Ic

AI

a

ARRAA

KSCMjP

1

ceChktrlt

m

(

h0

Electrochimica Acta 125 (2014) 404–414

Contents lists available at ScienceDirect

Electrochimica Acta

j our na l ho me pa g e: www.elsev ier .com/ locate /e lec tac ta

nvestigation on physiochemical properties of Mn substituted spinelobalt oxide for supercapacitor applications

. Nirmalesh Naveen ∗, Subramanian Selladuraionics lab, Department of Physics, Anna University, Chennai, Tamil Nadu, India 600025

r t i c l e i n f o

rticle history:eceived 22 October 2013eceived in revised form 30 January 2014ccepted 30 January 2014vailable online 12 February 2014

eywords:upercapacitorobalt oxideanganese doping

ahn-teller distortionseudocapacitance and Cyclic voltammetry.

a b s t r a c t

Investigation on physiochemical properties and electrochemical performance of doped cobalt spineloxide was carried by doping manganese ions into the cobalt oxide spinel system at various concen-trations (5% - 20%) using co-precipitation method. The influence of Mn incorporation on the structureand physical properties of the cobalt oxide were investigated using XRD, FTIR and HRSEM. It was foundthat, with Mn addition unit cell volume increases and the crystallite growth of the electrode materials washindered. Jahn teller distortion associated with Mn3+ ions has made the spinel lattice more compressible,aiding facile insertion/exertion of electrolyte ions. From SEM observations, compact agglomerates foundin pure cobalt oxide changes to loosely packed agglomerates on Mn addition. From the XPS studies, man-ganese concentration in the doped samples were identified to be close to the initial doping percentage(MnxCo3-xO4; x = 4.61, 8.25, 14.13 & 18.10%). It also reveals the preferential octahedral occupancy of Mn3+

ions in cobalt oxide spinel lattice. Electrochemical characterization of as synthesized electrode materialswas performed with cyclic voltammetry (CV), chronopotentiometry (CP) and electrochemical impedancespectroscopy (EIS). Specific current density of the electrodes increased with increasing Mn content forfixed scan rates. Specific capacitance (SC) values were calculated for the electrode materials from CV andCP. Electrode material doped with 20% Mn exhibits the highest SC of 440 Fg−1. Low equivalent series

resistance (ESR) and reduced ion diffusion resistance was observed for Mn doped electrode materials.According to the results, 10% and 20% manganese doped cobalt oxide electrode materials demonstratessuperior capacitive behavior than other prepared materials. Mn addition has improved the compoundintegrity on cycling and also increased the overall electrochemical performance of the electrode materials.

© 2014 Elsevier Ltd. All rights reserved.

. Introduction

The ever demanding energy requirements and environmentaloncerns over fossil fuels have made the search for alternativenergy resources and energy storage device materials inevitable.apacitors are devices used to store electrical energy and deliverigh power instantaneously. Electrochemical capacitors (EC’s) alsonown as supercapacitors are superior to their conventional coun-erparts and batteries in higher energy density and power density

espectively. Their large specific capacitance (1000 Fg−1), long cycleife, reduced size and unique power discharge characteristics makeshem attractive for various applications such as hybrid power

∗ Corresponding author. Tel.: +91 9962239031 Postal Address Ionics Lab, Depart-ent of physics, Anna University, Chennai, Tamil Nadu, India 600025.

E-mail addresses: nirmalesh.naveen@gmail.com, nirmal123 321@yahoo.co.inA.N. Naveen).

ttp://dx.doi.org/10.1016/j.electacta.2014.01.161013-4686/© 2014 Elsevier Ltd. All rights reserved.

sources, portable electronic devices, starting power of fuel cells, etc.[1–5]. High rate charge-discharge nature and excellent reversibil-ity of the EC’s helps in tapping the transient energy availablefrom solar cells and windmills, thereby increasing the efficiency ofenergy harvesting [6]. Based on the mechanism of charge storage,supercapacitors are classified into Electric Double Layer Capaci-tors (EDLC’s) which arises due to electrostatic charge separation atthe electrode-electrolyte interface and pseudocapacitor where thecapacitance is due to fast reversible faradaic redox process [7–9].Remarkable specific capacitance exhibited by hydrous rutheniumoxide in aqueous acidic electrolyte has shifted the focus towardstransition metal oxides. Their toxicity and unavailability forcedthe researchers to search for alternate materials that are benignto the environment and less expensive with the equivalent elec-

trochemical performance of RuO2 [10]. Among the metal oxidessuch as MnO2, Co3O4, MoO, V2O5, NiO and Fe2O3; cobalt oxideand manganese oxide were extensively studied for their superiorcapacitive behavior [11–13]. Shortcomings like poor cycle life, cost

trochi

eavstpowtcdonwiMeietbclcacotpfbaecmr

tsiTsoscdf

2

2

uaa2auwreTwfi

A.N. Naveen, S. Selladurai / Elec

ffectiveness of cobalt oxide, Mn ion dissolution into electrolytend low specific capacitance of manganese oxide can be circum-ented by a synergistic combination of both oxides. Among otherpinel systems, manganese containing mixed metal spinel sys-ems is particularly interesting as they provide the possibility forartition of different metals with more than one oxidation stateccupying any of the two sites (tetrahedral and octahedral sites),hich in turn affects the structural, electrical and magnetic proper-

ies of the material [14]. One of the difficulties regarding the crystalhemistry of spinel oxides is the determination of the metal oxi-ation states and cation distribution among the tetrahedral andctahedral sublattices sites of the spinel structure. This determi-ation becomes more complex when two or more metallic cationsith relatively similar oxidation states share the lattice sites. This

s the case of manganese-cobalt oxide systems in which bothn and Co cations adopt several oxidation states. The depend-

ncy of any particular doped composition adopting the normal,nverse or partially inverted samples spinel structure lies in latticenergy, crystal field stabilization and covalency effects. Similarlyhe site preference for any particular dopant will arise from thealance of energy considerations. It is well known that for spinelobalt oxide, high spin Co2+ ions occupies tetrahedral site andow spin diamagnetic Co3+ ions occupy octahedral sites. In thease of MnxCo3-xO4, manganese (II, III, IV) ions share tetrahedralnd octahedral sites with cobalt ions [15]. Manganese-cobalt oxideomposites are less studied compared to the investigations maden individual oxides [1]. Prasad and Miura reported that elec-rodeposited manganese-cobalt oxide shows improved capacitiveerformance than manganese-nickel oxide [16], Zhao et al. [17]ound that manganese oxide specific capacitance can be increasedy doping cobalt ions. While Chang et al. [1,18] showed that Coddition can hinder the Mn ion dissolution into electrolyte, whichnhances the reversibility and stability of cobalt-manganese oxideomposite. Change in electrochemical characteristics of cobalt-anganese oxide with respect to calcination temperature was

eported by Li et al. [19].To the best of knowledge acquired, there are no reports on elec-

rochemical evaluation of manganese doped cobalt oxide spineltructure. Replacement of one metal ion by another creates anmpact on structural and electrical properties of the host material.he aim of the present work is to study the effect of Mn addition ontructure and electrochemical properties of spinel cobalt oxide. Inrder to determine the optimum level of doping that demonstratesuperior capacitive behavior, Mn ions are introduced at variousoncentrations ranging from 5% - 20%. To study and evaluate Mnoped cobalt oxide performance as a promising electrode materialor supercapcitor application.

. Experimental

.1. Material synthesis

Manganese doped cobalt oxide nanoparticles were synthesizedsing analytical grade cobaltous acetate, manganese acetate andmmonia solution (25%). First stoichiometric proportions of cobaltcetate (2 M) and manganese acetate for various doping range (5%-0%) were dissolved in extra pure de-ionized water. Secondly 25%mmonia solution was added to the continuously agitated solutionntil the pH of the solution reached 10. After stirring, the solutionas aged for 12 h at room temperature. Precipitated compound was

emoved from the solution mixture and washed several times with

thanol, water and acetone to remove by-products and impurities.he sample was dried at room temperature to remove the excessater and grounded thoroughly using an agate mortar to obtainne powders of metal hydroxide. Samples were heat treated at

mica Acta 125 (2014) 404–414 405

400 ◦C for 4 h and the products were labeled correspondingly withrespect to their initial doping percentages as CM-05, CM-10, CM-15 and CM-20. For comparison pure cobalt oxide was synthesizedfollowing the same procedure and labeled as PCO.

2.2. Physical characterization of the samples

X-ray diffractogram of the samples were recorded by Bruker X-ray diffractometer model. D2 PHASER K� radiation of copper target

with a wavelength of 1.5416 ´̊A was used as X-ray source. Detecteddiffraction angle (2�) was scanned from 10o to 80o with a stepsize of 0.02o. FTIR spectra of the samples were recorded between400 cm−1- 4000 cm−1 wavenumber using FTIR spectrophotome-ter (Perkin Elmer - 1600). Morphology of the doped and undopedsamples was imaged using FEI Quanta FEG 200 - High Resolu-tion Scanning Electron Microscope (HRSEM). Mn concentrations incobalt spinel structure were analyzed using Energy dispersive X-raydetector. X-ray photoelectron spectra were recorded using KratosAXIS Ultra DLD X-ray photoelectron spectrometer with aluminiumanode (monochromatic K� X-rays of 1.486 eV energy) as source andoperating at 160 eV pass energy. All binding energy values werecharge-corrected to the C 1s signal which was set at 284.6 eV. XPSspectra were analyzed and fitted using CasaXPS software (version2.3.16).

2.3. Fabrication of electrode and electrochemical performancetest

Prepared electrode material, activated carbon and polyvinyli-dene fluoride (PVDF) binder were mixed together in the ratio of85:5:10 (wt. %). A slurry of the mixture was made using N-Methyl-2-pyrrolidone (NMP), which was coated onto a nickel foil currentcollector of (1 × 1 cm) 0.5 mm thickness (produced by Alfa Aesar)and dried for 4 h to remove the solvent. The mass of the loadedsamples lies within the range of 0.2 - 0.4 mg measured using a Shi-madzu analytical balance of accuracy 0.01 mg. Cyclic Voltammetry(CV), Chronopotentiometry (CP) and Electrochemical Impedancestudies (EIS) were performed using a CHI 661 C electrochemicalworkstation employing a standard 3-electrode cell configurationwith platinum wire as counter electrode and Standard CalomelElectrode (SCE) as a reference electrode. The measurements wereperformed using aqueous 3 M KOH electrolyte at ambient condi-tions.

3. Results & Discussion

3.1. XRD studies

The phase identity of the manganese doped and undoped sam-ples were determined by X-ray diffraction. Fig. 1 displays thepowder X-ray diffraction (PXRD) patterns of PCO (a), CM-05 (b),CM-10 (c), CM-15 (d) and CM-20 (e) samples. Eight obvious diffrac-tion peaks corresponding to (111), (220), (311), (222), (400), (422),(511) and (440) planes; not only the peak positions but also theirrelative intensities, were identified for the face centered cubicphase Co3O4 crystalline structure with space group FD-3 m (JointCommittee on Powder Diffraction Standards (JCPDS) file no. 78-1970). It was observed that the position of characteristic peaksof the doped samples is consistent with that of the pure cobaltoxide (a-e). 1. In the limit of instrument sensitivity, no other peakswere observed in the XRD patterns except those attributed to

Co3O4, which suggests that doped manganese ions have been wellincorporated into the Co lattice site without distorting the crys-tal symmetry. This also implies the well-dispersion of Mn into thelattice of Co3O4, and that also excluded conglomeration of MnOx,

406 A.N. Naveen, S. Selladurai / Electrochimica Acta 125 (2014) 404–414

FC

wisaclr

rMeihMsi

M

tsme

d

Ahitp

TP

ig. 1. Powder XRD patterns of (a) PCO, (b) CM-05, (c) CM-10, (d) CM-15 and (e)M-20.

hich is further confirmend from XPS studies in section 3.4. Hence,t can be concluded that there is a possible formation of MnxCo3-xO4olid solution. Variation in lattice parameter, cell volume and aver-ge particle size with respect to increase in Mn concentration werealculated and tabulated in Table 1. An increase in cell volume andattice parameter is expected for doped materials, since the ionicadii of manganese ions are relatively larger than cobalt ions (ionic

adii of Co2+- 0.65 ´̊A, Mn2+- 0.67 ´̊A, Co3+ - 0.61 ´̊A, Mn3+- 0.65 ´̊A,n4+- 0.53 ´̊A). Among the prepared materials, CM-10 has the high-

st cell volume corresponding to the high concentration of Mn3+

ons (discussed below in the XPS section 3.4) occupying the octa-edral site. Manganese ions exist in three oxidation states Mn(II),n(III) and Mn(IV) in a spinel system, as revealed from the XPS

tudies the following equilibrium among the cobalt and manganeseons proposed by Junhua Li et al. was modified here.

nocta3+/Mnocta

4++Cotetra2+↔ Mntetra

3++Coocta2+↔ Mntetra

2+

+Coocta3+

The balance between these cations would be determined byheir structural and redox properties [20]. The data from Table 1uggest that added Mn has hindered the crystallite growth of theaterial resulting in reduced average particle size. Debye–Scherrer

quation used for calculating the average crystallite size is:

= 0.9 ∗ �/ ̌ ∗ cos � (1)

Where d is the crystallite size, � is the X-ray wavelength (1.542´̊ ), � is the Bragg diffraction angle and � is the full width at the

alf maximum (FWHM) of the diffraction peak. Similar results of

ncreasing cell volume and reduced average crystallite site on addi-ion of manganese to cobalt oxide spinel system has been reportedreviously [21–23]. For sample CM-10 with 10% Mn doping showed

able 1hysical parameters of the as synthesized electrode materials.

Sample name � (FWHM) Cell parameter a = b = c ( ´̊A)

PCO 0.37488 8.06

CM-05 0.3752 8.08

CM-10 0.55332 8.19

CM-15 0.39212 8.07

CM-20 0.42433 8.07

Fig. 2. FTIR spectra of pure cobalt oxide (PCO) and Mn doped cobalt oxide samples(MnxCo3-xO4;x = 4.61, 8.25, 14.13 & 18.10%) recorded between 400-4000 cm−1.

the maximum reduction in particle size accompanied with peakbroadening in XRD pattern

Elastic strain of the samples was calculated using the formulaE = �/2 cot � [24]. An obvious decrease in elastic strain with anincrease in particle size was noticed from Table 1. Microstrain-ing in the lattice is due to the replacement of cobalt ions withMn ions causing point defects, change in bond length and bondangle that leads to lattice being more compressible. This evolutioncan be understood in terms of Jahn teller distortion observed formanganese (Mn3+) ions with mean oxidation state less than ∼ 3.5stretching the octahedra leading to higher compressibility [25–27].

3.2. Infrared (IR) studies

IR absorption spectra of pure (PCO) and doped cobalt oxide(MnxCo3-xO4; x= 4.61, 8.25, 14.13 & 18.10) are shown in Fig. 2. Twovery strong peaks at 566 cm−1 and 665 cm−1observed are char-acteristics of OB3 (B - Co3+/Mn3+ in an octahedral site) and ABO(A–Co2+ in the tetrahedral site) vibrations in the Co3O4 spinel lat-tice respectively [28]. This further confirms the formation of phasepure spinel cobalt oxide in agreement with the XRD studies. Thebroad absorption band in the region around 3419 cm−1 was dueto co-ordinated/entrapped water absorbed from moisture duringstorage process [29]. Band at 1623 cm−1 was attributed to the angu-lar deformation of absorbed water molecules. A minor peak at1023 cm−1 was due to C-O stretching vibrations. The peaks around2371 cm−1 is a characteristic of asymmetric vibration (C=O) of CO2which was absorbed from the air during heat treatment of metal

−1

oxides. A small band around 2923 cm originates from the stretch-ing vibrations of �C-H [4].

FTIR spectrum of all the doped samples has same characteris-tic peaks that of pure cobalt oxide (PCO) which depicts the similar

Cell volume ( ´̊A3) Particle size ( ´̊A) Elastic strain E

524 223 0.001092528 223 0.001090549 151 0.001610527 213 0.001140525 197 0.001233

A.N. Naveen, S. Selladurai / Electrochi

Table 2Chemical composition of the prepared oxide materials evaluated using XPS.

Element at.%

PCO CM-05 CM-10 CM-15 CM-20

Co 11.98 11.12 13.95 13.29 12.86

cgi(

3

tcedtLoacrticsEi

3

miamt&cytsba1aTat

TX

�p

Mn - 0.54 1.28 2.21 2.91O 88.02 88.34 84.77 84.50 84.23Doped - 4.61 8.25 14.13 18.10

hemical bonding nature. Interaction between incorporated man-anese ions with native cobalt ions can be inferred from the changen intensity of FTIR spectra and slight shift in peaks towards redlower frequency region) corresponding to M-O bond.

.3. Morphological studies

The HRSEM image of the materials is presented in Fig. 3. Fromhe images it can be seen that pure cobalt oxide was made of moreompact agglomerates compared to the loosely packed agglom-rates observed for doped samples as an outcome of manganeseoping. It also justifies the role played by manganese in hinderinghe crystallite growth of the compound as mentioned in section 3.1.oosely packed agglomerates means an increase in the surface areaf the doped electrode materials thereby increasing the wettingbility of the electrode materials. Significant reduction in parti-le growth was observed for 10% doping that offers higher porousegion for facile ion transport. Introduction of manganese ions inhe cobalt oxide crystal field would have caused localized changesn the electrostatic field (which in general tends to bring the particleloser); resulting in reduced particle growth. Insets of the figureshow sample morphology being retained at higher magnifications.DS data (not shown here) further confirms the presence of Mnons in the doped electrode materials.

.4. X-ray photoelectron spectroscopy (XPS)

Chemical state and composition of a material can be deter-ined quantitatively using XPS studies. Amount of manganese

ons doped were quantitatively estimated from XPS spectra forll the samples and tabulated in Table 2. It has been found thatanganese concentration in the final product is almost close to

he initial doping percentage (MnxCo3-xO4; x = 4.61, 8.25, 14.13 18.10). XPS survey spectra of cobalt oxide and 10% Mn dopedobalt oxide (CM-10) is shown in the Fig. 4 (a) & (b). Detailed anal-ses on the oxidation state of the elements were obtained fromhe Co 2p, Mn 2p, Mn 3s and O 1s high resolution spectra. Co 2ppectra of PCO seen from Fig. 4 (c), exhibits two major peaks ofinding energy 780.25 eV and 795.5 eV corresponding to Co 2p3/2nd Co 2p1/2 respectively, with a spin orbit splitting energy of5.25 eV. Additionally two shake up satellite peaks (S1 & S2) char-

cteristics of spinel phase of cobalt oxide (Co3O4) were observed.able 3 summarizes the quantitative details of Co 2p spectra forll the samples. Appearance of first satellite peak 3.5–6.5 eV abovehe Co 2p3/2 peak is the characteristics of Co2+ ions, and second

able 3PS analysis data from Co 2p spectra of all samples.

Sample Co2p �a (eV)

Co 2p1/2 (eV) Co 2p3/2 (eV)

PCO 795.5 780.25 15.25

CM-05 795.88 780.47 15.41

CM-10 794.87 779.72 15.15

CM-15 795.76 780.51 15.25

CM-20 794.95 779.77 15.18

a–spin orbit splitting, �S1b–difference in binding energy between Co 2p3/2 and first sateeak.

mica Acta 125 (2014) 404–414 407

satellite peak 9–10 eV above the Co 2p1/2 main peak is characteris-tic of Co3+ ions [15]. This further confirms the Co3O4 spinel phaseformation of cobalt oxide as determined from the XRD studies. Allthe values reported in the table are in agreement with previousliteratures [15,31, and 42]. Further Co 2p spectra of PCO is decon-voluted into two doublets corresponding to Co2+ and Co3+ ionsshown in Fig. 6 (c). For doped samples CM-10 and CM-20 Co 2ppeaks shifts slightly towards lower binding energy side. The oxida-tion state of the substituted manganese ions was determined fromresolved Mn 3s high resolution spectra shown in Fig. 5. The Mn3s core level spectra should usually show a peak splitting and adoublet due to the parallel spin coupling of the 3s electron withthe 3d electron during the photoelectron ejection. The energy sep-aration between the two peaks is related to the mean manganeseoxidation state. Since a lower valence implies more electrons in the3d orbital, more interaction can occur upon photoelectron ejection.Consequently, the energy separation between the two componentsof the Mn 3s multiplet will increase. The inverse trend will beobserved when the manganese valency increases [30]. Mean oxi-dation state calculated from the Mn 3s spectra were tabulated inTable 4. For better understanding of the oxidation state of man-ganese ions occupying tetrahedral and octahedral sites, Mn 2phigh resolution spectrum is deconconvoluted and studied. Fig. 5(d) shows the deconvoluted Mn 2p spectrum of CM-10 sample dis-playing the presence of Mn2+ and Mn3+ ions. Similar deconvolutionwhich was done for other doped samples and values were tabu-lated in Table 4. From the Mn 2p spectra three kinds of manganesespecies were identified Mn (II) at 640.62–640.76 eV, Mn(III) at641.20–641.47 eV and Mn(IV) at 642.62–642.85 eV [31] exist on thesurface of the doped materials. Information about the percentageamount of mixed oxidation state of manganese ions can be obtainedfrom area under the deconvoluted peaks. For the prepared sam-ples, with increase in manganese doping concentration, Co3+/Co2+

ratio increases which are consistent with previously reported lit-eratures [22] .According to this literature, for MnxCo3-xO4 solidsolution there may exist an electron transfer between Co2+ andMn4+. Thus, a part of Co2+ and Mn4+ may be changed into Co3+

and Mn3+, respectively. Increase in Co3+/Co2+ ratio with Mn con-centration and change in oxidation state of manganese ions from5 to 10% doping can be ascertained based on this transformation.For most of the doped samples Mn3+ ions substitute for Co3+ ionsin the octahedral site, this is in good agreement with literaturesreporting the predominant tendency of Mn3+ ion to occupy octa-hedral site in spinel systems [14,15,20]. Fig. 6 (a) & (b) shows thehigh resolution XPS spectra of O 1s for pure and CM-10 sample. O 1sspectra of pure sample exhibited a shoulder peak on the right sideand the left side, much variation in the O 1s spectra of CM-10 fromcobalt oxide can be seen from the figure. O 1s spectra was furtherdeconvoluted into three peaks, one about 529.32 eV correspondsto metal oxygen bond (M–O, M-Co, Mn), second peak at 531.41 eV

is a typical of oxygen in an OH group and strongly supports thepresence of a hydroxyl species such as CoOOH and MnOOH, andthird at 532.85 eV corresponds to surface bonded water (H–O–H)[15].

Satellite peaks �S1b �S2c

S 1 (eV) S 2 (eV)

787.46 804.97 7.21 9.47784.31 805.43 3.84 9.55787.22 804.64 7.5 9.77787.78 804.95 7.27 9.19785.15 805.18 5.38 10.23

llite peak, �S2c - difference in binding energy between Co 2p1/2 and second satellite

408 A.N. Naveen, S. Selladurai / Electrochimica Acta 125 (2014) 404–414

Fig. 3. Typical HRSEM images of (a) PCO, (b) CM-05, (c) CM-10, (d) CM-15 and (e) CM-20 (inset shows the surface morphology at higher magnification).

Table 4Determination of manganese ions oxidation state from Mn 2p and Mn 3s spectra.

Sample Mn 3s �a Mn 2p 3/2 Manganese oxidation state

Peak 1 Peak 2 Peak 1 Peak 2

Position FWHMb % Conc. Position FWHM % Conc.

CM-05 82.64 85.59 2.95 640.76 2.94 40.86 642.71 3.51 59.14 2,4CM-10 84.08 87.15 3.07 640.62 3.40 21.08 641.20 4.81 78.92 2,3CM-15 82.27 87.69 5.42 641.26 4.93 56.58 642.85 5.0 43.42 3,4CM-20 81.99 86.77 4.78 641.47 4.34 39.63 642.62 4.89 60.37 3,4

�a–spin orbit splitting for Mn 3s, FWHMb- full width at half maximum.

A.N. Naveen, S. Selladurai / Electrochimica Acta 125 (2014) 404–414 409

Fig. 4. XPS survey spectra of (a) PCO and (b) CM-10. (c) and (d) shows the resolved Co 2p spectra for pure cobalt oxide (PCO) and CM-10 samples respectively.

Fig. 5. Mn 3s spectra (a-d) of manganese doped cobalt oxide samples.

410 A.N. Naveen, S. Selladurai / Electrochimica Acta 125 (2014) 404–414

F ateriac

3

3

tF

Fi

ig. 6. (a) & (b) showing resolved O1s spectra of PCO and 10% doped cobalt oxide morresponding component peaks for CM-10 sample shown in (c) & (d).

.5. Electrochemical studies.

.5.1. Cyclic Voltammetry (CV) studiesCyclic Voltammogram (CV) is a prominent technique used

o study the capacitive behavior of the electrode materials.ig. 7 shows the CV recorded for the samples swept within the

ig. 7. Cyclic Volatammograms of Mn doped electrodes recorded at 50 mVs−1 andnset showing the CV of undoped cobalt oxide in 3 M KOH aqueous electrolyte.

l. High resolution photoelectron spectra of Co 2p and Mn 2p deconvoluted into its

potential window 0 - 0.45 V for pure cobalt oxide (PCO) and 0 -0.4 V for all other doped samples at 50 mV scan rate in 3 M KOH. CVcurves exhibiting redox peaks deviating from the ideal rectangularshape of electric double layer capacitance indicate that the majortype of charge storage mechanism is pseudocapacitive in nature.Inset of the Fig. 7 shows the CV curve of PCO at 50 mVs−1 scanrate. For cobalt oxide a large anodic peak centered at 0.35 V wasdue to the super imposition of two peaks corresponding to transi-tion between (Co (II) → Co (III) and Co (III) → Co (IV)) its oxidationstates. In general, cobalt oxide exhibits two redox couples due tothe conversion of Co3O4 to CoOOH going from +2 → +3 oxidationstate and CoOOH to CoO2 going from +3 → +4 oxidation state whichis illustrated in (2) & (3)

Co3O4 + OH− + H2O ↔ 3CoOOH + e− (2)

CoOOH + OH−↔ CoO2 + H2O + e− (3)

Transitions between these states are highly reversible [32]. But,the CV curve of the pure cobalt oxide (PCO) shows only one reduc-tion peak at 0.22 V and a small hump at 0.18 V. Reason for thehump and asymmetry in the anodic and cathodic current densitywas due to irreversibility in the transition of cobalt ions duringreduction. Change in capacitive behavior of cobalt oxide with anincrease in manganese doping concentration can be seen fromFig. 7. The current density increases with an increase in manganese

concentration except for CM-05 which is lower relative to PCO. Anexciting increase in the current density of the doped samples (CM-10, CM-15 & CM-20) can be attributed to the addition of manganeseions into the cobalt spinel lattice. Manganese substitution is highly

A.N. Naveen, S. Selladurai / Electrochimica Acta 125 (2014) 404–414 411

F PCO, (

atsmnafcrpt&

M

M

ig. 8. Shows the CV recorded for the samples between the scan rate 5-100 mV (a)

dvantageous in terms of charge storage, it has been establishedhat for manganese materials, redox reaction occurring at theurface and bulk of the material are the major charge storageechanism. Large value of current in the voltammograms are

ot associated with electric double layer charging, but are likelyttributed to the Mn(III)/Mn(II) and Mn(IV)/Mn(III) transitions. Sur-ace faradaic reaction involves surface adsorption of electrolyteations (K+) on the electrode surface. Within the bulk of the mate-ial redox reactions occur by intercalation and deintercalation ofrotons or alkali metal cations [33]. Redox reaction mechanism forhe pseudocapacitance of the MnxCo3-xO4 oxide is described by (4)

(5)

nCo2O4 + H2O + OH−→ 2CoOOH + MnOOH + e− (4)

nOOH + OH−→ MnO2 + H2O + e− (5)

b) CM-05, (c) CM-10, (d) CM-15 and (e) CM-20 in 3 M KOH aqueous electrolyte.

These chemical transformations are highly reversible [34].Cobalt oxide is considered to be less electroactive compared tomanganese oxide in terms of pseudocapacitance [11,18]. Hence,introduction of manganese ions in place of cobalt oxide hasincreased the specific capacitance of the electrode materials. In thecase of CM-05, a decrease in capacitance is observed, which maybeattributed to the majority of Mn2+ ions occupying the electrodesurface as determined from the XPS studies. Mn2+ on oxidationgoes to Mn3+ state, at low scan rates Mn3+ ions have enough timefor disproportionation yielding Mn4+ species and Mn2+ species. Ithas been previously reported that Mn2+ ions readily dissolute into

KOH electrolytes leading to formation of Mn defects and capaci-tance fades with increased immersion time [35,36]. Narrowing ofthe large anodic peak for Mn doped samples compare to PCO anddistinct pair of resolved reduction peaks for the sample CM-10 is

412 A.N. Naveen, S. Selladurai / Electrochimica Acta 125 (2014) 404–414

Table 5Quantitative measures of the electrodes capacitive performance.

Sample name Max. SC fromCV (Fg−1)

Max. SC fromCP (Fg−1)

Columbicefficiency %

PCO 343 311 74CM-05 274 217 78CM-10 394 359 81

sboarhpftsfMvdv[garcitapttdbifiwestr

S

utSwCccoWp[Boib

Fig. 9. Specific capacitance calculated for Mn doped cobalt oxide materials at var-

CM-15 436 400 76CM-20 440 404 79

een from Fig. 7. As discussed in the section 3.4 the interactionetween Co2+ and Mn4+ in a MnxCo3-xO4 system, leading to partf Co2+ and Mn4+ transforming to Co3+ and Mn3+ resulting in theppearance of additional reduction peak in the CV and increasedeversibility for the Mn doped electrode materials (x > 5%). Forigher concentration (CM-15 & CM-20), two distinct reductioneaks merges together to form a single broader peak owing toast charge transfer rate of doped manganese ions observed fromhe Fig. 7. During redox reaction, manganese ions undergo tran-ition between its trivalent and tetravalent state which is highlyast and reversible [37]. Similarly, Chang et al. have reported that

n ions substituted for cobalt in Nickel-Cobalt spinel lattice wasery sensitive to charge-transfer within the given potential win-ow compared to Co and Ni ions of same valence state and greateralence variation occurred within the selected potential window21]. Fereydoon Gobal & Sanaz Jafarzadeh had reported that man-anese oxide deposited over cobalt oxide led to increase in currentst the cost of reduced potential window [37]. Similarly, in our case,educed potential window over PCO and an increase in the specificurrent was observed for the doped electrode materials. Increasen specific current is maybe due to the manganese oxide abilityo store electrical charge by simultaneous injection of electronsnd charge-compensating cations into the solid when it is incor-orated with other transition metal oxides [10]. Fig. 8 compareshe CV’s recorded for pure and Mn doped samples swept betweenhe scan rate 5 mV- 100 mV within their respective potential win-ow. It can be seen in all the samples the shape of the CV curve haseen preserved up to a scan rate as high as 100 mV, demonstrat-

ng the excellent high rate charge transfer and fast redox kineticsrom rapid ion insertion/exertion of the electrode materials. Anncrease in potential difference between cathodic and anodic peak

as observed for all the samples, which is due to the polarization oflectrodes at high scan rates. To quantitatively evaluate the chargetorage ability of the material, Specific Capacitance (SC) of the elec-rodes were calculated from CV measurements using the formulaeported earlier [33]

C = 1v × m(Va − Vc)

∫ Vc

Va

IV dV (6)

The SC values were calculated graphically by integrating the areander the I-V curves and then dividing by the sweep rate � (V s−1),he mass of the material (m), and the potential window (Va to Vc).C calculated for all the prepared electrode materials at 5 mVs−1

ere tabulated in Table 5. It can be seen from Fig. 9 that electrodeM-20 exhibits the highest capacitive behavior with SC of 440 Fg−1

ompared to 343 Fg−1 calculated for PCO. The obtained specificapacitance is higher than those reported for nanoporous cobaltxide nanorods developed using the hydrothermal method byang et al. (281 Fg−1 at 5 mV) [38], Mn3O4 nano-octahedrons pre-

ared by Jiang et al. via hydrothermal route using EDTA (322 Fg−1)5] and mesoporous Mn-Co oxide derived from mixed oxalates by

−1

ryan et al. (383 Fg at 2 mV) [2]. In spite of the irregular morphol-gy, higher specific capacitance exhibited by the electrode materials clearly due to Mn addition and morphological changes inducedy it. It is evident from inset of the Fig. 9 that SC increases with

ious concentrations for scan rates from 5–100 mV in 3 M KOH aqueous electrolyte.Inset showing the increase in specific capacitance with increase in Mn doping con-centration.

increase in Mn doping concentration in the electrode materials.Decrease in capacitance with an increase in scan rate was observed(Fig. 9) for the materials, it is well known that at low scan rates theions have sufficient time to diffuse into the host material and utilizemaximum active species available but at high scan rates only sur-face ions participate in the redox reaction. Maximum capacitanceretention observed for CM-10 at 100 mVs−1 is attributed to thecompressibility of the cobalt spinel lattice due to Mn incorporationas discussed in section 3.1 and reduced agglomeration observedfrom HRSEM images. Jahn teller distortion exhibited by Mn3+ ionsin MnO6 octahedral making the lattice highly compressible andreduces the strain caused by ion intercalation/deintercalation dur-ing the redox process.

3.5.2. Galvanostatic charge-discharge studiesRate capability is one of the important factors for evaluating

the power applications of supercapacitors. Chronopotentiogram isa complimentary technique used to study the charge-dischargecharacteristics of the electrode materials. The constant currentgalvanostatic charge-discharge curves of the samples at a currentdensity of 1 Ag−1 was shown in Fig. 10. Charge-discharge curves canbe clearly divided into two regions (i) A linear variation of poten-tial with time (0–0.25 V) parallel to the vertical axis was due to thedouble layer capacitive behavior and (ii) a sloped variation of poten-tial vs time (0.25–0.35 V) was due to the pseudocapacitive behaviorarising from the redox activity of the species [39]. From the charge-discharge profiles, the major type of charge storage mechanism inas synthesized electrode materials were found to be pseudocapaci-tance which is consistence with the CV studies. SC’s were calculatedfrom the charge-discharge curves using the formula (7)

SC = I ∗ td/m ∗ �V (7)

Where, I is the constant current applied, td is the dischargetime, m is the mass of the material loaded and �V is the potentialdifference (0.35 V). From Table 5, SC’s calculated from chronopo-tentiogram were in agreement with those calculated from CV

technique. The Columbic efficiency of the electrode materials iscalculated using the formula

� = (td/tc) ∗ 100% (8)

A.N. Naveen, S. Selladurai / Electrochimica Acta 125 (2014) 404–414 413

Fa

tiociittdFaciictf

Fig. 11. Specific capacitance of pure cobalt oxide (PCO) and manganese doped cobalt

Fr

ig. 10. Charge-discharge curves of the prepared electrode materials compared at current density of 1 Ag−1 in 3 M KOH aqueous electrolyte

Where, � is the columbic efficiency, td is the discharge time andc is charging time from charge-discharge curves. Data from Table 5ndicates that CM-10 having the maximum columbic efficiencyf 81% as expected from its nearly symmetrical charge-dischargeurves. As discussed earlier, incorporated Mn ion aids in feasibleon and electron transport inside the bulk of the material result-ng in better reversibility and increased efficiency compared tohe pure cobalt oxide (PCO). Long charge-discharge characteris-ics of the materials are as important as any other parameters thatefine material capability for long and high power deliverance.ig. 11 shows the long cycling profile of the electrode materialst 3 Ag−1; it can be seen that specific capacitance retention ofobalt oxide has been significantly enhanced by manganese ionncorporation. Among the materials, CM-10 & CM-20 exhibits max-

mum capacitance retention; decrease in capacitance below 200ycles observed for doped samples is due to Mn ion dissolution intohe electrolyte. Previously, similar capacitance fading was reportedor manganese oxide electrode materials [2]. On cycling, ion

ig. 12. (a) Complex plane impedance plots (Nyquist plots) of all the oxides before 1000 cespective high frequency region.

oxide electrode materials as a function of cycle number recorded at a current densityof 3 Ag−1.

diffusion resistance increases with increase in number of cyclesfor PCO seen from the lower frequency region of the nyquist plotin Fig. 12 (b). Mn doped samples exhibits lower ion diffusion resis-tance and better capacitance retention than PCO, visible from thelower frequency side of the figure. This can be attributed to the mor-phological changes induced by Mn ions as observed from HRSEMimages and dissolution of surface Mn2+ ions shortening the dif-fusion length. Mn doping has also increased the compressibilityof lattices which accommodates the strain produced during inter-calation/deintercalation of electrolyte ions, thereby increasing thecompound stability over long cycling process.

3.5.3. Electrochemical Impedance StudiesInformation about kinetic features of the ions and electron in

the electrode at the electrode-electrolyte interface are attainedfrom electrochemical impedance spectra (EIS). Samples were

harge -discharge process and (b) after cycling; insets shows the impedance at their

4 trochi

s1si1hergdpatittfdspfcvmtpsfacesarototeo

4

caXtstfMfosto1dra

[[[[

[

[

[

[[

[

[

[

[

[[

[[

[[[

[

[

[[

[[[[

[

[39] K.-Q. Ding, J. Chin. Chem. Soc 55 (2008) 543.

14 A.N. Naveen, S. Selladurai / Elec

ubjected to AC impedance measurement in the range of Hz–1 MHz at their respective open circuit potentials. Fig. 12hows the complex plane impedance plots or Nyquist plots drawnmaginary part Z” against the real part Z’ for before and after000 charge-discharge process. Near absence of semicircle in theigh frequency region depicts the low internal resistance of thelectrode materials and diffusion controlled rate kinetics of theedox process. Lower internal resistance (high conductivity) is ofreat importance; less energy will be wasted during the charge -ischarge process [40]. At high frequency region, intercept at realart of the axis gives equivalent series resistance (ESR) which is anggregation of contributions from (i) discontinuity in the chargeransfer process at the solid oxide/liquid electrolyte interface (ii)ntrinsic resistance of the oxides (iii) contact resistance betweenhe active material and current collector and (iv) resistance due tohe faradaic process [41]. All the materials show very low ESR seenrom the high frequency region of the nyquist plot. The ESR valueecreases with increase in Mn content except for CM-05. It alsoupports that Mn incorporation has led to the fast charge transferrocess within the material and reduced ESR values. It is evidentrom the spectra that the decrease in capacitance of sample CM-05ompared to PCO was also due to the relatively higher resistancealue. Slope of the line at low frequency region gives a qualitativeeasurement of the resistance offered to the diffusion of ions into

he solid oxide electrode known as Warburg resistance. Nyquistlots of the samples after long charge-discharge process werehown in Fig. 12 (b) and inset shows the impedance plot at highrequency region. It can be seen that ESR value of CM-10 remainedlmost the same as before cycling while other electrodes valuehanges slightly. Among the materials, CM-10 and CM-20 showednhanced ion diffusion into the electrode material over cycling,een from the linear line (large slope value) along the imaginaryxis at low frequencies. Pure cobalt oxide (PCO) offers highestesistance to OH− ion diffusion due to the compact agglomerationf the particles observed from the SEM images. It can be concludedhat, incorporation of Mn ions has increased the electrode stabilityver long cycling and facile transport of electrolyte ions intohe electrode materials. Hence, these findings demonstrate thatlemental doping is an effective way to improve the performancef pseudo-capacitive metal oxides.

. Conclusion

In this present work we had successfully doped different con-entrations of Mn ions into the spinel cobalt oxide system withoutny secondary phase formation, confirmed from XRD, FTIR andPS studies. Mn doping has played a significant role in reducing

he crystallite size and increasing the compressibility of the latticeites. SEM images showed that the compact agglomerates of PCOurned to loosely packed crystallites on Mn doping, making it easieror ion diffusion. XPS analysis revealed the presence of Mn(II),

n(III) & Mn(IV) species in the spinel cobalt oxide system. Resultsrom electrochemical studies revealed the enhanced performancef the Mn doped samples over PCO. On Mn addition current den-ity, reversibility, columbic efficiency and capacitance retention ofhe electrodes has been increased. Overall capacitive performancef the cobalt oxide electrode has been lifted by Mn incorporation.

0% and 20% manganese doped cobalt oxide electrode materialsemonstrates superior electrochemical performance. The aboveesults indicate that Mn doped cobalt oxide electrode materialsre promising electrode material for supercapcitor applications.

[[[

mica Acta 125 (2014) 404–414

Acknowledgement

Financial support from Anna University by providing Anna Cen-tenary Research Fellowship (ACRF) for A. Nirmalesh Naveen wasgreatly appreciated (Lr.No.CR/ACRF/2013/37)

References

[1] J.-K. Chang, W.-C. Hseih, W.-T. Tsai, J. Alloys and Compounds 461 (2008) 667.[2] Yong Yan, Bingcai Wu, Cuihong Zheng, Daolai Fang, Materials Sciences and

Applications 3 (2012) 377.[3] M.-S. Song, Kyung Min Lee, Yu Ri Lee, In Young Kim, Tae Woo Kim, Jayavant

Laxman Gunjakar, S.-J. Hwang, J. Phys. Chem. C 114 (2010) 22134.[4] X.-H. Xia, J.-P. Tu, Y.-Q. Zhang, Y.-J. Mai, X.-L. Wang, C-D. Gu, X.-B. Zhao, RSC

Adv 2 (2012) 1835.[5] Hao Jiang, Ting Zhao, Chaoyi Yan, Jan Ma, Chunzhong Li, Nanoscale 2 (2010)

2195.[6] Linrui Hou, Changzhou Yuan, Long Yang, Laifa Shen, Fang Zhang, Xiaogang

Zhang, RSC Adv 1 (2011) 1521.[7] Y.-C. Chen, Y.-K. Hsu, Y.-G. Lin, Y.-K. Lin, Y.-Y. Horng, L.-C. Chen, K.-H. Chen,

Electrochimica Acta 56 (2011) 7124.[8] Jie Zhang, Dong Shu, Tianren Zhang, Hongyu Chen, Haimin Zha, Yongsheng

Wang, Zhenjie Sun, Shaoqing Tang, Xueming Fang, Xiufang Cao, J. of Alloys andCompounds 532 (2012) 1.

[9] Yu Il Yoon, Jang Myoun Ko, Int. J. Electrochem. Sci 3 (2008) 1340.10] J.-M. Luo, Bo Gao, X.-G. Zhang, Materials Research Bulletin 43 (2008) 1119.11] Banafsheh Babakhani, G. Douglas, Ivey, Electrochimica Acta 56 (2011) 4753.12] H.-Y. Wu, H.-W. Wang, Int. J. Electrochem. Sci 7 (2012) 4405.13] Dewei Wang, Yuqi Li, Qihua Wang, Tingmei Wang, J Solid State Electrochem 16

(2012) 2095.14] A. Paul, Wright, Srinivasan Natarajan, M. John, Thomas, Chem. Mater 4 (1992)

1053.15] J.L. Gautier, E. Rios, M. Gracia, J.F. Marco, J.R. Gancedo, Thin Solid Films 311

(1997) 51.16] Kalakodimi Rajendra Prasad, Norio Miura, Electrochemistry Communications

6 (2004) 1004.17] G.-Y. Zhao, C.-L. Xu, H.-L. Li, J. Power Sources 163 (2007) 1132.18] J.-K. Chang, M.-T. Lee, C.-H. Huang, W.-T. Tsai, Materials Chemistry and Physics

108 (2008) 124.19] Qiang Li, K.-X. Li, J.-Y. Gu, Hui Fan, J. Physics and Chemistry of Solids 69 (2008)

1733.20] Junhua Li, Xi Liang, Schicheng Xu, Jiming Hao, Applied Catalysis B; Environ-

mental 90 (2009) 307.21] S.-K. Chang, K.-T. Lee, Zulkarnain Zainal, K.-B. Tan, Nor Azah Yusof, Wan

Mohamad Daud Wan Yusoff, J.-F. Lee, N.-L. Wu, Electrochimica Acta 67 (2012)67.

22] Qinghong Zhang, Xianhong Liu, Wenqing Fan, Ye Wang, Applied catalysis B;Environmental 102 (2011) 207.

23] Banafsheh Babakhani, G. Douglas, Ivey, Electrochimica Acta 56 (2011) 4753.24] C.R. Indulal, G.S. Kumar, A.V. Vaidyan, R. Raveendran, J. Nano- Electron. Phys. 3

(2011) 170.25] C. Popescu, L. Pinsard-Gaudart, N. Dragoe, J. Appl. Phys. 112 (2012) 053503.26] Naoufal Bahlawane, Peter Antony Premkumar, Janine Feldmann, Katharina

Kohsé-Hiönghaus, Chem. Vap. Deposition 13 (2007) 118.27] J.M. Paulsen, J.R. Dahn, Solid State Ionics 126 (1999) 3.28] Mustafa Aghazadeh, J. Appl Electrochem 42 (2012) 89.29] Lijing Xie, Kaixi Li, Guohua Sun, Zhongai Hu, Chunxiang Lv, Jianlong Wang,

Changming Zhang, J Solid State Electrochem 17 (2013) 55.30] Mathieu Toupin, Thierry Brousse, Daniel Bélanger, Chem. Mater 16 (2004)

3184.31] X.-H. Xia, J.-P. Tu, X.-L. Wang, C.-D. Gu, X.-B. Zhao, Chem. Commun 47 (2011)

5786.32] Xingyi Wang, Qian Kang, Dao Li, Catalysis Communications 9 (2008) 2158.33] Banafsheh Babakhani, G. Douglas, Ivey, J. Electrochimica Acta 55 (2010)

4014.34] Jamie Gomez, E. Egwu, Kalu, J. Power Sources 230 (2013) 218.35] Dongfang Yang, J. Power Sources 198 (2012) 416.36] Fereydoon Gobal, Sanaz Jafarzadeh, J Solid State Electrochem 16 (2012) 1561.37] Guoxiu Wang, Xiaoping Shen, Josip Horvat, Bei Wang, Hao Liu, David Wexler,

Jane Yao, J. Phys. Chem. C 113 (2009) 4357.38] Fang Zhang, Liang Hao, Luojiang Zhang, Xiaogang Zhang, Int. J. Electrochem. Sci

6 (2011) 2943.

40] G. Sumanta Kumar Meher, Ranga Rao, J. Phys. Chem. C 115 (2011) 25543.41] G. Sumanta Kumar Meher, Ranga Rao, J. Phys. Chem. C 115 (2011) 15646.42] Beng Jit Tan, J. Kenneth, Klabunde, M.A. Peter, Sherwood, J. Am. Chem. Soc 113

(1991) 855.