A Near-Neutral Chloride Electrolyte for Electrically Rechargeable Zinc-Air Batteries

Near-neutral electrolytes based on zinc chloride and ammonium chloride are examined for rechargeable zinc-air battery application. TheeffectsofpHvalue,saltconcentration,andpolyethyleneglycolandthioureaadditivesareinvestigatedandachlorideelectrolyteisdeveloped.Thereversiblezincdepositionandzincstrippingprocessesarestudiedbycyclicvoltammetrywithrotating-discelectrodetechnique.Thezincanodeandaircathodebehaviorsinnear-neutralchlorideelectrolytearecharacterizedbyquasisteady-statepolarizationandimpedancespectroscopy.Prototypedzinc-airbatterywithnear-neutralchlorideelectrolytescansustainmorethan1000hoursandhundredsofdischarge-chargecycleswithminimizedzincdendriteformationandnocarbonateformationproblem,underdischarge-chargecapacityrangingfrom20to120mAh.Thenear-neutralchlorideelectrolyteprovidesasaferandmorerobustalternativetotraditionalalkalineelectrolyteforrechargeablezinc-airbatteries.

Zinc-air batteries (ZABs) are efficient electrochemical energy storage devices with the advantages of high specific energy, safety and low cost. Primary ZABs batteries are commercially available to power hearing aids, railroad track circuits, warning lights and remote signals. 1 These types of devices typically discharge at low rate over a long period of time. For broad applications and environmentally benignity, it is of great interest to upgrade the primary ZABs into electrically rechargeable ones. However, secondary ZABs batteries have not been commercialized due to some technical challenges.
Electrolyte carbonation is one of the technical hurdles of electrically rechargeable ZABs. The traditional alkaline electrolyte reacts with CO 2 and results in the reduction of ionic conductivity. The precipitated carbonate particulates block the diffusion channels of air electrode. [2][3][4] This problem can be avoided by employing non-alkaline electrolytes, where no reactions with CO 2 occur. 5,6 Another technical challenge comes from the zinc dendrite formation. During battery charging, zinc tends to grow on protruded surfaces and results in zinc dendrites after prolonged charging/discharging cycles. 7 Zinc dendrites could penetrate through the separator and eventually lead to short circuit when it reaches the air cathode. Generally, two mechanisms of additive are commonly known: (i) ligating with metal species and forming complexes; or (ii) adsorbing on the electrode surface and inhibiting metal nucleation. 8 In this study, a near-neutral electrolyte was developed for electrically rechargeable zinc-air battery application. The effects of salt concentration, pH value and additives on zinc electrodeposition and stripping were studied in ZnCl 2 -based electrolyte baths by cyclic voltammetry and ZAB prototype tests. A suitable chloride-based nearneutral electrolyte suppressing zinc dendrite formation was developed for electrically rechargeable ZABs. The air cathode catalyst was commercial manganese oxide (MnO 2 ) with conductive carbon materials due to its good oxygen reduction reaction (ORR) activity and low cost. 9,10 The zinc anode and air cathode was characterized by quasi steady-state polarization curves and impedance spectroscopy. The near-neutral electrolyte was demonstrated to be a more technically viable alternative to traditional alkaline counterpart.
Cyclic voltammetry (CV), linear sweep voltammetry (LSV) and electrochemical impedance spectroscopy (EIS) were conducted through a potentiostat/galvanostat (Autolab PGSTAT302N) in threeelectrode configuration. The working electrode for CV was a platinum rotating disk electrode (3 mm in diameter) at a rotating rate of 400 rpm. The working electrodes for LSV and EIS were either zinc plate or air cathode, with working area of 0.5 cm 2 . The counter electrode and reference electrode were platinum foil and Ag/AgCl (3 M KCl), respectively. The bare chloride bath consisted of 0.51 M ZnCl 2 and 2.34 M NH 4 Cl. The potential sweeping rates for CV and LSV was 20 mV s −1 and 5 mV s −1 , respectively.
The catalyst slurry consisted of 66 wt% MnO 2 , 22 wt% Vulcan XC-72 and 12 wt% Nafion. The slurry was roll-pressed on a SGL carbon paper and dried at 70 • C for 1 h to achieve a loading of 4.5 mg cm −2 . The working area of the air cathode and polished zinc sheet anode was 2 cm 2 . A zinc air battery prototype was assembled by an acrylic cell fixture with the cell volume of 30 mL and the distance between anode and cathode of 30 mm. The electrolyte was ZnCl 2 -NH 4 Cl bath with or without additives and the pH value was adjusted by NH 4 OH. Battery testing and cycling experiments (Maccor 4300) were performed with the recurrent galvanic pulse method at 25 • C and under ambient air condition. Surface morphology of zinc anodes was observed by field emission scanning electron microscopy (SEM, JEOL JSM-6700F). Crystalline structure of zinc anodes and air cathodes were characterized by X-ray diffraction (XRD, Bruker D8 GADDS).

Results and Discussion
The hydrogen evolution reaction occurs near −1.15 V vs. Ag/AgCl in a 2.34 M NH 4 Cl solution ( Figure 1). The Zn reduction occurs at   12 which is thermodynamically stable in ZnCl 2 -NH 4 Cl solutions. 13 As given in Figure 1, the peak position of the zinc stripping loop is significantly shifted to −0.82 V (cf. −0.66 V for bare ZnCl 2 bath), indicating that the zinc deposition and stripping processes are more reversible than those in the bare ZnCl 2 bath.
With increasing pH values, both the reduction potential of Zn 2+ and peak positions of the zinc stripping loops shift to the negative direction, and the full width at half maximum (FWHM) of the zinc stripping loops are narrowed ( Figure 2). These observations imply more sluggish reduction of zinc-ammonia complex but more facile zinc stripping process with increasing pH values. Tetrahedral [Zn(NH 3 ) 4 ] 2+ exists in equilibrium with octahedral [Zn(NH 3 ) 6 ] 2+ in the ZnCl 2 -NH 4 Cl bath. 12 A high pH value is postulated to favor the forward reaction of Eq. 1: Reduction of [Zn(NH 3 ) 6 ] 2+ requires higher activation energy than [Zn(NH 3 ) 4 ] 2+ , thus explaining the cathodic shift of zinc reduction  potential with the increasing pH value. The pH values of all ZnCl 2 -NH 4 Cl baths were adjusted to be 6.0, unless explicitly stated. The CVs show that zinc deposition and stripping processes are slightly affected by the addition of 1000 ppm PEG, as reflected from a zinc reduction potential shift from −1.041 V to −1.043 V and a mild depression of the zinc stripping loop ( Figure 3). In contrast to PEG, the incorporation of 1000 ppm thiourea exhibits pronounced inhibition effect, where zinc deposition initiates at −1.071 V. Simultaneous addition of PEG and thiourea further shifts the Zn reduction potential to −1.092 V. Interestingly, the peak current of Zn stripping is reduced from 257 mA cm −2 for the bare chloride bath to 85 mA cm −2 in the same bath containing PEG and thiourea. These results demonstrate that PEG and thiourea, when used concurrently, are effective additives in inhibiting zinc deposition and stripping processes. Such an inhabitation effect would be beneficial for mitigating zinc dendrite formation in ZABs. Figure 4 shows the impedance spectra of zinc anode and air cathode. Polarization resistance (R p ) is defined by the difference between the low-frequency intercept and high-frequency intercept of an impedance spectrum. The impedance spectrum of zinc anode is featured by a depressed impedance arc and a high-frequency inductive loop. Compared to R p of 4.9 cm 2 under OCP condition which was −1.00 V, R p of zinc anode is decreased by 41% and 92% when polarized at 100 mV and −100 mV potential biases, respectively, implying that zinc deposition is a more sluggish process than zinc dissolution. R p of air cathode are about one order of magnitude larger than those of zinc anode, cf. Figures 4d-4f to Figures 4a-4c. To better understand the reaction mechanisms of air cathode, all air-cathode impedance spectra were fitted with relevant equivalent circuits, as given in the insets of Figures 4d-4f. The cathodic polarized impedance spectrum is modeled by a resistor R s , which models the lumped resistance, in series with a Voigt-type R p /Q sub-circuit where R p is the polarization resistance and Q is the constant phase element (CPE) describing the non-ideal capacitive feature of the electrochemical cell. The OCP and anodic polarized impedance spectra have a high-frequency depressed arc and a low-frequency impedance tail which is described by a CPE connected in series. The open impedance tail implies the infinite diffusion characteristic of the air cathode, possibly due to the sluggish oxygen transport in the chloride electrolyte. R p of air cathode polarized at −0.3 V (cathodic bias of 550 mV) is 78.3 cm 2 , comparable with 69.0 cm 2 that is under an anodic bias of 350 mV. The low-frequency diffusional impedance of air cathode under anodic bias implies that the oxygen evolution reaction is impeded by some reaction steps with large time constants, such as adsorption/dissociation of oxygen molecule and surface diffusion of oxygen ad-atoms. Figure 5 shows the i-η responses and the Tafel plots of air cathode and zinc anode, respectively. Current responses at specific potentials    Table II). i 0,c of air cathode is slightly larger than the i 0,a . This polarization characteristic of air cathode indicates, while at low η region ORR is more facile than OER, OER has better reaction kinetics at high η regions than ORR. Zinc dissolution is thermodynamically and kinetically a more facile process than zinc deposition. The half-cell reactions of zinc anode and air cathode of ZAB in the near-neutral chloride electrolytes would be described by: Discharging:  Figure 6a reveals the salt concentration effects on ZAB performance without pH adjustment. The 0.51 M ZnCl 2 -2.34 M NH 4 Cl bath provides the most efficient discharge-charge cycling behavior. Figure 6b shows the influence of PEG and/or thiourea on ZAB performance, where the pH value of 0.51 M ZnCl 2 -2.34 M NH 4 Cl baths is adjusted to 6.0. Under a rate of 10 mA discharging and 5 mA charging, the discharge-charge potential gap is 1.4 V and the discharge-charge efficiency is 36%. The ZABs can sustain 21, 15, 22 and 26 dischargecharge cycles when running at bare chloride bath, chloride bath with thiourea, chloride bath with PEG, and chloride bath with both thiourea and PEG, respectively. The degradation rate of the discharge-charge potential gap of ZAB with the ZnCl 2 -NH 4 Cl-PEG-thiourea electrolyte is 7%. Figure 7 shows the ZAB performances at various discharge-charge rates and capacities. While the ZnCl 2 -NH 4 Cl-PEG-thiourea electrolyte is only capable for 26 discharge-charge cycles under 10 mA/40 mAh, it can sustain 91 cycles (1092 h equivalent), 100 cycles (1200 h equivalent) and 120 cycles (1440 h equivalent) under the 5 mA/20 mAh, 2 mA/8 mAh, and 1 mA/4 mAh, respectively. Under a rate of 1 mA and a capacity of 4 mAh, the discharge-charge potential gap is 1.1 V, the discharge-charge efficiency is 45% and the degradation rate of discharge-charge potential gap is negligible up to 1200 h (Figure 7d).
The longevity of ZAB is, however, paralyzed when it is subjected to relatively large discharge-charge cycling capacity. At a rate of 5 mA, while the ZAB performs stably for 42 discharge-charge cycles (504 h equivalent) under 20 mAh, it deteriorates significantly after 432 h under 60 mAh and dies quickly after 360 h under 120 mAh (Figure 8). The amount of zinc species available in the electrolyte is  15.3 mmol. As an example of discharge-charge capacity at 120 mAh, 2.2 mmol of Zn is dissolved/deposited on each discharging-charging cycle. This corresponds to the depth of discharge (DOD) of 14%. Figure 9 further demostrates that the high-rate capability remains a technical challenge for ZAB with chloride electrolyte. Better bifunctional catalysts in near-neutral aqueous media than MnO 2 -VC would increase the rate capability of ZABs. Figure 10 shows the XRD patterns of zinc anode after 6 dischargecharge cycles and a 7 th discharge process in the ZnCl 2 -NH 4 Cl electrolyte. The phases are Zn (JCPDS 04-0831), ZnCl 2 (JCPDS 16-0850) and Zn(NH 3 ) 2 Cl 2 (JCPDS 24-1435). Figure 11 shows the zinc anodes after 6 discharge-charge cycles and a 7 th discharge, where any soluble salts were removed by soaking the dismantled zinc anodes in large quantity of water. No reaction products were detected in the ZnCl 2 -NH 4 Cl electrolyte except the Zn. ZnO was detected in ZnCl 2 -NH 4 Cl-PEG electrolyte. Trace of ZnO appears on zinc anodes in ZnCl 2 -NH 4 Cl-thiourea and ZnCl 2 -NH 4 Cl-PEG-thiourea-PEG electrolytes. Based on the XRD results, it is proposed that the immediate ZAB discharging product is zinc-ammonia complex (Eq. 3) while the final reaction product is ZnO (Eqs. 4 and 5). Figure 12 shows the surface morphology of zinc deposits after 7 discharge-charge cycles where the battery test was stopped after the 7 th charging stage. The zinc deposits from the ZnCl 2 -NH 4 Cl are porous and cloudy. The zinc deposits are relatively compact and granulated for ZnCl 2 -NH 4 Cl-PEG and consist of mixed granulates and cross-linked flakes for ZnCl 2 -NH 4 Cl-thiourea. It is interesting to note that compact and uniform ripple-like zinc deposits are formed in ZnCl 2 -NH 4 Cl-PEG-thiourea. Figure 13 shows the surface morphology of zinc anode after 4 discharge-charge cycles and a 5 th discharge battery test. The discharged zinc anode in the bare ZnCl 2 -NH 4 Cl bath consists of flakelike structures on the zinc anode surface. The addition of PEG to the chloride bath significantly flattens the zinc anode surface, whereby the inset of Figure 13b shows zinc deposits composing of compact  PEG was known as inhibitor against hydrogen embrittlement during metal electrodeposition. 15,16 Banik and Akolkar found the addition of PEG-200 (PEG, M.W. = 200) substantially lowered the i 0 but not the cathodic transfer coefficient (α c ) of zinc deposition from zinc halide solutions. 17 PEG-200 suppressed the activation-controlled zinc dendrite propagation by absorbing on the electrode surface and forming a passivation layer. 17 In this study, PEG functions as a surface smoothing agent and prevents deleterious zinc dendrite formation during prolonged operation of rechargeable ZABs. The addition of 4-8 wt% of thiourea in acidic zinc electroplating bath reduced the corrosion rate in sea water by 30%. 18 Thiourea would perform as an inhibition agent to increase the overpotential of zinc dissolution/deposition during the ZAB discharge-charge cycles, thus leading to fine structures of zinc deposits. In contrast, the zinc dendrite growth in alkaline electrolyte is so significant that it will penetrate to the air cathode and short the battery cell. Zinc deposits generated from alkaline zincate solutions were very rough and showed dendritic and heavy spongy morphologies. 19,20 With the elimination of carbonate formation and zinc dendrite problems that are inherent to traditional alkaline electrolyte, the near-neutral chloride electrolyte casts light on the commercialization and realization of electrically rechargeable zinc-air batteries.

Conclusions
A class of near-neutral chloride baths containing zinc chloride (ZnCl 2 ) and ammonium chloride (NH 4 Cl) in the presence of polyethy-lene glycol and thiourea additives was developed for rechargeable zinc-air battery application. A chloride bath consisting of 0.51 M ZnCl 2 , 2.34 M NH 4 Cl, 1000 ppm PEG, 1000 ppm thiourea, and with the pH value of 6.0 showed satisfactory electrochemical and zinc air battery performance. The incorporation of PEG and thiourea additives posed inhibitory effects on zinc deposition from the chloride baths. The reaction kinetics of air cathode was almost one order-ofmagnitude more sluggish than that of zinc anode, as reflected from the impedance spectroscopy and exchange current densities. For both zinc anode and air cathode in three-electrode configuration, the cathodic and anodic branches of the quasi steady-state polarization curves were asymmetric. Rechargeable zinc-air battery tests proved that this type of chloride electrolyte system can sustain more than 1000 h and hundreds of discharge-charge cycles, under discharge-charge capacity ranging from 20 to 120 mAh. No zinc dendrite formation was observed after prolonged zinc-air battery test.