Enhancement of Stability by Positive Disruptive Effect on ...lncae.cicata.ipn.mx/wp-content/uploads/2018/02/... · inhibited,13 and (v) synthesizing hexacyanoferrates with two external

Enhancement of Stability by Positive Disruptive Effect on Mn−FeCharge Transfer in Vacancy-Free Mn−Co Hexacyanoferrate Througha Charge/Discharge Process in Aqueous Na-Ion BatteriesM. A. Oliver-Tolentino,*,† J. Vazquez-Samperio,†,‡ S. N. Arellano-Ahumada,§ A. Guzman-Vargas,‡

D. Ramírez-Rosales,§ J. A. Wang,‡ and E. Reguera*,†

†Laboratorio Nacional de Conversion y Almacenamiento de Energía-CICATA, Instituto Politecnico Nacional, Calzada Legaría 694,Col. Irrigacion, Mexico D.F. 11500, Mexico‡ESIQIE-Departamento de Ingeniería Química, Laboratorio de Investigacion en Materiales Porosos, Catalisis Ambiental y QuímicaFina, Instituto Politecnico Nacional, UPALM Edif. 7 P.B. Zacatenco, GAM, Mexico, D.F. 07738, Mexico§ESFM-Departamento de Física, Instituto Politecnico Nacional, UPALM Edif. 9 Zacatenco, GAM, Mexico, D.F. 07738, Mexico

*S Supporting Information

ABSTRACT: Several materials have been studied as electro-des for aqueous batteries that use sodium as alkali ion; theseinclude Prussian blue analogue or hexacyanoferrates. Theinhibition or disruption on metal−metal charge transfer playsan important role for improving electrochemical stability ofthe material. The stability improvement is achieved when twoexternal metals are coordinated to N ends in the Na-richhexacyanoferrates. Additionally, the presence of vacancies inthe material is another important factor that influences itsstability. In this study, NaxCo1−yMny[Fe(CN)6] has beensynthesized at different Mn/Co ratios by precipitation usingcitrate as a chelating agent to obtain a material withoutvacancies. Its electrochemical behavior during redox processesand the correlation with the electronic interaction between external metal sites in the framework through the interaction ofspins have been studied too. To discuss the effect of the presence of [Fe(CN)6]

n‑ vacancies on the electrochemical process, wesynthesized a material without citrate for obtaining materials with low ferrocyanide vacancies. The vacancy-free Co0.55Mn0.45HFversus n-CoMnHF, were compared in this work. These studies reveal that manganese hexacyanoferrate is unstable. The partialsubstitution of Co by Mn modifies the metals spin ordering and consequently, the interaction between metals coordinated to Nin the cyanide linker. Such partial substitution, with a Mn/Co ratio of 1:1 (Co0.55Mn0.45HF), improves the electrochemicalstability and enhances the discharged potential as well. On the other hand, when vacancies are present, the n-CoMnHFcompound showed a decrease in its crystallinity as well as in its external metal interaction. Both changes may be due to thepresence of coordinated water, which modifies electrochemical performance. A spontaneous hopping from Mn to Fe duringoxidation in n-CoMnHF was detected, but this phenomenon was disrupted in Co0.55Mn0.45HF. Such charge transfer inhibitionwas associated with the modification of electron delocalization on Fe (LS); which was caused by the external metals; mainly byCo.

■ INTRODUCTION

Research about energy storage has grown in recent yearsmotivated by the need to ensure that the use of renewableenergy will be economically and technologically viable, andenvironmentally friendly.1 Rechargeable batteries have shownpromising results because of their high storage capacities,especially lithium-ion batteries used in electronic devices.However, due to the high cost and low abundance of lithium,many studies have focused on using other ions, such assodium, magnesium, calcium, and zinc, to replace lithium inrechargeable batteries.2,3 Interesting results about theseelectrochemical storage systems in aqueous and nonaqueousmedia have been reported.4 In particular, aqueous batteries

have shown promising results due to the easier desolvationprocess of alkali ions and the lower viscosity in aqueouselectrolyte solutions as compared to the organic one.5

Different materials have been studied as electrodes foraqueous batteries that use sodium as an alkali ion; theseinclude: rock salt NaxMnO2,

6 the polyanionic compoundNa2FeP2O7,

7 a NASICON-type structure specially doped withTi,8,9 and Prussian blue analogue (PBA).10−15 In particular, theeasy diffusion of Na+ accompanied by solvation water within

Received: June 8, 2018Revised: August 20, 2018Published: August 22, 2018

Article

pubs.acs.org/JPCCCite This: J. Phys. Chem. C 2018, 122, 20602−20610

© 2018 American Chemical Society 20602 DOI: 10.1021/acs.jpcc.8b05506J. Phys. Chem. C 2018, 122, 20602−20610

Dow

nloa

ded

via

INST

PO

LIT

EC

NIC

O N

AC

ION

AL

on

Sept

embe

r 20

, 201

8 at

20:

16:1

8 (U

TC

).

See

http

s://p

ubs.

acs.

org/

shar

ingg

uide

lines

for

opt

ions

on

how

to le

gitim

atel

y sh

are

publ

ishe

d ar

ticle

s.

open framework cavities has been reported in the context ofPBAs used as cathodes in aqueous batteries.16

Both PBAs or hexacyanoferrates are crystalline openframework materials that show the general molecular formulasAxM[P(CN)6], where P and M are cations octahedrallycoordinated to C and M to N, respectively. The number ofcations in balance charge (A) depends on the oxidation state ofP and M. Since Wessells reported the insertion of sodium innickel and copper hexacyanoferrate for use in batteries,17

different strategies for improving the electrochemical behaviorof hexacyanoferrate as a cathode in aqueous sodium ionbatteries have been reported; among these are (i) increasingelectrolyte concentration, which increases the activity of Na+

ions,14 to slightly improve working potential (ii) usingpolyethilenglicol in an aqueous electrolyte solution to increasethe operation potentials of batteries,18 (iii) adding surfactant tothe electrolyte, which lead to molecules getting adsorbed onthe electrode surface via electrostatic absorption, to efficientlysuppress the evolution of hydrogen or oxygen,19 (iv)synthesizing low-defect Prussian blue with vacancy-free[Fe(CN)6]

n‑ to increase electrochemical stability by blockingactive sites in the lattice where the coordination water isinhibited,13 and (v) synthesizing hexacyanoferrates with twoexternal metals coordinated to N ends to promote animprovement in electrochemical properties; in NiCu hexacya-noferrate, it was associated with tunable reaction potentials,20

and in NiCo hexacyanoferrate, it was attributed to themodification of electron density.21 However, these materialsonly exhibited a specific capacity near 60 and 80 mA h g−1,respectively. In this context, MnCo hexacyanoferrate, whichcan reach a specific capacity near 120 mA h g−1, is a materialwith promising results. However, Pasta et al.22 reported thesynthesis of NayCo1−xMnx[Fe(CN)6] with the purpose ofdecreasing the amount of Co in the framework to decreaseelectrode cost. Their results revealed the presence of (i) low[Fe(CN)6]

n‑ vacancies, (ii) a spontaneous electron hoppingfrom Fe to Mn during the electrochemical process, and (iii)poor electrochemical stability associated with manganese dueto the oxidization of this metallic ion to Mn3+ producing alocalized strain attributed to contraction in Mn−N bond takesplace and also the Mn3+ exhibited a Jahn−Teller distortion. Onthe other hand, Kurihara et al.23 reported that in theMn1−yCoy[Fe(CN)6] system at y > 0.33, the spontaneouselectron transfer between Mn and Fe is inhibited, which isassociated with suppressed redox voltages of Mn and Co,probably because of the anomalous spin states of Co and Mnin PBA however, this phenomenon is not fully understood.For this reason, the present study is divided into two parts.

In the first, we synthesize NaxCo1−yMny[Fe(CN)6] at differentMn/Co ratios using citrate as a chelating agent to obtain amaterial without vacancies; electrochemical behavior duringredox processes are correlated with electronic interactionbetween external metal sites in the framework through theinteraction of spins using EPR spectroscopy. In the second

part, to discuss the effect of the presence of vacancies on theelectrochemical process, we synthesized a material withoutcitrate, thus obtaining materials with low ferrocyanidevacancies. The results provide evidence of the disruption ofmetal−metal charge transfer between Mn and Fe in vacancy-free Mn−Co hexacyanoferrate.

■ EXPERIMENTAL SECTIONSynthesis of Samples. NaxCo1−yMny[Fe(CN)6] was

obtained by coprecipitation method from the mixture ofsolution A (0.1 M of Na4Fe(CN)6), solution B (0.1 M ofsodium citrate and 0.1 M of Mn(NO3)2, and/or Co(NO3)2) inthe Y-type micromixer, using a peristaltic pump at 20 rpm. Thesolution was stirred for 12 h at room temperature; the resultingprecipitate was washed with distilled water and ethanol severaltimes and was finally dried under vacuum at 50 °C for 24 h.The samples were labeled CoHF, Co0 .8Mn0.2HF,Co0.55Mn0.45HF, Co0.3Mn0.7HF, and MnHF.

Physical Characterization. The elemental composition ofhexacyanoferrates was determined by optical emissionspectrometry using inductively coupled plasma (OES-ICP),with a PerkinElmer OPTIMA 8300 spectrophotometer. X-raypowder diffraction (XRD) patterns were collected with aBruker D8 Advance diffractometer in the Bragg−Brentanoconfiguration using CuKα radiation (λα = 1.5418 Å). TheRaman spectra were obtained on a Thermo Scientific DXRspectrometer with a 532 nm laser. Electron ParamagneticResonance (EPR) spectra of powder samples were recordedusing a Bruker Elexsys E-500-II EPR spectrometer operating atX-band frequency (9.4186 GHz), equipped with 100 kHz fieldmodulation and phase sensitive detection to obtain the firstderivative signal. EPR measurements were carried out at 300 Kand at 77 K using a liquid-N2-immersion dewar. The in situinfrared was recorded with an FTIR PerkinElmer spectropho-tometer using a SP-02 spectroelectrochemical cell fromSpectroelectrochemistry Partners; the cell was mounted on aPike MIRacle ATR system. Mossbauer spectra were recordedat 77 K with a WissEl Elektronik GmbH MRG500 conven-tional constant acceleration spectrometer, equipped with akrypton proportional detector. The c-radiation source was57Co of 925 MBq (25 mCi) within a rhodium matrixmaintained at room temperature. Chemical isomer shift (IS)data are given relative to α-Fe. Absorption spectra were fittedby using the NORMOS program.

Electrochemical Characterization. Working electrodeswere prepared by stirring 80 wt % of hexacyanoferrate powder,10 wt % of amorphous carbon (Timcal SuperP Li), and 10 wt% polyvinylidenedifluoride (Aldrich) in N-methyl-2-pyrroli-done, then the mixture was coated on a carbon plate (Fuel Cellgrade) and dried at 40 °C in vacuum for 24 h. The amount ofhexacyanoferrate deposited was 5 mg cm−2. The electro-chemical data were recorded with a Biologic potentiostat−galvanostat SP300 using a three electrodes cell, where Ag/AgCl (1 M KCl) electrode was used as a reference electrode. A

Table 1. Sample Composition, a Lattice Parameter, the Mossbauer Spectra Isomer Shift, and Quadrupole Splitting Values

sample formula a (Å) IS (mm/s) QS (mm/s)

CoHF Na1.84Co[Fe(CN)6]0.96 10.37 −0.1894 0.1841Co0.8Mn0.2HF Na1.88Co0.8Mn0.2[Fe(CN)6]0.97 10.41 −0.1897 0.1872Co0.55Mn0.45HF Na1.88Co0.55Mn0.45[Fe(CN)6]0.97 10.46 −0.1903 0.1932Co0.3Mn0.7HF Na1.88Co0.3Mn0.7[Fe(CN)6]0.97 10.51 −0.1908 0.1952MnHF Na1.92Mn[Fe(CN)6]0.98 10.57 −0.1911 0.1972

The Journal of Physical Chemistry C Article

DOI: 10.1021/acs.jpcc.8b05506J. Phys. Chem. C 2018, 122, 20602−20610

20603

large, partially charged hexacyanoferrate on a carbon electrodewas employed as counter electrode, which acts as reversible ionsinks. All electrochemical experiments were carried out in a 1M NaNO3 solution.

■ RESULTS AND DISCUSSIONStructural Characterization. The chemical composition

of every sample calculated with OES-ICP and its respectivevalues are shown in Table 1. All samples showed a highamount of sodium and Mn, or Co/Fe ratios close to 1, whichindicates that the structure has a low amount of vacancies of[Fe(CN)]4−.The X-ray patterns (Figure S1 in Supporting Information)

reveal peak-splitting of the cubic crystalline phase at 2θ= 24.5°,38.8°, 49.5° and 55.93° corresponding to the (220), (420),(440) and (620) planes, respectively. These results indicate thecrystallization of materials in a monoclinic lattice with a P21/nspace group, as previously reported.24 This result was verifiedby Le Bail fitting method (Figure 1). The splitting is due a to

high amount of sodium in the structure, produced using thecitrate-chelating method, which decreases kinetic reaction inthe formation of hexacyanoferrate.13 Na+ ions are locatedasymmetrically at the N-coordinated corners with smaller Na−N distances, thus inducing a distortion on an elementary cell(inset in Figure 1) associated with the cooperative displace-ment of (NaOH2)

+ groups in an alternating cubic [111]direction.24 The lattice parameter a decrease as the amount ofcobalt increases (see Table 1). This could be associated withthe fact that the atomic radius of Co2+ (0.885 Å) is smallerthan that of Mn2+ (0.970 Å).The CN− group has the ability to act as an σ-donor by

donating electrons to the metal coordinated to the N end. Thiselectron subtraction occurs through the 5σ orbital, which has acertain antibonding character. The metal coordinated to C endexhibited π bonding interaction, which involves the t2gelectrons of the metal with the π and π* orbitals of the

ligand.25 This phenomenon allows the oxidation state ofinternal and external metals in the cyano complex to be sensedby infrared (Figure S2) and Raman spectroscopy (Figure 2).

The FTIR spectrum of every material reveals a band ca.2070 cm−1 assigned to M2+−CN−FeII links. The substitutionof Mn2+ (Z/r2 = 3.287) by Co (Z/r2 = 3.652) in the structurepromotes the increase of ν(CN) stretching vibration; due tothe polarizing power (Z/r2) of cobalt, the charge subtraction atthe N end through the 5σ orbital was increased. The splittingin the ν(CN) vibration in MnHF and the asymmetric signalobserved in Co0.55Mn0.45HF and Co0.8Mn0.2HF, can beassociated with not all −CN− bridges being equivalentdue to the decrease of local symmetry by the improvement ofNa+-framework interaction, as it has been reported for zinchexacyanoferrate in dehydrated form.26 Fe cation inhaxacyanoferrate materials presented local Oh point groupsymmetry with an inversion center and two stretching modesA1g and Eg.

27 The vibration mode A1g appears near to 2128cm−1, whereas, at 2095 cm−1, the Eg mode verifies the presenceof M2+−CN−FeII links in all samples.28,29 The band near to2080 cm−1 can be attributed to the T1u vibration, whichindicates a localized structural distortion promoting deviationsfrom ideal Oh to D4h symmetry.27 This band decreasedinversely to cobalt’s amount in relation to Mn due to the highpolarizing power of cobalt, which inhibits symmetry distortionin the framework. The presence of bands at a lower Ramanshift is attributed to the band T1u being able to split into twocontributions, as it has been reported.30

To determine the effect of symmetry distortion of externalmetal on iron coordinated to C in the cyanide group, wecarried out 57Fe Mossbauer analyses. Spectra at roomtemperature are shown in Figure S3. The low value of anisomer shift (IS ≈ −0.19 mm/s) is assigned to FeII−C in lowspin (LS) with electronic configuration t2g

6eg0; the singlet is

due to the t2g levels, which are full of d electrons (S = 0).Additionally, quadrupole splitting caused by any electric fieldgradient was detected (QS).27 However, the results of this

Figure 1. Powder X-ray diffraction and Le Bail profile fitting (redline) for sample Co0.55Mn0.45HF. Inset: Crystalline structure, whereNa+ cations (white spheres), oxygen from water (red spheres). Theframework negative charge is accumulated on the N atom (bluesphere) coordinated to Co/Mn (green spheres). C atom (blacksphere) is coordinated to Fe atom (yellow sphere).

Figure 2. Raman spectra of hexacyanoferrates for (a) CoHF, (b)Co0.8Mn0.2HF, (c) Co0.55Mn0.45HF, (d) Co0.3Mn0.7HF, and (e)MnHF.

The Journal of Physical Chemistry C Article

DOI: 10.1021/acs.jpcc.8b05506J. Phys. Chem. C 2018, 122, 20602−20610

20604

paper show quadrupole splitting QS ≈ 0.19 mm/s, which canbe interpreted as an anisotropic charge involving an importantdistortion in the octahedral environment of the [Fe(CN)6]

4−

block.31 The replacement of cobalt by manganese in theframework (Table 1) decreased the IS value related to thehigher polarizing power of cobalt. The charge subtraction overN ends therefore increased, and therefore increased the π-back-donation from the iron atom toward the CN ligand. Thedecrease of QS can be attributed to a lower cell distortion.These results agree with the Raman experiments discussedabove.The spin interaction between manganese and cobalt in the

structure can be analyzed by EPR spectroscopy (Figure 3); due

to the iron in the open framework material having S = 0, itwould not show any influence on the EPR response. All EPRspectra of MnHF and CoMnHF compounds showed a singletsignal with an average line width of 20 gauss, with theexception of CoHF compound, which was EPR-silent. The Mncoordinated to N exhibited electronic configuration in highspin t2g

3eg2 with S = 5/2,

32such that, in EPR experiments, threeKramer’s doublets, ±5/2, ± 3/2 and ±1/2, were expected.However, at room temperature (Figure 3A), the MnHFcompound shows a broad singlet EPR signal with g = 2.023.Here, the degeneracy of Kramers doublets was removed hadbeen removed through the applied magnetic field and EPRspectra coming from transitions between energy levels of thesedoublets. The resonance at g around 2 arises from thetransition between energy levels of S= ± 1/2 Kramer’s doublet.The high possibility of only the transition of the ground

doublet (±1/2) appearing is because the exited doublets (±5/2and ±3/2) are occupied. If this is so, the lifetime of the excitedstates is generally so short because of the relaxation to thelattice and transitions between them being too broad to beobservable. Therefore, in practice, resonance is almost alwaysrestricted to the ground doublet.25 The shape and line width ofthese signals imply the presence of dipole coupling andexchange interaction. This behavior could be attributed toferromagnetic interaction, which indicates that two magneticorbitals are orthogonal, that ground state of the system hasparallel electron spins.33

In contrast, the CoHF compound did not show any EPRsignal, which is associated with the antiferromagnetic orderingof Co2+ spin (S = 3/2). This behavior in nature is aconsequence of the Pauli principle, where two nonorthogonalmagnetic orbitals lead to antiparallel spin ordering.33

The electrons occupying the orbitals t2g in FeII are partiallydelocalized on the neighboring (Co/Mn) as has been reportedfor Prussian blue.34 In MnHF, which contains Mn2+ (t2g

3eg2),

the t2g and eg orbitals are both exactly half full; the t2g electronwith a fraction of spin from FeII can interact with electrons inthe eg orbitals of Mn, thus promoting a spin order parallel to

Figure 3. Electronic paramagnetic resonance spectra: (A) roomtemperature and (B) 77 K, for (a) CoHF, (b) Co0.8Mn0.2HF, (c)Co0.55Mn0.45HF, (d) Co0.3Mn0.7HF, and (e) MnHF.

Figure 4. (A) Cyclic voltammetry at 1 mV s−1, where the thin line is 1st cycle and the thick line is 2nd cycle for MnHF (e), 20th cycle forCo0.3Mn0.7HF (d), 40th cycle for Co0.55Mn0.45HF, (c) 45th cycle for Co0.8Mn0.2HF (b), and 50th for CoHF (a). (B) Galvanostatic experiment at1C, for (a) CoHF, (b) Co0.8Mn0.2HF, and (c) Co0.55Mn0.45HF.

The Journal of Physical Chemistry C Article

DOI: 10.1021/acs.jpcc.8b05506J. Phys. Chem. C 2018, 122, 20602−20610

20605

each other. However, for CoHF with Co2+ (t2g5eg

2), theinteraction occurs in the t2g orbital, provoking an antiparallelorder to each other.On the other hand, as Co concentration decreases, the EPR

spectra intensity increases until obtaining a maximum intensityfor MnHF. However, this increase is not proportional to Coconcentration. More specifically, the EPR signals ofCo0.3Mn0.7HF and Co0.55Mn0.45HF exhibit practically thesame intensity (see Figure S4), which suggests that spinalignment is modified by Mn−Co interaction.To investigate other mechanisms that may be involved in the

evolution of the EPR spectra with different Co:Mn ratios, theCoMnHF compounds were measured at 77 K (Figure 3B).EPR behavior for CoHF is similar to that observed at 300 K,while the EPR signal decreased for MnHF. This fact can beattributed to cell contraction when the sample is cooled, whichshortens the distance between Mn coordinated to N ends,modifies the spin alignment, and produces a ferrimagneticinteraction. A similar phenomenon is observed for theCo0.3Mn0.7HF sample. These results are ascribed to thepresence of a high amount of manganese, which exhibited ahigh lattice parameter (a), as discussed above. Whereas in thematerial with a low amount of manganese, such asCo0.8Mn0.2HF, the EPR signal was increased, which isindicative of some Co spins of the framework ordering in aparallel way to those of Mn spins. This spin ordering isprobably due to the smaller size in the lattice parameter, suchthat the distance between the external metals does not changesignificantly when the sample is cooled, thus avoiding analignment of spins that cause a ferrimagnetic interaction asrevealed by MnHF and Co0.3Mn0.7HF samples. The parallelalignment between the Co and the Mn spins for materials witha low amount of Mn is confirmed by the EPR spectrum ofCo0.55Mn0.45HF, which was increased 1.73 times with respectto its own intensity at 300 K. This signal increase implies thatan extra number of paramagnetic entities contribute to theEPR signal, which suggests that electron partial delocalizationin the Co2+−NC−FeII−CN−Mn2+ chain is modulated towardCo sites due to their high charge subtraction ability of the ironmodifying the spin orientation.Electrochemical Evaluation. The cyclic voltammetry

profile of each material at 1 mVs−1 in 1 M NaNO3 solutionis shown in Figure 4A. CoHF exhibited two faradaic processes,which agrees with recent reports.21,22 The first at formalpotential (Ef = 0.35 V vs Ag/AgCl) is attributed to anelectrochemical process of the Co3+/Co2+ redox couple. This isaccompanied by a change in electronic configuration, fromhigh (HS: t2g

5 eg2) to low spin (LS: t2g

6eg0); the redox potential

associated with the FeIII/FeII redox couple occurred at Ef =0.87 V vs Ag/AgCl. The increment in formal potential for Fe(LS) with respect to other hexacyanoferrates17 is due to thehigh polarizing power of Co3+ formed in the first redoxprocess, increasing charge subtraction on N ends, andpromoting an increment in the π back bonding interactionon the iron atom and on the cyanide ions. This last decreasedthe absolute energy of the fully filled t2g

6 orbitals (FeII in lowspin configuration), favoring the 2+ oxidation state of the ironcenters.35 The anodic peak at 0.43 V vs Ag/AgCl is associatedwith Na-ion desertion along the [111] direction during theoxidation process, which induces a structural change in the unitcell from monoclinic to cubic.22 The peak to peak separationfor FeIII/FeII redox couple is ΔEp = 60 mV, indicating a goodelectrochemical reversibility, while ΔEp = 100 mV is observed

in the Co3+/Co2+ redox couple. This semireversible process isdue to the stability of Co3+ LS configuration (t2g

6eg0), where a

transfer of charge density from t2g to eg orbitals should takeplace. The semireversibility can be verified by the electro-chemical profile observed between 0.55 and 0.75 V/vs Ag/AgCl, which has been attributed to the gradual rearrangementof the electronic structure during the Co3+ (LS)/Co2+(HS)redox process.13 Any change was observed in the electro-chemical profile after 50 cycles. During the anodic sweep,MnHF showed a faradaic process at 0.52 V vs Ag/AgCl (thinline), which was attributed to an oxidation from FeII (t2g

6eg0)

to FeIII (t2g5eg

0) due to spontaneous electron hopping from Mnto Fe through the cyanide linker, as reported.22 This behaviorcould be explained by the metal to metal charge transfermechanism, as reported for manganese PBA.36

The structural changes in the unit cell during sodiumextraction take place at 0.58 V vs Ag/AgCl. The anodic peak at1.1 V vs Ag/AgCl may be attributed to with the oxidationprocess of Mn2+ (t2g

3 eg2) to Mn3+ (t2g

3 eg1) (see Figure 4A).

The presence of iron 3+ in the structure reduces the electrondensity around [the/a] metal coordinated to nitrogen, whichincreases potential when manganese oxidation takes place. Theabsence of [a cathodic peak associated with manganesereduction, as well as a faradaic process in the second cycle(thick line), indicates the poor stability of MnHF in aqueousmedia during the redox process, which is verified because thesolution turns yellowish. This fact suggests that a fraction ofthe [Fe(CN)6]

3− complex anion is partially decomposed.As discussed previously, the external metal substitution of

manganese by cobalt into the Co0.3Mn0.7HF, Co0.55Mn0.45HF,and Co0.8Mn0.2HF samples showed a Mn2+−NC−FeII−CN−Co2+ chain, which modulates spin delocalization around FeII

(LS). The metal substitution modifies electron density alongthe chain, which affects electrochemical behavior. In general,the electrochemical profiles exhibited two faradaic processes, atlow potential <0.7 V vs Ag/AgCl and at high potential >0.7 Vvs Ag/AgCl for Co0.8Mn0.2HF, any well-defined peak wasdetected in the oxidation process 1, indicating an increment inrearrangement of the electronic structure due to the presenceof Co3+ (LS). However, electrochemical reversibility in redoxprocess 2 decreased (ΔEp= 88 mV), suggesting that Fe (LS)did not solely participate in the faradaic process due tomanganese’s contribution. This can be verified inCo0.3Mn0.7HF, where the faradaic process at high potential isirreversible by the presence of Mn3+ formed during theoxidation process, which increases the hybridization betweenthe Mn and N orbitals and decreases stability. On the otherhand, the Co0.55Mn0.45HF showed better electrochemicalreversibility in every redox process, which indicates fastreaction kinetics (which decreases energy loss in the battery).The charge/discharge experiments at 1C (60 mA g−1) for

CoHF, Co0.8Mn0.2HF, and Co0.55Mn0.45HF are shown inFigure 4B; minor current losses after several cycles of 50, 45,and 40 (thick line), respectively, were observed.The CoHF material exhibited complex galvanostatic

behavior associated with the spin transition in cobalt duringthe redox process; voltage discharge began at 0.8 V vs Ag/AgCl. On the other hand, the CoMnHF compound showedtwo principal and defined plateaus at 0.3−0.58 V vs Ag/AgCland 0.9−1.0 V vs Ag/AgCl. The second process showed ahigher slope than the first. This can be associated with sodiuminsertion/desertion of process two, which takes place in a solidsolution state in a cubic cell; whereas the sodium insertion/

The Journal of Physical Chemistry C Article

DOI: 10.1021/acs.jpcc.8b05506J. Phys. Chem. C 2018, 122, 20602−20610

20606

desertion in the first process occurs by the presence of two-monoclinic and cubic phases, as reported for similar frameworkmaterials.37 The increase in manganese content increasesdischarge voltage at 0.95 and 1.0 V/SCE for Co0.8Mn0.2HF andCo0.55Mn0.45HF, respectively. The specific capacity of materialshere studied is in the following order: 118.81 mAh g−1

(CoHF) > 113.84 mAh g−1 (Co0.8Mn0.2HF) > 112.82 mAhg−1 (Co0.55Mn0.45HF). On the basis of our results, the materialwith the best electrochemical properties is the Co0.55Mn0.45HFsystem, which exhibits higher discharge potential and greaterstability than other materials with manganese, according to theresults of the cyclic voltamperometry. This is associated withcharge density subtraction in the Co2+−NC−FeII−CN−Mn2+

chain, toward the cobalt due to its polarizing power; promotinga decrease in distance between the external metals, modifyingthe electron delocalization between orbitals t2g of Fe and egorbitals of Mn, which can modulate the kinetics electrontransfer to and from eg orbitals.Effect of Vacancies on the Metal−Metal Charge

Transfer. A new synthesis was made with the aim ofunderstanding the influence of vacancies on material behaviorand electronic and electrochemical properties. Considering thebetter electrochemical properties shown by Co0.55Mn0.45HF(free of vacancies), a similar compound was synthesized usingsodium chloride instead of sodium citrate. A solution 1 M ofsodium chloride was used for the new synthesis and the solidobtained was labeled as n-CoMnHF. Its chemical compositionwas Na1.65Mn0.50Co0.50[Fe(CN)6]0.87, revealing the presence ofcrystalline defects by [Fe(CN)6]

4− vacancies. XRD peak-splitting, which is characteristic of monoclinic phases, was notclearly observed for n-CoMnHF (Figure 5A) due to thepresence of a lower amount of sodium with respect toCo0.55Mn0.45HF. A important magnetic difference betweenthem was observed from the EPR spectra (Figure 5B). TheCo0.55Mn0.45HF compound has a strong and a wide singletEPR signal that implies a ferromagnetic order as discussed

above. However, the n-CoMnHF compound showed a weaksinglet signal with a hyperfine splitting superimposed on it,which is characteristic of Mn2+ with an electronic spin of S =1/2 and a nuclear spin of I = 5/2 (see inset in Figure 5B). Thehyperfine splitting implies that the distance between Mn andCo ions was increased, weakening dipole−dipole and exchangeinteractions. This could be attributed with the presence ofdefects in the frameworks by [FeII(CN)6]

4− vacanciesproduced during synthesis. These vacancies promote thecompletion of coordination spheres with water molecules bysome Mn and Co sites. The deficiency of cyanide ligandsinhibits electron delocalization in the iron center with theexternal metals.The EPR spectrum of n-CoMnHF compound at 77 K

reveals a singlet with a hyperfine splitting superimposed on it.The increase in the EPR signal intensity at 77 K compared tothe signal intensity of the same compound at 300 K impliestypical paramagnetic behavior.38 The electrochemical experi-ments (Figure 5C) exhibited a lower specific capacity (87 mAhg−1) than that found in Co0.55Mn0.45HF (112.82 mAh g−1),attributed to the presence of ferrocyanide vacancies39,40 andlower discharge voltage (0.9 V Ag/AgCl) due to the lowpresence of Mn3+ species; whereas Figure 5D showed highelectrochemical stability of Co0.55Mn0.45HF with a fractioncapacity retention of 80% while n-CoMnHF exhibited a valueof 60% after 100 charge/discharge cycles at 1 C; these resultsshowed that the improvement in the stability of hexacyano-ferrates without vacancies can be associated with electronicinteraction between external metals through electron delocal-ization, which is inhibited by the presence of coordinated waterand not just to the blockage of active sites, as previouslyreported.41

Electrochemical Mechanism. In order to elucidatewhether the disruptive effect of spontaneous charge transfersbetween Mn and Fe depend only on the Mn/Co ratio, as wasreported before,42 in situ infrared spectroscopy was carried out

Figure 5. Comparison between Co0.55Mn0.45HF (free of vacancies) and n-CoMnHF (with vacancies): (A) X-ray diffraction, (B) EPR spectra atroom temperature, (C) galvanostatic experiments at 1 C, and (D) fraction capacity retention vs cycle number at 1C.

The Journal of Physical Chemistry C Article

DOI: 10.1021/acs.jpcc.8b05506J. Phys. Chem. C 2018, 122, 20602−20610

20607

during the charge/discharge process in Co0.55Mn0.45HF and n-CoMnHF, and the results are shown in Figure 6. The samplen-CoMnHF (Figure 6A) exhibited a minimum of %T near2065 cm−1, indicating that the metal in the oxidation state of2+ is coordinated to the cyanide linker through C and N.During sodium desertion, an IR vibration at 2150 cm−1 isobserved in the first redox process to around 0.45 vs Ag/AgCl,indicating the presence of Fe(III) LS and suggesting thatmetal−metal charge transfer between Fe and Mn takes place at0.35 V vs Ag/AgCl, as discussed above. An evident IR bandassociated with changes in metal oxidation states coordinatedto N appeared ca. 2100 cm−1. On the other hand, inCo0.55Mn0.45HF (Figure 6B), the band attributed to thepresence of metal coordinated to N ends in oxidation state3+ (2100 cm−1) is observed in the first faradaic process duringsodium desertion, whereas the IR band at 2150 cm−1,attributed to FeIII(LS), appears until 0.75 V vs Ag/AgCl,where the second faradaic process starts. The featuresdiscussed above suggest that the interaction between cobaltand manganese, through spin delocalization of the Fe center,disrupts electron transfer from Fe to Mn.The mechanism proposed for Co0.55Mn0.45HF can be

explained by the charge density distribution of the Co−NC−Fe−CN−Mn chain because, during the sodium desertionprocess, the oxidation of Co2+(HS) to Co3+(LS) and theoxidation of Mn2+(HS) to Mn3+(HS) occur simultaneously.Co3+ increases the subtraction charge over FeII (LS) throughcyanide and decreases the electron density around Mn. Thisfact modulates the electron delocalization toward the cobaltsite, decreasing the capacity to add one electron on the egmanganese orbitals, which inhibits the charge transfer from Feto Mn. On the other hand, the [Fe(CN)6]

n‑ vacancies in n-CoMnHF favors the presence of coordinated water to theexternal metal, as in the chains OH2−Mn-NC-Fe and OH2−Co-NC-Fe, and additionally decrease the interaction with themetal coordinated to nitrogen, inducing electron hopping fromFe to Mn.

■ CONCLUSIONS

The high sodium content within the hexacyanoferrateframework produces cell distortion and some changes inoctahedral symmetry. Higher distortion was exhibited inMnHF and it decreased by the incorporation of Co, whichpossesses high polarizing power, thus modulating the electron

delocalization of Fe (LS) through the cyanide linker in theMn−NC−Fe−CN−Co chain. The Co0.55Mn0.45HF sample,free of vacancies, showed the change from an antiparallel spinorder of Co to a parallel spin order caused by the presence ofMn. The metal to metal charge transfer between Mn and Fewas disrupted in a vacancy-free framework, due to the decreasein the addition capacity of one electron on eg manganeseorbitals during redox process; this was attributed to theinteraction between Co and Mn through the cyanide linker,which improves the electrochemical properties of redoxstability and discharge voltage. Co0.55Mn0.45HF material, freeof vacancies, exhibited a fraction capacity retention of 80%while n-CoMnHF exhibited a value of 60% after 100 cycles ofcharge/discharge.

■ ASSOCIATED CONTENT

*S Supporting InformationThe Supporting Information is available free of charge on theACS Publications website at DOI: 10.1021/acs.jpcc.8b05506.

XRD patterns of Cox Mn1−xHF samples, infrared

spectrum, Mossbauer spectrum, and EPR signal vs

manganese amount with respect to cobalt content in

hexacyanoferrates (PDF)

■ AUTHOR INFORMATION

Corresponding Authors*(M.A.O.-T.) E-mail: [email protected].*(E.R.) E-mail: [email protected].

ORCIDM. A. Oliver-Tolentino: 0000-0001-8454-0837J. A. Wang: 0000-0002-7007-8212E. Reguera: 0000-0002-4452-9091NotesThe authors declare no competing financial interest.

■ ACKNOWLEDGMENTS

The authors are grateful to CONACYT (Project 225115) forthe acquisition of the EPR spectrometer and INFRA 2014-225161. In particular, M.A.O.-T. is grateful to ProjectCONACYT-255354 for their economical support.

Figure 6. Contour maps of in situ infrared experiments during charge process in (A) n-CoMnHF and (B) Co0.55Mn0.45HF, where R is Co or Mn.

The Journal of Physical Chemistry C Article

DOI: 10.1021/acs.jpcc.8b05506J. Phys. Chem. C 2018, 122, 20602−20610

20608

■ REFERENCES(1) Pan, H.; Hu, Y.-S.; Chen, L. Room-Temperature StationarySodium-ion Batteries for Large-Scale Electric Energy Storage. EnergyEnviron. Sci. 2013, 6, 2338−2360.(2) Yabuuchi, N.; Kubota, K.; Dahbi, M.; Komaba, S. ResearchDevelopment on Sodium-Ion Batteries. Chem. Rev. 2014, 114,11636−11682.(3) Hwang, J.-Y.; Myung, S.-T.; Sun, Y.-K. Sodium-Ion Batteries:Present and Future. Chem. Soc. Rev. 2017, 46, 3529−3614.(4) Kim, H.; Hong, J.; Park, K.-Y.; Kim, H.; Kim, S.-W.; Kang, K.Aqueous Rechargeable Li and Na Ion Batteries. Chem. Rev. 2014, 114,11788−11827.(5) Fernandez-Ropero, A. J.; Zarrabeitia, M.; Reynaud, M.; Rojo, T.;Casas-Cabanas, M. Toward Safe and Sustainable Batteries:Na4Fe3(PO4)2P2O7 as a Low-Cost Cathode for RechargeableAqueous Na-Ion Batteries. J. Phys. Chem. C 2018, 122, 133−142.(6) Yuan, A. B.; Tian, L.; Xu, W. M.; Wang, Y. Q. Al-Doped SpinelLiAl0.1Mn1.9O4 with Improved High-Rate Cyclability in AqueousElectrolyte. J. Power Sources 2010, 195, 5032−5038.(7) Manickam, M.; Singh, P.; Thurgate, S.; Prince, K. RedoxBehavior and Surface Characterization of LiFePO4 in LithiumHydroxide Electrolyte. J. Power Sources 2006, 158, 646−649.(8) Zhang, F.; Li, W.; Xiang, X.; Sun, M. Nanocrystal-AssembledPorous Na3MgTi(PO4)3 Aggregates as Highly Stable Anode forAqueous Sodium-Ion Batteries. Chem. - Eur. J. 2017, 23, 12944−12948.(9) Lei, P.; Wang, Y.; Zhang, F.; Wan, X.; Xiang, X. Carbon-CoatedNa2.2V1.2Ti0.8(PO4)3 Cathode with Excellent Cycling Performance forAqueous Sodium-Ion Batteries. ChemElectroChem 2018 ,DOI: 10.1002/celc.201800379.(10) Oliver-Tolentino, M. A.; Vazquez-Samperio, J.; Cabrera-Sierra,R.; Reguera, E. Materials for Aqueous Sodium-ion Batteries: CationMobility in a Zinc Hexacyanoferrate Electrode. RSC Adv. 2016, 6,108627−108634.(11) Wessells, C. D.; Huggins, R. A.; Cui, Y. CopperHexacyanoferrate Battery Electrodes with Long Cycle Life and HighPower. Nat. Commun. 2011, 2, 550.(12) Wu, X. Y.; Sun, M. Y.; Shen, Y. F.; Qian, J. F.; Cao, Y.l.; Ai, X.P.; Yang, H. X. Energetic Aqueous Rechargeable Sodium-Ion BatteryBased on Na2CuFe(CN)6−NaTi2(PO4)3 Intercalation Chemistry.ChemSusChem 2014, 7, 407−411.(13) Wu, X.; Sun, M.; Guo, S.; Qian, J.; Liu, Y.; Cao, Y.; Ai, X.;Yang, H. Vacancy-Free Prussian Blue Nanocrystals with HighCapacity and Superior Cyclability for Aqueous Sodium-Ion Batteries.ChemNanoMat 2015, 1, 188−193.(14) Li, W.; Zhang, F.; Xiang, X.; Zhang, X. High-efficiency Na-storage Performance of Nickel-Based Ferricyanide Cathode in High-Concentration Electrolytes for Aqueous Sodium-ion Batteries.ChemElectroChem 2017, 4, 2870−2876.(15) Zhang, D.; Zhang, J.; Yang, Z.; Ren, X.; Mao, H.; Yang, X.;Yang, J.; Qian, Y. Nickel Hexacyanoferrate/Carbon Composite as aHigh-Rate and Long-Life Cathode Material for Aqueous HybridEnergy Storage. Chem. Commun. 2017, 53, 10556−10559.(16) Chen, L.; Shao, H.; Zhou, X.; Liu, G.; Jiang, J.; Liu, Z. Water-Mediated Cation Intercalation of Open-Framework Indium Hex-acyanoferrate with High Voltage and Fast Kinetics. Nat. Commun.2016, 7, 11982.(17) Wessells, C. D.; Peddada, S. V.; McDowell, M. T.; Huggins, R.A.; Cui, Y. The Effect of Insertion Species on Nanostructured OpenFramework Hexacyanoferrate Battery Electrodes. J. Electrochem. Soc.2012, 159, A98−A103.(18) Niu, L.; Chen, L.; Zhang, J.; Jiang, P.; Liu, Z. Revisiting theOpen-framework Zinc Hexacyanoferrate: The role of TernaryElectrolyte and Sodium-ion Intercalation Mechanism. J. Power Sources2018, 380, 135−141.(19) Hou, Z.; Zhang, X.; Li, X.; Zhu, Y.; Liang, J.; Qian, Y.Surfactant Widens the Electrochemical Window of an AqueousElectrolyte for Better Rechargeable Aqueous Sodium/Zinc Battery. J.Mater. Chem. A 2017, 5, 730−738.

(20) Li, W.; Zhang, F.; Xiang, X.; Zhang, X. Nickel-SubstitutedCopper Hexacyanoferrates as Superior Cathode for Aqueous Sodium-Ion Batteries. ChemElectroChem 2018, 5, 350−354.(21) Li, W.; Zhang, F.; Xiang, X.; Zhang, X. ElectrochemicalProperties and Redox Mechanism of Na2Ni0.4Co0.6[Fe(CN)6] Nano-crystallites as High-Capacity Cathode for Aqueous Sodium-IonBatteries. J. Phys. Chem. C 2017, 121, 27805−27812.(22) Pasta, M.; Wang, R. Y.; Ruffo, R.; Qiao, R.; Lee, H.-W.; Shyam,B.; Guo, M.; Wang, Y.; Wray, L. A.; Yang, W.; et al. Manganese−Cobalt Hexacyanoferrate Cathodes for Sodium-ion Batteries. J. Mater.Chem. A 2016, 4, 4211−4223.(23) Kurihara, Y.; Moritomo, Y. Electrochemical, Structural, andElectronic Properties of Mn−Co Hexacyanoferrates Against LiConcentration. Jpn. J. Appl. Phys. 2014, 53, 067101.(24) Song, J.; Wang, L.; Lu, Y.; Liu, J.; Guo, B.; Xiao, P.; Lee, J.-J.;Yang, X.-Q.; Henkelman, G.; Goodenough, J. Removal of InterstitialH2O in Hexacyanometallates for a Superior Cathode of a Sodium-IonBattery. J. Am. Chem. Soc. 2015, 137, 2658−2664.(25) Bertran, J. F.; Pascual, J. B.; Ruiz, E. R. The CN Stretch ofHexacyanometallates as a Sensor of LigandOuter CationInteractionsI. Ferricyanides and Cobalticyanides. Spectrochim.Acta A: Molecular Spectroscopy 1990, 46, 685−689.(26) Oliver-Tolentino, M.; Ramos-Sanchez, G.; Guzman, G.; Avila,M.; Gonzalez, I.; Reguera, E. Water Effect on Sodium Mobility inZinc Hexacyanoferrate During Charge/Discharge Processes inSodium ion-based Battery. Solid State Ionics 2017, 312, 67−72.(27) Ojwang, D. O.; Grins, J.; Wardecki, D.; Valvo, M.; Renman, V.;Haggstrom, L.; Ericsson, T.; Gustafsson, T.; Mahmoud, A.; et al.Structure Characterization and Properties of K-Containing CopperHexacyanoferrate. Inorg. Chem. 2016, 55, 5924−5934.(28) Samain, L.; Gilbert, B.; Grandjean, F.; Long, G. J.; Strivay, D.Redox Reactions in Prussian blue Containing Paint Layers as a Resultof Light Exposure. J. Anal. At. Spectrom. 2013, 28, 524−535.(29) Gomez, A.; Reguera, E. The Structure of Three CadmiumHexacyanometallates(II): Cd2[Fe(CN)6]·8H2O, Cd2[Ru(CN)6]·8H2O and Cd2[Os(CN)6]·8H2O. Int. J. Inorg. Mater. 2001, 3,1045−1051.(30) Fernandez-Bertran, J.; Reguera, E.; Dago, A.; Lopez-Hernandez,C.; Infante, G. Synthesis and Characterization of Two Complexes ofGlycine with Lanthanum Hexacyanoferrate (III) andHexacyanocobaltate(III). Polyhedron 1996, 15, 315−319.(31) Reguera, E.; Fernandez-Bertran, J. Effect of the Water ofCrystallization on the Mossbauer Spectra of Hexacyanoferrates (IIand III). Hyperfine Interact. 1994, 88, 49−58.(32) Manzan, R. S.; Donoso, J. P.; Magon, C. J.; Silva, I. d. A. A.;Russel, C.; Nalin, M. Optical and Structural Studies of Mn2+ DopedSbPO4-ZnO-PbO Glasses. J. Braz. Chem. Soc. 2015, 26, 2607−2614.(33) Ohkoshi, S.-I.; Iyoda, T.; Fujishima, A.; Hashimoto, K.Magnetic Properties of Mixed Ferro-Ferrimagnets Composed ofPrussian Blue Analogs. Phys. Rev. B: Condens. Matter Mater. Phys.1997, 56, 11642−11652.(34) Mayoh, B.; Day, P. Charge Transfer in Mixed-Valence Solids.Part VIII. Contribution of Valence Delocalisation to the Ferromag-netism of Prussian Blue. J. Chem. Soc., Dalton Trans. 1976, 1483−1486.(35) Scholz, F.; Dostal, A. The Formal Potentials of Solid MetalHexacyanometalates. Angew. Chem., Int. Ed. Engl. 1996, 34, 2685−2687.(36) Bhattacharjee, A.; Saha, S.; Koner, S.; Ksenofontov, V.; Reiman,S.; Gutlich, P. Metal-to-metal Electron Transfer and MagneticInteractions in a Mixed-Valence Prussian Blue Analogue. J. Magn.Magn. Mater. 2006, 302, 173−180.(37) Asakura, D.; Okubo, M.; Mizuno, Y.; Kudo, T.; Zhou, H.;Ikedo, K.; Mizokawa, T.; Okazawa, A.; Kojima, N. Fabrication of aCyanide-Bridged Coordination Polymer Electrode for EnhancedElectrochemical Ion Storage Ability. J. Phys. Chem. C 2012, 116,8364−8369.(38) Pathak, N.; Gupta, S. K.; Ghosh, P. S.; Arya, A.; Natarajan, V.;Kadam, R. M. Probing Local Site Environments and Distribution of

The Journal of Physical Chemistry C Article

DOI: 10.1021/acs.jpcc.8b05506J. Phys. Chem. C 2018, 122, 20602−20610

20609

Manganese in SrZrO3:Mn; PL and EPR Spectroscopy Complimentedby DFT Calculations. RSC Adv. 2015, 5, 17501−17513.(39) Wu, X.; Shao, M.; Wu, C.; Qian, J.; Cao, Y.; Ai, X.; Yang, H.Low Defect FeFe(CN)6 Framework as Stable Host Material for HighPerformance Li-Ion Batteries. ACS Appl. Mater. Interfaces 2016, 8,23706−23712.(40) Wu, X.; Wu, C.; Wei, C.; Hu, L.; Qian, J.; Cao, Y.; Ai, X.;Wang, J.; Yang, H. Highly Crystallized Na2CoFe(CN)6 withSuppressed Lattice Defects as Superior Cathode Material forSodium-Ion Batteries. ACS Appl. Mater. Interfaces 2016, 8, 5393−5399.(41) Wu, X.; Luo, Y.; Sun, M.; Qian, J.; Cao, Y.; Ai, X.; Yang, H.Low-Defect Prussian Blue Nanocubes as High Capacity and Long-Life Cathodes for Aqueous Na-Ion Batteries. Nano Energy 2015, 13,117−123.(42) Kurihara, Y.; Morimoto, Y. Electrochemical, Structural, andElectronic Properties of Mn−Co Hexacyanoferrates Against LiConcentration. J. Appl. Phys. 2014, 53, 067101-1-6.

The Journal of Physical Chemistry C Article

DOI: 10.1021/acs.jpcc.8b05506J. Phys. Chem. C 2018, 122, 20602−20610

20610