Metal-organic frameworks (MOFs) as precursors towards TiOx/C composites for photodegradation of organic dye

A series of TiOx/C composites were prepared by pyrolyzing a Ticontaining MOF MIL-125, which exhibited excellent catalytic activities towards the photodegradation of methylene blue (MB) in aqueous solution. The best photocatalytic performance was achieved in MIL125 pyrolyzed at 1000 C due to its high surface area, reduced Ti3O5 composition, and conductive carbon matrix. This work represents a novel approach towards TiOx/C composites as photocatalysts.

The harvest of solar energy has become an important procedure in the research of clean energy and environmental sustainability, such as water splitting, 1 degradation of organic pollutants, 2 and photocatalytic reduction of CO 2 . 3 One of the bottlenecks in this procedure is the development of photocatalysts that can make full use of the whole solar spectrum with high energy conversion efficiency. Metal oxides are an important family of photocatalysts, among which TiO 2 is the widest studied one due to its semiconductor property, high natural abundance, and low toxicity. 4 In the last decade, a lot of research has been done to improve the photocatalytic activity of TiO 2 , such as morphology control, 5 crystal facet engineering, 6 doping with other elements, 7,8 etc., with the aim for better absorption of solar energy and smaller chance of electron-hole recombination. Recently, TiO 2 /C composites have attracted lots of attention because of (1) higher surface area provided by the porous carbon matrix with better adsorption of reagents; (2) extended light absorption range and higher absorption intensity; and (3) longer electron-hole recombination time, all of which are benecial to the photocatalytic activities. 9,10 The current methods for the preparation of TiO 2 /C composites include simple mixing, thermal oxidation, sol-gel, hydrothermal, deposition, etc. 10 Novel approaches are needed to have better control and ne tuning of the composition, texture, and polymorphs of TiO 2 /C composites.
Ti-containing MOFs have been reported in the literature. [30][31][32][33] Their photocatalytic activities have been demonstrated through CO 2 reduction, 31 water splitting, 34 and degradation of organic dye. 33 However, because of the labile coordination bonds in MOFs, 35,36 using Ti-containing MOFs directly as photocatalysts has the risk of material decomposition, especially in aqueous environments where most of the photocatalytic reactions are carried out. Alternatively, Ti-containing MOFs can be pyrolyzed into TiO 2 /C composites with much higher stabilities. A previous work has been reported on pyrolyzing Ti-modied IRMOF-3. 37 However, due to the complicated postsynthetic modication of MOF precursor with titanium isopropoxide, the Ti content in the nal pyrolyzed product is only 4.3 wt% which greatly limits its photocatalytic activity. In this communication, we report the facile synthesis of robust TiO x /C composites using Ti-containing MOFs as the sacricial precursors. Thanks to the hybrid and crystalline structure of the MOF precursors, the chemical composition and pore texture of TiO x /C composites can be readily adjusted, rending a series of photocatalysts which exhibit excellent photocatalytic activities towards the degradation of methylene blue.
MIL-125 is one of the earliest Ti-containing MOFs reported so far. 30 It is composed of 1,4-benzenedicarboxylate (BDC) bridged by Ti/O clusters as the secondary building units (SBUs).
In this study, MIL-125 was selected as the MOF precursor, which was pyrolyzed under Ar atmosphere at temperatures of 400, 600, 800, and 1000 C, respectively, with the products named as T4, T6, T8, and T10 (see ESI † for experimental details).
The powder X-ray diffraction (PXRD) pattern of synthesized MIL-125 matches well with the one simulated from published single crystal structure (Fig. 1). The white powder of MIL-125 turned into blackish aer pyrolysis, indicating the formation of carbonaceous materials. The crystallinity of MIL-125 was completely lost in T4, and the PXRD peak at 25.2 belonging to (101) plane of anatase TiO 2 (JCPDS card 21-1272) can be easily identied, revealing the decomposition of crystalline MIL-125 framework and evolution of TiO 2 /C composites under this temperature (400 C, Fig. 1). For the sample of T6 which was pyrolyzed at a high temperature (600 C), more peaks representing anatase have emerged, indicating better crystallinity of anatase phase. Increasing the pyrolysis temperature to 800 C trigged the phase transition of anatase to rutile, rending new peaks at 27.3, 36.0, and 41.2 belonging to (110), (101), and (111) planes of rutile TiO 2 (JCPDS card 21-1276) in T8, respectively. This phase transition is anticipated as anatase is a metastable polymorph of TiO 2 which can be converted to a more stable rutile polymorph at higher temperatures. 38 Neither anatase nor rutile polymorph can be found in T10 which was pyrolyzed at an even higher temperature (1000 C). Instead, the newly emerged peaks match well with the PXRD pattern of g-Ti 3 O 5 (JCPDS card 40-0806), 39 whose formation can be attributed to the carbothermal reduction of TiO 2 under this condition. 40,41 Scanning electron microscope (SEM) images of pyrolyzed MIL-125 indicate a variable distribution of particles of about 0.1-1 mm in size and clusters of particles packed together (Fig. S1 †). Even distributions of C, O, and Ti in pyrolyzed samples can be conrmed by the EDS elemental mapping (Fig. S2 †). Transmission electron microscope (TEM) images of T4 reveal the structure of carbonaceous matrix imbedded with crystalline nanoparticles with the size of around 5 nm, which increases to around 10 nm in T6 with a lattice spacing of d ¼ 0.352 nm representing the (101) plane of anatase polymorph (Fig. 2). The evolution of rutile polymorph in T8 was conrmed by the lattice spacing of 0.325 nm indicating the (110) plane of rutile. Larger particles with the size of 20-50 nm can be found in T10, whose lattice spacing of 0.292 nm matches well with the ( 112) plane of Ti 3 O 5 .
The chemical state changes of C and Ti elements in MIL-125 pyrolyzed at different temperatures were traced through X-ray photoelectron spectroscopy (XPS). As can be seen from the C 1s XPS spectra (Fig. 3a), the shake-up peak at around 291 eV coming from pi-pi* transition becomes more distinct for samples pyrolyzed at higher temperatures, indicating a higher degree of graphitization in these samples. Ti mainly exists as Ti 4+ (TiO 2 ) in samples being treated at lower temperatures (T4 and T6), while higher portions of Ti 3+ and Ti 2+ can be found in samples obtained at higher temperatures (T8 and T10, Fig. 3b). This conclusion agrees well with the PXRD and TEM data in supporting the formation of reduced TiO 2 (such as Ti 3 O 5 ) in MIL-125 pyrolyzed at higher temperatures.
Thermogravimetric analyses (TGA) of the pyrolyzed samples were carried out under an air ow to burn off the carbon matrix and yield white TiO 2 powder as the nal products (Fig. S3 †). 37 The highest Ti content was observed in T10 (50.2 wt%), followed by T6 (40.7 wt%), T8 (38.6 wt%), and T4 (37.0 wt%). The Ti contents in our samples are much higher than the previous report of 4.3 wt% 37 reinforcing the attractiveness of preparing TiO x /C composites using Ti-containing MOFs directly as precursors. It is interesting to note that the TGA curve of T10 exhibits an upward curve at temperatures below 400 C. This can be attributed to the oxidation of Ti 3 O 5 into TiO 2 because the calculated weight increase (5.6 wt%) assuming this process matches well with the measured weight increase (5.3 wt%). 42 Fig. 1 PXRD spectra of samples and reference. Thermal stability of the pyrolyzed samples is also demonstrated as combustion in air won't happen until the temperature is higher than 400 C, which grants the applications of these materials in conditions up to 400 C. The surface area and pore size distribution data were obtained by analyzing N 2 sorption isotherms collected at 77 K (Fig. 4). The isotherm of repeated MIL-125 exhibits a Type I shape indicating the microporous structure which reects the crystal model and matches well with the literature. 30 A slight hysteresis between adsorption and desorption branches was found in MIL-125, which can be attributed to the mesoporous voids formed among crystalline particles. The surface area of MIL-125 was calculated to be 1321 m 2 g À1 using BET model (1563 m 2 g À1 assuming Langmuir model). All the pyrolyzed samples, however, exhibit Type IV isotherms with much smaller surface areas (Table 1), possibly due to the loss of crystallinity and crack of framework during pyrolysis. In T4 and T6, the pore size still remains within microporous range (<2 nm). Higher pyrolysis temperatures help to increase the surface area and trigger the formation of both smaller pores ($0.8 nm) and larger pores (>3 nm) in T8 and T10. It is clearly demonstrated herein that the surface area and pore size of TiO x /C composites can be readily changed through pyrolysis temperature using Ticontaining MOFs as precursors. Although the surface areas of our samples are merely moderate, the versatile synthetic approaches of using MOFs as precursors to prepare carbon materials, such as including supplementary carbon sources within MOF cavities, 24 give enormous opportunities for the netune of porous texture that is rarely seen in other synthetic methods.
The photocatalytic activities of our samples were measured through photodegradation of methylene blue (MB), a common dye used for the simulation of pollutant in industrial effluent streams (see ESI † for experimental details). Two extra samples were added as comparisons. One is P25 TiO 2 that is commercially available with mixed anatase/rutile crystallites commonly   used as the benchmark photocatalyst. The other one is T10 washed with HF (T10/HF) to remove Ti 3 O 5 affording porous carbon. Before subject to irradiation, all the sample solutions were le under stirring in dark for 1 hour to reach physisorption equilibrium of MB (Fig. S4a †). More MB up to 25% can be physically adsorbed into the samples with a higher surface area such as T10, while there is barely any physisorption of MB observed in P25 due to its low surface area (Fig. S4b †). All the pyrolyzed samples exhibited photocatalytic activities towards the degradation of MB (Fig. S5 †). Aer 60 minutes of irradiation, 38.8% of MB was degraded under the catalysis of T10, while only 32.9% of MB was degraded with P25 as the catalyst (Fig. 5a and Table 1). The photodegradation data at irradiation time # 20 minutes were used to t the kinetics as rst order model (Fig. 5b). The apparent rst-order rate constant k was obtained that can reect the speed of photodegradation. T10 has the largest k, which is 32% higher than that of P25. It can be safely concluded that compared to the benchmark P25 TiO 2 , at least one pyrolyzed sample (T10) reported herein has a better photocatalytic activity towards the photodegradation of methylene blue. The poorest activity was observed in T10/HF, where most of the Ti 3 O 5 has been removed through HF wash. This fact conrms that the superior photocatalytic activity of T10 should come from the imbedded Ti 3 O 5 nanoparticles.
Several factors need to be considered in order to enunciate the relationship between the structure of pyrolyzed samples and their photocatalytic activities. The most obvious factor is the content of TiO x as they serve as the catalytic sites. T10 has the highest Ti content among the pyrolyzed samples which may justify its good activity. However, its better activity compared to pure TiO 2 needs further explanation. Since the photodegradation described herein is heterogeneously catalyzed, the access of substrates towards active sites is another important factor. Compared to pure TiO 2 , pyrolyzed samples have TiO x as nanoparticles spread apart and supported on porous carbon matrix that make it easier for the substrate (MB in this case) to access and react. This point is commonly accepted in heterogeneous catalysis 43 and is supported by the high surface area of T10. However, surface area should not be the sole factor in determining photocatalytic activities. The surface areas of T8 and T10 are similar while T8 only demonstrates sluggish reaction kinetics. A closer observation of material structure reveals a huge difference of TiO x between T8 and T10. Ti exists as TiO 2 in T8, with a substantial portion of rutile polymorph that is catalytically less active than its anatase analogue. 44 Although it has been reported that TiO 2 with anatase/rutile mixed phases has better photocatalytic activities due to efficient electron-hole separation, 45,46 such enhanced catalytic activity was not observed in T8 possibly due to the separation of anatase and rutile nanoparticles in carbon matrix that prevents close contact and impairs synergetic effect. In T10, nevertheless, TiO 2 is carbothermally reduced into Ti 3 O 5 with higher density of oxygen vacancies for better trapping of electrons affording more efficient photocatalysis. 47 Although Ti 3 O 5 has been extensively studied for its phase transition behaviour, [48][49][50] there is seldom any report for its photocatalytic performance 51 and this work is among the initial studies to the best of our knowledge. The last factor that needs to be taken into consideration is the conductivity of carbon matrix. One of the properties of TiO x /C composites is the electron conductivity provided by the carbon matrix which can afford longer electron-hole recombination time for more efficient catalysis. From this perspective, samples pyrolyzed at higher temperatures (such as T10) should have better activities because of the higher electron conductivity brought by higher degree of graphitization proved by XPS previously.
In summary, a series of TiO x /C composites were prepared through pyrolysis of a Ti-containing MOF MIL-125. They exhibited considerable photocatalytic activities towards the photodegradation of methylene blue. Among these materials, T10 has the best activity due to its high surface area, reduced Ti 3 O 5 composition, and conductive carbon support. This work demonstrates the potency of preparing TiO x /C composites as photocatalysts using Ti-containing MOFs as precursors. More work will be focused on ne-tuning the porosity, polymorph, and elemental doping of TiO x /C composites through the judicious selection of functional MOF precursors and pyrolysis conditions.