Fabrication of Hierarchical Nanocomposites through a Nature- Mimic Method: Depositing MoS2 Nanoparticles on Carbon Nitride Nanotubes by Polydopamine Coating

The combination of 1D nanotubes with 0D nanoparticles to integrate a nanocomposite structure has attracted increasing research interest, while the interfacial interaction plays an important role in such composites. This paper presents a facile and universal approach for the fabrication of hierarchical NP-NT nanocomposites with improved photocatalytic performances by mussel chemistry. Polydopamine (PDA) serves as the biomimetic adhesive layer and then connects MoS2 nanoparticles with g-C3N4 nanotubes (CNNTs). The obtained nanocomposites were characterized by FT-IR spectra, XRD, SEM, TEM, XPS, UV-vis DRS, and PL. Compared with unmodified CNNTs, the as-prepared MoS2-PDA-CNNT composites exhibited an enhanced photocatalytic properties for the degradation of methylene blue (MB) under visible light irradiation. This research would provide a green and versatile method to construct hierarchical structured nanocomposites with high catalytic activity.


Introduction
For the past few years, graphic carbon nitride (g-C 3 N 4 ) attracts increasing interest in the research areas of photocatalytic water splitting [1,2], organic synthesis [3], and degradation of pollutants [4][5][6][7]. Generally, the g-C 3 N 4 has been widely utilized in many research areas due to its steady physical-chemical properties, narrow band gap, and inexpensive preparation cost [8]. However, the drawbacks of g-C 3 N 4 such as its bulk layered structure, rapid photogenerated recombination of electron-hole (e − -h + ) pair, and low efficiency in the use of visible light induce a low visible-light activity, which limited its further applications [9]. The nanostructural design of g-C 3 N 4 and the development of heterostructured nanocomposites based on g-C 3 N 4 are two main methods for improving the photocatalytic performance of g-C 3 N 4 under visible light irradiation. By the first method, metals [10], metal oxides [11], metal sulfides [12], and some other components [13] were usually used to modified g-C 3 N 4 . Besides, g-C 3 N 4 with hierarchical nanostructures such as mesopores [9], nanorods [14],, nanosheets [15] (hollow), nanospheres [16,17], and nanotubes [18,19] was also successfully constructed by the researchers. As one known, a tubular nanostructure can provide both internal and external exposed active sites for photocatalytic reactions, promote photogenerated electron transfer, and reduce the resistance to mass transfer of photocatalytic reactions. Huang and coauthors [20] fabricated porous g-C 3 N 4 nanotubes by heating precursors synthesized by melamine recrystallization from H 2 SO 4 /methanol mixed solution. Compared with bulk g-C 3 N 4 obtained by direct pyrolysis, the porous g-C 3 N 4 nanotubes exhibit improved photocatalytic activity for the photocatalytic water splitting into hydrogen under visible light irradiation.
Over the past decade, integrating 1D nanotubes with 0D nanoparticles (NP) to develop a hybrid structure has aroused great interest. This hybridization often leads to intriguing structural, electromagnetic, electrochemical, and photochemical properties which were not shown by the separate components alone [21,22]. For instance, Li's group [21] deposited carbon nitride nanotubes with platinum nanoparticles (Pt/CNNT) by a facile one-pot solvothermal treatment method. Pt/CNNTs as prepared exhibited improved visible light photocatalytic activity toward the water splitting as well as pollutant degradation [21].
Nevertheless, it is still so challenging to strongly bind the NPs onto NTs due to the weak binding interaction between them [23]. Traditionally, researchers have introduced oxygen-containing functional groups (such as -OH, -COOH, or -C=O) through a complex process of surface modification to strongly anchor the NPs on the NTs.
In nature, 3,4-dihydroxy-L-phenylalanine (dopamine, DA) was abundant in marine mussel adhesive proteins (MAPs). Interestingly, it is via the oxidation of the catechol group in a slightly alkaline solutions (pH > 7:5), and DA can easily self-polymerize and be coated on many types of organic or inorganic substrates [24]. Yu et al. [25] employed polydopamine (PDA) to modify the surface of bulk graphic carbon nitride, and the PDA-g-C 3 N 4 exhibited increasing photocatalytic activity. Due to its versatile adhesion ability, PDA is utilized as the connecting layer to decorate substrates with NPs. In addition, polydopamine can provide an efficient platform for complementary interface reactions. In particular, functional groups such as the amine (-NH 2 ) and catechol/quinine group (-OH/=O) at the ring can serve as good ligands for binding metal ion and reducing noble metal ions to noble metal NPs [25].
Recently, molybdenum sulfide-(MoS 2 -) based catalysts have attracted interest in the field of photocatalysis. It was utilized in photocatalytic water splitting into hydrogen [26] as well as in the degradation of various dyes [27], due to its narrow band gap [28] (visible to near-infrared), high absorption of visible light, low toxicity, and abundant of active sites. Meanwhile, because of its too narrow band gap and low quantum efficiency, MoS 2 has been widely used as a cocatalyst for creating heterostructure with Bi 2 O 3 [29], MoO 3 [30], ZnO [31], and especially g-C 3 N 4 [26,32].
Inspired by the aforementioned researches, we modified carbon nitride nanotubes (PDA-CNNTs) by polydopamine coating. Due to this biomimetic adhensive PDA layer, MoS 2 nanoparticles were easily deposited. Specifically, the structure, morphology, and photocatalysis properties of the asprepared MoS 2 -PDA-CNNT nanocomposites have been systematically studied.

Preparation
2.2.1. Preparation of g-C 3 N 4 Nanotubes. The preparation of the holey g-C 3 N 4 nanotubes (CNNTs) is according to the published literature with some modification [33]. Typically, melamine (520 mg) was added in ethylene glycol (20 mL) and stirred for dissolution under room temperature. The 0.12 mol L -1 nitric acid aqueous solution (60 mL) was added dropwise to the above solution. The reaction is stopped after stirring for 10 min. After collecting the generated precipitation by centrifugation and washing three times using ethanol, vacuum drying at 55°C for 8 h was applied to remove the volatile solvent. The obtained white precursor was annealed at 350°C for 1 h using a heating rate of 8°C·min -1 . The as-prepared samples are denoted as CNNTs.
2.2.2. Preparation of Dopamine/g-C 3 N 4 Nanotubes. PDA-CNNTs were fabricated by a solution method. In brief, of dopamine-hydrochloride added to the aqueous dispersion of CNNTs for 24 h, CNNTs (500 mg) were dispersed into 100 mL deionized H 2 O, and the suspension was under continuously ultrasonic treatment. After about 30 min, a certain amount of dopamine-hydrochloride was added to the above suspension. And then the mixture suspension was for 60 min under continuously stirring at room temperature. A desired concentration of Tris-HCl solution (100 mL) was added, and 1 mol L -1 NaOH solution was added to adjust the pH (pH = 8:5) of the mixture. The polydopamine coated g-C 3 N 4 nanotubes (PDA-CNNTs) were obtained by centrifugation [25].

Preparation of MoS 2 /Dopamine/g-C 3 N 4 Nanotubes.
PDA-CNNTs were redispersed in 10 mmol L -1 Tris buffer solution (500 mL, pH 8.5), and then 500 mg nano-MoS 2 nanoparticles were added to the solution with magnetically stirred vigorously at 60°C for 24 h. And then the unreacted dopamine and MoS 2 nanoparticles were removed by Tris-HCl buffer and deionized water three times, and solid samples were collected by centrifugation. Finally, the composite was dried at 60°C in air for 24 h and is denoted as MoS 2 -PDA-CNNTs.

Characterization.
The scanning electron microscopy (SEM, Philips XL30 ESEM, and Hitachi S-4800 instrument at an operation voltage of 20.0 keV and 0.7 keV) and transmission electron microscopy (TEM; Tecnai G2 F20 with accelerating voltage of 200 kV) were performed to observe the surface morphology and microstructures of as-prepared samples. Fourier transform infrared spectra (FTIR) of the samples were measured on FTIR Spectrometer (FTIR; Bruker Tensor 27). The power X-ray diffraction (XRD) was tested by an XRD-6100X diffractometer (Shimadzu, Japan) with a Cu Kα source. The scan rate is 2°min -1 (from 5°to 70°). The Xray photoelectron spectroscopy (XPS) was tested by XPS spectroscopy (Thermo Fisher Scientific, UK). The UV-vis diffuse reflectance spectra (UV-vis DRS) were measured on a UV-3600 Plus spectrometer (Shimadzu, Japan). The photoluminescence measurements (PL) were recorded on fluorescence spectrophotometer (F-7000, Hitachi, Japan) with 273 nm laser illumination.

Photocatalytic Experiments.
The photocatalytic degradation efficiency of pure CNNTs and composites was evaluated by removal of methylene blue (MB) under visible light irradiation [25]. The visible light source was a 500 W Xe lamp (λ > 420 nm), and the illumination distance was kept at about as-prepared photocatalyst (100 mg) was dispersed in 20 mg·L -1 MB solution (100 mL). Prior to light irradiation, the system was under magnetically stirring in the dark for 30 min to ensure the as-prepared photocatalyst reach the adsorption equilibrium. After certain time intervals (30 min), the reaction solution was collected by centrifugation to remove the as-prepared photocatalyst for subsequent analysis. The residual concentrations of MB were monitored by UV-vis spectroscopy (UV-762; Shanghai China), and the photodegradation efficiency was calculated.

Results and Discussion
The tentative formation mechanism of MoS 2 -PDA-CNNTs was schematically illustrated in Scheme 1. The process included four steps: melamine dissolving, self-assembly to 1D fibrous, thermalized to tubular structure, biomimetic coating, and nanoparticle depositing. Melamine could be slightly dissolved in ethylene glycol, and then dropwise addition of an aqueous solution of nitric acid to the glycol solution saturated with melamine gradually leads to a white precipitation consists of a fibrous structure. In the above process, we can easily protonate the melamine by reacting the amine in melamine with HNO 3 , thereby reducing its solubility in ethylene glycol. The protonated melamine selfassembled with nitrate ions and turned into fibrous structures. With a heating process, the protonated melamine forms tris-s-triazine rings and polymerizes to form molecular ribbons. Then, the molecular ribbons overlap each other and form stable stacking layers of a certain thickness through p-p electron interaction. Next, the stacking layer tends to roll up into nanotubes due to the driving force for minimizing surface free energy. The fibrous structures transformed to nanotubular [33]. Next, the freshly prepared CNNTs were subject to partially crosslinking reaction with DA in a Tris buffer of pH = 8:5. The partial crosslinking process caused by a Schiff base/Michael addition reaction, that is, the interaction between the catechol group of DA/PDA and the primary amine group of CNNTs. After applying the PDA layer onto the CNNTs, MoS 2 nanoparticles were coated onto the surface inspired by dopamine chemistry, and the MoS 2 -PDA-CNNTs were obtained.
The morphology and microstructure of the obtained pure CNNTs and the composites were investigated using scanning electron microscopy (SEM) and transmission electron microscopy (TEM). As can be seen from Figures 1(a) and 1(b), the outer diameter of as-prepared CNNTs is ranging from 300 to 500 nm, and the length of CNNTs reaches several millimeters. The corresponding TEM image (Figure 1(e)) further confirmed the morphology of the tubular elements with pore penetration across the tube wall (as arrows marked in Figure 1(b)) [33]. The TEM image reveals that the thickness of CNNTs wall was measured to approximately 30 nm. The unique properties of these CNNTs will make them very suitable for photocatalytic reaction. Moreover, there are enough published references to prove that the thin tubular structure can promote photocharge separation and inhibit the charge recombination. At the meantime, the tubular structure and the pores into the tube wall should promote the mass transfer of reactants and products, which will also facilitate the photocatalytic reactions. After modifying PDA and subsequent MoS 2 on the surface of CNNTs, the overall tubular morphology of CNNTs was still preserved, as shown in the SEM (Figure 1(d)) and TEM images (Figure 1(e)). The corresponding TEM image also indicates that some MoS 2 protruded nanoparticles have been successfully grafted onto the surface of PDA-CNNTs. The energydispersive X-ray spectroscopy (EDS) indicated that only the elements C to O mass ratio were increased while the N was 3 Journal of Nanomaterials decreased due to the PDA coating. In addition, Mo and S element was indeed distributed in the MoS 2 -PDA-CNNTs [34]. Figure 2 shows FTIR spectra of CNNTs and composites. The characteristic peaks ranging from 1200 to 1650 cm -1 were observed in the FTIR spectrum of CNNTs, corresponding to the typical stretching vibration signals of the g-C 3 N 4 molecular structure [35]. The broad peaks at about 3051 and 1244 cm -1 were assigned to the stretching vibration of the C-NH-C linking groups due to partial polycondensation during the g-C 3 N 4 formation [36]. In addition, the strong peaks at about 812 cm -1 are assigned to the breathing mode of the tri-s-triazine ring of g-C 3 N 4 [35]. After the PDA and . This result may be mainly due to the partial crosslinking process caused by a Schiff base/Michael addition reaction, that is, the interaction between the catechol (-OH) group of DA/PDA and the primary amine (-NH 2 ) group of CNNTs [37]. The X-ray diffraction (XRD) measurement was conducted to investigate the crystal structures of pure CNNTs, MoS 2, and composites, and the results were presented in the Figure 3. The XRD patterns of pure CNNTs show the presence of two mainly diffraction peaks at 13.35°and 27.69°, which were corresponded to the (100) and (002) planes, respectively. The strong diffraction peak at 27.69°of CNNTs was ascribed to the interlayer stacking reflection of conjugated aromatic systems, while the weak diffraction peaks at 13.35°represent the inplane trigonal nitrogen linkage of tri-s-triazine units. The calculated interlayer distance of CNNTs is 0.322 nm according to the XRD patterns [25]. According to PDF card no. 01-075-1539, the main diffraction peaks which were presented at 2theta 14.10°, 32.89°, 33.68°, 39.50°, and 58.73°matched well with the (002), (100), (101), (103), and (110) planes, respectively, attributable to MoS 2 [38], which were also showed in the spectra of MoS 2 -PDA-CNNT composites. Figure 4 shows XPS spectra of CNNTs and nanocomposites, which are used to investigate the surface chemical properties of the as-prepared samples. Compared to original CNNTs, we have observed new peaks belong to the S and Mo element in the MoS 2 -PDA-CNNT composites, implying that the MoS 2 nanoparticles have been successfully decorated on the CNNTs surface. Figures 4(b)-4(f) show the highresolution XPS spectra of Mo 3d, S 2p, C 1 s, N 1 s, and O 1 s. The C 1 s XPS spectra was deconvoluted into the three distinct peaks, C=C peak (at about 284.8 eV), C-N/C-OH peak (at about 286.3 eV), and C=N peak (at about 288.3 eV) (Figure 4(b)). Figure 4(c) shows that the N 1 s XPS spectra were deconvoluted into three peaks, which are C=N-C (at about 399.   Figure 5(b) shows fluorescence spectra of CNNTs and nanocomposites with an excitation wavelength of 330 nm under room temperature. All the as-prepared samples show a wide range luminescence from 410 nm to 550 nm, and the peaks positions are centered at about 450 nm. This fluorescence phenomenon indicates that the photogenerated electrons and holes inevitably recombine in the semiconductors. It is worth noting that, compared with CNNTs and PDA-CNNTs, the PL intensity of MoS 2 -PDA-CNNTs is significantly reduced, which suggested that the recombination degree of photogenerated carriers in the MoS 2 -PDA-CNNTs sample is weakened [18].
The photocatalytic activity of the sample for degradation of MB is shown in Figure 6(a). When only light is provided without adding photocatalyst, almost no MB is degraded. The photodegradation activity of pure CNNTs is low, and only less than 60% of MB is photodegraded after 180 minutes. The sample MoS 2 nanoparticles showed the lowest photodegradation activity under 180 minutes of visible light irradiation. In previous research, PDA was proved to generate hydroxyl radicals under UV light irradiation [41]. To investigate its photo stability under visible light irradiation, the PDA particles as ref [41] were prepared and also catalyzed MB degradation. As shown in Figure S2, PDA presented high photo stability and negligible photocatalytic function. Measurements on MoS 2 /CNNT samples show that the simply hydrothermal treating of the CNNTs with MoS 2 NPs had a tent of effect on the photocatalysis activity. The degradation rate of MB increased slightly for MoS 2 /CNNTs. The PDA-modified CNNTs exhibited more obviously increase photocatalytic efficiency. Previous researches revealed the fact that PDA played a crucial role in the increasing photocatalytic activity and separating the electron-hole pairs [25]. Due to the synergic effect of the    Journal of Nanomaterials slight MB photodegradation activity under visible light is due to the photosensitization of MB dye itself [42]. Besides, a colorless pollutant RhB was utilized in degradation to demonstrate the photocatalytic activity of MoS 2 -PDA-CNNT composite ( Figure S3). The results were similar with MB degradation, which also shown that the MoS 2 -PDA-CNNTs have a well photocatalytic activity comparing with either pure g-C 3 N 4 or MoS 2 .
When the content of PDA was increased ( Figure 6(b)), the photocatalytic activity of MoS 2 -PDA-CNNT composites was enhanced that the degradation efficiency is gradually from 62% to 89%. The MoS 2 -PDA-(30%-) CNNT sample presented the highest photocatalytic efficiency. This phenomenon is attributed to the fact that PDA can promote the light harvesting of the system through the electron and proton redox coupling method, thereby improving the photocatalytic activity. However, excessive PDA on the surface will produce a photogenerated electron shielding effect, which will hinder g-C 3 N 4 from absorbing and utilizing the light source, reducing the photocatalytic activity of the system [25]. The reusing stability of MoS 2 -PDA-CNNT composites was investigated via the cycle photodegradation of the MB under visible light irradiation (λ > 400 nm). After 5 cycles, only a little loss of degradation efficiency is observed, while the MoS 2 -PDA-CNNTs still remain as photocatalytic activity as high as 89%, as shown in Figure 7. We also examined the stability of the used samples by the XRD patterns ( Figure S4). It was found that the diffraction peak of the MoS 2 -PDA-CNNT composites had no obvious discrepancy before and after five times of photocatalytic reaction. The results demonstrate that the photocatalyst owns good photocatalytic stability under visible light. Figure 8 shows the schematic band structure diagram and the possible photocatalytic mechanism of MoS 2 -PDA-CNNTs. Upon the irradiation with visible light, the electrons in the valence band edge (VB) of g-C 3 N 4 and MoS 2 are excited and transferred to the conduction band edge (CB) and generating holes in the VB of g-C 3 N 4 and MoS 2 . The VB position values for g-C 3 N 4 and MoS 2 are about -1.12 eV and -0.1 eV (vs. NHE), respectively [43]. The VB potential difference between g-C 3 N 4 and MoS 2 facilitates the transfer of electrons from g-C 3 N 4 to MoS 2 . Besides, due to the interface modification of PDA with high electron transfer and separation ability, the electrons in the CB of g-C 3 N 4 can quickly transfer to the CB of MoS 2 . As a result, the recombination of photogenerated electrons and holes is hindered, and effective charge separation is achieved. The electrons in the CB of MoS 2 could react with O2 to form radicals (·OH or ·O2-). The radicals generated in the CB of MoS 2 and holes photogenerated in the VB of g-C 3 N 4 are able to oxidize the MB to CO 2 and H 2 O. The enhanced efficiency of photogenerated charge separation promotes the photocatalytic MB degradation activity of MoS 2 -PDA-CNNTs.

Conclusions
In conclusion, a hierarchical NP-NT nanocomposite with enhanced photocatalytic properties was successfully constructed through a facile and versatile approach by mussel chemistry. Polydopamine (PDA) serves as the biomimetic adhesive layer which connects MoS 2 nanoparticles with g-C 3 N 4 nanotubes (CNNTs). The MoS 2 -PDA-CNNT composites presented improved visible light photocatalytic properties toward degradation of methylene blue (MB) due to the synergic effect of their unique tubular nanostructure, PDA coating layer, and MoS 2 nanoparticles. With the PDA ratio increasing, the photocatalyst light-harvesting performance was gradually enhanced. This research would provide a green, versatile strategy to fabricate hierarchical-structured nanocomposites with high catalytic activity.

Data Availability
All data used to support the findings of this study are available from the corresponding author upon request.

Conflicts of Interest
The authors declare no competing financial interests.