Simulation of anodizing current–time curves and the morphology evolution of TiO₂ nanotubes obtained in phosphoric electrolytes

Abstract

Although anodic TiO₂ nanotubes (ATNTs) have been investigated for many years, the growth kinetics need to be further studied to work out a comprehensive fully theoretical model. Here, the simulation and separation of anodizing current–time curves are applied to investigate the morphology evolution in mixed NH₄F/H₃PO₄ electrolytes for the first time. The total anodizing current was separated into ionic current and electronic current. The influence of NH₄F concentration on the ionic current, electronic current and morphology evolution was systemically studied. The linear relationship between nanotube length and ionic current indicates that the ionic current contributes to oxide growth. It is the first time the special surface layer of nanotubes and two-hole nanotubes in mixed electrolytes have been observed. Furthermore, a formation mechanism for the two-hole nanotubes is proposed based on the plastic flow model and oxygen bubble mould.

---

Introduction

Anodic TiO₂ nanotubes^1,2 (ATNTs) have attracted considerable attention due to their potential applications in dye sensitized solar cells,³ biosensors⁴ and supercapacitors⁵ as well as their mysterious formation mechanism.^6–10 Much effort has been made to investigate the formation mechanism and some theories were proposed accordingly, including field-assisted dissolution theory,^11–13 plastic flow model^14–17 and oxygen bubble mould.^17,18 The generally accepted mechanism is a field-assisted dissolution process (TiO₂ + 6F⁻ + 4H⁺ → [TiF₆]²⁻ + 2H₂O) for the pore formation.^11–13 However, Diggle and Thompson et al.^19,20 indicated that the field-assisted dissolution is chemical in nature and does not contribute to the anodizing current across the barrier oxide. Moreover, direct measurement of the field-dependent oxide dissolution rate has been never done or reported.²⁰ In 2006, Garcia-Vergara et al.²¹ concluded that pore morphology was influenced by oxygen bubbles. Zhu et al.^2,6,10,18,22 have proposed an oxygen bubble mould, which emphasizes that ionic current contributes to the growth of oxide while electronic current leads to oxygen evolution. These theories are able to make the qualitative prediction of pore formation, but up to now none has been worked out a comprehensive fully theoretical model that directly translates into quantifiable experimental data and explains the relationship between pore formation and anodizing current.^2,15

Generally, ATNTs are fabricated in NH₄F electrolyte.^23–26 However, there are few researches on the mixed electrolytes,²⁷ let alone quantitatively discuss the anodizing current and morphology evolution in mixed electrolytes. Here, Ti anodization process in three different electrolytes was studied, including pure H₃PO₄ electrolyte, pure NH₄F electrolyte and mixed NH₄F/H₃PO₄ electrolytes. And the total anodizing current was separated into two parts: ionic current and electronic current by a novel method. The results show that the total anodizing current, ionic current and electronic current increase with the addition of NH₄F concentration in mixed electrolytes. The ionic current contributes to oxide growth and there is a linear relationship between the ionic current and nanotube length. The present results can provide unique insights into the fabrication of TiO₂ nanotubes, and help to better understand the growth kinetics of the porous anodic oxides.

Experimental

Titanium foils (100 μm thick, purity 99.5%) were polished in the polishing solution before anodization, which was a mixture of HF (≥40%), HNO₃ (65–68%) and deionized water (1 : 1 : 2 in volume). The chemical polishing can renew the foil surface by removing the effect of history and exposing a fresh microstructure. Then, the samples were rinsed thoroughly by deionized water and dried in the air prior to anodization.

The fluoride-free electrolytes, prepared by using reagent grade chemicals, were ethylene glycol (EG) solutions containing 2 wt% H₃PO₄ and 2 vol% H₂O. Besides, different concentrations of ammonium fluoride were added to the above original electrolyte (SE, i.e., EG solutions containing 2 wt% H₃PO₄ and 2 vol% H₂O). The fluoride-containing electrolyte with 0.5 wt% NH₄F and 2 vol% water was prepared for comparison. Anodization was carried out by employing a two-electrode electrochemical cell, in which the polished foils of Ti were used as anode and a Pt mesh as cathode. The first group of pretreated samples were all anodized in corresponding solutions under 70 V at 30 °C for 500 s. Another group of pretreated samples were all anodized in corresponding solutions under 10 mA at 30 °C for 500 s. The anodizing area of Ti foil was maintained constant (∼2 cm²). At the end of the anodization, the samples would be immersed in deionized water in order to clear the electrolyte and dried in the air. All samples were examined by the field-emission scanning electron microscope (FESEM, Zeiss Supra 55). The length and diameter of nanotubes were measured directly on the FESEM.

Results and discussion

Anodizing current density, voltage and concentration of NH₄F

Fig. 1a shows the current density–time curves obtained in 2 wt% H₃PO₄ electrolyte with n wt% NH₄F (n = 0, 0.25, 0.50, 0.75, 1.00, 1.25 and 1.50, respectively) under 70 V. The current density–time curves in Fig. 1a, except pure H₃PO₄ electrolyte (n = 0), show the typical curve with three stages: an initial exponential decay, an increase of current and a quasi-steady state.^28–31 With the addition of NH₄F, the durations of first stage (the decrease of current density) and second stage (the increase of current density) decrease accordingly. Compared with other four groups of electrolytes, the duration of first two stages is extremely long when the NH₄F concentration is 0.25 wt%, while the duration is extremely short when the NH₄F concentration is 1.50 wt%. Furthermore, the minimum value of current density increases from 0.45 mA cm⁻² to 30.05 mA cm⁻² and the equilibrium current density increases from 3.06 mA cm⁻² to 93.52 mA cm⁻², when NH₄F concentration increases from 0.25 wt% to 1.50 wt%. However, the current density–time curve tends to fluctuate at the later duration when NH₄F content reaches to 1.25 wt%. Especially, when NH₄F content reaches to 1.50 wt%, the anodizing current density continues to increase, leading to the breakdown of nanotubes eventually. Fig. 1b shows the voltage–time curves obtained in 2 wt% H₃PO₄ electrolyte with n wt% NH₄F (n = 0, 0.25, 0.50, 0.75, 1.00, 1.25 and 1.50, respectively) under 10 mA. The voltage–time curves in Fig. 1b show the typical anodizing curve with three stages, except those in the electrolytes with 0 wt% and 1.50 wt% NH₄F.^30,31 Similarly, the durations of first stage (the increase of voltage) and second stage (the decrease of voltage) decrease with the addition of NH₄F. At the same time, the equilibrium voltage decreases with the addition of NH₄F. However, when NH₄F content reaches to 1.50 wt%, the voltage–time tends to fluctuate.

Fig. 1 (a) The current density–time curves in different mixed electrolytes obtained under 70 V, (b) the voltage–time curves in different mixed electrolytes obtained under 10 mA.

The relationship between anodizing current density–time curves and the morphology of ATNTs has been explored by many groups.^11–13 However, the increase and quasi-steady state of the current density cannot be clearly explained based on the field-assisted dissolution theory. Our group has made much effort on the quantitative analysis of current density–time curves and proposed a quantitative model based on oxygen bubbles mould^17,18 and oxide flow model.^14–17 Yang et al.³² have investigated anodizing current to find out the quantitative relationship between anodizing current and the morphology evolution by a theoretical equation, where the anodizing current is divided into ionic current J_ion and electronic current J_e. At stage I, J_ion contributes to the growth of barrier oxide layer, while the increase of barrier oxide leads to the decrease of J_ion.^30,33–36 At stage II, the generation of electronic current is attributed to the formation of anion contaminated layer (ACL)^30,31 due to the contamination of F⁻ and OH⁻ anions.^35,36 And the increase of J_e causes the evolution of oxygen bubbles, which behaves as a mould and therefore contributes to the formation of porous nanostructure.³⁷ At stage III, J_ion and J_e arrive at the equilibrium state. Actually, it is impossible to observe oxygen bubbles in nanopores with modern analysis techniques.^38,39 Because oxygen evolution has stopped after anodization, it is impossible to observe oxygen bubbles under the FESEM or TEM. However, it is a common belief that the existence of electronic current is bound to lead to oxygen evolution.^{24,25,30,36,39–43}

The total anodizing current density–time curve can be described by a theoretical expression shown below:^32,33

where, U₁ is the voltage across the electrolyte and oxide layer, R_el is the resistance of electrolyte. α and β are constants with physical significance. θ and C_ion are the correction factors of ionic current, i_e0 represents the initial electronic current, and η represents the correctional parameter of time delay.^22,32 The ionic current can be seen at the first two items of the expression, and the last item represents electronic current over time. The fitted functional formula of each curve is given as:

Obtained in 2.00 wt% H₃PO₄ electrolyte with 0.25 wt% NH₄F:

i = 49.10 e^−0.2414t + 0.6350 + 9.455 × 10⁻¹² e^{26.42(1−e−0.01351t)}, R² = 0.9963 (2)

Obtained in 2.00 wt% H₃PO₄ electrolyte with 0.50 wt% NH₄F:

i = 72.89 e^−0.3221t + 2.237 + 4.752 × 10⁻²⁵ e^{58.18(1−e−0.03546t)}, R² = 0.9960 (3)

Obtained in 2.00 wt% H₃PO₄ electrolyte with 0.75 wt% NH₄F:

i = 59.42 e^−0.3577t + 4.770 + 4.258 × 10⁻⁵ e^{12.60(1−e−0.05346t)}, R² = 0.9811 (4)

Obtained in 2.00 wt% H₃PO₄ electrolyte with 1.00 wt% NH₄F:

i = 71.94 e^−0.3605t + 7.905 + 0.006402 e^{7.924(1−e−0.05612t)}, R² = 0.8603 (5)

Obtained in 2.00 wt% H₃PO₄ electrolyte with 1.25 wt% NH₄F:

i = 81.67 e^−0.4269t + 8.844 + 0.4018 e^{4.126(1−e−0.09079t)}, R² = 0.5799 (6)

Obtained in 2.00 wt% H₃PO₄ electrolyte with 1.50 wt% NH₄F:

i = 50.52 e^−2.144t + 28.86 + 8.765 e^{2.648(1−e−0.003080t)}, R² = 0.9562 (7)

R reflects the goodness of fit and R² is the coefficient of determination.

Fig. 2 shows the comparison between measured current density–time curves and fitted ones. The measured current density–time curves fit well with the fitted curves when NH₄F concentrations are 0.25 wt%, 0.50 wt% and 0.75 wt% (R² > 0.98).

Fig. 2 Comparison between measured current density–time curves and fitted ones in different mixed electrolytes with n wt% NH₄F (n = (a) 0.25, (b) 0.50, (c) 0.75, (d) 1.00, (e) 1.25 and (f) 1.50, respectively).

Ionic current and electronic current

According to the fitting functional formula, the total anodizing current density is separated into two parts: the ionic current and electronic current. The theoretical formulas are as below:^32,33

Fig. 3 shows the fitted ionic current density–time curves and electronic current density–time curves in different electrolytes. It can be easily found that the inflection point in Fig. 1a is caused by the occurrence of electronic current. Both ionic current density and electronic current density increase with NH₄F concentration. The main reason is that the conductivity of electrolytes is high when the NH₄F concentration is high. Meanwhile, the generation of electronic current becomes quicker and quicker with the increase of NH₄F concentration. More F⁻ anions lead to a bigger chance of anionic incorporation into the barrier oxide, which causes the original electronic current. More anionic incorporation into the barrier oxide results in higher electronic current along with oxygen evolution. However, the electronic current density keeps increasing exponentially when NH₄F concentration reaches to 1.50 wt%, which leads to the rupture of nanotubes eventually. The electronic current in pure NH₄F electrolyte is larger than that in mixed electrolytes except the electrolyte with 1.50 wt% NH₄F. Furthermore, the ionic current in pure NH₄F electrolyte is close to that in mixed electrolyte with 0.75 wt% NH₄F. It is demonstrated that the H₃PO₄ in mixed electrolytes leads to the decrease of ionic current and electronic current. As a result, when NH₄F concentration is low (n = 0.25) in mixed electrolyte, the current density is close to zero, as shown in Fig. 1. Electronic current in fluoride-containing electrolytes is much larger than that in fluoride-free electrolytes because F⁻ anions generally migrate faster than PO₄³⁻ anions. Hence, the existence of PO₄³⁻ anions in mixed electrolytes hinders the movement of F⁻ anions. Table 1 shows the ionic current density and electronic current density recorded at 400 s during anodization. Here, we define E (E = J_e/J) as the efficiency of anodization, to study the effect of NH₄F on current in mixed electrolytes. At the same time, the normalization of E mainly eliminates the effect of dimensions in order to draw the scatter diagrams in Fig. 9c.

Fig. 3 Comparisons of ionic current density–time curves (a) and electronic current density–time curves (b) in different mixed electrolytes and pure NH₄F electrolyte.

Table 1 The steady-state ionic current density (J_ion,s), steady-state electronic current density (J_e,s) and anodizing efficiency (E) in different mixed electrolytes

NH4F content (wt%)	Jion,s (mA cm^−2)	Je,s (mA cm^−2)	E	Normalization of E
0.25	0.6350	2.817	0.8160	1
0.50	2.237	8.792	0.7971	0.8752
0.75	4.770	12.63	0.7259	0.4049
1.00	7.905	17.69	0.6912	0.1757
1.25	8.844	24.88	0.7378	0.4835
1.50	28.86	57.19	0.6646	0

---

Morphology evolution of ATNTs in different mixed electrolytes

Fig. 4 and 5 respectively show the surface morphology of nanotubes obtained under 10 mA and 70 V in different electrolytes with n wt% NH₄F (n = 0.25, 0.50, 0.75, 1.00, 1.25 and 1.50, respectively). The surface morphology and porosity of nanotubes in different mixed electrolytes are significantly different. The porosity of nanotubes increases significantly with the NH₄F concentration under constant voltage or current. The electronic current increases with the addition of NH₄F as shown in Table 1, and the electronic current is attributed to the generation of oxygen bubbles. As a result, larger electronic current contributes to more oxygen evolution as well as high porosity.

Fig. 4 The surface morphology of nanotubes obtained under 10 mA in different electrolytes with n wt% NH₄F (n = (a) 0.25, (b) 0.50, (c) 0.75, (d) 1.00, (e) 1.25 and (f) 1.50, respectively).

Fig. 5 The surface morphology of nanotubes obtained under 70 V in different electrolytes with n wt% NH₄F (n = (a) 0.25, (b) 0.50, (c) 0.75, (d) 1.00, (e) 1.25 and (f) 1.50, respectively).

Kruse et al.⁴³ indicated that the evolving rate of the anodic oxygen bubble is sufficiently high due to the residual concentration. However, it can be found that the surface is half-opening in Fig. 4a and is irregular in Fig. 5f. On the one hand, when the NH₄F concentration is low, there is too little anionic incorporation into the barrier oxide to cause enough electronic current to give rise to oxygen evolution, leading to the half-opening-pore in Fig. 4a. On the other hand, when NH₄F concentration is high, the electronic current will keep increasing exponentially as shown in Fig. 3b, leading to the oxide breakdown and the different morphology in Fig. 5f. Actually, the morphology in Fig. 5f is called wall-separated highly oriented free standing nanotube structure,²⁷ which is indistinct in pure NH₄F electrolyte, although the conditions of anodization are different. The above result indicates that short anodizing time and high voltage have the similar effect with long anodizing time and low voltage in mixed electrolyte at a certain degree.

Fig. 6 shows the FESEM image of ATNTs obtained in 0.50 wt% NH₄F electrolytes with n wt% H₃PO₄ (n = 0, 2, respectively). The nanotubes in Fig. 6b are larger and shorter than those in Fig. 6a, which demonstrates that the addition of H₃PO₄ contributes to the increase of nanotube diameter.⁸ Hazra et al.²⁷ also found that larger tube diameter can be observed for the samples prepared by mixed electrolyte. The ionic current contributes to the growth of nanotubes and therefore the nanotubes in pure electrolyte are longer. Furthermore, the critical thickness of compact oxide in the fluoride-containing electrolyte is much thinner than that in fluoride-free.¹⁸ In mixed electrolytes, the electronic current tends to appear later and become lower, leading to the increase of the critical thickness. Therefore, the nanotube diameter is larger in mixed electrolytes.

Fig. 6 The FESEM image of ATNTs obtained in 0.50 wt% NH₄F electrolyte with n wt% H₃PO₄ (n = (a) 0, (b) 2, respectively).

Fig. 7 and 8 respectively show the cross-sections of ATNTs obtained under 70 V and 10 mA in different electrolytes with n wt% NH₄F (n = 0.25, 0.50, 0.75, 1.00, 1.25 and 1.50, respectively). The nanotube lengths increase with NH₄F content, while the nanotube diameters decrease with NH₄F content. The length and diameter of nanotubes are listed in Tables 2 and 3. Here, we define ζ (l/d) as the ratio of length to diameter, to study the effect of NH₄F on the cross-section morphology in mixed electrolytes. At the same time, the normalization of ζ mainly eliminates the effect of dimensions in order to draw the scatter diagrams in Fig. 9c.

Fig. 7 FESEM images of cross-sections of ATNTs obtained under 70 V in different electrolytes with n wt% NH₄F (n = (a) 0.25, (b) 0.50, (c) 0.75, (d) 1.00, (e) 1.25 and (f) 1.50, respectively).

Fig. 8 FESEM images of cross-sections of ATNTs obtained under 10 mA in different electrolytes with n wt% NH₄F (n = (a) 0.25, (b) 0.50, (c) 0.75, (d) 1.00, (e) 1.25 and (f) 1.50, respectively).

Table 2 The geometry parameters and anodizing currents for nanotubes obtained in different electrolytes under 70 V

NH4F content (wt%)	Nanotube length (μm)	Nanotube diameter (nm)	The lowest current (mA cm^−2)	Equilibrium current (mA cm^−2)	ζ	Normalization of ζ
0.25	1.090	209.6	0.45	3.06	5.20	0
0.50	1.820	199.5	2.05	10.95	9.12	0.189
0.75	2.547	191.0	4.8	17.08	13.34	0.394
1.00	2.895	182.1	8.65	24.32	15.90	0.518
1.25	3.483	136.1	14.7	36.43	25.59	0.987
1.50	3.484	134.7	30.05	93.52	25.86	1

---

Table 3 The geometry parameters and anodizing voltages for nanotubes obtained in different electrolytes under 10 mA

NH4F content (wt%)	Nanotube length (μm)	Nanotube diameter (nm)	The highest voltage (V)	Equilibrium voltage (V)	ζ	Normalization of ζ
0.25	1.394	237.7	90.63	74.29	5.865	0
0.50	1.619	212.1	80.15	66.82	7.633	0.1570
0.75	1.757	175.9	61.44	52.81	9.989	0.3664
1.00	1.859	174.9	60.00	51.93	10.63	0.4233
1.25	2.050	147.9	42.96	47.06	13.86	0.7103
1.50	1.798	105	39.68	38.39	17.12	1.000

---

Fig. 9 The fitting curves of nanotube length–ionic current density (a) and nanotube diameter–equilibrium voltage (b), the scatter diagrams of E–NH₄F concentration and ζ–NH₄F concentration (c).

In Fig. 9, there is a linear relationship between nanotube length (l) and steady-state ionic current density (J_ion), which demonstrates that nanotube length increases with the increase of ionic current. The fitting function is given as below, which may have a positive influence on predicting the length of nanotube length.

l = 0.25797J_ion + 1.10858, R² = 0.9361 (10)

High NH₄F concentration leads to high conductivity of electrolyte and high steady-state ionic current, as shown in Fig. 3a. Hence, nanotube length increases with NH₄F concentration. The electronic current will occur earlier and increase with NH₄F concentration as shown in Fig. 3b, leading to the decrease of critical thickness. As a result, the wall thickness becomes thicker, which determines the diameter of nanotubes.^30,31 In Fig. 9b, there is a linear relationship between nanotube diameter (d) and equilibrium voltage (U) and it is given as below:

d = 3.04468U + 11.33636, R² = 0.97411 (11)

Compared the nanotube diameters in Tables 1 and 2, the nanotube diameters (∼237.7 nm) are larger under constant current than those (∼209.6 nm) under constant voltage when the NH₄F concentration is 0.25 wt%, while the diameters are smaller under constant current in other mixed electrolytes (n = 0.75 wt%, 1.00 wt%, 1.25 wt% and 1.50 wt%). The main reason is that the equilibrium voltage is 74.29 V (n = 0.25), is higher than 70 V (the applied voltage under potentiostatic conditions), which leads to larger nanotube diameter.

In general, when NH₄F concentration increases, the efficiency of anodization (E) decreases as shown in Table 1, while the ratio of length to diameter (ζ) increases as is shown in Tables 2 and 3. With higher NH₄F concentration, higher ionic current leads to the increase of nanotube length while higher electronic current results in the decrease of nanotube diameter. When NH₄F concentration is low, E is high and ζ is low, leading to short nanotubes with large diameter. On the contrary, when NH₄F concentration is high, E is low and ζ is high, leading to long nanotubes with small diameter. As is shown in Fig. 9c, when NH₄F concentration is 0.75 wt%, E is closest to ζ. From the above experimental results, it can be deduced that the geometric parameters of nanotubes can be controlled by analyzing E and ζ, which has profound influences on the research of morphology evolution of ATNTs.

Special surface layer

Fig. 10e and f show the schematic of ATNTs on Ti substrate without and with the additional special surface layer. There are mainly three layers in pure NH₄F electrolyte, including the inner layer of nanotube (ILN), outer layer of nanotube (OLN) and fluoride-rich layer (FRL), as shown in Fig. 10e. In general, when ATNTs are fabricated in fluoride-containing electrolyte, it is believed that TiO₂ nanotube wall consists of two different parts: ILN and OLN.¹ Actually, an additional interface layer exists (FRL) between the bottom of nanotubes and the Ti substrate.⁴⁴ However, compared to Fig. 10e, there is another layer on the upper part of ATNTs, called special surface layer here, as shown in Fig. 10f. All the nanotubes obtained in mixed electrolytes show the special surface layer in Fig. 7 and 8, while the nanotubes only show FRL in pure NH₄F electrolyte as shown in Fig. 6a. The migration of fluoride species was reported to have twice the speed of the oxygen ion movement and thus lead to their accumulation at the TiO₂/Ti interface,⁴⁵ resulting in the formation of FRL.^46,47 Previous studies demonstrated that the chemical dissolution of FRL would take place due to its water solubility. This chemical dissolution of the FRL was suggested to be the main reason for the poor adhesion strength.^1,48 However, the separated nanotubes are linked by the entire surface layer due to the existence of special surface layer in mixed electrolytes. The special surface layer consists of two parts: upper part and surface of nanotubes, as is shown in Fig. 10f. The upper part is mainly attributed to rupture or dissolution of nanotubes as is shown in Fig. 10c. Some of the nanotubes which fall off from the upper part are left on the surface, as is shown in Fig. 10d. On the one hand, when NH₄F concentration is high, the barrier oxide grows very fast, leading to uneven stress.⁴⁹ On the other hand, the electronic current increases with NH₄F concentration, resulting in intense release of oxygen, which may accelerate the rupture of the upper part of nanotubes. Certainly, the ionic current increases with NH₄F concentration, making nanotubes become long and unstable. Due to the inexistence of special surface layer in pure NH₄F electrolyte, H₃PO₄ plays a leading role in the formation of the special surface layer in mixed electrolytes. It is known to all that compact oxide will be obtained in pure H₃PO₄ electrolyte.^17,32,44 PO₄³⁻ anions are less active than F⁻ anions.⁹ At initial anodization, NH₄F plays a leading role in the formation of ATNTs. After the ionic current and electronic current reach the balance at later anodization, H₃PO₄ addition leads to the formation of the special surface layer. In Fig. 10a, there are few holes on the surface and some holes are closed. The existence of PO₄³⁻ anions hinders the migration of F⁻ and O²⁻ anions, leading to the accumulation of F⁻ and O²⁻ anions in the contaminated layer. The special surface layer is a natural electrolyte access for electrolytes. The electrolytes run through the electrolyte access and reach the gap base around the nanotube. Then it will touch the Ti substrate and electrolyte/Ti interface will appear.⁶ It is indicated that if field-assisted dissolution contributes to the porous nanostructure, the bottom of nanotubes will be dissolved and incomplete. New barrier oxide will form in the electrolyte/Ti interface like those formed at initial anodization and ribs will form around the nanotubes.⁶ Compared Fig. 10c and d, when NH₄F concentration is high, there are more ribs, which indicates that high NH₄F concentration is beneficial to the growth of ribs.

Fig. 10 (a–d) The special phenomena of the surface and cross-sections of nanotubes, the schematic of ATNTs on Ti substrate (e) without and (f) with the additional special surface layer.

Fig. 10b shows the two-hole nanotubes, which results in the non-uniform thickness of nanotubes in the vertical direction. According to oxygen bubble mould and plastic flow model,^14,18 the newly-formed barrier oxide grows from the pore base to the nanotube wall. More oxygen will be released to form pore base with higher NH₄F concentration, as shown in Fig. 4. Meanwhile, H₃PO₄ in the ACL hinders the evolution of oxygen bubbles, making them hard to separate out. Higher NH₄F content and the inhibition of PO₄³⁻ anions result in the accumulation of oxygen bubbles on the surface of oxide. Then two oxygen bubbles released subsequently can act as a mould to form two separate nanopores due to the stress from the newly-formed oxygen bubbles. Thus, it may form multihole nanotubes according to above analysis, which can be demonstrated by the three-hole nanotubes as shown in Fig. 10b.

Conclusions

Anodization process in mixed electrolytes under potentiostatic and galvanostatic conditions was investigated. With the increase of NH₄F concentration in mixed electrolytes, the total anodizing current increases under potentiostatic condition, while the total anodizing voltage decreases under galvanostatic condition. Through quantitatively separating anodizing current into ionic current and electronic current, it is demonstrated that ionic current increases with NH₄F concentration, resulting in the increase of nanotube length, and electronic current increases with NH₄F concentration, leading to the decrease of nanotube wall thickness. Furthermore, there is a linear relationship between ionic current and nanotube length. The present results show that there is a special surface layer (compact oxide) on the surface of ATNTs obtained in mixed electrolytes, which provides a natural electrolyte access and makes it easy for the formation of ribs. The present results also demonstrate that two separate oxygen bubble can act as two moulds to form two separate nanopores.

Acknowledgements

This work was financially supported by the National Natural Science Foundation of China (Grant No. 51577093, 61171043) and the Undergraduate Research Training “Millions Talents” Plan and the Extra-curricular Academic Scientific Research Fund of Nanjing University of Science and Technology. Authors Shiyi Chen and Maoying Liao contributed equally to this work.

References

1. D. L. Yu, X. Zhu, Z. Xu, X. Zhong, Q. Gui, Y. Song, S. Zhang, X. Chen and D. D. Li, ACS Appl. Mater. Interfaces, 2014, 6, 8001.
2. Y. L. Zhang, W. Cheng, F. Du, S. Zhang, W. Ma, D. Li, Y. Song and X. Zhu, Electrochim. Acta, 2015, 180, 147.
3. Y. L. Liu, Y. Q. Cheng, K. Q. Chen, Z. Y. Peng, G. J. Yang, G. S. Zakharovab and W. Chen, RSC Adv., 2014, 4, 45592.
4. L. X. Zheng, Y. C. Dong, H. D. Bian, C. Lee, J. Lu and Y. Y. Li, Electrochim. Acta, 2016, 203, 257.
5. H. Wu, D. Li, X. Zhu, C. Yang, D. Liu, X. Chen, Y. Song and L. Lu, Electrochim. Acta, 2014, 116, 129.
6. B. Chong, D. Yu, M. Gao, H. Fan, C. Yang, W. Ma, S. Zhang and X. Zhu, J. Electrochem. Soc., 2015, 162, H244.
7. D. Regonini, A. Groff, G. D. Sorarù and F. J. Clemens, Electrochim. Acta, 2015, 186, 101.
8. C.-Y. Chen, K. Ozasa, F. Kitamura, K. Katsumata, M. Maeda, K. Okada and N. Matsushita, Electrochim. Acta, 2015, 153, 409.
9. Y. Li, Y. Qin, Z. Y. Ling, X. Hu and Y. Shen, RSC Adv., 2015, 5, 136.
10. Y. L. Zhang, D. L. Yu, M. Q. Gao, D. Li, Y. Song, R. Jin, W. Ma and X. Zhu, Electrochim. Acta, 2015, 160, 33.
11. K. M. Deen, A. Farooq, M. A. Raza and W. Haider, Electrochim. Acta, 2014, 117, 329.
12. H. Li, J. W. Cheng, S. W. Shu, J. Zhang, L. Zheng, C. K. Tsang, H. Cheng, F. Liang, S. T. Lee and Y. Y. Li, Small, 2013, 9, 37.
13. G. D. Sulka, J. Kapusta-Kolodziej, A. Brźozka and M. Jaskula, Electrochim. Acta, 2013, 104, 526.
14. D. J. LeClere, A. Velota, P. Skeldon, G. E. Thompson, S. Berger, J. Kunze, P. Schmuki, H. Habazaki and S. Nagata, J. Electrochem. Soc., 2008, 155, C487.
15. J. E. House and K. R. Hebert, Nat. Mater., 2009, 8, 415.
16. K. R. Hebert, S. Albu, I. Paramasivam and P. Schmuki, Nat. Mater., 2012, 11, 162.
17. X. M. Zhong, D. L. Yu, S. Y. Zhang, X. Chen, Y. Song, D. D. Li and X. F. Zhu, J. Electrochem. Soc., 2013, 160, E125.
18. S. Y. Zhang, D. L. Yu, D. D. Li, Y. Song, J. F. Che, S. Y. You and X. F. Zhu, J. Electrochem. Soc., 2014, 161, E135.
19. J. W. Diggle, T. C. Downi and C. W. Goulding, Chem. Rev., 1969, 69, 365.
20. J. Oh and C. V. Thompson, Electrochim. Acta, 2011, 56, 4044.
21. S. J. Garcia-Vergara, K. E. Khazmi, P. Skeldon and G. E. Thompson, Corros. Sci., 2006, 48, 2937.
22. Y. Wang, D. L. Yu, B. Chong, D. D. Li, Y. Song, S. Y. Zhang, W. H. Ma and X. F. Zhu, J. Electrochem. Soc., 2014, 161, H891.
23. X. J. Xu, X. S. Fang, T. Y. Zhai, H. B. Zeng, B. D. Liu, X. Y. Hu, Y. Bando and D. Golberg, Small, 2011, 7, 445.
24. D. Regonini, C. R. Bowen, A. Jaroenworaluck and R. Stevens, Mater. Sci. Eng., R, 2013, 74, 377.
25. K. Y. Wang, G. H. Liu, N. Hoivik, E. Johannessen and H. Jakobsen, Chem. Soc. Rev., 2014, 43, 1476.
26. Q. Gui, D. Yu, S. Zhang, H. Xiao, C. Yang, Y. Song and X. Zhu, J. Solid State Electrochem., 2014, 18, 141.
27. A. Hazra, B. Bhowmik, K. Dutta, V. Manjuladevi, R. K. Gupta, P. P. Chattopadhyay and P. Bhattacharyya, Sci. Adv. Mater., 2014, 6, 714.
28. W. Lee and S. J. Park, Chem. Rev., 2014, 114, 7487.
29. P. Roy, S. Berger and P. Schmuki, Angew. Chem., Int. Ed., 2011, 50, 2904.
30. X. F. Zhu, Y. Song, L. Liu, C. Wang, J. Zheng, H. Jia and X. Wang, Nanotechnology, 2009, 20, 475303.
31. X. F. Zhu, H. Han, Y. Song, H. T. Ma, W. X. Qi, C. Lu and C. Xu, Acta Phys. Sin., 2012, 61, 228202.
32. R. Q. Yang, L. F. Jiang, X. F. Zhu, Y. Song, D. L. Yu and A. J. Han, RSC Adv., 2012, 2, 12474.
33. B. Chong, D. Yu, R. Jin, Y. Wang, D. D. Li, Y. Song, M. Gao and X. F. Zhu, Nanotechnology, 2015, 26, 145603.
34. G. E. Thompson, Thin Solid Films, 1997, 297, 192.
35. J. M. Albella, I. Montero and J. M. Martinez-Duart, Electrochim. Acta, 1987, 32, 255.
36. Y. Li, Z. Y. Ling, X. Hu, Y. Liu, Y. S. Liu and Y. Chang, RSC Adv., 2012, 2, 5164.
37. X. F. Zhu, Y. Song, D. L. Yu, C. Zhang and W. Yao, Electrochem. Commun., 2013, 29, 71.
38. P. Y. Deng, X. D. Bai, X. W. Chen and Q. L. Feng, J. Electrochem. Soc., 2004, 151, B284.
39. D. Regonini, A. Satka, A. Jaroenworaluck, D. W. E. Allsopp, C. R. Bowen and R. Stevens, Electrochim. Acta, 2012, 74, 244.
40. D. L. Yu, S. Y. Zhang, X. F. Zhu, H. Ma and H. Han, J. Solid State Electrochem., 2014, 18, 2609.
41. R. Jin, H. Fan, Y. Liu, W. Ma, H. Lu, P. Yang and W. Ma, Electrochim. Acta, 2016, 188, 421.
42. G. Patermarakis and K. Moussoutzanis, Electrochim. Acta, 2009, 54, 2434.
43. A. Pauric, S. Baig, A. Pantaleo, Y. Wang and P. Kruse, J. Electrochem. Soc., 2013, 160, C12.
44. A. Mazzarolo, M. Curioni, A. Vicenzo, P. Skeldon and G. E. Thompson, Electrochim. Acta, 2012, 75, 288.
45. H. Habazaki, K. Fushimi, K. Shimizu, P. Skeldon and G. E. Thompson, Electrochem. Commun., 2007, 9, 1222.
46. P. Roy, S. Berger and P. Schmuki, Angew. Chem., Int. Ed., 2011, 50, 2904.
47. S. P. Albu, A. Ghicov, S. Aldabergenova, P. Drechsel, D. LeClere, G. E. Thompson, J. M. Macak and P. Schmuki, Adv. Mater., 2008, 20, 4135.
48. C. C. Chen, H. W. Chung, C. H. Chen, H. P. Lu, C. M. Lan, S. F. Chen, L. Luo, C. S. Hung and E. W. G. Diau, J. Phys. Chem. C, 2008, 112, 19151.
49. Ö. Özgür Çapraz, P. Shrotriya, P. Skeldon, G. E. Thompson and K. R. Hebert, Electrochim. Acta, 2015, 167, 404.

---

Footnote

† Electrochemical Society Active Member.

---