Lihua Li, Lili Yu, Zhaoyong Lin, and Guo Wei Yang

ACS Appl. Mater. Interfaces, Just Accepted Manuscript • DOI: 10.1021/acsami.6b00966 • Publication Date (Web): 17 Mar 2016 Downloaded from http://pubs.acs.org on March 21, 2016

“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Lihua Li, Lili Yu, Zhaoyong Lin, Guowei Yang *  State Key Laboratory of Optoelectronic Materials and Technologies, Nanotechnology  Research Center, School of Materials Science & Engineering, Sun Yat-sen University,  Guangzhou 510275, Guangdong, P. R. China.  *Corresponding author: stsygw@mail.sysu.edu.cn.  

Abstract  

The reduced TiO2-graphene oxide heterostructure as an alternative broad  spectrum-driven efficient water splitting photocatalyst has become a really interesting  topic, however, its syntheses has many flaws, e.g. tedious experimental steps,  time-consuming, small scale production and requirement of various additives, for  example, hydrazine hydrate is widely used as reductant to the reduction of graphene  oxide, which is high toxicity and easy to cause the second pollution. For these issues,  herein, we reported the synthesis of the reduced TiO2-graphene oxide heterostructure  by a facile chemical reduction agent-free one-step laser ablation in liquid (LAL)  method, which achieves extended optical response range from ultraviolet to visible  and composites TiO2-x (reduced TiO2) nanoparticle and graphene oxide for promoting  charge conducting. 30.64% Ti3+ content in the reduced TiO2 nanoparticles induces the  electronic reconstruction of TiO2, which results in 0.87 eV decrease of the band gap for the visible light absorption. TiO2-x-graphene oxide heterostructure achieved  drastically increased photocatalytic H2 production rate, up to 23 times with respect to  the blank experiment. Furthermore, a maximum H2 production rate was measured to  be 16 mmol/h/g using Pt as a co-catalyst under the simulated sunlight irradiation (AM  1.5G, 135 mW/cm2 ), the quantum efficiencies were measured to be 5.15% for  wavelength λ=365±10 nm and 1.84% for λ=405±10 nm, and overall solar energy  conversion efficiency was measured to be 14.3%. These findings provided new  insights into the broad applicability of this methodology for accessing fascinate  photocatalysts.

Introduction

Among numerous photocatalysts, titanium dioxide (TiO2) has been extensively  studied due to its stable chemical and physical properties, abundance, non-toxity  nature. Nevertheless, the practical applications of pure TiO2 are hampered by the  broad band-gap, fast recombination of photo-generated electron–hole pairs and slow  transfer rate of electrons, which then result in low hydrogen conversion efficiency  (STH) and quantum efficiency (QE). Enhancing light absorption and efficient charge  separation are crucial for solar-driven hydrogen evolution reaction. Therefore, band  gap engineering and conductivity improvement are required.

For these issues, two approaches have been developed to overcome these  disadvantages of TiO2 on the basis of energy band engineering. One approach is  doping and self-doping for extending the optical response range from ultraviolet to  visible. For example, there has been the doping of TiO2 with N, Cr, Fe, C, V, Mo and  so on1-4. However, it has been done with limited success because of the increased  recombination centers, the thermal instability, the need for exorbitant ion-implantation  facility and the hydrogen production is relatively low. Reduced TiO2 (TiO2-x), i.e.  self-doping, which contains disorders (Ti3+ or oxygen vacancies) was reported to be  an effective way to induce visible light absorption5,6. A variety of self-doping routes  have been developed to synthesis dark blue Ti3+/TiO2, such as heating TiO2 under  vacuum or reducing conditions(e.g. H2), thermal treatment7 , Ar+ -ion bombardment8 ,  laser-irradiated at the sintered rutile target9 . Although these methods above have their  own virtues, they are limited by the harsh conditions, tedious experimental steps,  time-consuming and small scale production, for inner Ti3+ can only be generated  during prolonged and high temperature (973 K) heat treatment5 . Additionally, the  surface Ti3+ and oxygen defects on the TiO2 are usually not stable enough in air, for  the surface Ti3+ is easily oxidized6 .

Another approach is to develop heterostructure materials formed by jointing  multiple semiconductor components together. In this case, the foreign semiconductor  is requested to be compatible with TiO2 and possess high solar-light-harvesting  capability. Moreover, the heterostructure between TiO2 and semiconductor can  efficiently separate electron and hole to reduce the recombination of photo-generated  carries10-12. There seems to be a consensus that graphene-based hybrid structures are  ideal candidates for its exceptionally high surface area (greater than 2600m2 g-1),  superior electrical conductivity, high thermal conductivity (5000Wm-1K -1), good  superior mobility of charge carriers (200000cm2V -1 s -1) 13-15, and plenty of oxygen  containing functional groups (-COOH, epoxide and -OH), which could be employed  as an excellent backbone to support nanoparticles. The incorporation of TiO2 onto  graphene or graphene oxide (GO) sheets with homogenous dispersion can provide  more active sites. Electron transporter channel could be constructed between graphene  and TiO2 nanoparticle, which efficiently reduces the recombination of the  photoinduced charge carriers and leads to a high photo-conversion efficiency16-18. For  instance, Wang et al. proposed that photoinduced electrons in the space-charge  regions could be transferred into graphene, while the holes remain on TiO2, thereby  retarding the recombination of electrons and holes19. In additional, GO would be more  stable after reduced, and it is not easy for the less-functional group graphene combined with other substances. Furthermore, the oxygen functional group hampers  the carrier mobility of graphene and thereby reduces the photocatalytic activity. Hence,  there is commonly require special chemical reduction agents to reduce GO. Hydrazine  hydrate is widely used as the reductant to the reduction of GO, but it is high toxicity  and easy to cause the second pollution.19,20

Based on the above discussion, the reduced TiO2-graphene oxide heterostructure  represents a good choice for settling the problems. To our best knowledge, researchers  have reported the reduced TiO2-graphene oxide do a neoteric photocatalyst with  decent catalytic performance in organic decomposition or H2 production, nevertheless,  there are still some intractable problems have not been fully resolved, such as the  reductivity degree of graphene dependent the photocatalytic performance of the  reduced TiO2-graphene oxide, the fountainhead of its visible light response, the  searching for appropriate reductant to graphene reduction and developing rapid, clean  and effective synthesis routes.  

In this contribution, we report the preparation of reduced TiO2-graphene oxide  heterostructure by a facile chemical reduction agents free one-step laser ablation in  liquid (LAL) method, which achieves extended optical response range from  ultraviolet to visible and composites TiO2-x nanoparticle and graphene oxide for  promoting charge conducting. To our knowledge, it is for the first time that the  reduced TiO2-graphene oxide heterostructure synthesized by the one-step LAL  procedure. Importantly, a maximum H2 production rate is measured to be 16  mmol/h/g using Pt as a co-catalyst under the simulated sunlight irradiation (AM 1.5G,135 mW/cm2 ), the quantum efficiencies are measured to be 5.15% for wavelength  λ=365±10 nm and 1.84% for λ=405±10 nm, respectively, and the overall solar energy  conversion efficiency is measured to be 14.3%. Furthermore, we study the reductivity  degree of graphene dependent the photocatalytic performance of the reduced  TiO2-graphene oxide heterostructure, which has almost not been specified in other  reports. The photocatalytic properties are highly enhanced by the synergetic effect of  Ti3+ self doping and well dispersion of TiO2 on graphene for reinforcing light  absorption and charge-transfer kinetics. Therefore, these results suggest the broad  applicability of this methodology for accessing fascinate photocatalysts.Experimental

Catalyst preparation. Commercial titanium dioxide (TiO2) with a purity of  99.99% is from Sigma-Aldrich (China), graphene oxide (GO) is from JCNANO in  China. All chemicals were of analytical grade and used as received without any  further purification. All solutions are prepared with deionized water. 25 mg GO was  dispersed into 100ml deionized water and ultraphonic for an hour, forming 0.25  mg/ml GO solution. The laser irradiation in liquid technique has been reported in our  previous works24-28. In this case, firstly, a certain volume of GO solution (4 ml-16ml)  and 50 mg TiO2 nanoparticles are firstly placed in a 30ml glass bottle filled with a  certain of deionized water and ultraphonic for 5 minutes, making TiO2 and GO mixed  uniformly. 4mg GO was mixed with 100mg TiO2 by ultrasonic for an hour in ethanol,  which was denoted as Sample 1. According to the different dosage of GO in the  synthesis, the samples are denoted as Sample 2 (1 mg GO/50mg TiO2), Sample 3 (2 mg GO/50mg TiO2), Sample 4 (4 mg GO/50mg TiO2). Then, the aqueous solution is  irradiated by a Q-switched Nd:YAG laser device with a wavelength of 532 nm,  repeating frequency of 10 Hz, pulse width of 10 ns, energy density of 580 mJ /pulse  and spot size of 4 mm in diameter. During laser irradiating, the liquid environment is  maintained at ambient temperature and pressure. The laser irradiation process lasts for  15 min. Then, the colloidal solution were synthesized and then kept in a dry box at 60℃  to evaporate water for a series of latter measurements.

Working electrode preparation. In brief, 5 mg of catalyst powder was  dispersed in 2 mL ethanol solvent with 45 µL of Nafion solution (5 wt%,  Sigma-Aldrich), and then ultrasonic for about 30 minutes to generate homogeneous  ink. Next, 16 µL of the dispersion was transferred onto graphite sheet, leading to a  catalyst loading of ~0.2 mg cm-2. Finally, the as-prepared catalyst film was dried at  room temperature.

Materials characterization. The morphology of the as-synthesized products are  characterized by scanning electron microscopy (SEM) with Thermal Field Emission  Environmental SemEdsEBSD. Transmission electron microscopy (TEM) is carried  out with a JEOL JEM-2010HR instrument (Japan) at an accelerating voltage of 200kV.  Crystal structure was characterized by X-ray Powder diffraction (XRD) (Rigaku  D-max 2200 VPC (Japan) with Cu Kα radiation (λ=1.54056 Å, accelerating voltage is  40 kV, emission current is 26 mA)), and a scanning rate of 6° min-1 is employed. The  electron paramagnetic resonance spectra (EPR) are recorded on a Bruker A300 EPR  spectrameter at X-band frequency of 9.443 GHz, sweep width of 500 Guass, and center field of 3390.00 Guass at 100K. Put same amount (3mg) of samples in the  glass tube for test. High-resolution X-ray photoelectron spectra (XPS) and valence  band (VB) XPS data are collected on ESCALab250 using the reference of C 1s (284.6  eV) with the Al-Kα (20.0 eV) radiation. The optical absorption of the as-synthesized  powder is characterized by using the UV-Vis 3150 Spectrophotometer (Japan). Raman  measurements were recorded by an in Via Raman Microscope (Renishaw, England)  with a He-Ne laser with a wavelength of 514 nm and maximum output power of 6  mW. Fourier transform infrared (FTIR) spectroscopy was recorded by a FTIR  spectrometer (Nicolet6700 Thermo Scientific, USA). Inductively coupled  plasma-atomic emission spectrometry (ICP-AES) (Thermo Instrument System  Inc.USA) was used to characterize the concentration of Ti in Sample 3.  Electrochemical measurements are performed with a CH Instrument Workstation  760E potentiostat at room temperature. All measurements were carried out in 20%  methyl alcohol solution and conducted in a convertional three-electrode cell by using  a Ag/AgCl (sat.KCl) electrode as the reference electrode, a graphite sheet as the  counter electrode, and the sample on the graphite electrode as the working electrode.  Electrochemical impedence spectroscopy (EIS) measurements were performed in an  alternating current frequency range of 10 KHz to 10 mHz.

Photocatalytic characterization. The photocatalytic activity of the samples is  performed by measuring the hydrogen evolution experiments. The reactions were  carried out in a 250 mL Pyrex top-irradiation photoreactor connected to a closed  gas-circulation system. A 300W top-irradiated Xenon lamp with different filters was used as the light source (10 cm away from the photoreactor). The average intensity of  irradiation was determined by CEL-NP2000-10 spectroradiometer. Approximately  0.1g of the photocatalyst was dispersed by ultrasonic for 5 minutes and a magnetic  stirrer in 100ml methanol solution (20vol%), 1.25 mL of chloroplatinic acid  hexahydrate (H2PtCl6·6H2O, 40 wt% Pt) aqueous solution (corresponding to 1 wt% Pt)  was added into the system. The suspension was vacuum pumped for about 20 minutes  to remove air. Before the photocatalytic test, the solution was irradiated for about an  hour to photo deposit Pt on the surface of TiO2. Environmental temperature is 25℃ and external circulation cooling water was used to decreasing the reactor’s  temperature. The H2 evolution rate was analyzed by an online gas chromatograph (GC,  TCD detector, dye air driver and N2 carrier). Magnetic stirring (450rpm) was used  during the water splitting experiments to ensure homogeneity of the suspension.  Illumination processing for 5 to 8 hours. The catalysts were recycled by 70 ℃ drying  in dry box, weighted and reused. We optimize the catalyst composition by measuring  hydrogen production rate under AM 1.5G and visible light illumination (λ> 400 nm)  for different amount of GO.

Quantum efficiency (QE) and solar to hydrogen conversion (STH)  calculations. The quantum efficiency were measured by applying a Xe lamp (300W)  with λ± 10nm band-pass filters (365nm, 405nm, 580nm, 670nm) irradiated for 5  hours. The average intensity of irradiation was determined by CEL-NP2000-10  spectroradiometer. The number of incident photons was calculated by:  N = Eλ hc

The QE was calculated using the following equation: QE = 2 × the number of evolved H molecules the number of incident photons (N) × 100%

The STH were evaluated by using AM 1.5G (irradiation area is 10.18 cm2 and the  light power density is 100mW/cm2 ) solar simulator as the light source with 0.1 g  Sample 3. After 5h of illumination, 11000µmol H2 was detected by GC. The total  input solar energy was E !"#$ = 5 × 3600 × 10.18 × 100 × 10*+ = 18324 J. The  energy generated by water splitting is  E. = 11000 × 10*/ × 6.02 × 10+ × 2.46 × 1.609 × 10*12 = 2621.08, 2.46eV  is the free energy of water splitting.  STH = Energy of generation of hydrogen gy water splitting Solar energy irradiating the reaction cell × 100% = E.E !"#$ × 100% = 2621.08 18324 × 100% = 14.3%

Results and Discussion

Characterization of catalysts. Morphology and structure characterization  results of TiO2 raw nanoparticles and Sample 3 are shown in Figure 1. As displayed  by SEM images (Figure 1(a-b)), laser ablation does not induce any marked change in  the morphology of the samples apart from the formation of several nanospheres,  though Sample 3 become dark blue, as shown in the inset. We can see that both of  TiO2 raw nanoparticles and Sample 3 are nanoparticles with relative even size of  15-20 nm. TEM image (Figure 1(c)) reveals that TiO2 had favorable dispersibility in  the matrix of the sheet of GO, suggesting that the well contact between GO and TiO2.  To verify samples’ phase, we measured XRD pattern of them. Figure 1(d-e) verified that samples are a mixture of anatase TiO2 and rutile TiO2. Unexpectedly, we cannot  index obvious change among samples, which is consistent with the SEM images.  From the above discussion, we think that the laser ablation has not broken the  long-range order structure of TiO2, which is beneficial to the excellent hydrogen  generation stability.

The blue color of TiO2 is always readily discernible by the presence of Ti3+ cations21,22. To test for the presence of Ti3+, low temperature EPR spectra were  recorded. As found from Figure 2(a), Sample 1, Sample 2 and Sample 4, show EPR  signal at g=2.001, Sample 3 shows a signal at g=2.004 while pure TiO2 shows no  signal, and these signals would be ascribed to surface defects23, typically due to  trapped electrons on oxygen vacancy (Ov). From Figure 2(b), under the laser  irradiation and addition of GO, another broad absorption peak is observed at around  g= 1.949-1.952 owe to in situ Ti3+ . 24,25 The EPR spectra indicate that there is no Ti3+ present on the surface of samples, for the surface Ti3+ would tend to adsorb  atmospheric O2, which would be reduced to O2- and exhibits signal at g≈2.02.26

In addition, X-ray photoelectron spectroscopy (XPS) analysis didn’t detect any  Ti3+ peaks, which further confirms that Ti3+ was present in the bulk of the sample  rather than at its surface (Figure 3(a)), which should improve the stability of hydrogen  evolution performance. It has been reported that Ti3+-doping and the Ov in the TiO2 resulted in some new energy levels below the conduction band27-29, that is forming  new electron pathway, extending the optical response range from ultraviolet to visible  and trapping the charge carriers, facilitating the trapping of charge carriers, helping to increase the concentration of photoinduced electrons and suppress charge  recombination 30. As we know, XPS reveals the elements and associated chemical  bonds in the top few atomic layers of the material. Ti-2p–Sample 3 spectrum (Figure  3(a)) shows two main peaks centered at 458.84 eV and 464.64 eV, which are ascribed  to Ti2p3/2 and Ti2p1/2, respectively, shift to lower binding energy compared to TiO2,  owing to the increasement of the electron density of Ti atom in TiO2, indicating the  existence of Ti-O-C bonds on the surface of TiO2. Remarkably, the Ti 2p peaks  cannot well de-convoluted into Ti3+2p and Ti4+2p, which is in accordance with EPR  data. The maximum of the relative Ti3+ species content in the bulk of Sample 3 is  estimated up to 30.64% calculated according to ICP-AES, and the high concentration  of Ti3+ can induce the electronic reconstruction of TiO2.

The C1s spectra also indicate the existence of chemical bonding between TiO2 and  GO/RGO (Figure 3(b-d)). Notably, the C1s spectra of GO (Figure 3(b)) were  deconvoluted into four peaks: the non-oxygenated rings belong to C-C and C=C  bonds (284.73 eV); the C in C-OH bonds (286.46eV); the carbonyl C(C=O) (286.9  eV); and the carboxylate carbon (O=C-OH) (288.5 eV)31. Inset in Figure 3(b) shows  C1s spectra of GO and GO-LAL. Obviously, after the action of laser, each peak area  ratio has changed. The C1s spectra of Sample 2 can be deconvoluted into four peaks  283.6 eV, 284.7 eV, 286.3 eV and 288.7 eV (Figure 3(c)), which corresponding to  Ti-C bonds, C-C and C=C bonds, C-OH bonds and O=C-OH bonds respectively. It is  not easily to find an obvious Ti-C peak of Sample 1 in Figure 3(d), we believe that  due to LAL process, Ti-C bonds have been formed. After the laser irradiation, relative content of C=O bonds decreased and Ti-C bonds formed, which would create strong  adhesion of TiO2 over the GO surface. The weaker peaks due to oxygenate groups  indicate a considerable de-oxygenation and the reduction of GO to RGO. The  formation of Ti-C and Ti-O bonds would introduce shallow trap into the band of TiO2 to narrow the band gap of TiO2, extending the optical response range from ultraviolet  to visible, and can also supply a transmission electron bridge between TiO2 and  GO/RGO32. Furthermore, we estimated the relative content of carbon not bound to  oxygen and Ti-C bonds by comparing the XPS peak areas of carbon in bonds. As  shown in Table 1, Sample 3 has the highest relative content of carbon not bound to  oxygen of 61.47% and Ti-C bonds of 5.7%, suggesting that the reduction degree of  GO has peaked in Sample 3. The higher reduction degree of GO, the more conductive  is, for RGO is conductive-taking, promoting the transfer rate of electrons, and also  conduce to maintaining a good dispersion of TiO2 on GO/RGO nanosheets. Noticing  that Sample 3 shows the most optimistic photocatalytic performance.