[12, 13] The XRD patterns of Ca–Zn-containingproducts formed after 91 and 200 Ah of cell discharge are pre-sented in Figure 2. The patterns in Figure 2 d and Figure 2 eshow the pronounced peaks emerged at 14.15 and 28.558,which are characteristic of calcium zincate (standard XRDdatabase ICDD file No. 24-222). This strongly suggeststhat the formed calcium zincate has the structure of

Ca(OH)2·2 Zn(OH)2·2 H2O. This finding reaffirms previously per-formed studies on the chemistry of calcium zincate phase.[19, 20]

Thus, the true scavenging reaction observed in this study ap-pears to that depicted in Reaction (6) above.

In order to understand the mechanism of the reaction be-tween Ca(OH)2 and ZnO in greater detail, two additional ex-periments were performed. In the first experiment, 1 g of ZnOand 1 g of Ca(OH)2 were added to 200 mL of double-distilledwater and stirred for 24 h. Meanwhile, the second experimentwas based on Reactions (5) and (6), and therefore 22 g of ZnOwas mixed with 11 g of Ca(OH)2 in 200 mL of 20 wt % KOH andstirred for 24 h. The solutions were then aged for 6 h to trackthe formation of calcium zincate. All the precipitates weredried in a vacuum oven at 325 K, and the dry powders and theelectrolytes used in the experiments were examined by usingAAS and XRD.

In the first experiment, AAS measurements (with a detectionlimit of 2 mg L�1 for Zn) indicated no formation of any detecta-ble amount of calcium zincate. Similarly, XRD measurementsclearly showed that the electrolyte precipitate was a mixture ofCa(OH)2 and ZnO (Figure 3). This suggests that calcium zincateis barely formed through the hydrolysis of ZnO and Ca(OH)2 inthe presence of ample water, as the solubility of both ZnO andCa(OH)2 in water is low (1.6 and 1.73 g L�1 at 20 8C, respective-ly),[29] and at least an energy of 6.917 kJ mol�1 is needed tocomplete the conversion at room temperature.[19, 20] It shouldbe mentioned that calcium zincate can be produced by contin-uous stirring and aging of the mixture of Ca(OH)2, water, andZnO at a high temperature[30] or by ball milling.[31, 32]

In the second experiment, both Ca(OH)2 and ZnO werefound to dissolve sparingly into the KOH solution upon stirringfor 6 h, suggesting possible calcium zincate formation. TheXRD examination (Figure 2 c) confirmed a partial formation ofcalcium zincate. AAS measurements showed that the zinc levelin the solution was 86 g L�1, suggesting that almost all ZnOwas consumed to form calcium zincate, and the coexistence of

Figure 2. XRD patterns of a) Ca(OH)2, b) ZnO, c) precipitate of the electrolytefrom the second experiment (described in Section 2.1), d, e) precipitates ofthe electrolyte with 91 and 200 Ah discharging, respectively. The back-ground peaks on the 2q axis represent the data from the ICDD fileNo. 24-222.

ChemElectroChem 2015, 2, 134 – 142 www.chemelectrochem.org � 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim137

Articles

water and excess OH� is the key factor in provoking the reac-tion between ZnO and Ca(OH)2 and the dissolution of theformer. OH� may act as a catalyst to promote the hydrolysis ofZnO. The dissolution and dissociation of ZnO in aqueous KOHfollow the following route depicted in Reactions (7)–(10):[13, 33]

ZnOþ Hþ þ OH�,Zn2þ þ O2�Hþ þ ðOHÞ�,ZnðOHÞ2 ð7Þ

ZnðOHÞ2 þ 2 Hþ,Zn2þ þ 2 H2O ð8Þ

ZnðOHÞ3� þ 3 Hþ,Zn2þ þ 3 H2O ð9Þ

ZnðOHÞ42� þ 4 Hþ,Zn2þ þ 4 H2O ð10Þ

It is expected that, in the presence of abundant OH� andfree water, ZnO will dissolve quickly to produce zincate-richcompounds, such as K2Zn(OH)4, which may couple withCa(OH)2 to form calcium zincate. We also found that Ca(OH)2

cannot be completely consumed when the KOH concentrationis higher than 37 %, as XRD patterns always showed both ZnOand Ca(OH)2 peaks in the precipitates. This is consistent withthe finding of Sharma[19, 20] who concluded that water activityin aqueous KOH with concentrations higher than 34 wt % issmaller than that in equilibrium with calcium zincate, Ca(OH)2,and ZnO. Thus, Ca(OH)2 cannot suppress the presence of ZnOin aqueous KOH anymore.

2.2. Empirical Calculation Approach for Calcium ZincateFormation

A linear relationship was found between the zinc content inthe electrolyte and the cell discharge capacity based on repro-ducible titrations, AAS, and electrochemical experiments(Figure 4). To affirm the reliability of these results, mathemati-cal simulations were performed by using various types of soft-ware: MATLAB, Mathematica, and Maple 16. The computationalresults strongly confirmed the abovementioned linearity. The

energy equivalent capacity of the cell was determined fromthe amount of reactive zinc in a volume unit (kg L�1) and thetotal quantity of electricity obtained from the zinc involved inelectrochemical reactions. The simple linear relationship(Figure 4) suggests that the conversion of dissolved zinc intozincate in KOH is straightforward, and no other second-orderreactions or side reactions occur in parallel. In addition, theslope of the plot can be used to judge the quality of the celldesign, that is, the flatter the slope (the energy conversion effi-ciency is high), the better the cell design. Based on theCa(OH)2-to-Zn ratio in Reaction (6), Sharma’s reports on the ki-netics of calcium zincate formation,[19, 20] the report on the de-composition kinetics of supersaturated zincate solutions in ref-erence,[16] the linear relationship between the zinc content inelectrolyte and discharge capacity (Figure 4) and the abovementioned XRD results, an empirical calculation approachquantifying the formation of calcium zincate can beestablished.

The accurate ratios of ionized Zn to Ca(OH)2, as well as theamount of Zn converted to zincate, can be calculated directly,according to the molecular weight ratios in Reactions (4) and(6). To estimate the water loss resulting from the conversion ofZn to zincates, assuming that all Zn dissolved in KOH electro-lyte is converted into zincate form prior to the initiation ofZnO precipitation, Equations (I) and (II) were set:

CB ¼ C1DC þ C2 ðIÞ

WL1 ¼ CB � mZincate=Zn � aH2O � f c � V E ðIIÞ

Equation (I) is established based on the linear relation be-tween the Zn content in electrolyte and the cell discharge ca-pacity, which is found in the AAS measurements. CB is the dis-solved Zn concentration in the electrolyte in grams per liter(g L�1) and C1 is the dissolution rate of metallic to ionic Zn inthe electrochemical cell. It presents the conversion rate ofchemical energy into electrochemical energy and can be repre-sented by the slope of the plot of Zn content in the electrolyteversus the electrolyte discharge capacity of the cell. Its unit is

Figure 3. XRD pattern of precipitate from the first experiment (described inSection 2.1).

Figure 4. Linear relationship between zinc concentration in the electrolyteand the discharge capacity of the cell. The results are fitted linearly as:y = 1.2099x + 7.8915 (R2 = 0.9953).

ChemElectroChem 2015, 2, 134 – 142 www.chemelectrochem.org � 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim138

Articles

gram per Ampere hour (g Ah�1), and the value is 1.21 in thiswork (see Figure 4). DC represents the electrolyte discharge ca-pacity (in Ah L�1), and its value can be obtained directly fromthe electrochemical testing (i.e. total discharge capacity of thecell divided by the electrolyte volume). C2 is the concentrationof Zn dissolved in electrolyte, which did not contribute to thegeneration of energy through the cell circuit (i.e. zinc sponta-neously dissolved by corrosion reactions), and it can be repre-sented by the intercept of the plot presenting the Zn contentin the electrolyte versus the electrolyte capacity.

Equation (II) is derived based on Reaction (6) and is con-firmed by XRD, AAS results, and references.[13, 14] It is appliedfor the calculation of water loss, owing to metallic Zn conver-sion into ionic form presented as zincates. WL1 is the water lossin grams (g), owing to metallic Zn conversion into zincates. CB

is the dissolved Zn content in the electrolyte, and has thesame meaning as mentioned above. mzincate/zn, is the metalliczinc-to-zincate conversion factor (molecular weight ratio of zin-cate to zinc), which is 1.52 in this work; aH2O represents thepercentage weight of water in zincate, 0.233 (see Refs. [13, 14]).Regarding fc, it is the correction factor to account for waterloss during the cell operation time, owing to evaporation, fil-tration, and so forth (equal to 1.3 in this work), and was ob-tained from the measurements. VE is the volume of the electro-lyte in liters (L).

Equation (III) is utilized to estimate the water loss afterCa(OH)2 treatment:

WL2 ¼ WCa � eCaZincate=CaðOHÞ2 � bH2O � f c ðIIIÞ

where WL2 (in g) represents the water loss, owing to calciumzincate formation, WCa is the amount of Ca(OH)2 added intoelectrolyte. eCaZincate/Ca(OH)2

is theCa(OH)2-to-calcium zincate con-version factor, which can be cal-culated from the molecularweight ratio of calcium zincateto Ca(OH)2; the value is 4.17.bH2O is the water weight percent-age in calcium zincate, valued as0.291 based on Reaction (6), andfc has the same aforementionedmeaning.

2.3. Effect of Zincate Scaveng-ing by Calcium Hydroxide

The effect of the zincate scavenging by adding Ca(OH)2 to theelectrolyte was first estimated theoretically by using the de-rived formula in Equation (IV):

RT ¼ WCa �F=x ðIVÞ

where RT (in g) is the theoretically estimated amount of zinc re-moved from the electrolyte after Ca(OH)2 treatment; F repre-

sents the purity of Ca(OH)2, which is 95 % in this work. x is themolecular weight ratio of Zn to Ca(OH)2, 0.57; note that it wascalculated as 2 Zn/Ca(OH)2.

The actual scavenging data was then calculated with Equa-tion (V) by using the experimentally measured AAS data:

RA ¼ ðCB�CEÞ � VE ðVÞ

where RA (in g) is the actual amount of reduced zinc from theelectrolyte and CE (in g L�1) is the concentration of zinc in elec-trolyte after Ca(OH)2 treatment. The meanings of VE and CB

have already been mentioned above. On the basis of the de-rived formulas [Equationa (I)–(V)] , the two most important for-mulas used to calculate the amount of Ca(OH)2 reacted withzincate were developed as Equations (VI) and (VII):

WCa ¼ R � x ðVIÞ

WCa ¼ ðCB�CEÞ � VE � x ðVIIÞ

where R (in g) is the amount of dissolved Zn removed fromthe electrolyte. The other symbols are same as explainedabove.

Although further experiments pertinent to the quantificationof water compensation in the course of calcium zincate forma-tion are needed, the cell tests performed in accordance withthe abovementioned empirical calculation theory showedpromise, as the electrolyte capacity appeared to increase asthe cell service life expanded. Table 1 shows that both the the-oretically predicted and experimentally measured values forscavenged zinc dissolved in electrolyte are in good agreement,

suggesting the validity of the applied theoretical approachand that scavenging zincate by means of calcium hydroxide isa selective process that results in calcium zincate. The forma-tion of calcium zincate through the hydrolysis of zinc oxide,which is then followed by a direct reaction with calcium hy-droxide, is proven to be very slow and can, therefore, be ne-glected. Upon adding Ca(OH)2, the zincate anion concentrationin KOH electrolyte can be reduced by approximately 40 %(Figure 5 and Figure 6). Correspondingly, both the dischargecapacity and the electrolyte loading can be increased by morethan 40 % (Figure 7), leading to a significantly improved cellperformance (Figure 8) without the necessity to change the

Table 1. Theoretically evaluated and experimentally obtained values for zinc reduction in KOH electrolyte withand without adding Ca(OH)2.

Reduction in Zn based on theoretical approach Reduction in Zn based on AASTerm Ca(OH)2/

Zn ratioUsed Ca(OH)2

weightCa(OH)2

purityElectrolytevolume [L]

Zn beforetreatment [g]

Zn aftertreatment [g]

Input data 0.57 75 0.95 1.25 256.6 158.2

Theoretically estimated Znreduction [g]

125 Actual reduction in Znamount [g]

123

ChemElectroChem 2015, 2, 134 – 142 www.chemelectrochem.org � 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim139

Articles

electrolyte. The small reduction (2.1 %) in zinc content upon fil-tration observed in Figure 6 suggests the formation of someinsoluble zincate species, most likely zinc oxide, with particlesizes bigger than 0.45 mm after a long period of nonstop run-ning. The formation of insoluble zinc oxide can be mitigatedor even eliminated through the aging and stirring of a usedelectrolyte mixed with calcium hydroxide, as was done in thesecond experiment described above.

2.4. Importance of Free Water for Ca(OH)2 Treatment andCell Performance

In a KOH solution, most water molecules typically interact withOH� anions. In this context, free water can be defined as watermolecules that are mobile enough to link to other chemicalgroups or species rather than hydroxide anions and whose ac-tivity is not significantly reduced compared to that in purewater. It was found that, if no small amount of water (less than1/200 volume fraction) was added to a continuously runningcell with 45 % KOH as the electrolyte, the service life of the cell

was shorter. In turn, the same cells running with added freewater demonstrate longer service times of over 250 h, and nonoticeable amount of precipitated ZnO is found by means offiltration and EDS, as well as by AAS measurements for zincand oxygen contents in the precipitates. This clearly provesthat the presence of free water in the electrolyte plays an im-portant role in the conversions of zinc into zincate and zincateto ZnO. Such water is believed to be necessary for the releaseof OH� ions and then to increase their mobility in KOH inorder to fuel and facilitate anodic Reactions (1) and (4). In par-allel, in the presence of free water, an equilibrium chain of re-actions is known to be built up in saturated and supersaturat-ed KOH solutions, that is, [Zn(OH)4]2�$[ZnO(OH)2]2�$ZnO.[16, 34]

Therefore, the presence of such free water will hold back theprecipitation of ZnO, increasing the [Zn(OH)4]2� concentrationin the solution and facilitating the scavenging. It should bekept in mind that KOH crystallizes when its concentrationreaches values above 48 %, that is, in the entire absence of

Figure 5. Reduction in Zn concentration in electrolyte upon adding Ca(OH)2

to scavenge zincate.

Figure 6. Zinc content change (in %) in the electrolyte upon filtration andtreatment with Ca(OH)2. Green columns represent the percentage of zinccontent in the tank, whereas violet columns represent its reduction (in %)after treatment.

Figure 7. Electrolyte loading and capacity gain upon adding Ca(OH)2 to scav-enge zincate. Green columns represent the electrolyte loading in Wh L�1,whereas blue represents the discharge capacity in Ah L�1. The numbersmarked on the columns present either the maximum loading or the dis-charge capacity that the electrolyte has reached.

Figure 8. Cell discharge polarization curves (potential versus current, solidcircles) and corresponding output power of the electrolyte (dashed lines)before (black) and after (red) treatment with Ca(OH)2.

ChemElectroChem 2015, 2, 134 – 142 www.chemelectrochem.org � 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim140

Articles

free water. The use of highly concentrated KOH to gain poweroutput is expected to result in increased cell resistance. Thiswill decline the cell efficiency and yield additional heat thatmay potentially damage the polymer materials in the separatorand gas diffusion layer, thus causing a running failure of theZAFC. Meanwhile, using Ca(OH)2 to scavenge the zincate alsorequires water [see Reaction (6)] . Therefore, water loss andcompensation should be taken into account in all of the abovereactions. Sections 2.1 to 2.3 present a theoretical way to cal-culate the amount of water needed to maintain the equilibriaof the reactions. Water that has to be added into the electro-lyte to compensate for its loss during conversion processesconsists of two components: 1) water consumed for convert-ing Ca(OH)2 into calcium zincate and 2) water used in the con-version reactions of zinc to zincate and zincate to ZnO. Equa-tions (II) and (III), shown above, are proposed to evaluate thetotal water loss during these two processes. Figure 9 shows

how adding water affects the removal efficiency of zincate inan electrolyte with 37 % of KOH, which was discharged at187 Ah L�1. It is clearly seen that, without adding water, the de-viation from the theoretical line (representing removal of Zn inan ideal system) appears to begin at about 30 % Zn present inthe electrolyte. This suggests that increasingly more zincatespecies dissolved in KOH cannot be converted into calcium zin-cate as more Ca(OH)2 is added to the solution. The situation isdrastically improved when water is added into the system, asthe deviation is only observed when more than approximately60 % Zn is found in the electrolyte. However, more work isneeded to better understand when and how water should beadded, as Figure 9 also shows that not all zincate ions wereconverted into calcium zincate after adding water into the

electrolyte. This is consistent with the XRD results in Figure 2 d,where the presence of minor quantities of Ca(OH)2, Zn(OH)2,and ZnO in calcium zincate is seen.

One of the fundamental assumptions used in establishingthe above theoretical calculation approach is that the metalliczinc dissolved into highly concentrated KOH (>34 %) can becompletely converted into zincate ion form (i.e. no significanttrace amount of ZnO is produced in the course of conversion),detained in the saturated and/or supersaturated zincate solu-tions, and then the zincate ions can fully react with calcium hy-droxide to form calcium zincate. The appearance of deviationbetween the simulated and experimental results in Figure 9suggests that, as the cell undergoes discharge, the dissolvedzinc may not be completely converted into the zincate form,or the zincate may not be entirely scavenged by calcium hy-droxide. This apparent deviation occurs when the electrolytecapacity reaches values over 200 Ah L�1 or the cell dischargecurrent density is over 200 mA cm2. This suggests that moreZnO is present in the electrolyte, which was also revealed byXRD measurements. As described in Section 2.1, the dissolvedzinc passes through several stages in the KOH solution untilgetting converted into zincate ions, and then forms variouscomplexes and oxides.[16, 34] ZnO can be scavenged by calciumhydroxide easily when it is first converted to zincate in thepresence of excess OH� bound with mobile water. Light scat-tering spectra showed that no evidence of colloid formationwas found in saturated and/or supersaturated zincate solu-tions. For the zincate electrolytes with low discharge capacity,the peaks presenting the particle size of Zn-containing specieswere centralized at one or two small values, whereas theygradually became scattered into a larger value range for theelectrolytes with discharge capacities over 120 Ah L�1, suggest-ing that zincate species dominate throughout the entireanodic Zn dissolution, and complex Zn-containing particlescan be formed through two stages (homogeneous and hetero-geneous ones), which is fully consistent with the findings ofDirkse[35] as well as Debiemme-Chouvy and co-workers.[16, 34]

The results presented respectively by the theoretical curve andthe experimental points in Figure 9 are fairly consistent witheach other for zinc contents in the electrolyte below 30 %, cor-responding to the state in which most of zinc dissolved inKOH electrolyte should exist, that is, mainly in the form ofZn(OH)4

2�. Note that the presence of zincate ions and ZnO wastraced, simultaneously, by means of titration, AAS, and XRDmeasurements. This suggests that the best scavenging effectcould be obtained, as all the dissolved zinc was present in theform of Zn(OH)4

2�. The deviation only started to skew severelywhen zinc-oxide precipitation began. We interpret this findingas a result of a faster formation of ZnO when it reaches itssecond stage, that is, heterogeneous formation. At this point,its formation rate accelerates, whereas its scavenging is muchslower than that for the zincate. Dissolving ZnO by means ofwater and OH� is known to be very slow. Therefore, the pa-rameters set up in Equations (I)–(VII) may need to be fine-tuned in order to be more accurate and reliable, which shouldbe done by considering more detailed information on eachstage of ZnO formation. The work can be engaged by initiating

Figure 9. Effect of adding water on zinc removal from electrolyte by meansof Ca(OH)2. Solid line indicates theoretical values of Zn removal (in %),whereas solid circles and empty triangles stand for actual measured values,without and with added water, respectively.

ChemElectroChem 2015, 2, 134 – 142 www.chemelectrochem.org � 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim141

Articles

kinetic and dynamic studies of the zincate solution, as well assetting up an XPS/Auger spectroscopy database for differentzincate/ZnO species produced at different cell-dischargestates; although, it seems rather challenging, because theavailable literature reports are quite scarce to date.

4. Conclusions

The utilization of calcium hydroxide to scavenge the zincatespecies forming in a Zn–air cell has been proven to be a highlyeffective approach to improve the cell performance and in-crease the electrolyte capacity (40 %). A fundamental calcula-tion approach to quantify the Ca zincate formation and watercompensation has been established and tested in this study.By using Ca(OH)2 to scavenge Zn-containing species in theconcentrated KOH solution, it could mostly facilitate the re-moval of zincate, rather than ZnO, from the electrolyte. How-ever, the presence of excess free mobile H2O and OH� helpsdissolve ZnO and form zincate. Replenishing the water wasdemonstrated to increase the percentage of zincate scavengedby means of Ca(OH)2.

The results obtained in this study demonstrate that the pro-posed approach can widely be applied to all zinc–air systemsduring their discharge cycle. However, a concern still remainsas for how calcium hydroxide will affect the charging cycle,which is important for the development of next-generation(electrochemically rechargeable) zinc–air batteries and fuelcells. In this context, determining the best method for recover-ing zinc from calcium zincate becomes an exciting challenge.Literature reports indicate that calcium zincate is unstable inconcentrated KOH solutions above 34 %. If a calcium hydroxide“filter” were removed from the battery cell and placed in anelectroplating cell with KOH concentration above 34 %, calciumzincate would start spontaneously releasing some zincate ions.The released zincate ions could be electroplated as zinc ona cathodic substrate, thereby stimulating more zincate dissolu-tion from calcium zincate, which is seen as a high-potentialin situ approach for zinc regeneration to be explored and de-veloped in the future. Moreover, as calcium zincate was recent-ly found to provide a beneficial effect on the improved per-formance of Ni–Zn batteries, searching for new preparationroutes and applications for calcium zincate is also worthpursuing.

Acknowledgement

The authors sincerely acknowledge Mrs. Helen Zhu and Dr. Jing-wei Hu from AFCC (Mercedes Benz-Ford fuel cell research center,Canada) for their full support during the completion of this re-search project. S.A.K. acknowledges the Marie Curie InternationalIncoming Fellowship from the European Commission.

Keywords: calcium hydroxide · electrolyte discharge capacity ·fuel cells · scavenging · zincates

[1] J. M. Tarascon, Phil. Trans. R. Soc. A 2010, 368, 3227 – 3241.[2] J. P. Meyers, Getting Back Into Gear : Fuel Cell Development after the

Hype, Interface, 2008, Winter, 36 – 39.[3] B. Peng, J. Chen, Coord. Chem. Rev. 2009, 253, 2805 – 2813.[4] F. Cheng, J. Chen, Chem. Soc. Rev. 2012, 41, 2172 – 2192.[5] J. Christensen, P. Albertus, R. S. Sanchez-Carrera, T. Lohmann, B. Kozin-

sky, R. Liedtke, J. Ahmed, A. Kojic, J. Electrochem. Soc. 2012, 159, R1 –R30.

[6] Z. Zhang, C. C. Zuo, Z. H. Liu, Y. Yu, Y. X. Zuo, Y. Song, J. Power Sources2014, 251, 470 – 475.

[7] C. Z. Shu, E. D. Wang, L. H. Jiang, G. Q. Sun, Int. J. Hydrogen Energy2013, 38, 5885 – 5893.

[8] J. S. Lee, S. T. Kim, R. G. Cao, N. S. Choi, M. L. Liu, K. T. Lee, J. P. Cho, Adv.Energy Mater. 2011, 1, 34 – 50.

[9] B. Richter, D. Goldston, G. Crabtree, L. Glicksman, D. Goldstein, D.Greene, D. Kammen, M. Levine, M. Lubell, M. Savitz, D. Sperling, EnergyFuture : Think Efficiency, American Physical Society, College Park, 2008.

[10] D. J. MacKay, Sustainable Energy - Without the Hot Air, UIT Cambridge,Ltd, Cambridge, 2009.

[11] K. Kinoshita, Electrochemical Oxygen Technology, John Wiley & Sons, Inc. ,New York, 1992, pp. 19 – 112.

[12] D. Linden, T. B. Reddy, Handbook of Batteries, 3rd ed. , McGraw-Hill, NewYork, 2002.

[13] X. G. Zhang, Corrosion and Electrochemistry of Zinc, Plenum Publisher,New York, London, 1996.

[14] S. I. Smedley, X. G. Zhang, J. Power Sources 2007, 165, 897 – 904.[15] O. Haas, F. Holzer, K. M�ller, S. M�ller in Handbook of Fuel Cells, Vol. I

(Eds. : W. Vielstich, H. A. Gasteiger, A. Lamm), John Wiley & Sons, NewYork, 2003, pp. 382 – 408.

[16] C. Debiemme-Chouvy, J. Vedel, J. Electrochem. Soc. 1991, 138, 2538 –2542.

[17] F. R. McLarnon, E. J. Cairns, J. Electrochem. Soc. 1991, 138, 645 – 664.[18] G. A. Roberts, I. Rehmanji, US 2010/0196768 A1.[19] R. Sharma, J. Electrochem. Soc. 1986, 133, 2215 – 2219.[20] R. Sharma, J. Electrochem. Soc. 1988, 135, 1875 – 1882.[21] R. J. Wang, Z. H. Yang, B. Yang, X. M. Fan, T. T. Wang, J. Power Sources

2014, 246, 313 – 321.[22] J. M. Rubio-Caballero, J. Santamaria-Gonzalez, J. Merida-Robles, R.

Moreno-Tost, M. L. Alonso-Castillo, E. Vereda-Alonso, A. Jimenez-Lopez,P. Maireles-Torres, Fuel 2013, 105, 518 – 522.

[23] PerkinElmer Corporation, PerkinElmer Atomic Absorption Spectropho-tometer: System Description and Maintenance, Norwalk, CT, 1982.

[24] Elmer Corporation, Analytical Methods for Atomic Absorption Spectro-photometry, Norwalk, CT, 1982.

[25] J. G. Webster, The Measurement, Instrumentation and Sensors Handbook,CRC Press, Boca Raton, 1998.

[26] Z. Kowalski, W. Kubiak, A. Kowalska, Anal. Chim. Acta 1982, 140, 115 –121.

[27] The Merck Index : An Encyclopedia of Chemicals, Drugs, and Biologicals,14th ed. (Ed. M. J. O’Neil), Merck Inc. , New Jersey, 2006.

[28] F. Ziegler, C. A. Johnson, Cem. Concr. Res. 2001, 31, 1327 – 1332.[29] W. M. Haynes, CRC Handbook of Chemistry and Physics, 93rd ed. , CRC

Press, Taylor & Francis Group, Boca Raton, 2012 – 2013.[30] S. W. Wang, Z. H. Yang, L. H. Zeng, Mater. Chem. Phys. 2008, 112, 603 –

606.[31] X. M. Zhu, H. X. Yang, X. P. Ai, J. X. Yu, Y. L. Cao, J. Appl. Electrochem.

2003, 33, 607 – 612.[32] C. C. Yang, W. C. Chien, P. W. Chen, C. Y. Wu, J. Appl. Electrochem. 2009,

39, 39 – 44.[33] C. F. Baes, R. E. Mesmer, The Hydrolysis of Cations, CRC Handbook, 89th

ed. , John Wiley & Sons, New York, 1974.[34] C. Debiemme-Chouvy, M.-C. Bellissent-Funel, R. Cortes, J. Vedel, J. Elec-

trochem. Soc. 1995, 142, 1359 – 1364.[35] T. P. Dirkse, J. Electrochem. Soc. 1981, 128, 1412 – 1415.

Received: July 23, 2014Revised: September 14, 2014Published online on November 4, 2014

ChemElectroChem 2015, 2, 134 – 142 www.chemelectrochem.org � 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim142

Articles