Controllable synthesis of small-sized Pd nanoparticles on acetic acid-modified halloysite for enhanced toluene oxidation Authors: Haoyan Zhang, Yinmin Zhang, Zhifei Hao, Qihang Zhao, Yanbing Gong, Tianjia Chen, Yongfeng Zhang, Quanzhi Tian, Hao Sun Published: 2026 ABSTRACT AbstractThe particle size of catalyst active components is a key factor for catalytic activity. In this work, functionalized halloysite (Hal) provides a synthetic strategy for the preparation of catalysts loaded with small Pd particles. In this study, halloysite nanotubes were employed as catalyst supports, and the ac-Hal support was synthesized via mild treatment of halloysite nanotubes with acetic acid (ac). The ac-Hal-Pd catalyst was prepared by the coprecipitation method and exhibited excellent catalytic activity for toluene oxidation, with a temperature for 90% toluene conversion (T90) of 179 °C. Transmission electron microscopy (TEM), CO pulse chemisorption, and CO adsorption in situ diffuse reflectance infrared Fourier transform spectroscopy (DRIFTS) analyses revealed that the Pd particles on the surface of the ac-Hal-Pd catalyst had a smaller particle size and more uniformly dispersed Pd particles. X-ray photoelectron spectroscopy (XPS) and O2-temperature-programmed desorption (O2-TPD) analyses indicate that the reduction in the particle size of Pd particles increases the proportion of Pd0 on the catalyst surface and the concentration of adsorbed oxygen, which is also a key factor for the enhancement of catalytic activity. In situ DRIFTS spectra of NH3 adsorption reveal that ac-Hal-Pd possesses a higher abundance of Lewis acid sites, which is favorable for the catalytic reaction. In situ DRIFTS results suggest that toluene oxidation over ac-Hal-Pd proceeds through the formation of benzyl alcohol/alkoxy species, benzaldehyde, benzoate/benzoic acid species, and maleic anhydride intermediates, which are finally oxidized to CO2 and H2O.Supplementary InformationThe online version contains supplementary material available at https://doi.org/10.1038/s41598-026-58558-z.Keywords: Acetic acid treatment, Pd catalyst, Halloysite nanotubes, Toluene oxidationSubject terms: Chemistry, Materials science, Nanoscience and technology Introduction Co-firing in coal-fired boilers is the simplest and most effective method for the clean resource utilization of organic solid waste and biomass1,2. However, the flue gas generated during combustion contains large amounts of toluene, whose emission poses a serious threat to ecosystems and human well-being. Therefore, developing a toluene removal technology is of great significance for co-firing power generation as well as the disposal of organic solid waste and biomass3. Compared with other technologies such as adsorption4, absorption5, membrane separation6, condensation, biodegradation7, photocatalytic decomposition8, and non-thermal plasma oxidation, catalytic oxidation of toluene is a highly feasible multi-pollutant control method with significant practical potential9,10. The development of highly active and stable catalysts is essential for achieving efficient low-temperature catalytic oxidation of volatile organic compounds (VOCs)11. Generally speaking, supported noble metals and metal oxides are two commonly used catalysts for VOC catalytic oxidation. Notably, supported noble metal catalysts (e.g., Pt and Pd) have attracted extensive attention and research in both academic and industrial circles due to their high activity and stability12,13. However, the high cost and agglomeration tendency of noble metals will seriously affect their catalytic performance14. Therefore, the active components need to be uniformly and highly dispersed on appropriate supports to maximize atomic utilization efficiency and enhance catalytic stability15. Pd-based catalysts have garnered extensive attention in VOC abatement, mainly because of their excellent stability and catalytic performance.16. In the absence of ion doping, the activity of Pd-based catalysts is mainly determined by three key factors: the valence state of Pd, support properties, and Pd particle size17,18. Notably, the smaller the Pd particle size, the higher the proportion of Pd0, which in turn enhances catalytic activity19. This phenomenon arises from the fact that smaller particles can expose a greater number of active sites, thereby improving catalytic performance20,21. Several studies have reported that Mg-doped Pd catalysts exhibit excellent catalytic activity in the hydrogenation of 5-hydroxymethylfurfural (HMF)22. The underlying mechanism lies in the modification of Pd’s charge distribution induced by Mg doping. Meanwhile, compared to undoped low-activity Pd catalysts, the introduction of Mg reduces the catalyst particle size and increases the number of active sites. Consequently, regulating the particle size of Pd-based catalysts is recognized as one of the effective strategies to tune their catalytic activity. Halloysite is a naturally occurring layered silicate mineral with a hollow nanotube structure23,24. Due to the difference in dispersed functional groups on its inner and outer tube walls, it exhibits distinct chemical properties between the inner and outer surfaces, enabling the controllable and highly dispersed loading of active components on both surfaces. Furthermore, with abundant reserves and low costs for mining and purification, halloysite nanotubes are considered ideal support materials for catalysts25. Therefore, in existing reports, numerous researchers have fabricated halloysite nanotubes into catalysts, which have been applied in various reactions such as catalytic reduction, oxidation reactions, cross-coupling reactions, and photocatalytic degradation reactions, demonstrating excellent performance26. We have conducted relevant research on halloysite-supported palladium catalysts for toluene oxidation. Based on previous work, this study modified the surface of halloysite via acetic acid treatment, and prepared a Pd-supported halloysite catalyst (ac-Pd/Hal) by coprecipitation. Its catalytic performance, stability, and water resistance in toluene oxidation were evaluated. The results show that when the acetic acid concentration is 4 mol/L, ac-Pd/Hal exhibits the optimal catalytic activity, with a T90 (temperature for 90% toluene conversion) as low as 179 °C. Meanwhile, to clarify the structure-activity relationship of the catalyst during the reaction, a series of characterizations were performed. The results indicate that the enhanced catalytic activity stems from the reduced particle size of the active component Pd during catalyst preparation, which increases the utilization efficiency of the active component. Experimental Material Halloysite (Hal) was purchased from Guangdong Jina New Materials Technology Co., Ltd., China. Palladium nitrate dihydrate (Pd(NO3)2·2H2O, Pd content ≥ 40.0%) was obtained from Merck KGaA, Germany. Glacial acetic acid (analytical grade, purity 99%) and sodium hydroxide (content ≥ 96.0%) were purchased from Tianjin Fengchuan Chemical Reagent Technology Co., Ltd., China. Toluene (analytical grade, purity 99%) was acquired from Tianjin Xinbotai Chemical Co., Ltd., China. Deionized water (18.2 MΩ·cm) was self-prepared in the laboratory. Catalyst preparation Initially, halloysite was subjected to acetic acid treatment (Fig. 1). 3 g of halloysite nanotubes (Hal) were added into 300 mL of acetic acid solution, stirred for 1 h, followed by ultrasonic treatment for another 1 h. After ultrasonication, the mixture was continuously stirred at a constant temperature of 60 °C for 3 days. The product was centrifugally washed with deionized water until the supernatant was neutral, then placed in a vacuum drying oven at 70 °C for 12 h for subsequent use. Depending on the concentration of acetic acid, the acetic acid-modified halloysite was designated as ac-Hal-x, where x represents the concentration of acetic acid (in mol/L). Catalyst preparation flow chart. The Pd loading process was similar to that reported in our previous work. The detailed experimental procedures are as follows: 2 g of acetic acid-modified halloysite was dispersed in 20 mL of deionized water, and 6.25 mL of Pd(NO₃)₂ solution (200 mg/50 mL) was added thereto. After stirring for 30 min, ultrasonic treatment was performed for another 30 min. Subsequently, the mixture was continuously stirred at room temperature for 4 h, followed by the addition of a 1 mol/L NaOH solution to adjust the pH of the mixture to 9, with continued stirring for 1 h. The product was centrifugally washed several times until the supernatant was neutral after centrifugation, then placed in a vacuum drying oven at 70 °C for 12 h. The dried sample was reduced under a H2/Ar flow of 100 mL/min at 400 °C for 4 h, with a heating rate of 5 °C/min. Catalytic activity measurement The catalytic degradation of toluene was conducted in an atmospheric pressure fixed-bed flow reactor. 200 mg of catalyst powder was loaded into a quartz reaction tube with an inner diameter of 7 mm and a length of 650 mm, which was sealed with quartz wool. The quartz tube was then installed in the reactor, and toluene gas (concentration: 1000 ppm) was fed at a flow rate of 100 mL/min, corresponding to a weight hourly space velocity (WHSV) of 30,000 mL·g− 1·h− 1. The reaction temperature was increased from 100 °C to 230 °C at an interval of 10 °C. At each temperature point, the catalyst bed was maintained for 30 min to reach a stable reaction state before the outlet gas was analyzed. Furthermore, in the water resistance test of the catalyst, we maintained the total gas flow rate at a constant 100 mL/min, and introduced water vapor with volume fractions of 5% and 10% into the reaction system via the bubbling method. All catalytic activity tests were repeated three times under the same reaction conditions, and the reported T50 and T90 values were obtained from reproducible measurements. The conversion rate of toluene was calculated by Eq. (1) Where x was toluene conversion, Cin and Cout were toluene concentration in reactor inlet gas and outlet gas(mg/m3), respectively. It is generally accepted that the combustion reaction of VOCs in the presence of excess oxygen follows a first-order kinetic mechanism when their concentration is below 20%27. Therefore, the apparent activation energy (Ea) of the catalyst can be calculated using the Arrhenius equation (Eq. 2): where C is a constant, T is the reactor temperature (K), R is the universal gas constant (J·mol− 1·K− 1) and r is the toluene oxidation rate (mol·s− 1·g− 1). where Ftoluene is the molar flow rate of toluene and mcat is the catalyst mass. Only the data in the low-conversion region were used for Arrhenius fitting. Catalyst characterization X-ray diffraction (XRD) analysis of the catalysts was performed using a Rigaku X-ray diffractometer with CuKα radiation (λ = 1.5406 Å). The operating voltage and current were 40 kV and 40 mA, respectively, with a scanning range of 2θ = 5°–90° and a scanning speed of 2°/min. Transmission electron microscopy (TEM) images were acquired using a FEI Tecnai G-20 microscope. High-resolution TEM (HRTEM) and energy-dispersive X-ray spectroscopy (EDX) mapping analyses were conducted on a FEI Tecnai F20 microscope at an accelerating voltage of 200 kV. The valence states of surface elements in the catalysts were analyzed via X-ray photoelectron spectroscopy (XPS) using a Thermo Fisher Scientific ESCALAB 250 spectrometer. In high-resolution XPS narrow scans, the C1s peak (binding energy: 284.8 eV) was used as an internal standard for charge correction of the binding energies of surface elements. The actual palladium (Pd) content in each catalyst was determined by inductively coupled plasma atomic emission spectroscopy (ICP-AES). Hydrogen temperature-programmed reduction (H2-TPR) analysis was carried out using an AutoChem II 2920 chemisorption analyzer equipped with a thermal conductivity detector (TCD). In situ diffuse reflectance infrared Fourier transform spectroscopy (DRIFTS) was performed using a Bruker INVENIO S spectrometer equipped with a mercury cadmium telluride (MCT) detector to investigate CO adsorption behavior, NH3 adsorption on acid sites, and surface intermediates during toluene oxidation. Results and discussion Structure and micromorphology Structure and micromorphology provide the most direct means of investigating the distribution of active components on catalysts. As observed from the TEM images in Fig. 2, the dispersion of Pd particles on halloysite nanotubes varies with different preparation treatments. The TEM images of the raw-Hal (a), nc-Hal (b) and ac-Hal (c). The TEM and HR-TEM images of the nc-Hal-Pd (d) and ac-Hal-Pd (e, f). The elemental mapping images of the ac-Hal-Pd (g) and nc-Hal-Pd (h). The effect of acid treatment on the catalyst surface was analyzed by transmission electron microscopy (TEM). The TEM images of raw-Hal, nc (nitric acid)-Hal and ac-Hal are presented in Fig. 2a-c. The results showed that both the inner and outer walls of raw-Hal nanotubes were smooth. The outer walls of nc-Hal nanotubes were partially etched with a slight degree of corrosion. In contrast, the outer walls of ac-Hal nanotubes exhibited significant corrosion, showing a rough and uneven morphology. Meanwhile, partial corrosion was also observed on the inner walls of ac-Hal nanotubes, leading to a widened inner diameter in some regions. Figure 3a and b shows the N2 adsorption-desorption isotherms of two supports, nc-Hal and ac-Hal. Both supports exhibit obvious H3-type hysteresis loops in the relative pressure range of 0.4–1.0. Their isotherms are basically consistent and belong to Type IV isotherms. ac-Hal has a higher specific surface area, but the difference is not significant (Table 1). Although ac-Hal and nc-Hal show similar BET surface areas and pore-structure parameters, their surface morphologies and surface charges are clearly different. Acetic acid, as a weak organic acid, induces a milder surface modification of halloysite and produces a rougher nanotube surface without severe destruction of the tubular structure. TEM observations suggest that acetic acid treatment mainly affects the outer surface of halloysite, while partial modification of the inner lumen also occurs. Therefore, we do not claim a strictly selective modification of either the inner or outer surface. The improved Pd dispersion on ac-Hal is mainly attributed to the roughened surface and more favorable negative surface charge, which enhance the anchoring and electrostatic adsorption of Pd²⁺ species. Since no direct evidence confirms the formation of new oxygen-containing functional groups, this effect is described as surface modification rather than the generation of additional functional groups. N2 adsorption-desorption isotherms (a) and corresponding BJH pore size distribution curves (b) of the ac-Hal and nc-Hal. The Zeta potential of raw-Hal, ac-Hal and nc-Hal (c). Table 1BET surface area, average pore diameter and pore volume of the ac-Hal and nc-Hal.SampleBET surface area (m2/g)average pore diameter (nm)average pore volume (cm3/g)ac-Hal34.213.810.086nc-Hal34.163.810.094 BET surface area, average pore diameter and pore volume of the ac-Hal and nc-Hal. For the catalysts synthesized by supporting Pd on the two supports, morphological characterization was performed, and their TEM and high-resolution TEM (HR-TEM) images are shown in Fig. 2d–f. The particle sizes of Pd nanoparticles in nc-Hal-Pd and ac-Hal-Pd are 2.83 nm and 2.56 nm, respectively, based on the TEM particle-size distributions shown in Fig. S1. It should be noted that the TEM-derived particle sizes were calculated from the nanoparticles observed in representative TEM images. Therefore, these values inevitably contain sampling uncertainty and are mainly used as a reference for comparing the relative Pd particle sizes of nc-Hal-Pd and ac-Hal-Pd. The lattice plane spacings of ac-Hal-Pd and nc-Hal-Pd are 0.242 nm and 0.223 nm, respectively, both corresponding to the (111) plane of metallic Pd.28. Elemental mapping shows that Pd species in ac-Hal-Pd are distributed along the halloysite nanotubes, whereas larger Pd aggregates are observed in nc-Hal-Pd. Zeta potential measurements at pH 7 further show that nc-Hal is positively charged, leading to electrostatic repulsion with Pd2+, whereas ac-Hal has a slightly negative surface charge, which favors the adsorption and anchoring of Pd2+ species. Overall, TEM, elemental mapping, and Zeta potential results indicate that acetic acid treatment facilitates Pd dispersion on halloysite. Combined with the ICP-OES data (Table 2), this indicates that the active component Pd in ac-Hal-Pd has better dispersion and more uniform particle size. Table 2Theoretical and measured Pd loading amounts of the Pd-supported catalysts based on Hal acidified with different concentrations of acetic acid.SampleTheoretical (%)Actual (%)nc-Hal-Pd0.5%0.471.0 M-ac-Hal-Pd0.5%0.391.5 M-ac-Hal-Pd0.5%0.442.0 M-ac-Hal-Pd0.5%0.362.5 M-ac-Hal-Pd0.5%0.393.0 M-ac-Hal-Pd0.5%0.473.5 M-ac-Hal-Pd0.5%0.494.0 M-ac-Hal-Pd0.5%0.494.5 M-ac-Hal-Pd0.5%0.485.0 M-ac-Hal-Pd0.5%0.415.5 M-ac-Hal-Pd0.5%0.456.0 M-ac-Hal-Pd0.5%0.49 Theoretical and measured Pd loading amounts of the Pd-supported catalysts based on Hal acidified with different concentrations of acetic acid. The surface accessibility and adsorption configuration of Pd sites on nc-Hal-Pd and ac-Hal-Pd were further investigated by CO adsorption in situ DRIFTS. As shown in Fig. 4, during the CO purging process, four peaks at 2170 cm− 1, 2120 cm− 1, 2178 cm− 1, and 2119 cm− 1 in Fig. 4a and c are attributed to the gaseous adsorption peaks of CO29. In contrast, four characteristic peaks at 2094 cm− 1, 2080 cm− 1, 2039 cm− 1, and 2098 cm− 1 correspond to the linear adsorption peaks of CO30,31. Among these, the linear adsorption peak of CO (2098 cm− 1) on the surface of the ac-Hal-Pd catalyst is the most prominent, indicating better dispersion of Pd particles on the surface of the ac-Hal-Pd catalyst. For CO adsorption under N2 purging atmosphere, as shown in Fig. 4b and d, only three absorption peaks at 2024 cm− 1, 2000 cm− 1, and 1974 cm− 1 remain in Fig. 4b, which are all attributed to the bridged adsorption of CO on Pd particles30,32. In Fig. 4d, however, not only gaseous adsorption peaks of CO (2178 cm− 1, 2122 cm− 1) are observed, but also linear adsorption (2040 cm− 1) and bridged adsorption structures (2099 cm− 1, 1974 cm− 1) of CO on Pd particles are detected. These CO adsorption features suggest that more accessible surface Pd sites are present on ac-Hal-Pd than on nc-Hal-Pd, which is consistent with the TEM observations. Under the CO purge atmosphere: in-situ DRIFTS spectra of the ac-Hal-Pd (a) and nc-Hal-Pd (c) catalysts; Under the N2 purge atmosphere: in-situ DRIFTS spectra of the ac-Hal-Pd (b) and nc-Hal-Pd (d) catalysts. CO pulse chemisorption was further used to quantitatively evaluate Pd dispersion. The results show that ac-Hal-Pd exhibits a higher Pd dispersion of 24.12% and a smaller chemisorption-derived Pd particle size of 3.87 nm than nc-Hal-Pd (Fig. 5 and Table 3). Although the Pd particle size derived from CO pulse chemisorption is larger than that obtained from TEM, both methods show the same trend, confirming that acetic acid modification promotes the formation of smaller and more accessible Pd particles. Pulse CO adsorption signals over ac-Hal-Pd and nc-Hal-Pd catalysts. Table 3Pd dispersion, metal surface area, and Pd particle size determined by CO pulse chemisorption on ac-Hal-Pd and nc-Hal-Pd catalysts.SamplePd dispersionMetallic surface areaPd particle diameternc-Hal-Pd15.98%0.3559 m²/g5.84 nmac-Hal-Pd24.12%0.5372 m²/g3.87 nm Pd dispersion, metal surface area, and Pd particle size determined by CO pulse chemisorption on ac-Hal-Pd and nc-Hal-Pd catalysts. Catalytic oxidation performance of toluene In this study, the catalytic activity was comprehensively evaluated by testing toluene conversion at different temperatures, activation energy, long-term stability, and water resistance of the catalysts. We conducted catalytic performance tests for toluene combustion on all prepared samples. The test results, presented in Fig. 6a; Table 4, were evaluated using T90 and T50 as the evaluation criteria. The results show that when the acetic acid concentration was 4.0–5.5 mol/L, the T90 value decreased to 179 °C, which is 6 °C lower than that of our previously reported mineral-supported Pd-based catalyst (185 °C)33. A comparison of the reported catalytic systems is presented in Table 5. It should be noted that the comparison is only semi-quantitative because the catalytic performance is strongly affected by reaction conditions, including toluene concentration, WHSV/GHSV, noble-metal loading, support composition, and gas composition. Therefore, a direct ranking based only on T90 values is not fully appropriate. Under the present reaction conditions of 1000 ppm toluene and a WHSV of 30,000 mL·g− 1·h− 1, ac-Hal-Pd exhibits competitive low-temperature activity among reported Pd-based catalysts. Because ac-Hal and nc-Hal show nearly identical BET surface areas and pore-structure parameters (Table 1), the activity difference is not mainly governed by surface area. Based on the structural characterization above, the enhanced activity of ac-Hal-Pd is mainly associated with improved Pd anchoring/dispersion and higher Pd utilization efficiency rather than an increase in BET surface area. In particular, considering its low Pd loading and the low-cost halloysite support, the catalyst shows promising Pd utilization efficiency for toluene oxidation. Nevertheless, further comparison under identical reaction conditions is still required for a more rigorous evaluation. The catalytic performance of acx-Pd-Hal in toluene oxidation; (b) The activation energy of ac-Hal-Pd and nc-Hal-Pd; The long-term stability (c) and water-tolerant tests (d) of ac-Hal-Pd and nc-Hal-Pd (The long-term stability and water -tolerant tests were conducted at 180 ℃ with WHSV and toluene concentrations of 30000 mL·g− 1·h− 1 and 1000 ppm, respectively). Table 4The parameters of toluene conversion of ac-Hal-Pd with the different concentrations of acetic acid.SampleToluene Conversion (°C)T50T901.0 M-ac-Hal-Pd1811901.5 M-ac-Hal-Pd1771892.0 M-ac-Hal-Pd1831902.5 M-ac-Hal-Pd1811873.0 M-ac-Hal-Pd1761853.5 M-ac-Hal-Pd1791864.0 M-ac-Hal-Pd1741794.5 M-ac-Hal-Pd1741805.0 M-ac-Hal-Pd1771805.5 M-ac-Hal-Pd1731796.0 M-ac-Hal-Pd175180 The parameters of toluene conversion of ac-Hal-Pd with the different concentrations of acetic acid. Table 5Reported works about oxidation of toluene over Pd-based catalysts.CatalystPd loading(wt %)Toluene concentrationWHSV(mL·g− 1·h− 1)T90Ref.Pd/W20-Al2O30.51000 ppm180,000210 ℃ 34 0.5%Pd/SnO2-R0.51000 ppm20,000240 °C 35 PdAg@S-10.551000 ppm20,000179 °C 36 Pd0.90Co0.10@Z50.71600 ppm60,000199 °C 37 0.47Pd-0.46Pt/α-MnO20.471000 ppm60,000152 °C 13 Pd/meso-TiO2−CeO20.75700 ppm30,000251 °C 15 Pd8Ce2/A-Pal0.3931000 ppm30,000185 °C 38 LPMO-900#4.141000 ppm60,000215 ℃ 39 ac-Hal-Pd0.4911000 ppm30,000179 ℃This work Reported works about oxidation of toluene over Pd-based catalysts. Activation energy is a key indicator for describing the catalytic performance of catalysts, which can directly reflect the difficulty of the catalytic reaction. The lower the activation energy, the lower the energy barrier that needs to be overcome during toluene oxidation, and the higher the catalytic performance. The apparent activation energies of ac-Hal-Pd and nc-Hal-Pd were recalculated from the Arrhenius plots in the low-conversion region. As shown in Fig. 6b and Table S1, ac-Hal-Pd exhibits a lower apparent activation energy of 73.5 kJ/mol than nc-Hal-Pd of 82.1 kJ/mol. The lower Ea value indicates that toluene oxidation over ac-Hal-Pd requires a lower energy barrier, which is consistent with its superior low-temperature catalytic activity. This result further supports that acetic acid modification promotes the formation of more active Pd species and improves the efficiency of toluene oxidation. Besides conversion, stability is also an important indicator of catalyst performance. Meanwhile, water vapor in the reaction atmosphere can affect catalytic performance, so it is necessary to test the long-term stability and water resistance of the catalysts. Figure 6c shows the conversion changes of nc-Hal-Pd and ac-Hal-Pd during continuous reaction at 180 ℃ (T90) for 24 h. The test results indicate that both catalysts exhibit good short-term stability at 180 °C. However, it should be noted that the 24 h stability test is still insufficient for evaluating practical industrial durability, and longer-term tests under more realistic reaction conditions are required in future work. In addition, Fig. 6d presents the conversion of the catalysts under a water vapor atmosphere. The results show that when the volume fraction of water vapor is 5% and 10%, the performance of both catalysts decreases significantly. Specifically, when the water vapor fraction reaches 10%, the conversion of ac-Hal-Pd decreases from 99% to 87%. However, after turning off the water vapor gas flow, the catalytic activity of the catalysts can recover to the state before water vapor was introduced, indicating that the catalysts exhibit reversible water inhibition and recoverable catalytic activity under humid conditions. Although toluene was selected as a representative aromatic VOC in this work, the applicability of ac-Hal-Pd to other aromatic VOCs, such as benzene, ethylbenzene, and xylene, still requires further systematic investigation because different aromatic VOCs may show different adsorption and activation behaviors. Surface electronic states of the catalysts XPS analysis was performed to investigate the surface chemical states of Pd and O species in the two catalysts. Figure 7 shows the XPS spectra of nc-Hal-Pd and ac-Hal-Pd. As can be seen from Fig. 7a, the surface oxygen species mainly exist in three forms: surface adsorbed oxygen species (Oads: 531.5–531.8 eV), lattice oxygen (Olatt: 529.8 eV), and oxygen in adsorbed water molecules (532.5 eV)40,41. The ratio of Oads to Olatt (Oads/Olatt) in ac-Hal-Pd is 31.08, which is higher than that in nc-Hal-Pd (9.03) (Table 6)42. Electron paramagnetic resonance (EPR) spectroscopy was further employed to investigate oxygen-vacancy-related defects on the catalyst surfaces. As shown in Fig. 8, both catalysts exhibit a broad and symmetric signal at g = 2.003, which is generally associated with oxygen-vacancy-related paramagnetic defects. To provide a more reliable comparison, quantitative EPR analysis was performed under the same measurement conditions. The oxygen-vacancy-related spin concentration of ac-Hal-Pd was calculated to be 1.023 × 1013 spins g− 1 which is higher than that of nc-Hal-Pd (7.368 × 1012 spins g− 1). Since EPR detects unpaired electron spin centers rather than all oxygen vacancies, the spin concentration is described here as the concentration of EPR-detectable oxygen-vacancy-related defects. Therefore, the higher spin concentration of ac-Hal-Pd indicates that it contains more EPR-detectable oxygen-vacancy-related sites, which is consistent with the XPS result showing a higher Oads/Olatt ratio43,44. The XPS patterns of ac-Hal-Pd and nc-Hal-Pd catalysts. Table 6Summarized XPS data of the catalysts.SampleBinding Energy (eV)Molar ratioPd 3d5/2Pd 3d3/2O 1sPd0Pd2+Pd0Pd2+OlattOadsOads/OlattPd0/Pd2+nc-Hal-Pd335.12336.44340.68342.62530.92532.119.032.49ac-Hal-Pd335.31336.80340.53342.87529.24531.5531.0810.56 Summarized XPS data of the catalysts. EPR spectra of ac-Hal-Pd and nc-Hal-Pd. In both catalysts, Pd species mainly exist as metallic Pd0, accompanied by a small amount of oxidized Pd2+ species. From the Pd 3d spectra, two sets of asymmetric signals appear for both catalysts, corresponding to metallic Pd0 (Pd 3d5/2) at 335.11 eV and 340.69 eV, and oxidized Pd2+ (3d3/2) at 336.46 eV and 342.6 eV (Fig. 7b)45. Compared with nc-Hal-Pd, the binding energy of the Pd 3d orbital in ac-Hal-Pd shifts to the higher binding energy direction, which is consistent with previous research results. In Pd-based catalysts, Pd0 is an important factor determining catalytic activity: the higher the proportion of Pd0, the stronger the catalytic activity. The ratio of Pd0 to Pd2+ (Pd0/Pd2+) in ac-Hal-Pd is 10.56, which is higher than that in nc-Hal-Pd (2.49). This shows that the reduction in Pd particle size increases the proportion of Pd0, corresponding to the higher catalytic performance of ac-Hal-Pd. Toluene catalytic oxidation reaction mechanism The mechanism of toluene catalytic oxidation is of great guiding significance for the development of catalysts. In this work, the reaction mechanism of toluene oxidation over the catalysts was discussed in depth by investigating the adsorption–desorption behavior of various gases on the catalyst surface and employing in situ infrared spectroscopy. The adsorption/desorption processes of toluene on the catalysts were investigated by toluene temperature-programmed desorption (toluene-TPD) tests (Fig. 9a-b, Fig. S2). A large amount of toluene desorbed at temperatures below 300 ℃. Among them, the toluene desorption amount of nc-Hal-Pd was greater than that of ac-Hal-Pd due to its larger specific surface area. This indicates that the adsorption capacity of toluene on the catalyst surface is weak, and toluene desorbs without participating in the oxidation reaction. It is worth noting that the peak temperature of the toluene desorption curve for ac-Hal-Pd is 107 ℃ lower than that for nc-Hal-Pd. This observation suggests relatively weak toluene adsorption on the ac-Hal-Pd surface. As can be seen from Fig. 9b, the generation of CO2 signals is mainly concentrated in the range of 400 ℃–800 ℃, and the CO2 signal area of ac-Hal-Pd is larger than that of nc-Hal-Pd. This shows that more CO2 desorbed from ac-Hal-Pd during the Toluene-TPD process, confirming that toluene is more likely to react on the ac-Hal-Pd catalyst surface. The O2 signal in the products during the toluene-TPD process is shown in Fig. S2. The O2 signal varies with temperature, reflecting the combined effects of oxygen desorption and oxygen consumption during toluene-related surface reactions. Owing to the complexity of this signal, Fig. S2 is used as supplementary information for the toluene-TPD analysis rather than as the primary evidence for identifying active oxygen species. Spectra of Toluene (a)and CO2 (b) in Toluene TPD products of ac-Hal-Pd and nc-Hal-Pd catalysts; H2-TPR (c) and O2-TPD (d) profiles of ac-Hal-Pd and nc-Hal-Pd catalysts. The H2-TPR tests of the two catalysts are shown in Fig. 9c. The H2 consumption curves of the two catalysts show no significant differences, which should be attributed to the H2 reduction during the preparation process. Meanwhile, the H2-TPR profiles of both ac-Hal-Pd and nc-Hal-Pd show distinct negative peaks around 80–90 °C, which correspond to the decomposition of the PdHₓ species generated by the adsorption of H₂ on Pd at ambient temperature46. O2-TPD tests were performed to investigate the reaction mechanism between toluene and the oxygen species on the catalyst surface (Fig. 9d). The O2 signal curve of ac-Hal-Pd occupies a larger area at temperatures below 500 ℃, indicating that it has more Oads. This is consistent with the XPS characterization results. In addition, the lattice oxygen desorption temperature of ac-Hal-Pd is lower than that of nc-Hal-Pd (729 °C vs. 755 °C). Generally, a lower lattice oxygen desorption temperature suggests that lattice oxygen species are more weakly bound and can be more readily activated during the redox process, which is usually associated with enhanced oxygen mobility and oxidation ability. Therefore, the lower lattice oxygen desorption temperature of ac-Hal-Pd indicates more active lattice oxygen species, which may facilitate the deep oxidation of toluene.47. The types of acid sites on the two catalysts were investigated by NH3 adsorption in situ DRIFTS, as shown in Fig. 10. The infrared absorption bands observed at 1625 and 1260 cm− 1 are attributed to NH3 adsorbed on Lewis acid sites48,49, whereas the bands at 1693, 1497, and 1407 cm− 1 correspond to NH3 adsorbed on Brønsted acid sites50. Lewis acid sites can promote the adsorption and polarization of toluene-related intermediates, especially oxygen-containing intermediates such as benzaldehyde and benzoate species. The retention of Lewis acid sites on ac-Hal-Pd at the reaction temperature may facilitate the adsorption and subsequent oxidation of intermediates, thereby contributing to the enhanced catalytic activity. These results indicate that for ac-Hal-Pd, Lewis acid sites constitute the predominant acid sites at ambient temperature, and a considerable abundance of Lewis acid sites is retained upon heating to T90. In contrast, for nc-Hal-Pd, both Lewis and Brønsted acid sites coexist at ambient temperature; however, the Lewis acid sites disappear upon heating. Therefore, the presence of Lewis acid sites is a critical factor contributing to the high catalytic activity of ac-Hal-Pd. In situ DRIFTS spectra of NH₃ adsorption on ac-Hal-Pd and nc-Hal-Pd at ambient temperature (30 °C)and 180 °C. To gain a deeper understanding of the toluene oxidation process, In situ DRIFTS was used to monitor surface intermediates formed during toluene oxidation over ac-Hal-Pd, providing insight into the possible reaction pathway. The catalyst was exposed to 1000 ppm toluene gas with 20% O2 in the carrier gas. Before the test, the catalyst was heated to 300 ℃ at a programmed rate under N2 protection, held for 1 h, and then cooled to room temperature to remove physisorbed substances on the catalyst surface. As shown in Fig. 11, C-H stretching vibration peaks belonging to the aromatic ring were observed at 3038 cm− 1 and 3042 cm− 1. In addition, a distinct alkyl C-H stretching vibration peak was detected at 2938 cm− 1, especially associated with the methyl or methylene groups of toluene51. The bands at 3038, 3042, and 2938 cm− 1 indicate the presence of adsorbed toluene-related species during the reaction. In situ DRIFTS results of toluene oxidation on ac-Hal-Pd with different exposure times at 180 ℃. With the extension of test time, new characteristic peaks appeared at 1388, 1475, 1520, 1600, 1617, and 1654 cm− 1. Among them, the peaks at 1388, 1600 and 1617 cm− 1 are typical characteristic peaks of carboxylic acids, corresponding to the intermediate product benzoic acid52. The peaks at 1475 and 1520 cm− 1 are assigned to C-O stretching vibrations, indicating the conversion of methyl groups in toluene to C-O bonds53. The peak at 1654 cm− 1 corresponds to the C = O stretching vibration of benzaldehyde54. After 20 min of reaction, the characteristic peaks at 1731, 1800, 1860, and 1944 cm− 1 are attributed to the symmetric and asymmetric stretching vibrations of C = O in maleic anhydride55,56. A characteristic peak of the final product CO2 was observed at 2352 cm− 157. Combining the In situ DRIFTS data and toluene-TPD results, the reaction proposed pathway of toluene over the catalyst is proposed as follows: Toluene is adsorbed on the catalyst surface and oxidized to benzyl alcohol, benzaldehyde, and benzoic acid. These intermediates further react to form maleic anhydride, which is finally converted to CO2 and H2O (Scheme 1). Proposed reaction pathway for toluene oxidation over ac-Hal-Pd based on in situ DRIFTS analysis. Conclusions A Pd-based catalyst was prepared by acetic acid modification of halloysite nanotubes. When the acetic acid concentration was 4 mol/L, ac-Hal-Pd exhibited a T90 value of 179 °C for toluene oxidation, lower than that of our previously reported mineral-supported Pd-based catalyst. Structural and surface analyses indicate that acetic acid modification produces a rougher and more favorable halloysite surface, which promotes the formation of smaller and more uniformly dispersed Pd nanoparticles. XPS, quantitative EPR, O2-TPD, and NH3-DRIFTS results further suggest that the improved catalytic activity is associated with higher Pd utilization efficiency, more surface active oxygen/oxygen-vacancy-related defects, and abundant Lewis acid sites. This work provides a mild weak-acid modification strategy for regulating Pd dispersion on halloysite supports and offers guidance for the design of low-cost mineral-supported noble-metal catalysts for VOC oxidation. Further studies on long-term durability and broader VOC applicability are still required for practical application. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1 Author contributions H.Z. and Y.Z. conducted the experiments, performed the data analysis, and carried out the material characterization. The conceptual framework was developed by Y.Z. and Z.H., who also contributed to the critical revision of the manuscript. H.Z. wrote the original draft of the manuscript and participated in subsequent revisions. Y.Z. secured the funding necessary for this work. Q.Z., Y.G., T.C., Y.Z., Q.T., and H.S. reviewed the manuscript. All authors have read and approved the final version of the manuscript. Funding The authors gratefully acknowledge the financial support provided by the National Natural Science Foundation Project of China (52064040), the National Natural Science Foundation Project of China (22206090), the Natural Science Foundation of Inner Mongolia (No. 2022MS05010 and No. 2022QN02009), the central government guides local funds for scientific and technological development (2023ZY0001), the Major Special Project of Jungar Banner (No. 2025ZD-02) and the Group Project of Developing Inner Mongolia through Talents from the Talents Work Leading Group under the CPC Inner Mongolia Autonomous Regional Committee (2025TYL04). Data availability The data that support the findings of this study are available from the corresponding author, upon reasonable request. Declarations Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Contributor Information Yinmin Zhang, Email: zhangyinmin@imut.edu.cn. Zhifei Hao, Email: haozhifei@imut.edu.cn. References Associated Data Supplementary Materials Supplementary Material 1 Data Availability Statement The data that support the findings of this study are available from the corresponding author, upon reasonable request.