Evolution of TiO2 Nanotubular Morphology Obtained in Ethylene Glycol/Glycerol Mixture and its Photoelectrochemical Performance

The evolution of TiO2 nanotubular morphology, synthesized in a mixture of fluorinated ethylene glycol and glycerol electrolyte, was studied as a function of the anodization time. The samples were characterized by FEG-SEM, XRD, XPS, UV-Vis and EIS. The formation of single-or double-walled TiO2 nanotube structure can be efficiently controlled by the anodization time. For anodization times less than 30 minutes, a compact oxide layer is formed, followed by double-walled nanotube formation up to 120 minutes and single-walled nanotubes up to 240 minutes. XPS analyses show that the samples obtained with short anodization time have a high carbon content and oxygenated surface species compared to the longer-time anodized sample; however, binding energy peaks for Ti 2p remained invariant. The performances of TiO2 nanotubular photoelectrodes were evaluated in photoelectrochemical water splitting where TiO2 nanotubes anodized for 120 minutes presented the best performance that was related to their optimal morphology and charge transportation.


Introduction
Nano-architectured TiO 2 semiconductors have been extensively studied in the last few decades mainly due to their advantageous properties such as high adsorption ability, excellent photoelectrochemical activities and large specific surface area [1][2][3] .TiO 2 -based materials are widely applied in sensors [4][5][6][7][8][9] , pollutant degradation 10 , hydrogen generation 11 , bio-applications 12,13 , and photocatalysis [14][15][16][17][18] .Owing to their one-dimensional nature, TiO 2 nanotubes offer a high surface area (1000 m 2 g -1 ) with electron diffusion higher than that of similar systems based on nanoparticles 19 .Electrochemical anodization in fluorinated electrolytes is one of the most promising and effective methods employed to obtain highly ordered nanotube arrays of TiO 2 and many other materials with controllable morphologies [20][21][22] .Optimizing anodic film topography and its properties are some of the main interests for smart material application 23 .
By controlling parameters such as anodization potential [24][25][26] anodization time 27,28 , temperature 22,29 , as well as the electrolyte composition 30 and its pH value 31 , different nanotubular structures can be obtained.The first and second generations of electrolytes were based on aqueous electrolytes containing hydrofluoric acid and fluoride salts, respectively 32,33 .In the third generation, non-aqueous, organic solvents containing fluoride ions were used; this allowed controllable TiO 2 nanotubes to be obtained due to the high viscosity of organic electrolytes 34 .The use of viscous electrolytes allows wall smoothness by damping local spikes in the flux of reaction species by locally different pH values within the tubes.However, viscous electrolytes result in to lower current density, local acidification and amount of competing chemical dissolution 35 .
The main studied organic electrolytes consist of ethylene glycol 28,34,36 and glycerol [35][36][37] .The latter is used due to its high viscosity that influences the diffusion of ionic species, the kinetics of nanotube formation and its morphology; therefore it does not show current fluctuations during anodization.However, the chemical dissolution of TiO 2 in glycerol is slow 35,37 .
Roman et al. 36 studied the influence of the applied potential and anodization time on the nanotube diameter and length.They found that the diameter tends to increase with the water content increasing in glycerol-based electrolytes 36 .An ethylene-glycol-based electrolyte presented a higher viscosity than that of a glycerol-based one (13.1 and 11.8 cP, respectively) 35 .High viscosity solvents decrease the mobility of ionic species; thereby, decreasing the growth rate.However, anodization in these solvents helps decreasing the current fluctuations; therefore, smoother nanotube walls are formed.In case of less viscous electrolytes, the growth rate is high, but higher chemical dissolution rate of TiO 2 causes the formation of ripples on the tube walls 35,39 .These results warrant to use ethylene-glycol/glycerol mixture (where one contributes to increase the growth rate while other to obtain smooth walled nanotubes) and to investigate their effect on the evolution of nanotubular morphology.Nischk et al. 35 showed that for the same anodization potential and time, the walls of TiO 2 nanotubes obtained in a glycerol-based electrolyte are rough; those obtained in ethylene-glycolbased electrolyte are smooth.Additionally, in glycerol-based electrolytes, the outer diameter was wider and the length of the nanotubes was shorter than that of the nanotubes obtained in the ethylene-glycol-based electrolyte.
Albu et al. 38 demonstrated the formation of double-walled TiO 2 nanotubes in ethylene glycol electrolyte at a voltage of 120 V, showing the well-defined inner and outer walls in the tubular morphology.Macak et al. 39 reviewed different anodization processes for Ti-based nano-architectured morphologies.However, the focus of these studies lay on the synthesis of nanotubes rather than their photoelectrochemical performance.Despite the fact that TiO 2 nanotubes are broadly synthesized by anodization in glycerol or ethylene glycol, few studies report TiO 2 nanotubes obtained by anodization in glycerol mixed with ethylene glycol 36 .Therefore, in order to obtain an optimized nanotubular morphology, it is interesting to study the evolution of TiO 2 nanotubes in glycerol mixed with ethylene glycol solution, and to optimize them for improved photoelectrochemical applications.
The aim of this study is to identify the nano-structures formed during the anodization procedure.The effects of anodization time on the morphology, crystal structure and surface properties of the prepared nanotubes were investigated with the final aim of applying them in photoelectrochemical water splitting.

Experimental
Nanotubes were grown on Ti foil grade 2 -ASTM-F67 40 (99.8 -purity: wt.%, 0.7 mm thickness, Realum), with dimensions of 1 cm × 5 cm.This foil was cleaned by ultrasonication for 15 min in degreased and deionized water (DI) in sequence, and then dried under cold air.The Ti foil was used as the working electrode and a platinum foil as the counter electrode.The distance between these electrodes was about 2 cm 27 .The anodization was performed at 60 V at a potential ramp of 1 V s -1 , in organic electrolyte consisting a mixture of glycerol (Synth 99.5%) and ethylene glycol (Synth -purity: 99.0%) in 2:1 (v/v) ratio, containing 0.25 wt% ammonium fluoride (Synth -purity: 98.0%) dissolved in 2 wt% of distilled water.The pH of the solution was pH 4.5-5.0 and the anodization time was varied up to (a) 5 minutes, (b) 15 minutes, (c) 30 minutes, (d) 60 minutes, (e) 120 minutes, (f) 180 minutes and (g) 240 minutes.The current transient (I-t) curves for the anodization were obtained by system acquisition (Multimeter Inimipa ET-2076A) of data recorded from the source meter connected to the software TCXX.After anodization, the samples were annealed in a muffle Furnace at 450 °C for 3 h with a heating ramp of 10 ºC/min to crystallize the TiO 2 nanotubes 34 .

Characterization
The morphology was studied with a Field Emission Gun Scanning Electron Microscope (FEG-SEM) equipment MIRA3 by TESCAN operated at 10 and 15 kV.The crystal structure was determined by X-ray diffraction (XRD) using XRD 6000 by SHIMADZU, equipped with a graphite mono-chromator and rotating copper anode operated at 40 kV and 30 mA.The X-ray source consisted of Cu radiation (1.54184 Å) selected with a Ni filter.The measurements were performed with a step of 0.04° and a counting time of 0.60 s per step.The optical properties of the TiO 2 nanotubes were investigated by UV-Vis diffuse reflectance spectroscopy using the Carry 5000 spectrophotometer.The spectra were recorded in the wavelength range of 200-800 nm and contributions from scattering were removed using Kubelka-Munk function 41,42 .The X-ray photoelectron spectrometer (XPS) was conducted using a XPS AXIS ULTRA from Krator.The measurement carried out with Al K α radiation (1486.6 eV).Survey spectra were determined by 160 eV pass energy and the specific element pass energy was fixed at 40 eV.The element energy position was corrected using Ag energy, as a reference, obtained from a pure Ag sample.
The photoelectrochemical measurements were performed using an Autolab potentiostat (PGSTAT 100N).The experiments were performed in a quartz cell using a three-electrode setup employing TiO 2 nanotubes as the working electrode, a Pt wire as the counter electrode and an Ag/AgCl/3.5MKCl electrode as reference electrode.The Ag/AgCl reference electrode was placed in a Lugging capillary to avoid a possible contamination of the working solution by the chloride coming from the reference compartment.The electrical contacts for all of the samples were made using copper wires in a homemade airtight Teflon reactor.Before each experiment, the reference electrode was properly washed with deionized water to avoid contamination of the electrolyte with chloride ions.The linear sweep voltammetry was performed at a scan rate of 10 mV/s in the dark and under illumination.A solar simulator (Newport) incorporated with xenon lamp (Oriel) was used as a radiation source and the radiation was filtered by an AM 1.5G filter.The intensity of the light was Materials Research calibrated by a photodiode with a known responsivity to 100 mW/cm 2 (i.e. 1 sun illumination).These electrodes were thermally treated at 450 °C which contributes to avoid organic contamination.Prior to each measurement, the photoanodes were cleaned by 3 successive cyclic voltammetry scans, applying potentials from 1 V to -1 V vs Ag/AgCl with a scan rate of 50 mV/s 43,44,45 .Prior to Electrochemical impedance spectroscopy (EIS) measurements, 3 successive CV runs were carried out.The EIS experiments were performed in the same experimental setup with equal areas shown to the electrolyte for all samples.The frequency range was 100 kHz -100 mHz with a signal amplitude of 10 mV (peak to peak).The data was acquired under 1 Sun illumination at open circuit potential (OCP) for the respective electrodes.The waiting time to stabilize the system for OCP under 1 Sun illumination was 2 min and these values were ranged from -0.75 to 0.80 V vs Ag/AgCl.For TiO 2 the stable V oc values are obtained within less than 10 sec. of illumination.Since even 10 sec illumination is enough for the open circuit potential (V oc ) stability; however, we followed experimental protocol of giving more time to stabilize the V oc under illumination before performing the EIS experiments 43,44,45

Results and Discussion
Figure 1 shows the current transient (I-t) curve obtained during the maximum time period of the anodization process.At the beginning of the anodization process, up to ~5 minutes, the anodizing current density increases abruptly to a value of ca.1.8 mA, followed by a rapid decrease.During this anodization time, the oxidation of Ti to Ti 4+ occurs due to the hydrolysis reaction resulting in the formation of an initial oxide barrier 46,47 .In addition, during this stage, the curve essentially follows the fluoride-free case 48 , and if samples are removed from the electrolyte, a compact oxide layer will be formed 46 .After the negatively-charged cation vacancies result from dissolution in the oxide, the latter migrates to the metal/oxide barrier due to the potential gradient at the interface 49 .Meanwhile, the oxidation of Ti creates vacancies as well, hence facilitating the hopping of cations into the Ti vacancies 46 .This event slightly increases the current for a short time, resulting 32 , in the pitting of the oxide layer due to field-assisted dissolution.In Figure 1 from 15-30 minutes, a minor increase in the current density can be clearly seen, which is related to the pitting event.Thus, irregular nanoscale pores are expected to be formed initially, which penetrate the initial compact oxide.Subsequently, a steady-state current over time can be observed after 60 minutes of anodization time.After that, the anodization current was approximately maintained at a steady value until 240 minutes (Figure 1).Compared to the literature, the current density and the time to achieve the steady state current density are slightly different to those observed in glycerol or ethylene glycol electrolytes.This is due to the fact that we have used a mixture of both of them; ethylene glycol because of its higher viscosity and higher current density, and glycerol to avoid current density fluctuations during the anodization process.We expect that the combined effect of both may be advantageous for easy growth of TiO 2 nanotubes 37 .According to Albu et al. 38 and Macak et al. 39 the synthesis of different morphologies of TiO 2 has been shown utilizing either ethylene glycol or glycerolbased electrolytes.We have used a mixture of these (viz.ethylene glycol and glycerol with a low concentration of fluoride) with the aim of investigating the evolution of tubular morphology and its applications in photoelectrochemical water splitting.
To investigate the evolution of nanotubular morphology, we conducted a series of experiments via the removal of samples after (a) 5 minutes, (b) 15 minutes, (c) 30 minutes, (d) 60 minutes, (e) 120 minutes, (f) 180 minutes and (g) 240 minutes from the anodization bath.Figure 2 shows the FEG-SEM images of the TiO 2 nanotubes prepared at different anodization times.After 5 minutes (Figure 2 (a)) of anodization, a thin oxide layer (105 nm) was formed on the surface.By increasing the anodization time to 15 and 30 minutes (Figure 2 (b) and (c), respectively), the thickness of the oxide films was increased to 218 nm and 850 nm, respectively.For the sample prepared at 30 minutes, the formation of slightly regular pits can be observed compared to those obtained at shorter time.However, tubular morphology could not be yet obtained.This may be ascribed to the fact that in these three samples, the surface barrier layer formed on the substrate surface of Ti foil is not dissolved during the anodization process, thus leading to the formation of the compact oxide layer.In addition, the I-t curve of Figure 1 clearly shows that up to ~15 minutes, the current density still decreases, and from 15 minutes to 30 minutes, a hump can be seen; yet steady current is not achieved.Thus, these conditions are not suitable for the formation of a tubular morphology.
These results are in line with the literature, as the formation of highly ordered nanotubes occurs when the current density becomes steady 46 .Figures 2(d)-(g) show the morphology of TiO 2 nanotubes after anodization times of 60 minutes, 120 minutes, 180 minutes and 240 minutes, respectively.In Figure 1, the current density during these anodization times remained steady, indicating that the dissolution rate is in equilibrium with the oxidation rate, and hence the formation of nanotubular morphology can be inferred 46 .These figures clearly confirm the formation of a nanotubular morphology; however, it is clear that the top layers of these samples do not show a well-defined tubular structure.It is known that during initial stage of anodization, the compact oxide, i.e., the random pore layer or the initiation layer often remains as remnants on the tube tops called nanograss and is frequently observed previously in the literature 39,50 .Efforts have been directed toward removing these layers [50][51][52] .Therefore, we have tried the methodology reported elsewhere to remove that layer 53,54 .
Interestingly, after removal of the surface nanograss in Figures 2(d) and 2(e), a double-walled feature is revealed.The top-view images clearly show that these nanotubes are comprised of inner and outer shells.These features have been observed earlier in the literature 54 .It was found that the inner shell presented a large amount of F and C species; hence, the formation of the double-walled feature was suggested to be related to the residual amount of the trapped electrolyte species 55 .
Albu et al. 38 explained the thickness of nanotubes in the form of an outer and inner shell, with a variation of thickness along the tube.Additionally, a fluoride-rich layer between the bottom of the nanotubes and the Ti substrate was observed, which was attributed to high-field migration of electrolyte anions, in particular small fluoride species that compete with O 2-migration 39,55,56 through the amorphous tube bottom during the tube growth.On the other hand, the double-walled features disappeared for the samples prepared at 180 and 240 minutes (Figures 2(f) and (g)) of anodization time.Therefore, single-walled or doublewalled TiO 2 nanotubes can controllably be obtained by the optimizing the anodization time.In Figure 2 (e) in the bottom of the nanotubes obtained after 120 min of anodization, the formation of double-walled nanotubes can be confirmed.However, after 180 minutes of anodization (Figure (f &  g)) the formation of double-wall was not observed and the outer diameter of these nanotubes is smaller than those prepared for shorter time durations; suggesting that the chemical etching occurred to dissolve the outer-wall of the nanotubes.These results indicate the effectiveness of the anodization method in producing well-defined TiO 2 nanotubes and controlling the nanotubular morphology via anodization time 57 .
Figure 3 compares the cross-sectional images of the TiO 2 nanotubes obtained at 60 minutes, 120 minutes, 180 minutes, and 240 minutes of anodization times.It can be seen that the tube length increases with increasing time.
The geometrical parameters as a function of anodization time obtained from the FEG-SEM analyses are displayed in Figure 4.The length of the nanotubes increases from ~650 nm to ~6 µm, and the average length growth rate was found to be ~25 nm/min.Furthermore, the outer diameter and the inner diameter vary from 210 to 170 nm and 100 to 88 nm, respectively.This indicates that the nanotubes' diameter is extremely dependent on the anodization time 51 .The decrease in the diameters for larger anodization times can be related to the formation of single-walled nanotubes.When compared to a similar work in electrolyte without the addition of glycerol for similar anodization conditions, the nanotubes obtained here are found to be shorter; however, the outer diameter increased 26,34 .These results suggest that the addition of glycerol to the electrolyte results in decreasing the growth rate of TiO 2 nanotubes; yet it increases the diameter (Figure 4).
The as-anodized nanotubes are generally amorphous; therefore, thermal treatment is required to crystallize them 58 .Figure 5 shows the XRD patterns of the samples annealed at 450 °C for 3 h.The characteristic peak related to the anatase phase of TiO 2 appears at 2θ = 25.2°.The peaks at 38.3º, 40.4° and 53.4° correspond to Ti substrate KOH 30,34,58 .It should be noted that for short anodization times such as from 5-30 minutes, the characteristic anatase peak is hardly visible.On the other hand, its relative intensity compared to the Ti peaks increases with anodization time.These results are in line with the observed oxide thickness, as for long time of anodization, a thicker nanotubular layer is formed (Figure 4).The crystalline properties and optical absorption are two important factors for the photoelectrochemical performance of TiO 2 nanotubular photoanodes.Subsequently TiO 2 nanotubes with a better optical absorption and higher crystallinity will result in a better photoelectrochemical activity.
UV-Vis absorption spectra were obtained to assess the optical absorption of the TiO 2 nanotubes obtained at different anodizing times.For the sake of simplicity, UV-Vis spectra of only a few samples are shown in Figure 6.The optical bandgap energy was found to be ~3.1-3.2 eV for all samples, and is in line with the reported value for TiO 2 59,60 . It can be seen that the absorption from the sample prepared for    240 minutes is higher than that prepared for shorter times of anodization.This is due to the fact that the nanotube obtained for longer times of anodization is longer (Figure 3), therefore exhibiting high absorption 59 .
The XPS spectra of the TiO 2 nanotubes obtained at 5 and 120 minutes of anodization are presented in Figure 7.These spectra are obtained from the as-anodized and thermally treated samples.The survey spectra (Figure 7a) show that the dominant signals are from Ti, O and C, and the relative intensities of C1s for 5 min samples are higher than those of the samples prepared for 120 minutes.The exposure of the samples to air, as well as the anodization solution itself, can be reasons for the carbon content.However, for the anodic thin film obtained at 5 minutes, the higher carbon content might be related to the high carbon diffusion at the thin oxide/metal barrier.For the Ti 2p spectra (Figure 7b), the core lines for 2p 1/2 and 2p 3/2 are centered at 464 eV and oxide 455.8 eV respectively, related to Ti 4+ 61 .For all the samples, the Ti 2p spectra do not differ in binding energies, indicating the formation of TiO 2 for the applied anodization conditions.Additionally, the deconvoluted O ls spectra (Figure 7c) present two peaks.The peak at ~531 eV corresponds to the oxide ions of the TiO 2 , and another peak related to the oxygenated species at higher binding energy side centered at ~532.7 eV corresponds to the oxidized hydrocarbons.However, for the sample prepared for 5 min higher energy peak present larger area as compared to the oxide ion peak of TiO 2 (~531 eV).It may be due to the fact that for 5 min of anodization time the porous oxidized layer is thin and non-uniform; helping to incorporate higher concentration of oxygenated species from the solution to the substrate.
Subsequently, the samples obtained were tested for photoelectrochemical water splitting in 1 M KOH (aq) electrolyte under AM 1.5 G (1 sun illumination) conditions.The linear sweep voltammetry (LSV) curves are presented in Figure 8.The best photoelectrochemical performance is acquired from the sample obtained at 120 min of anodization time.Comparing its optical properties with samples prepared at longer times of anodization against their photoelectrochemical performance, it is clear that better optical absorption properties (Figure 6) and improved photoelectrochemical performance cannot be achieved simultaneously.Moreover, the photoelectrochemical performance of the sample prepared at 60 min is also better than that of samples prepared at 180 and 240 min (Figure 8).In accordance with the Beer-Lembert law, more photons will be harvested by the longer nanotubes.On the other hand, these nanotubes will have more recombination centers and higher series resistance, which result in a decreasing photoelectrochemical performance.
Thus, to ensure efficient electron collection, there should be a balance between the tube length and the optical properties of  TiO 2 nanotubes.Furthermore, the lower photoelectrochemical performance of the samples prepared at shorter anodization times (≤ 30 min) can be rationalized on the basis of their compact morphology, resulting in a lower surface area as well as lower optical absorption abilities.Therefore, the 1D nature of TiO 2 nanotubes and their optimal light absorption are important factors in improving the photoelectrochemical performance.
To obtain further insight into the photoelectrochemical properties of TiO 2 nanotubes as a function of anodization time, the semiconductor/electrolyte interfacial charge transportation properties were investigated by electrochemical impedance spectroscopy (EIS).Figure 9 displays the Nyquist plots obtained from the TiO 2 nanotubes under AM 1.5G (1 sun illumination) conditions at open circuit potential.It can be seen that all of the samples present an incomplete semicircle, which is the characteristic of a capacitive system exhibiting non-ideal behavior.One time constant corresponding to the native passive anodic film was observed; hence, EIS data were fitted to an equivalent circuit (Figure 9) consisting of a series resistance R s and single constant phase element (CPE) parallel with polarization resistance R p .Based on the nanotubular morphology, i.e. porous structure, adding CPE Z Y j for the TiO 2 nanotubes-electrolyte interface is physically true 20 .The impedance of the CPE is given by: Where Y 0 is the admittance of an ideal capacitance, ω is the signal frequency and n is an empirical constant, ranging from 0-1 [62][63][64] .Fitting the Nyquist plots into the proposed equivalent circuit, the values of the circuit elements have been deduced and are displayed in Table 1.
To visualize R p and admittance (Y 0 ) of CPE as a function of anodization time better, we have plotted R p and Y 0 against anodization time (Figure 9(b)).Y 0 increases with the anodization time (Table 1), showing that the interface becomes more capacitive; in other words, the double layer capacitance increases with anodization time.This is due  to the fact that longer nanotubes offer large area for the electrolyte; hence, capacitance increases with time.These results corroborate the formation of a thick anodic film when the anodization time is increased; the increase in the relative intensity of the XRD peak of anatase TiO 2 (Figure 5) and an increase in the length of nanotubes from the FEG-SEM images (Figure 3) have clearly confirmed that.Furthermore, the charge transfer resistance (R p ) decreases from 5 minutes up to 120 minutes of anodization time, and then slightly increases upon any further increase in anodization time.These results are in line with the photoelectrochemical performance of these samples, as the best performance was obtained for the sample anodized for 120 minutes (Figure 8).On the other hand, the sample prepared for 5 min of anodization time had the highest charge transfer resistance as well as the lowest photocurrent response.Additionally, the samples prepared at 240 and 180 minutes had lower charge transfer resistance compared to those prepared at shorter times (≤ 30 minutes), mainly due to their large absorption as well as nanotubular nature.Hence, for improved photoelectrochemical performance, the optimal charge transportation strongly depends on the tubular morphology as well as the optical properties of the TiO 2 nanotubes.Based on the LSV curves and Nyquist plots, the sample prepared for 120 minutes of anodization time had the best photoelectrochemical properties, and has the potential to be used for technical applications.

Conclusions
In this work, the evolution of the tubular morphology of TiO 2 nanotubes was studied by varying anodization time in glycerol mixed ethylene glycol electrolyte-based fluorinated electrolyte.The tubular morphology was formed after 60 minutes of anodization time; before that, a compact oxide layer was observed.The diameter of the nanotubes slightly increases after 120 minutes due to the formation of single-walled nanotubes.However, their length increases linearly with anodization time.The formation of the single-or double-walled nanotubular morphology of TiO 2 is efficiently controlled by the anodization time.For improved photoelectrochemical performance from the TiO 2 nanotubes, the tube length should be sufficiently long to harvest most of the photons, yet short enough to decrease the recombination carriers.The performance of TiO 2 nanotube photoanodes evaluated in photoelectrochemical water splitting demonstrated that TiO 2 nanotubes anodized for 120 minutes presented the best performance.Also, Y 0 increased with the anodization time, showing that the interfacial capacitance increased with anodization time.

Acknowledgments
Thanks are due to CAPES / UDELAR (047/2013) and FAPERGS PqG (2235-2551 / 14-4) for the financial support.We also thank C-LABMU, UEPG, for carrying out the FEG-SEM analysis and LCMicro-UCS and INPE for their contribution in this work.

Figure 1 .
Figure 1.Current density-time behavior for titanium anodized at 60 V in 0.25 g NH 4 F and 2 g H 2 O for 240 minutes.

Figure 5 .
Figure 5. XRD patterns of TiO 2 samples annealed at 450 °C during 3 h with a heating and cooling rate of 10 °C min -1 .

Figure 6 .
Figure 6.UV-Vis diffuse reflectance spectra of TiO 2 nanotubes obtained at different anodization times

Figure 8 .
Figure 8. Linear sweep voltammetry curves of TiO 2 nanotubes obtained at different anodization times.

Figure 9 .
Figure 9. (a) Nyquist plots of TiO 2 nanotubes obtained at different anodization times, where the symbols represent the experimental data and the solid lines represent circuit fitting; and (b) values of R p and Y 0 obtained from the circuit fitting as a function of anodization time.

Table 1 .
EIS parameters for the samples.