Article Electrochemistry, 85(4), 179–185 (2017)

Effect of Concentrated Electrolyte on Aqueous Sodium-ion Batterywith Sodium Manganese Hexacyanoferrate CathodeKosuke NAKAMOTO,a Ryo SAKAMOTO,a Masato ITO,b

Ayuko KITAJOU,b and Shigeto OKADAb,*a Interdisciplinary Graduate School of Engineering Sciences, Kyushu University,6-1 Kasuga-koen, Kasuga, Fukuoka 816-8580, Japan

b Institute for Materials Chemistry and Engineering, Kyushu University,6-1 Kasuga-koen, Kasuga, Fukuoka 816-8580, Japan

*Corresponding author: s-okada@cm.kyushu-u.ac.jp

ABSTRACTFrom the viewpoint of the cost and safety, aqueous sodium-ion batteries are attractive candidate for large-scaleenergy storage. Although the operating voltage range of the aqueous battery is theoretically limited to 1.23V by theelectrochemical decomposition of water, the voltage restriction is a little bit eased in real aqueous battery system bythe charge/discharge overvoltage. Effect of the concentrated electrolyte on the operation voltage was studied inaqueous Na-ion battery with Na2MnFe(CN)6 hexacyanoferrates cathode and NaTi2(PO4)3 NASICON-type anode, inorder to increase the discharge voltage. According to the cyclic voltammetry, the electrochemical window of diluted1mol kg−1 NaClO4 aqueous electrolyte is only 1.9V, whereas the corresponding electrochemical window ofconcentrated 17mol kg−1 NaClO4 aqueous electrolyte is widen to 2.8V. This wide electrochemical window of theconcentrated aqueous electrolyte allows the Na2MnFe(CN)6//NaTi2(PO4)3 aqueous sodium-ion system to workreversibly. By contrast, the framework of Na2MnFe(CN)6 cathode was destroyed by the hydroxide anion generatedin diluted 1molkg−1 electrolyte.

© The Electrochemical Society of Japan, All rights reserved.

Keywords : Aqueous Sodium-ion Battery, Concentrated Electrolyte, Sodium Manganese Hexacyanoferrate, WaterDecomposition

1. Introduction

There is growing interest in large-scale, highly cost-effectiveenergy storage systems that utilize intermittent renewable energysources such as solar and wind power efficiently. To date, lithium-ionbatteries (LIBs) as well as sodium sulfur batteries have beencommercialized as large-scale energy storage systems around theworld. LIBs are advantageous for their high voltage and high energydensity, but they have serious drawbacks in safety, cost, andconductivity due to their reliance on non-aqueous electrolyte. Inaddition, the yearly production of lithium for battery grade is limitedby the surface area of salt lake, because high quality lithium is mainlyprovided by sun-drying of salt lakes. Current concern to Li shortageby the strong demand of LIBs for electric vehicles brings the priceincrease of lithium. Minor-metal-free sodium sulfur batteries areattractive for large-scale energy storage due to their low initial cost,whereas the problem is the high running cost necessitated by their300°C operating temperature. Therefore, sodium-ion batteries (SIBs)operating at room temperature by the similar intercalation mechan-ism to that of LIBs are highly desired because of the abundance ofsodium resources. Nevertheless, SIBs neither meet sufficient safetyrequirement nor have a significant cost-down effect as long as theyuse non-aqueous electrolytes. For example, non-aqueous electrolytesoften induce exothermic decomposition by the release oxygen gasfrom metal oxide cathode in SIBs.1,2 Besides, conventional non-aqueous electrolytes are generally costly and have poor ionicconductivity. To overcome these disadvantages, aqueous sodium-ionbatteries with aqueous electrolyte that are safe, cost effective, andoffer high ionic conductivity have been proposed as attractivealternatives for large-scale energy storage.

The last but not least problem associated with aqueous electro-lytes is the low operating voltage to avoid the electrochemical

decomposition of water. Most recently, however, several groupshave reported that certain kinds of aqueous electrolytes without freewater molecule have wider electrochemical window and reasonedthat the water molecules hydrated to lithium ion do not engage inelectrochemical decomposition. For example, 21m lithium bis(tri-fluoro methanesulfonyl)imide (LiTFSI) aq.,3 21m LiTFSI + 7mlithium trifluoromethane sulfonate (LiOTf ) aq.,4 saturated LiNO3

aq.,5 and 0.25M Li-PO4 buffer with saturated disodium propane-1,3-disulfonate (PDSS) aq.,6 where m represents a molality unit(molality [m] = mole of solute/weight of solvent [mol kg¹1]) andM represents a molarity unit (molarity [M] = mole of solute/volumeof solution [mol L¹1]), respectively. In such highly concentratedaqueous Li electrolytes, not only the water activity but also thesolubility of the electrode is reduced by water solvation and thicksolid electrolyte interface (SEI).

These positive results in aqueous LIBs encouraged us to studythe concentration effect of aqueous electrolytes containing sodiumions on the practical electrochemical window of water. As highlyconcentrated electrolytes of aqueous Na electrolytes, 2M Na2SO4

aq.,7 5M NaNO3 aq.,8 and 10M NaClO4 aq.9,10 have been alreadyreported. Among them, 2M Na2SO4 aq. molarity is insufficient.Besides, the deterioration of the NASICON-type NaTi2(PO4)3 (NTP)anode for aqueous SIBs by the corrosive side reaction in NaNO3

electrolyte has been reported.11 Accordingly, we focused on NaClO4

as a highly soluble conventional sodium salt, because its concen-tration can be increased to 10M, which corresponds to 17m.12 Sinceour first report about NTP,7 it has been used as actually unique anodefor aqueous SIB in many literatures.13–20 NTP has the 2.1V (vs.Na/Na+) flat voltage plateau derived from the Ti4+/Ti3+ redox. It isslightly under H2 evolution potential at pH = 7, nevertheless it isavailable as anode for aqueous SIB.

Electrochemistry Received: November 26, 2016Accepted: January 20, 2017Published: April 5, 2017

The Electrochemical Society of Japan http://dx.doi.org/10.5796/electrochemistry.85.179

179

On the other hand, many Prussian blue analogues (PBAs) havebeen reported as cathode for SIB. They can contain varying amountsof sodium ions in the open sites of the cyano-bridged jungle gym-type metal organic framework (MOF) and the sodiated molecularformula is written as NaxM[MB(CN)6]y·zH2O (0 ¯ x ¯ 2, 0 ¯ y ¯ 1,0 ¯ z). In particular, many hexacyanoferrates (HCFs, MB = Fe) witha variation of M (M = Ni,21,22 Cu,23 Co,24 Fe,25 and Mn10) havebeen reported as cathode for aqueous SIB. According to the previousreports for M = Ni and Cu, the redox center in aqueous electrolytewas only the [Fe2+(CN)6]4¹/[Fe3+(CN)6]3¹, and neither Ni nor Cuwas active and their capacities were only about 60mAhg¹1.On the other hand, for M = Co and Fe, Co2+/Co3+ and Fe2+/Fe3+

as well as [Fe2+(CN)6]4¹/[Fe3+(CN)6]3¹ are active even in aqueouselectrolyte. Recently, it was reported that NMHCF (M =Mn)in aqueous electrolyte also has two active redox couples,[Fe2+(CN)6]4¹/[Fe3+(CN)6]3¹ and Mn2+/Mn3+, in half-cell config-uration with active carbon as counter electrode and Ag-AgCl asreference electrode.

In this study, in order to confirm the high-concentration effectin NaClO4 aq. electrolyte, we have selected sodium manganesehexacyanoferrate (NMHCF) as a cathode and NTP as an anode,respectively. In addition, we investigated the operation anddeterioration mechanism of NMHCF//NTP aqueous SIB by tracingpH, electrode elution, precipitation in electrolytes.

2. Experimental

2.1 Preparation of cathodeAs a cathode active material, NMHCF was synthesized by the co-

precipitating method described previously.26,27 Briefly, Na4Fe(CN)6·10H2O (99%, Sigma-Aldrich) and excess NaCl (99.5%, NacalaiTesque) were dissolved in a solution of distilled water and ethanol.An aqueous solution of MnCl2·4H2O (99.0%, Wako Pure ChemicalIndustries Ltd.) was dropped into the solution and stirred vigorouslyfor 2 h. The suspension was filtered and washed by a water/ethanolsolution three times, and a light green precipitate product wasobtained. The precipitate product was dried at 100°C in vacuumovernight, after which a final light blue NMHCF powder wasobtained. The lattice structure of the product was characterized withan X-ray powder diffractometer (XRD, 50 kV and 300mA, Cu-KA,RINT2100HLR/PC, Rigaku Corp.). The morphology of the productwas confirmed by scanning electron microscopy (SEM, JSM-6340F,JEOL Ltd.). The chemical composition of the NaxMn[Fe(CN)6]y·zH2O was ascertained using inductively coupled plasma-atomicemission spectroscopy (ICP-AES, Optima 8300, PerkinElmer Co.,Ltd.), atomic adsorption spectrophotometry (AAS, Z-5310, HitachiHigh-Tech Science Corp.), and thermogravimetric analysis (TGA,Thermo Plus TG8110, Rigaku Corp.). To improve electronicconductivity, NMHCF powder and acetylene black (AB, DenkaCo., Ltd.) were mixed in a mortar, and the resulting NMHCF/ABmixture had a weight ratio of 70/25.

2.2 Preparation of anodeAs an anode active material, NTP was prepared by a solid-state

method previously reported.16 A stoichiometric mixture of Na2CO3

(99.7%, Kishida Chemical Co., Ltd.), TiO2 (99%, Sigma-Aldrich),and NH4H2PO4 (99%, Wako Pure Chemical Industries Ltd.) waspotted into an alumina container along with 3mmT zirconia balls.The precursor was subsequently prepared by wet ball-milling inacetone at 400 rpm for 1 h. After the acetone was evaporated toobtain a dry mass, the precursor was ground and decomposed at350°C for 3 h in air to eliminate the ammonia. The resulting powderwas ground, pressed into cylindrical pellets, and calcined at 800°Cfor 12 h in air. In a previously reported procedure,7 AB was added toNTP powder, the mixture was vigorously ball-milled in an aluminacontainer along with 3mmT zirconia balls, and the carbon-coated

powder was subsequently annealed at 800°C for 1 h in an Ar + H2

atmosphere furnace. The resulting NTP/AB had a weight ratio of70/25.

2.3 Electrochemical evaluationCathode and anode pellets were fabricated by mixing NMHCF/

AB or NTP/AB with a polytetrafluoroethylene (Polyflon PTFE F-104, Daikin Industries Ltd.) binder in a weight ratio of 95/5 andsubsequently punched into discs. The cathode and anode activematerials were approximately 20 and 30mg cm¹2 weight loading,respectively. The cathode and anode pellets were approximately200µm thick. These disc pellets were sandwiched by sheets oftitanium mesh (Thank Metal Co., Ltd.). The aqueous electrolytesused here were NaClO4 aqueous solution with various concentrationof 1, 7, 14, and 17m. Ag-AgCl/saturated KCl (RE-6, BAS Inc.)and NTP disc were used as the reference and counter electrodes,respectively. The cathode/anode weight balance for this ion-typecell is 2:3, and the cathode/anode capacity balance is approximately2:3. A three-electrode electrochemical cell (half cell) and a two-electrode cell (full cell) using aqueous electrolytes were used tomeasure the cyclic voltammetry (CV) and the galvanostatic charge/discharge cycle test. The CV was performed with a Versastat 3(AMETEK Inc.). Galvanostatic charge/discharge tests were carriedout with a cycler (Nagano & Co., Ltd.) at various constant currentdensities.

2.4 Investigation of the properties of electrochemical cellduring charge/discharge process

To confirm the lattice behavior of NMHCF and the valencechanges of Mn and Fe in NMHCF during the charge/dischargecycle, ex-situ XRD and ex-situ X-ray photoelectron spectroscopy(XPS, JPS-9010MC/IV, JEOL Ltd.) were carried out. The measure-ment points on the first charge/discharge cycle were the initial state,charged states up to 0.9V, 1.2V, and 1.3V, which were followedby discharged states down to 0.7V and 0.2V vs. Ag/AgCl,respectively. At these points, the electrochemical cells weredismantled, and the cathode pellets were removed from the titaniummesh, washed with water in a glove box filled with Ar, and vacuum-dried at room temperature for 1 h before the measurements. At thesame points, to confirm the pH and mass of metal ions eluted intothe electrolyte, ICP-AES and pH measurements were performed.Here, the ratios of cathode pellet and electrolyte in the electrochem-ical cell were approximately 2mg and 1mL, respectively.

3. Results and Discussion

3.1 Cathode propertiesAs shown in Fig. 1(a), there were not any other peaks except

cubic NMHCF (ICSD # 75-4637) with a Pm-3m in the diffractionprofile of the obtained light blue powder. The morphology of theobtained NMHCF sample is shown in Fig. 1(b). Instead of the200 to 500 nm-sized nanocrystal cubes previously reported,24,25,28

approximately 200 nm-sized round particles were observed, whichwere morphologically similar to particles reported by J. Song.26 Onthe basis of elemental analysis by ICP-AES for Mn and Fe, AAS forNa, and TGA for water content, the NMHCF molecular formulawas determined as Na1.24Mn[Fe(CN)6]0.81·1.28H2O. Although thecrystal water content in the sample depends on the temperature andhumidity,21 the theoretical capacity of this hydrate can be roughlyestimated as 120mAhg¹1 on the basis of 1.24 Na+ extraction/insertion.

3.2 Cyclic voltammetryFigures 2(a) and (b) show the CV profiles of the titanium mesh

current collector in 1m and 17m NaClO4 aqueous electrolytes,respectively. On the other hand, Figs. 2(c) and (d) show the CV

Electrochemistry, 85(4), 179–185 (2017)

180

profiles of NMHCF and NTP in 1m and 17m NaClO4 aqueouselectrolytes, respectively. The following potential values aremeasured against the Ag-AgCl reference electrode. The graycolored ranges in Figs. 2(a) and (b) indicate the theoreticalelectrochemical window of water at each pH value of the aqueouselectrolytes. As shown in Fig. 2(a), oxidation/reduction (ox./red.)peaks were observed at 1.29/¹0.62V in 1m NaClO4 aq. electrolyte.By comparison with the theoretical water ox./red. potential(0.618V/¹0.612 at pH = 7), the oxygen overpotential (definedas GO2,1m) on the titanium mesh was approximately 0.67V in 1mNaClO4 aq. electrolytes shown in Fig. 2(a). However, there wasno hydrogen overpotential (GH2,1m µ 0.0V) on the titanium mesh.Hereby, a practical stability window of 1m NaClO4 aq. electrolytewas estimated to be approximately 1.9V. On the other hand, asshown in Fig. 2(b), ox./red. peaks were observed at 1.54/¹1.24Vin 17m NaClO4 aq. Electrolyte. By comparison with the theoreticalwater ox./red. potential (0.677/¹0.553V at pH = 6), GO2,17m andGH2,17m on the titanium mesh current collector were approximately0.86 and 0.69V in concentrated 17m NaClO4 aq. electrolyte,respectively. Thus, a practical stability window of 17m NaClO4 aq.electrolyte was approximately enlarged to 2.8V. The gray coloredranges in Figs. 2(c) and (d) show the practical stability window ofeach aqueous electrolyte estimated by Figs. 2(a) and (b).

As shown in Fig. 2(c), on the NMHCF cathode side in 1mNaClO4 aq. electrolyte, two pairs of ox./red. peaks were observed:the first at 0.58/0.39V derived from Fe3+/Fe2+, the second largerpeak at 1.17–1.23/1.00V derived from Mn3+/Mn2+, and the thirdlone oxidative peak from 1.36V derived from O2 evolution. Onthe NTP anode side in 1m NaClO4 aq. electrolyte, as shown inFig. 2(c), one pair of ox./red. peaks was observed at ¹0.67/¹0.93V derived from Ti4+/Ti3+ redox reactions, and a lonereduction peak was observed at ¹1.09V derived from H2 evolution.As in the O2 evolution on the cathode side, the NTP redox reactionmay contain H2 evolution partially because the Ti4+/Ti3+ reductionpeak neighbors the H2 evolution peak. In contrast to the case with1m NaClO4 aq. electrolyte in Fig. 2(c), redox peaks in 17m NaClO4

aq. electrolyte were observed at higher potential, as shown inFig. 2(d). Similar tendency was also reported in an aqueous lithium-ion battery with highly concentrated electrolyte,3 and it is explainedthat redox potentials are shifted higher as guest cation activityincreases. The same explanation holds true for the present aqueoussodium-ion system, because sodium-ion activity in concentrated17m NaClO4 aq. electrolyte is higher than in diluted 1m electrolyte,according to a previous report.29 On the NMHCF cathode side, twopairs of ox./red. peaks were observed the first peak at 0.64/0.51Vderived from Fe3+/Fe2+, the second peak at 1.20–1.37/1.03V

Figure 1. (a) XRD patterns and (b) SEM image of as-prepared NMHCF powder.

Figure 2. (a) CV profiles of titanium mesh in 1m NaClO4 aq., (b) CV profiles of titanium mesh in 17m NaClO4 aq., (c) CV profiles ofNMHCF and NTP in 1m NaClO4 aq., and (d) CV profiles of NMHCF and NTP in 17m NaClO4 aq., respectively.

Electrochemistry, 85(4), 179–185 (2017)

181

derived from Mn3+/Mn2+, and a very small third lone peak at1.49V derived from the O2 evolution. On the NTP anode side, a pairof ox./red. peaks was observed at ¹0.55/¹0.84V. The H2 evolutionpeak in Fig. 2(b) was well separated from the NTP reduction peak inFig. 2(d), suggesting that the NTP anode in 17m NaClO4 aq. couldwork reversibly without any H2 evolution.

3.3 Cathode properties during charge/discharge processFigure 3(a) show the charge/discharge profiles of the NMHCF//

NTP full cell and NTP anode half cell in 1m NaClO4 aq. at a rateof 2.0mAcm¹2. On the other hand, Fig. 3(b) show them in 17mNaClO4 aq. The gray colored ranges in Figs. 3(a) and (b) showthe practical stability window of each electrolyte estimated fromFigs. 2(a) and (b). In association with the results of CV in Fig. 2(c),on the NMHCF cathode side in 1m NaClO4 aq. as shown inFig. 3(a), three voltage plateaus derived from Fe2+/Fe3+, Mn2+/Mn3+ redox and O2 evolution were observed on the first charge,followed by two very small plateaus derived from Fe3+/Fe2+ andMn3+/Mn2+ reduction were observed on discharge. The plateaus inthe second cycle were smaller than those in the first. This resultsuggests that the dissolution of the NMHCF cathode into 1mNaClO4 aq. is more serious than NTP anode. Because the cyclabilityof the NTP//Zn half cell in 1m electrolyte were not so bad as shownin Fig. 3(a). (The clear traces of the NMHCF cathode dissolution

into 1m NaClO4 aq. are illustrated in Table 1). On the other hand,on the NMHCF cathode side in 17m NaClO4 aq. as shown inFig. 3(b), two plateaus derived from Fe2+/Fe3+ redox and Mn2+/Mn3+ redox were observed on the first and second charge,and the first charge/discharge capacities were 124/116mAhg¹1,respectively. These profiles and capacities were consistent with theprevious reports26,27 in non-aqueous electrolyte and the theoreticalcapacity (120mAhg¹1) of NMHCF hydrate, respectively.

The reversible capacity on the NTP anode side was higher in17m NaClO4 aq. than in 1m NaClO4 aq. This positive effect canbe explained by the smaller amount of dissolved oxygen inconcentrated aqueous electrolyte than that in diluted aqueouselectrolyte17,30 and larger hydrogen overpotential (GH2,17m = 0.69V)in 17m than that in 1m NaClO4 aq. as estimated in Fig. 2(b).Figures 3(c) and (d) compare the XRD patterns of NMHCF cathodepellets worked in 1m and 17m NaClO4 aq. at each state of chargepoint on the first charge/discharge cycle in Figs. 3(a) and (b),respectively. According to the XRD patterns of NMHCF in 1mNaClO4 aq. as shown in Fig. 3(c), characteristic peaks wereobserved in the range from initial to the charged state up to 0.9V,but the peaks disappeared on the followed charge and discharge.In addition, at the 0.2V fully discharged state, slight characteristicpeaks were observed again. The jungle-gym structure of NMHCF in1m NaClO4 aq. was broken on charge over 0.9V, but the crystal

Figure 3. (a) First and second charge/discharge curves of NMHCF//NTP full cell and NTP//Zn half cell in 1m NaClO4 aq., (b) First andsecond charge/discharge curves of NMHCF//NTP full cell and NTP//Zn half cell in 17m NaClO4 aq., (c) XRD profiles of NMHCF on thefirst cycle in 1m NaClO4 aq., and (d) XRD profiles of NMHCF on the first cycle in 17m NaClO4 aq., respectively.

Electrochemistry, 85(4), 179–185 (2017)

182

structure was partially recovered on discharge less than 0.7V. On theother hand, in the case of 17m NaClO4 aq., the sharp characteristicXRD peaks of the NMHCF cathode were kept at each point on thefirst cycle as shown in Fig. 3(d) and the lattice behavior in 17mNaClO4 aq. on cycle was in good agreement with the previousreport26 of the NMHCF cathode in non-aqueous electrolyte exceptthe generation of new rhombohedral phase after 1 cycle. Theaqueous circumstance may disturb the generation of rhombohedralanhydride.

3.4 Electrolyte properties during charge/discharge processTo estimate the charge/discharge reaction of NMHCF/NaClO4

aq./NTP full cell on the first cycle, dissolved amounts of thetransition metals (Fe, Mn and Ti) and pH value were monitoredduring first cycle. However, Ti ion was never detected in either 1mor 17m NaClO4 aq. in any state, which means that there are nopossibilities of dissolving from titanium mesh current collector orthe NaTi2(PO4)3 anode. As shown in Table 1, in the initial state, thepH value of the 1m NaClO4 aq. electrolyte was shifted to 4 andboth Fe and Mn ions were detected in 1m NaClO4 aq. electrolyte. Itis assumed that soluble Na1.24Mn2+[Fe2+(CN)6]0.81·1.28H2O at theEveritt’s salt (ES) state31 partially eluted into the 1m diluted NaClO4

aq., so part of the NMHCF structure was broken, then yellow[Fe2+(CN)6]4¹ ferrocyanide ion, Mn2+ and H3O+ hydronium ioneluted into the 1m NaClO4 aq. The acidic pH value at the initialstate can be explained by the elution of H3O+.27,31,32 At the chargedstate up to 0.9V in 1m NaClO4 aq., detected metal ions amount inthe electrolyte and the coloration did not change much compared tothe initial state, but pH value shifted from 4 to 6. H+ concentrationis not increased, because the voltage of NMHCF cathode is muchlower than the potential of O2 evolution as shown in Fig. 3(a).Meanwhile, the voltage plateau of the NTP anode is out of theelectrochemical window, so the alkaline shift of the pH (from 4 to 6)at 0.9V must be caused by OH¹ increasing along H2 evolution onthe NTP anode. But, charged state up to 1.3V in 1m NaClO4 aq.,the pH value goes back to acidity (from 6 to 2). It is reasonablebehavior, because at this voltage range, the cathode voltage locatedon the third voltage plateau jumped over the electrochemicalwindow and it brings H+ increasing along O2 evolution. Conversely,there is no OH¹ increasing along H2 evolution in 17m NaClO4 aq.,because the NTP anode voltage is kept in electrochemical windowduring cycling. So, the pH value in 17m NaClO4 aq. becomes lowerthan in 1m NaClO4 aq. For the charged states up to 1.3V in 1mNaClO4 aq., Fe ion increased, but Mn was not detected. The colorof 1m NaClO4 aq. was changed from yellow to yellow-green andgreen precipitation appeared at the bottom of the beaker-type cell.It is assumed that OH¹, which was generated continuously on theanode, decomposed the partially Na+ de-inserted Na0.43Mn2+-[Fe3+(CN)6]0.81 cathode structure,33 and as a result green-yellow[Fe3+(CN)6]3¹ ferricyanide ion was segregated and green Mn2+Owas precipitated at the bottom of the beaker-type cell, according tothe following equation;

Na0:43Mn½FeðCNÞ6�0:81 þ 2NaOH

! 0:81Na3½FeðCNÞ6� þMnO# þ H2O

On the other hand, according to Nernst equation, H2 evolutionpotential shifts to 0.30V (= 0.058¦pH = 0.058 © (6–0.8)) whenpH = 0.8 in 17m NaClO4 aq., so H2 gas should be easily generatedon NTP anode. However, due to the larger overpotential (GH,17m µ0.69V) than the Nernst shift, there was no OH¹ generation leading todecomposition of NMHCF, and coloring of the electrolyte was notobserved.

3.5 Confirmation of cathode operation in 17m NaClO4 aq.Figure 4 shows XPS spectra of (b) Fe and (c) Mn, while (d) shows

the strongest (100) XRD peak of NMHCF at each state of chargepoint: (i) initial state, (ii) charged state up to 0.9V, (iii) charged stateup to 1.3V, (iv) discharged state down to 0.7V, and (v) dischargedstate down to 0.2V on (a) the first charge/discharge profile ofNMHCF cathode in 17m NaClO4 aq. As shown in Fig. 4(b), thebinding energies (BE) of Fe3/2p and Fe1/2p increased on charge on thelower voltage plateau from (i) to (ii) and decreased on discharge from(iv) to (v). It strongly suggests the [Fe2+(CN)6]4¹/[Fe3+(CN)6]3¹

redox plays a role in the lower voltage region. On the other hand, theBE of Mn3/2p and Mn1/2p shifted slightly higher from (ii) to (iii) andlower from (iii) to (iv). It means the higher voltage plateau corre-sponds to Mn2+/Mn3+ redox instead of [Fe2+(CN)6]4/[Fe3+(CN)6]3¹

redox as shown in Fig. 4(c). Assuming that all valence state of Feand Mn are 2+ in the initial composition of Na1.24Mn[Fe(CN)6]0.81·1.28H2O, in fully desodiated state (Mn[Fe(CN)6]0.81·1.28H2O), thevalence state of Fe becomes 3+, meanwhile the valence state ofMn becomes 2.43+. The smaller XPS peak shift of Mn (Fig. 4(c))than that of Fe (Fig. 4(b)) can be explained by the smaller Mnvalence change between 2+ and 2.43+ on cycle.

Moreover, the complicated lattice behavior as shown in Fig. 4(d)is caused by the Jahn-Teller instability of Mn3+ in MnN6 octahedraand it is consistent with the previously reported NMHCF structuralchanges among monoclinic ((i), (v)), cubic ((ii), (iv)), and tetragonal(iii) during cycling in non-aqueous electrolyte.26

3.6 Cycle performances of NMHCF half cellFigure 5(a) shows the cycle performance of NMHCF half cell

dependence on NaClO4 concentration in aqueous electrolyte. Thetendency of a higher NaClO4 concentration aqueous electrolyteto derive better cyclability of NMHCF cathode is evident fromFig. 5(a) and the best cycle performance was obtained in theNMHCF half cell with the most highly concentrated (17m)electrolyte as shown in Fig. 5(b). In brief, highly concentratedelectrolyte brings stable cyclability by extending the practicalstability window. As shown in Fig. 5(b), the fraction of capacityretention at the 100th cycle against the first cycle at rates of 0.5, 2.0,and 5.0mAcm¹2 were 43, 74, and 92%, respectively. As previousstudies3,30,34 about aqueous batteries have revealed, the cycleperformance of the concentrated electrolyte at a higher rate is better

Table 1. Behavior of the dissolved amounts of the transition metal ions and pH values in the diluted and concentrated aqueous electrolytesof NMHCF/NaClO4 aq./NTP full cell on the first charge/discharge cycle.

State of chargeon the first cycle

1m NaClO4 aq. (pH = 7) 17m NaClO4 aq. (pH = 6)

pH Fe/mol% Mn/mol% Coloration Precipitation pH Fe/mol% Mn/mol% Coloration Precipitation

Initial state 4 4.0 7.3 Yellow No 5 0.0 0.0 Transparent No

Charged up to 0.9V 6 6.8 8.0 Yellow No 5 0.0 0.0 Transparent No

Charged up to 1.3V 2 27 0.0 Yellow green Green 0.5 0.0 0.0 Transparent No

Discharged down to 0.7V 2 26 0.0 Yellow green Green 0.8 0.0 0.0 Transparent No

Discharged down to 0.2V 2 15 0.7 Yellow green Green 0.8 0.0 0.0 Transparent No

Electrochemistry, 85(4), 179–185 (2017)

183

than a lower rate condition. This is probably because the largecharge/discharge overpotential at the high rate condition suppressesthe side reaction of the water decomposition. Conversely, it seems tobe difficult to suppress the water decomposition completely only bythe concentrated electrolyte.

3.7 Full cell performancesThe charge/discharge profile and cyclability of the NMHCF//

17m NaClO4 aq.//NTP aqueous sodium-ion battery are shown inFig. 6(a) and its inset, respectively. Owing to the 2.8V practicalstability window of 17m NaClO4 aq., the aqueous sodium-ionbattery exhibits 1.3, 1.5, and 1.8V plateaus, and 117mAhg¹1 firstdischarge capacity at a rate of 2.0mAcm¹2. As shown in the inset,capacity retention at the 50th cycle against the first cycle was 81%.As shown in Fig. 6(b), the aqueous sodium-ion battery exhibits 117,98, 69, and 56mAhg¹1 discharge capacities at current densities of2.0, 5.0, 10, and 20mAcm¹2, respectively. In spite of the thicker200µm electrode pellet, which is two to four times thicker thanelectrodes of commercial lithium-ion batteries (i.e., 50 to 100 µm),this relatively good rate capability is attributed to larger ionicconductivity and lower viscosity of the aqueous electrolyte than theconventional organic electrolyte. Besides, the 3D open structure ofNMHCF and NTP can be regarded as suitable framework for largeNa ion insertion.

4. Conclusions

We investigated the practical stability windows of 1m and 17mNaClO4 aq., and revealed that the potential ranges of eachelectrochemical window were 1.9 and 2.8V, respectively. By virtueof the extended stable potential window of 17m NaClO4 aq., theaqueous sodium-ion battery of Na1.24Mn[Fe(CN)6]0.81·1.28H2O//NaTi2(PO4)3 was able to exhibit 1.3, 1.5, and 1.8V dischargeplateaus and 117mAhg¹1 of discharge capacity at the first cycle,good rate capability, and good cyclability without serious degrada-tion despite the incomplete crystallinity of the NMFCF cathodewith some defects. Conversely, the narrow stability window of 1mNaClO4 aq., led to severe deterioration after the first cycle. Thedeterioration of the aqueous battery with 1m electrolyte was mainlyattributed to H2 generation by the reductive water decomposition onNTP anode and to the elution of the cathode by the alkalization ofthe aqueous electrolyte. Higher concentrations of electrolyte underhigher rate condition supported more stable operation in thisaqueous sodium-ion system. Therefore, we were able to confirm thatNa1.24Mn[Fe(CN)6]0.81·1.28H2O//NaTi2(PO4)3 with highly concen-trated electrolyte is one of the promising candidates for large-scaleenergy storage.

Figure 5. (a) Cyclability dependence on the NaClO4 concentration of aqueous NMHCF half cell at a rate of 2.0mA/cm2 and (b) Cyclabilitydependence on the current density of NMHCF half cell with 17m NaClO4 aqueous electrolytes.

Figure 4. (a) Typical charge/discharge profile of the NMHCF cathode in 17m NaClO4 aqueous electrolyte. The 5 sampling points forex-situ XPS and XRD measurements on the first cycle are as follows; (i) initial state, (ii) charged up to +0.9V, (iii) charged up to +1.3V,(iv) discharged down to +0.7V, and (v) discharged down to +0.2V vs. Ag/AgCl. XPS spectra of (b) Fe and (c) Mn, while (d) shows detailedXRD patterns of the NMHCF at selected potentials.

Electrochemistry, 85(4), 179–185 (2017)

184

Acknowledgments

This work was financially supported by the Elements StrategyInitiative for Catalysts and Batteries Project and the CooperativeResearch Program of “Network Joint Research Center for Materialsand Devices”, MEXT, Japan.

References

1. P. G. Balakrishnan, R. Ramesh, and T. Prem Kumar, J. Power Sources, 155, 401(2006).

2. J. Zhao, L. Zhao, N. Dimov, S. Okada, and T. Nishida, J. Electrochem. Soc., 160,A3077 (2013).

3. L. Suo, O. Borodin, T. Gao, M. Olguin, J. Ho, X. Fan, C. Luo, C. Wang, and K.Xu, Science, 350, 938 (2015).

4. L. Suo, O. Borodin, W. Sun, X. Fan, C. Yang, F. Wang, T. Gao, Z. Ma, M.Schroeder, A. von Cresce, S. M. Russell, M. Armand, A. Angell, K. Xu, and C.Wang, Angew. Chem., Int. Ed., 55, 7136 (2016).

5. C. Wessells, R. Ruffo, R. A. Huggins, and Y. Cui, Electrochem. Solid-State Lett.,13, A59 (2010).

6. K. Miyazaki, T. Shimada, S. Ito, Y. Yokoyama, T. Fukutsuka, and T. Abe, Chem.Commun., 52, 4979 (2016).

7. S.-I. Park, I. D. Gocheva, S. Okada, and J. Yamaki, J. Electrochem. Soc., 158,A1067 (2011).

8. H. Qin, Z. P. Song, H. Zhan, and Y. H. Zhou, J. Power Sources, 249, 367 (2014).9. P. R. Kumar, Y. H. Jung, B. Moorthy, and D. K. Kim, J. Electrochem. Soc., 163,

A1484 (2016).10. M. Pasta, R. Y. Wang, R. Ruffo, R. Qiao, H.-W. Lee, B. Shyam, M. Guo, Y. Wang,

L. A. Wray, W. Yang, M. F. Toney, and Y. Cui, J. Mater. Chem. A, 4, 4211 (2016).11. K. Nakamoto, Y. Kano, A. Kitajou, and S. Okada, J. Power Sources, 327, 327

(2016).12. J. D. Toner and D. C. Catling, Geochim. Cosmochim. Acta, 181, 164 (2016).13. L. Chen, J. Liu, Z. Guo, Y. Wang, C. Wang, and Y. Xia, J. Electrochem. Soc., 163,

A904 (2016).14. F. Sagane, J. Electrochem. Soc., 163, A2835 (2016).

15. W. Wu, A. Mohamed, and J. F. Whitacre, J. Electrochem. Soc., 160, A497 (2013).16. W. Wu, J. Yan, A. Wise, A. Rutt, and J. F. Whitacre, J. Electrochem. Soc., 161,

A561 (2014).17. W. Wu, S. Shabhag, J. Chang, A. Rutt, and J. F. Whitacre, J. Electrochem. Soc.,

162, A803 (2015).18. X. Li, X. Zhu, J. Liang, Z. Hou, Y. Wang, N. Lin, Y. Zhu, and Y. Qian,

J. Electrochem. Soc., 161, A1181 (2014).19. G. Pang, C. Yuan, P. Nie, B. Ding, J. Zhu, and X. Zhang, Nanoscale, 6, 6328

(2014).20. B. Zhao, B. Lin, S. Zhang, and C. Deng, Nanoscale, 7, 18552 (2015).21. C. D. Wessells, S. V. Peddada, R. A. Huggins, and Y. Cui, Nano Lett., 11, 5421

(2011).22. X. Wu, Y. Cao, X. Ai, J. Qian, and H. Yang, Electrochem. Commun., 31, 145

(2013).23. X. Wu, M. Sun, Y. Shen, J. Qian, Y. Cao, X. Ai, and H. Yang, ChemSusChem, 7,

407 (2014).24. X. Wu, M. Sun, S. Guo, J. Qian, Y. Liu, Y. Cao, X. Ai, and H. Yang,

ChemNanoMat, 1, 188 (2015).25. X. Wu, Y. Luo, M. Sun, J. Qian, Y. Cao, X. Ai, and H. Yang, Nano Energy, 13,

117 (2015).26. J. Song, L. Wang, Y. Lu, J. Liu, B. Guo, P. Xiao, J. J. Lee, X. Q. Yang, G.

Henkelman, and J. B. Goodenough, J. Am. Chem. Soc., 137, 2658 (2015).27. L. Wang, Y. Lu, J. Liu, M. Xu, J. Cheng, D. Zhang, and J. B. Goodenough, Angew.

Chem., Int. Ed., 52, 1964 (2013).28. X. Wu, C. Wu, C. Wei, L. Hu, J. Qian, Y. Cao, X. Ai, J. Wang, and H. Yang, ACS

Appl. Mater. Interfaces, 8, 5393 (2016).29. J. W. Morales, H. R. Galleguillos, F. Hernández-Luis, and R. Rodríguez-Raposo,

J. Chem. Eng. Data, 56, 3449 (2011).30. J.-Y. Luo, W.-J. Cui, P. He, and Y.-Y. Xia, Nat. Chem., 2, 760 (2010).31. L. T. T. Kim, C. Gabrielli, H. Perrot, J. Garcia-Jareno, and F. Vicente, Electrochim.

Acta, 84, 35 (2012).32. J. J. García-Jareño, A. Sanmatías, F. Vicente, C. Gabrielli, M. Keddam, and H.

Perrot, Electrochim. Acta, 45, 3765 (2000).33. R. Koncki and O. Wolfbeis, Anal. Chem., 70, 2544 (1998).34. Y. Yamada, K. Usui, K. Sodeyama, S. Ko, Y. Tateyama, and A. Yamada, Nat.

Energy, 1, 16129 (2016).

Figure 6. (a) First and second charge/discharge curves at a rate of 2.0mAcm¹2 and (b) the rate capability of NMHCF/17m NaClO4 aq./NTP full cell.

Electrochemistry, 85(4), 179–185 (2017)

185