Synthesis of Cu , N-Doped TiO 2 Nanotube and Study on Photoelectric Properties

Cu,N-TiO2 nanotube (Cu,N-TNT) is prepared through a novel magnetron sputtering and anodic oxidation method. Then the morphology, structure and physicochemical property of Cu,N-TNT was analyzed by XRD, SEM, TEM, EDX and UV-vis-DR. The results indicate that the evenly doped copper is beneficial to the transformation of the TNT from anatase to rutile and play a key role in the morphology of the Cu,N-TNT. The doped Cu and N in the TNT influence the growth orientation of the TiO2 crystals, which result in the lattice distortion and wider the interplanar spacing 60s-Cu,N-TNT has less band gap and stronger absorption intensity in visible region than other Cu,N-TNT samples, which make the combination rate of photogenerated electron and photogenerated hole decrease greatly, thus beneficial to its physicochemical property.


Introduction
Owing to its cheap, non-toxic and high stability proprieties，TiO2 has been regarded as an important semiconductor nanometer material and could be applied in the field of solar cell, photocatalyst, gas sensor, etc [1][2][3][4][5][6][7].However, the shortcomings of its wide band gap and the fast recombination of the photogenerated electron-hole pairs greatly restricted its practical application [8,9].To overcome the shortcomings, intensive efforts such as implanting an adequate amount of cation or anion and the doping of nonmetallic or metallic elements had been made to extend the work spectrum of TiO2 to the visible (VIS) regions, and to enhance the electron-hole separation [10][11][12][13].At present, it is demonstrated that metal ions are used as dopants can enhance remarkably photocatalytic activity and strong absorption in visible light regions [14,15].Moreover, as an low-cost and abundant metal, copper have tremendous potential in doping for TiO2 [16] .However, to our knowledge, single Cu ion doping has limited improvement in the visible light absorption efficiency and can not meet the needs of practical application of the nanotubes [17,18].Furthermore, it is said that TiO2 modified by nonmetal (such as N, F and S) could narrow the band gap and inhibit the indirect recombination of photogenerated electrons and holes [19][20][21].
Therefore, it is necessary to implant simultaneous two kinds of atoms into TiO2.
So far, many kinds of preparing methods such as ion injection, chemical vaporous deposition and heat-treating process have been used to modify metal ions in TiO2.However, the preparation procedure of these method are usually complicated and the doping amount is relatively low.Furthermore, these solution-based methods can lead to secondary pollution and influence on the morphology of nanotube [22][23][24][25].
On the contrary, the samples prepared by magnetron sputtering have the advantage of composition and thickness uniformity, while the preparation of nanotubes by anodic oxidation is simple and have good repeatability [26][27][28].Therefore, magnetron sputtering and anodic oxidation was used in this paper to prepare Cu,N-TiO2, where the doping amount of Cu can easily be controlled by the sputtering time and N can be doped into the nanotubes by changing the sputtering atmosphere.Furthermore, the photoelectric property of the co-doped TiO2 nanotube arrays were also studied.

Preparation of Cu,N-Ti nanofilms
Cu (99.9 wt.%) and Ti target (99.9wt.%) targets were used to prepare Cu,N-Ti nanofilms on titanium sheets at room temperature (25℃) by magnetron sputtering in a multifunctional magnetron sputtering instrument (JGP560B).Firstly, titanium sheet (30×10×0.1 mm) were ultrasonically cleaned with acetone, ethyl alcohol, and deionized water in sequence, respectively, and then is baked at 105 ℃.Secondly, the mixed gas (the flow rate of Ar:N2=70:1) was flowed into the sputtering chamber when the pressure of the sputtering chamber reached to 8 × 10 −4 Pa.The titanium sheets were pre-puttered for 2 min in order to remove metal oxides and other impurities.The distance between the specimen and the target material is about 30 mm, and the chamber pressure was 1 Pa during sputtering.The thickness of Cu, N-Ti nanofilms was controlled by the power and time of sputtering.The 2 μm Cu,N-TiO2 nanofilms were generated using Ti and Cu target by alternately sputtering.In the process, the total sputtering time of Ti target is 1 hour while the sputtering time of Cu target vary from 20s, 40s, 60s, 80s, and 100s.The Cu,N-Ti nanofilms were calcined in Ar at 450 ℃ for 2 h with a heating rate of 2 ℃/min .

Preparation of Cu,N-TiO2 nanotubes
The above Cu,N-Ti nanofilms and a platinum foil was used as the working electrode and the counter electrode in the electrolyte of a mixture including water, ethylene glycol and 0.02 M NH4F at 25 ℃, respectively.The samples that were generated when the anodization voltage was 30 V, and the anodization time was 45 minutes at 25 ℃ were ultrasonically cleaned with deionized water.The Cu,N-TiO2 nanotubes were calcined in Ar at 450 ℃ for 2 h with a heating rate of 2 ℃/min .

Materials characterization
The morphology and selected area electron diffraction (SAED) patterns of the prepared samples were examined using field emission scanning electron microscopy (FESEM) (Tescan MIRA3 LMH) at 10 kV and transmission electron microscopy (TEM) (JEOL JEM 2100F) at 200 kV.The composition of the prepared samples was determined using energy dispersive spectroscopy (EDS) (OXFORD X-Max 80), and the crystal structure of the prepared structures was analyzed with Cu Kα radiation on a Rigaku D/max 2500 X-ray diffractometer with patterns recorded in a range of 20-80° (Rigaku Smart Lab).Optical absorption of the Cu,N-TNT in the wavelength range of 300-800 nm was measured using a UV-vis spectrophotometer (HITACHI: U-3900).The binding energy was determined using X-ray photoelectron spectroscopy (XPS) (Escalab 250).

Photoelectrochemical activity
Photoelectrochemical activity of the prepared TiO2 nanotubes (or modified TNT) was carried out in a three-electrode cell.The prepared TNT or modified TNT , Pt foil, and SCE (saturated calomel electrode) act as working, counter, and reference electrodes in 0.5 M Na2SO4 electrolyte, respectively.A 500 W Xenon lamp (TrustTech, Beijing, China) with a cut off filter (λ＞420 nm) was used as the light source.The photocurrent was conducted by a CHI660D electrochemical workstation.However, for the 60s-Cu,N-TNT, two new peaks centered at 27.4 o and 36.1 o , which belongs to (110) and (101) planes of rutile TiO2 (JCPDS No.21-1276) appears.In general, it is widely known that it could not be observed the diffraction characteristic peaks of rutile TiO2 when TNT were calcined at 450℃ [29,30].However, it occurs the diffraction characteristic peaks of rutile TiO2 when TNT doped with Cu and N simultaneously were calcined at 450 ℃.The results demonstrate that Cu,N co-doping promote the phase transition from anatate to rutile of TiO2, which should result from decreasing the transition temperature from anatate to rutile of TiO2 according to reported literatures [31].Furthermore, it is observed that besides the peaks belongs to anatase and rutile of TiO2, titanium diffraction peaks can also be detected, but no peaks belongs to copper or copper oxide appears, demonstrating that the Cu element may be exist in the form of ions.At the same time, it is found there is similar grain growth orientation among Cu,N-TNTs that the intensity of (101) plane of anatate TiO2 become weaker and the intensity of (004) plane of anatate TiO2 become stronger with increasing the sputtering time of Cu for Cu,N-TNT samples in Fig. S1.The results indicate that Cu,N-TNT with increasing copper content promote the optimal growth of anatase (004) plane.

Results and discussion
The morphological features of the prepared nanotubes were investigated using FESEM.Fig. 2 (a) and (b) show the top-view and side-view of the TNT.It is presented that TNT had a complete nanotube arrays structure, whose average length, diameter and thickness is 2 μm, 50 nm, and 15 nm, respectively.From the topology images of Fig. 2 (c)-(h), it is observed that the tube diameter and tube length changed little after N or Cu doping, confirming that Cu or N doping has little effect on the morphology of the nanotubes.Furthermore, in the Cu doped nanotubes, no copper or copper oxide nanoparticles can be observed, further demonstrating that the Cu element maybe exist in the form of ions, and consist with our XRD results.Interestingly, when more Cu added in the N-nanotubes, the tubular structure collapses and cannot be maintained after annealing at 450 ℃, which confirms buy the SEM images in Fig. S2, meaning that TNT nanobutes were destroyed with the increase of Cu content in Cu,N-TNT samples.The above phenomena could result from that excess copper content for Cu,N-TNT samples inhibits the generation of nanotubes during the anodic oxidation process.The special selected HRTEM and SADP of 60s-Cu,N-TNT were separately listed in Fig. 4. The diffraction pattern of the selected region is listed in Fig. 4 (b) that is consistent with rutile TiO2 and the lattice distance is 3.52 Å that agrees with (110) planes of rutile TiO2, which suggest that 60s-Cu,N-TNT contains rutile phase composition and is identical with the results in Fig. 1.Meantime, it is observed that the lattice space is 2.37 Å that agrees with (004) plane of anatase TiO2 and there is a large area of (004) plane in Fig. 4 (d), which indicated that promoted the optimal growth of (004) crystal plane among TNT and modified samples [33][34][35].Furthermore, there are obviously lattice distorted in Fig. 4  The UV-Vis spectra of TNT, 60s-Cu-TNT, N-TNT, 60s-Cu,N-TNT is listed in Fig. 5.It is found that the absorption edge of TNT is ～387 nm, and the absorption peak generates bathochromic shift in ultraviolet visible absorption spectrum.The absorption edge follows in the order of TNT< Cu-TNT < N-TNT< 60s-Cu,N-TNT.
The band gap of 60s-Cu,N-TNT is 2.93eV, which has a 0.27 eV shift than that of TNT, whose band gap is 3.2 eV when calculating from the equation: (ahv) 2 = k(hv-Eg) (a: absorption coefficient, h: Planck constant, v: light frequency, Eg: band energy) [36,37].The results indicate it is an effective method to decrease the band gap of TNT and increase the absorption intensity in visible region after Cu and nitrogen element doped.
The above phenomena could put down to that the crystal structure of TiO2 is destroyed to form impurity level between conduction band and valence band in the process of TNT doped with Cu and N, which could prevent photogenerated-electron and photogenerated-hole recombine and result in less band gap and stronger absorption intensity in visible spectra [38].Therefore, the Cu,N-TNT film was further are studied in detail.
The EDS analysis for section graphs of 60s-Cu,N-TNT is shown in Fig. S3.The results indicate that the major ingredient of 60s-Cu,N-TNT is Ti and O, and while the content of evenly distributed copper and evenly distributed nitrogen is 1.0 at.%Cu and 5.1 at.%N, which ascertain that TNT was successfully modified by Cu and N element.Cu + [39] .The N1s at 399.8 eV and 399.(namely 60s-Cu,N-TNT).The above phenomena could result from morphological difference, namely, TNT was destroyed for 80s-Cu,N-TNT and 100s-Cu,N-TNT, which give rise to decreasing contact area between Cu,N-TNT samples and electrolyte [41,42].

Conclusion
Cu,N-TNT was generated by magnetron sputtering and anodic oxidation, which has the larger absorption edge, stronger absorption intensity in visible region, less band gap than TNT.Moreover, the photocurrent density of 60s-Cu,N-TNT is large with excellent sensitivity, which could be applied widely as promising photoanode material.As a result, the present work provides a good guideline to the design of photoanode with excellent performance.

Fig. 1
Fig. 1 XRD patterns of the prepared TiO2 nanotubes (a)TNT; (b)N-TNT; (c)60s-Cu-TNT; (d)60s-Cu,N-TNT X-ray diffraction (XRD) was carried out on TiO2 nanotubes, N doped TiO2 nanotubes Cu 60 s doped TiO2 nanotubes and N-Cu 60s co-doped TiO2 nanotubes, which are referred to as TNT, N-TNT, 60s-Cu-TNT and 60s-Cu,N-TNT, respectively to confirm their phase structure and crystal transformation.Fig. 1 shows the XRD patterns of the nanotubes calcined at 450 ℃.It is obviously observed that all the

Fig. 2 and
Fig. 2 SEM images of the TiO2 nanotubes arrays; (a) overview of TNT; (b) side view of TNT; (c) overview of N-TNT; (d) side view of N-TNT; (e) overview of Cu-TNT; (f) side view of Cu-TNT; (g) overview of Cu, N-TNT; (h) side view of Cu, N-TNT Further analysis the morphology difference of the nanotubes after Cu or N doping, the morphological features of the prepared nanotubes have been investigated using TEM.Fig. 3a to h show the TEM and HRTEM images TNT, N-TNT, Cu-TNT

Fig. 6
Fig. 6 XPS spectra of 60s-Cu-TNT (a)Full spectrum diagram of 60s-Cu-TNT; (b)Cu2p; (c)N1s; (d)Ti2p; (e)O1s XPS spectra were measured to further investigate the chemical bonding state through of the elements.Figure 6(a) shows a general scan spectrum of XPS over a large energy range at low resolution of the prepared 60s-Cu-TNT.It is evident that the 60s-Cu,N-TNT is composed of Ti, O, N, C and Cu elements , which is consistent with EDS results in the supporting information in Fig.S3.Trace amounts of C is attribute to

Fig. 8
Fig. 8 Transient photoelectric response maps of (a)20s-Cu,N-TNT; (b)40s-Cu,N-TNT; (c)60s-Cu,N-TNT; (d)80s-Cu,N-TNT; (e)100s-Cu,N-TNT It is listed that the relationship between photocurrent density and the doping time of copper of Cu,N-TNT samples in Fig.8.It is seen that the photocurrent density of Cu,N-TNT electrode increases firstly then decreases with increasing percentage composition, and reached the maximum rate when the doping time of copper is 60s