A TiO2/C catalyst having biomimetic channels and extremely low Pt loading for formaldehyde oxidation

This study presents a TiO2/C hybrid material with biomimetic channels fabricated using a wood template. Repeated impregnations of pretreated wood chips in a Ti precursor were conducted, followed by calcination at 400–600 °C for 4 hours under a nitrogen atmosphere. The generated TiO2 nanocrystals were homogenously distributed inside a porous carbon framework. With an extremely low Pt catalyst loading (0.04–0.1 wt%), the obtained porous catalyst could effectively oxidize formaldehyde to CO2 and H2O even under room temperature (conv. ∼100%). Wood acted as both a structural template and reduction agent for Pt catalyst generation in sintering. Therefore, no post H2 reduction treatment for catalyst activation was required. The hierarchal channel structures, including 2–10 nm mesopores and 20 μm diameter channels, could be controlled by calcination temperature and atmosphere, which was confirmed by SEM and BET characterizations. Based on the abundant availability of wood templates and reduced cost for low Pt loading, this preparation method shows great potential for large-scale applications.

Formaldehyde (HCHO) is an important chemical feedstock that is widely used in the production of industrial resins, polymers and coating materials. However, formaldehyde is highly toxic and volatile, and it has been listed as a human carcinogen group I by the International Agency for Research on Cancer (IARC). 1 Formaldehyde is slowly emitted from building and furnishing materials, which can cause serious health problems. Developing an efficient and feasible catalyst for the removal of formaldehyde pollutant from indoor environment has received a growing concern recently. Inorganic porous materials are good candidates for application in the removal of formaldehyde either by physical absorption or chemical catalytic oxidation. 2,3 Inorganic porous materials have multi-shapes of pore structures and large surface areas, which allow formaldehyde molecules access into holes. Moreover, the inorganic framework of the material has high thermal stability, suggesting that the material can be thermally regenerated and reused for many times. Many inorganic porous material-based formaldehyde absorbents, such as active carbon, 4 porous Al 2 O 3 (ref. 5) and zeolite, 6 have been investigated. For example, different porous Al 2 O 3 materials were prepared by using the template method, which showed broad pore size distribution and good formaldehyde adsorptiondesorption behavior. 7,8 However, the porous inorganic adsorbents have limited adsorption capacities, resulting in only a short effective time.
Formaldehyde oxidation over porous material-supported catalysts is a promising technology because formaldehyde can be continuously oxidized to CO 2 and H 2 O. [9][10][11] The porous structure can facilitate fast diffusion and mass transport of formaldehyde and products. Furthermore, the high surface area of the porous material provides a large number of active sites for formaldehyde adsorption and catalyst loading. Many catalystloaded porous materials or porous transition metal oxides were prepared and displayed excellent formaldehyde oxidation performance; some examples include Pt on ZSM-5 and NaY zeolites, 12 [30][31][32] and even small organic molecules including salicylic acid and aspartic acid. 33 However, these organic templates generally have a high price, which hinders their potential for large-scale applications. Another issue is that the noble metal catalyst loading is still high (1-2.5%) to reach effective formaldehyde oxidation.
Biomass materials have low cost and natural hierarchal structures in the range from millimeters to nanometers, 34 which can be used as natural templates for porous material fabrication. TiO 2 with a porous structure has been prepared by biotemplating of naturally grown biomass materials, such as wood and bamboo; 22,35 however, there is no reported example on the use of the material as a catalyst in reactions such as organic oxidation. Herein, we presented a wood templated TiO 2 / C material having biomimetic channel structures. The obtained material was different from the supported catalysts or coreshell structure materials, in which the catalyst layers are only coated inside or outside the porous structure. Wood templated TiO 2 /C exhibited a hybrid structure; this resulted in the generation of TiO 2 nanocrystals, which were homogenously distributed inside the porous carbon framework because the Ti precursor penetrated into wood tissues before catalyst sintering. With an extremely low loading of Pt, this type of hierarchal structured material showed highly effective formaldehyde oxidation behavior even under room temperature. Fig. 1 shows the preparation process and schematic illustration of the obtained TiO 2 /C hybrid porous structure. Simple impregnation of wood chips in the organic titanium precursor and the following drying process were repeated alternately. Aer the Ti precursor penetrated into wood tissues, the Pt precursor also inltrated into wood and then was calcinated under a controlled temperature and atmosphere.

Results and discussion
The morphology of biomimetic channel-containing TiO 2 /C material prepared by the wood template is shown in Fig. 2, showing multi-scaled hierarchal structures. As shown in Fig. 2A, the channel structure is formed aer sintering at 400 C. This could be produced by the xylem vessels in wood. During the impregnation process, the organic titanium precursor Ti(OBu) 4 diffused into the xylem vessels and tissues or was coated on the surface of vessel walls (Fig. 1). Furthermore, pre-formed TiO 2 sol could also be lled into the vessels of wood; aer solvents evaporated, the TiO 2 sol was xed on the vessel walls. The TiO 2 precursor layer could be increased by repeating the impregnation and drying processes alternatively. With sintering at a high temperature, the organic wood components, including hemicellulose, cellulose and lignin, were converted to porous carbon. The TiO 2 precursor was transformed into TiO 2 and the whole material maintained the physical structure of the wood template. The formed TiO 2 /C channels exhibited diameter of ca. 20 mm, which was the same as the size of xylem vessels in wood. Some pits were observed in the TiO 2 channels. Pits are cavities located on wood cell walls, which are essential components for water-transport system in wood. 36 The pitted structure was maintained perfectly through the impregnation method aer sintering, which increased the porosity and helped improve the catalytic performance over the obtained material. As shown in the magnied images ( Fig. 2C and D), the pits exhibit a cylindrical shape with diameter of ca. 10 mm.
Different samples were prepared under the sintering temperatures of 400, 500 and 600 C in this study. The samples showed similar structures, i.e., hierarchical channels. The obtained morphologies under sintering at 500 and 600 C are shown in Fig. 2E and F. The SEM images verify the successful preparation of a porous TiO 2 material templated by wood.  Before sintering, the catalyst (Pt) precursor was loaded into the channel structure of the material. The loading process was similar to the preparation of a TiO 2 layer, but a solution of H 2 PtCl 6 dissolved in isopropanol was used. The desired amount of catalyst could be reached by repeating the impregnation and drying processes. The advantage of the impregnation method for catalyst loading is the homogenous dispersion of the catalyst. In order to investigate the chemical composition, elemental mapping was performed using energy-dispersive spectrometry (EDS) (shown in ESI, Fig. S1. †). For the sample calcined at 400 C, the results showed that the TiO 2 component included 25.3% and 74.5% of carbon and 0.1% of Pt loading amount. The exact Pt contents in the prepared catalysts were measured by an inductively coupled plasma mass spectrometer (ICP-MS), showing that the samples calcined at 400, 500 and 600 C contained 0.092%, 0.095% and 0.098% weights of Pt, respectively. In addition, the catalysts with different amounts of Pt loading were studied, and Pt loadings were found to be 0.063% and 0.041% (shown in ESI Table S1 †).
X-ray diffraction (XRD) was employed to study the crystalline phase of prepared porous TiO 2 /C materials. Fig. 3 shows the XRD patterns of different samples calcined at 400, 500 and 600 C. As the calcination temperature increased, the intensity of XRD peak increased, indicating the increase in crystallinity of TiO 2 compared to the amorphous phase. The XRD peaks could be identied by comparing with the standard diffraction peaks of anatase and rutile TiO 2 , showing that the samples contained both anatase and rutile phases of TiO 2 . The XRD peaks at 2q ¼ 25.25 (101) and 2q ¼ 48.03 (200) veried the formation of the anatase phase, while the peaks at 2q ¼ 27.42 (110) and 2q ¼ 36.08 (101) indicated the presence of the rutile phase. 37,38 Signicant phase transformation from anatase to rutile was observed under a high sintering temperature. The ratio of anatase to rutile can be estimated by the intensity of characteristic diffraction peaks of both phases. According to the Spurr-Meyers equation, the weight percentages of anatase (X A ) and rutile (X R ) can be calculated as follows: 39,40 (1) Here, I A is the anatase phase intensity at 2q ¼ 25.25 ; I R is the intensity of rutile peak at 2q ¼ 27.42 .
The calculation results showed that the anatase phase accounted for 63.5% in the sample sintered at 400 C, but this value signicantly decreased to 16.8% when a high sintering temperature of 600 C was applied ( Table 1). The average crystalline size of obtained TiO 2 was calculated by Debye-Scherrer equation, indicating that the crystal size increased at a high calcination temperature.
To investigate the phase transition behavior with the wood template, TiO 2 samples were prepared using the sol-gel method without wood template but with the same calcination conditions. The XRD characterization showed a high content ratio of anatase phase in the TiO 2 samples prepared using the sol-gel method (shown in ESI, Fig. S2. †). However, in wood templated TiO 2 preparation, all samples showed anatase and rutile mixed phases. These results indicated that the wood template actually promoted the anatase-to-rutile phase transition in TiO 2 powders because the transformation behavior can be affected by many factors such as impurities in wood, grain sizes of formed TiO 2 crystals, preparation methods and annealing conditions. Fig. 4 shows the nitrogen adsorption isotherms and pore size distribution curves of the obtained porous TiO 2 material templated by wood. The isotherm curves of all samples can be classied as type-IV isotherms based on the Brunauer-Deming-Deming-Teller classication, 41 indicating the mesoporous structure in the obtained TiO 2 /C material. The isotherms show hysteresis loops that have an H3-type shape, 42 suggesting the presence of a layered structure with slit-like pores (multichannel structures in this material). These results were consistent with the structure observed using SEM. The specic surface area was 41.5 m 2 g À1 for the sample sintered at 400 C and the value slightly decreased with the high sintering temperatures of 500 and 600 C. The pore volumes of the obtained TiO 2 /C material were in the range of 0.055-0.033 cm 3 g À1 . The specic surface areas and pore volumes of obtained samples in this study were similar to those of previously reported biomass templated materials. 35,43 The corresponding pore size distribution curves show that the wood templated TiO 2 materials have meso-pores with width of 2-10 nm (shown in ESI, Table S2. †). It should be noted that some pores exhibited large size of 12-18 nm. The large-sized pores could be formed by wood template removal during sintering. The micro-sized   46 The Pt 4f XPS spectra veried the reduction of Pt during the sintering process by wood because there was no post reduction treatment by H 2 gas in this study. Wood acted as both a structural template and reduction agent of Pt catalyst in the preparation process. The Ti 2p spectra are shown in Fig. 5C. The peaks at ca. 464.0 and 458.2 eV were ascribed to Ti 2p 1/2 and Ti 2p 3/2 of Ti 4+ in TiO 2 , respectively. The O 1s spectra are shown in Fig. 5D, which include many overlapped peaks. The tting peak at ca. 530 eV represents the binding energy of O 1s of TiO 2 lattice oxygen (Ti-O-Ti). 20,45 A signicant shoulder peak at ca. 533 eV appeared beside the Ti-O-Ti peak in the spectrum of the sample sintered at 400 C. This shoulder peak included two tting peaks: one at 531.6 eV, which can be ascribed to the oxygen in Ti-OH group on the TiO 2 surface, and the other at 533.8 eV, which can be ascribed to organic oxygen in C-O bonds. 47 These results indicated that there is a small amount of organic residues in the sample aer sintering at 400 C in air for 4 hours. The C-O peak disappeared in the spectra of the samples sintered at 500 and 600 C, which indicated that the organic components of wood were completely oxidized at high sintering temperatures. The Ti-OH peak also reduced with increase in the sintering temperature. The Ti-OH group, which was certicated to be active in the oxidation of formaldehyde, could be formed by the adsorption of oxygen or H 2 O molecules on the surface of TiO 2 . Therefore, the sample that has a large ratio of Ti-OH groups could have high catalytic activity.
In order to investigate mass changes during the sintering process, thermogravimetric analysis (TGA) was performed under air atmosphere. The wood samples aer the impregnation of TiO 2 and Pt precursors were used for TGA measurements. With the increase in temperature, TiO 2 and catalyst precursors were transformed into crystalline and metallic forms, and wood was decomposed to carbon. The TGA curve with the temperature up to 400 C is shown in Fig. 6A. The sample mass continuously decreased with increasing temperature and constant temperature sintering. The mass loss rate curve shows that the temperature at the highest decomposition rate of wood is 350 C. However, the mass loss rate curve with the temperature up to 500 C indicates two highest decomposition rate temperatures of 350 and 500 C, as shown in Fig. 6B. The possible reason is that wood could be pyrolyzed and carbonized under a low temperature of 350 C and oxidized at a high temperature of 500 C. Compared with the TGA curve of the original wood template, the result showed that the highest   decomposition rate temperatures of wood chips were 350 and 380 C (shown in ESI, Fig. S3. †). It should be noted that mass was continuously lost at the constant temperature of 500 C (shown in the curve in Fig. 6B). The organic residues could be completely oxidized if there was sufficient reaction time. The TGA curve with the temperature up to 600 C has similar result compared with the curve of 500 C. The sample mass was stable aer 1 hour sintering at 500 or 600 C, indicating complete removal of wood template.
The catalytic behavior of the prepared porous TiO 2 material was evaluated in a xed bed reactor. Gaseous HCHO and humid air gas (water vapor 50%) were introduced into the reactor tube, and the HCHO concentration was measured at the outlet of the reactor. The concentrations of HCHO in the stream owing in and out of the reactor were measured. As shown in Fig. 7A, the HCHO concentration signicantly decreases from the initial 104 ppm to ca. 0.2-0.5 ppm at the reaction temperature of 35 C. These results showed high catalytic oxidation performance of formaldehyde over the catalyst with a low Pt loading (0.09%). Anatase phase TiO 2 was prepared by the sol-gel method without a wood template. The catalytic performance evaluation showed that 44 ppm of HCHO still remained in the downstream of catalyst bed, which indicated low conversion of formaldehyde oxidation. The high catalytic performance of wood templated porous TiO 2 can be ascribed to the channelled micro-nano hierarchal porous structure and the synergism of anatase-rutile mixed phases. The chemical analysis of oxidation products in the out stream was conducted. It veried that HCHO was completely oxidized to CO 2 because no CO was detected, as shown in Fig. 7B.
The conversions of formaldehyde oxidation reaction on different prepared catalysts under different reaction temperatures (20-35 C) and humidities are summarized in Fig. 7C and D. The results indicated that the wood templated TiO 2 catalyst materials have high catalytic performances in our test time. The sample sintered at 400 C exhibited higher formaldehyde oxidation conversion than the samples sintered at 500 and 600 C. This result could be explained by the crystalline phase of TiO 2 and the amount of hydroxyl groups (Ti-OH) on the surface of the TiO 2 material. Previous studies show that anatase TiO 2 has higher catalytic activity than rutile TiO 2 . 44,45,48 Therefore, the sample sintered at a low temperature exhibited a high content of anatase phase, resulting in high catalytic activity. In addition, the high amount of surface hydroxyl groups led to high oxidation conversion of formaldehyde, which was veried by reported studies. 48,49 The high Ti-OH ratio conrmed by XPS for the sample sintered at 400 C also improved the catalytic performance of formaldehyde oxidation. The formaldehyde oxidation performance was related to the humidity of air (Fig. 7D). When dry air was introduced in the oxidation reaction, the conversion slightly decreased. This was because the moisture in air could be adsorbed and reacted with the surface of TiO 2 , resulting in increase in the concentration of hydroxyl groups on the catalyst surface. 50 The catalysts with different Pt loading amounts were also evaluated in formaldehyde oxidation in a xed bed reactor (Fig. 7E). The catalyst with 0.092% Pt content showed the highest oxidation conversion (close to 100%). By slightly decreasing the Pt loading amount to 0.063%, the catalyst still showed high performance with the conversion higher than 95%. However, the conversion decreased to around 85% when 0.041% Pt was loaded. Compared with the performance of the sol-gel catalyst without a wood template, the conversion was as low as 60%. The durability of wood templated catalyst was investigated as well. As shown in Fig. 7F, the porous TiO 2 catalyst calcined at 400 C maintains high performance in a long test time of 12 hours, showing only 2% decrease in conversion. Fig. 8 illustrates the reaction process between moisture and TiO 2 surface. The formed hydroxyl groups were active sites for formaldehyde adsorption. The oxidization reaction started with the adsorption of formaldehyde on the TiO 2 surface and oxygen on the Pt catalyst surface. The absorbed formaldehyde molecules were then transformed to formic acid over the Pt catalyst surface. With the elimination of H 2 O and CO 2 , the catalyst surface could be restored to the initial state, as shown in Fig. 8B. The wood templated channel structure largely improved the reactant and product diffusion and catalytic performance. Therefore, this prepared biomimetic TiO 2 material showed good catalytic performance even at an extremely low Pt catalyst loading.

Conclusions
TiO 2 catalyst materials containing biomimetic channels were successfully prepared by a wood template. SEM images veried the hierarchal channel structure (width ca. 20 mm) formed by xylem vessels in wood. BET analysis showed that the prepared TiO 2 material exhibited 2-10 nm mesopores, which could be formed by small structures in wood-like pits. XRD showed that the prepared TiO 2 is a mixture of anatase and rutile phases. XPS certied the presence of metallic Pt, which is the active component in the oxidation of formaldehyde. The results of catalytic performance evaluation showed that the prepared TiO 2 catalysts have high oxidation conversion of formaldehyde.

Fabrication method
Birch wood was cut into chips with size of 4 mm Â 2 mm Â 1 mm. The wood chips were pre-treated with mixed organic solvent (volume of benzene to absolute ethanol was 2 : 1) under reux for 6 hours to remove the extractives in wood. Aer solvent was evaporated at 80 C, the wood chips were washed with deionized water at 80 C for several times and then dried in oven overnight.
The Ti precursor was prepared by partial hydrolysis of titanium butoxide. Titanium butoxide was dissolved in isopropanol to form 20 wt% solution with stirring (noted as solution A). Water (2 wt%)-isopropanol solution (10 mL) was used as solution B, which was slowly added to 10 mL solution A with stirring. Aer aging for 1 hour with stirring, the clear and transparent Ti precursor solution was obtained. Pre-treated wood chips were impregnated in the precursor solution under ux for 4 hours and then removed from the solution and dried at 120 C for 2 hours. This process was repeated until the weight of wood chips increased by 20%.
Catalyst precursor solution was obtained by dissolving chloroplatinic acid (H 2 PtCl 6 ) in isopropanol (nal H 2 PtCl 6 concentration, 0.1 wt%). The wood chips aer Ti precursor solution treatment were impregnated in catalyst precursor solution at 50 C for 2 hours with stirring and then dried in an oven overnight. With repeated impregnation and drying processes, different amounts of Pt catalyst loadings could be achieved. Finally, the obtained wood chips were sintered in N 2 ow. Different samples were prepared at the sintering temperatures of 400, 500 and 600 C with the heating rate 10 C min À1 . Fig. 1 shows the synthesis process of wood templated TiO 2 catalyst material.

Characterizations
The amount of loaded Pt was measured by an inductively coupled plasma mass spectrometer (ICP-MS, NexIon, Perkin-Elmer). Before measurement, the samples were dissolved in aqua regia (HNO 3 : HCl ¼ 1 : 3 v/v) with the heating of microwave oven. Aer cooling, HF solution was added and mineralized to obtain a clear solution. The morphology of obtained porous TiO 2 catalyst material was characterized by scanning electron microscopy (SEM, Zeiss Ultra60 FE-SEM instrument) coupled with INCA 200 X-Sight de Oxford instruments for energy-dispersive X-ray spectroscopy (EDS) analysis. The formed TiO 2 crystalline phase was veried by powder X-ray diffraction (PXRD). PXRD was conducted using a PANalytical X'Pert X-ray diffractometer coupled with X'Celerator detector and Cu Ka (l ¼ 1.5418Å). X-ray photoelectron spectroscopy (XPS) analysis was performed with a Thermo K-Alpha instrument equipped with monochromatic AlKa radiation at 1486.6 eV X-ray source. N 2 physisorption isotherms were collected at 77 K on a Quadrasorb system from Quantachrome Instruments. Before the measurements, the sample ($50 mg) was degassed under dynamic vacuum at 150 C for 18 hours. TGA analysis results were obtained using an STA 6000 Simultaneous Thermal Analyzer (PerkinElmer) with the heating rate of 10 C min À1 and air ow of 10 mL min À1 .

Catalytic formaldehyde oxidation test
The catalytic formaldehyde oxidation test was performed in a xed-bed quartz ow reactor with a gas mixture of air containing 100 ppm HCHO at a low rate 50 cm 3 min À1 . The humidity of air gas was controlled at 50% by bubbling through water. The concentrations of HCHO, CO and CO 2 were analyzed by gas chromatography (GC).
The conversion of HCHO was calculated according to the following equation: HCHO conversion ¼ 1 À HCHO concentration in out-stream HCHO concentration in in-stream Â 100%

Conflicts of interest
There are no conicts to declare.