A kinetic model for determining morphology transitions and growth kinetics of titania nanotubes during anodization of titanium in ethylene glycol based electrolytes

Highlights

• Growth kinetics of titania nanotubes can be divided into 3 successive stages.

• Dissolution kinetics of initial barrier oxide determines the growth mode of the tubes.

• The barrier oxide dissolution time can be determined experimentally and theoretically.

• Temperature dependence of the barrier oxide dissolution process obeys Arrhenius law.

• Nanotube morphology and thickness can be controlled with the developed kinetic model.

Abstract

During porous anodization of titanium different surface morphologies can be observed. However, the influence of this morphological variance on nanotube growth has not been defined with a consistent model. In this study, we aim to develop a kinetic model-approach that combines the growth process with the final morphologies of the different nanotubular titania structures that are formed during anodization in ethylene glycol based electrolytes. Accordingly, we divide the nanotube growth process into three sequential stages. Morphological transitions and growth kinetics during anodization are investigated individually for all stages. In the first stage, nanotubes grow under gradually dissolving initial barrier oxide. Nanopore/nanotube transition occurs at the second stage after the complete dissolution of the initial barrier oxide. Nanograss formation starts and progresses at the third stage of the anodization. Growth of nanotubes under the gradually dissolving top barrier layer (Stage 1) obeyed the field-assisted growth model with a growth rate equivalent to 0.7 μm C⁻¹ cm². However, the completion of the chemical dissolution of top barrier layer totally changes the growth rate leading to a shortening of the nanotubes (Stage 3). After experimental determination of the chemical shortening rate of the nanotubes (CSR) at this stage, a kinetic model has been produced to determine the top barrier layer dissolution time (BDT). The experimental determination of the temperature dependency of BDT allowed us to calculate the activation energy of the process and determination of BDT values for different processing temperatures by extrapolation. Extrapolated BDT values for different temperatures showed well consistency with the experimental results and validate that open tube-top nanotubes can be obtained at calculated anodization durations.

Introduction

Titanium dioxide (TiO_2, titania) nanoporous structures have gained great interest due to their unique properties, making them the material of choice in various applications, including biomaterials, solar cells, photocatalysis, sensors [[1], [2], [3]]. Although the anodic formation of non-porous barrier oxide films on titanium has been known for decades [4,5], the revolutionary study of Zwilling et al. in 1999 that reported the possibility of production of self-organized nanoporous titania by anodization in fluoride containing electrolytes opened a new pathway; and numerous studies have been conducted on titanium anodization in fluoride containing electrolytes [6].

Porous anodization of titanium can be divided into three generations according to the fluoride sources and solvents of the electrolyte. First-generation titania nanotubes that are grown in acidic, HF containing aqueous electrolytes have a limited thickness of 500 nm because of the aggressive nature of the electrolyte [7,8]. Second generation titania nanotubes that are produced in buffered (neutral), aqueous solutions of fluoride salts also have a limited thickness up to several microns [9,10]. On the other hand, third generation titania nanotubes that are grown in neutral organic solutions may have a thickness in the range of hundreds of microns as the high viscosity of the organic electrolytes reduces ion migration, and results in lower local acidification and dissolution caused by fluoride ions [[11], [12], [13]]. Paulose et al. reported over 1000 μm thick titania nanotube film (longest reported) by anodization in ethylene glycol solution containing NH₄F and water [14]. Today, studies have been mainly focused on the anodization in these types of organic electrolytes, and with fine tuning of conditions, titania nanotubes with various sizes, shapes, and morphologies have been obtained [3,15]. Major anodization parameters are well-studied. In general, higher anodization potential results in larger tube diameter and higher growth rate [16]; increase in anodization time results in thicker films (especially in organic electrolytes) [17]; higher anodization temperature increases ion migration, thus growth rate of the oxide film, and also has a mild effect on tube diameter [11].

For the explanation of the formation of nanotubular structures on titanium, the field-assisted dissolution model that is developed initially for aluminum has been extensively used. According to this model, tubular film growth is controlled by the balance between field-assisted formation of oxide at the metal-oxide interface and the dissolution at the oxide-electrolyte interface (tube bottoms) [18]. Even though this growth model had been used for decades, in some cases, it was not adequate to explain abnormal volume expansions during titania nanotube formation. In recent years, a different growth model, the field-assisted flow, was proposed to consider this volume expansion. In this model, it is proposed that the volume expansion and resulting stress that occurred during field-assisted growth is relieved by the plastic flow of the oxide. According to this model, the oxide formed at the tube bottom moves upwards by the flow mechanism to form nanotube walls [19,20]. Besides, small sized fluoride ions migrate through the tube bottom oxide layer twice as fast as oxygen under the effect of the electric field and accumulate at the outer part of the oxide, forming a fluoride rich layer (FRL). This layer also moves upwards by the flow of the oxide [21]. The common opinion about the transition from nanopores to nanotubes is the dissolution tendency of this fluoride rich layer (FRL) located at the pore junctions (triple point) and outer part of the tube walls, which is also well-supported by the flow model [21].

Porous titania films' growth stages can be followed by current time behavior under constant anodization voltage. At stage I, a barrier oxide layer is formed immediately before pore initiation, and a sharp decay on the current is observed [22]. At stage II, fluoride ions in the electrolyte stimulate the pitting in the oxide layer (pore initiation), and an increase in the anodization current is seen as the resistance of the anodic film decrease [23]. At stage III; nanotubes start to grow from initiated pores, and the current goes into a gradual decline period as the nanotubular oxide film grows thicker [24]. Porous film thickness increases with time as more current passes through the system. A disordered top layer (nanograss), which is not favored in many applications because of its irregular and loose structure, may form under extended anodization duration [25]. Some studies showed that, by protecting the top barrier oxide layer, longer tubes without nanograss could be obtained, which is attributed to restricted chemical dissolution of the grown nanotubes by preventing contact with the electrolyte. Kim et al. and Song et al. obtained longer nanotubes without nanograss by retarding chemical dissolution with a prior thin protective rutile film grown on the specimen via polishing and heat-treatment, respectively [25,26]. Albu et al. achieved the same progress by coating a semi-permeable photoresist before anodization, allowing ion migration and tube growth but acting as a resistance to chemical etching and protecting the tube tops [27]. Some groups studied to suppress nanograss formation by tuning anodization conditions such as water content of the electrolyte [28], addition of organic additives [29], changing the viscosity of the electrolyte [30]. On the other hand, in several other studies nanograss structure is removed by ultrasonication [31] or stripping with scotch tape [32], which can severely damage nanotube films.

Kinetics of nanotube formation and growth has been investigated by using different techniques and approaches. Xie et al. investigated the effect of fluoride concentration, temperature, and applied potential on nanotube formation. They suggested a model for simulating optimal growth conditions (fluoride concentration and temperature) for nanotube formation [33]. Butail et al. investigated nanotube growth rate versus temperature and voltage. They found that nanotube growth rate increases exponentially with temperature under constant voltages. The results supported that nanotube growth is thermally activated and obeys the Arrhenius equation [34]. Cortes et al. suggested a mathematical model for the growth of titania nanotubes with variables including voltage, fluoride and water concentrations, and time; and tried to verify the model with the experimental data taken from the literature. Even though the model showed consistency with some data sets, it gave a very high error ratio in others [35]. Zhang et al. conducted a systematic study for investigating time dependent final nanotube lengths at different temperatures and voltages, and observed a linear dependence at all anodization temperatures. They also reported that temperature dependence of the nanotube lengths follows the Arrhenius relationship, with a linear activation energy- anodization potential relation [36]. Suliali et al. studied anodization reaction kinetics and produced a model for simulating anodic current during anodization. They tested and verified the model at different anodization potentials and different NH₄F concentrations. They reported that anodic current increases with anodiation potential and NH₄F concentration, and it is the integral function of the water quantity during anodization [37].

Even though important findings have been obtained in the studies summarized above, a kinetic model-approach is still lacking that combines the growth process with the final morphologies of the different nanotubular structures. Although nanotube growth under the initial barrier layer was suggested in an early work [24] and mentioned in several studies [15,38], it has not been addressed in kinetic studies. On the other hand, it has been well observed that extended anodization durations cause irregular nanograss structure. However, attempts have been mainly focused on suppressing the nanograss formation [[25], [26], [27], [28], [29], [30]] or removing the so-formed nanograss structure [31,32], instead of investigating the formation kinetics and controlling the process. In several works, the initial barrier oxide layer, open tube-top nanotubes, and nanograss structures were shown in the same study; but they were interpreted as different morphologies caused by different anodization conditions, instead of different stages of the growth process [17,39]. In most of the kinetic studies, the final length of the nanotubes have been used as the main parameter for defining kinetics of the growth process [[33], [34], [35], [36]], and different characteristics of growth stages have not been considered. Therefore, kinetic models for nanotube growth have not provided consistent results that always satisfy the experimental findings.

In this study, we propose a new approach for determining the formation and growth kinetics of different titania structures during anodization in ethylene glycol based electrolytes. Accordingly, we divide the growth process into three stages (I. Growth under initial barrier layer, II. Nanopore to nanotube transition, III. Nanograss formation) and evaluate the anodization stage dependent nanotube growth kinetics. As will be explained and verified later, the reason for this division is the inability of the application of the same growth mechanism to different stages of the growth. At the first stage, the nanotube growth kinetics under chemically and gradually dissolving protective barrier oxide are investigated, employing total charge passed through the system during anodization. At the second stage, after the accomplishment of the chemical dissolution of the top barrier layer, a well-defined nanoporous structure appears. This structure is stable for a short period of time and converts into nanotubes. At the third stage, nanograss formation starts on the top of the nanotubes. For the determination of the growth kinetics of nanotubes and conversion of them into nanograss, the simultaneous influence of field-assisted nanotube growth and chemical dissolution of the tube tops (shortening of the tubes) are taken into account. Morphology transitions during anodization and growth kinetics at the proposed stages are investigated with experimental studies and verified to be compatible with the suggested model. This model approach is produced based on determining chemical dissolution kinetics of the initial barrier oxide layer formed on the nanotubes, which is a temperature-dependent process and found to fit well with the Arrhenius equation. By using this approach, it is shown that morphology (barrier top, open tube-top or nanograss) and thickness of the nanotube can be estimated under modeled condition; and desired morphology and nanotube thickness can be obtained by tuning anodization temperature and duration.