Distinct Role of Surface Hydroxyls in Single-Atom Pt1/CeO2 Catalyst for Room-Temperature Formaldehyde Oxidation: Acid-Base Versus Redox.

The development of highly efficient catalysts for room-temperature formaldehyde (HCHO) oxidation is of great interest for indoor air purification. In this work, it was found that the single-atom Pt1/CeO2 catalyst exhibits a remarkable activity with complete removal of HCHO even at 288 K. Combining density functional theory calculations and in situ DRIFTS experiments, it was revealed that the active OlatticeH site generated on CeO2 in the vicinity of Pt2+ via steam treatment plays a key role in the oxidation of HCHO to formate and its further oxidation to CO2. Such involvement of hydroxyls is fundamentally different from that of cofeeding water which dissociates on metal oxide and catalyzes the acid-base-related chemistry. This study provides an important implication for the design and synthesis of supported Pt catalysts with atom efficiency for a very important practical application-room-temperature HCHO oxidation.

Introduction

Formaldehyde (HCHO) is one of the main sources of hazardous indoor air pollution, which is generally emitted from building and furnishing materials. Long-term exposure to HCHO may cause adverse health effects on humans such as skin irritation, eczema, asthma, and vomiting, even at low concentrations.[1,2] In 2006, HCHO has been classified as a major indoor air pollutant and group I carcinogen, as specified by IARC.[3] Therefore, effective removal of HCHO from indoor air is imperative to improve indoor air quality and safeguard human health. Various techniques for HCHO removal have been reported, including physical adsorption,[4] plasma technology,[5] photocatalytic oxidation,[6,7] and catalytic oxidation.[8−10] Compared with other HCHO purification technologies, catalytic oxidation is more desired as it can efficiently transfer HCHO to CO2 and H2O without any harmful byproducts and secondary pollutants.[11]

The catalysts for room-temperature HCHO oxidation are mainly transition-metal oxides (Co3O4, TiO2, CeO2, and MnO2)[11,12] and supported noble metals (Pt, Pd, Rh, and Au).[9,13] The latter ones, especially Pt-based, exhibit superior catalytic activity than non-noble metal oxide catalysts at relatively low temperatures.[14−16] The valence state of Pt in supported Pt-based catalysts has a significant effect on HCHO oxidation activity.[9] Until now, previous studies verified that metallic Pt was more reactive than oxidized Pt in HCHO oxidation at room temperature.[17−20] Single-atom catalysts (SACs) have received much attention recently because they allow the maximum utilization efficiency of the precious metal atoms for surface reactions. Many SACs (Pt1/FeOx,[21] Pt1/θ-Al2O3,[21,22] Pd1/La–Al2O3,[23] Pt1/CeO2,[24] Ir1/FeOx,[25] and Rh1/ZrO2[26]) have been reported for CO oxidation,[21−24] water-gas shift reaction,[25] and the direct oxidation of methane.[26] However, there are few studies on SACs for catalyzing the oxidation of HCHO at room temperature. Tang et al. found that Ag/K/Na/Rb atoms anchored in the tunnel openings of hollandite-type manganese oxide can completely oxidize HCHO at 393 K.[27−31] Jia et al. found that single-atom Au supported on α-MnO and CeO2 supports can also achieve the complete oxidation of HCHO at 353 K.[32,33] Later, they further explored the addition of Mn to modify the single-atom Pt1/TiO2 catalyst for the catalytic oxidation of HCHO at room temperature[34] and reported that atomic-level dispersed Pt/TiO2 catalysts prepared by adding alkali metals can also catalyze the oxidation of HCHO at room temperature.[16] In addition, Sun et al. prepared SACs by decreasing Pt loading for use in HCHO oxidation.[35] However, over these reported Na–Pt1/TiO2,[16] Mn–Pt1/TiO2,[34] and Pt1/TiO2[35] catalysts, hydrogen reduction is required before HCHO oxidation at room temperature, leading to the reduction of Pt cations and agglomeration of metallic Pt. So far, there is no report on atomically dispersed Pt2+ catalyst with well-defined structure for HCHO oxidation at room temperature.

The conversion of HCHO usually involves the initial formation of dioxymethylene (DOM, H2CO2) and the subsequent formation of formate species (HCOO), before its complete oxidation to CO2, which is generally regarded as the kinetic relevant step.[34−36] The surface hydroxyl (OH) groups on the catalyst surface formed from cofeeding water were found to be essential for achieving high performance for the HCHO oxidation,[16,37] affecting all steps involved in HCHO conversion. The surface hydroxyls on varied catalysts have been reported to either act as an adsorption site[38−42] or promote the conversion of DOM into formate,[43] as well as the direct oxidation of HCOO to CO2.[16,44,45] Chen et al.[43] reported a facile pathway for DOM conversion to HCOO catalyzed by surface OH from water dissociation over a TiO2-based catalyst, where the bridging OH groups are involved in DOM formation, while the terminal OH groups are responsible for the transformation of DOM into HCOO and its further oxidation at room temperature. In their previous work, both O2 and H2O were found necessary for the reaction,[13,43,46] and no activity was observed in the absence of H2O, due to the lack of active surface OH groups that oxidize HCOO through the direct oxidation mechanism (HCOO + OH → H2O + CO2).[16,44] However, the recycle and fate of the H atom from water dissociation are not well considered in the mechanism, which was then assumed to be consumed by O2 to avoid the accumulated H poisoning. On the other hand, water dissociation at room temperature is generally regarded as a process that requires the close proximity of acidic and basic sites, where an acidic site binds the molecule and a basic site abstracts a proton.[47] Accordingly, reducible oxide or metal-reducible oxide support interface (MSI) that is known to dissociate water would result in two distinct OH groups: a terminal hydroxyl, that is, a coordinatively unsaturated hydroxyl (CUS-OH), and a bridge hydroxyl or multicoordinated hydroxyl.[48,49] The “terminal” one resides at an acidic site (e.g., a surface CUS cation site like Ce4+[49] and Ti4+ site[47,48]) or oxygen vacancy, which would most likely possess the oxygen atom from the predissociated water molecule, thus behaving more like basic OH–. Accordingly, the other “bridging” hydroxyl resides at a multiple-coordinate or “bridging” O-site (a basic site), which might form by abstraction of a proton from the adsorbed water molecule, probably being more acidic. Hence, the terminal and bridge hydroxyls from water dissociation on the reducible oxide or at the metal-oxide interface are still more related with the acid–base chemistry, rather than the redox one. In principle, these acid or base surface hydroxyls cannot be used alone as an active site for low-temperature oxidation such as the HCHO removal mentioned above, unless the lattice oxygen is activated through the formation of hydroxyl groups as revealed in the FeO(111)–Pt(111) interface[50] or steam-treated Pt1/CeO2.[24] Therefore, when dealing with HCHO reactions involved with surface hydroxyls on metal oxides or MSI, one deals with two main concepts: (i) oxidation–reduction; and (ii) acid–base reaction, which can be related but inherently different. However, such distinct property and catalytic role of the varied surface OH have not been systematically evaluated or well clarified in the previous work on HCHO conversion, to the best of our knowledge.

Recently, Nie et al. reported that steam treatment on single-atom Pt1/CeO2 catalyst significantly enhances the catalytic activity for CO oxidation,[24] thanks to the generation of new surface hydroxyl sites, that is, the activated OlatticeH sites in the vicinity of Pt. Unlike the SACs mentioned earlier,[16,34,35] this catalyst neither requires pre-reduction in H2 to activate the catalyst nor needs co-feeding water to achieve high performance. Hence, it would be worth exploring whether this atomically dispersed Pt2+ catalyst with a well-defined structure could also exhibit excellent catalytic activity to fulfill HCHO oxidation at low temperatures. Herein, in this work, we employed single-atom Pt1/CeO2 catalysts with and without steam treatment and systematically evaluated their distinct catalytic performance for HCHO oxidation. Various characterization techniques such as CO temperature-programmed reduction (CO-TPR), in situ DRIFTS, and density functional theory (DFT) calculation were employed to reveal the relation between the surface properties and the performance of catalysts and to unravel the reaction pathways at room temperature. Moreover, the inherent property of each elemental step (i.e., acid–base vs redox) was clarified, and the fundamental difference between the varied surface OH groups as well as their catalytic roles was unambiguously distinguished.

Results

Structure Characterization

Pt1/CeO2 and Pt1/CeO2–S catalysts were synthesized as previously reported.[24] The Pt content measured in these two catalysts is 0.61 wt %. The Brunauer–Emmett–Teller surface area and the textural structure of bare CeO2, CeO2–S, Pt1/CeO2, and Pt1/CeO2–S catalysts are listed in Table S1. The surface area of Pt1/CeO2–S is equivalent to that of Pt1/CeO2. However, the surface area and pore volume of CeO2–S are much smaller than that of both Pt1/CeO2 and Pt1/CeO2–S, consistent with the earlier reports that the presence of Pt can help stabilize the crystal structure and preserve the surface area of CeO2 under severe conditions.[51,52] The X-ray diffraction (XRD) patterns of Pt1/CeO2 and Pt1/CeO2–S (Figure S1) exhibit peaks only belonging to the ceria structure (JCPDS no. 65-5923), with no peaks for Pt detected. X-ray photoelectron spectroscopy (XPS) demonstrated that Pt remained in ionic form (Pt2+) over both catalysts (Figure S2). Raman spectroscopy (Figure S3) revealed the formation of the Pt–O–Ce bond, as evidenced by the featured peaks at 550 cm–1 (the one at 657 cm–1 attributed to Pt–O). CO adsorption by in situ DRIFTS analysis (Figure S4) revealed that only ionic Pt2+ was present, as confirmed by the typical bands at 2094 and 2090 cm–1.[52−55] The CO spectra did not show lower wavenumber features (<2000 cm–1) either, which are assigned to the bridge or threefold adsorbed CO on larger metal Pt nanoparticles.[56] The combination of the characterization above suggests that atomic dispersion of Pt2+ was successfully realized on both Pt1/CeO2 and Pt1/CeO2–S and that the steam treatment can maintain the structure of SACs, in agreement with the previous report.[24]

The oxygen and surface hydroxyl species of the catalysts were further characterized by CO-TPR analysis (Figure ). H2 and CO2 may be produced simultaneously from the water–gas shift due to the interaction of CO with the surface hydroxyl groups in the catalysts,[57−59] while CO2 is only emitted through the reduction of surface lattice oxygen from Pt–O–Ce and CeO2.

Figure 1

CO-TPR profiles of Pt1/CeO2 and Pt1/CeO2–S catalysts.

As for Pt1/CeO2, the reduction peak of the surface lattice oxygen in the vicinity of Pt (Pt–O–Ce bond) is centered at 397 K, while the surface OH groups start reacting with CO to produce CO2 and H2 at about 493 K (peaked at 608 K). Instead, the peak of the surface lattice oxygen is shifted down to 360 K on the Pt1/CeO2–S catalyst. Meanwhile, more surface OH groups were generated, which can react with CO at about 453 K (peaked at 553 K). The steam treatment significantly activated both surface oxygen and hydroxyl species, which should play a key role to fulfill the catalytic oxidation of HCHO at low temperatures.

Catalytic Performance for HCHO Oxidation

The light-off curves of HCHO oxidation were measured over various catalysts with a weight-hourly space velocity (WHSV) of 100,000 mL·g–1·h–1 to evaluate their distinct catalytic performance. Bare CeO2–S and CeO2 have a high onset temperature around 373 K (Figure S5) and achieve only 24 and 52% of HCHO conversion, respectively, at 473 K. The steam treatment of the CeO2 support results in rather decreased reactivity, probably due to the collapsed pore structure and the subsequently reduced surface area (from 80 to 12 m2·g–1, Table S1). The Pt1/CeO2 catalyst exhibits similar reactivity in HCHO oxidation (Figure a), that is, the onset temperature remains at 393 K and complete formaldehyde oxidation occurs at 495 K. However, the steam treatment counterpart, that is, Pt1/CeO2–S exhibits a striking low-temperature reactivity. 100% conversion of HCHO is achieved at 288 K. Even at a very low temperature, such as 278 K, the catalyst is still active enough to reach 46% conversion. Previous studies have shown that SACs can achieve good performance in catalyzing the oxidation of HCHO at room temperature only after being reduced with hydrogen at a certain temperature (Table S2),[16,34,35] in which the contribution from metallic Pt nanoparticles cannot be excluded.[19,20] Interestingly, the reduced catalyst of Pt/CeO2–R does not exhibit similar room-temperature catalytic activity, with T90 (temperature required to achieve 90% conversion) shifted to 393 K. The reaction rate (noted as r value) and the turnover frequency (TOF) in HCHO oxidation were calculated and are listed in Table S2. The reaction rate of Pt1/CeO2–S is 290 μmol·gPt–1·s–1, and the TOF is 0.056 s–1 at 298 K, which are much higher than those of SACs and metallic Pt-based catalysts reported in the literature (Table S2). It is worth noting that this is the first time an atomically dispersed Pt catalyst is reported, without the requirement of hydrogen reduction being highly active in HCHO oxidation at room temperatures.

Figure 2

(a) HCHO conversion over Pt1/CeO2 and Pt1/CeO2–S catalysts as a function of temperature. Reaction conditions: 400 ppm HCHO, 20 vol % O2, and N2 as balance gas; total flow rate: 100 mL·min–1 and WHSV: 100,000 mL·g–1·h–1. (b) Relative humidity (RH) effect on the activity of Pt1/CeO2–S catalyst at 298 K. Reaction conditions: 400 ppm HCHO, 20 vol % O2, and N2 as balance gas; total flow rate: 100 mL·min–1 and WHSV: 222,000 mL·g–1·h–1. (c) HCHO conversion as a function of time-on-stream on the Pt1/CeO2–S catalyst at 298 K. Reaction conditions: 400 ppm HCHO, 20 vol % O2, and N2 as balance gas; RH: 25%, total flow rate: 100 mL·min–1, and WHSV: 240,000 mL·g–1·h–1. (d) Arrhenius plots for HCHO conversion over Pt1/CeO2 and Pt1/CeO2–S catalysts. Reaction conditions: 1400 ppm HCHO, 20 vol % O2, and balance N2; RH: 25% and total flow rate: 100 mL·min–1. Conversions in all the tests involved were kept below 15%.

It should be noted that the superior reactivity of Pt1/CeO2–S exhibited in Figure a is realized in the absence of cofeeding water, which has been reported indispensable for the room-temperature reactivity observed for the TiO2-based[13,43] or alkaline-doped noble metal catalysts.[16,44] Nevertheless, in practical application, the catalysts should be adaptable in both dry and humid environments, so it is of significance to investigate the effect of RH on HCHO catalytic oxidation at room temperature. Under 25% RH (Figure b), the reactivity of Pt1/CeO2–S is 60% higher than that under dry conditions. The higher moisture content with RH at 50 and 75% could still promote the activity but to a less extent (about 20%). Extensive research has shown the improved catalytic oxidation of HCHO by cofeeding water.[36,38,42,60] The presence of H2O significantly improves the DOM transformation into HCOO[43] or promotes the adsorption of HCHO and converts it into formate.[61] The promotional effect must be related to the generation of surface OH from water dissociation,[36,38,42,60] which will be discussed in detail later. However, a further increase in the RH results in lower reactivity, probably due to the competitive adsorption between HCHO and H2O molecules on the catalyst.[11,13,61,62] The above results indicate that the presence of a suitable amount of water vapor is greatly beneficial for the oxidation of HCHO at room temperature, and the Pt1/CeO2–S catalyst has good tolerance to a high concentration of moisture. However, under 25% RH (Figure S6), the reactivity of Pt1/CeO2 is lower than that under dry conditions, which may be due to the fewer active sites of the Pt1/CeO2 catalyst and the competitive adsorption between HCHO and H2O molecules on the catalyst. Moreover, the Pt1/CeO2–S catalyst demonstrated excellent stability, which shows stable conversion (55%) during a time-on-stream run of 60 h at room temperature in the presence of 400 ppm HCHO (Figure c), and the SEM results further confirm that there is no significant change in morphology (Figure S7).

Reaction Mechanism of Catalytic HCHO Oxidation

HCHO temperature-programmed surface reaction (HCHO-TPSR) experiments were performed to evaluate HCHO oxidation as a function of temperature (298–573 K) on Pt1/CeO2 and Pt1/CeO2–S, respectively. Over the Pt1/CeO2 catalyst, it is not until 423 K that CO2 production from the oxidation of surface HCHO starts to occur (Figure S8), accompanied by oxygen consumption in the same temperature range of 423–503 K (Figure S9). A small amount of CO and H2 was also detected, probably due to the direct dehydrogenation route (HCHO → H2 + CO). By contrast, the production of CO2 took place even at room temperature over Pt1/CeO2–S catalyst, and no obvious CO and H2 were observed over the entire temperature range studied (Figure S8). However, on the Pt1/CeO2–S catalyst, the CO2 signal is much lower than that on the Pt1/CeO2 catalyst. This is because most of HCHO is oxidized at room temperature, before the introduced O2/He reaches a stable baseline. The drastic enhancement of reactivity brought by steam treatment is in good agreement with the earlier activity tests shown in Figure a. To further evaluate the new HCHO oxidation pathway, we carried out kinetic studies on both Pt1/CeO2 and Pt1/CeO2–S catalysts (Figure d). The apparent activation energy (Ea) of the reaction on the Pt1/CeO2 catalyst is 61 ± 3 kJ·mol–1, while an apparent activation energy of about 14 ± 2 kJ·mol–1 is obtained over the Pt1/CeO2–S catalyst, thus providing further evidence that the reaction pathway was changed on the steam treatment catalysts.

The reaction mechanism of HCHO oxidation on Pt1/CeO2 and Pt1/CeO2–S catalysts was then investigated using in situ HCHO-DRIFTS. With the introduction of HCHO/N2 at 303 K, there was no peak associated with molecularly adsorbed HCHO (1722 cm–1)[13] on the Pt1/CeO2 catalyst (Figure S11a). While many other complex peaks appear, which can be assigned as listed in Table S3, the structures of the intermediates of the key elementary steps in the oxidation of HCHO are shown in Figure S10. On the Pt1/CeO2 catalyst, HCOO and DOM species are formed and are dominant after the catalyst was exposed to a flow of HCHO/N2 for 30 min at 303 K (Figure S11a). However, when the O2/N2 gas is introduced to the system, the HCOO and DOM bands do not induce any change. This indicates that HCHO transforms to HCOO and DOM at room temperature but cannot be further reacted. This is consistent with the reaction activity that HCHO reacts at a higher temperature on the Pt1/CeO2 catalyst (Figure a). Hence, we set the HCHO-DRIFTS experiment at 453 K to further observe the changes in the intermediate species (Figure S11c,d). It should be noted that with the continuous introduction of HCHO/N2, HCOO and CO are produced. After the catalyst was exposed to a flow of O2/N2 (Figure S11d), the vibration band of HCOO and CO decreased significantly. Therefore, we conclude that on the Pt1/CeO2 catalyst, the oxidation of HCHO only occurs at higher temperature (>423 K), under which DOM can readily be converted to HCOO or CO intermediates and subsequently fully oxidized to CO2, as is consistent with the HCHO-TPSR result (Figures S8 and S9).

The same in situ HCHO-DRIFTS experiments were also performed for the Pt1/CeO2–S catalyst. After the catalyst was exposed to a flow of HCHO/N2 (for 30 min) at 303 K, a variety of notable bands appeared as shown in Figure (thin line), mainly corresponding to the related DOM and HCOO intermediates.[58,63−65] It should be noted that with the HCHO adsorption after 10 min, the obvious peaks of HCOO appear and increase rapidly with time. Meanwhile, the peak at 3660 cm–1 that can be attributed to the surface hydroxyl species[49] gradually becomes more negative, indicating that its consumption is accompanied by the transformation of HCHO to HCOO. Moreover, there is much less DOM compared to that of Pt1/CeO2 under the same conditions. The introduction of O2/N2 does not cause a change in DOM peaks (Figure S12). Instead, the vibration bands of HCOO decrease significantly, concomitant with a notable increase in CO2. Accordingly, H2O species formed during oxidation gradually increase with O2 purging, as evidenced by a broad absorption band ranging from 3550 to 3000 cm–1 (the band at 1658 cm–1 was due to water adsorbed on the catalyst).[13,66] This confirms that the surface hydroxyl facilitates the formation of HCOO intermediates, which can be further oxidized to CO2 on Pt1/CeO2–S at room temperature. On the other hand, even in the flowing HCHO/O2/N2 for 30 min (Figure thick line), the bands of HCOO species do not disappear but just diminish in intensity compared with that in the flow of HCHO/N2. This suggests that another kind of reaction site exists on Pt1/CeO2–S, where formate species can form but are never consumed, hence actually acting as the spectator during the HCHO conversion. To distinguish the property of distinct surface sites, we also conducted the in situ HCHO-DRIFTS experiments on the CeO2 and CeO2–S catalysts and observed the HCHO absorption at 303 K. It should be mentioned that neither of these support oxides is active to convert HCHO to CO2 at room temperature, similar to Pt1/CeO2. However, HCHO adsorption on both CeO2 catalysts rapidly formed DOM and HCOO species (Figure S13a,b), while less DOM appeared on the surface of CeO2–S. The results above on CeO2 and CeO2–S suggest that the surface lattice oxygen of bare reducible metal oxides can oxidize HCHO into HCOO but is not active enough to proceed further to produce CO2.

Figure 3

In situ HCHO-DRIFTS of the Pt1/CeO2–S catalyst as a function of time in a flow of HCHO/N2 (thin line) or HCHO/O2/N2 (thick line) at 303 K (HCHO/N2 reaction conditions: 400 ppm HCHO and N2 as balance gas; total flow rate: 100 mL·min–1; the HCHO/O2/N2 reaction condition: 400 ppm HCHO, 20 vol % O2, and N2 as balance gas; total flow rate: 100 mL·min–1).

Discussion

Property of Elemental Reactions over HCHO Conversion

To fully understand what happened over these varied catalysts, it is necessary to shed light on the intrinsic properties of these elementary steps toward HCHO conversion. In principle, the transformation from HCHO to DOM is redox-neutral, as the formal oxidation state of C remains at 0 during the transformation. Hence, it does not need the involvement of oxidative sites but is assisted by the acid–base pair. More specifically, the O atom of HCHO can be attached to the surface coordinated CUS cation sites as Lewis acid, while the C atom can react with the lattice O or surface OH on CeO2 or Pt1/CeO2 as base sites, resulting in the formation of DOM. This is why DOM can readily form on all the catalysts employed in the work because plenty of such acid–base pair sites are present on their surfaces. In contrast, the oxidation process does take place when HCHO/DOM is converted to the HCOO intermediate and final product CO2, where the oxidation state of C increased to +2 and +4, respectively. The results above on CeO2 and CeO2–S suggest that the surface lattice oxygen of bare reducible metal oxides can oxidize HCHO into HCOO but is not active enough to proceed further to produce CO2 at room temperature. In other words, the rate control step over HCHO conversion should be the oxidation of HCOO to CO2, rather than the initial formation of DOM or HCOO, as also consistent with the earlier reports.[34−36] Atomically dispersed Pt2+ on CeO2 without steam treatment did not further activate the surface lattice oxygen, thus lacking the ability to further oxidize HCOO species either. The striking performance over Pt1/CeO2–S is due to its strong oxidative reactivity, which is active enough to oxidize HCOO to CO2 at room temperature. In the previous work,[24] it has been revealed that steam treatment over Pt1/CeO2 can activate the surface lattice oxygen, generating two neighboring active hydroxyls in the vicinity of Pt, which is highly active to perform CO oxidation at low temperature. We speculated that the active lattice OH on the SAC after steam treatment also promotes the oxidation of DOM and the HCOO intermediate, thus enabling the full oxidation of HCHO even below room temperature. Accordingly, the spectator sites on Pt1/CeO2–S mentioned above should be related with the surface lattice oxygen or hydroxyl located away from the atomically dispersed Pt2+, the reactivity of which should be similar to those on CeO2 and CeO2–S. To better understand the nature of this new pathway of HCHO oxidation by active surface hydroxyls on the atomically dispersed Pt1/CeO2–S catalyst, DFT calculations were conducted.

Density Functional Theory Calculations

The detailed structure of the active center on the atomically dispersed Pt1/CeO2 surface can be seen in the previous work,[24] where H2O molecules can fill the oxygen vacancy (VO) surface, generating two neighboring active hydroxyls (OlatticeH) in the vicinity of Pt12+. Such a structure is also employed to interpret the mechanism of HCHO oxidation in this work. As shown in Figure , HCHO oxidation starts with the adsorption of HCHO at the Pt1 site over the Pt1/CeO2(111)–2OH surface (Note S1 and Table S4), forming the monodentate configuration with an adsorption energy of −46.6 kJ·mol–1. The adsorbed HCHO reacts with the neighboring bridging OlatticeH group, forming H2COOlatticeH*. Our DFT calculation shows that this step is highly exothermic (−223.8 kJ·mol–1) with a small activation barrier of 12.5 kJ·mol–1. The formed H2COOlatticeH* deprotonates with the assistance of surface oxygen generating H2COOlattice* and another OlatticeH group. It is found that this deprotonation step is slightly exothermic (−17.4 kJ·mol–1) with a small activation barrier of 18.5 kJ·mol–1. Formate (HCOOlattice) is then formed via the C–H bond breaking of H2COOlattice* with an activation barrier of 62.2 kJ·mol–1. Next, an oxygen molecule (O2*) is physically adsorbed at the surface vacancy site. O2* reacts with HCOOlattice*, forming OOH*, group and releases CO2 into the gas phase. The calculated activation barrier for CO2 formation is 107.8 kJ·mol–1 with a reaction energy of −207.7 kJ·mol–1, indicating that this step is the kinetically rate-determining step, although it is thermodynamically favorable. Finally, OOH* reacts with a surface OlatticeH group and forms H2O in the gas phase and completes the entire catalytic HCHO oxidation cycle. All the structures of reaction intermediates and transition states in HCHO oxidation over the Pt1/CeO2(111)–2OH surface are given in Figure .

Figure 4

Calculated energy profiles of HCHO oxidation on Pt1/CeO2(111)–2OH surface; the structures of the intermediates and transition states (TSs) of the key elementary steps are shown in the cycle. (Blue, yellow, red, gray, and white circles denote Pt, Ce, O, C, and H atoms, respectively).

For comparison, HCHO oxidation over Pt1/CeO2 without steam treatment was also calculated, which is completely distinct from that on Pt1/CeO2–S due to the lack of active surface hydroxyls. As discussed in the earlier HCHO-DRIFTS and HCHO-TPSR experiments (Figures S8, S9, and S11c,d), the oxidation of HCHO only occurs at higher temperature (>423 K) on the Pt1/CeO2 catalyst, where DOM can be readily converted to HCOO or CO before complete oxidation to CO2. On the Pt1/CeO2(111) surface, HCHO is easily adsorbed on the Pt1 site to form DOM with an adsorption energy of −30.0 kJ·mol–1 (Figure S14). The adsorbed HCHO reacts with the neighboring active oxygen, forming H2COOlattice (DOM). The DFT calculation shows that this step is highly exothermic (−177.2 kJ·mol–1) with a small activation barrier of 0.4 kJ·mol–1. Next, DOM can be decomposed to HCO and HOs, and further dehydrogenation of HCO leads to the formation of CO and H2O. Subsequently, CO can be oxidized to CO2 by the adsorbed oxygen molecule (O2*) at the surface vacancy site. The calculation shows that the decomposition of DOM is the rate-control step, which requires quite a high energy barrier of 226.3 kJ·mol–1. This also explains why the Pt1/CeO2 catalyst needs a high temperature (>423 K) to fully oxidize HCHO, rather than room temperature.

Roles of Distinct Surface Hydroxyls

From the above calculations, the active OlatticeH, which is essentially the activated lattice oxygen by the steam treatment, plays a key role to realize the low-temperature oxidation of HCHO. It first reacts with the adsorbed HCHO to generate formate species, leaving behind one oxygen vacancy that activates the oxygen molecule to enable its full oxidation to CO2. The mechanism involved in this work has nothing to do with the nanoparticles of noble metals that dissociate O2 to oxidize the adsorbed HCHO and the intermediates.[13,15] Moreover, these “oxidative” hydroxyl groups, being highly active and regenerated during the reaction, are fundamentally different from that previously reported on the surface OH on TiO2- or CeO2-based noble metal catalysts or formed by the dissociation of cofeeding water.[67−71] It has been reported that Ir/TiO2 and Pt/TiO2[16,44] are highly active in the room-temperature HCHO oxidation only under the co-presence of both O2 and H2O, while there was no activity in the absence of H2O, due to the lack of the dissociation source to generate active surface OH groups. In our case, the active OlatticeH can be regenerated with the input of oxygen, and the reaction cycle can be completed without the need of cofeeding water, though a suitable amount of water can significantly enhance the reactivity in the room-temperature HCHO oxidation.

Furthermore, we speculate that this active OlatticeH on Pt1/CeO2–S is also intrinsically distinct from the surface OH formed through water dissociation on CeO2, CeO2–S, or Pt1/CeO2. As mentioned earlier, Chen et al.[43] reported a facile pathway for DOM conversion to HCOO, catalyzed by the surface hydroxyl on TiO2-based catalyst, where the bridging OH groups are involved in DOM formation, while the terminal OH groups are responsible for the transformation of DOM into formate species and its further oxidation. This is not the case in our work. Indeed, it is well known that the reducible oxides or MSIs are active to dissociate water and generate surface hydroxyls on the surface at room temperature, which is generally regarded as a process that requires the proximity of acidic and basic sites.[47] Hence, the “terminal” hydroxyl from the heterolytic water dissociation usually resides at an acidic site[72] (e.g., a surface CUS cation site like Ce4+ and Ti4+ site) or oxygen vacancy and most likely possesses the oxygen atom from the predissociated water molecule, thus behaving more like basic OH–1. Accordingly, the other “lattice” or “bridging” hydroxyl resides at a multiple coordinate lattice or “bridging” O site (a basic site), which might form by the abstraction of a proton from the adsorbed water molecule, probably being more acidic. The terminal OH generated on the reducible oxide or at MSI mainly functions as a basic site, rather than the oxidative catalytic center. Hence, it may facilitate the formation of formate from DOM through the Cannizzaro mechanism,[64] as discussed next, but should not be able to catalyze the further oxidation of formate to CO2. That is why, no matter how long the cofeeding water lasts, and how much more of such surface hydroxyl is generated with steam treatment, there would never be any CO2 produced upon HCHO being absorbed on CeO2 or CeO2–S at room temperature.

Although the surface hydroxyl on CeO2 and CeO2–S cannot function as an oxidative site to directly oxidize HCHO into CO2, they can convert adsorbed HCHO or DOM to formate species through the Cannizzaro mechanism, which is a typical process catalyzed by base catalysts. As shown in Figure S13b, the in situ HCHO-DRIFTS experiments over the CeO2–S catalyst did show a much higher intensity of HCOO peak than that of DOM, accompanied by the consumption of terminal hydroxyl groups (3700 cm–1, one-coordinated OH[49]). The role of such terminal hydroxyl should be similar to those on nonreducible oxides such as γ-Al2O3 that catalyze the formation of the formate intermediate only through the Cannizzaro mechanism, thanks to the lack of active lattice oxygen. It is well known that on the surface of γ-Al2O3 there exists rich terminal and bridge OH, which could function as typical weak base and acid sites, respectively.[73] With HCHO adsorbed on γ-Al2O3, the formation of HCOO can proceed through the Cannizzaro mechanism, where the basic hydroxyl groups could attack DOM to produce formate, and the released hydride reduces another DOM to form methoxyl species.[64] Indeed, as shown in our in situ HCHO-DRIFTS experiment over γ-Al2O3 (Figure S13d), significant HCOO intermediates are found to be formed, with the consumption of this hydroxyl group (3738 cm–1, one-coordinated OH).[74,75] On the other hand, no matter on CeO2, CeO2–S, or atomically dispersed Pt1/CeO2, the bridging lattice oxygen in the “bridge” hydroxyl does have the oxidative function but just does not get more active compared to those bare ones without water dissociation. Hence, they can only oxidize adsorbed HCHO/DOM to formate species but are incapable of completing the rate-controlling step, that is, the further oxidation of formate to CO2. Again, it should be reminded that there remained a notable amount of formate species from the in situ HCHO-DRIFTS experiments on the Pt1/CeO2–S even after 30 min flow of O2 and N2 mixture, which should be due to the lattice oxygen or hydroxyl away from the Pt atom that is the spectator rather than the active site responsible for the full oxidation of HCHO (Note S1 and Table S4).

It is worth noting that the Pt1/CeO2–S catalyst, the atomically dispersed Pt2+ catalyst with a well-defined structure, is reported for the first time to be capable of catalyzing the oxidation of HCHO totally to CO2 and H2O at room temperature. It is the “oxidative” hydroxyl on Pt1/CeO2–S generated from steam treatment that is highly active to catalyze the oxidation of DOM and facilitate the further oxidation of formate species. The mechanism involved in this work is fundamentally distinguished from the earlier reports involved with the surface hydroxyl from water dissociation over the reducible oxide-based catalysts[38−43] or from those supported noble metal catalysts containing mainly nanoparticles or clusters where O2 is activated to oxidize the adsorbed HCHO and the intermediates.[15,35] Moreover, the Pt1/CeO2–S catalyst is catalytically active at extremely low temperatures, and robust in harsh environments, demonstrating the great potential in the application to eliminate the volatile organic compounds.

Conclusions

In summary, this work demonstrates that steam treatment of the Pt1/CeO2 catalyst leads to highly active and stable HCHO oxidation at room temperature. Such high performance can be attributed to the fact that steam treatment greatly facilitates the activation of both surface hydroxyl groups and surface lattice oxygen in the vicinity of Pt. Various characterization techniques demonstrated that formate species are the dominant intermediates after the adsorption of HCHO molecules on the Pt1/CeO2–S surface, which can be further oxidized to form the final products by the activated O2. The mechanism through the oxidative OlatticeH on Pt1/CeO2–S is fundamentally distinguished from those with the surface hydroxyl on the metal oxides generated from water dissociation at room temperature, which is more related with acid–base chemistry. In short, the study of the Pt1/CeO2–S catalyst with the ability of effective HCHO removal at room temperature together with good catalyst stability and tolerance to moisture can provide important applications and contribute to the design of catalysts with rich hydroxyl species.