Formaldehyde (HCHO) as a primary indoor gas pollutant can increase the risk of asthma, dyspnea, or even teratogenesis and carcinogenesis [1]. The concentration of indoor HCHO is usually in the sub-ppm or ppb range. But in some newly refurbished rooms, the concentration of HCHO may reach several hundred ppm [2]. Numerous available methods have been adopted for HCHO elimination [3]. Catalytic oxidation at room temperature is an effective strategy since no external energy (light, heat, etc.) is demanded. HCHO molecules can be oxidized by reactive oxygen species (ROS) and mineralized into CO₂ and H₂O over catalysts [4–7]. Principally, noble-metals have been regarded as the most active catalyst as they are capable of O₂ activation at room temperature [8,9]. However, due to the rarity of noble-metals, modulating the microstructures of noble-metals plays a crucial role in determining the catalytic efficiency.

Geometrically, single-atom catalysts (SACs) have realized the maximum atomic utilization efficiency [10]. The unique low-coordination environments facilitate the adsorption of reactants. Our previous work [11] has demonstrated that N-doped carbon octahedron supported iridium single atoms (Ir₁-N-C) with 1.6 wt% Ir loading can achieve high HCHO removal and conversion efficiency (~97%) in 760 min at 20 ℃. When the Ir loading decreases to 0.5 wt%, the HCHO conversion efficiency only reaches 42.7%. The weak O₂ activation capacity over low-loading Ir single atoms leads to an incomplete conversion of HCHO. O₂ could only be activated into ^•O₂^–, which slows down the reaction rate. How to improve O₂ activation capacity of low-loading Ir₁-N-C has become a challenge. Previous studies have explored two routes for improving the catalytic activity of SACs: increasing the density of active sites [12–14] and improving the intrinsic activity of the active sites [15]. Therefore, seeking the effective ways to improve the intrinsic activity of Ir single atoms may be a feasible strategy. Combing inorganic particles with M₁-N-C (M: transition metal) could optimize the electronic structure and the high catalytic activity is expected. In the preparation of M₁-N-C by pyrolysis of metal–organic frameworks (MOF), the inorganic particles (such as transition metal carbide, nitride and oxide, metal particles, etc.) were generated simultaneously. These inorganic particles may play a positive role in catalytic activity. Cheng et al. [16]. found that the iron nitride combined with the iron and nitrogen co-doped carbon layer (Fe_xN@Fe-N-C) selectively enhanced the activity of Fe-N-C toward CO₂ reduction reaction (CO₂RR). The introduction of Fe_xN could regulate the balance between the adsorption and desorption of CO. The same group also reported the promoted O₂ reduction reaction over Co@Co-N-C compared with that over Co-N-C due to the downshift of the d-band center of Co nanoparticles [17,18]. Nano-structured ZrO₂ was reported to cooperate with single-atom Ni-N-C to drive electrochemical CO₂RR, where the Ni-N₄ species was the real active sites and the nano-ZrO₂ accelerated the protonation of intermediate products. In addition, Zhang et al. [19]. constructed the hybrid catalyst containing ZrO₂ nanoclusters and Fe-N-C. ZrO₂ nanoclusters can improve O₂ adsorption capacity and ORR activity, which was attributed to the strong interaction between ZrO₂ and the isolated Fe single atom.

Motivated by the strategy, a catalyst containing ZrO₂ nanoparticles coupled Ir single atoms in N doped carbon (Ir₁-N-C/ZrO₂) was prepared by in situ pyrolysis of UiO-66-NH₂ adsorbed Ir ions. The reference Ir₁-N-C was obtained by etching ZrO₂ in HF solution. The presence of ZrO₂ nanoparticles was expected to promote the catalytic activity of Ir₁-N-C. The electronic structure of Ir₁-N-C/ZrO₂ was characterized by X-ray photoelectron spectroscopy (XPS), electron spin resonance (ESR) and density functional theory (DFT) calculations. The effect of ZrO₂ on O₂ adsorption and activation was investigated by temperature-programmed technique and DFT calculations. Furthermore, in situ diffuse reflectance infrared Fourier transform spectroscopy demonstrated the HCHO oxidation mechanism.

Ir₁-N-C/ZrO₂ was derived from UiO-66–NH₂ adsorbed Ir ions (Fig. 1a). N atoms of −NH₂ can readily trap the adsorbed Ir ions and form isolated Ir-N₄ coordination during the pyrolysis process. The [Zr₆(μ₃-O)₄(μ₃-OH)₄]¹²⁺ nodes were transformed to ultrasmall ZrO₂ nanoparticles and H₂BDC-NH₂ linker were decomposed into N-doped carbon skeleton [11].

According to the XRD patterns (Fig. 1b), Ir₁-N-C/ZrO₂ displays the typical tetragonal zirconia structure (t-ZrO₂, PDF No. 49-1642) [20]. The broad diffraction peak at 23° is corresponded to the (002) plane of amorphous graphitic carbon [21]. The t-ZrO₂ crystallites are dispersed in graphitic carbon. The average crystallite size of t-ZrO₂ was determined to be 5.7 nm using Scherrer’s equation and the (111) peak. For Ir₁-N-C prepared by HF etching, the XRD characteristic peaks of t-ZrO₂ disappear and Ir₁-N-C displays a typical XRD patterns of graphitic carbon.

Ir₁-N-C/ZrO₂ inherits the recognizable octahedral configuration of Ir/UiO-66−NH₂ (Fig. S1 in Supporting information). This is further verified by aberration-corrected high angle annular dark-field scanning transmission electron microscope (HAADF-STEM, Fig. 1c). The average diameter of Ir₁-N-C/ZrO₂ is approximately 200–300 nm. The magnified HAADF-STEM image (Fig. 1d) revealed that the particle sizes of ZrO₂ nanoparticles are about 4.0 nm. This is because the carbon skeleton could build a barrier to hinder the aggregation of ZrO₂ nanoparticles during pyrolysis. And the lattice fringe spacing of 0.295 nm is assigned to the (011) plane of a typical t-ZrO₂ [22]. The bright spots within the yellow ring areas are assigned to the isolated Ir atoms, indicating atomic dispersion of Ir on the carbon support. All of C, N, Zr, O and Ir elements are uniformly distributed within the entire octahedral matrix, as demonstrated by the STEM energy-dispersive X-ray spectroscopy (EDX) element mapping (Fig. 1e). Meanwhile, After HF etching, the HAADF-STEM image of Ir₁-N-C maintains a similar octahedral morphology. The atomically isolated Ir atoms can be clearly observed on the N-C support (Fig. S2 in Supporting information), indicating that the ZrO₂ nanoparticles were completely removed.

In order to investigate the changes of electronic properties of Ir single atoms, the surface chemical states of Ir₁-N-C/ZrO₂ were analyzed by XPS and ESR technique. Ir₁-N-C/ZrO₂ exhibits a spin-orbit doublet of the Zr 3d core level into 3d_5/2 and 3d_3/2 levels with an energy gap of 2.4 eV. The 3d_5/2 and 3d_3/2 binding energies locate at 182.3 eV and 184.7 eV (Fig. 2a). Compared with the high-resolution spectra of Zr 3d (182.2 for 3d_5/2 and 184.6 eV for 3d_3/2) of N-C/ZrO₂ (Fig. S3 in Supporting information), the Zr 3d XPS spectrum of Ir₁-N-C/ZrO₂ shifts to higher binding energy. Combined with the high-resolution XPS spectra of O 1s, the characteristic peaks located at 529.8 eV and 531.2 eV are corresponded to lattice oxygen (O_lat) and surface adsorption oxygen species (O_ads), respectively [23], which proves that ZrO₂ in Ir₁-N-C/ZrO₂ conforms to the stoichiometric ratios [24]. Furthermore, the O 1s high-resolution XPS spectra of Ir₁-N-C/ZrO₂ and N-C/ZrO₂ were analyzed. The surface molar ratio of O_ads/O_lat of Ir₁-N-C/ZrO₂ is 1.3, higher than that of 1.1 of N-C/ZrO₂ indicating that the introduction of Ir single atom favors the formation of surface adsorbed oxygen. The fine C1s spectrum shows three carbon species: sp²-hybridized structure (284.6 eV), C-N groups (286.0 eV), and carboxyl groups (288.7 eV) [23] (Fig. 2b). C atoms in the carbon carrier mainly exist in the form of sp² hybridization, which is consistent with the graphitized carbon carrier characterized in the XRD pattern. The high-resolution broad N1s spectrum is fitted into four components, which are associated with pyridinic N (Py-N, 398.6 eV), pyrrolic N (Pr-N, 400.1 eV), graphitic N (G-N, 401.1 eV), and oxidized N (O-N, 403.1 eV) [17] (Fig. 2c). The ratio of Py-N/total N of Ir₁-N-C/ZrO₂ and Ir₁-N-C were 0.41 and 0.44, respectively. Furthermore, the high-resolution Ir 4f spectra of Ir₁-N-C can be deconvoluted into Ir 4f_7/2 and Ir 4f_5/2 spin-orbit doublets centered at 61.9 eV and 64.9 eV, which could be attributed to the predominant oxidation state of Ir^δ+ [25] (Fig. 2d). The Ir 4f_7/2 and Ir 4f_5/2 spin-orbit doublets of Ir₁-N-C/ZrO₂ shift to lower binding energies (61.8 eV and 64.8 eV), indicating that the electron density close to Ir single atoms increases. Combined with the above results, it can be speculated that Ir species in Ir₁-N-C/ZrO₂ exist in the form of single atoms, without Ir nanoparticles or clusters. Ir single atoms mainly exist in the coordination of Ir-N₄. The surface chemical states of Ir single atoms changed, suggesting that the Ir-C-Zr channel formed.

ESR spectroscopy was further used to study the unpaired electrons of Ir₁-N-C/ZrO₂. As shown in Fig. 2e, for Ir₁-N-C, a single Lorentzian line with a g value of 2.002 was observed (Fig. 2e), which can be assigned to the unpaired electron of N-doped carbon [26]. When introduced ZrO₂ nanoparticles, the intensity of the narrow peak of g = 2.002 decreased significantly. The decreasing ESR signal reveals that less lone electrons resided in Ir₁-N-C/ZrO₂ and electrons become more localized. Based on DFT calculation, the Mulliken charge of Ir atom for Ir₁-N-C/ZrO₂ is positive (1.05) (Fig. 2f), lower than that of Ir₁-N-C reported in our previous work (1.12) [11]. The results are consistent with these of the high-resolution Ir 4f spectra. These above results mean that after interact with ZrO₂, the Ir is more electron deficient, so there forms a localized electronic state around Ir atom, which provides an electron donor active site for the subsequent catalytic reaction.

The promotional effect of ZrO₂ nanoparticles on HCHO oxidation was investigated. The HCHO oxidation was evaluated under ambient conditions (Atmospheric pressure, T = 20 ℃) in terms of HCHO removal (η) and HCHO conversion (X). As shown in Figs. 3a and b, Ir₁-N-C/ZrO₂ displays the highest η and X values. After 120 min, the η values of all samples are higher than 80%, which can be ranked as Ir₁-N-C/ZrO₂ > Ir₁-N-C + ZrO₂ (Fig. S4 in Supporting information) > Ir₁-N-C > N-C/ZrO₂. The X values of these samples were quite different. Ir₁-N-C/ZrO₂ exhibits a remarkable HCHO oxidation performance with X of 95.6%, compared to Ir₁-N-C (20.6%), indicating that ZrO₂ nanoparticles play a positive role in promoting HCHO oxidation performance. In contrast, the pristine N-C/ZrO₂ without Ir species exhibits inferior catalytic activity (η = 94.1%, X = 15.8%). This could be ascribed to the porous structure and N species favorable for the efficient HCHO adsorption. ZrO₂ mainly contributes to HCHO conversion. But the X values gradually decreased within 2 h (Figs. S5, S6 and S7 in Supporting information), indicating that Ir single atoms were the main reactive active sites. In addition, the physical mixture of Ir₁-N-C + ZrO₂ shows lower catalytic activity (η = 86.3%, X = 55.7%), suggesting the interaction between isolated Ir atoms and ZrO₂ nanoparticles instead of mechanical mixing in Ir₁-N-C/ZrO₂.

The intrinsic activities of Ir₁-N-C/ZrO₂ and Ir₁-N-C catalysts were also compared by calculating the specific rates (R) values. R value was calculated when X value reaches 20% (Fig. 3c). R value of Ir₁-N-C/ZrO₂ was calculated as 1285.6 mmol g_Ir⁻¹ h⁻¹ when the weight hourly space velocity (WHSV) was set as 360,000 mL h⁻¹ g_cat⁻¹ after 120 min, surpassing that of Ir₁-N-C (21.4 mmol g_Ir⁻¹ h⁻¹) when the WHSV was set as 60,000 mL h⁻¹ g_cat⁻¹. The R value of Ir₁-N-C/ZrO₂ was also higher than the values of 43.5 mmol g_Ir⁻¹ h⁻¹ (1.8 wt% Ir/Al₂O₃), 79.2 mmol g_Ir⁻¹ h⁻¹ (1.5 wt% Ir/Al₂O₃) [27] and 159.0 mmol g_Ir⁻¹ h⁻¹ (0.95 wt% Na-Ir/TiO₂) [28] reported in previous studies (Table 1). Based on these above results, the Ir₁-N-C/ZrO₂ outperformed most Ir-based catalysts reported hitherto. The long-term stability of the catalyst is also an important factor for the evaluation of the catalyst. Fig. 3d shows the catalytic activity of Ir₁-N-C/ZrO₂ maintained as long as 1000 min. The HCHO conversion is about 95%, indicating that Ir₁-N-C/ZrO₂ has good catalytic long-term stability. In addition, the effect of humidity on the catalytic performance of Ir₁-N-C/ZrO₂ was measured under the relative humidity of 20%, 50% and 75% (Figs. 3e and f). The HCHO conversion still kept above 80% when the RH increased to 75%, indicating the catalyst has excellent moisture resistance. When the RH increased from 50% to 75%, more water molecules competed with HCHO molecules over the surface of Ir₁-N-C/ZrO₂. The decreasing adsorbed HCHO molecules may result in a low η value. However, when the RH increased from 50% to 75%, the increasing water molecules may react with the active oxygen atoms and form hydroxyl groups. The highly reactive hydroxyl groups may promote the conversion of adsorbed HCHO molecules into CO₂.

To understand the mechanisms underlying the promoting effect of ZrO₂ nanoparticles on HCHO oxidation, the relationship between the structure and catalytic activity of Ir₁-N-C/ZrO₂ and Ir₁-N-C was investigated. N₂ sorption–desorption experiment was conducted to analyze the texture structure of Ir₁-N-C/ZrO₂ and Ir₁-N-C (Fig. S8 in Supporting information). Both Ir₁-N-C/ZrO₂ and Ir₁-N-C samples show typical type Ⅳ adsorption and desorption isotherms. The isotherm of Ir₁-N-C/ZrO₂ has a H2-type hysteresis ring while the isotherm of Ir₁-N-C has a H3-type hysteresis loop, which leads to the different pore structures of the two samples. The Ir₁-N-C/ZrO₂ sample mainly possesses a large number of mesoporous pores with an average pore size of 12.7 nm. The pore size ensures that the HCHO and O₂ molecules could get to the active sites in the catalyst. Table S2 in Supporting information lists the specific surface area (S_BET), pore volume (V_p) and pore size (d_p) parameters of Ir₁-N-C/ZrO₂ and Ir₁-N-C. HCHO conversion efficiency over Ir₁-N-C/ZrO₂ and Ir₁-N-C was normalized by being divided by S_BET, and the results were 0.0685% and 0.0357% g/m², respectively. It can be seen that S_BET is not the only factor affecting the catalytic performance.

The oxygen species of Ir₁-N-C/ZrO₂ was studied by O₂-TPD, as shown in Fig. 4a. The O₂-TPD spectra can be divided into three regions: the desorption of surface adsorbed oxygen (< 200 ℃), the desorption of the surface lattice oxygen (200–400 ℃) and the desorption of intrinsic lattice oxygen in bulk phase (> 400 ℃), respectively [29]. A small peak at 109 ℃ is observed for Ir₁-N-C/ZrO₂ from the 5-fold magnified figure, whereas no obvious peaks were observed for Ir₁-N-C and N-C/ZrO₂ at lower temperature below 200 ℃. This indicates that more easily migrated surface adsorbed oxygen species were activated over Ir₁-N-C/ZrO₂. In the range of 200–400 ℃, Ir₁-N-C/ZrO₂ and N-C/ZrO₂ show two obvious desorption peaks, respectively. The typical desorption peak of N-C/ZrO₂ was located at 341 ℃, while 285 ℃ for Ir₁-N-C/ZrO₂. Compared to that of N-C/ZrO₂, the desorption peak of surface lattice oxygen of Ir₁-N-C/ZrO₂ appear at a lower temperature and the peak intensity become stronger, indicating that the interaction between Ir single atom and ZrO₂ promotes the activation of surface lattice oxygen over Ir₁-N-C/ZrO₂.

The adsorption behavior of O₂ molecule on the surface of Ir₁-N-C/ZrO₂ was further investigated by DFT calculation. O₂ molecule prefers to be adsorbed on Ir sites by the end-on adsorption pattern with the formation of Ir−O bonds and the adsorption energy of −1.47 eV (Fig. 4b), higher than that over Ir₁-N-C (−0.43 eV) [11]. The O-O bond is elongated to 1.35 Å. In comparation with Ir₁-N-C reported in our previous work, the high adsorption energy and elongated O-O bond manifested the promotional effect of ZrO₂ nanoparticles on O₂ adsorption. Furthermore, ESR technique was used to trace the reactive oxygen species (ROS) after O₂ adsorption (Fig. 4c). The DMPO/^•O₂⁻ adduct with characteristic six split lines (a_H = 7.7G, a_N = 13.6 G) [30] was observed for Ir₁-N-C. The signal attenuated significantly when ZrO₂ was introduced to Ir₁-N-C, indicating that the generation of ^•O₂⁻ was significantly reduced or most ^•O₂⁻ species was transformed into other oxygen species. Combined with XPS and O₂-TPD results, it is reasonable to speculate that the surface reactive oxygen species were generated after O₂ activation. In addition, the introduction of ZrO₂ increased the adsorption energy of HCHO molecule from 0.25 eV to −1.38 eV (Fig. 4d) [11].

The O₂ dissociation behavior and the activation energy of the major reaction paths of HCHO oxidation at the interface of Ir₁-N-C/ZrO₂ were calculated using DFT calculations [31] (Fig. 5a). The results show that O₂ obtains electrons and formed ^•O₂⁻. Then ^•O₂⁻ dissociates into ^*O by the Ir-C-Zr channel rapidly. The activation energy of the adsorbed O₂ dissociation at the interface of Ir₁-N-C/ZrO₂ is −1.0 eV or −23.06 kcal/mol (1 eV = 23.06 kcal/mol); then two active oxygen atoms (^*O) are generated, which can oxidize HCHO to form formic acid. Specifically, an ^*O₁ produced by the dissociation of adsorbed O₂ begins to migrate to the interface between the Zr site of ZrO₂ (011) and Ir₁-N-C. Another ^*O₂ is closer to the Zr site of ZrO₂ (011). The ^*O₁ reacts with HCHO to form formic acid (HCOOH), the activation energy of which is 0.6 eV (13.84 kcal/mol). Furthermore, the ^*O₂ reacts with HCOOH to form CO₂ and H₂O, the activation energy of which is 0.4 eV (9.22 kcal/mol). Thus, the activation energy of the rate-determination step of the major reaction path is 13.84 kcal/mol, which is less than our previous report results [32], and the rate-determining step is that the reactive ^*O₁ near the Zr site of ZrO₂ (011) is transferred to the C atom in C-H bond of HCHO (HCHO + ^*O₁ → HC^*O₁OH). The formed reactive oxygen species could oxidize HCHO into DOM, formate species and CO₂ and H₂O rapidly.

In situ DRIFTS was further performed to monitor the intermediates generated in the catalytic process, revealing the reaction paths over Ir₁-N-C/ZrO₂ and Ir₁-N-C surfaces. All catalyst powders were dispersed onto an inert silica wool for testing. For Ir₁-N-C/ZrO₂ supported on silica wool, after the mixture of HCHO and air flow was introduced for 2 min, three obvious characteristic peaks appeared at 968, 1120 and 1290 cm⁻¹, which were attributed to ν(CO), ρ(CH₂) and τ(CH₂) of dioxymethylene (DOM) species [33,34], respectively. As reaction time was prolonged, the characteristic peak intensity of DOM decreased; especially the intensity of characteristic peak at 1290 cm⁻¹ decreased significantly. At the same time, the intensity of characteristic peak at 1560 cm⁻¹ attributed to ν_as(OCO) of increased gradually, indicating that DOM was further oxidized to formate species. The bands in the range of 3250–4000 cm⁻¹ originated from the gradual accumulation of water molecules and hydroxyl groups [34]. The DOM and formate species were the main intermediates generated in the reaction (Fig. 5b). In contrast, in terms of Ir₁-N-C supported on silica wool, the intensities of these characteristic peaks are extremely low under the same experimental conditions and the peaks at 1120 cm⁻¹ and 1290 cm⁻¹ vanished (Fig. 5c). The results indicate that only a small fraction of HCHO molecules were oxidized into DOM and the reactive oxygen species including ^•O₂⁻ and surface oxygen species generated on the surface of Ir₁-N-C/ZrO₂ could oxidize HCHO into DOM and formate species rapidly.

Based on the DFT calculations and in situ DRIFTS results, the HCHO degradation pathways were proposed over Ir₁-N-C/ZrO₂. ZrO₂ nanoparticles and Ir single atoms synergistically enhanced HCHO oxidation at room temperature. O₂ was activated to ^•O₂⁻ and surface oxygen species, which served as the key reactive oxygen species to oxidize HCHO to DOM, formate species and CO₂ and H₂O rapidly. However, on the surface of Ir₁-N-C, O₂ was mainly adsorbed and activated to generate ^•O₂⁻.

To summarize, an effective strategy was proposed to improve the catalytic activity of low-loading Ir single atoms at room temperature. By engineering ZrO₂ nanoparticles with Ir single atoms in N-doped carbon (Ir₁-N-C/ZrO₂), the enhancing effect of ZrO₂ nanoparticles on HCHO oxidation was demonstrated by experimental results and DFT calculations. The specific reaction rate of 0.25 wt% Ir₁-N-C/ZrO₂ could reach as high as 1285.6 mmol g_Ir⁻¹ h⁻¹, surpassing the Ir-based catalysts reported in previous studies. 0.26 wt% Ir_1--N-C only achieved a low specific reaction rate of 21.4 mmol g_Ir⁻¹ h⁻¹. ZrO₂ nanoparticles can regulate the electronic property of Ir single atom through charge redistribution, synergistically promoting the activation of surface adsorption oxygen and surface lattice oxygen to generate reactive oxygen species. The Ir-C-Zr channel also accelerated the dissociation of ^•O₂⁻ to active oxygen atom (^*O). The formed reactive oxygen species could oxidize HCHO to DOM, formate species and CO₂ and H₂O rapidly. In brief, our work offers a new way to obtain the low-loading single-atom catalyst with superior catalytic activity towards HCHO oxidation by rational design.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgments

This work was supported by the Strategic Priority Research Program of the Chinese Academy of Sciences, China (Nos. XDA23010300 and XDA23010000), National Science Foundation of China, China (Nos. 52200137 and 21725102), the Plan for "National Youth Talents" and GuangDong Basic and Applied Basic Research Foundation (No. 2021A1515110427).

Supplementary materials

Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.cclet.2023.109219.