Carbon-Modified Mesoporous Anatase/TiO₂(B) Whisker for Enhanced Activity in Direct Synthesis of Hydrogen Peroxide by Palladium

Abstract

The regulation of the interaction between H₂O₂ and its catalysts is a promising route to achieve high productivity and selectivity towards H₂O₂. Herein, mesoporous anatase/TiO₂(B) whisker (mb-TiO₂) modified with heterogeneous carbon was prepared as the support of Pd-based catalysts for the direct synthesis of H₂O₂. The morphology and structure of the catalyst were investigated by transmission electron microscopy, X-ray diffraction, Raman spectroscopy, Brunner-Emmet-Teller measurements, and X-ray photoelectron spectroscopy. The interaction between H₂O₂ and the support was studied by isothermal calorimeter. The carbon heterogeneous modification can weaken the interaction between H₂O₂ and the support, then accelerate the desorption of H₂O₂ and reduce the re-adsorption of H₂O₂ in the reaction medium. Meanwhile, the synergistic effects between TiO₂ and Pd nanoparticles are not influenced by the heterogeneous carbon distribution. The catalyst exhibits better performance for the synthesis of H₂O₂ compared with the corresponding unmodified catalyst; the productivity of H₂O₂ increases more than 40%, which can be ascribed to the decrease of further H₂O₂ conversion under the weakened interaction.

1. Introduction

As a green and excellent oxidant, hydrogen peroxide (H₂O₂) has been widely applied in the pulp/paper industry, water purification, and chemical synthesis [1]. With the increasing demand for H₂O₂ in the international market, the direct synthesis of H₂O₂ from hydrogen and oxygen is believed to be a promising route because of its remarkable advantages [2,3,4,5,6,7], such as its lower levels of environmental pollution and lower cost. However, three inevitable side reactions in the direct synthesis process, including the combustion of H₂ and the hydrogenation/decomposition of H₂O₂, reduce the productivity and selectivity towards H₂O₂, which is the major drawback of this process [2,4,8].

The surface property of the support has been proved to have an important effect on the performance of H₂O₂ synthesis. For example, the hydrophilic/hydrophobic moderation of the support is an efficient approach to improve the catalytic activity. As early as the 1970s, Pd catalysts supported on more hydrophobic materials were prepared to improve the productivity of H₂O₂ [9]. Hu et al. came to similar conclusions; they speculated further that more hydrophilic support might increase the re-adsorption of H₂O₂, resulting in the further conversion of H₂O₂ and the lower selectivity of H₂O₂ synthesis [8]. In addition, the crystal phase of the support surface can also lead to the enhancement of catalytic performance by the suppression of side reactions. TiO₂ is a promising candidate support for H₂O₂ synthesis and has been widely studied [10,11,12]. According to our previous work with assistance from molecular dynamic simulations [13], it has been indicated that introducing the proper amount of TiO₂(B) phase in the catalyst surface can accelerate the desorption of H₂O₂ from the catalyst due to the weak interaction between H₂O₂ and TiO₂(B), and consequently improve the productivity of H₂O₂ over Pd-TiO₂(B)/anatase catalyst.

In fact, the interaction between H₂O₂ and the support is the origin of the support effect according to the analysis mentioned above, and the selectivity and productivity of H₂O₂ is thus influenced by the moderation of the interaction between H₂O₂ and the support. Carbon-modified TiO₂ has been proved to weaken the adsorption of polar molecules and accelerate their transfer [14]. Meanwhile, the carbon modification method is facile, inexpensive, and minimally changes the structure of supports. Herein, we fabricated a heterogeneous carbon-modified catalyst for the direct synthesis of H₂O₂ on the basis of mesoporous anatase/TiO₂(B) whisker (mb-TiO₂). mb-TiO₂ whisker with excellent performance in various catalytic systems has been successfully prepared in our previous work [15,16,17,18]. Carbon heterogeneous surface modification can regulate the adsorption of H₂O₂ on the catalyst surface without changing the strong synergy between active sites and TiO₂. The experimental results showed that the Pd-supported, carbon-modified mb-TiO₂ (C-mb-TiO₂) catalyst exhibited higher productivity and selectivity towards H₂O₂ compared with the unmodified Pd/mb-TiO₂ catalyst. X-ray diffraction (XRD), Raman spectra, transmission electron microscopy (TEM), X-ray photoelectron spectroscopy (XPS), and isothermal microcalorimeter (IMC) characterization techniques were applied to analyze the structure–performance relationship.

2. Results and Discussion

2.1. Structural Characterization of the Catalysts

Figure 1 shows the XRD and Raman patterns of mb-TiO₂, C-mb-TiO₂, and the corresponding Pd-supported catalysts. Five distinct diffraction peaks are assigned to anatase phase of TiO₂ (JCPDF 21-1272) [19]. Meanwhile, the weak observed peaks at 2θ of 14.3°, 28.7° and 43.6° correspond to (001), (002) and (003) of TiO₂(B) phase (JCPDF 35-0088) [20]. Evidently, no extra diffraction peaks of carbon can be found in C-mb-TiO₂, which implies that either the amount of carbon is too little or the generated carbon is amorphous [21]. In addition, the diffraction peaks for Pd are indistinguishable, which may be caused by the low content or good dispersion, as previously reported [22]. The bicrystalline structure was also confirmed by Raman spectra. Peaks at 234 cm⁻¹, 431 cm⁻¹ and 469 cm⁻¹ are all indexed to the TiO₂(B) phase [13]. The specific crystal structure has been explored in our previous work [15]. As shown in Figure 1b, two Raman bands located at 1300–1450 cm⁻¹ (D-band) and 1500–1650 cm⁻¹ (G-band) are only observed in C-mb-TiO₂ rather than mb-TiO₂, which verifies the existence of carbon species on the surface of C-mb-TiO₂ [8]. The TG analysis shows that the content of carbon species of C-mb-TiO₂ is less than 1 wt % (seen in Figure S2), and only about one third of mb-TiO₂ surface is covered by the introduced carbon species according to our previous calculation [14].

N₂ adsorption-desorption isotherms were measured at 77 K to characterize the physical and chemical properties of supports and corresponding catalysts, as shown in Figure 2. The typical type IV isotherms according to IUPAC classification are observed for all samples, indicating the presence of a mesoporous structure [17]. The corresponding pore size distribution curves (seen in Figure 2) show a narrow distribution ranging from 5 to 9 nm with a maximum around 7 nm for all samples. Details about the BET surface area and average pore volume/size are summarized in

Table 1. Structural properties of the supports and catalysts.

Sample	Surface Area (m2·g−1)	Pore Volume (cm3·g−1)	Average Pore Size (nm)
mb-TiO2	82.2	0.22	10.5
C-mb-TiO2	81.8	0.22	10.4
Pd/mb-TiO2	78	0.20	10.2
Pd/C-mb-TiO2	79	0.19	10.1

. It can be seen that the surface area and pore volume of C-mb-TiO₂ are nearly identical to that of mb-TiO₂, indicating that almost nothing has changed in the initial structure of mb-TiO₂ after the carbon modification. Both the average pore size and the pore volume decrease after loading with Pd, which can be ascribed to the adhesion of metallic Pd nanoparticles inside the pores.

To obtain more information about the surface morphology, the samples were characterized by SEM and HRTEM. Figure S1 shows the SEM images of C-mb-TiO₂, which is fibrous whisker with a uniform width of 300–400 nm, and the morphology shows little change from the mb-TiO₂. The structure of Pd-supported C-mb-TiO₂ was further studied by HRTEM and ED mapping. Well-crystallized TiO₂(B) can be observed in Figure 3a; the nanocrystal with lattice fringes of 0.38 nm is assigned to d₀₀₃ of TiO₂(B) facet, and the lattice fringe spacing of 0.35 nm corresponds to d₁₀₁ of anatase [15], with the majority being the anatase phase. Furthermore, we find that the evident anatase phase is covered by a thin area without crystal lattice fringes (see the two yellow dot lines in Figure 3b), which can be ascribed to the generated amorphous carbon species [21], and is consistent with the results of the Raman analysis. Figure 3c shows that the Pd particles are closed to the carbon and anatase phases, but it is difficult to obtain the accurate positioning of Pd from HRTEM images. The loading of Pd is confirmed by elemental mapping, as shown in Figure 3d. The blue regions represent the presence of Pd on the surface of TiO₂ in the TEM maps. The relatively low levels of carbon cannot be distinguished due to the external introduction of the carbon source during the testing process. The average particle size of Pd on C-mb-TiO₂ (11.2 nm) is very close to that on mb-TiO₂ (10.5 nm), indicating that the size effect of the Pd particles can be ignored.

Figure 4 shows the Pd3d core-level spectra of the fresh catalysts. The binding energies at 336.6 and 341.9 eV are ascribed to Pd²⁺3d_5/2 and Pd²⁺3d_3/2, respectively [23]. The binding energy position of Pd⁰ and Pd²⁺ is unchanged between the two catalysts. The presence of Pd²⁺ in the fresh reduced sample is caused by TiO₂-assisted partial oxidation of Pd⁰ when exposed to air. As shown in Table S1, the ratio of Pd²⁺/Pd⁰ of the two catalysts is very similar, indicating that the property of active sites is not affected by the carbon heterogeneous modification [11,13]. According to the literature focusing on carbon riveted supported catalysts, the electronic structure of the supported metal would be altered obviously when the generated carbon species are in intimate contact with metallic particles [24,25]. Therefore, it can be deduced that most of the Pd particles deposit onto the TiO₂ phase.

2.2. The Adsorption Properties of H₂O₂ on the Supports

Isothermal calorimeter was used to investigate the adsorption properties of H₂O₂ on the supports. The adsorption thermodynamic data was calculated from the calorimetric profile recorded through in situ monitoring; the actual amount of adsorbed H₂O₂ was confirmed by the colorimetric method. Finally, their ratio (the adsorption energy of H₂O₂) can be used to represent the interaction strength between H₂O₂ and the support. The detailed process is described in Electronic Supplementary Information (ESI). Figure 5 shows the data of the adsorption properties of H₂O₂ on the support. It can be seen that the adsorption heat of H₂O₂ on C-mb-TiO₂ is obviously lower than that on mb-TiO₂, meaning that the adsorption of H₂O₂ on C-mb-TiO₂ is weaker. The value of adsorption heat of H₂O₂ indicates a reversible physical adsorption. In other words, the generated H₂O₂ is more easily desorbed from the C-mb-TiO₂ surface and H₂O₂ in the reaction medium is more difficult to re-adsorb on the support.

2.3. The Catalytic Activity of the Catalysts

The direct synthesis of H₂O₂ from H₂ and O₂ over Pd/mb-TiO₂ and Pd/C-mb-TiO₂ catalysts were investigated at 283 K and ambient pressure. As shown in Figure 6a, the conversion of H₂ shows little variation over time, while the selectivity and productivity of H₂O₂ decrease along with the time for both catalysts. The Pd/C-mb-TiO₂ catalyst exhibits better activity for H₂O₂ synthesis. The conversion of H₂ over Pd/C-mb-TiO₂ is approximately 35% higher than that over Pd/mb-TiO₂. The Pd/mb-TiO₂ catalyst shows H₂O₂ productivity of 1933 and 1717 mmol/g_Pd/h after 0.5 h and 1h, respectively, while the corresponding productivity on Pd/C-mb-TiO₂ is 2848 and 2411 mmol/g_Pd/h with an improvement of more than 40%. Meanwhile, the H₂O₂ selectivity over Pd/C-mb-TiO₂ is also promoted by 9% when compared with that over Pd/mb-TiO₂. The analysis above indicates that the catalyst performance is significantly improved due to the carbon modification. Notably, the catalytic performance of both the Pd/mb-TiO₂ and Pd/C-mb-TiO₂ are superior to the Pd/P25 catalyst [4,8,11], which is widely investigated for the direct synthesis of H₂O₂.

To further unveil the effect of the support, we also performed the decomposition and hydrogenation of H₂O₂ under N₂ and N₂/H₂ atmospheres, respectively, which are the two main side reactions in the direct synthesis of H₂O₂. As illustrated in Figure 6b, neither of the catalysts showed activity for the decomposition of H₂O₂ in pure N₂ flow. When the H₂/N₂ mixture was fed into the reactor, the H₂O₂ was catalytically hydrogenated to form H₂O, demonstrating that the hydrogenation of H₂O₂ is mainly responsible for the productivity and selectivity in this system. Similar results have also been observed for Pd/P25 [26]. Furthermore, it is observed that Pd/C-mb-TiO₂ exhibits much lower hydrogenation rate than Pd/mb-TiO₂. Willock et al. suggested that the production of H₂O₂ takes place at the interface between the particle and TiO₂ [27]. Therefore, it is reasonable for us to assume that the properties of both active sites and supports have great impact on the catalytic synthesis of H₂O₂. Additionally, in our previous work, we confirmed that the accelerated desorption of H₂O₂ from the support can decrease the hydrogenation rate of H₂O₂ and then improve the selectivity towards H₂O₂ over the corresponding catalyst [13]. As mentioned above, the presence of carbon species can significantly weaken the interaction between H₂O₂ and the catalyst under the premise of similar property of active sites. Therefore, it can be concluded that the weaker interaction can reduce the concentration of H₂O₂ on the surface of the catalyst and then decrease the hydrogenation rate of H₂O₂. In addition, the improvement of H₂O₂ desorption can provide more reactive sites for H₂O₂ formation. As a result, the enhanced selectivity of H₂O₂ and the conversion of H₂ improve the productivity of H₂O₂.

3. Materials and Methods

3.1. Synthesis of Carbon-Modified TiO₂ Whisker

First, the mesoporous and bicrystalline TiO₂ whisker was prepared according to our previous report [15]. Briefly, K₂Ti₂O₅ was suspended in 0.1 M H₂SO₄ solution with vigorous stirring to form H₂Ti₂O₅. The suspension was filtered and washed, following which it was dried at 353 K under vacuum and then calcinated at 723 K for 2 h in muffle to obtain mesoporous and bicrystalline TiO₂ whisker. The final catalyst was denoted as mb-TiO₂.

The carbon-modified mb-TiO₂ was prepared as follows [14]. 1 g of mb-TiO₂ was dispersed in 20 mL of dichloromethane containing 0.377 g of benzenephosphonous acid. After stirring for 24 h, the precipitation was filtered and washed with water/acetone (50/50 by volume) solution, and then dried in an oven at 373 K. The dried sample was treated at 673 K in Ar gas flow for 4 h and denoted as C-mb-TiO₂.

3 wt % Pd-supported catalysts were prepared by incipient wetness impregnation. A certain amount of PdCl₄²⁻ solution was slowly dropped into the support under stirring. The paste was left in natural environment for 4 h, following which it was dried at 383 K overnight. Prior to the reaction, all the catalysts were reduced at 573 K in H₂ (40 mL·min⁻¹) for 1 h. For comparison, commercial titania (P25, Degussa Co., Frankfurt am Main, Germany) was loaded with Pd using the same method.

3.2. Catalyst Characterization

X-ray diffraction (XRD) Spectrogram was collected on a D8 Advance X-ray diffractometer operating at a 40 kV with a current of 100 mA with Cu-Kα radiation. The data was collected in the 2θ range from 5° to 60° at a rate of 0.2 s step⁻¹. Raman spectroscopy (Horiba HR 800) with a multichannel air-cooled CCD camera detector was employed to determine the surface crystal phases of supports using a 514.5 nm He-Cd laser beam. Nitrogen adsorption/desorption at −196 °C were achieved by using the TriStarII 3020M machine (Micromeritics Instrument Co., Atlanta, GA, USA). Surface area was calculated by the BET method; pore volume was determined by N₂ adsorption at a relative pressure of 0.99. The Pd content of the catalysts was confirmed by ICP-AES (PerkinElmer Co., Waltham, MA, USA). SEM for the morphology of catalyst samples was taken by an America Quanta200 environmental scanning electron microscope. The information of dispersion and size of Pd nanoparticles were obtained by high-resolution transmission electron microscopy (HRTEM) performed on a Japan JEM-2100 at 200 kV. XPS was recorded by an AXIS UltraDLD spectrometer with the monochromatic Al Ka X-ray source. The interaction between H₂O₂ and the support was characterized on a TAM air Calorimeter from TA Instruments.

3.3. Catalytic Reaction

The direct synthesis of H₂O₂ was performed under atmosphere pressure in a glass tri-phase reactor. The reagent gases (H₂/O₂/N₂ = 9:36:15) with a total flow rate of 60 mL/min were premixed in a steel mixer, and then introduced into the reactor via a frit fixed at the bottom. The temperature of 60 mL slurry containing 50 mg catalyst and 0.38 mL H₂SO₄ was controlled at 10 °C through the water circulated jacket. Vigorous stirring (950 rpm) was adopted to minimize mass transfer resistance. The outlet of the reactor was linked to a gas chromatography equipped with a 5A molecular sieve packed column and a thermal conductivity detector (SP-6890, Rainbow Chemical Instrument Co., Jinan, China), which helped to continuously analyze the conversion of H₂. The concentration of H₂O₂ in the medium was determined by the colorimetric method after the complexation with a TiOSO₄/H₂SO₄ reagent using a UV-vis spectrophotometer (UV-2802S, Unico (Shanghai) Instrument Co., Shanghai, China) at a wavelength of 400 nm. The selectivity of H₂O₂ ($S_{{\mathrm{H}}_2{\mathrm{O}}_2}$) was calculated from the following formula:

$$S_{{\mathrm{H}}_2{\mathrm{O}}_2}=\frac{\mathrm{rate}\mathrm{of}{\mathrm{H}}_2{\mathrm{O}}_2\mathrm{formation}(\mathrm{mol}/\min )}{\mathrm{rate}\mathrm{of}{\mathrm{H}}_2\mathrm{conversion}(\mathrm{mol}/\min )}\times 100\%$$

(1)

The reaction of H₂O₂ hydrogenation and decomposition was also implemented. The process was similar to that described above except for the initial composition of reaction medium. H₂O₂ was added to ethanol solution with the concentration of 0.25 wt %, then the decomposition and hydrogenation of H₂O₂ were tested under a flow of pure N₂ (60 mL/min) and a flow of H₂/N₂ (9:51 mL/min), respectively. The same tests using bare supports or the absence of catalysts were also performed under the conditions mentioned above.

4. Conclusions

In this work, a simple heterogeneous carbon modification was applied to introduce carbon species on the surface of mb-TiO₂. Pristine mb-TiO₂ and carbon-modified mb-TiO₂ were used as the supports of Pd-based catalysts for the direct synthesis of H₂O₂. The interaction between H₂O₂ and the supports was characterized by isothermal microcalorimetry. It was demonstrated that the carbon-decorated TiO₂ exhibited weaker interaction with H₂O₂ than that achieved with the pristine TiO₂. Therefore, the carbon-modified catalyst can effectively accelerate the desorption of H₂O₂ from the catalyst surface and reduce the re-adsorption of H₂O₂ in the reaction medium, which benefits the decrease of side reaction and the improvement of H₂O₂ productivity. Meanwhile, the properties of metallic active sites are not influenced by the carbon modification. As a result, the Pd/C-mb-TiO₂ shows better performance for the direct synthesis of H₂O₂ compared with Pd/mb-TiO₂. More than 40% increase in the productivity of H₂O₂ was achieved, and the selectivity of H₂O₂ improved by 9%. In addition, the catalytic performance of both Pd/mb-TiO₂ and Pd/C-mb-TiO₂ are superior to Pd/P25, which is widely investigated for the direct synthesis of H₂O₂. Hence, heterogeneous carbon modification represents a promising approach to tune the properties of the catalysts, as demonstrated here with TiO₂-supported Pd catalyst for H₂O₂ synthesis.