Metallic 1T-TiS 2 nanodots anchored on a 2D graphitic C 3 N 4 nanosheet nanostructure with high electron transfer capability for enhanced photocatalytic performance

Photocatalysis is one of the most promising technologies for solar energy conversion. With the development of photocatalysis technology, the creation of low-dimensional structure photocatalysts with improved properties becomes more and more important. Metallic 1T-TiS 2 nanodots with a low-dimensional structure were introduced into environmentally friendly two-dimensional g-C 3 N 4 (2D-C 3 N 4 ) nanosheets by a solvothermal method. It was found that the ultrathin TiS 2 nanodots were uniformly anchored on the surface of the 2D-C 3 N 4 . The e ﬀ ective suppression of electron – hole recombination was realized due to the addition of the intrinsic metallic property of 1T-TiS 2 in the prepared nanocomposite. The 5 wt% TiS 2 /2D-C 3 N 4 nanocomposite exhibited the best photocatalytic performance and the degradation rate towards RhB was ca. 95% in 70 min, which showed an improvement of ca. 30% in comparison with 2D-C 3 N 4 . The results indicate that the obtained TiS 2 /2D-C 3 N 4 nanocomposite is a promising photocatalyst for practical applications.


Introduction
With the increasing scarcity of conventional energy resources and the deterioration of the environment, the development and utilization of renewable energy becomes more and more important. 1 Solar power can be converted by photocatalysts into chemical energy to degrade pollutants, which has attracted a lot of attention in the past decades. 2 Numerous researchers have devoted themselves to creating efficient photocatalysts to take full advantage of the redox ability of photogenerated carriers. TiO 2 was reported to be used for hydrogen generation via water splitting under ultraviolet light. [3][4][5][6] Nevertheless, the challenges of photocatalytic degradation mainly lie in the following two points: (1) the wide light absorption range and (2) the efficient separation of photogenerated electrons and holes. 7,8 In recent years, researchers have found the dimension reduction of the photocatalysts can shorten the diffusion length of photogenerated carriers. [9][10][11][12][13][14][15][16] Moreover, the introduction of conducting materials has been found to be able to suppress recombination of photogenerated electron (e)-hole (h) pairs of semiconductors, such as some noble metals (Ag, Au and Pt). [17][18][19][20][21][22] Platinum (Pt) with the lowest Fermi level is the most effective accepter of photogenerated electrons for the photocatalytic reaction. 23 However, its high material cost makes it uncompetitive for practical applications. Thus, the development of the substitute materials (noble metal-free) of the noble metals is very necessary, but still challenging to date.
Layered transition metal dichalcogenides (TMDs) such as MoS 2 , WSe 2 , TaS 2 etc., have received much attention owing to their excellent catalytic activity and low cost compared to noble metals. [24][25][26][27][28][29][30] As a prototype of TMDs, 1T-TiS 2 is composed of metal Ti layer sandwiched between two S layers forming edgesharing TiS 6 octahedra with strong covalent forces. 3 The adjacent S-Ti-S layers are coupled to each other by weak van der Waals interactions, providing the practical feasibility for exfoliating the bulk TiS 2 to ultrathin two dimensional TiS 2 nanosheets. 31,32 In addition, 1T-TiS 2 is a semimetal with excellent inplane conductivities. 33 It is noticeable that conductive 1T-TiS 2 can replace the noble metals to be used as the accepter of the photogenerated electrons and improve the separation efficiency of the photoexcited carriers of semiconductors.
Graphitic carbon nitride (g-C 3 N 4 ), a polymeric semiconductor with a low band gap of $2.7 eV, has an appropriate band structure for photocatalysis. 34 Two dimensional C 3 N 4 (2D-C 3 N 4 ) can be prepared according to our previous report. [35][36][37] The two dimensional structure of g-C 3 N 4 can contribute to the separation of photogenerated e-h pairs and then the carriers could handily transfer to the surface to suppress the recombination. [38][39][40][41][42][43] However, the separation efficiency of the photoexcited carriers of 2D-C 3 N 4 are still unsatisfactory in the absence of noble metals. To further improve the photocatalytic activity of 2D-C 3 N 4 , the introduction of some conductive materials is an effective method.
Herein, the exfoliated ultrathin TiS 2 nanodots were introduced into 2D-C 3 N 4 . Anchoring metallic TiS 2 nanodots on 2D-C 3 N 4 nanosheets via solvothermal method facilitated the fast transfer of the photogenerated carriers. The high photocatalytic performance of the nanocomposite resulted from the effective suppression of the e-h recombination was tested by degrading Rhodamine B (RhB). In the nanocomposites, uniformly distributed TiS 2 nanodots on surface of 2D-C 3 N 4 nanosheets did serve as the accepter of the photogenerated electrons. Additionally, the photocatalytic mechanism was also studied in detail by electron spin resonance (ESR).

Synthesis of ultrathin TiS 2 nanodots
Ultrathin TiS 2 nanodots were prepared via liquid phase exfoliation in a co-solvent (acetonitrile/IPA ¼ 19 : 1, by volume). Specically, 900 mg 1T-TiS 2 powder (Sigma-Aldrich) was dispersed in 300 ml co-solvent and the solution was sonicated in an ultrasonic bath at 25 C for 4 h. Then the mixture was centrifuged at 6000 rpm for 30 min to remove any unexfoliated bulk TiS 2 . The supernatant was further probe sonicated for another 4 h followed by centrifugation at 10 000 rpm for 30 min. The freeze-dried supernatant was named as TiS 2 nanodots.

Synthesis of ultrathin g-C 3 N 4 nanosheets
Ultrathin g-C 3 N 4 was synthesized as below. 2 g melamine were calcined at 550 C for 4 h with 2 C min À1 heating rate in the muffle furnace. The obtained sample was bulk g-C 3 N 4 . Then 400 mg bulk g-C 3 N 4 was ground and heated at 550 C for $30 min in the muffle furnace. Aer that, the obtained samples were heated at 550 C for another $30 min in the muffle furnace. Finally, the obtained ultrathin g-C 3 N 4 were white.

Synthesis of TiS 2 /2D-C 3 N 4 nanocomposites
The TiS 2 /2D-C 3 N 4 nanocomposite was prepared by solvothermal method in benzyl alcohol. An appropriate amount of ultrathin TiS 2 nanodots and 70 mg 2D-C 3 N 4 was dispersed in benzyl alcohol and sonicated for 20 min, respectively. Then a certain concentration of TiS 2 nanodots was mixed with pure 2D-C 3 N 4 and stirred for 2 h. Aer sonicating the mixture again for 20 min, the suspension was transferred to a stainless-steel autoclave and heated for 4 h at 140 C. Aer natural cooling, the solution was washed by ethanol and water twice respectively. The washed powder was then freeze dried and used as catalyst. The TiS 2 /2D-C 3 N 4 catalysts with different TiS 2 content were named as x% TiS 2 /2D-C 3 N 4 , where "x" represent the mass percentage of TiS 2 (x ¼ 1, 3, 5, 10 wt%).

Material characterization
Scanning electron microscope (SEM) images of 5% TiS 2 /2D-C 3 N 4 nanodots were taken on a JEOL-6500 scanning electronic microscope. Transmission electron microscope (TEM) and high resolution TEM (HRTEM) images were recorded on a JEOL-2100F at an accelerating voltage of 200 kV. X-ray diffraction (XRD) patterns of TiS 2 nanodots and the nanocomposites were performed on D/MAX2500V with Cu-Ka radiation (l ¼ 1.54056 A). Raman spectra of TiS 2 nanodots, 2D-C 3 N 4 and 5% TiS 2 /2D-C 3 N 4 nanocomposite were acquired with a RENISHAW in Via Raman Microscope using a 532 nm laser excitation. X-ray photoelectron spectroscopy (XPS) measurements were performed to analyze the presence of TiS 2 in the nanocomposites at ambient temperature using PHI Quantera with Al-Ka X-ray source. Shimadzu UV-2450 ultraviolet-visible spectrophotometer was used to collect the UV-vis absorption spectra of TiS 2 /2D-C 3 N 4 nanocomposites. Composition analysis of as-prepared materials was carried on fourier transforms infrared spectrometer (FT-IR) using Nicolet Nexus 470 spectrometer. The ESR spectra were conducted on a Bruker model ESR JES-FA200 spectrometer. Photoluminescence (PL) spectroscopy experiments were conducted at excitation wavelength 377 nm using Jobin Yvon HORIBA NanoLog spectrouorometer.

Photocatalytic activity measurements
For photocatalytic degradation experiment, the organic dye RhB was used as a model pollutant. In detail, 10 mg samples were added into 50 ml RhB (10 mg l À1 ) in a Pyrex photocatalytic reactor with a circulating water system to maintain a constant temperature (30 C). Before irradiation, the suspensions were magnetically stirred for 30 min in the dark to ensure that RhB could reach the absorption-desorption equilibrium on the photocatalyst surface. At a certain time intervals, 3 ml aliquots were sampled and centrifuged to remove the photocatalyst nanoparticles. Then the ltrates were analyzed by recording variations of the absorption band maximum (553 nm) in the UVvis spectra of RhB using a UV-vis spectrophotometer. The air velocity was 2 l min À1 and the photocatalytic reaction was performed under a 300 W Xe lamp with a 400 nm cutoff lter.

Results and discussions
The TiS 2 /2D-C 3 N 4 nanocomposites with different content of TiS 2 nanodots were prepared by solvothermal method in benzyl alcohol. The morphology of 2D-C 3 N 4 , TiS 2 nanodots and the asprepared 5% TiS 2 /2D-C 3 N 4 nanocomposite were examined by AFM and TEM, as shown in Fig. 1. 2D-C 3 N 4 nanosheets performed a micrometer-scale lateral size and thin thickness about 1-2 nm (Fig. 1a). The size and thickness of TiS 2 nanodots were nanometer grade, 5-10 nm and 1-3 nm (Fig. 1b and c), respectively. Ultrathin and small TiS 2 nanodots were uniformly anchored on the surface of thin 2D-C 3 N 4 nanosheets without hard aggregations. From the HRTEM image, the lattice with spacing of 0.29 nm of TiS 2 nanodots assigned to the (010) planes could be seen on the surface of 2D-C 3 N 4 nanosheets clearly. It demonstrated that the hybridization of 2D-C 3 N 4 nanosheets and TiS 2 nanodots did not disturb the crystal structure and morphology of TiS 2 .
The crystalline structure of the as-prepared TiS 2 /2D-C 3 N 4 nanocomposites with different content of TiS 2 nanodots were further characterized by XRD. In addition to the diffraction peaks at 12.5 and 27.7 attributed to the (100) and (002) crystal planes of 2D-C 3 N 4 , respectively, 44,45 many other peaks for intralayer crystal planes at 15.5, 34.2, 44.2, 53.9, 57.7 and 65.4 of TiS 2 /2D-C 3 N 4 nanocomposites were consistent with those of pure few layers of 1T-TiS 2 . 46 Raman spectra of 5% TiS 2 /2D-C 3 N 4 nanocomposite, 2D-C 3 N 4 and TiS 2 nanodots was taken to further demonstrate the presence of TiS 2 in the nanocomposites (Fig. 2c). The pure TiS 2 nanodots showed a peak at 227 cm À1 assigned to E g and a peak at 332 cm À1 with a shoulder at 380 cm À1 attributed to A 1g vibrational modes of 1T-TiS 2 . 47 However, except for the signicant peaks of 2D-C 3 N 4 , the peaks of TiS 2 were not very clear in 5% TiS 2 /2D-C 3 N 4 because of the low content of TiS 2 in the nanocomposites, which was in agreement with previous publications. 48 FT-IR spectra of pure 2D-C 3 N 4 and TiS 2 /2D-C 3 N 4 nanocomposites were obtained as well. Because of the low detective sensitivity and low content of TiS 2 in the nanocomposites, FT-IR spectra of all hybrid only showed typical stretching mode of 2D-C 3 N 4 heterocycles from 1250 to 1700 cm À1 and the vibrational mode of N-H bond at 3147 cm À1 respectively. 48 The chemical states and elemental compositions of TiS 2 nanodots, 2D-C 3 N 4 and 5% TiS 2 /2D-C 3 N 4 nanocomposites were further examined by XPS, as shown in Fig. 3. The C 1s spectra of both 2D-C 3 N 4 and 5% TiS 2 /2D-C 3 N 4 nanocomposites showed two peaks at 284.7 eV and 288.2 eV assigned to the sp 2 C]C bonds and N]C-N 2 bonds, respectively (Fig. 3a and c). The N 1s spectrum of 5% TiS 2 /2D-C 3 N 4 nanocomposite with four peaks in Fig. 3b is similar with that of 2D-C 3 N 4 in Fig. 3d. The peaks at 398.7 eV, 399.8 eV, 401.0 eV and 404.4 eV were ascribed to the sp 2 -hybridized nitrogen atoms in C]N-C, the tertiary nitrogen, the amino functions carrying hydrogen (C-N-H) and p-excitation, respectively. 49 The C 1s and N 1s signals conrmed the presence of 2D-C 3 N 4 in 5% TiS 2 /2D-C 3 N 4 nanocomposite, which is consistent with the results in FT-IR. The binding energies of Ti 2p peaks (458.5 eV for 2p 3/2 and 463.3 eV for 2p 1/2 ) and the weak binding energies of S 2À 2p 3/2 and 2p 1/2 peaks emerging at 161.0 eV and 162.2 eV in 5% TiS 2 /2D-C 3 N 4 nanocomposite were similar with the peaks in TiS 2 nanodots, which agreed well with those of 1T-TiS 2 . 50 It demonstrated the existence of TiS 2 nanodots in 5% TiS 2 /2D-C 3 N 4 composite.
The presence of TiS 2 nanodots in the nanocomposites was further studied by SEM images and corresponding EDS. The micro-morphology and element mapping of 5% TiS 2 /2D-C 3 N 4 nanocomposite were shown in Fig. 4. The 2D nanosheets could be seen clearly in Fig. 4a and d. Because TiS 2 nanodots and 2D-C 3 N 4 nanosheets were self-assembled together, it was difficult to distinguish them from morphology. From the element mapping of 5% TiS 2 /2D-C 3 N 4 nanocomposite, it could be seen    that TiS 2 nanodots were dispersed in the so nanosheets of 2D-C 3 N 4 uniformly. Ti and S elements of TiS 2 nanodots were distributed in the matrix C and N elements of 2D-C 3 N 4 in the nanocomposites, which was consistent with the TEM results. As the content of TiS 2 nanodots in composite was 5%, the signals of Ti and S were much lower than those of C and N. The UV-vis absorption of pure TiS 2 nanodots, 2D-C 3 N 4 and as-prepared TiS 2 /2D-C 3 N 4 nanocomposites were shown in Fig. 5a. The curve of 2D-C 3 N 4 displayed a starting absorption at the edge of visible light blending into the ultraviolet, which corresponded to the band gap of 2D-C 3 N 4 at 2.68 eV in Fig. 5b. And TiS 2 nanodots could absorb UV and visible light. TiS 2 /2D-C 3 N 4 nanocomposites performed similar absorption width with 2D-C 3 N 4 nanosheets but much higher absorption value attributed to the light absorption of TiS 2 , and the absorption intensity at the UV and visible light region were enhanced with increasing content of TiS 2 nanodots which indicated that TiS 2 /2D-C 3 N 4 nanocomposites obtained more photogenerated e-h pairs at 200-700 nm. In order to investigate the inuence of TiS 2 QDs on the separation efficiency of the photoexcited carriers in TiS 2 /2D-C 3 N 4 nanocomposites, the PL spectroscopy measurements were carried out for 2D-C 3 N 4 and 5% TiS 2 /2D-C 3 N 4 nanocomposites. As shown in Fig. 5c, the uorescence intensity of emission peak at about 460 nm for 5% TiS 2 /2D-C 3 N 4 nanocomposite was much weaker than that of 2D-C 3 N 4 due to the low recombination rate of photo-generated charge carriers. 51 Because of the higher UVvis absorption and lower uorescence emission intensity, more photo-generated charge carriers in 5% TiS 2 /2D-C 3 N 4 nanocomposite were trapped by highly conductive TiS 2 nanodots on 2D-C 3 N 4 nanosheets.
The photocatalytic performance of the nanocomposites was mainly evaluated by photo-degradation of RhB, because RhB is a typical organic dye which is a common water pollutant and could cause long-term environmental toxicity and short-term public health damage. Fig. 6a showed the evolution of the degradation rate of RhB along with time. All the curves had a similar linear downtrend which seemed to be zero-order kinetic process. Hence, the data points were tted according to the zero-order kinetic process: where k is the degradation rate constant, and b represents the residual composition at 0 min. The k, b and R 2 values of the tted curves for all the samples were listed in Table 1. According to R 2 values of the samples (higher than 0.982), the tted curves matched well with the obtained data. 52 So, the degradation behaviour surely followed the zero-order kinetic process. The degradation rate constant of 2D-C 3 N 4 was 0.00783 min À1 . Compared with blank 2D-C 3 N 4 , all the TiS 2 /2D-C 3 N 4 nanocomposites exhibited higher photocatalytic degradation rate under the same condition. The addition of TiS 2 nanodots to 2D-C 3 N 4 enhanced the photocatalytic activity obviously. As the amount of TiS 2 nanodots increase, the degradation rate of RhB was enhanced signicantly. 5% TiS 2 /2D-C 3 N 4 nanocomposite performed the highest degradation rate among all the catalysts (0.0128 min À1 ). When the loading amount of TiS 2 nanodots reached to 10 wt%, the photocatalytic degradation of RhB decreased as too much TiS 2 nanodots blocked the photo absorption of 2D-C 3 N 4 . The as-prepared 10% TiS 2 /2D-C 3 N 4 has higher b value than 5% TiS 2 /2D-C 3 N 4 nanocomposite because of weaker absorption ascribed to too high loadings of TiS 2 nanodots. The highest degradation rate of TiS 2 /2D-C 3 N 4 nanocomposites was ca. 95% in 70 min, which got an improvement of ca. 30% in comparison with that of 2D-C 3 N 4 . The photocurrent experiment was carried out to explore the charge separation and transfer efficiency (Fig. 6b). The measurements clearly showed that the introduction of TiS 2 nanodots enhanced the photocurrent of 2D-C 3 N 4 , in which the 5% TiS 2 /2D-C 3 N 4 nanocomposite exhibited the highest value. The photocurrent for 5% TiS 2 /2D-C 3 N 4 was about 2.5 times higher than that given by bare 2D-C 3 N 4 . The results above suggested that the introduction TiS 2 nanodots enhanced the charge separation greatly and suppressed the recombination of e-h pairs, thus an improving the photocatalytic performance.
Besides the high photocatalytic degradation efficiency, the 5% TiS 2 /2D-C 3 N 4 nanocomposite showed high stability aer ve times of cyclic experiments (Fig. 7a). In addition, the nanocomposite still preserved the chemical structure of 2D-   were consistent with nanocomposite before cycling. FT-IR spectra of TiS 2 /2D-C 3 N 4 nanocomposites before and aer cycling was similar as well.
As discussed above, the metallic TiS 2 nanodots in the system played a crucial role in the migration of electrons excited by the 2D-C 3 N 4 semiconductor, which could efficiently suppress the recombination of e-h pairs. To gain a further insight into the photocatalytic degradation process, ESR analysis was carried out to explore the active radicals using 5,5-dimethyl-1-pyrroline N-oxide (DMPO) as a spin trapping, as shown in Fig. 8. Obviously, the ESR results showed that, when TiS 2 /2D-C 3 N 4 nanocomposite were employed as photocatalyst, both O 2 c À and cOH could be generated under visible light irradiation. According to previous reports, the top of valence band (TVB) potential of 2D-C 3 N 4 was 1.79 V vs. RHE. 53,54 The band gap was calculated to be 2.68 eV (Fig. 5b), thus the bottom of conduction band (BCB) potential of 2D-C 3 N 4 was À0.89 V vs. RHE. Apparently, the potential energy of the photo-excited electrons at BCB potential was higher than that of O 2 c À /O 2 (À0.046 eV), thus the electrons possessed the ability to combine with O 2 into O 2 c À . For the generation of cOH, the holes at the TVB of 2D-C 3 N 4 cannot oxidize H 2 O or OH À into cOH due to the lower TVB position in comparison with OH À /cOH (1.99 eV) and H 2 O/cOH (2.38 eV). The cOH may be generated by the further reactions of strong reduction O 2 c À through an intermediate of cOOH: [55][56][57] O 2 c À + e À + H + / cOOH (1) Fig. 9 is a schematic illustration of the synergistic effect of the photocatalytic process. In a typical photocatalytic degradation process, the 2D-C 3 N 4 nanosheets could photogenerated e-h pairs under visible light rstly. Aerwards, the electrons kept at conduction band would be efficiently trapped by metallic TiS 2 nanodots as co-catalyst. The high concentration of electrons could react with O 2 to form active radicals O 2 c À . Due to the strong reduction of the photogenerated electrons, the formed O 2 c À also possessed very strong reduction which led to the further reactions and formation of cOH radicals through an intermediate of cOOH as listed from reaction (1) to (3). The radicals O 2 c À and cOH were well known as effective radicals for water pollutant treatment. In addition, the hole at valence band of 2D-C 3 N 4 could oxidize pollutant directly. The metallic nature of TiS 2 nanodots and the construction of tight interface between the two components were benecial for the separation and transfer of photo-excited carriers. Therefore, high concentration of photogenerated electrons could converge for photocatalytic reactions at TiS 2 nanodots instead of recombination at 2D-C 3 N 4 nanosheets.

Conclusion
We utilized a simple exfoliation synthesis to get ultrathin TiS 2 nanodots and the TiS 2 /2D-C 3 N 4 was synthesized by a temperate solvothermal method. The TiS 2 nanodots were tightly anchored on the surface of 2D-C 3 N 4 nanosheets and the inherent metallic nature of 1T-TiS 2 improved the photocatalytic degradation activity due to the effective suppression of the e-h recombination and the higher light absorption. Noteworthy, the content of the TiS 2 nanodots in the prepared complex played a role in the RhB photocatalytic degradation. The 5% TiS 2 /2D-C 3 N 4 exhibited the best photocatalytic performance and the degradation rate towards RhB was ca. 95% in 70 min, which got an improvement of ca. 30% in comparison with pure 2D-C 3 N 4 nanosheets. So, the combination of ultrathin metallic TMDs with 2D graphitic C 3 N 4 yields a promising photocatalyst for practical application.   9 The degradation of mechanism of as-prepared TiS 2 /2D-C 3 N 4 composites. Fig. 7 (a) Cycling runs for the photodegradation of RhB in the presence of 5% TiS 2 /2D-C 3 N 4 nanocomposite. The chemical and crystal structure of 5% TiS 2 /2D-C 3 N 4 nanocomposite before and after 5 cycles photodegradation of RhB, XRD pattern (b) and FT-IR spectra (c).