Titanium oxynitride microspheres with the rock-salt structure for use as visible-light photocatalysts

Novel photocatalysts (TiO2@TiO1 xNx) with the core–shell geometry were synthesized by controlled nitridation of TiO2 microspheres using ammonia gas. The oxynitride phases (TiO1 xNx) with a cubic rocksalt structure are exclusively formed on the surface of the TiO2 microspheres while the cores of the microspheres retain the TiO2 phase without nitrogen doping. Various spectroscopic data confirm the formation of the core–shell structure, denoted as TiO2@TiO1 xNx. The TiO2@TiO1 xNx materials showed superior photocatalytic activities for the decomposition of methylene blue as well as the generation of photocurrent under visible-light. It is evident that the TiO1 xNx phase is the key element to induce the photocatalytic activity. Specifically, the partial doping of nitrogen into the TiO matrix is crucial for visiblelight absorption.


Introduction
Titanium dioxide (TiO 2 ) has been extensively studied because of its wide applications in solar-driven environmental and energy technologies since Honda and Fujishima discovered water splitting on a TiO 2 electrode under UV light. 1 TiO 2 has several merits as a photocatalyst including large abundance, low toxicity, and excellent stability. 2,3 However, a major drawback is that it absorbs only UV light due to its large band gap energy (3.0 and 3.2 eV for the rutile and anatase phases, respectively). 4,5 A wide range of approaches have been explored to narrow the band gap of TiO 2 . Chen et al. reported that disordered TiO 2 nanoparticles (NPs) prepared by the hydrogenation method exhibit signicantly enhanced activity in the production of hydrogen (H 2 ) from water under visible-light. 6 The chemical composition of TiO 2 was also modied by doping nonmetal atoms such as carbon (C), hydrogen (H), nitrogen (N), uorine (F), and sulfur (S) into the oxygen (O) site. [7][8][9][10][11][12][13] The dopants typically generate impurity states above the valence band of TiO 2 , which results in the upshi of the valence band edge and the improvement in the solar absorption efficiencies of TiO 2 . In particular, N-doped TiO 2 (nominally TiO 2Àx N x ) materials with low N content ($1 wt%) have been thoroughly studied. 10,14 They are typically prepared by annealing TiO 2 at temperatures below 600 C under an ammonia (NH 3 ) atmosphere, which exhibit excellent catalytic activities in the visible light range. But their absorption is not fully covered in the solar spectrum. Moreover, TiO 2Àx N x is unstable aer going through the photocatalytic reaction and easily returns to parent TiO 2 .
Despite the extensive interest in N-doping in the TiO 2 matrix, there are few studies on N-doped titanium monoxide (TiO) materials with a cubic rock-salt structure due to the difficulty to control the doping level and the metallic character of TiO. 15, 16 Simon et al. reported the synthesis of N-doped TiO NPs using laser pyrolysis, which shows a large shi of the absorption threshold in the visible region. 17 Similar to TiO, titanium nitride (TiN) also adopts a cubic rock-salt structure and is considered to be metallic. In contrast to bulk TiN materials, TiN lms prepared by using a cathodic arc technique have a band gap of about 2.0 eV, indicating that TiN can be semiconducting if the particle size reduces to the nanometer scale. 18 Very recently, Zheng et al. prepared TiO x N y @TiN composites through the ash oxidation of commercial TiN particles, demonstrating a high photocatalytic activity for H 2 production under visible light irradiation. 19 However, the non-equilibrium quick oxidation of TiN particles in the presence of 2,4,6-trinitrophenol hampers the control of the optimal doping level.
Here we report a convenient and reproducible method to make a new visible-light active composite photocatalyst. Controlled nitridation of monodisperse TiO 2 microspheres, accomplished by owing ammonia gas, converted the shell into the TiO 1Àx N x phase and transformed the core into the amorphous TiO 2 phase. As a result, the nitrided product has the core-shell type structure, which is composed of two phases: a crystalline TiO 1Àx N x phase as the shell and an amorphous TiO 2 phase as the core. Hereaer, we denote it as TiO 2 @ TiO 1Àx N x . The TiO 2 @TiO 1Àx N x material exhibited excellent photocatalytic activities under visible-light, including the degradation of methylene blue and the production of photocurrent. The nitrogen doping in TiO 1Àx N x is considered to be responsible for absorbing visible light.

Synthesis of monodisperse TiO 2 microspheres
Monodisperse TiO 2 microspheres were prepared using TTIP as the titanium source. In brief, 3 mL of DDA, 4 mL of deionized water, and 300 mL of ethanol were added to a round-bottom ask, and the solution was cooled to À20 C and stirred for 1 h. Using a syringe, a solution containing 6 mL of TTIP and 8 mL of ethanol was injected into the cold solution. Aer vigorous stirring for 20 h at À20 C, white precipitates that formed gradually were separated from the solution by centrifugation. The precipitates were sintered in air at 500 C for 6 h, yielding 1.44 g of TiO 2 powders (89.2% yield based on TTIP). The powders are composed of evenly dispersed spherical microspheres and their average diameter is approximately 560 nm.

Synthesis of TiO 2 @TiO 1Àx N x microspheres
The TiO 2 microspheres were annealed in NH 3 at 700 C to yield the desired TiO 2 @TiO 1Àx N x microspheres. The ow of ammonia gas was controlled using a mass ow controller. The ow rate is xed at 300 standard cubic centimeters per minute. Six different samples were prepared by controlling the annealing time. The annealing was performed at 700 C for 1, 4, 7, 10, 24, and 48 h. The white microspheres gradually became blue when the sample was annealed under the ammonia atmosphere.

Characterization
Powder X-ray diffraction (XRD) data were collected using a Rigaku DMAX 2500 diffractometer (Cu Ka; Rigaku, Japan) operating at 40 kV and 150 mA. High resolution transmission electron microscopy (TEM) was performed using a JEOL JEM-2100F microscope (JEOL, Japan). Specimens for the TEM examinations were prepared by dispersing nely ground powders of the samples in anhydrous ethanol and then allowing a drop of the suspension to evaporate on a 400 mesh carboncoated grid. High resolution scanning electron microscopy (SEM) analyses were carried out using a Hitachi S-5500 microscope (Hitachi, Tokyo, Japan). Samples for the SEM analyses were prepared by dropping diluted samples in anhydrous ethanol on a lacey support grid. The samples were also subjected to chemical microanalyses using an Oxford Instruments INCA TEM 300 system (Oxford Instruments, Abingdon, UK) for energy dispersive X-ray (EDX) analysis. UV-visible absorption spectra of methylene blue (MB) solutions were recorded using a Perkin Elmer Lambda 950 spectrometer. The UV-visible absorption spectra of the powders were measured using an integrating sphere accessory by the diffuse reectance method. Raman spectra were obtained at 25 C using a LabRam HR Raman spectrometer (Horiba Jobin-Yvon) equipped with a liquid-nitrogen-cooled CCD multichannel detector. A 514 nm Ar-ion laser was used as the excitation source. Photoluminescence spectra were measured on a Hitachi F-7000 uorescence spectrophotometer. Thermal gravimetric analysis was carried out using a TGA 2050 instrument (TA Instruments). The sample was placed on a platinum pan for each run. The data were collected in air from 25 C to 700 C at a rate of 5 C min À1 . Adsorption and desorption measurements were carried out at 77 K using an ASAP 2420 instrument (Micromeritics, USA) with nitrogen as the adsorptive gas. The Brunauer-Emmett-Teller (BET) surface areas were calculated using P/P 0 ¼ 0.05-0.3 from the adsorption curve using the BET equation. The poresize distributions were obtained from the desorption curve using the density functional theory method. Prior to each sorption measurement, the sample was out-gassed at 300 C for 24 h in vacuo to completely remove the impurities. To investigate the elemental compositions, X-ray photoelectron spectroscopy (XPS; Theta probe AR-XPS System, Thermo Fisher Scientic, UK) analysis using a mono-chromated Al Ka X-ray source (hn ¼ 1486.6 eV) was performed at the Korea Basic Science Institute (KBSI) in Busan. The nitrogen contents of the TiO 2 @TiO 1Àx N x samples were analyzed using inductively coupled plasma atomic emission spectroscopy (ICP-AES, JY Ultima2C) at KBSI in Seoul.

Results and discussion
Monodisperse titanium oxynitride (TiO 2 @TiO 1Àx N x ) microspheres were prepared via two main pathways: (1) sol-gel synthesis of TiO 2 microspheres and (2) nitridation of the TiO 2 microspheres in NH 3 . Uniform TiO 2 microspheres were prepared with solutions of TTIP, DDA, deionized water, and ethanol. The solution temperature was maintained at À20 C during the sol-gel polymerization process to make TiO 2 microspheres with virtually identical diameters as well as to prevent agglomeration. Aer separating white precipitates from the solution, white precipitates were then heated at 500 C for 6 h, yielding white crystalline TiO 2 microspheres with an average diameter of 560 nm. The TiO 2 microspheres were then nitrided at 700 C in an NH 3 atmosphere for different periods of time. The white TiO 2 microspheres turned blue to deep blue, depending on the annealing time. This color change implies the formation of the rock-salt TiN 1Àx O x phase in the TiO 2 microsphere.
As illustrated in Fig. 1, the TEM and SEM images of representative TiO 2 @TiO 1Àx N x microspheres clearly reveal that the size of parent TiO 2 microspheres has noticeably shrunk but the spherical shape appears to be remained. The results implicate that the annealing is effective mostly on the surface of the TiO 2 microspheres at 700 C. The average diameter of TiO 2 @ TiO 1Àx N x microspheres nitrided for 24 h was reduced to about 420 nm, which was estimated from the TEM images of TiO 2 @TiO 1Àx N x in Fig. 1. However, the spherical shape has almost collapsed when the sample was nitrided for 24 h at 900 C ( Fig. S1 †), indicating that annealing at high temperatures completely converts the oxide phase into the nitride phase.
We rst investigated the structures of the TiO 2 @TiO 1Àx N x microspheres with powder X-ray diffraction (XRD). The XRD data, given in Fig. 2, show that the anatase TiO 2 phase is gradually transformed into TiN 1Àx O x with a cubic rock-salt structure as the nitridation time increases. The sample nitrided at 700 C for 4 h shows that two new XRD peaks at about 37.2 and 43.2 , which can be assigned to (111) and (200) reections in TiO 1Àx N x , appear along with the TiO 2 peaks. 20,21 This indicates that the anatase TiO 2 phase coexisted with the cubic TiO 1Àx N x phase for a short period of time. When the nitridation was prolonged to 10 h, the two peaks corresponding to cubic TiO 1Àx N x became more distinctive in the XRD data while other peaks from the anatase TiO 2 phase almost completely disappeared. This suggests that the cubic TiO 1Àx N x phase is predominantly formed on the surface of the TiO 2 microspheres nitrided for longer than 10 h. The average crystallite size of TiO 1Àx N x synthesized by nitriding the TiO 2 sample for 24 h, determined from half-peak widths of the XRD peaks by applying the Scherrer equation, is 17.1 nm. Lattice parameters were calculated using the (111) and (200) reections in the XRD data. Table S1 † lists the calculated lattice parameters for TiO 1Àx N x , along with the TiN and TiO standard data from the JCPDS le. Lattice constants of TiO 1Àx N x nitrided for 24 h are approximately in the middle of those for TiN and TiO, indicating that both O and N atoms are randomly disordered in the anion sites of the rock salt structure.
It is worth mentioning that the metallic TiN phase can be obtained only when the nitridation temperature was higher than 900 C. This suggests that the nitridation at 700 C results in a partial replacement of oxygen in TiO 2 with nitrogen rather than complete substitution. In the course of nitridation, TiO 2 is the only oxygen source. It is thus conceivable that anion exchange between nitrogen and oxygen occurs initially at the surface of TiO 2 . Hence, the TiO 2 microspheres annealed at 700 C are composed of partially nitrided TiO 1Àx N x and unreacted TiO 2 phases. Presumably the nitrided phase starts forming on the outer surface of a TiO 2 microsphere and gradually penetrates into the inner core via the oxygen exchange. As a result, the outer shell is predominantly composed of TiO 1Àx N x whereas the inner core mainly consists of TiO 2 , yielding the core/shell type microspheres. The XRD data suggest that the nitridation time and temperature strongly inuence the formation of TiO 1Àx N x at the surface and the shell thickness. It is interesting to note that any XRD peaks corresponding to the TiO 2 phase were not displayed in the samples nitrided for 24 and 48 h although the unreacted TiO 2 phase remains in the nitrided samples. The absence of the XRD peaks suggests that TiO 2 in the core part might exist as an amorphous state. Nitridation appears to induce the disordered state via the exchange between oxygen and nitrogen in the boundary, which results in the transformation of the crystalline TiO 2 phase into an amorphous TiO 2 state. 6,22 Another plausible cause is that X-ray might not penetrate enough into the inner core of the TiO 2 microsphere. The diffracted peaks are thus ascribed mainly to the TiO 1Àx N x phase on the surface.
Raman spectroscopy was employed to conrm the structural changes by the nitridation time (Fig. 3). The anatase phase of TiO 2 has six Raman-active vibrational modes including one A 1g mode at 527 cm À1 , two B 1g modes at 382 and 504 cm À1 , and three E g modes at 141, 154, and 654 cm À1 . 23 The samples nitrided at 700 C for 7 h showed virtually identical Raman patterns owing to the anatase TiO 2 phase, suggesting that the conversion of TiO 2 into TiO 1Àx N x is not complete at this stage. However, the E g mode at 154 cm À1 in the spectrum of the  sample nitrided for 10 h was signicantly blue-shied and noticeably broadened relative to that in the TiO 2 spectrum, indicating that particle sizes were changed and defects were produced by nitridation. The Raman spectrum of the TiO 2 microspheres nitrided for 24 h did not show any peaks corresponding to the anatase TiO 2 phase and exhibited very broad peaks at $210, $290, and $570 cm À1 due to the cubic TiO 1Àx N x phase. The rst-order Raman scattering is forbidden for the crystal with a cubic rock salt structure (space group: Fm3m) because every atom is located at a site of inversion symmetry. 24 Since TiO 1Àx N x contains a wide range of vacancies and defects, however, the forbidden rule can be relaxed and Raman peaks can thus be observed. 25,26 The very broad peaks are probably associated with acoustic and optical modes of TiO 1Àx N x . It is worth mentioning that the unreacted TiO 2 phase was not detected in the 24 h sample by Raman analysis. This indicates that structural changes occur aer nitridation, resulting in oxygen disorders primarily in the TiO 2 phase in the core. The Raman data obtained from the nitrided samples are consistent with the XRD results.
The XRD and Raman results discussed above show the formation of the cubic TiO 1Àx N x phase but do not provide any direct evidence on the presence of the unreacted TiO 2 phase. XPS data of the nitrided samples prepared by annealing at different dwell times were examined to assess the chemical environments and oxidation states of Ti and N. In particular, two important regions including Ti 2p and N 1s were carefully investigated for each sample. Fig. 4 shows the XPS spectra in the Ti 2p region (452-462 eV) of the TiO 2 @TiO 1Àx N x samples. The Ti 2p XPS spectra of the TiO 2 and nitrided samples exhibit drastic differences. The XPS data show that the binding energy was shied from higher to lower values with increased annealing time. The sharp peak at 458.3 eV is typical for the Ti 2p 3/2 in anatase TiO 2 , namely the Ti 4+ species. 27,28 A notable feature is that the peak at 458.3 eV is observed regardless of the annealing time, suggesting that the unreacted TiO 2 phase is present even if the sample was nitrided for over 24 h. The broader peaks in the range of 455 to 458 eV can be attributed to Ti 2+ , Ti 3+ , and Ti 4+ species. 29,30 For the sample nitrided for 24 h, the spectral curve was tted with ve peaks centered at 455.6, 457.2, 458.3, 461.4, and 464.0 eV, which is given in Fig. 5. The Ti 2p 3/2 XPS peaks are typically observed in the range of 455 to 460 eV. The lowest peak at 455.6 eV is ascribed to the Ti 2+ species that could be associated with TiO. A peak centered at 457.2 eV can be assigned to the Ti 3+ of TiN or N-Ti-O bonding in TiO 1Àx N x . 31 The peak at 458.3 eV correlates with the Ti 4+ in TiO 2 . Two peaks at 461.4 and 464.0 eV are due to the Ti 2p 1/2 spectra. One peak at 464.0 eV is ascribed to the Ti 4+ of TiO 2 while the other at 461.4 eV is associated with the N-Ti-O bonding in the oxynitride phase such as TiO 1Àx N x . The Ti 2p XPS spectra suggest that the nitrided sample contains   the multiple oxidation states of Ti and might be composed of TiO, TiO 1Àx N x , and TiO 2 .
The presence of N in the nitrided sample was clearly evidenced by the N 1s XPS spectra of TiO 2 @TiO 1Àx N x as shown in Fig. 6. The N 1s peaks are more distinctive with increasing nitridation time. For the sample nitrided for 7 h, a nearly single peak is observed. However, the samples with a longer nitridation time showed a broad peak. The inset in Fig. 6 is the N 1s XPS spectrum of the 24 h sample. The red curve is the tting data of experimental XPS data, which can be resolved into two peaks shown as blue curves. The rst peak at 396.5 eV is characteristic of the N dopant which corresponds to the Ti-N bonding and the second one at 398.7 eV can be attributed to the N-Ti-O species. 32 From the XRD, Raman, and XPS data, it is conceivable that the nitrided samples are composed of TiO 1Àx N x and TiO 2 . However, it is very difficult to determine how much TiO 2 phase in the parent TiO 2 microsphere remains intact during the nitridation procedure. Thus, TGA was employed to determine the nitrogen content, which is useful to estimate the approximate ratio of TiO 1Àx N x to TiO 2 in TiO 2 @TiO 1Àx N x . The TGA data of the nitrided samples were collected in air as a function of temperature. For comparative purposes, commercial TiN powder was also measured under the same conditions. All the samples were transformed into the rutile TiO 2 phase aer measurements, which were conrmed by XRD.
As illustrated in Fig. 7, the weight percentage of TiN increases to about 129%, which is close to the theoretical value for the conversion of TiN into TiO 2 . In contrast, the weight gain of the sample nitrided at 700 C for 24 h is about 115%, which is much lower than that of TiN. This clearly suggests that the unreacted TiO 2 phase remains in the nitrided samples. A notable feature is that the weight gains of the samples annealed for 24 and 48 h are not signicantly different, supporting that the temperature annealed at 700 C is not sufficient to completely convert TiO 2 into TiN.
Since the ratio of TiO 1Àx N x to TiO 2 in the nitrided sample cannot be determined by the TGA data alone, the EDX elemental analysis was employed to help estimating the relative compositions in TiO 2 @TiO 1Àx N x . Although the EDX data are not completely reliable due to several factors such as surface, elements, and the specimen features, the technique provides reasonably reliable elemental compositions particularly for insoluble inorganic materials like TiO 2 @TiO 1Àx N x . Weight percentages of Ti and N of the samples nitrided for 10, 24, and 48 h were obtained directly from the EDX data and oxygen contents were calculated assuming that the samples are solely composed of Ti, N, and O. Table S2 † gives weight percentages and atomic ratios for the three samples. As anticipated, the EDX data show that the nitrogen content increases as the annealing time increases.
Based on TGA and EDX data, the ratio of TiO 1Àx N x to TiO 2 for the sample nitrided at 700 C for 24 h is estimated to be 0.18. For this calculation, we assumed that TiO 2 @TiO 1Àx N x has a core-shell type structure because the nitridation occurs initially on the surface and gradually penetrates into the core. EDX data were taken to determine relative amounts of N and O. As illustrated in the inset of Fig. 7, the shell thickness and the core diameter of the TiO 2 @TiO 1Àx N x sample are 56 nm and 308 nm, respectively. Fig. 8 shows elemental mapping images and EDX line scan elemental proles of the sample nitrided at 700 C for 24 h, clearly conrming the presence of N, O, and Ti. The Ti intensity is strong and the intensity prole along the line is nearly domeshaped. On the other hand, the intensities of N and O are relatively weak and did not noticeably change along the line due to their low scattering power. The elemental mapping images indicate that N, O, and Ti are evenly dispersed on the spherical surface and also imply that TiO 1Àx N x is mostly located on the outer portion of the shell.  The inset is a simplified drawing of a TiO 2 @TiO 1Àx N x microsphere with a shell thickness of about 56 nm and core diameter of about 208 nm. The relative ratio of TiO 1Àx N x to TiO 2 was determined on the basis of the TGA and EDX data for the sample nitrided at 700 C for 24 h.
As illustrated in Fig. 8, conventional TEM is very difficult to differentiate between core and shell components in TiO 2 @ TiO 1Àx N x mainly because the L-shell peak of Ti (0.452 keV) and the K-shell peak of N (0.392 keV) were too close to distinguish their elemental intensities. We have thus investigated the structure and chemical composition of the bisected TiO 2 @ TiO 1Àx N x microsphere with scanning transmission electron microscopy (STEM), electron energy loss spectroscopy (EELS), and energy-dispersive X-ray spectroscopy (EDS). A TiO 2 @ TiO 1Àx N x microsphere nitrided at 700 C for 10 h was used and approximately bisected by a focused ion beam (FIB). Fig. 9 presents the bisected STEM image and the corresponding EDS elemental mapping images of Ti, O, and N. The Ti and O signals are evenly dispersed in the core and shell parts of the bisected sphere whereas the N signals are located exclusively on the shell of the sphere, indicating that the TiO 2 @TiO 1Àx N x microsphere has the core/shell structure. To further convince the compositions of the core and shell parts, we have investigated the STEM-EDS analysis of the bisected sample. As given in Fig. S2, † the Ti, O, and N signals were observed in the shell part but only Ti and O signals were detected in the core. This clearly demonstrates that the TiO 1Àx N x phase is formed on the shell by the nitridation but the core TiO 2 remains intact.
Diffuse reectance UV-visible spectroscopy (DRUVS) was employed to determine whether the nitrided sample absorbs light effectively in the visible light region. The DRUVS spectrum of TiO 2 @TiO 1Àx N x nitrided at 700 C for 24 h clearly shows that the sample absorbs much more light in the visible regions in comparison with that of TiO 2 (Fig. S3 †). Thus the photocatalytic activity of TiO 2 @TiO 1Àx N x nitrided at 700 C for 24 h was evaluated by measuring the degradation rate of methylene blue (MB) under visible light as a function of time. The absorbance of MB at 664 nm, which is the strongest peak of MB, was used to determine the concentration change of the MB solution. 33 As shown in Fig. 10(a), the strong absorption peak is continuously weakened with increasing irradiation time. This clearly indicates that the TiO 2 @TiO 1Àx N x catalyst efficiently decomposes MB under visible light. Similar efficient activities were also observed in all the samples nitrided at 700 C. Among them, the sample nitrided for 24 h has the highest performance on the MB decomposition, which is given in Fig. 10(b). Due to the large band gap energy (E g ¼ 3.2 eV) of TiO 2 , the core TiO 2 is inactive in decomposition of MB under visible light. Therefore, it is evident that the TiO 1Àx N x phase in the shell is the only    To elucidate whether hydroxyl radicals (cOH) produced from the illumination of TiO 2 @TiO 1Àx N x lead to the degradation of MB, we have employed the uorescence technique that is usually used to estimate the formation of hydroxyl radicals on the surface of the photocatalyst. [34][35][36] Fig. 11 shows the uorescence spectral change of the terephthalic acid in the presence of TiO 2 @TiO 1Àx N x nitrided at 700 C for 24 h with increasing irradiation time. A new peak at about 425 nm, due to the hydroxylation product (2-hydroxyterephthalic acid), is distinctively displayed in the illuminated spectrum. The uorescence intensity is gradually increased with increasing irradiation time. Based on these results, it is evident that the hydroxyl radicals generated from the TiO 2 @TiO 1Àx N x catalyst plays a signicant role in the decomposition of MB.
In addition to the chemical degradation of MB by the hydroxyl radicals, we also evaluated the capability of photocurrent generation of the nitrided samples under visible light. The photocurrent density was clearly observed, which was very reproducible in the repeated on/off cycles of excitation.
As illustrated in Fig. 11, the photocurrent of the nitrided samples is enhanced when the nitridation time increased. The sample nitrided for 24 h has the largest value of photocurrent density ($170 mA cm À2 ), which is comparable with that of carbon nitride (C 3 N 4 ). 37,38 This clearly demonstrates that the TiO 2 @TiO 1Àx N x catalyst is able to generate photocurrents effectively under visible light while TiO 2 does not produce photocurrents under the same conditions (Fig. 12).

Conclusions
The present study demonstrated the successful synthesis of new composite photocatalysts (TiO 2 @TiO 1Àx N x ) with enhanced visible light activity, which was accomplished by a controlled nitridation of TiO 2 microspheres. The nitrided products maintain a spherical shape and have a unique core/shell structure.
The core and shell parts are mainly composed of amorphous TiO 2 and crystalline TiO 1Àx N x phases, respectively. They exhibited high photocatalytic and photoelectrochemical activities under visible light irradiation. Our structural and spectroscopic results conrm that the origin of their activity is due primarily to the TiO 1Àx N x phase. Specically, the nitrogen dopant in TiO 1Àx N x appears to be responsible for photocatalytic activity under visible-light similar to the well-documented Ndoped TiO 2 phase (TiO 2Àx N x ). Our studies on the TiO 1Àx N x phase having a simple rock-salt structure promise a wide range of visible-light applications in both photocatalytic and photoelectrochemical systems.