Enhanced Formaldehyde Oxidation Performance of the Mesoporous TiO2(B)-Supported Pt Catalyst: The Role of Hydroxyls.

As one of the crystal phases of titania, TiO2(B) was first utilized as a catalyst carrier for the oxidation of formaldehyde (HCHO). The mesoporous TiO2(B) loaded with Pt nanoparticles enhanced the HCHO oxidation reaction whose reaction rate was 4.5-8.4 times those of other crystalline TiO2-supported Pt catalysts. Simultaneously, Pt/TiO2(B) exhibited long-term stable HCHO oxidation performance. The structural characterization results showed that in comparison with Pt/anatase, Pt/TiO2(B) had more abundant hydroxyls, facilitating increasing the content of oxygen species. Studies on the role of hydroxyls in HCHO oxidation of Pt/TiO2(B) illustrated that synergistic involvement of terminally bound hydroxyls and bridging hydroxyls in HCHO oxidation accelerated the transformation from HCHO to formate via dioxymethylene. Moreover, hydroxyls could avoid the accumulation of excessive formate on Pt/TiO2(B) and promote the rapid oxidation of CO. Accordingly, the hydroxyl groups could accelerate each substep of formaldehyde oxidation, which enabled Pt/TiO2(B) to exhibit excellent formaldehyde oxidation performance.

Introduction

Improvement in the lives of the masses arouses great public concern about the importance of health. Formaldehyde (HCHO) is considered as one of the most representative pollutants that can cause several diseases.[1,2] Among developed treatments, the oxidation of HCHO into nontoxic H2O and CO2 is an environmentally friendly and sustainable technology for HCHO removal.[3−5] To date, the supported noble metal catalysts have demonstrated more advanced catalytic performance compared with other types of HCHO oxidation catalysts. However, the high cost of noble metals severely limits the wide application of corresponding types of catalysts. Therefore, an attempt to promote the HCHO oxidation performance of catalysts is a feasible strategy for reducing the usage of noble metals.[6−8]

In the development course of catalysts, the regulation of the catalytic support is a relatively facile strategy to improve the catalytic performance of catalysts. TiO2, Al2O3, CeO2, activated carbon, cellulose triacetate, and so on were chosen as the support to investigate their HCHO oxidation performances.[9−12] It was found that the HCHO oxidation mechanisms on catalysts with various supports were distinct from each other. On this basis, several aforementioned materials were compounded together to give full play to their advantages.[13,14] Moreover, the approaches of enlarging specific areas, constructing hierarchical pores, and modifying surface properties have been proved to promote the dispersion of noble metals, adjust the electronic structure of active sites, accelerate the transfer of reactants, or enrich the hydroxyls of the support.[7,15−17] Although great progress has been made in the past, the strategy on the support regulation still attracts increasing attention for enhancing the HCHO oxidation performance of the catalyst.

TiO2 is the most commonly used support in HCHO oxidation, which is usually regarded as the research benchmark in this area.[6,18,19] Anatase, rutile, and brookite are three crystal phases of TiO2. Jiang et al. comparatively studied the turnover frequency (TOF) value for HCHO oxidation of the rutile TiO2-supported Pt catalyst, which was five times as high as that of the anatase TiO2-supported counterpart.[20] However, most research studies were carried out on anatase or anatase-rutile mixed phases because the anatase TiO2 with a large surface area was easy to be synthesized. As for the brookite TiO2, Yu et al. constructed hierarchical porous TiO2-supported Pt catalysts. They found that the presence of a small amount of brookite TiO2 in the support made negligible contribution to the HCHO oxidation performance of the catalyst.[21]

In addition to the aforementioned crystal phases, TiO2(B) represents the metastable phase of TiO2, which has the potential to be applied as a prospective lithium electrode, sensor material, and photocatalyst.[22−24] To the best of our knowledge, TiO2(B) has not been used as a catalyst support for HCHO oxidation until now. Compared to the three aforementioned TiO2 polymorphs in structure that have only one type of oxygen atom (i.e., three-coordinated one), TiO2(B) features two, three, or four-coordinated oxygen atoms on the surface.[25−27] This leads to different surface properties between TiO2(B) and other crystalline phases of titania. Previous research studies have demonstrated that the adsorbed water molecules could readily dissociate into hydroxyls on the TiO2(B) surface.[25] It is well acknowledged that hydroxyl groups play a crucial role in the oxidation of HCHO by supported noble metal catalysts, such as promoting the adsorption and transfer of oxygen, assisting the catalyst in adsorbing HCHO, directly reacting with formate to generate CO2 and H2O, and so on.[28−31] Accordingly, considering the unique properties of TiO2(B), it is urgent and meaningful to conduct related research on the TiO2(B)-supported catalyst for HCHO oxidation.

In the present work, mesoporous TiO2(B) was used as the support to load Pt nanoparticles (NPs) for synthesizing the HCHO oxidation catalyst (denoted as Pt/TiO2(B)). As expected, the Pt/TiO2(B) catalyst exhibited better HCHO oxidation performance than Pt NPs supported on anatase, rutile TiO2, and P25. The structural properties of various catalysts were systematically studied by Raman spectroscopy, X-ray photoelectron spectroscopy (XPS), Brunauer–Emmett–Teller (BET) analysis, scanning electron microscopy (SEM), transmission electron microscopy (TEM), CO-diffuse reflectance infrared Fourier transformed spectroscopy (DRIFTS), and so on. Simultaneously, the structure–performance relationship discussion was supported by HCHO-DRIFTS analyses.

Results and Discussion

Overall Structural Properties of Catalysts

The XRD pattern and Raman spectrum of the Pt/TiO2(B) catalyst are displayed in Figure . There are six obvious peaks at 2θ = 25.0°, 28.7°, 29.9°, 33.3°, 37.8°, 43.6°, and 48.5°, as shown from the XRD pattern in Figure a. These peaks are ascribed to the (110), (002), (−401), (310), (003), and (020) reflections of the TiO2(B) phase, respectively (PDF#35-0088). The characteristic peak of Pt species is not observed in the XRD pattern, suggesting the good dispersion of Pt NPs in the Pt/TiO2(B) catalyst.[32] It is acknowledged that the Raman technique is more sensitive than XRD in detecting the crystal information of TiO2(B).[33] In the Raman spectrum of Pt/TiO2(B) presented in Figure b, the main modes at 122, 197, 241, 248, 364, 407, 431, 468, 636, and 654 cm–1 are the direct expression of the corresponding characteristic Raman vibrations of the TiO2(B) phase.[26,33] The signal related to the anatase phase could not be observed in the Raman spectrum, which indicates that the support of Pt/TiO2(B) consists of pure TiO2(B) only.

Figure 1

X-ray diffraction (XRD) pattern (a) and Raman spectrum (b) of Pt/TiO2(B).

Figure displays nitrogen adsorption–desorption isotherms and pore size distributions of Pt/TiO2(B). The catalyst Pt/TiO2(B) has a type IV isotherm with a H3 hysteresis loop in a relative pressure p/p0 range of 0.4–0.6, suggesting the mesoporous structure of the Pt/TiO2(B) catalyst. As shown in Figure S1, all control catalysts exhibited type IV isotherms and possessed mesoporous structures.[34] The inset in Figure displays that the most probable pore size of the Pt/TiO2(B) catalyst is about 3.9 nm and its pore size distribution is relatively uniform. Furthermore, the detailed pore structural data of catalysts are shown in Table . The specific surface area of Pt/TiO2(B) is 53.0 m2 g–1, slightly higher than that of the control catalysts (40.1 m2 g–1 for Pt/anatase, 30.6 m2 g–1 for Pt/rutile, and 41.0 m2 g–1 for Pt/P25). In addition, the pore volume of the Pt/TiO2(B) catalyst is 0.16 cm3 g–1, comparable to those of the control catalysts (0.14–0.40 cm3 g–1).

Figure 2

Nitrogen adsorption–desorption isotherms and pore size distributions (inserted diagram) of the Pt/TiO2(B) catalyst.

Table 1

Comparison in Porous Structural Data and Reaction Rate of Pt/TiO2(B) with Control Samples

catalyst	SBET (m2 g–1)	VP (cm3 g–1)	DP (nm)	crystal phase	metal type and amount	T (°C)	C0 (ppm)/GHSV (h–1)	reaction rate (μmolHCHO g–1cat s–1)	reference
Pt/TiO2(B)	53.0	0.16	8.5	TiO2(B)	0.5% Pt	30	180/54,000	15.8	present work
Pt/anatase	40.1	0.40	19.5	anatase	0.5% Pt	30	180/54,000	1.89	present work
Pt/rutile	30.6	0.14	7.6	rutile	0.5% Pt	30	180/54,000	3.53	present work
Pt/P25	41.0	0.28	22.8	anatase/rutile	0.5% Pt	30	180/54,000	2.84	present work
Pt/TiNT	31.6			anatase	0.4% Pt	30	50/30,000	4.42	(35)
Pt/P25	51.1			anatase/rutile	0.4% Pt	30	50/30,000	0.84	(36)
Pt/TiO2	47.4			anatase	0.5% Pt	20	100/50,000	3.10	(37)
Pt/TiO2				rutile	1.0% Pt	40	375/60,000	1.40	(38)
Ag/TiO2	35.3	0.30	27.4	anatase	8.0% Ag	35	130/10,0000	0.11	(39)
Pt/ZSM-5	326.1			ZSM-5	0.4% Pt	30	50/30,000	1.36	(40)
Pd/TiO2	57.2			anatase/rutile	1.0% Pd	25	140/95,000	1.65	(16)
Pt/Fe2O3				α-Fe2O3	2.4% Pt	30	400/60,000	10.61	(41)
Ag/SBA-15	520.6	0.81	6.2	amorphous	7.0% Ag	30	1000/15,000	0.29	(42)
Au/Al2O3				γ-Al2O3	3.5% Au	30	500/35,400	2.51	(43)
Au/CeO2	130.0	0.21	3.4	fluorite	1.0% Au	25	200/55,000	4.09	(44)

Formaldehyde Decomposition Performance

Figure a shows the HCHO conversion of various catalysts as a function of reaction temperature. From 30 to 140 °C, the HCHO conversion of Pt/TiO2(B) is invariably higher than those of anatase, rutile, and P25 TiO2-supported Pt catalysts. Note that the HCHO conversion of Pt/TiO2(B) can reach 65% at 30 °C (i.e., room temperature), while those of the three control samples are only between 5 and 15%, indicating the excellent room-temperature HCHO oxidation performance of Pt/TiO2(B). As listed in Table , the corresponding HCHO oxidation rate of Pt/TiO2(B) is 4.5–8.4 times as high as that of Pt/anatase, Pt/rutile, and Pt/P25. Pt/TiO2(B) can achieve complete conversion of HCHO at 80 °C under the present experimental conditions. By comparison, the reaction temperatures of HCHO complete conversion on all control catalysts are above 140 °C. The HCHO conversion of Pt/anatase is slightly higher than that of Pt/rutile, but both of which are obviously lower than that of Pt/TiO2(B). Apparently, mesoporous TiO2(B) as the support has a significant advantage in catalytically oxidizing HCHO at room temperature, as confirmed by the superior reaction rate of Pt/TiO2(B) to those in the reported literature (as seen in Table ).

Figure 3

HCHO conversion as a function of reaction temperature (a) and long-term reaction results (b) of Pt/TiO2(B) and various control samples.

In general, the pore structure of catalysts directly influences the transport of reactants/resultants and the dispersion state of active species during catalyst preparation.[21,45,46] Accordingly, the catalytic performance of Pt/TiO2(B) was preliminarily analyzed with the pore structure data of catalysts. The porous structural data in Table show that the specific surface area of Pt/TiO2(B) is somewhat larger than those of Pt/anatase, Pt/rutile, and Pt/P25, while the pore volume value of Pt/TiO2(B) is between them. According to the similar porous properties of Pt/TiO2(B) and control samples, it could be determined that the geometrical factor of Pt/TiO2(B) is not the main factor for its excellent room-temperature HCHO oxidation performance.

Moreover, it can also be concluded from the above data that the structural properties and catalytic performance between the anatase TiO2-supported Pt catalyst and rutile TiO2-supported one are different but not significant. The catalytic differences between them have been discussed previously.[20,47] Thus, in the following content, Pt/anatase is selected as the primary control sample for comparatively studying the unique structural and catalytic characteristics of Pt/TiO2(B). Figure b displays the results of the long-term reaction of HCHO oxidation of Pt/TiO2(B) and Pt/anatase. Generally, there is only slight fluctuation in HCHO conversion of Pt/TiO2(B) and no significant deactivation during the process of stability evaluation, which is consistent with the results of Pt/anatase. This manifests the excellent long-term HCHO oxidation performance of Pt/TiO2(B) at room temperature.

Dispersion State of Pt Species

Field-emission scanning electron microscopy (FESEM) and TEM images of Pt/TiO2(B) are shown in Figure . Varied from spheroidal particles of Pt/anatase (Figure S2), Pt/TiO2(B) (Figure a) exhibits a whisker structure with several micrometers in length and a few dozen to 200 nm in diameter. As displayed from the TEM image in Figure b, Pt/TiO2(B) has an identical morphology to that in the FESEM image. The high-resolution TEM image in Figure c clearly shows that numerous dark spots are evenly dispersed in the Pt/TiO2(B) catalyst. One of the spots is randomly chosen to analyze its fringe patterns (Figure d). The lattice fringe spacing is 0.23 nm, which matches with the (111) plane of metallic Pt. Hence, all the dark spots can be identified as the Pt NPs. The element mapping shown in Figure S3 also confirms the good dispersion of Pt NPs on the catalyst surface. Moreover, the lattice fringe spacing of the support nearby Pt NP is 0.35 nm, which is assigned to the (110) plane of TiO2(B),[22,26] further confirming the Raman and XRD analyses.

Figure 4

FESEM image (a) and TEM images (b–d) of Pt/TiO2(B). The inset is the size distribution of Pt NPs of Pt/TiO2(B).

The statistical result of the Pt NP size of Pt/TiO2(B) inserted in Figure c shows that the most probable particle size of Pt NPs of Pt/TiO2(B) is ca. 1.8 nm, slightly smaller than that of Pt/anatase (2.2 nm, as seen in Figure S2). This indicates that Pt NPs can be well dispersed on both TiO2(B) and anatase, which is consistent with the above XRD analyses. The detailed structural properties of Pt NPs are further analyzed by CO-DRIFTS. Figure displays that both Pt/TiO2(B) and Pt/anatase have two distinct bands at 2056 and 2078 cm–1, which are assigned to the linear adsorption of CO on step and terrace sites of Pt NPs, respectively.[48−50] The similar shape in CO-DRIFT spectra between Pt/TiO2(B) and Pt/anatase indicates their analogous morphology of Pt NPs,[51] which is consistent with the observation by HRTEM analyses. Furthermore, the degree of Pt dispersion could be calculated according to the sphere model,[52] and the detailed process is described in the Supporting Information. Based on this, the TOF values of catalysts can be obtained and those of Pt/TiO2(B) and Pt/anatase are 4.4 × 10–3 and 0.65 × 10–3 s–1, respectively.[53] The dramatically higher TOF value indicates that Pt/TiO2(B) possesses higher intrinsic activity in room-temperature HCHO oxidation than Pt/anatase, which suggests the presence of different active sites for HCHO oxidation between Pt/TiO2(B) and Pt/anatase.

Figure 5

CO-DRIFT spectra of Pt/TiO2(B) and Pt/anatase.

The analyses in chemical valence of Pt NPs of Pt/TiO2(B) and Pt/anatase are presented in Figure a. All the XPS Pt 4f spectra could be divided into three 4f7/2-4f5/2 doublets at binding energies of 71.1–74.4, 72.3–75.6, and 73.8–77.1 eV assigned to Pt0, Pt2+, and Pt4+ species, respectively.[54,55] The content of corresponding species can be estimated by the area of the resolved peak, which is used to identify the chemical valent state of Pt NPs of the two catalysts. The percentage contents of Pt0, Pt2+, and Pt4+ species of Pt/TiO2(B) are 10.3, 64.0, and 25.6%, respectively, while those of Pt/anatase are 37.7, 41.7, and 20.6%, respectively. Pt/TiO2(B) contains a higher content of oxidized Ptδ+ species than Pt/anatase does, especially for Pt2+.

Figure 6

XPS spectra for Pt 4f (a), O 1s (b), and Ti 2p (c) of Pt/TiO2(B) and Pt/anatase.

As seen in Figure , temperature-programmed reduction (TPR) patterns of both Pt/TiO2(B) and Pt/anatase display two reduction stages from room temperature to 150 °C and from 200 to 300 °C, which are ascribed to the reduction of oxidized Pt species and Pt-support boundary species, respectively.[19,56] According to the reduction temperature of various species provided by TPR patterns, it is confirmed that all oxidized Pt species of two catalysts can be completely reduced to metallic Pt species under the preparation conditions. Moreover, the reduction temperature of Pt/TiO2(B) is not dramatically different from those of Pt/anatase, which demonstrates that the interaction of Pt species with TiO2(B) is the same as that with anatase. Additionally, the XPS spectra of Ti 2p of the catalysts in Figure c show that the binding energies of Ti ions in both Pt/TiO2(B) and Pt/anatase are the same; that is, a similar state of oxygen vacancies around Pt species to the Pt/anatase should exist in the Pt/TiO2(B).[32] Thus, it is inferred that the presence of adsorbed species on Pt species may be the main reason for the different chemical valent states of Pt NPs between Pt/TiO2(B) and Pt/anatase, because the reduced catalysts inevitably come into contact with air and moisture during the transfer and preservation. It is required to comparatively analyze the state of adsorbed species on the surface of Pt/TiO2(B) and Pt/anatase.

Figure 7

H2 temperature-programmed reduction (H2-TPR) patterns of Pt/TiO2(B) and Pt/anatase.

Figure b shows the XPS O 1s spectra of Pt/TiO2(B) and Pt/anatase. Both of them have three deconvoluted peaks at binding energies of 529.7, 532.0, and 533.5 eV, which are attributed to the bulk lattice oxygen (OI), surface oxygen species (OII), and surface hydroxyl oxygen (OH), respectively.[48,57,58] The ratios of OII/OI to OH/OI are calculated by the peak areas of OI, OII, and OH, which are useful when analyzing the amount of hydroxyls and oxygen species on the two catalysts. The OH/OI ratios of Pt/TiO2(B) and Pt/anatase are 0.615 and 0.126, respectively. Moreover, the OII/OI ratio of Pt/TiO2(B) is 0.609, higher than that of Pt/anatase (0.424). Unsurprisingly, the amounts of hydroxyls and oxygen species on Pt/TiO2(B) and Pt/anatase are different. Because there are more hydroxyls and oxygen species on Pt/TiO2(B), the different chemical valent Pt species between Pt/TiO2(B) and Pt/anatase as mentioned above can be explained.

The amount of oxygen species is generally considered to be one of the key factors directly influencing the HCHO oxidation performance of catalysts.[20,50,59] The large amount of oxygen species must be one of the pieces of evidence of highly efficient HCHO oxidation on Pt/TiO2(B). As for the increment in the amount of oxygen species of Pt/TiO2(B) relative to Pt/anatase, it is unlikely to be related to their dispersion states of Pt NPs as they have the same morphological structure of Pt NPs according to the HRTEM and CO-DRIFTS analyses. Alternatively, the increasing content of oxygen species on Pt/TiO2(B) could be attributed to the possession of a large number of hydroxyls. With the hydroxyl-rich supports or in a high-humidity environment, the corresponding catalysts could achieve the markedly increased content of oxygen species.[28,31,60] It is well accepted that hydroxyls can facilitate the adsorption and diffusion of O2 to active sites and consequently improve the oxygen activation.[30,61]

In addition to assisting oxygen activation, the hydroxyls can enhance the adsorption ability of HCHO on the catalyst due to their good affinity to HCHO molecules by the formation of hydrogen bonds.[62,63] In addition, the hydroxyls themselves can participate in the HCHO oxidation process. Hong et al. found that under anaerobic conditions (where oxygen was absent), the reaction of HCHO oxidation can still be performed in the presence of moisture.[31] Additionally, He et al. discovered that hydroxyls could react with formate (an intermediate of HCHO oxidation) to generate the carbonate/bicarbonate, which significantly improved the HCHO oxidation efficiency of the catalyst.[6] Hence, there is no doubt that Pt/TiO2(B) with abundant hydroxyls possesses such excellent HCHO oxidation performance.

Mechanism Discussion

The oxidation of HCHO is successively experienced through the intermediate species of dioxymethylene (DOM), formate (HCOO–), and CO, and finally being oxidized into CO2.[37,64] According to data analyzed above, although a high overall reaction rate of HCHO oxidation can be obtained on Pt/TiO2(B), it is still unclear which step of HCHO oxidation is improved. The understanding of this question is crucial for exploring the structure–activity relationship of catalysts and their further development. Herein, in situ HCHO-DRIFTS was used to detect the intermediate species of HCHO oxidation on the surface of various catalysts to study the differences in the reaction mechanisms between Pt/TiO2(B) and Pt/anatase.

In the initial stage, a mixed gas flow of HCHO, H2O, and N2 is injected into the in situ chamber. As shown in Figure a, the bands at around 1361 and 1576 cm–1 are assigned to the symmetric stretching vibration and the asymmetric stretching vibration of formate.[37,65] Moreover, the symmetric vibration of the −CH2– group at around 2861 cm–1 corresponding to DOM can also be observed,[66] demonstrating that the HCHO decomposition process from HCHO to formate via DOM also occurs on Pt/TiO2(B), which accords with that on Pt/anatase (as seen in Figure b). It is noteworthy that the signal intensity of the CO band at 2073 cm–1 is very weak in various DRIFT spectra of Pt/TiO2(B), while the intensity of the CO signal on Pt/anatase is obviously strong. After N2 purging in the second stage and O2 purging in the third stage, no signal about other new species arises in the DRIFT spectra of Pt/TiO2(B) and Pt/anatase. Apparently, the appearance of the signal of all intermediate species in DRIFT spectra demonstrates that the oxidation of HCHO on Pt/TiO2(B) follows the reaction pathway of HCHO → DOM → HCOO– → CO → CO2, which is consistent with that of Pt/anatase.

Figure 8

Dynamic changes of in situ HCHO-DRIFT profiles of Pt/TiO2(B) (a) and Pt/anatase (b) at the atmospheres of HCHO + N2 + H2O, N2, and O2.

Subsequently, the state of hydroxyls of two catalysts is further analyzed by comparing the hydroxyl-related signals in HCHO-DRIFT spectra. The magnified DRIFT spectra in Figure display that the bending vibrations of adsorbed water molecules (δ(H2O)) on Pt/TiO2(B) appear at a wavenumber of 1623 cm–1, while the corresponding wavenumber of Pt/anatase is 1652 cm–1. The red shift in the wavenumber of δ(H2O) is attributed to the increased interaction of H2O molecules with the substrate.[25,68] This suggests that Pt/TiO2(B) could have better affinity to water molecules than Pt/anatase, which is beneficial to increasing the adsorption of water on Pt/TiO2(B), as also demonstrated by its obviously higher signal intensity of bending vibrations of δ(H2O). As mentioned above, H2O is liable to dissociate into hydroxyl on TiO2(B).[25] Hence, the characteristics of enhanced H2O adsorption and dissociation can endow Pt/TiO2(B) with a large number of hydroxyls, echoing with the above XPS analyses.

Figure 9

Partial enlarged image of the DRIFT spectra of Pt/TiO2(B) (a) and Pt/anatase (b).

Moreover, Figure shows that the DRIFT spectra of both Pt/TiO2(B) and Pt/anatase have the negative peaks in the range of 3600–3800 cm–1, which are assigned to the stretching vibration of hydroxyls. These negative peaks have been regarded as evidence of hydroxyl involvement in HCHO oxidation.[6,31,69] As clearly displayed in Figure b, there is one negative peak at 3669 cm–1 in this region of DRIFT spectra of Pt/anatase, corresponding to bridging hydroxyls.[68,70] This indicates that the bridging hydroxyl-involved HCHO oxidation mechanism is followed on Pt/anatase, in accordance with many previous reports.[6,13,15] Dissimilarly, the DRIFT spectra of Pt/TiO2(B) (as seen in Figure a) show two negative peaks at 3729 and 3664 cm–1, ascribed to terminally bound hydroxyls and bridging hydroxyls, respectively.[68,70] It is confirmed that two types of hydroxyls participate in the HCHO oxidation on Pt/TiO2(B). Moreover, the intensity of terminally bound hydroxyls is obviously stronger than that of bridging hydroxyls, suggesting the dominant role played by the terminally bound hydroxyls in the HCHO oxidation of Pt/TiO2(B).

To the best of our knowledge (see the survey results in Table S1), the reports about the bridging hydroxyl involvement in the HCHO oxidation account for the majority, while very rare reports record the terminally bound hydroxyls or both. The type of hydroxyl as a factor influencing the HCHO oxidation performance receives little attention. Several studies have shown that the bridging hydroxyls interacted with HCHO to form DOM, and subsequently the terminally bound hydroxyls promoted the formation of formate from DOM through the Cannizzaro mechanism, and the latter reaction was considered to be thermodynamically and kinetically favorable.[66,67] This indicates that the synergetic involvement of bridging hydroxyls and terminally bound hydroxyls in HCHO oxidation is essential for elevating the reaction rates of HCHO → DOM → HCOO–. That is to say, the first two steps of HCHO oxidation can be accelerated over Pt/TiO2(B).

Further comparative analysis is conducted on the signal intensity of the other two key intermediate species (about 1580 cm–1 for formate and 2075 cm–1 for CO) as a function of experimental time, as plotted in Figure . During the initial stage of HCHO + H2O + N2 injection, the signal ascribed to formate is gradually enhanced, indicating that the formate increasingly accumulates on the surface of the two catalysts. As reported previously, the formate decomposition was the slowest step in formaldehyde oxidation, so the signal of formate was the most obvious in HCHO-DRIFT spectra compared with other intermediate species, and its rate was considered to be the rate-limiting step of HCHO oxidation.[6,37] In the second and third stages, the signal intensity of formate of both catalysts decreases slightly, indicating difficulty in decomposing the formate under the condition of only oxygen but no hydroxyl.[31] Thus, hydroxyls are essential for the fast formate decomposition. By comparison, the signal intensity of formate of Pt/TiO2(B) is always weaker than that of Pt/anatase during the whole experimental process, suggesting that the formate is not easy to accumulate on Pt/TiO2(B). As mentioned above, the coparticipation of two hydroxyl groups in HCHO oxidation can accelerate the formation of formate. Therefore, the weak signal intensity of formate in DRIFT spectra of Pt/TiO2(B) could be caused by the rapid decomposition of formate rather than the slow generation.

Figure 10

DRIFTS signal intensities of formate (a) and CO (b) as a function of time on Pt/TiO2(B) and Pt/anatase.

Moreover, as seen in Figure b, the CO signal intensity of Pt/anatase increases slowly with the emergence of formate and then decreases after O2 is introduced. The slow increase in CO signal intensity demonstrates the difficult decomposition of formate into CO. Additionally, the CO signal intensity begins to decrease with N2 purge, corresponding to the oxidation of CO by oxygen species on Pt/anatase. This indicates that the further reaction of CO has actually proceeded from the first stage. After pure O2 injection, the CO signal intensity decreases to a low value. Different from Pt/anatase, the CO signal of Pt/TiO2(B) has insignificant intensity at the first stage and instantly disappears upon stopping the HCHO injection. Obviously, the consumption rate of CO on Pt/TiO2(B) significantly exceeds the generation rate of CO from formate. This could be attributed to the existence of more oxygen species on Pt/TiO2(B) as analyzed by the above XPS measurement. Moreover, the hydroxyls on Pt/TiO2(B) can also contribute to the transport of CO species due to the weakened interaction between CO and the hydroxylated surface.[71,72]

Based on the above analyses and discussion, the role of hydroxyls in HCHO oxidation of Pt/TiO2(B) is clarified. As shown in Figure , the presence of abundant hydroxyls is beneficial to oxygen activation for formation of more active species on Pt/TiO2(B). Moreover, two types of hydroxyls (bridging hydroxyl and terminally bound hydroxyl) are synergistically involved in the HCHO oxidation, which effectively improves the reactions of formaldehyde to DOM and DOM to formate with Pt as the active site. Additionally, hydroxyls also contribute to the fast decomposition of formate and consequently avoid the accumulation of excessive formate on Pt/TiO2(B). Finally, the formed CO is liable to be transported to active sites on the hydroxylated surface and rapidly oxidized by a large number of reactive oxygen species on the Pt. It is noteworthy that all the steps of HCHO oxidation are connected in series, and consequently the delay in any step will result in a decrease in the overall reaction rate of HCHO oxidation. By consecutively analyzing each HCHO oxidation step, it can be determined that the reaction rate of all steps of HCHO oxidation on Pt/TiO2(B) is accelerated, which explains the excellent catalytic performance of Pt/TiO2(B).

Figure 11

Scheme for HCHO oxidation on Pt/TiO2(B).

Conclusions

In the present work, a mesoporous TiO2(B)-supported Pt catalyst was prepared to first apply in the HCHO oxidation. The as-synthesized Pt/TiO2(B) catalyst not only exhibited superior HCHO oxidation performance to other crystalline TiO2-supported Pt catalysts, but also demonstrated excellent catalytic stability. The roles of hydroxyls in different steps of HCHO oxidation were analyzed successively, and it was found that all steps could be accelerated. The hydroxyls on Pt/TiO2(B) significantly contributed to improving the HCHO oxidation performance.

Experimental Section

Chemicals and Catalyst Preparation

Anhydrous potassium carbonate (K2CO3), anhydrous calcium chloride (CaCl2), and chloroplatinic acid (H2PtCl6·6H2O) were purchased from Sinopharm Chemical Reagent Co., Ltd. Nitric acid (HNO3) was purchased from Chemical Reagent Co., Ltd. Anatase TiO2 was purchased from Jingrui New Material Co., Ltd. Rutile TiO2 was purchased from Nano New Material Technology Co., Ltd. P25 was purchased from Evonik Degussa (China) Co., Ltd. TiO2·nH2O was provided by Jiangsu Taibai Group Co., Ltd. All the chemicals were used as received.

Mesoporous TiO2(B) was synthesized according to the procedures as described in our previous work.[26,73] Briefly, potassium titanate (K2Ti2O5) was preliminarily prepared by calcination from a mixture of K2CO3 and TiO2·nH2O in a Ti/K molar ratio of 0.95 at 860 °C for 4 h. Then, the K2Ti2O5 powders were separately placed in an autoclave filled with CaCl2 saturated solution for being treated at 180 °C. After 24 h, the powders were ion-exchanged to remove potassium ions by HNO3 aqueous solution (pH = 2). The collected powders were calcined at 500 °C for 2 h to achieve the mesoporous TiO2(B). The content of potassium residual was below 0.2 wt % according to previous work,[22] most of which were distributed in the subsurface of mesoporous TiO2(B) due to no detection on potassium by XPS analyses.

Loading of Pt NPs (0.5 wt %) on mesoporous TiO2(B) was carried out by wet impregnation using H2PtCl6 as the precursor. The impregnated powders were calcined in air at 500 °C for 2 h and subsequently reduced in H2 at 300 °C for 2 h to obtain the Pt/TiO2(B) catalyst. For comparison, anatase TiO2, rutile TiO2, and P25 were used as control samples and their corresponding catalysts were denoted as Pt/anatase, Pt/rutile, and Pt/P25, respectively.

Characterization

The Raman spectrum of TiO2(B) was provided by a Horiba HR 800 spectrometer equipped with a CCD camera detector with the wavelength of excitation light of λ = 514 nm. The signal was sampled in the range from 100 to 800 cm–1. Nitrogen adsorption–desorption isotherms at −196 °C of various catalysts were measured on an ASAP2020 physisorption analyzer (Micromeritics, USA). The surface area (SBET) of catalysts was calculated using the BET model, and the pore volume (VP) was obtained by nitrogen adsorption with a relative pressure of 0.99. Pore size distribution was determined by the Barrett–Joyner–Halenda model. Nanoscale morphology of catalysts was provided by field-emission scanning electron microscopy (FESEM, JEM-7600 F) and high-resolution transmission electron microscopy (HRTEM, JEM-2010). X-ray photoelectron spectroscopy (XPS, ESCALAB 250) equipped with a monochromatic Al Kα X-ray beam was utilized to study the chemical state of various elements on the surface of the catalyst. The binding energies of elements were referenced to the C 1s peak (284.6 eV) of adventitious hydrocarbons. H2-TPR experiments were carried out in a TP-5000 micro fixed-bed flow system. Before the test, the catalyst powders (0.1 g) were pretreated in a flow of air (30 mL·min–1) at 500 °C for 0.5 h followed by cooling to room temperature. The atmosphere of the system was switched to a mixed gas of 10% H2/N2. After stable flow for 0.5 h, the system started to be heated to 500 °C at a heating rate of 10 °C·min–1. The H2 consumption during the test was monitored by the TCD detector.

In situ diffuse reflectance infrared Fourier transformed spectroscopy (in situ DRIFTS) was performed on a Nicolet-6700 spectroscope apparatus (Thermo Electron) equipped with a diffuse reflectance accessory and an MCT detector. Herein, the DRIFTS test contained HCHO-DRIFTS and CO-DRIFTS depending on different adsorbents. The operational processes of HCHO-DRIFTS were taken to investigate the reaction path of HCHO oxidation over various catalysts. The as-prepared catalyst powders were placed in the in situ chamber and pretreated at 300 °C for 0.5 h in the atmosphere of nitrogen with a flowing rate of 30 mL·min–1. After being cooled down to 30 °C, the chamber was introduced with a mixed gas stream of HCHO + H2O + N2, which was produced by bubbling. After 1 h, pure N2 and pure O2 passed through the chamber sequentially to study the consumption and oxidation of different intermediates. The spectra under reaction conditions were recorded after 32 scans with a resolution of 4 cm–1. Moreover, CO-DRIFTS was used to analyze the dispersion state of Pt NPs. The catalyst was in situ reduced in 10% H2/N2 at 300 °C for 0.5 h and cooled down to 30 °C in an N2 gas flow. Then, CO was introduced into the chamber. After 0.5 h, the excessive and weakly adsorbed CO was removed by being purged with a flowing pure N2 stream. The spectra from 1900 to 2200 cm–1 were recorded after 32 scans with a resolution of 4 cm–1.

Catalytic Evaluation

The HCHO oxidation reaction was carried out in a fixed-bed flow reactor. Catalyst powders were added in a quartz tube with an inner diameter of 6 mm. The feed gas consisted of 180 ppm HCHO, H2O vapor (relative humidity of 35%), 20 vol % O2, and balance N2, produced by passing air stream through a bubbler containing a 37% HCHO aqueous solution in bath at 4 °C. The gas hourly space velocity for each experiment was 54,000 h–1. During the reaction, the reaction temperature of HCHO oxidation was controlled between 30 and 140 °C by a temperature-programmed furnace. The HCHO concentrations of feed gas and exhaust gas were analyzed online by a gas chromatograph, which were synchronously checked by the CO2 detector. No CO was detected in exhaust gas by the CO detector. The sample was analyzed after at least 30 min of stable reaction. HCHO conversion of various catalysts was calculated according to the change of HCHO concentration in the feed gas and exhaust gas.