The Effects of AlPO-n Additives as Catalytic Support on Pd-Catalytic Hydrogenation of 2-Amylanthraquinone Process

Abstract

The Pd-catalyzed hydrogenation of anthraquinone to synthesize hydrogen peroxide is an important process in the chemical industry. A Pd catalyst with high dispersion is the key to hydrogenation activity and selectivity. For the first time, this team introduced the AlPO-5 zeolite to SiO₂ powder to prepare a finely dispersed Pd catalyst with higher efficiency than the conventional Pd/SiO₂ catalyst. Based on previous research, other aluminophosphate molecular sieves (AlPO-n) with different properties from AlPO-5 were introduced to the SiO₂ support, and then the synthesized hydrogenation catalysts were characterized by the BET, XRD, H₂-TPR, NH₃-TPD, H₂-TPD, XPS and HRTEM methods and then tested for the hydrogenation of 2-amylanthraquinone in a continuous stirred tank reactor (CSTR). It was demonstrated that the Pd-AlPO-31/SiO₂ catalyst exhibits superior H₂O₂ yield (8.4 g·L⁻¹) and selectivity (96%) among all prepared Pd-AlPO-n/SiO₂ and Pd/SiO₂ catalysts. The characterization results suggest that the dimensions and structure of AlPO-31 micropore channels are responsible for highly dispersing Pd particles, preventing the accumulation of Pd particles, and intensifying H₂ adsorption in its micropore channels, which is significant for the catalytic activity.

1. Introduction

Hydrogen peroxide, as an environmentally friendly oxidant, is vastly produced by the anthraquinone (AQ) process in the chemical industry [1,2,3]. H₂ molecules first gain access to the surface of Pd particles, and then they are dissociated by Pd. The generated H• spills over onto the oxide [4]. Two carbonyl groups on AQ capture H• to yield anthrahydroquinone (AQH₂), which can be oxidized to produce H₂O₂. During the process of the hydrogenation of 2-amylanthraquinone (AAQ) shown in Scheme 1 [5,6], several byproducts caused by deep hydrogenation, such as tetrahydro-amylanthrahydroquinone (H₄AAQH₂), octahydro-amylanthrahydroquinone (H₈AAQH₂) and anthrone (AAN), are generated, since AAQ and AAQH₂ absorbed on the Pd surface are degraded. H₄AAQH₂ is considered an effective product for generating H₂O₂. Furthermore, the formation of amyloxoanthrone (OXO) is favored under acidic conditions due to the tautomerization equilibrium [7,8].

Palladium-supported catalysts are commonly used in the anthraquinone hydrogenation process due to their high activity, selectivity and stability [9], and the support is usually silica or alumina, but the acidity of alumina leads to the increased production of byproducts [10]. Several methods have been proposed to avoid the negative effects of its acidity, such as modification with Na₂SiO₃, NaH₂PO₄ or other alkali metal salts [11,12]. Li et al. compared the performance of Pd/SiO₂/Cordierite with that of Pd/Al₂O₃/Cordierite. It was found that Pd/SiO₂/Cordierite shows higher selectivity and stability due to its weak surface acidity and regular structure [8]. As to the preparation of Pd/SiO₂, one of the drawbacks of using silica as a support is usually the dispersion of Pd particles due to the weak connection between Pd particles and the silica support, which makes the modification of silica necessary.

Novel aluminophosphate molecular sieves (AlPO-n) are similar to zeolites in some properties and have been utilized as catalysts or catalyst supports [13,14]. For the first time, our group [15] prepared Pd catalysts supported on an AlPO-5/SiO₂ complex support, which exhibited better activity and selectivity than the conventional Pd/SiO₂ catalyst for the 2-ethylanthraquinone hydrogenation process owing to the Pd-O-P link. However, the mechanism by which AlPO-5 disperses Pd particles remains confusing, as it is not reasonable that defects such as Pd-O-P exist on synthesized AlPO-5.

Yu et al. [16] used ESR and ESEM spectroscopic methods to deduce the locations of Pd(II) in the micropores of SAPO-5 and SAPO-11 molecular sieves in the presence of various adsorbates. The results indicated that Pd(II) enters the micropore channels of SAPO-5 and SAPO-11, which confines the sintering of Pd particles. Therefore, it is reasonable to conclude that AlPO-n zeolites with similar dimensions and structures increase the dispersion of Pd particles on AlPO-n/SiO₂ complex supports by the same method, and Pd particles in AlPO-n micropore channels are able to reduce the degradation of AAQ and AAQH₂, since the sizes of these molecules are too big to enter the micropore channels.

Recently, Pd catalysts for anthraquinone hydrogenation for H₂O₂ production were synthesized and discussed. Li et al. [17] loaded Pd on P-doped alumina (Pd/Al₂O₃-xP) and applied it to H₂O₂ production, showing that Pd/Al₂O₃-1P, containing 1 wt.% phosphorus, exhibited the best catalytic performance and reusability. Chen et al. [18] used a self-made spherical alumina support to distribute fine Pd particles for hydrogenating anthraquinone. Liang et al. [19] compared a Pd/AC catalyst and a Pd-Ce/AC catalyst, indicating that the addition of Ce not only reduced the diameter of Pd particles but also enhanced the content of Pd²⁺ in the final catalyst. Belykh et al. [20] discovered that Pd-P nanoparticles are more active in the hydrogenation of 2-ethyl-9,10-anthraquinone than Pd₆P crystalline phosphide.

In this work, in order to explore the effects of the structure of AlPO-n on hydrogenation catalysts, other AlPO-n, such as AlPO-8 with 14-ring pores, AlPO-11 with 10-ring pores and AlPO-31 with 12-ring pores, with different properties from AlPO-5, were introduced to prepare AlPO-n/SiO₂ complex supports. The synthesized Pd-AlPO-n/SiO₂ catalysts were characterized by the BET, XRD, H₂-TPR, NH₃-TPD, H₂-TPD, XPS and HRTEM methods and then tested for the hydrogenation of 2-amylanthraquinone in a CSTR. The impacts of the structure of AlPO-n on Pd dispersion and H₂ adsorption are discussed.

2. Results and Discussion

2.1. Characterization of the Supports and Catalysts

The XRD patterns of synthesized aluminophosphate molecular sieves are shown in Figure 1, which match those of AlPO-5, AlPO-8, AlPO-11 and AlPO-31 in the literature [21,22,23,24]. Compared with the simulated standard molecular sieve, the characteristic peaks of synthesized samples of AlPO-5, AlPO-8, AlPO-11 and AlPO-31 are consistent. X-ray analysis of the investigated samples (Figure 1) confirmed that the adjusted parameters and volumes of the elementary cells of synthesized aluminophosphate zeolites (

Table 1. Parameters of the structure of an elementary cell of aluminophosphate zeolites.

Sample	a, Å	b, Å	c, Å	Ve.c, Å
AlPO-5	13.706	13.705	8.470	1377.535
AlPO-8	33.162	14.814	8.376	4114.809
AlPO-11	13.369	18.710	8.451	2113.882
AlPO-31	20.830	20.830	4.999	1878.340

) correspond to standard values for these materials with the allowance for errors. It should be noted, however, that in the AlPO-8 sample, there were traces of VPI-5 material, the synthetic precursor of AlPO-8, upon final thermal treatment. In addition, in Figure 1, the simulated XRD spectra of AlPO-11 are shifted. The simulated XRD spectrum of AlPO-11 contains all possible crystal faces by full-spectrum fitting, as ideal AlPO-11 shows a hexagonal, rod-like morphology. However, synthesized AlPO-11 shows a more cubic-like and less hexagonal morphology, indicating that a crystal face shows less reflection on XRD spectra at 7°. The synthesized samples of AlPO-5, AlPO-8, AlPO-11 and AlPO-31 exhibit similar typical diffraction peaks of AFI, AET, AEL and ATO (the names of the structures of AlPO-5, AlPO-8, AlPO-11, AlPO-31, respectively) topologies.

The XRD patterns of the prepared catalysts are presented in Figure 2. Typical diffraction peaks of the catalyst samples Pd-AlPO-5/SiO₂, Pd-AlPO-8/SiO₂, Pd-AlPO-11/SiO₂ and Pd-AlPO-31/SiO₂ are consistent with the AFI, AET, AEL and ATO topologies, respectively, which confirms that the catalysts contained the AlPO-n zeolites. However, during the preparation of AlPO-n zeolites and AlPO-n/SiO₂, especially during the preparation of AlPO-8 and AlPO-31 (along with AlPO-8/SiO₂ and AlPO-31/SiO₂), their structures easily change into AlPO-11, as the components of their precursors are the same. Meanwhile, there is no characteristic peak of Pd species appearing in the catalysts, indicating that the Pd species, which exist on the surfaces or in the pores of zeolites, have high dispersion. In addition, the Pd content of the catalyst is 0.2 wt.%, which makes it difficult to characterize by XRD.

The N₂ adsorption isotherms of the catalysts (Pd-AlPO-5/SiO₂, Pd-AlPO-8/SiO₂, Pd-AlPO-11/SiO₂ and Pd-AlPO-31/SiO₂) were performed, and the results are shown in Figure 3. All isotherms were Type VI because the main component of the support was mesoporous SiO₂ (over 80%). However, a Type I isotherm with a remarkable nitrogen uptake at low P/P₀ for all samples was found, which is a typical signature of microporous AlPO-n zeolites [23]. The nitrogen uptake continues to greatly increase in a high P/P₀ range, and an H₄ hysteresis loop [24] is measured, which indicates that the part with the largest presence was the textural mesoporosity originating from the silica sol. It was also shown that the sorption volume of the mesopore was much larger than that of the micropore because the mass ratio of the AlPO-n zeolite to the catalyst was only 20 wt.%. The pore properties of AlPO-5, AlPO-8, AlPO-11 and AlPO-31 are listed in

Table 2. The pore properties of the AlPO-5, AlPO-8, AlPO-11 and AlPO-31 zeolites.

No.	AlPO-n Zeolite	S* (m2·g−1)	VS (mL·g−1)	Vμ (mL·g−1)	Dmicro (nm)
1	AlPO-5	48	0.276	0.111	0.73
2	AlPO-8	145	0.25	0.078	0.83
3	AlPO-11	35	0.11	0.051	0.49
4	AlPO-31	57	0.25	0.038	0.53

S* is the specific surface of the mesopore.

[25,26], in which the volumes of the micropores of the zeolites are much less than those of the catalysts (the sorption volumes of Pd-AlPO-5/SiO2, Pd-AlPO-8/SiO₂, Pd-AlPO-11/SiO₂, and Pd-AlPO-31/SiO₂ were 0.69 cc/g, 0.65 cc/g, 0.60 cc/g and 0.53 cc/g).

The reducibility of the catalysts was investigated by H₂-TPR, and the results are shown in Figure 4. There is only one reduction peak for each one from −20 °C to 200 °C, attributed to the reduction of PdO species supported on SiO₂ after calcination in this case. The H₂-TPR spectra of Pd-AlPO-11/SiO₂ and Pd-AlPO-31/SiO₂ illustrate that the PdO loaded on the two complex supports was reduced at 25 °C, which is lower than the reduction temperature of PdO of other catalysts, indicating that sizes of the Pd particles loaded on AlPO-11/SiO₂ and AlPO-31/SiO₂ are smaller and more finely dispersed. In addition, there are obviously two reduction peaks or broad peaks in the samples of Pd-AlPO-5/SiO₂, Pd-AlPO-8/SiO₂ and Pd/SiO₂, while smaller Pd particles were reduced at 145 °C, and larger Pd particles were reduced at 165 °C. The broader peaks at 165 °C suggest that the Pd particle-size distribution is wider and the Pd dispersion is poorer, which is consistent with the TEM and CO-pulse chemisorption results in

Table 3. Pd loading, dispersion and specific surface area of all catalysts.

Sample	Pd Loading (wt.%)	Pd Dispersion (%)	Specific Surface Area (m2/g)
Pd/SiO2	0.1964	5.21	185.0
Pd-AlPO-5/SiO2	0.1971	7.33	187.7
Pd-AlPO-8/SiO2	0.1952	7.52	187.5
Pd-AlPO-11/SiO2	0.1960	18.45	186.4
Pd-AlPO-31/SiO2	0.1970	20.20	186.2

.

Furthermore, the H₂-TPR spectra of PdO on Pd-AlPO-11/SiO₂ and Pd-AlPO-31/SiO₂ exhibit more intense reduction peaks than others, which matches well with the TEM images of catalysts. In addition, there is no strong interaction link of Pd-O-P, and the reduction temperatures of PdO on all samples are supposed to be relatively high, according to the literature [26,27], which is clear evidence of the high crystallinity of synthesized AlPO-n.

With the assistance of HRTEM analyses, the images of the catalysts and the size distributions of Pd particles are shown in Figure 5 and Figure 6. Pd particles are located on SiO₂, as well as AlPO-n zeolites. The average Pd particle sizes of Pd-AlPO-11/SiO₂ and Pd-AlPO-31/SiO₂ are 1.3 nm and 1.4 nm, which are so small that the HRTEM method is only able to locate Pd particles’ positions. The particles of the other samples are larger than 5 nm. The Pd dispersion measured by CO-pulse chemisorption is shown in Table 3, which matches well with the HRTEM results. This proves that the different structure of AlPO-n correlates with the size of Pd particles. In addition, it identifies the fact that the sizes of micropore channels in AlPO-n are different [26]. Except for SiO₂ in the support, the chemical properties of AlPO-n zeolites surface are the same, meaning that only their micro-channel structure causes the differences in Pd dispersion among all of the catalysts.

To compare the supports’ H₂ desorption ability, which affects their catalytic selectivity, H₂-TPD analysis was performed, and the results of all samples are shown in Figure 7. The H₂ desorption temperature of Pd-AlPO-31/SiO₂ is 15 °C, which is lower than that of Pd/SiO₂ and other Pd-AlPO-n/SiO₂ catalysts, exhibiting a superior H₂ desorption ability. This proves that the bonding force of H₂ in AlPO-31′s structure is relatively weak.

In order to exclude the possible influence of the Lewis acid on the support, which may be introduced by the residual aluminum phosphate because of low crystallinity, the catalysts were characterized by NH₃-TPD spectra, shown in Figure 8. Typically, there is only one peak with high intensity on each curve, which is attributed to weak acid sites. This result is in accordance with the investigation by Guo et al. [28], which is attributed to the Si-OH group on the surface of the modified catalyst, showing that finely dispersed Pd particles on AlPO-11/SiO₂ and AlPO-31/SiO₂ are not due to residual aluminum phosphate.

XPS measurements were conducted to further explore the electronic effect of the AlPO-n zeolite on the electronic structure of the catalysts, which are shown in Figure 9, and the type of background is Shirley. The binding energies of 335.1 eV and 340.4 eV are assigned to Pd⁰ 3d5/2 and Pd⁰ 3d3/2. The other peaks at 336.1 eV and 341.4 eV belong to Pd 3d5/2 and Pd 3d3/2, corresponding to Pd²⁺ [29]. It is worth noting that the ratios of Pd⁰/Pd²⁺ were 1.99, 2.45, 3.02, 1.31 and 2.68 for Pd/SiO₂, Pd-AlPO-5/SiO₂, Pd-AlPO-8/SiO₂, Pd-AlPO-11/SiO₂ and Pd-AlPO-31/SiO₂, respectively, which indicates that the reduction difficulty order of the palladium precursor is Pd-AlPO-8/SiO₂, Pd-AlPO-31/SiO₂, Pd-AlPO-5/SiO₂, Pd/SiO₂ and Pd-AlPO-11/SiO₂. In the anthraquinone hydrogenation reaction, a higher Pd⁰/Pd²⁺ ratio means that a greater proportion of reduced palladium states on the surface of the catalyst is formed under the same reducing conditions, which is able to improve the hydrogenation efficiency [30,31]. This is the very reason that Pd-AlPO-31/SiO₂ showed higher activity than others.

2.2. Catalytic Performance

The catalytic activity and selectivity of Pd/SiO₂ and Pd-AlPO-n/SiO₂, with AlPO-n/SiO₂ as a control group, were evaluated, and the results are shown in Figure 10 and Figure 11. AlPO-n/SiO₂ showed no catalytic activity in the hydrogenation of anthraquinone, as only Pd is able to activate hydrogen molecules in the reaction. Therefore, the Pd dispersion of catalysts plays a key role in the catalytic activity. In Figure 10, Pd-AlPO-31/SiO₂ shows the highest hydrogenation efficiency among the catalyst samples, and its value reaches 8.4 g/L. This is partially due to the fact that the Pd particles of Pd-AlPO-31/SiO₂ are smaller than those of Pd-AlPO-5/SiO₂, Pd-AlPO-8/SiO₂ and Pd/SiO₂, as demonstrated in Figure 5. Decreasing the size of the Pd particles alters the surface’s atomic coordination structures, such as by increasing the specific surface area and introducing unsaturated coordination sites. When the particle size decreases, the defects on the catalyst surface increase. Therefore, it improves the ability of catalytic hydrogen molecule activation. However, Pd-AlPO-11/SiO₂, with a smaller Pd particle size, had lower hydrogenation activity than Pd-AlPO-31/SiO₂ and even Pd-AlPO-5/SiO₂ and Pd/SiO₂, whose Pd particles sizes are quite larger than Pd-AlPO-11/SiO₂.

In Figure 11, Pd-AlPO-31/SiO₂ exhibits higher selectivity than Pd/SiO₂ and other Pd-AlPO-n/SiO₂ catalysts, where the selectivity of Pd-AlPO-31/SiO₂ is 96%. Typically, the selectivity of Pd-AlPO-11/SiO₂ is 65%, equivalent to that of Pd-AlPO-8/SiO₂, making it reasonable that Pd-AlPO-11/SiO₂ with a small particle size presented a low hydrogenation efficiency. During the evaluation, the activity of Pd-AlPO-5/SiO₂ reached 7.8, which is the second-highest one among all catalysts. Even though the size of Pd particles on AlPO-5/SiO₂ is the largest (average of 6.1 nm), based on the TEM images shown in Figure 5 and Figure 6, the catalytic selectivity of Pd-AlPO-5/SiO₂ is fairly high (90.3%), resulting in its considerably high catalytic activity.

According to Yu et al. [32], the adsorption of molecules such as oxygen, water, benzene, ammonia, carbon monoxide, pyridine and hydrazine causes the migration of Pd ions to a position inside the micropores of SAPO-5 and SAPO-11. In that case, Pd(NH₃)₄²⁺ migrates into the structure of AlPO-n, where the aggregation of Pd particles is suppressed. As the diameter of the AlPO-31 structure is 0.53 nm, while that of AlPO-11 is 0.39 nm–0.63 nm, both of them optimize the effect of suppression on Pd particle migration. During the reaction, molecules of both H₂ and 2-amylanthraquinone are adsorbed on the surface of Pd nanoparticles located on SiO₂ [5]. Only H₂ molecules are able to transport into the micro-channels of AlPO-n [25], where they are dissociated by Pd particles inside AlPO-n micropore channels. The dissociated H• effuses from AlPO-n micropore channels and then combines with 2-amylanthraquinone on the surface of Pd nanoparticles located on SiO₂.

Moreover, the better catalytic activity of Pd-AlPO-31/SiO₂ is attributed not only to the high dispersion of Pd by AlPO-31, matching the results of HRTEM and CO-pulse analyses, but also to H₂ adsorption in aluminophosphate molecular sieves, which is determined by the micropore characteristics of AlPO-n. According to Grenev et al., for the independence of principal sorption processes and the presence of various sizes of structures, the adsorption isotherms of H₂ on the AlPO-5 zeolite are higher than those of H₂ on others, which means that the micropore channels of AlPO-5 are able to adsorb more H₂, which will be dissociated by Pd particles inside AlPO-5 to generate more H∙, migrating to Pd on SiO₂ to hydrogenate 2-amylanthraquinone. Therefore, the catalytic activity of Pd-AlPO-5/SiO₂ is higher than the other catalysts’ activity, except for Pd-AlPO-31/SiO₂. The local isotherm of the density of adsorbed H₂ on AlPO-31 is higher than that of the others, meaning that adsorbing a unit volume of H₂ inside AlPO-31 is the most effective, even though the adsorption isotherm of H₂ on the AlPO-31 zeolite is only higher than that on AlPO-8 [25,26]. For the AlPO-8 zeolite, both the adsorption isotherms of H₂ and the local isotherm of the density of adsorbed H₂ are the lowest, reducing its potential to dissociate H₂, which determines its low catalytic activity.

However, the potential production of deeply hydrogenated byproducts by absorbed H₂ should be avoided in the whole process. According to the H₂-TPD spectra of different catalysts shown in Figure 7, the absorbed H₂ on Pd-AlPO-31/SiO₂ was desorbed at 150 °C, while that on the others was desorbed at a higher temperature, which proves that the bonding force of H₂ in AlPO-31′s structure is relatively weak. In fact, not only does the structure of AlPO-31 suppress the aggregation of Pd particles, but the adsorption of H₂ on catalysts’ pore structure is also restrained, which shifts the thermodynamic equilibrium to not form byproducts. On the contrary, the H₂ desorption ability of AlPO-11 is relatively feeble, resulting in the production of massive amounts of deeply hydrogenated byproducts in the reaction, which is disadvantageous for reaction selectivity.

3. Experimental Section

3.1. Catalyst Preparation

• Preparation of AlPO-5. The molar ratio of the starting gels was 1.2 TEA: Al₂O₃: P₂O₅: 50 H₂O. The gels were prepared by adding pseudoboehmite (Sasol, Sandton, South Africa) to a solution of phosphoric acid (Sinopharm Chemical Reagent Co., Shanghai, China) and stirred for 3 h. Tetraethylamine (TEA, Sinopharm Chemical Reagent Co., Shanghai, China) was added to the mixture and stirred for 24 h. The resulting gels were introduced into Teflon-lined stainless-steel autoclaves and heated statically at 200 °C (with a heating rate of 2 °C·min⁻¹) for 36 h. After hydrothermal treatment, the zeolite was separated from the mother liquor, rinsed repeatedly with distilled water and dried. The dried zeolite was finally calcined at 600 °C (with a heating rate of 2 °C·min⁻¹) for 6 h.
• Preparation of AlPO-8. The molar ratio of the starting gels was DPA: Al₂O₃: P₂O₅:34 H₂O. The gels were prepared by adding pseudoboehmite to a solution of phosphoric acid and stirred for 3 h. Then, Dipropylamine (DPA, Sinopharm Chemical Reagent Co., Shanghai, China) was added to the mixture and stirred for 24 h. The resulting gels were introduced into Teflon-lined stainless-steel autoclaves and heated statically at 144 °C (with a heating rate of 2 °C·min⁻¹) for 12 h. After hydrothermal treatment, the zeolite was separated from the mother liquor, rinsed repeatedly with distilled water and dried. The dried zeolite was finally calcined at 600 °C (with a heating rate of 2 °C·min⁻¹) for 6 h.
• Preparation of AlPO-11. The molar ratio of the starting gels was 4 DPA: Al₂O₃: P₂O₅:40 H₂O. The gels were prepared by adding pseudoboehmite to a solution of phosphoric acid and stirred for 3 h. Then, Dipropylamine (DPA, Sinopharm Chemical Reagent Co., Shanghai, China) was added to the mixture and stirred for 24 h. The resulting gels were introduced into Teflon-lined stainless-steel autoclaves and heated statically at 200 °C (with a heating rate of 2 °C·min⁻¹) for 24 h. After hydrothermal treatment, the zeolite was separated from the mother liquor, rinsed repeatedly with distilled water and dried. The dried zeolite was finally calcined at 600 °C (with a heating rate of 2 °C·min⁻¹) for 6 h.
• Preparation of AlPO-31. The molar ratio of the starting gels was 4 DPA: Al₂O₃: P₂O₅:40 H₂O. The gels were prepared by adding pseudoboehmite to a solution of phosphoric acid and stirred for 3 h. Then, Dipropylamine (DPA, Sinopharm Chemical Reagent Co., Shanghai, China) was added to the mixture and stirred for 24 h. The resulting gels were introduced into Teflon-lined stainless-steel autoclaves and heated statically at 200 °C (with a heating rate of 2 °C·min⁻¹) for 48 h. After hydrothermal treatment, the zeolite was separated from the mother liquor, rinsed repeatedly with distilled water and dried. The dried zeolite was finally calcined at 600 °C (with a heating rate of 2 °C·min⁻¹) for 6 h.
• Preparation of complex supports. First, 12 g of silica sol (ludox-30) and 1 g of AlPO-n were mixed and injected into an oil column at 90 °C, in which the spherical complex supports were generated and then dried in a vacuum oven at 60 °C for 10h. The supports were calcinated at 700 °C (with a heating rate of 2 °C·min⁻¹) for 12 h, while they were used at 120–200 mesh to prepare catalysts.
• Preparation of catalysts. A total of 2 g of the prepared support was impregnated in 4 mL of a Pd(NH₃)₄(NO₃)₂ aqueous solution (1.00 mg·ml⁻¹, calculated by the quality of Pd) for 2 h, followed by drying and calcination at 400 °C (with a heating rate of 2 °C·min⁻¹) for 2 h. All catalysts were reduced by H₂ at 120 °C for 2 h before the reaction.

3.2. Characterization of the Catalysts

The phases of synthesized AlPO-n zeolites and AlPO-n/SiO₂ complex supports were analyzed by X-Ray Diffraction in an X’pert Pro-1 instrument (PANalytical B.V., Almelo, The Netherlands), with Cu Kα radiation (λ = 1.5418 Å) operated at 40 kV and 40 mA. The patterns were scanned from 5 to 45°, with a step size of 0.003° and a speed of 1.0°·min⁻¹. The specific surface areas and porosities of the catalysts, which were degassed under vacuum conditions at 200 °C for at least 2 h in advance, were measured by nitrogen physisorption at liquidized N₂ temperature in a Quantachrome Autosorb-iQ₂ instrument (Quantachrome, Boynton beach, FL, USA). The specific surface area and pore-size distribution of catalysts were determined by the BET and DFT methods, respectively.

High-resolution transmission electron microscopy (HRTEM) measurements were carried out with a JEOL JEM-2000 EX device (JEOL, Tokyo, Japan) operated at an accelerating voltage of 120 kV. Pd loading was determined by inductively coupled plasma–atomic emission spectroscopy (ICP-AES). The dispersion of palladium was measured by CO-pulse chemisorption experiments on a Quantachrome CHEMBET Pulsar adsorption instrument. Before the analysis, the catalysts were reduced in H₂/Ar at 120 °C for 1 h. The adsorption measurements were performed at 45 °C in a flow of He.

H₂ temperature-programmed reduction (H₂-TPR) of catalysts was performed on the flow system of a Quantachrome CHEMBET3000 adsorption instrument (Quantachrome, Boynton beach, FL, USA) equipped with a TCD detector. Samples were dried in an argon flow at 150 °C for 1h before the reduction step. The samples were reduced in a 10 vol.% H₂/Ar flow system at a rate of 10 °C·min⁻¹ from −20 °C to 500 °C.

H₂ temperature-programmed desorption (H₂-TPD) of the samples was performed on a CHEMBET3000 chemisorption analyzer (Quantachrome, Boynton Beach, FL, USA) from the Quantachrome company. First, 0.1 g of each catalyst sample was put into a U-shaped quartz reactor and heated from 100 °C to 300 °C with a heating rate of 10 °C·min⁻¹, and the temperature was kept at 300 °C for 2 h in He flow with a rate of 40 mL·min⁻¹ to remove surface impurities. The samples were cooled down to 50 °C, then H₂ (10 vol.% H₂ in He) was introduced into the detection system, and the sample adsorbed NH₃ to saturation and was then purged with helium until the baseline was stable. The adsorbed H₂ of the sample was gradually desorbed in He flow by raising the temperature from 50 °C to 800 °C at 10 °C·min⁻¹, and a constant TCD signal was obtained.

NH₃ temperature-programmed desorption (NH₃-TPD) of the samples was performed on a CHEMBET3000 chemisorption analyzer from the Quantachrome company. First, 0.1 g of each catalyst sample was put into a U-shaped quartz reactor and heated from 100 °C to 300 °C with a heating rate of 10 °C·min⁻¹, and the temperature was kept at 300 °C for 2 h in He flow at a rate of 40 mL·min⁻¹ to remove surface impurities. The samples were cooled down to 100 °C, then NH₃ (20 vol.% NH₃ in He) was introduced into the detection system, and the sample adsorbed NH₃ to saturation and then purged with helium until the baseline was stable. The adsorbed NH₃ of the sample was gradually desorbed in He flow by raising the temperature from 100 °C to 800 °C at 10 °C·min⁻¹, and a constant TCD signal was obtained.

X-ray photoelectron spectra were acquired on a Thermo ESCALAB 250Xi XPS system (Thermo Fisher Scientific, Waltham, MA, USA) using an Al Kα X-ray anode at a base pressure of about 1.7 × 10⁻¹⁰ mbar. Energies were calibrated using the C 1s peak (284.6 eV). The catalyst samples were pre-reduced ex situ under a flow of H₂ (40 mL·min⁻¹) at 100 °C for 1 h, followed by cooling under an Ar atmosphere. The reduced samples were transferred into the XPS cell under an Ar atmosphere to avoid exposure to air.

3.3. Evaluation of Catalytic Performance

The catalytic activity and selectivity of all catalysts (1 g for each) were tested in a stainless-steel slurry reactor with a volume of 150 mL at 40 °C. The working solution was prepared by dissolving 224 g of AAQ in 1 L of a solvent containing mixed Diisobutylcarbinol (DIBC) and C₉-C₁₀ aromatic hydrocarbons with a volume ratio of 2:3. The H₂ volumetric flow rate was set at 14 mL·min⁻¹, while the liquid flow rate was set at 2 mL·min⁻¹. Five milliliters of the catalyst-free hydrogenation product was collected once an hour and oxidized by air for 30 min at room temperature. H₂O₂ was then extracted with deionized water, and its concentration was analyzed by a titration method with a KMnO₄ solution.

The catalyst activity can be expressed by the following simplified equation:

$$B=\frac{5c_{{KMnO}_4}\times V_{{KMnO}_4}\times M_{H_2{O}_2}}{2V}$$

(1)

where B is the hydrogenation efficiency (g·L⁻¹) to specify how much H₂O₂ is produced by hydrogenating a unit volume of the working solution; C_KMnO4 is the KMnO₄ solution concentration (mol·L⁻¹); V_KMnO4 is the KMnO₄ solution volume (mL); V is the hydrogenated working solution volume (mL); and M_H2O2 is the relative molecular mass of H₂O₂. As the Pd loading of each catalyst was slightly different, the hydrogen efficiencies were normalized.

The catalyst selectivity in this reaction can be expressed by the following simplified equation:

$$S=\frac{n_{\left(AAQ\right)}+n_{\left(H_{4}AAQ\right)}}{n_{0\left(AAQ\right)}}\times 100\%$$

(2)

S is the selectivity of active quinones (AAQ and H₄AAQ), and n₀ and n are the molar contents of components in the initial working solution and in the accumulated re-oxidized solution, respectively. A volume of 0.01 mL of AAQ solution samples extracted from the reactor before and after the reaction were analyzed by high-performance liquid chromatography (HPLC, Agilent, City of Santa Clara, CA, USA), and their compounds were analyzed by comparing HPLC digital fingerprints.

4. Conclusions

Reaction activity and selectivity are crucial for designing catalysts in the process of 2-amylanthraquinone hydrogenation to produce H₂O₂. In this work, a series of Pd-AlPO-n/SiO₂ and Pd/SiO₂ catalysts were prepared by the impregnation method, among which the Pd-AlPO-31/SiO₂ catalyst presented the highest H₂O₂ yield (8.4 g·L⁻¹) and a superior catalytic selectivity of 96%. This is attributed to the suppression of smaller Pd particles on AlPO-31/SiO₂, as proven by the results of HRTEM, H₂-TPR, XPS and H₂-TPD characterization. Moreover, it was proven that Pd(NH₃)₄²⁺ enters the micropore channels of AlPO-n, suppressing the accumulation of Pd particles, especially for AlPO-31 and AlPO-11. Among the tested AlPO-n molecular sieves, AlPO-5 has the best H₂ adsorption performance, while the density of H₂ inside the micropore channels of AlPO-31 is the highest, benefitting the catalytic activity of Pd inside its micropore channels for dissociating H₂. The catalytic selectivity of Pd-AlPO-11/SiO₂ is poor and produces more deeply hydrogenated byproducts, although AlPO-11 exhibits similar micropore properties.