Plasma Nitriding of TiO2 Nanotubes: N-Doping in Situ Investigations Using XPS.

The nitrogen doping of titanium dioxide nanotubes (TiO2 NTs) was investigated as a result of well-controlled plasma nitriding of TiO2 NTs at a low temperature. This way of nitrogen doping is proposed as an alternative to chemical/electrochemical methods. The plasma nitriding process was performed in a preparation chamber connected to an X-ray photoelectron spectroscopy (XPS) spectrometer, and the nitrogen-doped TiO2 NTs were next investigated in situ by XPS in the same ultrahigh vacuum (UHV) system. The collected high-resolution (HR) XPS spectra of N 1s, Ti 2p, O 1s, C 1s, and valence band (VB) revealed the formation of chemical bonds between titanium, nitrogen, and oxygen atoms as substitutional or interstitial species. Moreover, the results provided a characterization of the electronic states of N–TiO2 NTs generated by various plasma nitriding and annealing treatments. The VB XPS spectrum showed a reduction in the TiO2 band gap of about 0.6 eV for optimal nitriding and heat-treated conditions. The TiO2 NTs annealed at 450 or 650 °C in air (ex situ) and nitrided under UHV conditions were used as reference materials to check the formation of Ti–N bonds in the TiO2 lattice with a well-defined structure (anatase or a mixture of anatase and rutile). Scanning electron microscopy microscopic observations of the received materials were used to evaluate the morphology of the TiO2 NTs after each step of the nitriding and annealing treatments.


■ INTRODUCTION
Titanium dioxide (TiO 2 ) is widely used as a photocatalyst because of its high chemical stability (chemical inertness), low cost (earth abundance), and photocatalytic activity under UV light excitation and it is widely used in many applications such as the decomposition of organic pollutants, selective oxidation, hydrogen evolution, solar cells, and others. 1−4 In comparison with other forms of nanostructured TiO 2 materials, nanotubes (NTs) are attractive candidates as photocatalytic materials because of their strong light-scattering effects and high surfaceto-volume ratio. 5,6 These highly ordered, vertically oriented tubular structures (with a specific crystal structure) feature a high degree of electron mobility along the tube axis 7 perpendicular to the titanium substrate, which greatly reduces interface recombination as a result of structural distortion at the bottom cap region of the NTs (mixed valence states of Ti: Ti 3+ and Ti 4+ ). 8 Despite these numerous advantages, the photocatalytic activity of TiO 2 is limited to UV radiation because of its large band gap (∼3.2 eV for anatase and ∼3 eV for rutile). This indicates that only about 4−7% of the solar spectrum can be absorbed. 9 Therefore, electron−hole pair generation can only be achieved by UV light irradiation (λ < 380 nm), and much current research aimed to reduce the TiO 2 band gap by doping or band gap engineering. 9−11 Usually, doping with different types of transition metal cations, surface modification with noble metals, or doping with nonmetal anions are used to increase the visible light absorption or suppress the recombination of photogenerated carriers. 11−13 The incorporation of nitrogen atoms into the TiO 2 structure is still one of the most popular and effective methods for enhancing the photocatalytic performance. 14,15 This process can be successfully implemented, among others, by plasma nitriding, in particular of nanostructures. Lin and co-authors showed the possibility of using plasma treatment to surface functionalization of various materials including TiO 2 for different applications. 16 Therefore, under optimal conditions of doping processes, oxygen atoms in the TiO 2 lattice are   replaced by nitrogen ions. After this occurs, it seems that the new localized N 2p states, which are located above the valence band (VB) (a mixture of p states of the nonmetal dopant and the O 2p states of the TiO 2 structures modified by substitutional or interstitial doping), narrow the band gap of TiO 2 and finally lead to a shift in optical response toward the visible range. 10,15,17,18 Such a phenomenon has been noted by Asahi and Morikawa, who showed the presence of nitrogen atoms as substitutional or interstitial dopants 19,20 based on first principle calculations and experimental X-ray photoelectron spectroscopy (XPS) spectra for the N complex species introduced into TiO 2 . Table 1 shows some examples of XPS investigations of TiO 2 NTs after various methods of nitrogen doping, where an N 1s peak in the range of 396−404 eV was usually observed. 21 As shown in the table, the XPS data for the N−TiO 2 NT system are still under debate.
The literature data presented suggest that different strategies based on chemical, electrochemical, and physical processes of incorporating nitrogen into a TiO 2 NT lattice lead to the following observations: 21 (1) nitrogen peaks at 396.0−398.0 eV are attributed to substitutional nitrogen, where the typical binding energy (BE) for Ti in a TiN compound appears (structural effectphotosensitive element, light absorber); (2) usually, peaks at 399.0−403.0 eV are assigned to chemisorbed nitrogen (a surface effect) or interstitial/ substitutional nitrogen depending on the TiO 2 functionalization methods (a structural effectphotosensitive element, light absorber).
Some of these results are unusual and differ significantly from the source data. 34 Moreover, the authors of the references cited in Table 1 postulated in all cases that the N-doped TiO 2 NTs exhibited higher photoelectrochemical or photoelectrocatalytic activity than the undoped TiO 2 NTs. This is probably related to the surface modification of this type of nanostructures, where locally catalytically active sites may be formed because of an increase in the amount of molecularly or chemisorbed nitrogen (surface effect). Moreover, the surface of the TiO 2 NTs has a hydrophilic character, 35,36 which improves the photocatalytic properties through the formation of reactive oxygen species on the TiO 2 NTs. 36 Another factor favoring the photocatalytic activity of such systems are the geometrical parameters of the TiO 2 NTs, such as length, average pore size, and wall thickness, which are determined by the formation conditions of such oxide nanostructures (the type of electrolyte, anodic voltage, and anodization time). Mazierski and co-authors showed some correlation between the degradation efficiency of phenol under UV−vis light and preparation conditions of the NTs, where the photoactivity of TiO 2 NTs increased with increasing anodization voltage and time in the range of 20−40 V and 30−60 min, respectively. 15 Another geometrical factor that can significantly affect the photoactivity of TiO 2 NTs is their length. Macak et al. showed that the photoresponse in ultraviolet and visible light is strongly dependent on this parameter. An increased TiO 2 NT length from 500 nm to 6.1 μm shows a drastic decrease of the UV response, while a significant increase of the visible response was obtained. 17 The next parameters having an important impact on photocatalytic activity are the amount of nitrogen incorporated into the TiO 2 NTs 15 and crystallinity of the annealed tubes at 450 or 500°C (formation of the anatase phase). 9,18,22 All these factors, including a high geometrical surface area and axial charge transport along the tubes, are crucial here. 28 Therefore, these parameters may be significant to the efficiency of the light absorption process, which is particularly crucial if we would like to use TiO 2 NTs as a photoanode for the photoelectrochemical decomposition of water, as Fujishima and Honda reported for the first time in 1972. 37 At present, despite the fact that photoelectrochemical decomposition of water is a promising method for generating hydrogen, inexpensive and environmentally friendly materials based on TiO 2 structures are still being sought. When designing new photoanodes based on TiO 2 , it should be remembered that the effective photocatalytic and photoelectrocatalytic performance of such materials is limited by a number of factors, including band gap energy, the position of the conduction band and VB, activation energy, and the recombination process.
Therefore, in this work, we present experimental XPS data related to this point based on new strategies for doping TiO 2 NTs using a low-temperature plasma nitriding process. For this purpose, a preparative vacuum chamber equipped with an N 2 plasma source was used, which was connected to the XPS surface analysis system. This treatment guarantees the purity of the process and its repeatability in relation to chemical/electrochemical methods, which are commonly used for nitrogen incorporation in TiO 2 . Thus, titanium oxide NTs are particularly suitable for such investigations because the strictly controlled electrochemical procedures make it possible to produce nanotubular substrates having a well-defined geometry, uniform chemical composition, and crystalline structure that can be controlled by heat treatment. 10,35 ■ RESULTS AND DISCUSSION The formation of titanium oxide NTs perpendicular to the substrate is well confirmed by the scanning electron microscopy (SEM) images ( Figure 1). They exhibit an average diameter of ∼110 nm and a height of ∼1 μm (Figure 1a). 35,38,39 The NTs are open at the top but closed at the bottom (Figure 1b,c). The bottom surface of the TiO 2 NT layer ( Figure 1b) is characterized by a series of regularly spaced "bumps" forming a regular pattern, where the individual NT bottom is clearly

ACS Omega
http://pubs.acs.org/journal/acsodf Article visible. The SEM observations clearly show the bottom layer, which acts as a thin barrier between the Ti substrate and the tubular structure. After the mechanical removal of the nanoporous oxide layer from the Ti substrate, one can distinguish the areas of NT growth (Figure 1a). The shape of these areas corresponds to the shape of the bottoms of the NTs. This phenomenon is related to the mechanism of the formation and growth of self-organized TiO 2 NTs in electrolytes containing fluorides, which has been described in detail in many works. 3,4,10,15 Usually, the nanoporous oxides obtained by anodic oxidation were found to be amorphous in structure, and so, the TiO 2 NTs obtained were annealed in air at 450 or 650°C in order to transform the amorphous structures to their crystalline phases: anatase and a mixture of anatase and rutile, respectively. This may significantly affect certain properties of the oxide layers such as mechanical stability, 40,41 electronic properties, 8 and wettability. 35,42 The heat treatment did not cause any distinct changes in the diameter or shape of the TiO 2 NTs (see Figure 1d) but can modify the thickness of the TiO 2 nanoporous layer, 35 as discussed below. The chemical state of the titanium oxide NTs after anodization was determined by XPS measurements. Table 2 shows the photoelectron BEs of Ti 2p 3/2 and O 1s signals for the sample received at 25 V. The XPS spectrum of Ti 2p exhibits two dominant peaks located at 458.8 and 464.5 eV, which correspond to the Ti 2p 3/2 and Ti 2p 1/2 signals characteristic for the Ti 4+ state of titanium. 43 The main peak at 530.3 eV for the O 1s spectra comes from the Ti−O−Ti lattice. The calculated atomic ratio O 530.2 /Ti 458.8 was 1.93, which suggests that the stoichiometry of titanium dioxide at the surface region was disturbed by fluorine atoms coming from the electrolyte and carbon-like contaminations originating from the air environment. 41 To confirm the chemical state of the titanium, we also determined the modified Auger parameter, which is defined as α′ = E b (Ti 2p 3/2 ) + E k (Ti LMM), where E b and E k are the binding and kinetic energies of the dominant core electron and Auger electron lines for a particular element, respectively. 44 The Auger parameter is a valuable tool when assigning chemical states for a wide variety of surface species. 44 The determined α′ parameter = 873.1 was found to be close to the literature values of 872.9 and 873.0, 45−47 which indicates that stoichiometric titanium oxide is formed during the anodic oxidation process (see Table 2).
Subsequently, the TiO 2 NTs were subjected to a nitriding process under vacuum conditions, directly after the anodization process. The surface morphology of the tubes is shown in Figure  2 at low (a) and high magnification (b). The nitriding process did not change the shape or diameter of the TiO 2 NTs. The SEM images at low and high magnifications show the typical morphology of such materials (hollow cylinders). The nanopores in TiO 2 form a free-standing array of NTs, which exhibits a tendency to form a hexagonal, closely packed structure. 4,35 The elemental composition of the N-doped TiO 2 NTs after anodic oxidation and plasma nitriding was determined by XPS. Figure 3 shows HR XPS spectra of the C 1s, O 1s, Ti 2p, and N  (Figure 3d). The main signal at 400.2 eV can be attributed to chemisorbed nitrogen. The second peak located at 399.0 eV is probably due to the substitution of oxygen with nitrogen in the structure of the titanium dioxide, and the last peak at 401.2 eV can be assigned to the appearance of NO x impurities in the TiO 2 lattice, as suggested by data in the literature. [18][19][20][21]32,48,49 The XPS data collected for the unheated NTs were also confirmed by experimental investigations of a pure Ti foil covered with a passive oxide film after the same nitrogen plasma treatment (see Table 3). Deconvoluted Ti 2p spectra ( Figure  4a) showed a contribution by typical titanium oxides formed in air on the Ti surface: Ti 4+ (458.8 eV), Ti 3+ (457.7 eV), and Ti 2+ (454.8 eV) 50 in addition to nitrides (454.8, overlap peak) and oxynitrides (456.5 eV). 32 The N 1s region (Figure 4b) exhibits three singlets at 396.7, 399.2, and 400.4 eV, which correspond to the TiN, oxynitrides (substitutional nitrogen: Ti−O−N), and chemisorbed nitrogen, respectively. This indicates that Ti−N and Ti−O−N species may form under the proposed nitriding conditions. Sait and co-workers also showed the presence of these kinds of nitrogen compounds in TiO 2 , which were functionalized by the plasma-enhanced chemical vapor deposition method. 51 They observed three states of nitrogen in the N 1s spectra within the BE region of 400.0−396.6 eV. The first one (identified at 400.0 eV) was attributed to chemisorbed nitrogen, the second (at 397.8 eV) to Ti−N bonds, and the last (at 396.6 eV) to the titanium oxynitrides. 51 Some divergence in the location of N 1s peaks are due to the fact that the authors of this work used the nitriding process based on chemical reactions,  Figures 3d and 4b, respectively) clearly shows the C−N bound species to be one of the main chemical nitrogen states of the N 1s spectra. This is the case even though the nitriding process and XPS analysis of both samples were carried out in situ under vacuum conditions (10 −5 mbar). Macak et al. noticed that such behavior may be caused by an unintentional carbon doping of the TiO 2 NTs during their formation in glycerol-based electrolytes. 17 This observation was confirmed by the in-depth distribution of elements (nitrogen, oxygen, titanium, carbon, and fluorine) within the near surface zone of the nitrogen-doped NTs (Figure 5a). A careful inspection of these profiles revealed the relative in-depth distributions of nitrogen, carbon, and fluorine to be similar. The maximum concentration of these elements corresponds to the surface zone of the NTs, whereas their in-depth concentrations stabilize at a constant level, which indicates that carbon and fluorine dope the nanoporous TiO 2 structure during the anodic oxidation process, whereas nitrogen does so during the plasma treatment. It is also important that the chemical state of the main nitrogen N 1s peak does not change with the depth and is close to a BE of 400.0 eV (see Figure 5b).
A similar result was obtained for the NTs annealed at 450°C and then nitrogen plasma-treated ( Figure 6). However, no F 1s signal was registered because the fluorine compounds thermally decompose during the annealing process, which is consistent with the observations reported elsewhere. 35,41 It was also noticed that the concentration of nitrogen for the TiO 2 NTs annealed at 450 and 650°C in the near-surface layer is lower than that for the nonheated NTs (Table 3). This is related to a change in the NT structure during annealing from the amorphous to the crystalline phase. Obviously, the defective structure of the amorphous phase favors an easier penetration by the nitrogen atoms into the bulk of the titanium oxide than does the crystal lattice of anatase or rutile. Thus, the substitution of oxygen by nitrogen atoms is much more effective in the pristine NTs.
Similar features of Ti 2p, N 1s, and O 1s XPS spectra for unheated NTs and those annealed at 450°C (ex situ) were reported by Siuzdak et al., who applied the electrochemical method, 24 and Macak et al., who used chemical and ion implantation methods 17,30 for doping a TiO 2 NT lattice. Such a result is associated with the hypothesis that after the introduction of nitrogen into the TiO 2 structure by a process of substitution at the sites of the oxygen atoms, Ti−N−O bonds are formed, 19,20 which are usually manifested in shifts of the N1s peak (400 eV, chemisorbed nitrogen, organic matrix) toward lower BEs. The BE peak position at 399.0 eV is larger than that of typical representative Ti−N bonds and is probably related to a reduction in nitrogen electron density because of the high electronegativity of oxygen, as postulated by Cong et al. 48 and Jiang et al. 32 The signal detected at 401.2 eV (unheated TiO 2 NT) is usually ascribed to a generic interstitial site in the TiO 2 lattice, where N−O bonds play a crucial role as dopants. 19,20,24,32 A definitely different result was obtained for the sample heated at 650°C (ex situ), where only a nitrogen signal at 399.8 eV was recorded, which probably corresponds to the chemisorbed nitrogen at the surface of the TiO 2 NTs (Figure 7). This result also suggests that it is much more difficult to introduce nitrogen into a crystal lattice of titanium oxide (a mixture of anatase and

ACS Omega
http://pubs.acs.org/journal/acsodf Article rutile 35,40 ), which also shows a decrease in the N/Ti atomic ratio (Table 3). This is only a surface effect which does not change the electronic properties of NTs modified by the nitrogen plasma treatment. The XPS results for this group of samples (reference materials) are summarized in Table 3 and Figure 7. Therefore, in the further part of this work, we focus on the functionalization of TiO 2 NTs using a plasma nitriding process combined with annealing at 450°C. The following experimental procedure was used: (1) plasma nitriding of TiO 2 NT (asprepared), (2) annealing at 450°C/2 h under ultrahigh vacuum (UHV) conditions, (3) plasma nitriding of the annealed TiO 2 NT at 450°C, and (4) annealing at 450°C/2 h under UHV conditions. All the steps of this procedure were monitored using the in situ XPS method. As a result of the applied procedure, clear shifts in the N 1s peak were observed after each step (see Figure 8). Usually, after plasma treatment alone, the maximum peak position of the nitrogen was close to 400.0 eV, but after heat treatment, negative shifts in BE were recorded, which is consistent with our previous observations (this work) and data in the literature. 17,18,21,24,30,32,33,48,49 The negative shift in the nitride direction (∼397.0 eV) is crucial as an effective way of incorporating nitrogen into the TiO 2 lattice. This phenomenon also caused a change in the modified Auger parameter (α′) to a lower value: 872.5 (459.0 − Ti 2p 3/2 + 413.5 − L 3 M 23 M 45 ),

ACS Omega
http://pubs.acs.org/journal/acsodf Article which suggests the formation of certain chemical bonds between titanium, nitrogen, and oxygen. Our results indicate that plasma nitriding successfully managed to introduce nitrogen into the structure of the NTs at an atomic level of 1%, of which 50% was titanium nitride using a single-stage nitriding process in combination with annealing at 450°C. In the case of the double-nitrided NTs, the amount of nitrogen increases almost twofold for the sample N 2 + 450°C + N 2 and by a factor of 1.5 for the sample N 2 + 450°C + N 2 + 450°C. The proposed method leads to substitutional N or NO x doping and interstitial N or NO x doping (see Figure 8).
The incorporation of nitrogen into the TiO 2 network induces specific changes in the electronic structure of this material, as can be observed in the VB XPS spectra. Figure 9 shows VB photoemission spectra before and after the plasma nitriding procedure with heat treatment at 450°C. The VB of TiO 2 is characterized by two broad peaks at ∼5 and ∼7 eV, related to the π and σ bonding orbitals, which are of a predominantly O 2p character. 45,52,53 Looking at the results, it is clearly visible that the VB band is shifted toward lower BE energies immediately after the nitriding process in comparison with the sample directly after anodization. The in situ heat treatment of the nitrided samples leads to a change in the shape of the VB spectra. These changes suggest that the combination of the two procedures (nitriding and annealing) provides a lasting effect in changing the electronic structure of titanium dioxide. Therefore, on the basis of these spectra, the band gap reduction of the functionalized TiO 2 NTs was determined by a linear extrapolation of the low BE VB emission edge. 54,55 The extracted values differ from each other in relation to those of the pristine TiO 2 NTs. In particular, the double nitriding process with annealing treatment caused a narrowing in the band gap energy of about 0.6 eV. The band gap reduction in the TiO 2 NTs with N 2 may be associated with a change in the ratio of nitrogen atoms to oxygen vacancies, which occurs during the nitriding process. 56 The Fermi level is then achieved within the gap between the hybridized N 2p−Ti 3d−O 2p states and the new N 2p−Ti 3d impurity states, narrowing the band gap value of the TiO 2 . 56 This indicates that the observed effects are related to structural changes in the titanium oxide lattice. It seems to be possible to obtain nitrogen in two forms: substitutional or interstitial, which may affect the photocatalytic properties of the TiO 2 NTs (compare Figure 8), depending on whether a onestage or two-stage plasma nitriding procedure is used.

■ CONCLUSIONS
The use of plasma nitriding has proven to be an alternative solution for incorporating nitrogen atoms into the TiO 2 NT lattice with respect to chemical/electrochemical methods. This process leads to structural changes in the TiO 2     Such behavior manifests in a shift of about 0.6 eV in the edge of the maximum energy for VB spectrum. In addition, our XPS results confirm that it is definitely easier to functionalize the amorphous structure of titanium dioxide with nitrogen than crystalline structures, as manifested in a decrease in the N/Ti atomic ratio with increasing annealing temperature (see Table  3). Of the much available literature, it seems there is no observable effect of NT geometry (diameter and wall thickness of NTs) on the formation of nitrogen bonds in the TiO 2 lattice. Therefore, controlled substitutional or interstitial doping in a  • TiO 2 NTs were fabricated in a one-step anodization process of Ti foil (0.25 mm thick and 99.5% + % purity Alfa Aesar) at 25 V for 3 h in an optimized electrolyte based on a mixture of glycerol and water (volume ratio 50:50) containing 0.27 M ammonium fluoride (NH 4 F) using a two-electrode electrochemical system. After anodization, the samples were rinsed with deionized water (24 h) and dried in air. Subsequently, thermal annealing in air (ex situ) was performed at 450 or 650°C for 3 h (reference materials). • Plasma nitriding was carried out inside an UHV preparation chamber (PREVAC, Rogoẃ, Poland) at 1− 1.5 × 10 −5 mbar pressure of flowing nitrogen. The N plasma was generated by a SPECS PCS-ECR plasma cracker source with a plasma power of 60 W and microwaves at a frequency of 2.45 GHz. The temperature of the plasma nitriding process was about 40°C. After a 4 h nitriding process, some of the samples were annealed in vacuum conditions (10 −8 to 10 −9 mbar) at 450°C for 2 h. Such functionalized samples were then transferred in situ to the XPS analysis chamber for chemical surface analysis.
A schematic diagram of the plasma nitriding process with heat treatment is shown in Figure 10.
Surface and Structure Characterization.
• The in situ XPS measurements were performed using a PHI 5000 VersaProbe (ULVAC-PHI) spectrometer with monochromatic Al Kα radiation (hν = 1486.6 eV) from an X-ray source operating at a 100 μm spot size, 25 W and 15 kV. The HR XPS spectra were collected with the hemispherical analyzer at a pass energy of 23.5 eV and an energy step size of 0.1 eV. The angle of incidence of the Xray beam at the sample surface was 45°. Avantage software (version 4.88) was used to evaluate the XPS data. The deconvolution of all the HR XPS spectra was performed using a smart-type background and a Gaussian peak shape with a 35% Lorentzian character. The measured BEs were corrected in reference to the energy of adventitious carbonC 1s peak at 285.0 eV. The recorded survey scans were also used to determine the modified Auger parameters. Therefore, an Ar + ion gun (EX05Microlab 350) was used to measure the composition depth profiles of the TiO 2 NTs modified by the plasma nitriding process.
The following parameters of ion etching were applied: 60 s etching steps, ion energy 2 keV, beam current 1.3 μA, crater size 36 mm 2 . The XPS spectra were excited at an energy of 1486.6 eV and recorded after each sputtering period. Separate tests showed that the sputtering rate was 0.0225 nm/s, which was determined by ion etching a silicon sample covered by a SiO 2 layer of known thickness. • An SEM microscope (FEI Nova NanoSEM 450) was used for the morphological characterization of the samples after their anodization, heat treatment, and functionalization in the vacuum preparation chamber (UHV system).
These experiments were carried out using the through lens detector of secondary electrons at a primary beam energy of 10 kV under a high vacuum (pressure 10 −6 mbar). SEM images were obtained at a long scan acquisition time of typically 30 s per frame after the inspection region was chosen.