Preparation of Visible Light Photocatalytic Graphene Embedded Rutile Titanium(IV) Oxide Composite Nanowires and Enhanced NO x Removal

: The quest for developing highly efﬁcient TiO 2 -based photocatalysts is continuing and, in particular, evolving a new strategy is an important aspect in this regard. In general, much effort has been devoted to the anatase TiO 2 modiﬁcations, despite there being only a few recent studies on rutile TiO 2 (rTiO 2 ). To the best of our knowledge, studies on the preparation and characterization of the photocatalysts based on the intentional inclusion of graphene (G) into rTiO 2 nanostructures have not been reported yet. Herein, we develop a new type of TiO 2 -based photocatalyst comprising of G included pure rTiO 2 nanowire (abbreviated as rTiO 2 (G) NW) with enhanced visible light absorption capability. To prepare rTiO 2 (G) NW, the G incorporated titanate electrospun ﬁbers were obtained by electrospinning and subsequently heat treated at various temperatures (500 to 800 ◦ C). Electrospinning conditions were optimized for producing good quality rTiO 2 (G) NW. The rTiO 2 (G) NW and their corresponding samples were characterized by appropriate techniques such as X-ray diffraction (XRD), scanning electron microscopy, high-resolution transmission electron microscopy and UV-vis diffuse reﬂectance spectroscopy to ascertain their material characteristics. XRD results show that the lattice strain occurs upon inclusion of G. We present here the ﬁrst observation of an apparent bandgap lowering because of the G inclusion into TiO 2 NW. While anatase TiO 2 NW exhibited poor visible light photocatalysis towards NO x removal, the rTiO 2 (G) NW photocatalyst witnessed a signiﬁcantly enhanced (~67%) photocatalytic performance as compared to anatase TiO 2 (G) NW. We concluded that the inclusion of G into rTiO 2 nanostructures enhances the visible light photoactivity. A plausible mechanism for photocatalysis is suggested. A plausible mechanism for the visible light photocatalytic activities for rTiO 2 (G) composite nanowire (CNW) comprising of ( a ) carbon inclusion in the TiO 2 structure and ( b ) band energy modiﬁcation due to carbon inclusion for the visible light photocatalytic activities for rTiO 2 (G) CNW.


Introduction
Ever since Fujishima and Honda demonstrated the successful photoelectrochemical splitting of water, titanium dioxide (TiO 2 ) has become one of the most widely studied semiconductor photocatalysts [1,2]. TiO 2 has been fervently utilized as the photocatalyst and in other prime applications (such as photovoltaic devices, sensors, environmental remediation, etc.) due to its unique combination of low cost, chemical stability, non-toxicity, high reactivity, excellent stability against photocorrosion, and ease for further functionalization [3,4]. Generally, TiO 2 exhibits varying photocatalytic activities which depend on its phase structure, crystallite size, specific surface areas, contact between G and TiO2 without blocking the active sites of TiO2. The TiO2-G composites prepared by mechanical mixing can hardly result in a uniform distribution of G onto the surface of TiO2 nanostructures and can have limited surface area for the photocatalytic activity. These factors can definitely weaken the effect of G for the improved photocatalytic activity of TiO2.

Synthetic Optimizations
The procedure for the fabrication of rTiO2(G)-CNW and the other studied photocatalysts are presented in Scheme 1. The typical procedure involves two steps; (i) preparation of G embedded titanate composite nanofibers (Titanate (G)-CNFs) and (ii) calcination of the Titanate (G)-CNFs. Our focus was to prepare high quality rTiO2 (G)-CNW having uniform diameters. Taking into account the morphology and diameter of the resultant electrospun Titanate (G)-CNFs (step i) and the temperature for hydrothermal treatment (step ii), we manipulated the various parameters involved in these two steps, namely, the recipe for the electrospinning dope and temperature for the calcination process. Typical experimental parameters that were selected for the electrospinning process towards the preparation of Titanate(G)-CNFs are listed in Table 1. For pure rTiO2(G)-CNW formation, the calcination temperature was chosen as 800 °C based on our knowledge obtained from the relevant literature [48]. The as-synthesized materials are designated based on these two steps and presented in Table 1.
In this study, our focus is to prepare G embedded TiO2 NWs rather the physical mixture between TiO2 NW and G. Hence, we designed the conditions at the electrospinning dope stage so as to chemically link the G moieties within the titanate framework. During the formation of titanate gel, carboxylated G was introduced. As a result, the condensation processes resulted in the formation of Ti-O-C-Ti-three dimensional networks. During the hydrolysis of titanium isopropoxide (TIIP), titanium hydroxides were formed by the nucleophilic substitution of the terminal isopropoxide groups. In addition, condensation between two hydroxyl groups in titanium hydroxides and the occurrence of cross reactions between titanium hydroxides and carboxylated G leads to Ti-O-C-Tithree dimensional networks. Of course, the rate of condensation between two hydroxyl groups in titanium hydroxides is expected to be much higher than the condensation reaction between carboxyl or hydroxyl groups in G and hydroxyl groups in titanium hydroxide. By varying the G precursor in Scheme 1. Preparation of rTiO 2 (G)-CNW through electrospinning-calcination processes.

Synthetic Optimizations
The procedure for the fabrication of rTiO 2 (G)-CNW and the other studied photocatalysts are presented in Scheme 1. The typical procedure involves two steps; (i) preparation of G embedded titanate composite nanofibers (Titanate (G)-CNFs) and (ii) calcination of the Titanate (G)-CNFs. Our focus was to prepare high quality rTiO 2 (G)-CNW having uniform diameters. Taking into account the morphology and diameter of the resultant electrospun Titanate (G)-CNFs (step i) and the temperature for hydrothermal treatment (step ii), we manipulated the various parameters involved in these two steps, namely, the recipe for the electrospinning dope and temperature for the calcination process. Typical experimental parameters that were selected for the electrospinning process towards the preparation of Titanate(G)-CNFs are listed in Table 1. For pure rTiO 2 (G)-CNW formation, the calcination temperature was chosen as 800 • C based on our knowledge obtained from the relevant literature [48]. The as-synthesized materials are designated based on these two steps and presented in Table 1.
In this study, our focus is to prepare G embedded TiO 2 NWs rather the physical mixture between TiO 2 NW and G. Hence, we designed the conditions at the electrospinning dope stage so as to chemically link the G moieties within the titanate framework. During the formation of titanate gel, carboxylated G was introduced. As a result, the condensation processes resulted in the formation of Ti-O-C-Ti-three dimensional networks. During the hydrolysis of titanium isopropoxide (TIIP), titanium hydroxides were formed by the nucleophilic substitution of the terminal isopropoxide groups. In addition, condensation between two hydroxyl groups in titanium hydroxides and the occurrence of cross reactions between titanium hydroxides and carboxylated G leads to Ti-O-C-Ti-three dimensional networks. Of course, the rate of condensation between two hydroxyl groups in titanium hydroxides is expected to be much higher than the condensation reaction between carboxyl or hydroxyl groups in G and hydroxyl groups in titanium hydroxide. By varying the G precursor in the electrospinning dope (Table 1), the titanate gel with different extents of G inclusion was successfully achieved. Furthermore, it should be noted that molecular weight of the polymer (PVP in this case) can also influence the electrospun fiber morphologies. Generally, molecular weight of the polymer decides the entanglement of polymer chains in solutions and hence the solution viscosity. In the electrospinning process, electrospinning dope prepared with the low molecular weight polymers can result in beads as compared to the high molecular weight polymers. Relatively high molecular weight polymers can lead to smooth electrospun fibers. However, too high concentrations of the high molecular weight polymers would lead to micro-ribbon or belt or flattened fiber-like morphologies. In this work, we aim to obtain cylindrical Titanate(G)-CNFs with uniform diameters, which in turn can lead to the formation of uniform TiO 2 NWs at the calcination stage. We employed a judicious selection of mixture of PVP with low (40,000) and high (1.30 × 10 6 ) molecular weight towards achieving our goal. Table 1 details the electrospinning conditions (Table 1).
It must be noted that the calcination temperature influences the phase transformation among the main polymorphs of TiO 2 (anatase, rutile and brookite) [49]. The main active crystallite phases of TiO 2 are rutile and anatase. When the calcination temperature is set above 600 • C, the rutile phase predominantly results [50]. Keeping this in view, we kept the calcination temperature far above 600 • C to specifically induce rutile phases. (Table 1)

Morphology
FESEM has been utilized to investigate the morphology of the resultant electrospun fibers obtained from various conditions as in Table 1 ( Figure 1). The SEM images of pristine Titanate NFs and Titanate(G)-CNFs are shown in Figure 1a-f. The as-spun titanate NFs (S1 and S2, Table 1) prepared with different molecular weight (40,000 and 1.30 × 10 6 ) of PVP (Figure 1a,b) showed variations in the morphologies. Titanate NFs (S1) prepared with low molecular weight (40,000) PVP ( Figure 1a) consisted of randomly distributed beads and a few NFs (having diameters in the range between 500 and 700 nm). On the other hand, mainly NFs could be seen for S2, the titanate NFs prepared with high molecular weight (1.30 × 10 6 ) (Figure 1b) PVP. The diameters of fibers varied significantly (in the range of 200 nm to 720 nm) and a few flattened fibers could also be seen for S2 (Figure 1b). The titanate(G)-CNFs (S3) prepared with only 80 mg of G and the high molecular weight PVP (1.30 × 10 6 ) had larger proportions of NFs with diameters in the range 80 to 100 nm and randomly distributed beads (Figure 1c). On the other hand, FESEM image of titanate(G)-CNFs (S4) ( Figure 1d) prepared with 80 mg of G in a mixture of high and low molecular weight of PVP, showed the predominant existence of NFs (having an average diameter of 600 nm) with smooth surfaces (Figure 1d). Hence, it is concluded that the usage of mixture of PVP having low and high molecular weight in the presence of G resulted in Titanate (G)-CNFs without beads. FESEM images of S5 and S6 (prepared with different amount of G (40 mg and 120 mg) in a mixture solution containing low and high molecular weight of PVP show the presence of NFs without beads (Figure 1e,f). Hence, we have successfully optimized conditions for producing Titanate (G)-CNFs without beads.
Catalysts 2019, 9, 170 5 of 18 ( Figure 1d). Hence, it is concluded that the usage of mixture of PVP having low and high molecular weight in the presence of G resulted in Titanate (G)-CNFs without beads. FESEM images of S5 and S6 (prepared with different amount of G (40 mg and 120 mg) in a mixture solution containing low and high molecular weight of PVP show the presence of NFs without beads (Figure 1e,f). Hence, we have successfully optimized conditions for producing Titanate (G)-CNFs without beads.  (Figure 2c and Figure 3c). The larger amount of G included in TiO2(G)-CNW is expected to prevent the dehydration of the inter-layered OH groups It is observed that calcination (550 and 800 • C) of the as-electrospun Titanate NFs (S1 and S2) and Titanate (G)-CNFs (S4-S5) influences the morphology of the resultant composites (Figures 2a-d and 3a-d). FESEM images of S1 (550) and S1 (800) (Figures 2a and 3a) show the presence of randomly distributed micro/nanoparticles with smaller proportions of NWs. FESEM images of S2 (550) and S2 (800) (Figures 2b and 3b) inform the presence of larger proportions of NWs as compared with the NF proportions in S1 (550) and S1 (800) (Figures 2a and 3a). SEM images corresponding to S4 (550) (Figure 2c (Figures 2c and 3c). The average diameter of S4 (550) and S4 (800) was found to be 380 nm and 260 nm, respectively. This observation is in accordance with the Ti-O-C-Ti-three dimensional networks generation during the titanate(G) CNF formation stage. On the other hand, SEM images of S5 (500) and S5 (800) (Figures 2d and 3d) revealed the presence of longer NWs as compared to the length of S4 (550) and S4 (800) (Figures 2c  and 3c). The larger amount of G included in TiO 2 (G)-CNW is expected to prevent the dehydration of the inter-layered OH groups during the annealing process [51], resist the structural integrity of NWs and create, to a large extent, stable Ti-O-C-Ti-three dimensional networks in the titanate(G) CNF. during the annealing process [51], resist the structural integrity of NWs and create, to a large extent, stable Ti-O-C-Ti-three dimensional networks in the titanate(G) CNF.    during the annealing process [51], resist the structural integrity of NWs and create, to a large extent, stable Ti-O-C-Ti-three dimensional networks in the titanate(G) CNF.

X-Ray Diffraction Analysis
The TiO2 NWs and TiO2(G)-CNWs obtained through annealing of the respective titanate fibers at 550 and 800 °C were characterized by X-ray powder diffraction technique, and their XRD patterns are shown in Figure 7a,b. The peaks corresponding to the crystalline phases of TiO2, anatase and rutile, are identified in accordance with the references for these phases, JCPDS 21-1272 for anatase and JCPDS 21-1276 for rutile, in all the diffractograms, and labelled in Figure 7a,b. Several crystalline properties, such as the mass fraction percentage of anatase (XA) and rutile (XR) phases, the average crystallite size (t), the unit cell parameters (a, b and c) and the unit cell volume (V) are calculated and presented in Table 2 [52]. XRD patterns of 550 °C annealed samples (S1(550), S2(550), S4(550) and S5 (550)) showed the predominant existence of peaks for the anatase phase, indexed as 2= 5.  (112). The anatase peaks were virtually absent for 800 °C annealed samples (S1(800), S2(800), S4(800) and S5(800). Thus, the obtained TiO2(G) CNWs after calcination at 800 °C contain pure rutile TiO2 and confirmed the successful synthesis of pure rTiO2(G) CNWs. However, there are no separate peaks for G in the diffraction patterns of S4 (550), S5 (550), S4 (800) and S5 (800), possibly due to the low amount and low intensity of G peaks. Moreover, the characteristic peak of G at 24.5° may be superimposed or masked by the main peaks of TiO2 phases [53].

X-Ray Diffraction Analysis
The TiO 2 NWs and TiO 2 (G)-CNWs obtained through annealing of the respective titanate fibers at 550 and 800 • C were characterized by X-ray powder diffraction technique, and their XRD patterns are shown in Figure 7a,b. The peaks corresponding to the crystalline phases of TiO 2 , anatase and rutile, are identified in accordance with the references for these phases, JCPDS 21-1272 for anatase and JCPDS 21-1276 for rutile, in all the diffractograms, and labelled in Figure 7a,b. Several crystalline properties, such as the mass fraction percentage of anatase (X A ) and rutile (X R ) phases, the average crystallite size (t), the unit cell parameters (a, b and c) and the unit cell volume (V) are calculated and presented in Table 2 [52]. XRD patterns of 550 • C annealed samples (S1 (550), S2 (550), S4 (550) and S5 (550) (112). The anatase peaks were virtually absent for 800 • C annealed samples (S1 (800), S2 (800), S4 (800) and S5 (800). Thus, the obtained TiO 2 (G) CNWs after calcination at 800 • C contain pure rutile TiO 2 and confirmed the successful synthesis of pure rTiO 2 (G) CNWs. However, there are no separate peaks for G in the diffraction patterns of S4 (550), S5 (550), S4 (800) and S5 (800), possibly due to the low amount and low intensity of G peaks. Moreover, the characteristic peak of G at 24.5 • may be superimposed or masked by the main peaks of TiO 2 phases [53].  In order to understand the influences of calcination and G inclusion on the microstructures of the TiO2 with different phase structures, XRD patterns for all samples were carefully analyzed. One can see that the crystallite sizes for TiO2 increased upon increasing the calcination temperature ( Table 2). The increased crystallite size is attributed to the thermally promoted crystallite growth at elevated temperatures. For the 550 °C calcined samples (S1 (550), S2 (550), S4(550) and S5 (550)), the lattice parameters were in accordance with a=b≠c corresponding to the tetragonal crystal lattice with a, b, c values 3.78.3.78 and 9.50 A (Table 2), respectively. The 800 °C annealed samples (S1 (800), S2 (800), S4 (800) and S5 (800)) retained the tetragonal structure with distinctly different lattice parameters for a, b and c ( Table 2). After 800 °C calcinations, the lattice parameters, especially a and b, increased from 3.78 Å to 4.94 Å. This observation is not normal since high temperature calcinations usually lead to lattice shrinkage due to the larger grain sizes. Also, the lattice parameter c significantly decreased from 9.50 to 2.96. Besides, the axial ratio, c/a, that represents the lattice symmetry, significantly decreased from 2.70 to 0.65 and, the unit cell volume decreased from ~136 to ~62. These alterations in lattice parameters observed upon high-temperature calcinations and G inclusions suggested the presence of defects [54]. Addition of a small amount of carbon impurity to TiO2 in a controlled manner can be referred to as C-doping of TiO2. There can be two possible doping  In order to understand the influences of calcination and G inclusion on the microstructures of the TiO 2 with different phase structures, XRD patterns for all samples were carefully analyzed. One can see that the crystallite sizes for TiO 2 increased upon increasing the calcination temperature ( Table 2). The increased crystallite size is attributed to the thermally promoted crystallite growth at elevated temperatures. For the 550 • C calcined samples (S1 (550), S2 (550), S4 (550) and S5 (550)), the lattice parameters were in accordance with a=b =c corresponding to the tetragonal crystal lattice with a, b, c values 3.78.3.78 and 9.50 A (Table 2), respectively. The 800 • C annealed samples (S1 (800), S2 (800), S4 (800) and S5 (800)) retained the tetragonal structure with distinctly different lattice parameters for a, b and c ( Table 2). After 800 • C calcinations, the lattice parameters, especially a and b, increased from 3.78 Å to 4.94 Å. This observation is not normal since high temperature calcinations usually lead to lattice shrinkage due to the larger grain sizes. Also, the lattice parameter c significantly decreased from 9.50 to 2.96. Besides, the axial ratio, c/a, that represents the lattice symmetry, significantly decreased from 2.70 to 0.65 and, the unit cell volume decreased from~136 to~62. These alterations in lattice parameters observed upon high-temperature calcinations and G inclusions suggested the presence of defects [54]. Addition of a small amount of carbon impurity to TiO 2 in a controlled manner can be referred to as C-doping of TiO 2 . There can be two possible doping carbon sites in TiO 2 ; (i) carbon at an oxygen site anion doping and (ii) carbon at a titanium site cation doping. The density functional theory predicted that cation-doped carbon atoms can result in a carbonate-type structure, whereas anion-doped carbon atoms (substitution on the oxygen site) do not involve in any significant structural changes [54]. The G included in this work is expected to make carbonate type structural modifications. Due to carbon substitution on the oxygen, a visible-light response could be induced due to the appearance of an unoccupied impurity state occurring in the band gap, which can be beneficial in promoting photocatalytic degradation reactions. We invoked the possibility of energy level and optical property changes due to the calcination and G inclusion in the prepared rTiO 2 (G) CNW.

Raman Spectroscopy
The results from XRD characterizations (Figure 7) implied that the existence of G and annealing temperatures influence the crystal phase of TiO 2 and also contribute to the inclusion of defects in crystal structure. To clarify these aspects, Raman spectra were recorded. Figure 8 shows the Raman spectra for S2, S4, S2 (800) and S4 (800) samples. Raman spectra of S2 and S4 did not show characteristic peaks for crystalline TiO 2 phase due to the amorphous nature and the presence of large extent of polymers. Raman spectrum of S2 (800) and S4 (800) (Figure 8) shows predominant characteristic peaks around~140 cm −1 (B1g),~445 cm −1 (Eg),~610 cm −1 (A1g), and a broad band around 240 cm −1 for second-order effect that correspond to the rutile phase [55]. There exists a blue shift for the first two modes (Eg(1), B1g (1)) in the Raman spectrum of S4 (800) as compared to the samples S2 (800) suggesting the formation of hybrid structure in the case of S4 (800) due to the inclusion of G in the TiO 2 crystal lattice. It must be noted that Raman peaks of the S2 (800) correspond to pure rutile TiO 2 , whilst that of S4 (800) correspond to the G included rutile TiO 2 . We attribute the blue shifts in the Raman modes (Eg(1), B1g(1)) ) for S4 (800) to the oxygen vacancies as Eg(1) and B1g(1), modes are sensitive to O-O interactions [56]. As a result of possible sharing of oxygen at the interfaces between TiO 2 nanocrystals and the TiO 2 /G, the TiO 2 crystal structure includes defects. carbon sites in TiO2; (i) carbon at an oxygen site anion doping and (ii) carbon at a titanium site cation doping. The density functional theory predicted that cation-doped carbon atoms can result in a carbonate-type structure, whereas anion-doped carbon atoms (substitution on the oxygen site) do not involve in any significant structural changes [54]. The G included in this work is expected to make carbonate type structural modifications. Due to carbon substitution on the oxygen, a visible-light response could be induced due to the appearance of an unoccupied impurity state occurring in the band gap, which can be beneficial in promoting photocatalytic degradation reactions. We invoked the possibility of energy level and optical property changes due to the calcination and G inclusion in the prepared rTiO2(G) CNW.

Raman Spectroscopy
The results from XRD characterizations (Figure 7) implied that the existence of G and annealing temperatures influence the crystal phase of TiO2 and also contribute to the inclusion of defects in crystal structure. To clarify these aspects, Raman spectra were recorded. Figure 8 shows the Raman spectra for S2, S4, S2(800) and S4(800) samples. Raman spectra of S2 and S4 did not show characteristic peaks for crystalline TiO2 phase due to the amorphous nature and the presence of large extent of polymers. Raman spectrum of S2(800) and S4(800) (Figure 8) shows predominant characteristic peaks around ~140 cm −1 (B1g), ~445 cm −1 (Eg),~ 610 cm −1 (A1g), and a broad band around 240 cm −1 for second-order effect that correspond to the rutile phase [55]. There exists a blue shift for the first two modes (Eg(1), B1g(1)) in the Raman spectrum of S4(800) as compared to the samples S2(800) suggesting the formation of hybrid structure in the case of S4(800) due to the inclusion of G in the TiO2 crystal lattice. It must be noted that Raman peaks of the S2(800) correspond to pure rutile TiO2, whilst that of S4(800) correspond to the G included rutile TiO2. We attribute the blue shifts in the Raman modes (Eg(1), B1g(1)) ) for S4(800) to the oxygen vacancies as Eg(1) and B1g(1), modes are sensitive to O-O interactions [56]. As a result of possible sharing of oxygen at the interfaces between TiO2 nanocrystals and the TiO2/G, the TiO2 crystal structure includes defects. Raman spectroscopy was also employed for molecular morphology characterization of carbon inclusion in TiO2 structure. As shown in Figure 8 (inset), two Raman peaks were located around ~1350 cm −1 and ~1600 cm −1 for the sample S4(800) that are attributed to D and G bands [57]. While the G band is common to all sp2 carbon forms and provides information on the in-plane vibration of sp2 bonded carbon atoms, the D band indicates the presence of sp3 defects. This result confirms the inclusion of G in the S4(800). The intensity ratio of the D and G bands (ID/IG) of was 0.757 for S4(800). The significant intensity of D band in S4(800) indicated that inclusion of G could result in chemical Raman spectroscopy was also employed for molecular morphology characterization of carbon inclusion in TiO 2 structure. As shown in Figure 8 (inset), two Raman peaks were located around 1350 cm −1 and~1600 cm −1 for the sample S4 (800) that are attributed to D and G bands [57]. While the G band is common to all sp2 carbon forms and provides information on the in-plane vibration of sp2 bonded carbon atoms, the D band indicates the presence of sp3 defects. This result confirms the inclusion of G in the S4 (800). The intensity ratio of the D and G bands (I D /I G ) of was 0.757 for S4 (800). The significant intensity of D band in S4 (800) indicated that inclusion of G could result in chemical linking to TiO 2 structure and the modify the sp 2 carbon network. Raman results (Figure 8) are in good agreement with the results of XRD (Figure 7).

Optical Properties
Diffuse Reflectance Spectroscopy UV−vis diffuse reflectance spectra (DRS) measurements (Figure 9i) were used to infer the optical characteristics and to determine the optical band gap energy of the synthesized materials. Generally, the G included samples (S4 (550) and S4 (800)) samples exhibited broad absorbance in the visible region as compared to S2 (550) and S2 (800). The augmentation of absorbance in the visible light region up to 800 nm for (S4 (550) and S4 (800)) is attributed to the presence of G species, which act as a photosensitizer. In addition, the extent of visible region absorbance varied is higher for S4 (800) as compared to S4 (550). The results informed that the rutile transformation induces more visible light absorptions. The band gap of TiO 2 was evaluated from the intercept of the straight-line portion of the Kubelka-Munk plots (Figure 9ii) [49].
The band gap energy of S2 (550), S2 (800), S4 (550) and S4 (800) was determined to be 2.98 eV, 2.89 eV, 3.00 eV and 2.40 eV. The trend in band gap shows that S4 (800) has the lowest band gap (2.40 eV, corresponding to a visible light wavelength~520 nm) among the studied materials. The decrease in the bandgap is ascribed to the existence of defect states induced by the Ti−O−Ti-C-skeleton included in the frame works and the increased percentage of the rutile phase, which has a smaller bandgap than the anatase phase [7,8,50,51].

Optical Properties
Diffuse Reflectance Spectroscopy UV−vis diffuse reflectance spectra (DRS) measurements (Figure 9i) were used to infer the optical characteristics and to determine the optical band gap energy of the synthesized materials. Generally, the G included samples (S4 (550) and S4 (800)) samples exhibited broad absorbance in the visible region as compared to S2 (550) and S2 (800). The augmentation of absorbance in the visible light region up to 800 nm for (S4 (550) and S4 (800)) is attributed to the presence of G species, which act as a photosensitizer. In addition, the extent of visible region absorbance varied is higher for S4 (800) as compared to S4 (550). The results informed that the rutile transformation induces more visible light absorptions. The band gap of TiO2 was evaluated from the intercept of the straight-line portion of the Kubelka-Munk plots (Figure 9ii) [49].
The band gap energy of S2 (550), S2 (800), S4 (550) and S4 (800) was determined to be 2.98 eV, 2.89 eV, 3.00 eV and 2.40 eV. The trend in band gap shows that S4 (800) has the lowest band gap (2.40 eV, corresponding to a visible light wavelength ~520 nm) among the studied materials. The decrease in the bandgap is ascribed to the existence of defect states induced by the Ti−O−Ti-C-skeleton included in the frame works and the increased percentage of the rutile phase, which has a smaller bandgap than the anatase phase [7,8,50,51]. Photoluminescence Photoluminescence (PL) emission spectroscopy was used to study the charge transfer behavior of photo-induced electrons and holes in the S2, S4, S2(800) and S4(800) samples ( Figure 10) which reveal the surface structure and excited state of the TiO2 semiconductor and the TiO2-G hybrid samples. It must be noted that S2(800) and S4(800) are the pure rutile TiO2 and G included TiO2 materials. Hence, the analysis of PL spectral of the S2, S4, S2(800) and S4(800) samples could reveal changes in the photoemission characteristics due to the G inclusion and rutile transformation. In general, a decrease in the PL intensity indicates the lowering of photo-induced electron-hole pair recombination rate and suggests the increase in photo-generated charged separation and hence increased photocatalytic activity. The spectra of all samples show a peak centered around 356 nm. There is an increase in PL intensity upon conversion of S2 to pure rutile TiO2 (Figure 10) samples S2 Photoluminescence Photoluminescence (PL) emission spectroscopy was used to study the charge transfer behavior of photo-induced electrons and holes in the S2, S4, S2 (800) and S4 (800) samples ( Figure 10) which reveal the surface structure and excited state of the TiO 2 semiconductor and the TiO 2 -G hybrid samples. It must be noted that S2 (800) and S4 (800) are the pure rutile TiO 2 and G included TiO 2 materials. Hence, the analysis of PL spectral of the S2, S4, S2 (800) and S4 (800) samples could reveal changes in the photoemission characteristics due to the G inclusion and rutile transformation. In general, a decrease in the PL intensity indicates the lowering of photo-induced electron-hole pair recombination rate and suggests the increase in photo-generated charged separation and hence increased photocatalytic activity. The spectra of all samples show a peak centered around 356 nm. There is an increase in PL intensity upon conversion of S2 to pure rutile TiO 2 ( Figure 10) samples S2 and S2 (800). This is consistent with the literature that suggests that pure rutile TiO 2 is less photoactive as compared to pure anatase and mixture of anatase and rutile TiO 2 . However, a significant decrease in PL intensity was observed between the samples S4 and S4 (800) (Figure 10). The lowering of PL emission intensity for S4 (800) as compared to S4 may be attributed to the role of G as an electron shuttle for rutileTiO 2 , preventing electron-hole recombination with a resulting higher photocatalytic activity. PL spectrum of S4 (800) has the lowest emission intensity, suggesting highest photo activity for rTiO 2 (G).
Catalysts 2019, 9, 170 12 of 18 and S2(800). This is consistent with the literature that suggests that pure rutile TiO2 is less photoactive as compared to pure anatase and mixture of anatase and rutile TiO2. However, a significant decrease in PL intensity was observed between the samples S4 and S4(800) (Figure 10). The lowering of PL emission intensity for S4(800) as compared to S4 may be attributed to the role of G as an electron shuttle for rutileTiO2, preventing electron-hole recombination with a resulting higher photocatalytic activity. PL spectrum of S4(800) has the lowest emission intensity, suggesting highest photo activity for rTiO2(G). Figure 10. Photoluminescence spectrum of S2, S4, S2(800) and S4(800).

Photodegradation of NOX under Visible Light
To compare the photodegradation of NOx between the samples (S2, S4, S2 ((550), S2 (800), S4 (550) and S4 (800), the NOx removal measurements were monitored under visible light irradiation, and the results are shown in Figure 11. The NOx removal efficiency is normalized in terms of area of light exposure and amount of photocatalyst used. Continuous lowering NOx concentration was observed for uncalcined (S2, S4) and calcined S2 (550), S2 (800), S4 (550) and S4 (800) samples. In the first instance, the photocatalytic activities of the uncalcined samples (S2 and S4) were compared (Figure 9). S2 and S4 show similar NOx removal tendencies over the time. The results inform us that inclusion of G at the titanate gel forming stage did not influence the NOx removal. On comparing the photodegradation efficiency between S2 and S2 (550), it is clear that the NOx removal efficiency did not improve significantly even after the conversion of titanate to anatase TiO2 (as confirmed through XRD data (Figure 7a). The result is consistent with the assumption that anatase TiO2 cannot show visible light photo degradation of NOx because of its high optical band gap (3.40 eV). The visible light photo degradation of S4 (550) was significant compared to S2 (550). This clearly informs that inclusion of G in the anatase TiO2 NW causes increased visible light degradation. This result corroborates the lowering of TiO2 optical band gap because of the inclusion of G within the TiO2 NW. A very interesting observation was made by comparing the NOx photodegradation performances of S2 (550), S2 (800), S4 (550) and S4 (800). NOx removal efficiency of S2 (550) and S2 (800) are closer to each other. Hence, the conversion of anatase TiO2 NW to rutile TiO2 NW did not improve visible light photo degradation of NOx. However, strikingly significant photodegradation of NOx was witnessed for S4 (800). The rate of removal of NOx was much higher for S4 (800). Thus, through our results we demonstrated that rTiO2 (G)NW can be used for the visible light assisted removal NOx. It has been notified through several reports that photocatalytic activity of the anatase is superior compared with the rutile polymorph due to the higher electron mobility, low dielectric

Photodegradation of NO X under Visible Light
To compare the photodegradation of NO x between the samples (S2, S4, S2 (550), S2 (800), S4 (550) and S4 (800), the NO x removal measurements were monitored under visible light irradiation, and the results are shown in Figure 11. The NO x removal efficiency is normalized in terms of area of light exposure and amount of photocatalyst used. Continuous lowering NO x concentration was observed for uncalcined (S2, S4) and calcined S2 (550), S2 (800), S4 (550) and S4 (800) samples. In the first instance, the photocatalytic activities of the uncalcined samples (S2 and S4) were compared (Figure 9). S2 and S4 show similar NO x removal tendencies over the time. The results inform us that inclusion of G at the titanate gel forming stage did not influence the NO x removal. On comparing the photodegradation efficiency between S2 and S2 (550), it is clear that the NO x removal efficiency did not improve significantly even after the conversion of titanate to anatase TiO 2 (as confirmed through XRD data (Figure 7a). The result is consistent with the assumption that anatase TiO 2 cannot show visible light photo degradation of NO x because of its high optical band gap (3.40 eV). The visible light photo degradation of S4 (550) was significant compared to S2 (550). This clearly informs that inclusion of G in the anatase TiO 2 NW causes increased visible light degradation. This result corroborates the lowering of TiO 2 optical band gap because of the inclusion of G within the TiO 2 NW. A very interesting observation was made by comparing the NO x photodegradation performances of S2 (550), S2 (800), S4 (550) and S4 (800). NO x removal efficiency of S2 (550) and S2 (800) are closer to each other. Hence, the conversion of anatase TiO 2 NW to rutile TiO 2 NW did not improve visible light photo degradation of NO x . However, strikingly significant photodegradation of NO x was witnessed for S4 (800). The rate of removal of NO x was much higher for S4 (800). Thus, through our results we demonstrated that rTiO 2 (G)NW can be used for the visible light assisted removal NO x . It has been notified through several reports that photocatalytic activity of the anatase is superior compared with the rutile polymorph due to the higher electron mobility, low dielectric constant, higher Fermi level, lower ability to adsorb oxygen, and higher degree of hydroxylation of the anatase and indirect electronic transitions [9,10]. Rutile is the TiO 2 polymorph that is most thermodynamically stable in a wide temperature range and at varied pressures and it is abundantly available; thus, there is need to evolve strategies to utilize a rutile polymorph for commercial and environmental applications. However, there are limited reports that focus on the photocatalytic activity of the rutile nanostructures.
Catalysts 2019, 9, 170 13 of 18 constant, higher Fermi level, lower ability to adsorb oxygen, and higher degree of hydroxylation of the anatase and indirect electronic transitions [9,10]. Rutile is the TiO2 polymorph that is most thermodynamically stable in a wide temperature range and at varied pressures and it is abundantly available; thus, there is need to evolve strategies to utilize a rutile polymorph for commercial and environmental applications. However, there are limited reports that focus on the photocatalytic activity of the rutile nanostructures. Toma et al. [58] studied the removal of NOx using TiO2 Degussa P25 powder and reported a maximum NOx removal efficiency of 4 X10 -4 ppm/m 2 /mg. The modified TiO2 samples prepared in the present study (S2, S4, S2(550), S2(800), S4(550), S4(800)) showed extremely high photocatalytic NOx removal efficiencies (>5.0 ppm/m 2 /mg) ( Figure) than TiO2 Degussa P25. Specifically, rTiO2(G) CNW samples (S2(800) and S4(800)) are highly efficient for the removal of NOx. Particularly, S4(800) outperforms S2(800) and other samples towards photocatalytic NOx removal. We ascribe the following reasons for the supreme performance of rTiO2(G) (800) (S4(800)). Firstly, the TiO2 in the S4(800) is nanostructured with wire like morphology ( Figure 3c) and expected to possess the advantage of high surface area. Secondly, the inclusion of G within TiO2 structure (Figure 7b) modifies the optical band structure (Figure 9ii). It is to be noted that TiO2 Degussa P25 powder exhibited NOx removal only under UV light irradiation. However, S4(800) exhibited excellent NOx removal efficiency under visible light irradiation [58]. The intentional inclusion of carbon impurities within TiO2 lattice has been reported to induce enhanced visible light photocatalytic activities [59,60]. The carbon dopant/impurity can exist as an anion by replacing the oxygen or as a cation occupying at interstitial lattice sites.
In the present work, the as-prepared rTiO2(G) CNW has three modifications in TiO2; (i) morphology (nanostructured as NW), (Figure 3) (ii) phase transformations (anatase to rutile) ( Figure  7) and (iii) inclusion of foreign atom (inclusion of carbon) ( Figure 5, Figure 7 and Figure 8). All the three modifications synergistically make rTiO2(G) CNW (S4(800)) as the visible light active and efficient photocatalyst. While the nanostructuring provides large surface area, the other two modifications (anatase-rutile phase transformation and carbon inclusion) contribute to the alteration of optical bad gap (Figure 9) leading to extended visible light absorption [61]. The G inclusion extends the visible light absorption and improves the photocatalytic properties through the role of electron mediator/sink. A plausible mechanism is also suggested for the photocatalysis of rTiO2 (G) CNW based on the above considerations and experimental observations (Scheme 2). The bulk Toma et al. [58] studied the removal of NO x using TiO 2 Degussa P25 powder and reported a maximum NO x removal efficiency of 4 × 10 −4 ppm/m 2 /mg. The modified TiO 2 samples prepared in the present study (S2, S4, S2 (550), S2 (800), S4 (550), S4 (800)) showed extremely high photocatalytic NO x removal efficiencies (>5.0 ppm/m 2 /mg) ( Figure 11) than TiO 2 Degussa P25. Specifically, rTiO 2 (G) CNW samples (S2 (800) and S4 (800)) are highly efficient for the removal of NO x . Particularly, S4 (800) outperforms S2 (800) and other samples towards photocatalytic NO x removal. We ascribe the following reasons for the supreme performance of rTiO 2 (G) (800) (S4 (800)). Firstly, the TiO 2 in the S4 (800) is nanostructured with wire like morphology ( Figure 3c) and expected to possess the advantage of high surface area. Secondly, the inclusion of G within TiO 2 structure (Figure 7b) modifies the optical band structure (Figure 9ii). It is to be noted that TiO 2 Degussa P25 powder exhibited NO x removal only under UV light irradiation. However, S4 (800) exhibited excellent NO x removal efficiency under visible light irradiation [58]. The intentional inclusion of carbon impurities within TiO 2 lattice has been reported to induce enhanced visible light photocatalytic activities [59,60]. The carbon dopant/impurity can exist as an anion by replacing the oxygen or as a cation occupying at interstitial lattice sites.
In the present work, the as-prepared rTiO 2 (G) CNW has three modifications in TiO 2 ; (i) morphology (nanostructured as NW), (Figure 3) (ii) phase transformations (anatase to rutile) ( Figure 7) and (iii) inclusion of foreign atom (inclusion of carbon) ( Figures 5, 7 and 8). All the three modifications synergistically make rTiO 2 (G) CNW (S4 (800)) as the visible light active and efficient photocatalyst. While the nanostructuring provides large surface area, the other two modifications (anatase-rutile phase transformation and carbon inclusion) contribute to the alteration of optical bad gap ( Figure 9) leading to extended visible light absorption [61]. The G inclusion extends the visible light absorption and improves the photocatalytic properties through the role of electron mediator/sink. A plausible mechanism is also suggested for the photocatalysis of rTiO 2 (G) CNW based on the above considerations and experimental observations (Scheme 2). The bulk anatase and rutile materials could show UV light induced photocatalytic activities because of the wide band gaps (~3.20 eV for anatase and~3.0 eV for rutile). On the other hand, the carbon inclusion in rTiO 2 (G) CNW generated mid gap energy states in between conduction and valence bands (Scheme 2) causing lower band gaps (~2.60 eV) ( Figure 9). Besides, the included carbon framework extends the visible light absorption (Figure 9) and suppresses the hole-electron pair charge recombination. The excited electrons were effectively transferred for photochemical processes by the carbon mediation, which in turn increases the lifetime of hole-electron pairs and photocatalytic activities.
anatase and rutile materials could show UV light induced photocatalytic activities because of the wide band gaps (~3.20 eV for anatase and ~3.0 eV for rutile). On the other hand, the carbon inclusion in rTiO2(G) CNW generated mid gap energy states in between conduction and valence bands (Scheme 2) causing lower band gaps (~ 2.60 eV) ( Figure 9). Besides, the included carbon framework extends the visible light absorption ( Figure 9) and suppresses the hole-electron pair charge recombination. The excited electrons were effectively transferred for photochemical processes by the carbon mediation, which in turn increases the lifetime of hole-electron pairs and photocatalytic activities.

Scheme 2.
A plausible mechanism for the visible light photocatalytic activities for rTiO2(G) composite nanowire (CNW) comprising of (a) carbon inclusion in the TiO2 structure and (b) band energy modification due to carbon inclusion for the visible light photocatalytic activities for rTiO2(G) CNW.
Until now, several fundamental studies on synthesis, modification and applications have been reported on 1D TiO2 nanostructured materials [52][53][54]. However, our present study is the first kind to demonstrate the visible light photoactivity of G composited rTiO2 nanostructures. We attribute the following reason for the visible light activity of rTiO2 (G)NW. The possible creation of oxygen vacancy in the TiO2 lattice lowers the band gap [62]. In the present case, the chemical interaction of carboxylated G precursor during the in-situ growth of TiO2 from titanium alkoxides can generate a loosely packed polymeric Ti−O−Ti-C-skeleton. Such Ti−O−Ti-C-skeleton formation makes the disturbed crystal growth can result in depletion of lattice oxygen. Therefore, the presence of G in the rTiO2(G) NW can possibly induce crystal disorder, and to create oxygen vacancies, resulting in the remarkably enhanced visible light activity. Extensive studies are underway to gain deeper insight on the mechanism of photodegradation and visible light augmentation for TiO2 based nanostructures.

Characterization
The morphology of the samples was examined by field-emission scanning electron microscopy (FE-SEM, SU8200, Hitachi, Japan) and field-emission transmission electron microscopy (FE-TEM, Titan G2 ChemiSTEM Cs Probe, FEI Company). The crystal phases of the samples were examined by X-ray diffraction analyzer (D/Max-2500, Rigaku, Japan). The scanning angle 2θ was varied from 5 o to 70 o with a step size of 0.02 and a dwell time of 1.5 sec. The working voltage, applied electric current and Cu Kα radiation were 40 kV, 200 mA and 1.5406 Å, respectively. The band gaps of the samples were determined using adsorption data of thin films recorded by UV-Vis-NIR diffuse reflectance spectroscopy (Cary 5000, Agilent, Santa Clara, CA, USA) in the wavelength region 200 to 700nm. Raman spectra of samples were obtained in the range 100 to 2,700 cm -1 using a Raman spectrometer Scheme 2. A plausible mechanism for the visible light photocatalytic activities for rTiO 2 (G) composite nanowire (CNW) comprising of (a) carbon inclusion in the TiO 2 structure and (b) band energy modification due to carbon inclusion for the visible light photocatalytic activities for rTiO 2 (G) CNW.
Until now, several fundamental studies on synthesis, modification and applications have been reported on 1D TiO 2 nanostructured materials [52][53][54]. However, our present study is the first kind to demonstrate the visible light photoactivity of G composited rTiO 2 nanostructures. We attribute the following reason for the visible light activity of rTiO 2 (G)NW. The possible creation of oxygen vacancy in the TiO 2 lattice lowers the band gap [62]. In the present case, the chemical interaction of carboxylated G precursor during the in-situ growth of TiO 2 from titanium alkoxides can generate a loosely packed polymeric Ti−O−Ti-C-skeleton. Such Ti−O−Ti-C-skeleton formation makes the disturbed crystal growth can result in depletion of lattice oxygen. Therefore, the presence of G in the rTiO 2 (G) NW can possibly induce crystal disorder, and to create oxygen vacancies, resulting in the remarkably enhanced visible light activity. Extensive studies are underway to gain deeper insight on the mechanism of photodegradation and visible light augmentation for TiO 2 based nanostructures.

Characterization
The morphology of the samples was examined by field-emission scanning electron microscopy (FE-SEM, SU8200, Hitachi, Japan) and field-emission transmission electron microscopy (FE-TEM, Titan G2 ChemiSTEM Cs Probe, FEI Company). The crystal phases of the samples were examined by X-ray diffraction analyzer (D/Max-2500, Rigaku, Japan). The scanning angle 2θ was varied from 5 • to 70 • with a step size of 0.02 and a dwell time of 1.5 sec. The working voltage, applied electric current and Cu Kα radiation were 40 kV, 200 mA and 1.5406 Å, respectively. The band gaps of the samples were determined using adsorption data of thin films recorded by UV-Vis-NIR diffuse reflectance spectroscopy (Cary 5000, Agilent, Santa Clara, CA, USA) in the wavelength region 200 to 700nm. Raman spectra of samples were obtained in the range 100 to 2700 cm −1 using a Raman spectrometer (inVia reflex, Reinshaw, Wotton-under-Edge, Glos, England), equipped with a 536 nm laser. Photoluminescence (PL) spectra of samples were obtained using a photoluminescence Spectrophotometer (Spectra Pro 2150i, Acton Research, Lakewood Ranch, FL, USA) under an excitation line of 280 nm.

Visible LightPphotodegradation of NO x
The experiment for the photodegradation of NO x was carried out by the gas-bag A method (standardized by the Korea Photocatalyst Association). The experimental is schematically presented along with the practical set up (Scheme 3). Distilled water was added to 0.5 g of the prepared photocatalyst to sufficiently disperse the photocatalyst. The photocatalyst was coated on a 10 cm × 10 cm glass plate and dried at 100 • C for 1 hour. Then, the photocatalyst coated glass plate was placed in a Tedlar bag having 3L volume. The standard nitrogen oxide gas was diluted with ordinary air to make 3L at a concentration of 1 vol ppm and the diluted nitrogen oxide gas was injected into the Tedlar bag. The Tedlar bag with the sample and the gas was placed in a stainless steel box. The sample was irradiated using the fluorescent lamp (Kumho Electric INC, three wavelength day light color, 17W) placed 30cm above the sample. The illuminance of the fluorescent lamp was 3,450 lm/m 2 on the sample surface. The concentration of NO x was monitored with a NO X analyzer (GV-100s, Gastec) every hour over a period of 6 hours. Raman spectra of samples were obtained in the range 100 to 2,700 cm -1 using a Raman spectrometer (inVia reflex, Reinshaw, Wotton-under-Edge, Glos, England), equipped with a 536 nm laser. Photoluminescence (PL) spectra of samples were obtained using a photoluminescence Spectrophotometer (Spectra Pro 2150i, Acton Research, Lakewood Ranch, FL, USA) under an excitation line of 280 nm.

Visible LightPphotodegradation of NOx
The experiment for the photodegradation of NOx was carried out by the gas-bag A method (standardized by the Korea Photocatalyst Association). The experimental is schematically presented along with the practical set up (Scheme 3). Distilled water was added to 0.5 g of the prepared photocatalyst to sufficiently disperse the photocatalyst. The photocatalyst was coated on a 10cm×10cm glass plate and dried at 100 °C for 1 hour. Then, the photocatalyst coated glass plate was placed in a Tedlar bag having 3L volume. The standard nitrogen oxide gas was diluted with ordinary air to make 3L at a concentration of 1 vol ppm and the diluted nitrogen oxide gas was injected into the Tedlar bag. The Tedlar bag with the sample and the gas was placed in a stainless steel box. The sample was irradiated using the fluorescent lamp (Kumho Electric INC, three wavelength day light color, 17W) placed 30cm above the sample. The illuminance of the fluorescent lamp was 3,450 lm/m 2 on the sample surface. The concentration of NOx was monitored with a NOX analyzer (GV-100s, Gastec) every hour over a period of 6 hours.

Conclusions
In conclusion, the graphene (G) included rutile TiO2 nanowires (rTiO2 NW) of uniform diameters were fabricated by electrospinning of a polymeric solution prepared with low and high molecular weights and subsequent calcination at 800 °C. The pure rutile phase transformation occurred and the inclusion of G into rTiO2 NW resulted in lattice strain and optical band gap narrowing. We attribute that the inclusion of G in titanate network and subsequent high temperature calcination (800 °C) resulted in the production of new energy states and narrowing the band gap. G with its electron accepting capability can suppress the recombination effectively. As a consequence, the visible light absorption and photoactivity greatly enhanced. Thus, our results provide a new insight into the improvement of photoactivity of rutile TiO2. We believe that our study would provide the basis for developing promising visible light photocatalysts based on rutile TiO2 and to explore the utility of commercially available and cheaper rutile TiO2 for prospective applications.

Conclusions
In conclusion, the graphene (G) included rutile TiO 2 nanowires (rTiO 2 NW) of uniform diameters were fabricated by electrospinning of a polymeric solution prepared with low and high molecular weights and subsequent calcination at 800 • C. The pure rutile phase transformation occurred and the inclusion of G into rTiO 2 NW resulted in lattice strain and optical band gap narrowing. We attribute that the inclusion of G in titanate network and subsequent high temperature calcination (800 • C) resulted in the production of new energy states and narrowing the band gap. G with its electron accepting capability can suppress the recombination effectively. As a consequence, the visible light absorption and photoactivity greatly enhanced. Thus, our results provide a new insight into the improvement of photoactivity of rutile TiO 2 . We believe that our study would provide the basis for developing promising visible light photocatalysts based on rutile TiO 2 and to explore the utility of commercially available and cheaper rutile TiO 2 for prospective applications.