Nonstoichiometry-Induced Enhancement of ElectrochemicalCapacitance in Anodic TiO2 Nanotubes with Controlled PoreDiameterV. C. Anitha,† Arghya Narayan Banerjee,*,† G. R. Dillip,† Sang Woo Joo,*,† and Bong Ki Min‡

†School of Mechanical Engineering, Yeungnam University, Gyeongsan 712-749, South Korea‡Center for Research Facilities, Yeungnam University, Gyeongsan 712-749, South Korea

*S Supporting Information

ABSTRACT: We report the fabrication of self-organized titania(TiO2) nanotubes (TNTs) with controlled pore diameters (140−20nm) by anodization for the application of electrochemical capacitorelectrodes. The areal capacitances obtained for 140 nm TNTs as0.23/0.13 mF cm−2 at a scan rate of 1/5 mV s−1 and it is enhancedto 5.5/2.9 mF cm−2 (at the same scan rates) by controlling the porediameter to 20 nm. In this study, role of pore diameter in thecapacitance behavior of TNTs is explained on the basis of effectivesurface area and presence of oxygen vacancies/titanium interstitials.With a decrease in the pore diameter, the surface area-to-volumeratio (and hence, active surface sites) increases, which leads togreater dissociation of Ti4+ into Ti3+ under high temperatureannealing and thus brings more nonstoichiometric defects like Ti3+

interstitials and oxygen deficiency within the lower dimensionalTNTs. This manifests higher charge conductivity and greater electrochemical performance of TNTs with lower diameters. Thesimplicity of anodization method and the excellent electrochemical properties make these vertical TNTs as an alternativecandidate for use in energy storage applications.

1. INTRODUCTION

Supercapacitors are being increasingly explored as a feasiblecharge storage technology in recent years.1−6 On the basis ofthe mechanism of charge storage, supercapacitors can becategorized into three general groups: (i) nonfaradic super-capacitors (electric double layer capacitors-EDLCs)7 that arebased on electrostatic charge diffusion and accumulation at theelectrode/electrolyte interface, (ii) faradic supercapacitors(pseudocapacitors)7,8 that are dominated by faradaic reactionson electrode materials,9,10 and (iii) hybrid supercapacitors.11

Additionally, the specific capacitance for both storagemechanisms can be enhanced by using a material with a highspecific surface area, such as nanostructured conductingpolymer and metal oxides porous or carbonaceous materi-als.12−14 Thus, understanding the surface and interfacecharacteristics is critical for improving the performance ofsupercapacitors.Nanostructured materials have significant role in the field of

electrochemical capacitors due to the combination of nanoscalefeatures with a highly defined geometry and high surface area.Over the past decade, TiO2 has been considered as asupercapacitor electrode material because of its semiconductingproperties and chemical stability.15−19 Self-organized titania(TiO2) nanotubes (TNTs) have been explored for use asbinder-free supercapacitor electrodes.18,20 This nanostructured

TNT surfaces have been fabricated using the optimizedelectrochemical anodization process using metallic titanium(Ti) foil as the substrate in fluoride containing electro-lytes.19−21 Electrochemical anodization method offers suitablyback-contacted nanotube layers on the substrate, which can beemployed directly as an electrochemical device.15,16 Ideally, Tifoil is used directly as a current collector, which provides directand uninterrupted charge transport pathways, while TiO2

nanotubes provide the active area for charge storage activity.19

The TNT structures also form surface electrical fields andreduce recombination by confining the injected electrons to thecentral zone of the tubes which is observed in dye-sensitizedsolar cell (DSSCs) applications.19,22 Vertically oriented TNTshave attracted much attention in charging storage systemsbecause of their capacity to offer high surface area and greatlyimproved electron transfer pathways in comparison tononoriented structures, which favor higher charge propagationin active materials.19−21,23

Generally, titania capacitors resembles to conventionalEDLCs, which act by a nonfaradic mechanism with a verylow specific capacitance of 10−40 μF cm−2 in the process of

Received: February 3, 2016Revised: April 20, 2016Published: April 25, 2016

Article

pubs.acs.org/JPCC

© 2016 American Chemical Society 9569 DOI: 10.1021/acs.jpcc.6b01171J. Phys. Chem. C 2016, 120, 9569−9580

charge− discharge. It has been typically suggested that TiO2only contributes a very low nonfaradic capacitance and almostno faradic capacitance in terms of the mechanism ofcharging.24,25 Recently, many research studies have beenconducted to improve the electrical conductivity of TNTs bya thermal treatment for inducing surface defects such as oxygenvacancies and reduction of Ti4+ to Ti3+.19,23,26 Also by changingthe crystallinity (anatase- rutile),23 doping with nonmetals,27

hydrogenated processes28 and by making nanocomposites withelectroactive materials has been reported to achieve theimproved capacitive behavior. Annealing is a simple methodto get better electrical conductivity and improvement of theelectrochemical capacitance.27 Luet al.29 reported largestspecific capacitance of 3.24 mF cm−2 at a scan rate of 100mV s−1 by hydrogenation of TNTs. The hydrogenated TiO2electrodes were obtained by calcination of anodic TiO2nanotube arrays in hydrogen atmosphere at high temperatures.It is widely believed that bare titania, due to high electricresistance and low specific surface area, shows low electro-chemical capacitance. However, it has been reported thatdecreasing the particle size of TiO2 to less than 10 nm results inpseudocapacitive behavior of the resultant material, allowing foran increase in the capacitance of anatase of up to 90−120 μFcm−2.23,30

It is well-known that high surface area, interconnectivity andhigh conductivity of the active materials will enhance thecapacitive properties.31,32 Generally, the specific capacitancedepends not only on the specific surface area of the materials,but also on other parameters, such as the pore size, pore densityetc.33,34 This article reports the effect of pore diameter of TNTson the electrochemical capacitance under various scan rates.Controlled pore diameters (20−140 nm) are fabricated viaoptimizing the anodization parameters which includes theanodization voltage, time, ratio of ethylene glycol (EG) and HFelectrolyte. In energy storage applications such as batteries andsupercapacitors, electrochemical properties of titania particlesmostly depend on the size of the particles rather than thephases. However, in this study, crystallinity of TNT’s isanalyzed by X-ray photoelectron spectroscopy (XPS), Ramanspectra, etc. The results of XPS and Raman revealed thatpresence of Ti3+ and oxygen vacancies depending upon thediameter of TNTs. The effect of pore diameter and surface area(from AFM results) on capacitance is also studied. To the bestof authors’ knowledge, there are no reports in the literatureregarding the variation of capacitance with pore diameter ofTNTs with a detailed explanation using weight fraction ofanatase/effective surface area/titanium interstitials-oxygenvacancies. Additionally, a possible mechanism for capacitance

Figure 1. SEM images of different pore diameters of TiO2 nanotubes (a) 140 nm, (b) 80 nm, (c) 35 nm, and (d) 20 nm. Insets show the lateralviews of the corresponding images. A slightly magnified version of part d is represented as black circular portions.

The Journal of Physical Chemistry C Article

DOI: 10.1021/acs.jpcc.6b01171J. Phys. Chem. C 2016, 120, 9569−9580

9570

is provided based on the pore size, surface area of TNTs, andelectrolyte ion size.

2. EXPERIMENTAL SECTION

2.1. Materials. Titanium plate (99% purity, 2 mm thick,Sigma-Aldrich), platinum plate (99% purity, 2 mm thick,Sigma-Aldrich), hydrofluoric acid (48%, Sigma-Aldrich), ethyl-ene glycol (99.8% purity, Sigma-Aldrich), acetone (95%,DUKSAN pure chemicals co. Ltd., Korea),and a dc powersupply (EP-5001, PNCYS, Korea) were used for the samplepreparations. Deionized (DI) water with a resistivity of 18.2MΩ-cm was used throughout the experiment.2.2. Fabrication of TiO2 Nanotubes (TNTs). The

electrochemical anodization was carried out on commerciallypurchased titanium plates (14 mm2 area) acting as the anode,with platinum plate as the cathode connected to a regulated dcpower supply (PNCYS, EP-5001). The electrolyte system usedfor the study was 0.5 wt % hydrofluoric acid (HF) containingethylene glycol (EG). Here, different volume ratios (1:1, 1:5,1:10) of 0.5 wt % HF:EG have been used at two differentanodization voltages (viz. 20 and 40 V) to tune the porediameter (∼140−20 nm) of TNTs.35,36 Among the differentanodization conditions systematic change in pore diameterswith good uniformity were selected to perform the electro-chemical capacitance studies. Generally, as-synthesized TiO2 isamorphous, which confines its application in various fields.37,38

So, to improve the crystallinity and capacitance, the as preparedTNTs were air-annealed at 600 °C for 1 h in a conventionalbox furnace (Lind Berg Blue M 10 kA, 120/24 V).2.3. Characterization of TNTs and Device Perform-

ance. Surface morphology and dimensional characteristics ofTNTs were carried out using Field Emission Scanning ElectronMicroscopy (FE-SEM, HITACHI, S-4800, Japan). Crystallinityof the TNT samples was analyzed using X-ray diffraction(XRD, PAnalytical Xpert Pro, USA). Raman spectroscopy(HORIBA Scientific, Xplora plus-France) measurements werecarried out using Raman spectroscopy with laser excitation of532 nm. X-ray photoelectron spectroscopy (XPS) measure-ments were performed using a spectrometer (K-ALPHA,Thermo Scientific, U.K.) with a pass energy of 30.00 eV.Atomic force microscopy (AFM) imaging and surface rough-ness/specific surface area measurements were carried out usingDigital Instruments (Nanoscope IIIa, USA) Scanning ProbeMicroscope Controller.Electrochemical measurements were carried out in 1 M

aqueous potassium chloride (KCl) solution. A typical three-electrode experimental cell equipped with TNTS as theworking electrode, a platinum plate as the counter electrode,and an Ag/AgCl (KCl saturated) as the reference electrode formeasuring the electrochemical properties of working electrode.An electrochemical workstation (CHI 760E, CH Instruments,Inc.) consisting of cyclic voltammetry (CV) and electro-chemical impedance spectroscopy (EIS) was employed to study

the specific capacitance/electrical resistance of the electrodes.CV was conducted over a voltage range of −0.8 to +0.1 V atvarious scan rates (1, 5, 10, 20, 50, 100, 200, and 500 mV s−1)for all TNT samples. EIS measurements were performedbetween 100 kHz and 1 Hz under a constant potential of −0.2V using a 5 mV rms sinusoidal modulation.

3. RESULTS AND DISCUSSION

3.1. SEM Analysis. Figure 1 represents the SEM images ofdifferent pore diameters of TiO2 nanotube surfaces preparedunder optimized conditions of anodization (as shown in Table1). The insets are showing the lateral view images ofcorresponding surfaces. All the dimensional measurementswere carried out using SEM micrographs and are shown inTable 1. Figure S1 represents the pore diameter sizedistribution of TNTs. The nanotubular length of all sampleswas ∼1.3 μm by adjusting the anodization time.24,36,39 Bytuning the anodization parameters, 20, 35, 80, and 140 nmaverage pore diameters of TNTs are formed and termed asTNT-20, TNT-35, TNT-80, and TNT-140, respectively.

3.2. XRD Measurements. Figure 2(a) represents the XRDpattern of different pore diameters (TNT-20, TNT-35, TNT-80, and TNT-140 nm) of TNTs which is air-annealed at 600°C for 1 h. Annealing leads to the formation of a mixture ofanatase and rutile phases. The relative weight percentage ofanatase phase in TNT samples were also calculated from theSpurr and Myers equation, eq 1:40,41

= +f I I1/[1 1.265( / )]R A (1)

Here, f is the weight fraction of anatase, while IR and IA denotethe rutile (110) and anatase (101) reflection intensities. Thecalculated percentage of anatase phase for different TNTs fromeqn 1 is plotted in Figure 2b. The 20 nm diameter TNTsshowed ∼48% of anatase weight fraction and with increasingpore diameter of TNTs, the anatase weight fraction increases as∼54%, ∼62%, and ∼66% for 35, 80, and 140 pore diameters ofTNTs, respectively. Annealing or calcination is the widely usedprocess for anatase−rutile transformation and also to generatedefects in TiO2 phases.

42 It has been reported that with gradualincrease in the annealing temperature from 400 °C to morethan 1000 °C, the titania phase gradually transforms from apure anatase phase to a pure rutile phase via an intermediateanatase−rutile mixed phase.43 This anatase-to-rutile phasetransformation occurs due to a nucleation and growth process,during which rutile nuclei form at the surface (grain boundary)of the anatase crystallites and (or) within the grains44−47ofanatase and grow in size by consuming the surrounding anatase,which is mainly due to the fact that rutile is more thermallystable than anatase.48 Hence it becomes apparent that, in eithercase, higher surface-to-volume ratio provides larger nucleationsites for rutile, and therefore, nanostructured titania produceshigher rutile phase at lower annealing temperature than thebulk counterparts.49 In the current case, 20 nm TNT depicts

Table 1. Tabulation of Anodization Conditions, Pore Diameter, Tube Density and the Corresponding Sample Codes

electrolyte systemHF to EG volume

ratioanodization voltage

(V)anodization time

(h)inner pore diameter

(nm)tube density

(μm‑2)samplecode

mixture of (0.5-wt %) HF andEG

1:1 40 6 140 ± 3 150 ± 4 TNT-140

1:1 20 11 80 ± 4 582 ± 4 TNT-801:10 40 12 35 ± 3 1123 ± 3 TNT-351:5 20 14 20 ± 4 2202 ± 4 TNT-20

The Journal of Physical Chemistry C Article

DOI: 10.1021/acs.jpcc.6b01171J. Phys. Chem. C 2016, 120, 9569−9580

9571

higher active surface area (detailed in section 3.5), andgradually decreases with increasing nanotube diameter. There-fore, 20 nm TNT provides maximum nucleation sites for rutile,and hence, under identical annealing temperature, 20 nm TNTshows highest rutile fraction (∼52%), and gradually decreasesto ∼34% for 140 nm TNT. Also, as stated earlier, annealing/calcination of titania results in the formation of some structuraldefects, such as the partial reduction of Ti4+ to Ti3+ and oxygenvacancies inside the TiO2 phase,

50 as discussed below3.3. Raman Characterizations. Phase changes of all

annealed samples were further analyzed by Raman spectrosco-py (Figure 3a). The Raman bands located at ∼156 (Eg), ∼211(Eg), ∼405 (B1g), ∼520 (A1g + B1g), and 637 cm−1 (Eg) areattributed to anatase peaks and ∼237 (B1g), ∼447 (Eg + A1g),and ∼610 cm−1 (A1g) to peaks for rutile phase23,38

(corresponding Raman-active modes are given in thebrackets).51,52 The strongest Eg mode at ∼156 cm−1 arisingfrom the external vibration of the anatase structure is wellresolved, which indicates that an anatase phase was formed inthe annealed TNTs and the long-range order was fairly formed.These results are consistent with the XRD spectra as shown inFigure 2a. Comparing the Raman spectra of four different porediameters of TNTs, it becomes apparent that the Raman bandsshift with respect to a change in the pore diameter. Forexample, the lowest Eg mode of anatase at ∼156 cm−1 of TNT-

140 gradually shifts toward ∼161 cm−1 in TNT-20 sample. Alsothe line width (full-width-at-half-maxima, fwhm) increases andintensity decreases accordingly (as shown in Figure 3b). Thecorresponding quantitative values are plotted in Figure S2(a) ofthe Supporting Information. The change in the position andshape of the Raman peaks is attributed to several factors likenonstoichiometric defect, induced disorder via minor phases,pressure, phonon confinement, strain, nonhomogeneity of thesize distribution etc.53,54 Among these factors, phononconfinement due to grain size effect and stress are argued tobe least effective for the Raman line shift of heat-treatednanostructured TiO2.

45 In any case, for nanotubular structures(as in the current case), grain size effect is always leastinfluential. Also the size-strain analysis from the XRD datadepicts negligible strain and crystallite size variations withinTNTs with different pore diameters. Therefore, the observedchange in the line position and shape of the Raman-activelowest anatase Eg peak (at 156 cm

−1) is considered to be due tooxygen nonstoichiometry, as XPS analyses clearly reveal anincrease in the oxygen deficiency as a decreasing function of thenanotube diameter (discussed in section 3.4). That means thatTNT-20 has the highest oxygen deficiency and also shows themaximum blue-shift of the line position, maximum incrementin the line width (fwhm), and minimum line intensity. As theoxygen nonstoichiometry decreases according to the followingtrend: TNT-20 > TNT-35 > TNT-80 > TNT-140 (discussedlater in details), the blue-shift of the line position and the linewidth also decrease according to the above trend (cf. FigureS2a) and as expected, the line intensity varies in the reverse

Figure 2. (a) XRD patterns of TNTs/140−20 nm pore diametersamples. (b) Calculated weight fraction of the anatase phase in TNTs/140−20 nm pore diameters from XRD patterns.

Figure 3. (a) Raman spectra of annealed TNTs with different porediameters and (b) Raman shift of the anatase Raman peak located at156 cm−1.

The Journal of Physical Chemistry C Article

DOI: 10.1021/acs.jpcc.6b01171J. Phys. Chem. C 2016, 120, 9569−9580

9572

trend (cf. Figure 3b). On the other hand, the rutile 447 cm−1

peak is blue-shifted with increase in the pore diameter (i.e., withdecrease in the nonstoichiometry) (cf. Figure S2b). The trendin the line width of this peak is not analyzed mainly because ofthe asymmetric nature of this peak to produce unreliable data.

All the above trends are consistent with the literature,45

indicating the signature of increasing oxygen deficiency as adecreasing function of the pore diameter of the TNTs. Thistrend is apparently because of higher active surface area in thelower dimensional TNTs (as shown later in the AFM

Figure 4. Ti 2p XPS spectra of TNTs (a) 20, (b) 35, (c) 80, and (d) 140 nm pore diameters.

Figure 5. O 1s XPS spectra of TNTs (a) 20, (b) 35, (c) 80, and (d) 140 nm pore diameters.

The Journal of Physical Chemistry C Article

DOI: 10.1021/acs.jpcc.6b01171J. Phys. Chem. C 2016, 120, 9569−9580

9573

Analyses), which tends to induce more oxygen vacancies via thereduction of Ti4+ to Ti3+ at the surface of the titania nanotubes.Also the considerable difference in the lowest Raman Eg peakintensity between the standard value (∼147 cm−1)52 and thecurrent samples (∼156−161 cm−1) is probably due to thebreakdown of long-range translational crystal symmetrymanifested by the nonstoichiometric defects within the as-synthesized TNTs.55

3.4. XPS Analysis. X-ray photoelectron spectroscopy(XPS) is used to analyze the Ti 2p and O 1s valenceenvironment in TNTs with pore diameters. Figure 4 and 5represents the high resolution XPS spectra of Ti 2p and O 1sphotoelectrons. All the binding energies (BEs) obtained in theXPS analysis are calibrated using the C 1s peak at bindingenergy 284.5 eV as the reference. Deconvolution and curvefitting of spectra were performed using CasaXPS Gaussian−Lorentzian peak shape. The BEs, FWHMs, and relativepercentage (area under the curves) of all the peaks arefurnished in Table 2. The BEs of Ti 2p located at ∼463.1 and∼457.4 are assigned to Ti4+ 2p1/2 and Ti4+ 2p3/2 states ofphotoelectrons.38 The energy difference between these twopeaks is around 5.7−5.8 eV, which is consistent with the valuesreported in the literature for Ti (IV).41 After deconvolution,two sub peaks appeared at BEs ∼458.7 and ∼462.7 eV, whichare assigned to Ti3+ 2p3/2 and Ti3+ 2p1/2, respectively.

38 Thepresence of this lesser oxidation states of Ti ions indicates theformation of higher percentage of loosely bound Ti3+−O bondsagainst stronger Ti4+ − O bonds, which enhances the chargetransfer process between anions and cations and, thus improvesthe electrochemical performance of TNTs.56 Also this lessoxidized Ti3+ states manifest higher oxygen deficiencies withinthe TNTs, as corroborated by Raman measurements shownearlier. A closer look of the XPS data reveals that thepercentage of Ti3+ states against the total Ti states (Ti4+ + Ti3+)increases with the decrease in the pore diameter, as shown in

Figure S3. This is because of the higher surface-to-volume ratio(shown later in AFM analyses) of the lower-dimensional TNTsto manifest higher reduction of Ti4+ to Ti3+ at the surface of thetitania nanotubes.50

Figure 5 represents the XPS spectra of O 1s states withdifferent pore diameters. The binding energy at ∼530.5 eV canbe assigned to the oxygen bound to Ti4+ ions in TNTs and theoxygen anions (O2−) are in the lattice (Ti−O−Ti).28,57 Thepeak located at ∼531.5 eV is attributed to oxygen vacancies.57

Generally, for an ideal fully oxidized, defect-free TiO2 surface,all the surface Ti cations are supposed to be in the +4 oxidationstate and having a 5-fold coordination to the oxygen anions.Heating the surface to high temperatures induces desorption ofsurface oxygen, producing oxygen vacancies.42,50 Apparently,low-dimensional nanostructures with higher surface area-to-volume ratio would induce more oxygen deficiency within thenanomaterial, and due to this reason, the 20 nm TNTs in thecurrent case showed highest percentage of O-deficiency(27.4%) and gradually reduced with increase in the porediameter under identical annealing temperature (cf. Figure S3).A schematic diagram is presented in Figure S4, which depictsthe effect of higher active surface area in lower dimensionalTNTs to induce higher oxygen vacancies and Ti3+ interstitials.Similarly, as stated earlier (detailed in section 3.2), 20 nm TNTprovides maximum nucleation sites for rutile, and hence, underidentical annealing temperature, 20 nm TNT shows highestrutile fraction (∼52%), and gradually decreases to ∼34% for140 nm TNT. The oxygen vacancies created inside the TNTstructures will help to increase the capacitance due to increasein the TiO2 conductivity via an increase in the carrierconcentration23 according to the following defect equilibria:52

Table 2. Various XPS Data of O and Ti Species within Different TNT Samples

sample

TNT-20 TNT-35 TNT-80 TNT-140

electronic level Ti 2p Ti+4 2p3/2 BE (eV) 457.4 457.5 457.4 457.5fwhm (eV) 0.84 0.85 0.85 0.85area (%)a 64.5 64.5 64.1 64.5

Ti+4 2p1/2 BE (eV) 463.1 463.2 463.4 463.2fwhm (eV) 1.75 1.76 1.77 1.76area (%)a 35.5 35.5 35.9 35.5

Ti+3 2p3/2 BE (eV) 458.7 458.7 458.6 458.7fwhm (eV) 0.92 0.83 0.98 0.83area (%)a 5.4 5.3 6.0 5.3

Ti+3 2p1/2 BE (eV) 462.7 462.6 462.8 462.5fwhm (eV) 1.14 1.10 0.86 1.03area (%)a 9.0 6.0 5.1 2.0

O 1s Ti−O BE (eV) 530.5 530.5 530.6 530.5fwhm (eV) 0.86 0.83 0.9 0.9area (%)b 27.1 29.9 38.8 39.5

oxygen deficiency BE (eV) 531.5 531.5 531.5 531.5fwhm (eV) 1.49 1.46 1.46 1.46area (%)b 27.4 26.3 22.0 19.3

Ti−OH BE (eV) 532.2 532.2 532.8 532.1fwhm (eV) 2.98 2.97 2.76 3.05area (%)b 45.5 43.7 41.9 38.4

O/Ti ratio 1.44 1.52 1.82 1.89aWith respect to the entire Ti 2p curve. bWith respect to the entire O 1s curve.

The Journal of Physical Chemistry C Article

DOI: 10.1021/acs.jpcc.6b01171J. Phys. Chem. C 2016, 120, 9569−9580

9574

↔ + +

+ ↔ + +

+ −

+ −

⎫⎬⎪⎭⎪

O g

g

O V 2e12

( )

2O Ti Ti 3e O ( )

X

Xi

0 02

2

0 Ti3

2 (2)

where O0, V0 , TiTi, Tii, and e denote lattice oxygen, oxygenvacancy, lattice Ti, interstitial Ti, and electron, respectively.Superscripts X, − and + denote effective neutral, negative, andpositive charge states, respectively. Because of thehighest O-deficiency being within TNT-20 sample, it shows betterelectrochemical properties against other samples, as shown insection 3.6.The O 1s BEs at ∼532.2 eV are assigned to surface adsorbed

hydroxyl or water molecules from atmospheric contaminationsduring sample preparations and handling.41 The highpercentage of Ti−OH concentration (∼38−45%, cf. Table 2)clearly depicts the higher OH−-terminated surfaces in thecurrent samples, which also increase with the decrease in thepore diameter, apparently because of the higher surfaceadsorption sites in the lower dimensional TNTs. Because ofthe higher percentage of this surface “titanol” group, thehydrophilicity of the samples also increases, which in turnincreases the electrochemical performance under aqueouselectrolyte.58 Finally, the elemental composition of the samples,obtained from the XPS survey spectra, clearly shows higher

oxygen deficiency at lower dimensional TNTs. For example,O/Ti ratio for TNT-140 (1.89, closer to the stoichiometricvalue of 2) decreases gradually to 1.44 (more nonstoichiom-etry) for TNT-20, thus corroborating the Raman and high-resolution XPS measurements.

3.5. AFM Analysis. The AFM images of TNT surfaces arerepresented in Figure 6. The observed surface roughness andcalculated relative percentage of increase in effective surfacearea with respect to 140 nm pore diameter sample (TNT-140)are tabulated in Table 3. TNTs of 20 nm pore diameter showmaximum (∼88.24%) of effective surface area as well asmaximum surface roughness (∼18 nm), both of which produce

Figure 6. AFM micrographs of TNTs with different pore diameters (a) 140, (b) 80, (c) 35, and (d) 20 nm pore diameters.

Table 3. Surface Roughness, Relative Increase of EffectiveSurface Area (with Respect to) 140 nm of TNTs and SpecificCapacitance of Different TNT Samples

porediameter of

TNTs(nm)

surfaceroughnessRa (nm)

relative increase of effectivesurface area with respect to

TNT-140 (%)

specific capacitanceat 1 mV/s‑1 scanrate (mF cm‑2)

140 ± 3 10 ± 1 0 0.2380 ± 4 13 ± 2 49.55 0.6835 ± 3 16 ± 2 62.15 1.1420 ± 4 18 ± 1 88.24 5.51

The Journal of Physical Chemistry C Article

DOI: 10.1021/acs.jpcc.6b01171J. Phys. Chem. C 2016, 120, 9569−9580

9575

highest active surface sites for improved electrochemicalperformance. It is well-known that high surface area of theactive materials will enhance the capacitive properties.31

However, only the surface of the pores that the ions canaccess will contribute to the double layer capacitance of theEDLCs.59 Since the electrical energy stored in EDLCs are givenby the separation of charged species in the electrical doublelayer across the electrode/electrolyte interface, the performanceof the EDLCs depend strongly on the electrolyte and thespecific surface area of the electrode materials.3.6. Electrochemical Performance. In general, the cyclic

voltammetry (CV) method is used to characterize thecapacitive behavior of the electrode materials. Cyclicvoltammetry tests of TNT samples were carried out atpotentials between −0.8 to +0.1 V in 1 M aqueous electrolyteof KCl solution. The CV responses of the different porediameters of TNT (140, 80, 35, 20) samples at a scan rate of 1,5, 10, 20, 50, 100, 200, and 500 mV s−1 are shown in Figure7a−d. It is noted that the shape of the CV curves arerectangular-like pattern and they have significantly changedwith the increase of scan rate, even up to 500 mV s−1. There isno redox peaks observed, therefore the nanotubes showelectrical double layer capacitance behavior.31,60 In all casesthe CV shape is not distorted which indicates that the TNTlayer is strongly adhered to Ti surface and no oxide layer

breakdown occurred at the specified scan rates.61 In particular,the CV for TNT-20 sample shown in Figure 7d shows asymmetrical and nearly rectangular shape for all scan rates from50 to 500 mV s−1. This indicates that the process is highlyreversible and the nanotubes capacitance behavior is based onthe electrical double layer principle.62,63

The specific capacitance (Cs) of the electrodes was calculatedfrom eqn 323,64

∫ν= −C A V V I V V1/ . ( ) ( ) dSV

V

2 11

2

(3)

where Cs is the specific capacitance measured in the potentialrange of (V2−V1), v is the scan rate (mV s−1), A is the area(cm2) of the working electrode and I(V) is the responsecurrent. The calculated specific capacitance of TNT samples areplotted against the scan rate and shown in Figure 8. The Cs forpore diameters of 140, 80, 35, and 20 nm at a scan rate of 1 mVs−1 is 0.23, 0.68, 1.14, and 5.51 mF cm−2 (Table 3),respectively. The specific capacitance value decreases withincreasing scan rate due to the diffusion limitations in thenanotube channels.38,60The highest Cs value of 5.51 mF cm−2 isobtained for 20 nm diameter TNT sample at a scan rate of 1mV s−1. The cyclic voltammograms reveals a significant increasein capacitance with decreasing the pore diameter of TNTs. TheCs result clearly shows the effect of nanotubular pore dimeters

Figure 7. Overlaid CV curves in 1 M KCl at different scan rates in (a) 140 nm, (b) 80 nm, (c) 35 nm, and (d) 20 nm pore diameters of TNTs.

The Journal of Physical Chemistry C Article

DOI: 10.1021/acs.jpcc.6b01171J. Phys. Chem. C 2016, 120, 9569−9580

9576

on the enhancement of specific capacitance of the TNTs. Thevertically oriented TNTs prepared by anodization showedenhanced capacitance properties by providing a direct pathwayfor electron transfer38 and this TNTs which were directlygrown on conductive Ti metallic substrate provides anintrinsically high surface area of these nanostructured materialsand ensures the exploitation of electrode materials with higherspecific capacitance. Also, annealing at 600 °C helps to generate

Ti3+ interstitials and oxygen vacancies which further leads tohigher charge/ionic transfer (TNT-80, 35 20 nm), andsubsequently results in improved capacitance values.Electrochemical double layer capacitors (EDLCs) store

energy at the electrolyte/electrode interface through reversibleion adsorption onto the electrode surface, thus charging the so-called “double layer capacitance”; no faradaic (redox) reactionis involved in the charge storage mechanism.65 Figure 9a showsa schematic diagram of the electric double layer structureshowing the arrangement of solvated anions and cations nearthe electrode/electrolyte interface. Figure 9b shows the electriccircuit representation of an electric double layer capacitance,where Φsurf(t) is the time-dependent surface potential of theelectrode, Relec is the electrode resistance, Cstern is the Stern layerand Cdif f is the diffuse layer specific capacitances in series,respectively.62 Solvated cations can migrate and are adsorbed tothe electrode surface due to electrostatic forces. The Sternlayer/compact layer contains immobile ions which are stronglyadsorbed to the electrode surface. There are no free chargeswithin the Stern layer, whereas the diffuse layer is situatedoutside the Stern layer where ions are mobile under thecoupled influence of electrostatic forces and diffusion. Since thecompact layers of ions residing on adjacent electrode wallscontribute considerably to the capacitance,66 therefore, theeffective surface area of the electrode and solvated electrolyte’s

Figure 8. Capacitance of TNTs with different pore diameters.

Figure 9. (a) Schematic diagram of an electrical double layer. (b) Equivalent circuit representation of the electrical double layer in series with theelectrode. (c) Surface roughness. (d) Effective surface area. (e) Capacitance variations with respect to the TNT pore diameters. The figure is adaptedwith permission from ref 62.

The Journal of Physical Chemistry C Article

DOI: 10.1021/acs.jpcc.6b01171J. Phys. Chem. C 2016, 120, 9569−9580

9577

ion size become the key parameters to control the overallspecific capacitance.66,67 Although the size of solvated ion (K+,Cl−) is on the angstrom level68 which is substantially smallerthan the pore sizes of TNTs in the present study, the imposedcompact ion layers from adjacent pore walls and the effectivesurface area of the TNT electrode affects the double-layerformation and plays a significant role in the capacitanceenhancement. Also, the roughness of the surface will providethe active surface sites which further increase the capacitanceper unit electrode area, as the capacitance is directlyproportional to the surface area and inversely proportional tothe charge separation. Parts c−e of Figure 9 represent thevariation of surface roughness, effective surface area, andcapacitance (at 1 mV s−1) with pore diameter of TNTs basedon the results furnished in Table 3. Evidently, the effectivesurface area and the surface roughness increase with thedecrease in the pore diameter, which leads to the increase in thespecific capacitance as a decreasing function of the TNT porediameters. Therefore, the better electrochemical performanceshown by the 20 nm TNT samples over the other TNTelectrodes is the combined effect of several physicochemicalproperties which include (i) high oxygen vacancy that induceshigh carrier concentration and charge conductivity, (ii) highinterstitial Ti3+ ions that creates more loosely bound Ti3+ − Obonds for better charge transfer, (iii) high surface “titanol”groups to increase the hydrophilicity of the sample surface toenhance interfacial properties for better electrochemicalperformance under aqueous electrolyte, (iv) high surfaceroughness that provides large active surface sites and competeswith the Debye length69 to favorably control the capacitanceper unit electrode area,70 and (v) high effective surface area thattends to closely align the electrochemical double layer parallelto the surface at every point to enhance the overall capacitance.3.6.1. EIS Measurements. Electrochemical impedance spec-

troscopy (EIS) is a powerful technique and widely used tostudy porous electrodes. It gives an information about theinternal resistance of the electrode material and the resistancebetween the electrode and the electrolyte.71 Figure 10compares the Nyquist plots of TNT samples with porediameters 140−20 nm. The high frequency regions of thespectra are shown as the inset. The high frequency regioncorresponds to the charge transfer limiting process and is

ascribed to the double-layer capacitance (Cdl) in parallel withthe charge transfer resistance (Rct) at the contact interfacestructure between electrode surface and electrolyte solution.72

The linear part in the low frequency region is related to theWarburg resistance (diffusive resistance) of the electrolyte intothe interior of the electrode surface and an ion diffusion/transport into the electrode surface.64 A slight semicircular arcwas observed in TNT-140 and TNT-80, indicating a chargetransfer limiting process, which is usually the result of a parallelcombination of internal resistance and capacitance.23 However,no such semicircular arc was observed for the samples TNT-35and TNT-20, which corresponds to good charge transfer of theworking electrode and indicates that there is no electricalresistance. These results reveal that the TNT samples with aparticular heating regime results in improved electricalconductivity of TNT and therefore high charge propagationin TNT.

4. CONCLUSIONSTiO2 nanotubes (TNTs) with different pore diameters (140,80, 35, and 20 nm) are fabricated on the surfaces of metallic Tiby controlling the electrochemical anodization parameters. Thesamples are annealed @ 600 °C, 1h and used as anelectrochemical supercapacitor electrode. The capacitancevalue increases with decreasing pore diameters (140−20 nm)and maximum aerial capacitances of 5.5 mF cm−2 at a scan rateof 1 mV s−1 is obtained in 20 nm TNTs. The XPS analysisshows that presence of titanium interstitials (Ti3+) and oxygenvacancies increases with the decrease in the pore diameter, andmanifested by the higher surface area-to-volume ratio of thelower dimensional TNTs. With a decrease in the pore diameter,the surface area-to-volume ratio (and hence, active surfacesites) increases, which leads to higher temperature-induceddesorption of surface oxygen to produce higher oxygenvacancies and greater dissociation of Ti4+ into Ti3+ underhigh temperature annealing. Hence 20 nm TNTs showedmaximum nonstoichiometric defects like Ti3+ interstitials andoxygen vacancies under annealing. This establishes highercharge conductivity/electrochemical performance of TNTswith smaller diameters. Raman spectra analysis revealed thatthe values of lowest anatase Eg and rutile peaks shift withrespect to a change in the TNT pore diameter, which isattributed to the induction of higher oxygen vacancies in thelower dimensional TNTs, which further helps to improve theelectrochemical capacitance in this sample. The AFM resultscorroborated this by revealing that the effective surface areaincreases with decreasing pore diameter and 20 nm TNTsshowed a maximum effective surface area (∼88.24% increase,with respect to the 140 nm TNTs) and highest average surfaceroughness (∼18 nm), which provided an improved capacitancevalues in 20 nm TNTs compared 140, 80, and 35 nm TNTs.These results demonstrated that vertically aligned TNTs havemany advantages such as high surface area, increased number ofdelocalized carriers and an improved charge transport proper-ties compared to the spherical nanoparticles. Therefore, it canbe concluded that the high surface area and highly activesurface site will enhance the ion diffusion, charge transfer andcapacitance values.

■ ASSOCIATED CONTENT*S Supporting InformationThe Supporting Information is available free of charge on theACS Publications website at DOI: 10.1021/acs.jpcc.6b01171.

Figure 10. Nyquist plots of TNT samples with different porediameters (140−20 nm), with the inset showing an enlargement of thehigh frequency region.

The Journal of Physical Chemistry C Article

DOI: 10.1021/acs.jpcc.6b01171J. Phys. Chem. C 2016, 120, 9569−9580

9578

Variation of nanotube’s pore diameter size distribution,shift of Raman peak at anatase and rutile peak with porediameter of titania nanotubes, variation of Ti3+/Ti ratio,oxygen vacancy, and O/Ti ratio, and schematic diagramdepicting the desorption of surface oxygen underannealing in higher and smaller pore diameter ofnanotubes (PDF)

■ AUTHOR INFORMATION

Corresponding Authors*E-mail: arghya@ynu.ac.kr; banerjee_arghya@hotmail.com.Telephone: +82-53-810-2453; Fax: +82-53-810-2062 (A.N.B.).*E-mail: swjoo@yu.ac.kr. Telephone: +82-53-810-3239. Fax:+82-53-810-2062 (S.W.J.).

NotesThe authors declare no competing financial interest.

■ ACKNOWLEDGMENTS

This work is funded by NRF-2015-002423 of the NationalResearch Foundation of Korea.

■ REFERENCES(1) Miller, J. R.; Simon, P. Electrochemical Capacitors for EnergyManagement. Science 2008, 321, 651−653.(2) Liu, D.; Garcia, B. B.; Zhang, Q.; Guo, Q.; Zhang, Y.; Sepehri, S.;Cao, G. Mesoporous Hydrous Manganese Dioxide Nanowall Arrayswith Large Lithium Ion Energy Storage Capacities. Adv. Funct. Mater.2009, 19, 1015−1023.(3) Yu, G.; Hu, L.; Vosgueritchian, M.; Wang, H.; Xie, X.;McDonough, J. R.; Cui, X.; Cui, Y.; Bao, Z. Solution-ProcessedGraphene/MnO2 Nanostructured Textiles for High-PerformanceElectrochemical Capacitors. Nano Lett. 2011, 11, 2905−2911.(4) Kuila, T.; Mishra, A. K.; Khanra, P.; Kim, N. H.; Lee, J. H. RecentAdvances in the Efficient Reduction of Graphene Oxide and ItsApplication as Energy Storage Electrode Materials. Nanoscale 2013, 5,52−71.(5) Chen, L.; Sun, L. J.; Luan, F.; Liang, Y.; Li, Y.; Liu, X. X. Synthesisand Pseudocapacitive Studies of Composite Films of Polyaniline andManganese Oxide Nanoparticles. J. Power Sources 2010, 195, 3742−3747.(6) Bae, J.; Song, M. K.; Park, Y. J.; Kim, J. M.; Liu, M.; Wang, Z. L.Fiber Supercapacitors Made of Nanowire-Fiber Hybrid Structures forWearable/flexible Energy Storage. Angew. Chem., Int. Ed. 2011, 50,1683−1687.(7) Simon, P.; Gogotsi, Y. Materials for Electrochemical Capacitors.Nat. Mater. 2008, 7, 845−854.(8) Ke, Y.-F.; Tsai, D.-S.; Huang, Y.-S. Electrochemical Capacitors ofRuO2 Nanophase Grown on LiNbO3(100) and sapphire(0001)Substrates. J. Mater. Chem. 2005, 15, 2122−2127.(9) Hu, L.; Choi, J. W.; Yang, Y.; Jeong, S.; La Mantia, F.; Cui, L.-F.;Cui, Y. Highly Conductive Paper for Energy-Storage Devices. Proc.Natl. Acad. Sci. U. S. A. 2009, 106, 21490−21494.(10) Lang, X.; Hirata, A.; Fujita, T.; Chen, M. Nanoporous Metal/oxide Hybrid Electrodes for Electrochemical Supercapacitors. Nat.Nanotechnol. 2011, 6, 232−236.(11) Halper, M.; Ellenbogen, J. Supercapacitors: A Brief Overview;MITRE Nanosystems Group: McLean, VA, 2006.(12) An, K. H.; Jeon, K. K.; Heo, J. K.; Lim, S. C.; Bae, D. J.; Lee, Y.H. High-Capacitance Supercapacitor Using a NanocompositeElectrode of Single-Walled Carbon Nanotube and Polypyrrole. J.Electrochem. Soc. 2002, 149, A1058−A1062.(13) Rajendra Prasad, K.; Miura, N. Electrochemically SynthesizedMnO2-Based Mixed Oxides for High Performance Redox Super-capacitors. Electrochem. Commun. 2004, 6, 1004−1008.

(14) Frackowiak, F.; Beguin, F. Electrochemical Storage of Energy inCarbon Nanotubes and Nanostructured Carbons. Carbon 2002, 40,1775−1787.(15) Jennings, J. R.; Ghicov, a; Peter, L. M.; Schmuki, P.; Walker, a B.Dye-Sensitized Solar Cells Based on Oriented TiO2 Nanotube Arrays:Transport, Trapping, and Transfer of Electrons. J. Am. Chem. Soc.2008, 130, 13364−13372.(16) Stergiopoulos, T.; Ghicov, a; Likodimos, V.; Tsoukleris, D. S.;Kunze, J.; Schmuki, P.; Falaras, P. Dye-Sensitized Solar Cells Based onThick Highly Ordered TiO2 Nanotubes Produced by ControlledAnodic Oxidation in Non-Aqueous Electrolytic Media. Nanotechnology2008, 19, 235602.(17) Mohapatra, S. K.; Raja, K. S.; Mahajan, V. K.; Misra, M. EfficientPhotoelectrolysis of Water Using TiO2 Nanotube Arrays byMinimizing Recombination Losses with Organic Additives. J. Phys.Chem. C 2008, 112, 11007−11012.(18) Macak, J. M.; Schmidt-Stein, F.; Schmuki, P. Efficient OxygenReduction on Layers of Ordered TiO2 Nanotubes Loaded with AuNanoparticles. Electrochem. Commun. 2007, 9, 1783−1787.(19) Fabregat-Santiago, F.; Randriamahazaka, H.; Zaban, A.; Garcia-Canadas, J.; Garcia-Belmonte, G.; Bisquert, J. Chemical Capacitance ofNanoporous-Nanocrystalline TiO2 in a Room Temperature IonicLiquid. Phys. Chem. Chem. Phys. 2006, 8, 1827−1833.(20) Bo, G.; Xiaogang, Z.; Changzhou, Y.; Juan, L.; Long, Y.Amorphous Ru1-yCryO2 Loaded on TiO2 Nanotubes for Electro-chemical Capacitors. Electrochim. Acta 2006, 52, 1028−1032.(21) Fang, H.-T.; Liu, M.; Wang, D.-W.; Sun, T.; Guan, D.-S.; Li, F.;Zhou, J.; Sham, T.-K.; Cheng, H.-M. Comparison of the RateCapability of Nanostructured Amorphous and Anatase TiO2 forLithium Insertion Using Anodic TiO2 Nanotube Arrays. Nano-technology 2009, 20, 225701.(22) Chang, K. H.; Hu, C. C. Hydrothermal Synthesis of Binary Ru-Ti Oxides with Excellent Performances for Supercapacitors. Electro-chim. Acta 2006, 52, 1749−1757.(23) Salari, M.; Konstantinov, K.; Liu, H. K. Enhancement of theCapacitance in TiO2 Nanotubes through Controlled Introduction ofOxygen Vacancies. J. Mater. Chem. 2011, 21, 5128−5133.(24) Xie, Y.; Zhou, L.; Huang, C.; Huang, H.; Lu, J. Fabrication ofNickel Oxide-Embedded Titania Nanotube Array for RedoxCapacitance Application. Electrochim. Acta 2008, 53, 3643−3649.(25) Conway, B. E.; Pell, W. G. Double-Layer and PseudocapacitanceTypes of Electrochemical Capacitors and Their Applications to theDevelopment of Hybrid Devices. J. Solid State Electrochem. 2003, 7,637−644.(26) Yong-gang, W.; Xiao-gang, Z. Preparation and ElectrochemicalCapacitance of RuO2/TiO2 Nanotubes Composites. Electrochim. Acta2004, 49, 1957−1962.(27) Endut, Z.; Hamdi, M.; Basirun, W. J. Supercapacitance ofBamboo-Type Anodic Titania Nanotube Arrays. Surf. Coat. Technol.2013, 215, 75−78.(28) Wu, H.; Xu, C.; Xu, J.; Lu, L.; Fan, Z.; Chen, X.; Song, Y.; Li, D.Enhanced Supercapacitance in Anodic TiO2 Nanotube Films byHydrogen Plasma Treatment. Nanotechnology 2013, 24, 455401.(29) Lu, X.; Wang, G.; Zhai, T.; Yu, M.; Gan, J.; Tong, Y.; Li, Y.Hydrogenated TiO2 Nanotube Arrays for Supercapacitors. Nano Lett.2012, 12, 1690−1696.(30) Wang, J.; Polleux, J.; Lim, J.; Dunn, B. PseudocapacitiveContributions to Electrochemical Energy Storage in TiO2 (Anatase)Nanoparticles. J. Phys. Chem. C 2007, 111, 14925−14931.(31) Xu, B.; Wu, F.; Chen, R.; Cao, G.; Chen, S.; Zhou, Z.; Yang, Y.Highly Mesoporous and High Surface Area Carbon: A HighCapacitance Electrode Material for EDLCs with Various Electrolytes.Electrochem. Commun. 2008, 10, 795−797.(32) Salari, M.; Aboutalebi, S. H.; Konstantinov, K.; Liu, H. K. AHighly Ordered Titania Nanotube Array as a SupercapacitorElectrode. Phys. Chem. Chem. Phys. 2011, 13, 5038−5041.(33) Fabregat-santiago, F.; Barea, E. M.; Bisquert, J.; Mor, G. K.;Shankar, K.; Grimes, C. a. High Carrier Density and Capacitance in

The Journal of Physical Chemistry C Article

DOI: 10.1021/acs.jpcc.6b01171J. Phys. Chem. C 2016, 120, 9569−9580

9579

TiO2 Nanotube Arrays Induced by Electrochemical Doping. J. Am.Chem. Soc. 2008, 130, 11312−11316.(34) Yan, X.; Tong, X.; Wang, J.; Gong, C.; Zhang, M.; Liang, L.Construction of Three-Dimensional Porous Nano-Ni/NiO NanoflakeComposite Film for Electrochemical Energy Storage. Mater. Lett.2013, 106, 250−253.(35) Anitha, V. C.; Menon, D.; Nair, S. V.; Prasanth, R.Electrochemical Tuning of Titania Nanotube Morphology in InhibitorElectrolytes. Electrochim. Acta 2010, 55, 3703−3713.(36) Anitha, V. C.; Lee, J.-H.; Lee, J.; Narayan Banerjee, A.; Woo Joo,S.; Ki Min, B. Biofilm Formation on a TiO2 Nanotube with ControlledPore Diameter and Surface Wettability. Nanotechnology 2015, 26,065102.(37) Hesabi, Z. R.; Allam, N. K.; Dahmen, K.; Garmestani, H.; El-Sayed, M. A. Self-Standing Crystalline TiO2 nanotubes/CNTsHeterojunction Membrane: Synthesis and Characterization. ACSAppl. Mater. Interfaces 2011, 3, 952−955.(38) Salari, M.; Aboutalebi, S. H.; Chidembo, A. T.; Nevirkovets, I.P.; Konstantinov, K.; Liu, H. K. Enhancement of the ElectrochemicalCapacitance of TiO2 Nanotube Arrays through Controlled PhaseTransformation of Anatase to Rutile. Phys. Chem. Chem. Phys. 2012,14, 4770−4779.(39) Anitha, V. C.; Banerjee, A. N.; Joo, S. W.; Min, B. K. Fabricationof Hierarchical Porous Anodized Titania Nano-Network withEnhanced Active Surface Area: Ruthenium-Based Dye AdsorptionStudies for Dye-Sensitized Solar Cell (DSSC) Application. J. Ind. Eng.Chem. 2015, 29, 227−237.(40) Spurr, R. a.; Myers, H. Quantitative Analysis of Anatase-RutileMixtures with an X-Ray Diffractometer. Anal. Chem. 1957, 29, 760−762.(41) De los Santos, D. M.; Navas, J.; Aguilar, T.; Sanchez-Coronilla,A.; Fernandez-Lorenzo, C.; Alcantara, R.; Pinero, J. C.; Blanco, G.;Martín-Calleja, J. Tm-Doped TiO2 and Tm2Ti2O7 PyrochloreNanoparticles: Enhancing the Photocatalytic Activity of Rutile with aPyrochlore Phase. Beilstein J. Nanotechnol. 2015, 6, 605−616.(42) Lu, G.; Linsebigler, a; Yates, J. T. Ti3+ Defect Sites onTiO2(110): Production and Chemical Detection of Active Sites. J.Phys. Chem. 1994, 98, 11733−11738.(43) Nakaruk, a.; Ragazzon, D.; Sorrell, C. C. Anatase-RutileTransformation through High-Temperature Annealing of TitaniaFilms Produced by Ultrasonic Spray Pyrolysis. Thin Solid Films2010, 518, 3735−3742.(44) Li, W.; Ni, C.; Lin, H.; Huang, C. P.; Shah, S. I. SizeDependence of Thermal Stability of TiO2 Nanoparticles. J. Appl. Phys.2004, 96, 6663−6668.(45) Gouma, P. I.; Mills, M. J. Anatase-to-Rutile Transformation inTitania Powders. J. Am. Ceram. Soc. 2001, 84, 619−622.(46) Zhang, W. F.; He, Y. L.; Zhang, M. S.; Yin, Z.; Chen, Q. RamanScattering Study on Anatase TiO2 Nanocrystals. J. Phys. D: Appl. Phys.2000, 33, 912−916.(47) Zhang, J.; Li, M.; Feng, Z.; Chen, J.; Li, C. UV RamanSpectroscopic Study on TiO2. I. Phase Transformation at the Surfaceand in the Bulk. J. Phys. Chem. B 2006, 110, 927−935.(48) Banerjee, A. N. The Design, Fabrication, and PhotocatalyticUtility of Nanostructured Semiconductors: Focus on TiO2-BasedNanostructures. Nanotechnol., Sci. Appl. 2011, 4, 35−65.(49) Choudhury, B.; Choudhury, A. Local Structure Modificationand Phase Transformation of TiO2 Nanoparticles Initiated by OxygenDefects, Grain Size, and Annealing Temperature. Int. Nano Lett. 2013,3, 55.(50) Xiong, L.-B.; Li, J- L.; Yang, B.; Yu, Y. Ti3+ in the Surface ofTitanium Dioxide: Generation, Properties and PhotocatalyticApplication. J. Nanomater. 2012, 2012, 831524.(51) Choi, H. C.; Jung, Y. M.; Kim, S. Bin. Size Effects in the RamanSpectra of TiO2 Nanoparticles. Vib. Spectrosc. 2005, 37, 33−38.(52) Anitha, V. C.; Banerjee, A. N.; Joo, S. W. Recent Developmentsin TiO2 as N- and P-Type Transparent Semiconductors: Synthesis,Modification, Properties, and Energy-Related Applications. J. Mater.Sci. 2015, 50, 7495−7536.

(53) Ohsaka, T.; Yamaoka; Shimomura, S. Effect of HydrostaticPressure on the Raman Spectrum of Anatase (TiO2). Solid StateCommun. 1979, 30, 345−347.(54) Scepanovic, M. J.; Grujic-Brojcin, M.; Dohcevic-Mitrovic, Z. D.;Popovic, Z. V. Characterization of Anatase TiO2 Nanopowder byVariable-Temperature Raman Spectroscopy. Sci. Sinter. 2009, 41, 67−73.(55) Pal, M.; Pal, U.; Jimenez, J. M. G. Y.; Perez-Rodríguez, F. Effectsof Crystallization and Dopant Concentration on the EmissionBehavior of TiO2:Eu Nanophosphors. Nanoscale Res. Lett. 2012, 7, 1.(56) Banerjee, A.; Joo, S.; Min, B. Photocatalytic Degradation ofOrganic Dye by Sol-Gel-Derived Gallium-Doped Anatase TitaniumOxide Nanoparticles for Environmental Remediation. J. Nanomater.2012, 2012, 1−14.(57) Chu, D.; Younis, A.; Li, S. Direct Growth of TiO2 Nanotubes onTransparent Substrates and Their Resistive Switching Characteristics.J. Phys. D: Appl. Phys. 2012, 45, 355306.(58) Rahimi, P.; Ghourchian, H.; Sajjadi, S. Effect of Hydrophilicityof Room Temperature Ionic Liquids on the Electrochemical andElectrocatalytic Behaviour of Choline Oxidase. Analyst 2012, 137,471−475−1865.(59) Xu, B.; Wu, F.; Chen, R.; Cao, G.; Chen, S.; Wang, G.; Yang, Y.Room Temperature Molten Salt as Electrolyte for Carbon Nanotube-Based Electric Double Layer Capacitors. J. Power Sources 2006, 158,773−778.(60) Aboutalebi, S. H.; Chidembo, A. T.; Salari, M.; Konstantinov,K.; Wexler, D.; Liu, H. K.; Dou, S. X. Comparison of GO, GO/MWCNTs Composite and MWCNTs as Potential Electrode Materialsfor Supercapacitors. Energy Environ. Sci. 2011, 4, 1855−1865.(61) Endut, Z.; Hamdi, M.; Basirun, W. J. An Investigation onFormation and Electrochemical Capacitance of Anodized TitaniaNanotubes. Appl. Surf. Sci. 2013, 280, 962−966.(62) Wang, H.; Pilon, L. Physical Interpretation of CyclicVoltammetry for Measuring Electric Double Layer Capacitances.Electrochim. Acta 2012, 64, 130−139.(63) Arulepp, M.; Permann, L.; Leis, J.; Perkson, a.; Rumma, K.;Janes, a.; Lust, E. Influence of the Solvent Properties on theCharacteristics of a Double Layer Capacitor. J. Power Sources 2004,133, 320−328.(64) Ramadoss, A.; Kim, S. J. Vertically Aligned TiO2 NanorodArrays for Electrochemical Supercapacitor. J. Alloys Compd. 2013, 561,262−267.(65) Simon, P.; Burke, A. A. Nanostructured Carbons: Double-LayerCapacitance and More. Electrochem. Soc. Interface 2008, 17, 38−43.(66) Chmiola, J.; Yushin, G.; Gogotsi, Y.; Portet, C.; Taberna, P. L.Anomalous Increase in Carbon Capacitance at Pore Sizes Less Than 1Nanometer. Science 2006, 313, 1760−1763.(67) Faraji, S.; Ani, F. N. The Development Supercapacitor fromActivated Carbon by Electroless platingA Review. RenewableSustainable Energy Rev. 2015, 42, 823−834.(68) Salitra, G.; Soffer, A.; Eliad, L.; Cohen, Y.; Aurbach, D. CarbonElectrodes for Double-Layer Capacitors I. Relations Between Ion andPore Dimensions. J. Electrochem. Soc. 2000, 147, 2486−2493.(69) Thomas, G. M. Does Controlled Roughening of a SurfaceIncrease Its Capacitance? Thesis. Worcester Polytechnic Institute2012.(70) Daikhin, L.; Kornyshev, A. A.; Urbakh, M. Double-LayerCapacitance on a Rough Metal Surface. Phys. Rev. E: Stat. Phys.,Plasmas, Fluids, Relat. Interdiscip. Top. 1996, 53, 6192−6199.(71) Carrara, S.; Bavastrello, V.; Ricci, D.; Stura, E.; Nicolini, C.Improved Nanocomposite Materials for Biosensor ApplicationsInvestigated by Electrochemical Impedance Spectroscopy. Sens.Actuators, B 2005, 109, 221−226.(72) Liu, K.; Hu, Z.; Xue, R.; Zhang, J.; Zhu, J. Electropolymerizationof High Stable poly(3,4-Ethylenedioxythiophene) in Ionic Liquids andIts Potential Applications in Electrochemical Capacitor. J. PowerSources 2008, 179, 858−862.

The Journal of Physical Chemistry C Article

DOI: 10.1021/acs.jpcc.6b01171J. Phys. Chem. C 2016, 120, 9569−9580

9580