Introduction

Microwave (MW) heating has attracted significant attention as a novel tool for accelerating chemical reactions up to the industrial scales1,2,3,4. Large MW heating-mediated reaction accelerations have often been reported in gas-solid catalytic reaction systems5,6,7,8,9. In contrast to conventional heating (CH), MW energy can be directly supplied to the solid catalysts, which results in focused heating of the solid catalyst. Further, MWs afford shorter reaction times at lower reaction temperatures and improve energy consumption efficiency.

Metal nanoparticles (NPs) interact strongly with MWs and exhibit enhanced catalytic reactions10,11,12,13,14,15,16,17,18,19,20,21,22. The selective heating of metal NPs by MWs has been considered as a critical factor for the reaction rate enhancements5. Several groups have demonstrated the lowering of the reaction temperatures of the NP-mediated catalysis under MW irradiation. For example, Guler et al. demonstrated that the catalytic dehydrogenation of ammonia by Mo NPs was accelerated by MWs, at a temperature that was 250 K lower than that with CH19. Similarly, Jie et al. applied a Fe NPs-based catalyst to the dehydrogenation of hydrocarbon fuel and demonstrated that the reaction could be facilitated by MWs18,21. The reaction under MWs prevented the generation of light alkane by-products. The authors suggested that the selective and local heating of the supported Fe NPs accelerates the dehydrogenation reaction at the Fe surface and prevents the unexpected side reaction due to thermal cracking.

The impact of the MW-based selective heating on the catalytic reactions has been extensively studied in Pt NPs systems. In their early works, Zhang et al. reported that CO2 reforming of CH4 progressed under MW heating at a temperature that was 100 K lower than that with CH12. Meanwhile, Perry et al. reported only a small enhancement of CO oxidation under MW irradiation using Pt/Al2O3 and Pd/Al2O310. They undertook a numerical approach and suggested that the rapid heat dissipation from the Pt NPs to gas-phase did not bring about a temperature gradient between the Pt NPs and the surroundings23. Silverwood et al. also reported the lack of acceleration in the CO oxidation when the Europt-1 (6 wt% Pt/SiO2) catalyst was employed14,15. While extensive studies have been previously conducted to probe these phenomena, the local temperatures of the NPs under MWs have not been well understood due to their very small sizes.

Gaining insights into the temperature distribution in the reaction system is vital for elucidating the mechanism of the reaction enhancement by MWs5. Infrared (IR) thermometers are generally used for measuring the external surface temperature of the catalyst bed in the gas-solid catalytic system. However, the internal temperature often becomes higher than the surface temperature, due to which, the measurement of the surface temperature alone can cause an overestimation of the MW-based reaction enhancements. To prevent the overestimation, thermocouples10,20 or fiber-optic thermometers8,9,20,24 are inserted into the center of catalyst beds for measuring its core temperature. In contrast, thermography measures the temperature distribution at the surface of the catalyst bed. Besides, numerical simulation can estimate and visualize the complete temperature distribution of the catalyst bed8,9,24.

The temperature of the NPs at the active site is indispensable for discussing the effects of the local heating on the catalytic reactions under MWs. Kabb et al. demonstrated the ex situ nano-thermometric analysis of the vicinity of Au NPs, which were locally-heated by MWs25. Tsukahara et al. have reported that in situ Raman spectroscopy can estimate the local temperature around Co particles with 0.1–3.0 μm diameters26. To achieve a similar objective in nano size, Ano et al. had applied a temperature-dependent luminescent lifetime of Rhodamine B as a molecular temperature probe27. However, Raman spectroscopy and molecular probe methods are available only in limited conditions. Raman spectroscopy is limited only to highly Raman-active materials, and the organic molecular probes decompose at the high temperatures. For these reasons, these methods are not suitable for determining local temperatures of metal NPs during gas-solid reactions.

Synchrotron X-ray analysis is a powerful tool for in situ analysis under MW heating28,29,30,31,32,33,34. In particular, X-ray absorption fine structure (XAFS) spectroscopy affords not only direct structural information of the metal NPs but also the electronic states and the local structures at sub-nanometer scales. Cozzo et al. have demonstrated the first in situ XAFS measurement under MW irradiation for detecting the gelation of Ce solution31. Furthermore, Van de Broek et al. have proposed that the Debye–Waller factor (σ2) in extended X-ray absorption fine structure (EXAFS) oscillations could be interpreted as the temperate of the Au NPs under Plasmonic heating35. Theoretically, the σ2 contains thermal disorder (σT2) and structural disorder (σs2)36,37,38,39. In their report, Van de Broek et al. treated the increase of σ2 into the σT2 only because the structure of Au NPs was consistent through the heating, which means that σ2 will directly be correlated with the temperature of the Au NPs. These noteworthy results indicate that the local temperature of metal NPs under MW heating can be determined by the in situ XAFS method.

In this work, we demonstrate the nano-thermometric analysis of supported Pt NPs by in situ XAFS spectroscopy under various MW heating conditions. We discuss the local temperature of Pt NPs (TPt, denoted in Fig. 1) from the σ2 values. The TPt exhibits larger temperature than the macroscale temperature distribution in the pellet, indicating the formation of the nanoscale local-high temperature at Pt NPs. We further demonstrates the effect of the local heating on the enhancement of catalytic dehydrogenation of 2-propanol gas. operando X-ray absorption near edge structure (XANES) spectroscopy also confirms the enhanced reduction of the oxidized Pt (PtOx) on Al2O3 by MW irradiation. Finally, we conclude that the local heating of Pt NPs contributions to the enhanced catalytic reactions by MWs.

Fig. 1: Schematic diagram of the in situ XAFS setup under MW heating using TM010-mode MW cavity.
figure 1

Tex, Tin, and TPt represent the external surface temperature of the catalyst bed, the inner temperature of that, and nanoscale local temperature of Pt NPs, respectively.

Results and discussion

Nano-thermometric analysis of Pt NPs by MW in situ EXAFS

Figure 2a, b shows the TEM image and the XRD patterns of Pt/Al2O3, respectively. We found that the average particle size was 1.9 ± 0.35 nm referring to the size distribution in Supplementary Fig. 9c. The diffraction peak at 2θ = 40° was assigned to the Pt (111) lattice plane, which indicates that highly dispersed Pt0 NPs were formed on Al2O3. Figure 2c, d shows in situ FT-EXAFS spectra of Pt/Al2O3 under CH and MW heating which are obtained by curve-fitting, respectively (Supplementary Figs. 58, Supplementary Tables 23). The peaks around 2.7 Å are attributed to the Pt-Pt bonding of Pt NPs. Under CH, the peak intensity gradually decreased as the temperature rises (Fig. 2c). In contrast, the peak intensity under MW heating decreased instantly (Fig. 2d).

Fig. 2: In situ EXAFS of Pt/Al2O3 under CH and MW heating.
figure 2

a TEM image and b XRD pattern of Pt/Al2O3 catalyst. c FT-EXAFS spectra obtained by in situ XAFS measurement of Pt/Al2O3 catalyst under CH and d under MW heating. e Relationship between curve-fit σ2 and Tex. f TPt transformed from Δσ2 using the linear equation shown in e. Δσ2 values were relative values compared to σ2 of Pt foil as reference.

The general EXAFS theory (Eq. 1) indicates that the Debye–Waller factor (σ2)40 is an essential factor in the amplitude dampening of EXAFS oscillation,

$$\chi \left( k \right) = S_0^2\mathop {\sum }\limits_i \frac{{N_iF_i\left( k \right)}}{{kR_i^2}}{\mathrm{e}}^{ - 2k^2\sigma _i^2}{\mathrm{sin}}[2kR_i + \varphi _i(k)]$$
(1)

where χ(k) is the EXAFS oscillation, k is the wave vector of the excited photoelectron, N is the number of atoms at the interatomic distance (R), F(k) is the back-scattering amplitude, φ(k) is the phase shift function. The σ2 consists of the thermal vibration factor (\(\sigma _T^2\)) and the static vibration factor (\(\sigma _S^2\)), as described in Eq. 237,39.

$$\sigma ^2 = \sigma _T^2 + \sigma _S^2$$
(2)

Therefore, it can be assumed that the \(\sigma _{\mathrm{T}}^2\) could be isolated if the \(\sigma _{\mathrm{S}}^2\) has a constant value, which means no significant structural changes in the Pt NPs through a heating procedure. We found that the TEM results of Pt/Al2O3 after MW heating indicate no changes in the Pt size distribution compared with those before MW heating (Fig. 2a and Supplementary Fig. 9b–d). Further, no significant differences were observed in the curve-fit Δσ2 before and after heating (Supplementary Fig. 9a). We also found no specific changes in other curve-fit parameters, N, R, and ΔE0 under both CH and MW heating (Supplementary Table 3 and Supplementary Fig. 8). These results support that the \(\sigma _{\mathrm{T}}^2\) is a crucial factor in the dampening of Pt-Pt peaks in Fig. 2c,d.

Figure 2e shows the plot of the curve-fit Δσ2 values against Tex. Notably, the σ2 values under MW heating increased instantly while the Δσ2 values under CH increased gradually. Further, a linear relationship was observed between the Δσ2 and temperature under CH (Fig. 2e). Van de Broek et al. have previously proposed that the σ2 values can be used to determine the local temperature of NPs35. Therefore, we interpreted the observed changes of the Δσ2 values obtained under MW heating as those of the local temperature of Pt NPs (TPt, denoted in Fig. 2f) by applying the linear relationship. The results indicated that the TPt was 101 K higher than 378 K, which is the external surface temperature of the pellet measured by the IR thermometer (Tex, Fig. 1).

To discuss the extent of the local heating of the Pt NPs acculately, the Tex was further corrected to the exact average temperature where the X-ray passes through the internal section of the catalyst pellet (Tin, Fig. 1) and the Tin was compared with the TPt. The temperature distribution of the whole catalyst pellet can be depicted by the coupled simulation8, toward which we applied COMSOL Multiphysics 5.4a software analysis with electromagnetic field and heat transfer modules. The temperature of the catalyst pellet surface under MW heating (18 W) was first measured by a microscopic thermography with 20 μm resolution (Fig. 3a). The MW setup was exactly reproduced with the one used for the in situ EXAFS (Fig. 3b). The thermographic image indicated the static temperature gradient is formed in the catalyst pellet surface; the center temperature was higher than the edge temperature without spontaneous hot spots. The radial temperature gradient was reproduced by the simulation with the effective thermal conductivity of the pellet as 0.29 W m−1 K−1. Figure 3c shows the simulated temperature mapping, and the line profile of the simulation result coincided with that of experimental result (Fig. 3d). The temperature distribution obtained by the simulation for Line 3 axisis is displayed in Fig. 3g. The simulation depicts the whole temperature distribution in the catalyst pellet as shown in Fig. 3e. Line 4 axisis indicates a path in the catalyst bed through which X-ray passes, and the average temperature of the Line 4 (Tin) was calculated as 453 K. When the Tex was 378 K, the average Tin attained 453 K (Supplementary Figs. 15 and 16). Meanwhile, there was no temperature gradient when CH was used (Supplementary Fig. 17). The average TPt was estimated as 479 K by the in situ XAFS method, and the temperature difference between TPt and Tin was 26 K. This indicates that the Pt NPs supported on Al2O3 are locally-heated under MWs.

Fig. 3: Determination of temperature distribution of Pt/Al2O3 catalyst pellet under MW heating.
figure 3

a Temperature mapping by the thermography under 18-watt MWs. b Schematic illustration of MW setup for measuring the temperature mapping of catalyst pellet by the thermography. c Reproduced temperature mapping by COMSOL Multiphysics simulation. d Temperature profiles of lines 1 and 2. e Simulated temperature mapping from a different angle. f Temperature profiles of lines 3 and 4. g The average temperature of line 4 (Tin) along with the X-ray beam during in situ EXAFS experiment.

Next, we turned to the catalytic conversion of 2-propanol by using the Pt/Al2O3 catalyst under MWs to study the effect of Pt local heating on the catalytic reactions. Figure 4 shows the reaction results under CH and MW heating. Acetone was generated as the main product by the dehydrogenation of 2-propanol at the Pt surface. For instance, the yield of acetone at 373 K was 9.6 % under CH, while the yield by MW attained 16.5 % at the same temperature (Supplementary Fig. 19). This reaction was considerably suppressed without Pt NPs (Supplementary Fig. 20). Since the yield at 398 K under CH was 22.0 %, the reaction enhancement by MWs corresponded to the lowering of the temperature by ~25 K. The production of propylene by the intra-molecular dehydration of 2-propanol was also enhanced under MW irradiation. These reaction enhancements should be attributed to the below two aspects of the high-temperature area: The first is the Pt NPs, and the other is the internal part of the catalyst pellet. A fiber-optic thermometer measured the core temperature (Tcore) of the pellet. The Tcore was higher than the Tex by +8 K (at 373 K) to +13 K (at 448 K) (Supplementary Fig. 22). However, these are too low to explain the above reaction enhancements (~25 K). Therefore, we considered that the reaction enhancement was attributable to both effects of the macroscale local heating and the nanoscale one.

Fig. 4: Catalytic conversion of 2-propanol by Pt/Al2O3 under MW and CH.
figure 4

a Product yields by catalytic conversion of 2-propanol with Pt/Al2O3. b Comparison of ratios of product yields by MW to those by CH.

We further applied the experiments to those with Pt/SiO2 catalyst to compare the effect of the metal oxide supports (Fig. 5). Figure 5a shows the average Pt size to be 6.3 ± 1.8 nm (n = 17), and Fig. 5b shows sharp peaks of Pt (111), Pt (200), and Pt (220) in the XRD pattern. These results indicate that the Pt/SiO2 contains larger Pt0 NPs compared to those in Pt/Al2O3. Further, the in situ XAFS results of the Pt/SiO2 indicated that the TPt attained 603 K when the Tex was 376 K (Fig. 5c-f, Supplementary Figs. 10, 11, Supplementary Table 4). The temperature difference was 227 K, which was larger than the Pt/Al2O3. The temperature distribution in the Pt/SiO2 pellet was simulated to obtain Tin (Supplementary Fig. 18), which was determined as 471 K when TPt and Tex were 603 K and 376 K, respectively. Therefore, the temperature difference between TPt and Tin was 132 K.

Fig. 5: In situ EXAFS of Pt/SiO2 under CH and MW heating.
figure 5

a TEM image and b XRD pattern of Pt/SiO2 catalyst. c FT-EXAFS spectra obtained by in situ XAFS measurement of Pt/SiO2 catalyst under CH and d under MW heating. e Relationship between curve-fit Δσ2 and Tex. f TPt transformed from Δσ2 using the linear equation shown in e. Δσ2 values were relative values compared to σ2 of Pt foil as reference.

The yield of the acetone generated using Pt/SiO2 under MWs at 373 K was 2.6 times higher than that with CH (Fig. 6, Supplementary Fig. 21). The yield of acetone generated under MW irradiation at 373 K was 5.9 % between those achieved with CH at 398 K (3.6 %) and 423 K (7.2 %). Thus, the reaction enhancement by MW irradiation corresponds to the decrease in the reaction temperature by ~50 K. Considering the macroscopic temperature distribution, the Tcore of the Pt/SiO2 catalyst bed was 7 K higher than Tex at 373 K (Supplementary Fig. 22), and thereby, the reaction enhancement by MWs is much larger than that expected by this temperature difference. We conclude that the lager reaction enhancement by Pt/SiO2 is attributable to the nano-sized high temperature of the Pt NPs in Pt/SiO2.

Fig. 6: Catalytic conversion of 2-propanol by Pt/SiO2 under MW and CH.
figure 6

a Product yields by catalytic conversion of 2-propanol with Pt/SiO2. b Comparison of ratios of product yields by MW to those by CH.

As a result, the local heating realized in the Pt/SiO2 system was more remarkable than that in the Pt/Al2O3 system (Table 1). Accordingly, the Pt/SiO2 catalyst exhibited a larger enhancement of the reaction rate by MWs than that with the Pt/Al2O3 catalyst. Since there were no significant temperature differences between Tex and Tcore between Pt/Al2O3 and Pt/SiO2, the larger acceleration in Pt/SiO2 catalyst can be explained by the local heating of the Pt NPs. We hypothesize the following four factors that explain the differences in the local heating between the Pt/Al2O3 and the Pt/SiO2.

  1. (1)

    Heat transfer from the Pt NPs to gas: The size of the Pt NPs (6.3 nm) on the SiO2 support is larger than that of the Pt/Al2O3 (1.9 nm). The surface area of the Pt NPs on SiO2, which are in contact with air, is smaller. Thus, the Pt NPs on SiO2 are not cooled as much as those on Al2O3.

  2. (2)

    Heat transfer from the Pt NPs to the supports: The thermal conductivity of the SiO2 support was 0.20 W m−1 K−1, which was lower than that of the Al2O3 support (0.29 W m−1 K−1) (Supplementary Table 7). Moreover, the interfaces of the Pt/supports should be different each other. The surface of γ-Al2O3 can be defective41, thereby disallowing a large contact area at the Pt/support interface where heat transfers occur.

  3. (3)

    Selectivity of the Pt heating by MWs: TG results (Supplementary Note 1, Supplementary Figs. 2, 3) showed that the Pt/SiO2 contains a smaller amount of adsorbed water as compared to the Pt/Al2O3. The Pt heating can be more efficient in the Pt/SiO2 under MWs because there is less microwave absorbers in the Pt/SiO2 system except for the Pt nanoparticles.

  4. (4)

    Structural and electric effects of the Pt NPs: Differences in the size, shape, and electrostatically-charging state of the Pt NPs can affect the local MW absorption properties.

Table 1 Summary of temperature differences in TPt, Tex, and Tin.

Effect of heat transfer from Pt NPs to gas-phase

The significant heat transfer to gas-phase has been considered to limit the occurence of local high temperatures at the supported Pt NPs23. The heat dissipation from the particle to gas-phase was predominant because the typical Pt NPs only come into contact with the support in a small area. And then, we have examined the effect of gas flow with different thermal conductivity on the TPt under MW irradiation.

The in situ XAFS spectra of the Pt/SiO2 pellet under N2 or He flow were measured using the gas flow type system under MWs (Supplementary Fig. 12, Supplementary Table 5). Figure 7 shows that TPt attained 502 K under N2 flow (10 mL min−1) while Tex was 341 K. Thus, their difference reached 161 K. The TPt under N2 was almost the same as that under air (Fig. 5f). On the other hand, the temperature difference between TPt and Tex attained only 110 K under He. The less local heating under He can be attributed to the higher thermal conductivity of the He (Supplementary Table 6). Besides, He plasma can also diminish the formation of the local high temperature of the Pt NPs. The He plasma was observed when high MW power, above 60 W, was applied to the catalyst bed (Supplementary Fig. 13). While the plasma was not observed below 50 W, where the in situ XAFS was conducted, the microplasma might be formed and absorb MW energy. This microplasma could contribute to the reduction in the MW energy concentration on the Pt NPs. We consider that the local high-temperature of Pt NPs is affected by the heat dissipation to the gas depending on their thermal conductivity as well as possible formation of microplasma.

Fig. 7: TPt of Pt/SiO2 in N2 or He flow conditions under MWs.
figure 7

Transformed from σ2 using the linear equation shown in Fig. 5e.

MW operando XANES of the reduction of PtOx/Al2O3

We conducted operando XANES spectroscopy to explore the local heating effect of PtOx NPs in the reduction of Pt oxide nanoparticles (PtOx) supported on Al2O3 with 2-propanol. For the reduction process, we supposed that the MW energy absorbed in PtOx NPs could be efficiently used for the reduction without the heat dissipation to surroundings (namely, 2-propanol and Al2O3). Figure 8a–c shows the normalized XANES spectra of PtOx/Al2O3 under CH. The white line intensities, which appeared at around 11565 eV, gradually decreased over the reaction time at 373 K. Further, the rate of this decrease became larger as the temperature increased. Under MW heating (Fig. 8d, e), notably, the peak intensity decreased drastically even at lower temperatures.

Fig. 8: Operando XANES of PtOx/Al2O3 during PtOx reduction reaction by 2-propanol.
figure 8

Operando XANES spectra obtained under ac CH and d, e MWs. f References. g Relative reduction of PtOx.

Figure 8g shows the relative rate of PtOx reduction estimated by the Linear Combination Fitting (LCF) method42. The reduction rates under MW heating at 353 K and 373 K were almost identical to those under CH at 423 K and 473 K, respectively. These results indicate that the MWs promote the PtOx reduction lowering the reaction temperature by 100 K than that in the CH condition. Also, these operando XANES results are consistent with the in situ EXAFS results that TPt was 101 K higher than Tex, and further indicate that the reaction enhancement is attributed to a thermal effect of the Pt local heating. We conclude that the local heating of Pt NPs is efficient for accelerating the chemical reaction with Pt itself, which constitutes one of the most practical approaches to obtain a dramatic enhancement in catalytic reactions under MW irradiation.

This work demonstrates for the first time, the assessment of the local high temperatures of supported NPs and their effect on catalytic reactions. The nano-thermometric analysis with in situ XAFS is a general method and can be applied to the other metal types of the nanoparticles as long as the structure and the oxidation state of the nanoparticle do not change during heating.As future work, the sizes of the supported Pt NPs should be precisely controlled for a comparative discussion of the support effect on the local heating of the Pt NPs. According to the previous reports, highly endothermic reactions can be more favorable for utilizing the advantages of high local temperatures generated by MWs5,12. Moreover, large MW effects have often been observed in catalytic reactions under high-temperature, such as the production of hydrogen from fossil fuels and other materials21. Efficient energy concentration at the active sites of the catalysts should be one of the critical strategies for exploring the microwave chemistry to achieve the efficient energy use for reactions and to enable conditions for reaction acceleration at apparent lower temperatures with the use of locally-heated sites.

We apply in situ XAFS spectroscopy under MW heating for the nano-thermometric analysis of supported Pt NPs. The temperature of Pt NPs (TPt) is discussed from the curve-fit σ2 values extracted from the structural information measurued by in situ EXAFS. The σ2 values are further converted to average TPt using the calibration curve obtained by CH. The average TPt is higher than Tex by 101 K for the Pt/Al2O3 pellet when Tex was 378 K. On the other hand, significant local heating is observed for the Pt/SiO2 system, where the TPt was 227 K higher than the Tex of 376 K. Furthermore, the dehydrogenation of 2-propanol with Pt/SiO2 proceeds faster compared to that with Pt/Al2O3. Operando XANES further analyzes the extent of the reduction of pre-oxidized PtOx/Al2O3 during dehydrogenation of 2-propanol. The reduction rate of PtOx under MWs proceeds at temperatures that were 100 K lower to those with the CH condition. The enhancement coincides with the 101 K temperature difference in Pt/Al2O3. Therefore, we conclude that the nano-temperature of Pt under MWs can be determined by the in situ XAFS method, and the Pt local heating has a significant impact on the enhancement of the catalytic reaction rates. These developed methods are examples that enable quantitative verification of the local high temperature formed by MW irradiation. Precise control of the local heating at the supported metal NPs allows the novel design of catalytic reaction processes with the efficient propagation of the MW energy to the active sites.

Methods

Catalysts preparation

Supported Pt NPs catalysts (3 wt%) were prepared by the impregnation method. γ-Al2O3 (97 %, STREM Chemicals Inc.) and SiO2 (99.9 %, Fujifilm Wako Pure Chemicals Co.) were calcined at 973 K for 5 h in the air. The samples were immersed in aqueous H2PtCl6 and dried up at 393 K to obtain the 3 wt% Pt precursors. Pt/Al2O3 and Pt/SiO2 were synthesized by the reduction of the precursors under H2 (36 mL min−1) flow at 773 K for 2 h. The calcination of the precursor synthesized PtOx/Al2O3 at 773 K for 5 h in air. Supported Pt NPs were analyzed by FE-TEM (JEM-2010F, JEOL Ltd.) and XRD (MiniFlex600, Rigaku Co. Ltd.). Dielectric properties of as-prepared and vacuum-dried (383 K, 72 h) catalysts were measured by the perturbation method42,43 using a 2.45 GHz cavity resonator (TM010 mode) equipped with a vector network analyzer (ZND, Rohde & Schwarz, Supplementary Table 1). The amounts of water adsorbed on Pt/Al2O3 and Pt/SiO2 were determined using a thermo-gravimetric analyzer (TGA-51, Shimadzu Co.) under 20 mL min−1 Ar flow (Supplementary Fig. 1).

In situ X-ray XAFS under CH and MW heating

XAFS experiments were conducted in BL 9C beamline at KEK-IMSS-PF (Tsukuba, Japan). Synchrotron radiation from the energy storage ring was monochromatized by Si(111) channel-cut crystals. The monochromator angle was calibrated using Pt foil. Ionization chambers were filled with 15% Ar-85% N2 mixed gas and 100% Ar for monitoring the incident X-ray (I0) and transmitted X-ray (I), respectively. The 11,057–12,662 eV energy range was used for detecting the Pt L3-edge absorption in the QXAFS-mode at 180 s for each scan. Each Pt/Al2O3 and Pt/SiO2 was uniformly mixed with 30 wt% boron nitride (BN) (Fujifilm Wako Pure Chemicals Co.) as a filler and it is pelletized in the quartz tube. The homogeneous pellet is used for following XAFS experiments without unexpected scatterings of the incident X-ray.

The in situ XAFS measurement under CH was applied using an in situ XAFS cell (KEK IMSS PF), which can heat a sample under gas flow (Supplementary Fig. 4a). The sample of 110 mg was pelletized in a metal ring and placed in the in situ XAFS cell. A heat insulator surrounds the in situ XAFS cell to maintain the homogeneous temperature. Two thermocouples measure the temperatures at the sample and flowing gas to keep the same temperature during XAFS measurement. The XAFS spectra were obtained in the temperature range between 298–673 K XAFS spectra were analyzed with REX2000 software (Rigaku Co., Japan) to obtain the k3-weighted EXAFS spectra and Fourier-transformed EXAFS (FT-EXAFS) spectra using the k = 3–16 Å−1 range. The Fourier filtering was limited to the R = 1.9–3.1 Å range during the curve fitting processes. For estimating precise Δσ2, curve fit was conducted with empirical F(k) and φ(k) extracted from the EXAFS spectra of Pt foil (Nilaco Co.), which were measured by in situ XAFS at 298–673 K under He flow condition to prevent oxidation of Pt.

The in situ XAFS measurement under MW irradiation was conducted using a semiconductor MW generator and the TM010-mode cavity resonator (Ryowa Electronics Co., Ltd., Supplementary Fig. 4b). The sample of 180 mg was pelletized in a quartz tube and they were placed in the MW cavity. The temperature was monitored by a quartz-transparent IR thermometer TMSH STM0050 (Japan Sensor Co., Ltd.) to obtain the external surface temperature (Tex, Fig. 1) of the pelletized catalyst. MWs were irradiated into the cavity at the constant actual power calculated by subtracting the reflected power of the MW from the incident MW power. Impedance matching was maintained using the slug tuner and frequency-auto-tracking system. The QXAFS spectra were obtained after Tex became constant. The XAFS analysis and curve fitting were conducted by the same method used for the CH conditions indicated above. More details of MW in situ EXAFS was indicated in the Supplementary Note 2.

For the gas flow conditions, the pelletized catalyst sample in a quartz tube was sealed with polyimide (Kapton®) film windows to obtain the MW in situ XAFS spectra under 10 mL min−1 N2, or He flow (Fig. 9a). The exact pellet temperature distribution under MW heating (Tin, Fig. 1) was determined by combining thermography and a coupled simulation of electromagnetic wave and heat transfer8,9,24.

Fig. 9: Microwave heating setup.
figure 9

a Schematic diagram of in situ XAFS setup in gas flow condition under MW heating. The catalyst pellet was placed in the center of the TM010-mode cavity resonator. The X-ray beam passed through the center of the catalyst pellet. The surface temperature of the catalyst pellet (Tex) was monitored by a IR thermometer from the top of the catalyst pellet. b Setup for dehydrogenation of 2-propanol by supported Pt NPs under MW irradiation. Tex represents the temperature of the external surface of the catalyst bed measured by the IR thermometer, while the Tcore indicates the core temperature measured by a fiber-optic thermometer.

Monitoring the temperature distribution of the catalyst

The temperature distribution in the catalyst pellet under MW heating was determined by a thermography and coupled simulation with electromagnetic field and heat transfer modules. High-resolution thermography (Thermalview X MCR32-XA0350-LWD1.25×, ViewOhre Imaging Co. Ltd., detection wavelength; 7.5–14 μm) was used to obtain the temperature mapping of the surface of the catalyst pellet in the 6.4 × 5.12 mm area with a 20 μm resolution with a temperature range of 293–773 K. The emissivity was determined by calibration with the black body reference (emissivity; 0.94). The sample and the black body (the sample sprayed with the black body) were heated at 423 K on a hotplate. The average temperature of the 9 points of the black body was obtained as about 423 K and obtained emissivity of the Pt/Al2O3 as 0.73 and that of the Pt/SiO2 as 0.78 at 423 K. These values were used for all experiments since there was no substantial difference in the emissivity between 373–473 K.

Coupled simulation analysis was conducted by the finite element method (FEM) using COMSOL Multiphysics 5.4a software (RF and Heat Transfer Modules). Physical properties used in the simulation are summarized in Supplementary Table 7. The simulation model and mesh are shown in Supplementary Fig. 14, which were coupled to electromagnetic waves and heat transfer modules to determine the 3D electric field and temperature distributions in the catalyst pellet. The 2D temperature mapping by thermography was reproduced by the simulation analysis by applying effective thermal conductivity (Supplementary Table 7), and the resulting to obtain the 3D temperature distribution in the catalyst bed. More details of the determination of temperature distribution was indicated in the Supplementary Note 3.

Catalytic dehydrogenation 2-propanol

The MW setup is shown in Fig. 9b. Pt/Al2O3 or Pt/SiO2 (0.6–1.0 mm, 200 mg) was packed in a quartz tube reactor (inner diameter of 10 mm). The height of the catalyst bed was 6.0 mm for Pt/Al2O3 and 8.5 mm for Pt/SiO2. 2-Propanol gas (16 mL min−1) was introduced into the quartz tube with Ar carrier gas (40 mL min−1), and the contact time (W F1) for the 2-propanol gas was 0.21 g h L−1. To prevent condensation, the pathway of the 2-propanol gas was heated to 383 K with a ribbon heater. The catalyst bed was heated by CH (electric heating furnace) or MWs. The reaction temperature of the electric furnace (Tokyo Garasu Kikai Co., Ltd.) was controlled with a PID temperature controller equipped with a thermocouple thermometer. For MW heating, another TM010-mode cavity with a Q-factor larger than the one of in situ XAFS system was employed. The Tex and the core temperature (Tcore) of catalyst bed were measured with an IR thermometer and fiber-optic thermometer (FSE-35225 Anritsu Meter Co., Ltd.). After pre-heating under Ar flow, 2-propanol was introduced into the reactor using a micro-feeder when the Tex and the Tcore became constant. The gas and liquid products were collected every 20 min and analyzed by gas chromatography (GC-8A, thermal conductivity detector, Shimadzu Co.) with a Gaskuropack54 column (GL Science Inc.) and GC2014 (Flame ionization detector, Shimadzu Co.) with an Inertcap Pure Wax column (GL Science Inc.), respectively. More details of the catalytic dehydrogenation of 2-propanol was indicated in the Supplementary Note 4.

Operando XANES of the reduction of PtOx/Al2O3 by 2-propanol

A pre-oxidized PtOx/Al2O3 sample was measured in the operando XAFS experiment under 2-propanol gas flow to monitor the PtOx reduction. The operando XANES spectroscopy measurements were conducted by using similar setups applied to in situ XAFS measurement under CH and MW heating, as indicated above (Fig. 9). In the operando measurement under MWs, the catalyst sample was packed in a smaller quartz tube (outer diameter; 9 mm, inner diameter; 7 mm), which was placed inside the larger quartz tube (outer diameter; 12 mm, inner diameter; 10 mm). 2-Propanol gas (5.87 kPa at 298 K) was introduced onto the catalyst bed with an N2 carrier gas (10 mL min−1). QXAFS spectra were obtained every 5 min to trace the progress of the reduction by 2-propanol. According to the previous reports, the Linear Combination Fitting (LCF)42 method was applied to evaluate the progress of the PtOx reduction. The decreases of the whiteline intensities were transformed to the relative reductions of PtOx, where the most reduced one (MW 373 K for 43 min) was 100%. The XANES spectra of the Pt foil (Nilaco Co.) and Sodium Hexahydroxyplatinate (IV) (Fujifilm Wako Pure Chemicals Co.) were also measured as references.