Fabrication of a Contamination-Free Interface between Graphene and

This is an open access article published under an ACS AuthorChoice License, which permits copying and redistribution of the article or any adaptations for non-commercial purposes.

Article http://pubs.acs.org/journal/acsodf

Fabrication of a Contamination-Free Interface between Graphene and TiO2 Single Crystals Huihui Liu, Dongbo Zhu, Hong Shi, and Xiang Shao* Department of Chemical Physics, CAS Key Laboratory of Urban Pollutant Conversion and Synergetic Innovation Center of Quantum Information and Quantum Physics, University of Science and Technology of China, Hefei, Anhui 230026, China S Supporting Information *

ABSTRACT: The uniform and seamless interface between graphene and semiconductors, that is, without adsorbates or contamination in between, is of importance for optimizing the electronic and catalytic performances of the combined system. In this work, we try to synthesize graphene directly over atomically flat TiO2 single-crystal surfaces using chemical vapor deposition (CVD) with acetylene as the carbon source. The facile synthetic conditions facilitate the formation of ultrathin polycrystalline graphene films with nanosize domains, while reasonably maintaining the terrace-and-step morphologies of the TiO2 surfaces. The established recipe can thus lead to the construction of a contamination-free interface between graphene and reducible oxides and also provide a well-defined platform for further investigations into the physicochemical properties of the graphene−oxide complex system from an atomic/molecular level.

1. INTRODUCTION TiO2-based photocatalysis has attracted immense interest because it may serve as a feasible solution for global environmental and energy issues.1,2 It is popularly acknowledged that under light irradiation, semiconductors produce free electrons and holes, which then diffuse to the surface and stimulate redox reactions.3 In this regard, the preferential recombination of the electron−hole pairs is usually imputed as the main reason for the low efficiency of the whole photocatalytic process. To inhibit the recombination, a wellestablished way is to use additives or decorators as charge trappers, for instance, transition metals in most cases.4,5 Recently, new carbonaceous materials, from zero-dimensional (0D) to three-dimensional (3D), such as carbon cages, nanotubes, graphene, and carbon black, have come into the family and gained more and more attention because they present outstanding improvements in a number of aspects.6−9 Graphene (Gr) in particular, a brand-new two-dimensional material, has generated research interest owing to its superior properties such as high thermostability, high transparency, high electric conductivity, and easy manufacturability10−13 and has been found to be able to promote the photocatalytic activity of TiO2 materials significantly.14 The promoting effect is usually attributed to the electron reservoir role of graphene, which results in the separation of the yielded charge carriers in TiO2 via the charge transfer process across the interface between the two materials.15 There are two basic strategies among the various preparations of graphene−TiO2 composites reported so far. One is physical mixing, that is, physical dispersion of one material into the other.16 In this case, the interface between © 2016 American Chemical Society

graphene and TiO2 is poorly controlled. During the preparation processes, adsorbates or contaminations on the TiO2 surface cannot be fully avoided. They stay as interlayers and thus dramatically influence the desired electron transfer efficiency. The second method is to grow TiO2 nanoparticles directly on graphene or graphene oxide surfaces.17 This method may build direct contact between TiO2 and graphene if the growth conditions are finely controlled but would result in a rather small contact area. In other words, only a small fraction of the TiO2 surface can be covered by graphene, which naturally leads to a limited electron transfer effect. In this respect, directly synthesizing graphene on molecularly clean TiO2 surfaces should be a promising strategy to achieve a uniform and contamination-free graphene−TiO2 (Gr−TiO2) interface. The configuration is expected to optimize the promoting effect of graphene over the TiO2-based photocatalytic activities.18,19 Syntheses of various graphene structures, either free-standing or supported, involve a great deal of effort. The discovered recipes include mechanical cleavage, chemical intercalation, and epitaxial deposition.12,20−22 Particularly with the chemical vapor deposition (CVD) method, high-quality and large-sized graphene films have been routinely prepared on polycrystalline transition metal surfaces and can be transferred to the substrate of choice.23−25 However, the complexity and high cost of the transfer process rendered the method unsatisfactory; these and the above-mentioned contamination issues necessitated the appeals for attempts to fabricate graphene directly over various Received: June 11, 2016 Accepted: July 21, 2016 Published: August 2, 2016 168

DOI: 10.1021/acsomega.6b00074 ACS Omega 2016, 1, 168−176

ACS Omega

Article

for preparing an atomically clean and flat TiO2 surface described above. (2) Introduce Ar and keep a flow rate of 100 sccm after cooling the substrate to 100 °C and slowly ramp up to about 650 °C at a heating rate of 10 °C/min. (3) Introduce about 5 sccm of H2 and 0.5 sccm of C2H2 together as the reaction gas and maintian this for a certain length of time (see text below for details), allowing the growth of graphene overlayers with different thicknesses. (4) Stop the reaction gas and anneal the sample in an Ar atmosphere at the same temperature for a certain length of time. The thickness and uniformity of the graphene samples can be controlled by varying the concentration of gas mixtures, the temperature, and the growth time. 2.3. Characterizations. To verify that the CVD product is graphene, examination of a number of characterizations has been performed. AFM and STM images were recorded with a Veeco Multi-mode V (Bruker) microscope. UV−vis absorption spectra were collected on a SolidSpec-3700 spectrophotometer (Shimadzu). Raman spectra were recorded on a LabRAM HR (HORIBA) microscope with a 514.5 nm laser. Optical microscopy (OM) observation was performed on a Leica DM6000 microscope. Scanning electron microscopy (SEM) images were recorded on a SIRION200 (FEI) microscope. Transmission electron microscopy (TEM) images were obtained on a JEM-2100F (JEOL) microscope. XPS data were collected using a laboratory-based Al Kα source (hν = 1486.7 eV) mounted in an ultrahigh vacuum chamber with a base pressure of at least 1 × 10−7 Pa (AXIS Ultra, Kratos Analytical).

dielectric substrates. For this goal, using CVD, slightly lowerquality graphene films have been synthesized directly on several unreducible dielectric insulator surfaces, such as Al2O3, ZrO2, SiO2, and MgO.26−29 In these fabrications, methane and hydrogen are normally used and a high pyrolysis temperature is applied. Such recipes, in principle, can be transferred to reducible oxide surfaces, but the serious corrosion effect under a highly reductive atmosphere has to be considered and balanced with extreme care. An example was given by Sun et al., who have recently grown micrometer-sized single-layer graphene using methane and hydrogen on SrTiO3 (STO) single-crystal surfaces30 while keeping the STO surface fairly noncorroded. Another option to suppress substrate corrosion is to use more-reactive molecules to replace methane as the carbon source, such as glucose, benzene, cyclohexane, and acetylene, which, however, in turn reduces the quality of the yielded graphene.29,31,32 By using acetylene and hydrogen as the gas source, Rümmeli et al. synthesized nanographene directly over MgO nanocrystals.29 The facile synthetic conditions manifest the possibility of growing graphene on reducible oxide substrates such as TiO2. In this work, we have prepared atomically smooth rutile-TiO2 (r-TiO2) single crystals by chemical etching and subjected them to CVD growth of graphene with acetylene as the carbon source. The advantage of using single-crystalline substrates lies in their well-defined surface structures, which facilitate atomiclevel characterization of their surface processes. It also provides a model system for a deeper understanding of graphene growth as well as the catalytic reaction mechanisms over nanosized particles. As will be shown below, while maintaining the step− terrace surface structures of the TiO2 substrates, the ultrathin graphene films fabricated were easily characterized by various techniques including atomic force microscopy (AFM), scanning tunneling microscopy (STM), electron microscopy (EM), and Raman spectroscopy. The seamless interaction established between graphene and TiO2 was carefully investigated with Xray photoelectron spectroscopy (XPS) and ultraviolet−visible light absorption spectroscopy (UV−vis). Unambiguous charge transfer from graphene to TiO2 is demonstrated at the ground state.

3. RESULTS Preparation of a clean and atomically flat TiO2 substrate is the first step in graphene synthesis. The r-TiO2(110) crystals prepared by following the HF etching and air-annealing treatments reported by Yamamoto et al.33 display a clear step−terrace structure and are free of particles, as confirmed by the AFM measurements shown in Figure 1a. The step height and terrace size are 0.4 and 200 nm, respectively, as shown by the line profile in Figure 1a, indicating the atomic smoothness of the as-prepared TiO2 (hereafter indicated as Pr-TiO2) surface. These prepared crystals are then loaded into the CVD

2. EXPERIMENTAL METHODS 2.1. Preparation of Atomically Smooth TiO2 Surfaces. Rutile-TiO2 (99.99% in purity) single crystals (10 × 5 × 0.5 mm3) with (110), (100), and (001) faces were purchased from Hefei Kejing Materials Technology Co. Atomically smooth surfaces were routinely achieved by following a recipe reported by Yamamoto et al.,33 which goes sequentially through ultrasonic cleaning with acetone (Sigma-Aldrich, 99.9%) and ultrapure water (Millipore, 18.2 MΩ·cm), etching with 20% HF (Alfa Aesar, 99%) for 10 min, rinsing with water, drying in a nitrogen stream, and annealing at 800 °C for 3 h in air. The atomic flatness and cleanness of the prepared surfaces were confirmed by AFM and XPS measurements repeatedly. All chemical reagents were directly used without further purification. 2.2. Synthesis of Graphene. CVD growth of graphene films on the as-prepared TiO2 substrates was performed in a horizontal quartz tube mounted inside a furnace (Kejing Co.). The schematic illustration of the setup can be found in Scheme S1. Synthesis of the graphene film was conducted as follows: (1) load the HF-etched TiO2 substrates into the quartz tube and anneal at 800 °C for 3 h. This is the same as the last step

Figure 1. AFM morphology evolution of r-TiO2(110) single-crystal surfaces as a function of the CVD growth time. (a) Pristine, (b) 30 min, (c) 60 min, and (d) 90 min. Line profiles I and II show the roughness of the pristine and CVD-processed surfaces, respectively. 169


ACS Omega

Article

Figure 2. XPS characterization of the CVD-processed r-TiO2(110) single crystals. (a) Full-range survey and (b) deconvoluted C1s core-level spectra of TiO2 after CVD growth for 60 min. (c) Comparison of C1s against the growth time.

is basically a graphene film but may contain a significant number of defects or domain boundaries. The on-surface characterizations already dictate that the CVD process can successfully prepare graphene directly over the TiO2 surface. However, further confirmation and disclosure of the graphene structure require more detailed characterizations, which necessitates the separation of the film from the as-grown substrate. This can be done by spin-coating a layer of polymethyl methacrylate (PMMA) over the sample surface followed by HF etching (see Figure S3 for illustration of the transfer procedure). With this method, the carbon film can be completely stripped from the TiO2 substrate together with the PMMA film and transferred using a pincette onto the substrate of choice. After the transfer, the PMMA can be simply removed by ultrasonic rinsing in acetone, which is a well-established recipe. Figure 3a,b shows the OM and SEM images, respectively, of a 60 min film that has been transferred onto the SiO2/Si surface. The latter has been ultrasonically cleaned in acetone before use. The graphene film regions can be easily identified from their darker contrast in both the images.

furnace for graphene synthesis using C2H2/H2/Ar as the loading gas (see the Experimental Methods section for detail). Figure 1b−d shows the AFM topographies of the r-TiO2(110) surfaces after the CVD process for different growth times. It is clearly observed that small added patches gradually develop over the surface (Figure 1b) and emerge as a continuous film over an extended growth time (Figure 1d); in the meantime, the step features are still maintained. Besides the AFM measurements, the patch-like films that formed at the early stage were confirmed also by STM and SEM measurements, as shown in Figure S1. The profile in Figure 1b, corresponding to the sample undergoing 30 min CVD growth, displays an increased surface roughness of up to around 0.7 nm (peak-topeak), much higher than that of the pristine TiO2(110) surface (peak-to-peak roughness of around 0.1 nm). This surface roughening may be attributed to the corrugation of the islands of deposits as well as the unavoidable slight etching of the TiO2 surface. When growing over a longer time or with higher C2H2 loadings, much thicker films can be fabricated. It took 90 min to synthesize the film shown in Figure S2, but the gas loading was about 10 times higher than that used for the film shown in Figure 1. It can be seen that the film has completely concealed the underneath step structures of the TiO2 substrate and developed into wrinkle-like features because of the large tension at the interface. To determine the chemical nature of the CVD-synthesized film on the r-TiO2(110) substrates, we have conducted XPS measurements over the as-prepared samples. Figure 2a shows the survey spectrum of a typical sample prepared using CVD, over 60 min. Only Ti, O, and C signals can be found on the sample surface, demonstrating the metal-free nature of the CVD process, as well as the cleanness of the etching process while preparing the substrate. Figure 2c shows the comparison of the C1s peak obtained for three different samples, using airintroduced carbon contaminations as the reference. Clearly, the carbon signals increase gradually with the extended CVD growth time, accompanied by decreasing signals of both Ti and O (not shown here), thereby dictating carbon as the main component of the produced film. Figure 2b shows the highresolution scan of the C1s XPS transition obtained over the sample examined in Figure 2a. From the deconvolution, one can immediately identify the strongest peak at 284.75 eV, corresponding to CC double bonding of carbon atoms within a conjugated honeycomb lattice, which coincides with the synthesized graphene over various other substrates.34 Other peaks at higher binding energies such as 286.4, 287.3, and 289.2 eV can be sequentially assigned as C−O, CO, and OC−O species, respectively. In fact, all these species and their distributions are perfectly in line with those observed on the single-layer graphene prepared on the STO(001) substrate.30 Therefore, we may safely propose that the CVD-fabricated film

Figure 3. Characterizations of the CVD-synthesized graphene films transferred onto the SiO2/Si surface. (a) OM and (b) SEM images of the 60 min film transferred onto the SiO2/Si substrate, where the dark contrast areas represent the graphene film. (c) AFM image obtained at the boundary of the transferred graphene film. The inserted height profile is taken along the green line, showing the thickness of the graphene film around 1.5 nm. (d) Diagram showing the linear correlation of graphene thickness with growth time. 170


ACS Omega

Article

Figure 4. Raman spectra collected for (a) the as-grown graphene films on r-TiO2(110) substrates and (b) the graphene films after being transferred to a SiO2/Si substrate.

position of G-band is slightly blue-shifted relative to that of the perfect single-crystalline graphene or graphite (1580 cm−1). This is another fingerprint character of the formation of nanosized graphene.36 Actually, according to Rümmeli et al., the peak ratio of G-bands over D-bands can provide an estimation of the grain size of the nanographene.29 By following their method, we can estimate that the average grain size of our polycrystalline graphene is about 7 nm (see Figure S5 for the deconvolution of the Raman spectrum as well as the calculation of grain size of the nanographene). Apart from the most intense D- and G-bands, the relatively weak and severely entangled features at 2680 (2D) and 2950 cm−1 (D + G) also depict an imperfect graphene film. More direct and detailed evidence of the nanocrystalline graphene structure can be obtained through high-resolution TEM investigations. To do this, the stripped graphene film has to be deposited onto special sample holders for TEM measurement. A detailed description of how to transfer the graphene onto the copper grids can be found in Figure S3. Figure 5a displays a low-magnification TEM view of the rather clean and large-area graphene film that was synthesized using CVD for 30 min. Focusing the electron beam over the edges of the graphene sheet, one can directly determine the number of graphene layers. The high-resolution TEM image shown in Figure 5b clearly reveals that the film is single-layer graphene.

Moreover, under the OM observation, transfer-induced wrinkles can be clearly recognized, manifesting that the asgrown graphene film is continuous and flexible. This is further confirmed by the AFM measurements conducted on the same sample. As shown in Figure 3c, the transferred graphene film shows no obvious morphology defects (such as holes and bumps). Furthermore, one can directly image the graphene− SiO2 boundary and measure the film thickness, as shown by the inserted height profile. In this way, the thickness of the graphene film and the growth time can be correlated. For instance, the apparent height of the graphene film grown over 30, 60, and 90 min was around 1.4, 2.0, and 2.5 nm, respectively, depicting a roughly linear correlation as shown in Figure 3d. We notice fluctuation in the film height even within the same carbon film. This may be caused by the inhomogeneous thickness of the graphene film itself but may also be induced by the trapped adsorbates that are dispersed unevenly underneath the graphene overlayer. Raman spectroscopy is usually regarded as the most reliable and nondestructive technique to characterize the structure and quality of carbon materials, in particular to investigate the ordering as well as defects in graphene. Therefore, we subjected both the as-prepared graphene−TiO2 samples and the transferred graphene on SiO2/Si substrates to Raman spectra characterization. As shown in Figure 4a, apparently, new features apart from those of the bare TiO2 can be recognized on the CVD-processed samples. Unfortunately, the strong background of the r-TiO2(110) substrate (1000−1800 cm−1) overlaps with the characteristic features of carbon materials, thus severely interfering with a detailed analysis of the graphene structure (see full-range Raman spectrum of the bare rTiO2(110) single crystal in Figure S4). By contrast, on the SiO2 substrate, the carbon-related Raman transitions are away from those for Si−Si and Si−O bonds and therefore are very suitable for analyzing graphene-related signals. As shown in Figure 4b, the graphene transferred onto the SiO2/Si substrates exhibit four recognizable Raman peaks, that is, at about 1335, 1600, 2680, and 2950 cm−1. The two most remarkable peaks at about 1335 and 1600 cm −1 can be safely assigned to the corresponding D- and G-bands of graphene, respectively. It has been extensively discussed in the literature that the G-band is related to the E2g vibration mode of sp2 carbon and its domain size and can be used to explain the degree of graphitization, whereas the D-band is associated with structural defects and, in particular, the disordering of the sp2 domains.35 Here in our samples, the G-bands display as sharp peaks whereas the D-bands are broad but also fairly strong, dictating that the film is dominated by graphitic structure yet with considerable amorphous carbon coexistence. In addition, the

Figure 5. TEM investigations of the graphene films stripped from the TiO2(110) substrates. (a) Low-magnification image showing a clean and uniform graphene film grown for 30 min. (b) High-resolution image of the film in (a) but collected at the edge position, revealing the monolayer thickness. (c) High-resolution image collected on a 90 min film, showing the lattice fringes and the randomly orientated domains. (d) Typical SAED pattern obtained on the 90 min graphene film in (c). 171


ACS Omega

Article

Figure 6. CVD synthesis of graphene on the (100) and (001) orientated r-TiO2 substrates. (a,e) AFM images of the pristine TiO2 surfaces after cleaning. (b,f) AFM images of the graphene−TiO2 film prepared using CVD synthesis for 30 min. (g,h) TEM images of the graphene films stripped from TiO2 (Gr−TiO2). (c) Raman spectra of the graphene films measured on different TiO2 substrates. (d) Raman spectra of the graphene films transferred onto the SiO2/Si substrates.

carbonaceous species. After graphene was transferred onto the SiO2/Si substrate, the prominent G- and D-bands of the graphene structures can be unambiguously identified, as shown in Figure 6d. In comparison with that on the (110) substrate, the graphene films prepared on (100) and (001) surfaces principally display the same Raman features but with slightly weaker signals. Figure 6g,h correspondingly shows the TEM images of the films synthesized over (100) and (001) substrates and transferred to Cu grids, where the graphitic layers can be clearly identified at the edges of the films. The TEM images also reveal that the quality of the graphene structures formed on the (100) and (001) r-TiO2 surfaces of is slightly lower than that on the (110) surface by using the same growth parameter. This difference, together with the reduced growth speed, shows the dependence of the terminated atomic structures of distinct surfaces, which may be attributed to the different reactivities against the acetylene molecules.

In Figure 5c, a more careful imaging was taken to see the lattice fringes. The interdistances were measured to be 0.25 nm, corresponding to the separations between (112̅0) planes of the graphite crystal.37 In the meantime, the polycrystalline nature of the film is immediately ascertained, with an average domain size of around 6−7 nm, a value consistent with the Raman estimations. The typical selected-area electron diffraction (SAED) pattern is shown in Figure 5d, which clearly displays a diffraction ring decorated with isolated spots, ascertaining the polycrystalline nature of the synthesized graphene film. The direct fabrication of graphene on the r-TiO2(110) surface manifests the feasibility of constructing a contamination-free graphene−TiO2 interface. It also implies that there may be little restriction between the substrate lattice and synthetic graphene adlayer. In fact, various substrates from metals to insulators have been successfully used as substrates for the growth of graphene, and the achievable crystallinity of graphene turns out to have little correlations with the graphene−substrate lattice match. We therefore continue to testify the growing of graphene over other r-TiO2 surfaces, that is, (001) and (100) faces using the same CVD method. Both of these latter surfaces are less stable than the (110) surface and terminated with very different atomic arrangements. By following a wet-etching process similar to that for the (110) surface, these two substrates can also be cleaned and flattened to atomic smoothness.33 The as-prepared pristine surfaces are characterized by AFM measurements, as shown in Figure 6a,e, respectively. Both surfaces are dominated by multiple-layer steps as high as 1 nm, whereas the enclosed terraces are nearly atomically flat. After the CVD growth, both surfaces show changes similar to that of the (110) substrate, that is, small islands developed accompanied by slight surface roughening, as shown by the AFM images in Figure 6b,f, respectively. Similar to the case of (110) substrate, these changes already indicate that the graphene film has probably been formed. In Figure 6c, the Raman spectra collected directly on the as-grown samples also show recognizable features that are related with the

4. DISCUSSIONS 4.1. Optimization of the Growth Conditions. The above evidence demonstrates that polycrystalline graphene with an average domain size of around 7 nm can be directly synthesized via cracking of acetylene over the three TiO2 single-crystal surfaces, disregarding the mismatch of the lattice symmetry and atomic arrangements. Researchers have recently discovered a variety of carbon sources for graphene synthesis, ranging from pure inorganic molecules to complex biomolecules.23,38 Most often, methane is used mixed with hydrogen as the feed for CVD growth of graphene on various transition metal surfaces. Frequently, high-quality graphene can be fabricated under relatively high operating temperatures, particularly taking special care over the control of hydrogen loading. The same recipe can be applied to synthesize high-quality graphene on refractory or unreducible oxides such as Al2O3,26 SiO2,28 and STO substrates30 as well. However, when a reducible substrate, such as the TiO2 single crystal studied here, was used, the corrosion of the substrate surface has to be carefully considered. 172


ACS Omega

Article

Figure 7. XPS spectra of the CVD-synthesized graphene−TiO2(110) samples. (a) Ti2p and (b) O1s evolutions against the CVD growth time. (c) Deconvoluted high-resolution Ti2p spectrum obtained on a sample grown for 90 min (Gt = 90 min). (d) Deconvoluted high-resolution O1s spectrum of the same sample in (c).

the loading gas, and a growth temperature of around 650 °C. Hydrogen is added to balance the decomposition of C2H2 to ensure better graphene quality; yet its partition is seriously controlled in the range that will not induce significant surface reconstruction of the TiO2 substrate. Finally, with these optimized parameters, we can achieve a growth speed of roughly 0.03 ML/min with single-crystalline domains of around 7 nm, which means that a nearly complete graphene film can be normally achieved after 30 min of synthesis. 4.2. Growth Mechanism and Role of the Substrate. The growth mechanism of graphene over semiconductor substrates can be drastically distinct from that on metal surfaces. In the latter case, there is a relatively strong interaction between the graphene lattice and the metal surface and hence exerts a self-catalyzing and epitaxial effect on the growth process. The formation of graphitic carbon is usually based on the dynamic equilibrium between cracking of the carbonaceous molecules, dissolving of carbon into the metal, and its segregation on the surface.41 However, on the semiconductors, particularly the oxides, much weaker interactions are expected between graphene and the substrate, and the dissolving− segregation mechanism would turn invalid. In this case, the oxide substrate itself is usually considered to play a catalyzing role. On either SiO2 or Al2O3 surface, the surface oxygen atoms are considered to be active sites, which may capture the cracked molecular fragments and facilitate the growth of polycyclic aromatic flakes.26 Liu et al. proposed that on SrTiO3 surface, both oxygen and metal ions are involved in anchoring the graphene fragments.30 Nevertheless, difficulties in characterization under the operational conditions prevent a detailed understanding of how the cracked fragments of methane bind to the oxide surface and develop into a two-dimensional graphene structure. Here in our case, C2H2 as the more active precursor in combination with more facile growth conditions facilitate a much better understanding of graphene formation. Biedrzycki et al. have recently conducted Fourier transform infrared spectroscopy (FT-IR) and electron paramagnetic resonance (EPR) studies to investigate the interaction of acetylene with TiO2(110) surface at room temperature.42 They evidenced a process that includes heterolytic dissociative chemisorption of acetylene, followed by a redox and cyclization reaction between the fragmented acetylene on the TiO2 surface. Particularly, the unsaturated Ti4+ (bridge O2−) ions on the rutile(110) surface are proposed to be the suitable acidic (basic) sites for binding the fragmental aromatic (hydrogen)

In our preliminary tests, the methane−H2 mixture with various compositions and flow rates were tried. However, when the temperature was raised above 900 °C, which is necessary for cracking of methane, the TiO2 surfaces were always severely affected and step−terrace structures could not be retained any more (see typical results as shown in Figure S6). This is not beneficial for the construction of graphene−TiO2 interface and also contradicts our purpose of maintaining well-defined TiO2 surfaces for subsequent model studies. Therefore, we resorted to C2H2 with the consideration that it is frequently adopted for the synthesis of carbon nanotubes and acetylene black films.39,40 The advantage of C2H2 is its higher reactivity and relatively low cracking temperature when compared with those of methane. In a recent study, Jain et al. reported that at room temperature and near-ambient pressure, C2H2 can directly form graphene fragments upon adsorption on TiO2 P25 (a mixture of rutile and anatase).18 Although the authors did not provide further evidence for real graphene evolution, the study offered a good starting point for synthesizing graphene over TiO2 surface using C2H2 as the precursor. As found in our experiments, by using C2H2 as the carbon source, the synthesizing temperature can be significantly reduced from 1000 °C (typical for methane) to around 650 °C. This brings unpreced- ented advantages that the TiO2 surface can be protected very well from corrosion. As shown by the AFM images in Figures 1a−d and 6b,f, the step−terrace structures are all very well maintained on the three r-TiO2 single-crystal surfaces, when the graphene films are sufficiently thin. On the other hand, however, the higher reactivity of C2H2 may also induce some disadvantages because it may cause a fast growth of various types of carbon materials with relatively poor control. It is well known that carbon nanotubes have been frequently synthesized with C2H2 using a similar CVD process. Although it did not show out in our preparations, this growth diversity cannot be fully ignored. In this regard, the relatively low quality of our asprepared graphene may be understood because its growth can be easily interfered by the various by-products under the reaction conditions without delicate controls. The controllable parameters include the loading component and loading speed as well as the growth temperatures. They may have significant influence on the growth speed and quality of the graphene layers and on the TiO2 surface structures. After a number of tests, we finally fixed the best parameters as 0.5 sccm of C2H2 mixed with 5 sccm of H2, using 100 sccm of Ar as 173


ACS Omega

Article

species, respectively.43 The proposals fit well with our STM observations conducted under ultrahigh vacuum conditions, where C2H2 molecules adsorb readily onto the (1 × 1) terminated rutile-TiO2(110) surface and react to form larger clusters at increased exposures (not shown here). Further illustrations of the whole mechanism would await more systematic STM works in combination with other spectroscopic characterizations, which will be carried out in the near future. 4.3. Interaction between Graphene and TiO2 Substrate. Finally let us look at the interaction between the fabricated graphene film and TiO2 substrate. The aim of this work was to construct a clean graphene−TiO2 interface and provide a model system for further investigations of the photocatalytic mechanism of the complex system. Therefore, the TiO2 single crystals were adopted as the substrates. HFetching followed by annealing in an inert Ar atmosphere at high temperatures ensures surface flatness and cleanness, as verified by AFM and XPS characterizations. Also, during the graphene synthesis, the surface was protected from any contaminations other than the carbonaceous species yielded from C2H2. The growth temperature applied was mild but high enough to keep the surface clean from added water species. On the other hand, it should also be proper to activate the C2H2 molecules on the TiO2 surfaces while preventing them from thermally cracking in the gas phase. Moreover, the small hydrogen loading was adopted for improving the quality of the synthesized graphene, but carefully controlled to prevent severe etching of the substrate as well as formation of large amount of surface hydroxyls. With all these considerations, a clean graphene− TiO2 interface can be safely concluded. Once the seamless graphene−TiO2 connection is built up, careful XPS characterizations would help us to look into the interactions between the two components in the complex. Figure 7a,b shows the detailed XPS spectra analyses of the Ti2p and O1s core-level signals measured on different samples, respectively. It should be noted here that the carbon peak was fixed at 284.8 eV, serving as the reference for all XPS transitions (see the C1s XPS spectra in Figure 3c). One can immediately recognize a positive shift of the core-level binding energy of both Ti2p and O1s on the TiO2 samples after CVD growth of the graphene layer. For the growth time of 30 min, the chemical shift is around 0.6 eV, whereas it reaches a maximum of 0.8 eV for 60 min and is constant for thicker films. In a controlled experiment, the graphene film was stripped and the substrate (hence termed as St-TiO2) was subject to XPS measurement again. Interestingly, both Ti2p and O1s were found to restore to the original values corresponding to the pristine TiO2. It is worth pointing out that the transfer process has an ignorable effect on the TiO2 surface, as long as the operation is carefully controlled within a few minutes (see Figure S7). Therefore, the upward shifts of the core levels is a direct measure of the charge transfer from graphene to TiO2, which has originated from the work function difference between the two materials.44 The transferred charges build up surface dipoles and lead to a prominent downward band-bending and a sufficient increase in the binding energies of both Ti2p and O1s of the TiO2 substrate. More detailed analyses of the Ti2p and O1s spectra are shown in Figure 7c,d, respectively. As can be found in Figure 7c, there is a small fraction of the total signal peaked at around 457.8 eV, corresponding to the Ti3+ ions that have been produced upon the reduction of TiO2 substrate. However, a majority of the Ti ions are still in +4 state. Furthermore, if we

look back at the deconvoluted C1s spectrum shown in Figure 2b, we can notice that there is no indication for any C−Ti species, manifesting the lack of carbon doping as well as surface C−Ti bonds. For the O1s spectrum, additional features are observed besides the TiO2-related major peak at 530.6 eV. Two shouldering peaks at 531.2 and 532.6 eV can be assigned to the surface hydroxyls and carboxyls, respectively.45 These species are either intrinsic on the defects of the graphene layers or introduced during the exposure to air. Their existence is consistent with the polycrystalline phase of the graphene film synthesized. Many studies consider that these oxygen species improve the interaction between the graphene film and TiO2 surface, sometimes leading to a strengthened charge transfer effect.46 The interactions between graphene and TiO2 can also be characterized by the changes in the light-absorption properties. Band-gap narrowing and enhanced absorption of visible light have been frequently reported for the variously prepared graphene−TiO2 complex systems.47,48 Here, we also carried out UV−vis spectroscopy measurements on our contamination-free graphene−TiO2 combined samples. As shown in Figure 8a,

Figure 8. (a) UV−vis spectra obtained on the pristine TiO2(110) substrate and with CVD-synthesized graphene films for different growth times. (b) Kubelka−Munk estimation of the band gaps of the pristine and graphene-covered TiO2 samples.

along with increase in the growth time, that is, the thickness of the synthesized graphene layers, the TiO2 crystal has displayed an increasingly enhanced adsorption of the visible light ranging from 400 to 800 nm. This is simply embodied by the deepened color of the TiO2 samples along with the growth time. Based on the Kubelka−Munk method,49 optical band gaps can be derived from the measured absorption data as shown in Figure 8b. The pristine TiO2 single crystal displays a band gap around 2.91 eV, which roughly coincides with the reported value of 3.0 eV. After growing graphene for 60 min, the band gap is slightly reduced to 2.87 eV. This small reduction, although very close to the error limit of this estimation method, reproduced very well on different batches of samples and can be confirmed clearly from the raw UV−vis data as shown in Figure 8a. We notice that the band gap narrowing also contains the contribution of TiO2 reduction. Therefore, we tried to strip the graphene film away 174


ACS Omega

■

and measured the UV−vis adsorption again. As shown by the blue dashed line in Figure 8b, after removing the graphene layers, the band gap narrowing is slightly suppressed when compared to that with graphene overlaid. This result dictates that there is a rather weak effect of graphene in modifying the band gap of TiO2, which is in line with the weak interactions between graphene and TiO2. The observed weak effect is in drastic contrast to the results of previous studies of mixed graphene oxide and TiO2 particles, where prominent band gap reduction was observed.14,44,46 The possible reason may be the extremely low mass ratio of graphene over TiO2 in our case. It may also be due to the relatively low quality of our synthesized graphene layers.

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acsomega.6b00074. Schematic illustrations of the CVD setup and the graphene transfer process, additional AFM/STM results of the control experiments (PDF)

■

REFERENCES

(1) Grätzel, M. Photoelectrochemical cells. Nature 2001, 414, 338− 344. (2) Chen, X.; Mao, S. S. Titanium dioxide nanomaterials: synthesis, properties, modifications, and applications. Chem. Rev. 2007, 107, 2891−2959. (3) Roy, P.; Berger, S.; Schmuki, P. TiO2 nanotubes: synthesis and applications. Angew. Chem., Int. Ed. 2011, 50, 2904−2939. (4) Subramanian, V.; Wolf, E. E.; Kamat, P. V. Catalysis with TiO2/ gold nanocomposites. Effect of metal particle size on the Fermi level equilibration. J. Am. Chem. Soc. 2004, 126, 4943−4950. (5) Poroshkov, V. P.; Gurin, V. S. Surface structures on titanium dioxide electrodes after electrochemical silver deposition. Surf. Sci. 1995, 331, 1520−1525. (6) Ioffe, I. N.; Chen, C.; Yang, S.; Sidorov, L. N.; Kemnitz, E.; Troyanov, S. I. Chlorination of C86 to C84Cl32 with Nonclassical Heptagon-Containing Fullerene Cage Formed by Cage Shrinkage. Angew. Chem. Int. Ed. 2010, 122, 4894−4897. (7) Javey, A.; Kim, H.; Brink, M.; Wang, Q.; Ural, A.; Guo, J.; Mcintyre, P.; McEuen, P.; Lundstrom, M.; Dai, H. High-κ dielectrics for advanced carbon-nanotube transistors and logic gates. Nat. Mater. 2002, 1, 241−246. (8) Geim, A. K.; Novoselov, K. S. The rise of graphene. Nat. Mater. 2007, 6, 183−191. (9) Long, Y.-M.; Zhou, C.-H.; Zhang, Z.-L.; Tian, Z.-Q.; Bao, L.; Lin, Y.; Pang, D.-W. Shifting and non-shifting fluorescence emitted by carbon nanodots. J. Mater. Chem. 2012, 22, 5917−5920. (10) Novoselov, K. S.; Geim, A. K.; Morozov, S. V.; Jiang, D.; Katsnelson, M. I.; Grigorieva, I. V.; Dubonos, S. V.; Firsov, A. A. Twodimensional gas of massless Dirac fermions in graphene. Nature 2005, 438, 197−200. (11) Zhang, Y.; Tan, Y.-W.; Stormer, H. L.; Kim, P. Experimental observation of the quantum Hall effect and Berry’s phase in graphene. Nature 2005, 438, 201−204. (12) Geim, A. K. Graphene: status and prospects. Science 2009, 324, 1530−1534. (13) Ohta, T.; Bostwick, A.; Seyller, T.; Horn, K.; Rotenberg, E. Controlling the electronic structure of bilayer graphene. Science 2006, 313, 951−954. (14) Lee, J. S.; You, K. H.; Park, C. B. Highly photoactive, low bandgap TiO2 nanoparticles wrapped by graphene. Adv. Mater. 2012, 24, 1084−1088. (15) Du, J.; Lai, X.; Yang, N.; Zhai, J.; Kisailus, D.; Su, F.; Wang, D.; Jiang, L. Hierarchically ordered macro−mesoporous TiO2−graphene composite films: improved mass transfer, reduced charge recombination, and their enhanced photocatalytic activities. ACS Nano 2011, 5, 590−596. (16) Zhang, Y.; Pan, C. TiO2/graphene composite from thermal reaction of graphene oxide and its photocatalytic activity in visible light. J. Mater. Sci. 2011, 46, 2622−2626. (17) Sun, L.; Zhao, Z.; Zhou, Y.; Liu, L. Anatase TiO2 nanocrystals with exposed {001} facets on graphene sheets via molecular grafting for enhanced photocatalytic activity. Nanoscale 2012, 4, 613−620. (18) Jain, S. M.; Biedrzycki, J. J.; Maurino, V.; Zecchina, A.; Mino, L.; Spoto, G. Acetylene oligomerization on the surface of TiO2: a step forward in the in situ synthesis of nanostructured carbonaceous structures on the surface of photoactive oxides. J. Mater. Chem. A 2014, 2, 12247−12254. (19) Kamegawa, T.; Yamahana, D.; Yamashita, H. Graphene coating of TiO2 nanoparticles loaded on mesoporous silica for enhancement of photocatalytic activity. J. Phys. Chem. C 2010, 114, 15049−15053. (20) Berger, C.; Song, Z.; Li, X.; Wu, X.; Brown, N.; Naud, C.; Mayou, D.; Li, T.; Hass, J.; Marchenkov, A. N.; Conrad, E. H.; First, P. N.; de Heer, W. A. Electronic confinement and coherence in patterned epitaxial graphene. Science 2006, 312, 1191−1196. (21) Lin, Y.-M.; Dimitrakopoulos, C.; Jenkins, K. A.; Farmer, D. B.; Chiu, H.-Y.; Grill, A.; Avouris, P. 100-GHz transistors from wafer-scale epitaxial graphene. Science 2010, 327, 662.

5. CONCLUSIONS In conclusion, by using acetylene mixed with hydrogen as the feeding gas for a simple CVD process, we have successfully synthesized graphene films directly on atomically flat rutileTiO2 single crystals. With various characterizations including AFM, TEM, and Raman spectroscopy, the polycrystalline nature of the graphene films with average domain size of around 7 nm is clearly revealed. Most importantly, the mild growth conditions do not influence the step−terrace structure observed on the pristine TiO2 single crystals that have been successfully used in this work as CVD substrates. Even though the mechanism of graphene formation is yet unclear at the present stage, the reactive adsorption of acetylene on the TiO2 surface followed by intermolecular heterocyclic reactions has been proposed as the possible growth route of graphene. Detailed XPS and UV−vis measurements confirmed the seamless connection between the synthesized graphene layers with the TiO2 substrates, but also clearly manifest the charge transfer from graphene to TiO2 materials in the absence of interface adsorbates. Not only do the results of the study demonstrate that direct fabrication of the graphene−reducible oxide is feasible, but the obtained samples also provide a suitable model system for further investigations of photocatalytic chemistry mechanism of this complex system.

■

Article

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Phone: +86-0551-63600765. Fax: +86-0551-63600765. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS We are grateful for the financial support of NSFC (21333001, 91545128) and MOST (2014CB932700). X.S. thanks the financial support of the Fundamental Research Funds for the Central Universities and the Thousand Talent Program for Young Outstanding Scientists of the Chinese government. The “Strategic Priority Research Program” of the Chinese Academy of Sciences (grant XDB01020100) is also acknowledged. 175


ACS Omega

Article

(22) Hofrichter, J.; Szafranek, B. N.; Otto, M.; Echtermeyer, T. J.; Baus, M.; Majerus, A.; Geringer, V.; Ramsteiner, M.; Kurz, H. Synthesis of graphene on silicon dioxide by a solid carbon source. Nano Lett. 2010, 10, 36−42. (23) Li, X.; Cai, W.; An, J.; Kim, S.; Nah, J.; Yang, D.; Piner, R.; Velamakanni, A.; Jung, I.; Tutuc, E.; Banerjee, S. K.; Colombo, L.; Ruoff, R. S. Large-area synthesis of high-quality and uniform graphene films on copper foils. Science 2009, 324, 1312−1314. (24) Yu, Q.; Lian, J.; Siriponglert, S.; Li, H.; Chen, Y. P.; Pei, S.-S. Graphene segregated on Ni surfaces and transferred to insulators. Appl. Phys. Lett. 2008, 93, 113103. (25) Gao, L.; Ren, W.; Xu, H.; Jin, L.; Wang, Z.; Ma, T.; Ma, L.-P.; Zhang, Z.; Fu, Q.; Peng, L.-M.; Bao, X.; Cheng, H.-M. Repeated growth and bubbling transfer of graphene with millimetre-size singlecrystal grains using platinum. Nat. Commun. 2012, 3, 699. (26) Song, H. J.; Son, M.; Park, C.; Lim, H.; Levendorf, M. P.; Tsen, A. W.; Park, J.; Choi, H. C. Large scale metal-free synthesis of graphene on sapphire and transfer-free device fabrication. Nanoscale 2012, 4, 3050−3054. (27) Scott, A.; Dianat, A.; Börrnert, F.; Bachmatiuk, A.; Zhang, S.; Warner, J. H.; Borowiak-Palen, E.; Knupfer, M.; Büchner, B.; Cuniberti, G.; Rümmeli, M. H. The catalytic potential of high-κ dielectrics for graphene formation. Appl. Phys. Lett. 2011, 98, 073110. (28) Chen, J.; Wen, Y.; Guo, Y.; Wu, B.; Huang, L.; Xue, Y.; Geng, D.; Wang, D.; Yu, G.; Liu, Y. Oxygen-aided synthesis of polycrystalline graphene on silicon dioxide substrates. J. Am. Chem. Soc. 2011, 133, 17548−17551. (29) Rümmeli, M. H.; Bachmatiuk, A.; Scott, A.; Börrnert, F.; Warner, J. H.; Hoffman, V.; Lin, J.-H.; Cuniberti, G.; Büchner, B. Direct low-temperature nanographene CVD synthesis over a dielectric insulator. ACS Nano 2010, 4, 4206−4210. (30) Sun, J.; Gao, T.; Song, X.; Zhao, Y.; Lin, Y.; Wang, H.; Ma, D.; Chen, Y.; Xiang, W.; Wang, J.; Zhang, Y.; Liu, Z. Direct growth of high-quality graphene on high-κ dielectric SrTiO3 substrates. J. Am. Chem. Soc. 2014, 136, 6574−6577. (31) Li, Z.; Wu, P.; Wang, C.; Fan, X.; Zhang, W.; Zhai, X.; Zeng, C.; Li, Z.; Yang, J.; Hou, J. Low-temperature growth of graphene by chemical vapor deposition using solid and liquid carbon sources. ACS Nano 2011, 5, 3385−3390. (32) Li, X.-H.; Kurasch, S.; Kaiser, U.; Antonietti, M. Synthesis of Monolayer-Patched Graphene from Glucose. Angew. Chem., Int. Ed. 2012, 51, 9689−9692. (33) Yamamoto, Y.; Nakajima, K.; Ohsawa, T.; Matsumoto, Y.; Koinuma, H. Preparation of atomically smooth TiO2 single crystal surfaces and their photochemical property. Jpn. J. Appl. Phys. 2005, 44, L511−L514. (34) Xue, Y.; Wu, B.; Jiang, L.; Guo, Y.; Huang, L.; Chen, J.; Tan, J.; Geng, D.; Luo, B.; Hu, W.; Yu, G.; Liu, Y. Low temperature growth of highly nitrogen-doped single crystal graphene arrays by chemical vapor deposition. J. Am. Chem. Soc. 2012, 134, 11060−11063. (35) Ferrari, A. C. Raman spectroscopy of graphene and graphite: disorder, electron−phonon coupling, doping and nonadiabatic effects. Solid State Commun. 2007, 143, 47−57. (36) Jawhari, T.; Roid, A.; Casado, J. Raman spectroscopic characterization of some commercially available carbon black materials. Carbon 1995, 33, 1561−1565. (37) Qu, D.; Zheng, M.; Zhang, L.; Zhao, H.; Xie, Z.; Jing, X.; Haddad, R. E.; Fan, H.; Sun, Z. Formation mechanism and optimization of highly luminescent N-doped graphene quantum dots. Sci. Rep. 2014, 4, 5294. (38) Ruffieux, P.; Wang, S.; Yang, B.; Sánchez-Sánchez, C.; Liu, J.; Dienel, T.; Talirz, L.; Shinde, P.; Pignedoli, C. A.; Passerone, D.; Dumslaff, T.; Feng, X.; Müllen, K.; Fasel, R. On-surface synthesis of graphene nanoribbons with zigzag edge topology. Nature 2016, 531, 489−492. (39) Huang, Z. P.; Xu, J. W.; Ren, Z. F.; Wang, J. H.; Siegal, M. P.; Provencio, P. N. Growth of highly oriented carbon nanotubes by plasma-enhanced hot filament chemical vapor deposition. Appl. Phys. Lett. 1998, 73, 3845−3847.

(40) Cai, F.; Chen, J.; Xu, R. Porous acetylene-black spheres as the cathode materials of dye-sensitized solar cells. Chem. Lett. 2006, 35, 1266−1267. (41) Shu, H.; Tao, X.-M.; Ding, F. What are the active carbon species during graphene chemical vapor deposition growth? Nanoscale 2015, 7, 1627−1634. (42) Biedrzycki, J. J.; Livraghi, S.; Corazzari, I.; Mino, L.; Spoto, G.; Giamello, E. On the Redox Mechanism Operating along C2H2 SelfAssembly at the Surface of TiO2. Langmuir 2015, 31, 569−577. (43) Mino, L.; Spoto, G.; Bordiga, S.; Zecchina, A. Rutile Surface Properties Beyond the Single Crystal Approach: New Insights from the Experimental Investigation of Different Polycrystalline Samples and Periodic DFT Calculations. J. Phys. Chem. C 2013, 117, 11186− 11196. (44) Du, A.; Ng, Y. H.; Bell, N. J.; Zhu, Z.; Amal, R.; Smith, S. C. Hybrid graphene/titania nanocomposite: interface charge transfer, hole doping, and sensitization for visible light response. J. Phys. Chem. Lett. 2011, 2, 894−899. (45) Huang, B.-S.; Tseng, H.-H.; Wey, M.-Y. Comparison of visiblelight-driven routes of anion-doped TiO2 and composite photocatalyst. J. Ceram. Soc. Jpn. 2009, 117, 753−758. (46) Wang, W.-S.; Wang, D.-H.; Qu, W.-G.; Lu, L.-Q.; Xu, A.-W. Large ultrathin anatase TiO2 nanosheets with exposed {001} facets on graphene for enhanced visible light photocatalytic activity. J. Phys. Chem. C 2012, 116, 19893−19901. (47) Zhang, H.; Lv, X.; Li, Y.; Wang, Y.; Li, J. P25-graphene composite as a high performance photocatalyst. ACS Nano 2010, 4, 380−386. (48) Zhang, Y.; Tang, Z.-R.; Fu, X.; Xu, Y.-J. TiO2−graphene nanocomposites for gas-phase photocatalytic degradation of volatile aromatic pollutant: is TiO2−graphene truly different from other TiO2−carbon composite materials? ACS Nano 2010, 4, 7303−7314. (49) Kubelka, P.; Munk, F. An article on optics of paint layers. Z. Tech. Phys. 1931, 12, 593−601.

176


Fabrication of a Contamination-Free Interface between Graphene and

Recommend Documents