CO₂ hydrogenation toward toluene methylation for selective para-xylene synthesis

Abstract

The selective production of high-value para-xylene (PX) through CO₂ hydrogenation-coupled toluene methylation represents a promising strategy for sustainable carbon utilization. Herein, we report a ZnZrO_x & HMCM-22 tandem catalyst that integrates CO₂ hydrogenation with shape-selective toluene methylation. The optimized catalyst achieves 91% xylene selectivity with 70.4% PX dominance at 10.3% toluene conversion. Controlled tetraethyl orthosilicate (TEOS) deposition passivates Brønsted acid sites in HMCM-22 supercages, suppressing toluene disproportionation while preserving methylation activity in sinusoidal channels. Mechanistic studies reveal that formaldehyde species, generated on ZnZrO_x and transferred to HMCM-22, act as kinetically favored methylating intermediates, enabling direct coupling between CO₂ hydrogenation and toluene methylation. This pathway outperforms conventional methanol-mediated routes in both activity and selectivity. Our work establishes a dual-optimization strategy—zeolite microenvironment engineering and reactive intermediate control—offering mechanistic insights and design principles for high-performance tandem catalysis in carbon recycling aromatic synthesis.

Introduction

Carbon dioxide, as an abundant C1 feedstock, presents a great opportunity for sustainable chemical synthesis when coupled with green hydrogen derived from renewable energy. Recent advances have demonstrated the feasibility of converting CO₂/H₂ mixtures into value-added products including methanol^1,2, light olefins^3,4,5,6, and aromatics^7,8,9. Although direct CO₂ hydrogenation to aromatics has been reported using tandem catalysis, the para-xylene (PX) selectivity within aromatic products remains suboptimal. This limitation contrasts sharply with the industrial demand for PX—a precursor for polyethylene terephthalate (PET) manufacturing—which currently relies predominantly on fossil-based routes such as naphtha cracking and methanol-mediated toluene alkylation. These considerations highlight the urgency to develop efficient strategies for synthesizing PX through CO₂-derived methylating agents with toluene, thereby enabling PX production via a two-step CO₂/H₂ pathway.

Conventional toluene methylation processes employing methanol confront intrinsic thermodynamic limitations when implemented over H-ZSM-5 zeolites at 400–500 °C. Under these harsh thermal conditions, the concurrent activation of methanol-to-olefins (MTO) side reactions precipitates a cascade of operational challenges: competitive methanol consumption by olefin synthesis pathways leads to methylating agent depletion, while concomitant carbon deposition accelerates catalyst deactivation through pore blockage and active site coverage^10,11. Furthermore, the kinetic rivalry between methylation and MTO pathways severely compromises reaction efficiency. This tripartite interplay of resource misallocation, catalyst instability, and suppressed target reaction kinetics collectively necessitates the development of catalytic architectures capable of circumventing methanol intermediation under thermodynamically favorable conditions.

Recent developments in tandem catalysis reveal promising alternatives through in situ-generated oxygenate intermediates. Oxide-zeolite composites (e.g., oxide: ZnZrO_x, ZnGa₂O₄, In₂O₃-ZrO₂, ZnCrO_x; zeolites: SAPO-34, SSZ-13, H-ZSM-5)^{12,13,14,15,16} enable CO₂ hydrogenation to lower olefins and aromatics via CH_xO intermediates through formate-mediated pathways. Notably, these systems operate at significantly lower temperatures than conventional methanol-based processes, as CH_xO intermediates exhibit lower activation barriers compared to free methanol. Our previous work further demonstrated that methanol synthesis in tandem systems is thermodynamically driven by downstream olefin formation—a critical coupling mechanism that facilitates intermediate-directed catalysis⁴. These findings suggest that judicious selection of reactive methylating intermediates could decouple methylation kinetics from MTO side reactions.

Emerging CO₂ hydrogenation-coupled toluene methylation strategies underscore the pivotal influence of zeolite architectural design on product selectivity^17,18,19,20. Whereas H-ZSM-5-based systems deliver moderate xylene yields, their inherent structural constraints—notably unidirectional channel geometry and pronounced Brønsted acidity—predispose the system toward competing toluene disproportionation and xylene isomerization pathways^21,22. In contrast, MCM-22 zeolite demonstrates distinctive structural merits for shape-selective alkylation, characterized by sinusoidal 10-membered ring (10-MR) channels that impose steric restrictions on bulkier isomers, coupled with chemically tunable supercage configurations enabling precise acid site modulation. The spatial isolation of active sites within this architecture further suppresses disproportionation side reactions. These collective structural attributes establish MCM-22 as an exemplary scaffold for constructing tailored methylation microenvironments with enhanced selectivity^23,24.

In this work, we develop a bifunctional ZnZrO_x & HMCM-22 (ZZO & M-22) tandem catalyst that synergistically integrates CO₂ hydrogenation and shape-selective methylation. Through strategic silica deposition in MCM-22 supercages, we selectively passivate Brønsted acid sites responsible for toluene disproportionation while preserving methylation-active sites in sinusoidal channels. Experimental results, advanced characterization via in situ synchrotron vacuum ultraviolet photoionization mass spectrometry (SVUV-PIMS) and density functional theory (DFT) calculations identify formaldehyde-like CH₂O* as the dominant methylating intermediate, exhibiting a 0.86 eV lower activation barrier compared to methanol-mediated pathways. The optimized catalyst achieves 91% xylene selectivity with 70.4% PX dominance among isomers at 10.3% toluene conversion. This study establishes a dual-optimization strategy combining zeolite microenvironment engineering with reactive intermediate control, providing a blueprint for developing industrially viable CO₂-to-aromatics systems.

Results and discussion

Catalytic performance

The catalytic performance of tandem catalyst systems was systematically investigated in a fixed-bed reactor under controlled reaction conditions. Temperature-dependent studies show a monotonic increase in toluene conversion from 3% to 32% with simultaneous attenuation of PX selectivity with increasing reaction temperature from 280 to 400 °C (Fig. 1a, and Supplementary Fig. 1a). Thermodynamic redistribution of xylene isomers favored meta-xylene (MX) formation, accompanied by decreasing benzene selectivity and progressive accumulation of 1,2,4-trimethylbenzene (TMB) with increasing reaction temperature. This temperature-response profile suggests competing mechanisms: enhanced alkylation kinetics promoting active methyl species-toluene coupling at elevated temperatures, versus thermally activated processes including xylene isomerization to MX and over-methylation to TMB. Comparative evaluation of individual components in the ZZO & M-22 tandem system demonstrated distinct functionalities ZZO exclusively mediated CO₂-to-methanol conversion while M-22 remained catalytically inert in isolation for CO₂ hydrogenation (Fig. 1b). The combined system achieved integrated CO₂ hydrogenation and toluene methylation without generating aromatization byproducts, with control experiments confirming toluene disproportionation as the dominant pathway on isolated M-22 with toluene as feedstocks.

Fig. 1: Catalytic Performance in CO₂ Hydrogenation.

a The effect of reaction temperature. b CO₂ hydrogenation in the presence of toluene over ZZO & M-22; CO₂ hydrogenation over ZZO, M-22, ZZO & M-22 and toluene disproportionation over M-22. c Influence of M-22-modified times by TEOS on the aromatics distribution and toluene conversion. d The Stability test. Standard reaction conditions: 3 MPa, 320 °C, GHSV of a mixture gas (CO₂/H₂/Ar = 24:72:4) = 12000 mL g⁻¹ h⁻¹, and WHSV of toluene = 1 h⁻¹. When only toluene: 3.0 MPa, WHSV of toluene = 1 h⁻¹, GHSV of mixture gas (CO₂/H₂/Ar = 24:72:4) = 12000 mL g⁻¹ h⁻¹. The color shadings represent the selectivity, and the red square represents the conversion.

Modification of M-22 via tetraethyl orthosilicate (TEOS) deposition induced three critical effects: pore architecture modulation favoring PX diffusion, passivation of external acidic sites, and suppression of disproportionation pathways. Progressive TEOS deposition cycles (1–4 cycles) enhanced PX selectivity from 43% to 64% while suppressing benzene selectivity from 10% to <1%, though with concomitant TMB accumulation (9%) due to restricted PX diffusion enabling secondary methylation (Fig. 1c). Five deposition cycles induced pore blockage, manifesting as activity loss with toluene conversion decreasing from 14% to 7%, establishing 4-cycle modification (ZZO & 4Si/M-22) as the optimal configuration. The ZZO & 4Si/M-22 catalyst achieves 91% xylene selectivity with 70.4% PX dominance among isomers at 10.3% toluene conversion. Compared to other catalyst systems, the ZZO & 4Si/M-22 can obtain a high xylene selectivity while achieving a high PX selectivity, and can completely inhibit the occurrence of toluene disproportionation side reactions (Supplementary Table 1). And the ZZO & 4Si/M-22 achieved a CO₂ conversion of 4.2% with a CO selectivity of 22% at 320 °C (Supplementary Fig. 1b). Long-term stability tests confirmed sustained performance (>120 h) with undiminished PX selectivity (Fig. 1d).

Process parameter optimization revealed pressure-dependent behavior with elevated pressure from 1 to 6 MPa, which enhances toluene conversion from 5% to 12% but promotes TMB formation with selectivity increasing from 4% to 12% through PX retention-induced over-methylation (Supplementary Fig. 2a). Gas hourly space velocity (GHSV) studies identified 12,000 mL g⁻¹ h⁻¹ as optimal, balancing toluene conversion (10.3%) and selectivity (PX 64%, TMB 9%), with lower GHSV to 6,000 mL g⁻¹ h⁻¹ inducing disproportionation pathways (Supplementary Fig. 2b). Weight hourly space velocity (WHSV) adjustments demonstrated an intrinsic activity-selectivity trade-off: increasing WHSV from 1 to 3 h⁻¹ improved PX selectivity from 64 to 80% but incurred 80% toluene conversion loss with decreasing from 10.3% to 2%, while simultaneously suppressing TMB formation (Supplementary Fig. 2c). These systematic investigations establish the ZZO & 4Si/M-22 system as an effective tandem catalyst for selective PX production, achieving precise control through synergistic component integration, strategic pore engineering, and optimized reaction parameters.

Structure characterization of the tandem catalysts

The ZZO solid solution oxide catalyst was synthesized through 20 mol% ZnO doping into ZrO₂. X-ray diffraction analysis confirms that the resultant ZZO solid solution adopts a characteristic tetragonal crystal structure, as evidenced by the diffraction pattern in Fig. 2a. As shown in the transmission electron microscopy (TEM) image (Fig. 2b), the ZZO material exhibits a nanoparticle morphology with uniform dimensions. Elemental mapping analysis further demonstrates the homogeneous distribution of Zn atoms within the ZrO₂ matrix (Fig. 2c), verifying the successful formation of a solid solution structure. The M-22 zeolite component displays well-defined nanosheet morphology, as observed in Fig. 2d. These crystalline sheets exhibit thicknesses of approximately 28 nm and lateral dimensions ranging from 300 to 500 nm. In the ZZO & M-22 tandem catalyst system, scanning electron microscopy (SEM) characterization reveals that ZZO nanoparticles are uniformly dispersed across the M-22 nanosheet surfaces (Fig. 2e). A critical observation from the high-resolution SEM image is the preservation of distinct structural features for both components, indicating that the interfacial interaction between ZZO and M-22 remains primarily physical rather than chemical. Additional structural validation is provided in Supplementary Fig. 3a, b, which present SEM images of the 4Si/M-22 modified sample. The morphology of M-22 nanosheets remains essentially unchanged after TEOS treatment, confirming that the structural integrity of the zeolite framework is maintained during the modification process. And the specific BET surface area, micropore area, external area and pore volume of M-22 all decreased, while the pore diameter remained almost unchanged. (Supplementary Table 2). This indicates that TEOS does not deposit extensively within the internal pores of the M-22, but predominantly at the entrances of all channels and on the external surface, thereby enhancing the shape selectivity of M-22.

Fig. 2: Structure characterization of the tandem catalysts.

a X-ray diffraction patterns of ZZO, M-22, and ZZO & 4Si/M-22. b The TEM and Aberration-corrected HRTEM images of ZZO. c Aberration-corrected high-angle annular dark-field scanning TEM images and elemental distribution of Zn and Zr for ZZO. d SEM image of M-22. e SEM image of ZZO & M-22.

Surface chemical properties of the tandem catalysts

CO₂ temperature-programmed desorption (CO₂-TPD) analysis reveals distinct adsorption behaviors between components (Supplementary Fig. 4a). The ZZO solid solution demonstrates sustained CO₂ adsorption up to 500 °C, while the zeolite component exhibits negligible CO₂ adsorption capacity. Quantitative evaluation of basic sites confirms that the physical integration of zeolite in the tandem catalyst exerts minimal influence on CO₂ adsorption performance of ZZO (Supplementary Table 3). The superior PX selectivity observed in ZZO & 4Si/M-22 tandem catalysts highlights the critical role of TEOS modification. This performance enhancement correlates with the decisive influence of zeolite acid properties (both quantity and type) on PX selectivity. To elucidate this relationship, comprehensive acid characterization was conducted using NH₃-TPD, pyridine Fourier-transform infrared (Py-FTIR) spectroscopy, ¹H MAS NMR, and ²⁷Al MAS NMR. NH₃-TPD profiles (Supplementary Fig. 4b) exhibit two characteristic desorption regions: weak acid sites (50–250 °C) and strong acid sites (250–500 °C). As shown in Supplementary Table 4, the total acid content of the ZZO & M-22 tandem catalyst closely matches the arithmetic sum of individual components, confirming that physical mixing preserves the intrinsic acidity of both ZZO and M-22.

Systematic investigation of acid evolution for M-22 under successive TEOS modifications reveals progressive acid suppression (Fig. 3a). Notably, both weak and strong acid contents stabilize after four modification cycles, indicating saturation of surface passivation. Py-FTIR analysis identifies characteristic bands at 1540 cm⁻¹ (Brønsted acid sites) and 1440 cm⁻¹ (Lewis acid sites), with quantitative analysis following established protocols (Fig. 3b)²⁵. Supplementary Table 5 demonstrates a marked reduction in both acid types with increasing modifications, particularly the near-complete elimination of Brønsted acidity after four cycles. This evidences TEOS-mediated encapsulation of acid sites within external surface pockets of M-22. ¹H MAS NMR spectra (Fig. 3c) provide further validation of TEOS-induced acid modulation. The diminishing signal at 1.8 ppm (assigned to external silanol groups, Si-OH²⁶) and attenuated resonance at 3.9 ppm (bridging hydroxyls, Si(OH)Al^27,28) correlate with modification cycles. Concurrent chemical shift variations in the 3.9 ppm band reinforce the role of TEOS in external surface acid passivation, as quantified in Supplementary Table 6.

Fig. 3: Surface chemical properties of the tandem catalysts.

a NH₃-TPD of M-22 with different Si modification times. b Pyridine-FTIR results for M-22 with different Si modification times. c ¹H MAS NMR characterizations of M-22 with different Si modification time. d ²⁷Al MAS NMR characterizations of M-22 with different Si modification times. Lewis Acid (LA) and Brønsted acid (BA).

Considering the intrinsic link between Brønsted acidity and framework aluminum (Al_F) in zeolites (Supplementary Fig. 5), ²⁷Al MAS NMR analysis was employed to probe Al speciation (Fig. 3d). Distinct resonances at 0 ppm (extra-framework Al_EF, 6-coordinated) and 56 ppm (framework Al_F, 4-coordinated) were identified²⁹. Deconvolution of the 56 ppm band resolves three components: 50 ppm (T6 + T7 sites), 56 ppm (T1 + T3 + T4 + T5 + T8 sites), and 61 ppm (T2 sites)³⁰. Structural analysis localizes T2 and T1 + T3 sites in supercages/surface pockets, T4 in inaccessible framework positions, T5 in sinusoidal channels with unstable and unfavorable for siting Al, and T8 in the sinusoidal channels connected to the bottom of the supercages, with thermodynamically stable bridging hydroxyls^31,32,33, with T6 + T7 residing on external surfaces³⁴. Progressive TEOS modification induces three critical changes: gradual attenuation of the Al_EF signal at 0 ppm; systematic reduction of overall Al_F intensity at 56 ppm; selective disappearance of the 61ppm component (T2 supercage sites) upon deconvolution (Supplementary Table 7, Supplementary Fig. 6a).

These observations suggest TEOS not only passivates external surface pocket acid sites but also modulates supercage acidity. This aligns with literature demonstrating toluene disproportionation (a benzene-forming side reaction) predominantly occurs in MWW-type supercages^35,36. Consistent with Fig. 1c, reduced benzene selectivity parallels TEOS-induced supercage deactivation. Complementary ²⁷Al MAS NMR analysis of precoked M-22 confirms analogous attenuation of 61 ppm signals (Supplementary Fig. 6b, c), validating efficacy of TEOS in suppressing toluene disproportionation through targeted acid site passivation.

Unraveling the key formaldehyde intermediates species and the reaction mechanism

The reaction mechanism was investigated through multiple complementary techniques. In situ DRIFTS analysis of ZZO during CO₂ hydrogenation revealed formate (HCOO*) and methoxy (H₃CO*) intermediates, with characteristic vibrations at 1595/1370 cm⁻¹ (OCO stretching) and 2931/2824 cm⁻¹ (C-H modes), respectively (Supplementary Fig. 7a). Isotopic substitution (CO₂ + D₂) caused rapid depletion of HCOO*/H₃CO* signals within 5 min, replaced by transient DCOO* (2165 cm⁻¹) and HD₂CO* (2050 cm⁻¹) species, confirming the formate pathway dominates CO₂ hydrogenation over ZZO (Supplementary Fig. 7b)¹. Isotope-switching (¹³CO₂) experiments confirmed CO₂ incorporation into products through 15 cm⁻¹ redshift (2880 → 2865 cm⁻¹) in methyl C-H vibrations and intensified aromatic signals (Fig. 4a)^21,37,38. Compared to ZZO, ZZO & 4Si/M-22 showed selective disappearance of H₃CO* species, while the HCOO* species still retain suggesting intermediate derived from formate species transfer to the zeolite component, thereby involving the formation of PX (Fig. 4b).

Fig. 4: Surface species and the key formaldehyde intermediates derived from the reactions.

a In situ DRIFTS spectra over the ZZO & 4Si/M-22 catalyst exposed to the stream of C₇H₈ + CO₂ + H₂ and subsequently switched to C₇H₈ + ¹³CO₂ + H₂. b In situ DRIFTS of CO₂ + H₂ over ZZO and ZZO & 4Si/M-22. c Quasi in situ¹³C CP MAS NMR experiment for ¹³CO₂ + H₂ on ZZO and ¹³CO₂ + H₂ + C₇H₈ on ZZO & 4Si/M-22. d Quasi in situ 2D ¹³C-1H HETCOR MAS NMR experiment for ¹³CO₂ + H₂ on ZZO. e Quasi in situ 2D ¹³C-1H HETCOR MAS NMR experiment for ¹³CO₂ + H₂ + C₇H₈ on ZZO & 4Si/M-22. Reaction conditions: 320 °C, 3 MPa, 30 min tube-sealing test. f The GC-MS chromatograms of the others aromatics in the products from alkylation of toluene with different alkylation reagents (CO₂ + H₂, formaldehyde and methanol) over the ZZO & M-22 and M-22 catalyst. g The GC-MS chromatograms of the others aromatics in the products from alkylation of toluene with CO₂ + H₂ over the ZZO & M-22 and ZZO catalyst. h In situ SVUV-PIMS detection of gas-phase products and i photoionization efficiency spectra of the signals at m/z = 30 and 32 over ZZO catalyst during CO₂ + H₂. (300 °C, 0.8 MPa, 9,000 mL g_cat⁻¹ h⁻¹ and H₂/CO₂ = 3/1). j The reaction pathway over the ZZO & M-22 catalyst.

Quasi in situ ¹³C NMR revealed surface species evolution (Fig. 4c): ZZO in ¹³CO₂ + H₂ showed formates (169 ppm), methoxy (52 ppm), and methanol (49 ppm)^39,40 while ZZO & 4Si/M-22 in ¹³CO₂ + H₂ + C₇H₈ exhibited depleted CH_xO signals alongside emerging xylene and trimethylbenzene signatures (20.9 ppm and 19.5 ppm)^41,42. ¹H-¹³C 2D heteronuclear correlation (HETCOR) NMR correlations confirmed these assignments formate (169/8.4 ppm), methoxy (52/3.7 ppm), and methanol (48/3.2 ppm) in ZZO (Fig. 4d), transitioning to PX and 1,2,4-trimethylbenzene signals (20.9/2.4 ppm and 19.5/1.84 ppm) in the tandem system (Fig. 4e). Collectively, these results demonstrate CH_xO* identified as the key intermediate enabling selective PX formation through spatial proximity effects in the tandem catalyst system, and the depletion of CH_xO species promotes formate conversion.

To further identify the CH_xO intermediates and their migration behavior from ZZO to MCM-22, controlled experiments were performed (Fig. 4f, 4g). Significantly, side product—specifically 2-(m-methylbenzyl)toluene (2MMBT), 4-(p-methylbenzyl)toluene (4PMBT), and 2-(p-methylbenzyl)toluene (2PMBT)—were detected with selectivity of 3% during CO₂ hydrogenation coupled with toluene over the ZZO & M-22 tandem catalyst (Fig. 4f, top panel; Supplementary Fig. 8a). This observation suggests that CH₂O species, generated from CO₂ hydrogenation, react with toluene, initially forming methylbenzyl alcohol intermediates. These intermediates then undergo further reactions with toluene to produce the observed side products. Crucially, when formaldehyde (CH₂O) was used as a methylating agent, replacing CO₂/H₂, identical side products were formed. In contrast, methanol, as a methylating agent, did not lead to the formation of these compounds. Furthermore, formaldehyde exhibited more favorable reaction kinetics compared to methanol, facilitating PX formation at lower temperatures and achieving higher PX selectivity (Supplementary Fig. 8b). In situ SVUV-PIMS analysis enabled the identification of key intermediates through their characteristic ionization thresholds (Fig. 4h, i): m/z = 30 (H₂CO, 10.88 eV) and m/z = 32 (CH₃OH, 10.85 eV), which are consistent with values reported in the literature^43,44. These findings confirm that CH₂O* are generated over ZZO, and that formaldehyde—not methanol—acts as the alkylating agent formed from CO₂ + H₂ over the ZZO & M-22 tandem catalyst. Reaction kinetics further support the formation of CH₂O* species formed over ZZO from CO₂ hydrogenation, with methanol being produced over ZZO (Fig. 1b). To confirm that alkylation occurs exclusively over MCM-22 and not ZZO, the reaction of CO₂ + H₂ with toluene was carried out over isolated ZZO catalyst (Supplementary Fig. 9). This experiment unequivocally demonstrated that alkylation requires the presence of the MCM-22 component. These results indicate the CH₂O* species, formed over ZZO, can migrate to M-22 site to contribute to the formation of PX.

Based on the results presented above, the following reaction pathway is proposed, as illustrated in Fig. 4j. Initially, CO₂ and H₂ are adsorbed and activated on the ZZO component, leading to the formation of HCOO* species. These HCOO* species then undergo hydrogenation to produce CH₂O* intermediates. The CH₂O* intermediates subsequently migrate to the MCM-22 zeolite, where they react with toluene to form methylbenzyl alcohol, with para-methylbenzyl alcohol being the dominant product. The para-methylbenzyl alcohol then primarily undergoes dehydration, followed by hydrogenation, to yield the target product, PX. In parallel, some of the methylbenzyl alcohol can react with an additional toluene molecule, resulting in the formation of side products—2MMBT, 4PMBT, and 2PMBT.

DFT calculations on the reaction pathway

The experimental findings collectively suggest that CH₂O* serves as the critical intermediate enabling kinetic coupling between the ZZO and M-22 components in the tandem catalytic system. To validate this hypothesis at the molecular level, density functional theory (DFT) calculations were conducted to compare three potential intermediates–CH₂O*, CH₃O*, and CH₃OH in their ability to migrate from ZZO into sinusoidal channels of M-22. For the methanol-mediated pathway (Fig. 5a m1-m6), the mechanism initiates with methanol adsorption at the Brønsted acid site in the T8 position of sinusoidal channel in M-22 (Fig. 5a). Subsequent dehydration through transition state mTS2 generates a methyl group (CH₃) and water, requiring a substantial activation barrier of 2.05 eV (Fig. 5b). Following water desorption (Fig. 5a m4), toluene enters the framework where the CH₃ group attacks the para-carbon of the aromatic ring (Fig. 5a m5), ultimately yielding desorbed PX (Fig. 5a m6).

Fig. 5: DFT calculations on the reaction pathway.

a Optimized geometries of all reaction intermediates and transition states. Reaction Gibbs free energy diagram: b Methanol reacting with toluene over M-22 zeolite, c CH₂O* reacting with toluene on M-22 zeolite. d The initial adsorption energy of CH₃O* relative to CH₂O* and CH₃OH in M-22.

In contrast, the CH₂O*–mediated pathway demonstrates superior thermodynamic feasibility (Fig. 5a f1-f11). Initial adsorption of CH₂O* at the T8 Brønsted site (Fig. 5a f1) is followed by protonation to form a hydroxymethyl cation (CH₂OH*) (Fig. 5c f2). Concurrent toluene ingress enables para-carbon attack via transition state fTS1 (Fig. 5c), forming H--HO-CH₂-Tol* (Fig. 5c f3-f4). The rate-determining step involves C-O bond cleavage through transition state fTS2 (1.19 eV barrier) (Fig. 5c f5), yielding CH₂-Tol* and water (Fig. 5c f6). Subsequent hydrogen transfer from surface H* species (Fig. 5c f8) via transition state fTS3 (Fig. 5c f9) generates CH₃-Tol*, culminating in PX desorption (Fig. 5c f11). Notably, CH₃O* exhibits substantially higher adsorption energies compared to both CH₂O* (+1.0 eV) and CH₃OH (+3.0 eV) (Fig. 5d), rendering it thermodynamically unfavorable for alkylation within M-22. The maximum activation barrier of CH₂O*-mediated pathway (1.19 eV) remains significantly lower than that of methanol barrier (2.05 eV), conclusively identifying CH₂O* as the kinetically favored intermediate.

In summary, this work demonstrates a tandem catalytic strategy for direct PX synthesis via CO₂ hydrogenation coupled with toluene methylation using a ZZO & M-22 system. Compartmentalized catalysis enables synergy: ZZO preferentially activates CO₂/H₂, while M-22 channels mediate C–C coupling via CH₂O* intermediates. This approach overcomes thermodynamic constraints of conventional toluene methylation, achieving 91% xylene selectivity (70.4% PX/xylene) at 10.3% toluene conversion. TEOS modification of M-22 is critical, passivating supercage Brønsted acid sites (Si(OH)Al) via TEOS deposition to suppress disproportionation side reactions. Mechanistically, CH₂O* intermediates kinetically couple CO₂ activation and alkylation, bypassing methanol dehydration. Controlled diffusion of CH₂O* from ZZO to M-22 optimizes thermodynamic-kinetic balance, facilitating confined electrophilic para-methylation. This establishes a paradigm for tandem catalyst design through zeolite architecture engineering, where precise intermediate dimension/configurational control governs selectivity. The strategy offers an efficient CO₂ valorization route and a carbon recycling alternative to conventional PX production, underscoring structure-directed catalysis in bridging CO₂ utilization with aromatics synthesis.

Methods

Materials

Zn(NO₃)₂ · 6H₂O were purchased Sinopharm (China), (NH₄)₂CO₃, m-xylene (99%), toluene (99.9%) and Zr(NO₃)₄ · 5H₂O, tetraethylorthosilicate (TEOS, 99.98%) and n-hexane (99%) were obtained from Macklin, CO₂/H₂/Ar [24/72/4(v/v/v)], H₂ (99.99%) and Ar (99.99%) were obtained from Lanzhou Yulong Gas Co., Ltd. HMCM-22 (Si/Al=15) were obtained from Nankai Catalysts.

Preparation of catalysts

The ZnZrO catalyst that doped 20% ZnO (Zn/ (Zn + Zr) molar ratio) was prepared as follows. 14.9 g of Zn (NO₃)₂ · 6H₂O and 8.6 g of Zr (NO₃)₄ · 5H₂O were dissolved in a flask by using 400 mL of deionized water, then the 250 mL aqueous solution of (NH₄)₂CO₃ (43.2 g) as precipitant was dropwise added into the aforementioned solution under vigorous stirring at 70 °C to form a precipitate. The obtained suspension was continuously stirred for 4 h at 70 °C, followed by cooling down to room temperature and filtering. The precipitate was washed three times with deionized water, dried at 110 °C for 4 h, and calcined at 500 °C in static air for 3 h.

Si/M-22 zeolite was prepared by tetraethoxysilane (TEOS) modification. Typically, the steps to modify the M-22 once were to immerse 1 g M-22 into the mix solution of 1 g TEOS and 20 ml of n-hexane for 4 h; after that, organic solvent was removed by evaporation at 80 °C and the powder was dried overnight at 110 °C, followed by calcination at 550 °C for 4 h in air. While the above steps were repeated, a new symbol of nSi/M-22 (n means the number of repetitions) was used. After repeating the procedure above for four times, the 4Si/M-22 was obtained. The tandem catalyst was typically prepared by physical mixing. The weight ratio of oxides and zeolites was 1:1. For preparation of tandem catalyst ZZO & M-22, ZZO and M-22 were grinded in an agate mortar for 5 min, pressed under 10 MPa and granulated into the required size in the range of 40-80 mesh.

M-22 with the pre-coked supercages, denoted as pre-coked M-22 was prepared following the procedure: M-22 was pretreated at 500 °C for 2 h and then cooled down to 350 °C under Ar flow (100 mL min⁻¹). Subsequently, the transformation of m-xylene (99%) was carried out at 350 °C, atmospheric pressure and 8 h⁻¹ of m-xylene WHSV with Ar flow (40 mL min⁻¹) for 12 h, to form pre-coked M-22.

Catalyst evaluation

Catalytic performance was evaluated on a high-pressure fixed-bed reactor equipped with a quartz sample tube. Typically, 0.52 g catalyst (40-80 mesh) were loaded into the quartz sample tube without mixed quartz sand. Before evaluating the catalytical performances, raised to the specified temperature under the protection of argon. For the investigation of toluene methylated with CO₂ and H₂, a mixture gas of CO₂/H₂/Ar [24/72/4(v/v/v)] was induced into a quartz sample tube, and then toluene was carried into reaction tube (90 °C) at a constant flow rate by using a piston pump. The reaction was conducted at 320 °C and 3.0 MPa, with a GHSV of 12,000 mL g⁻¹ h⁻¹ and a toluene weight hourly space velocity (WHSV) of 1 h⁻¹, and the data were collected at the reaction time of 12 h. For the contrast experiment of toluene methylated with methanol/Formaldehyde (s-Trioxane), a mixed solution of C₇H₈/CH₃OH or HCHO [5/1 (mol/mol)] was pumped into reaction tube with a toluene WHSV of 1 h⁻¹ at 320 °C /260 320 °C and 3.0 MPa, while 12,000 mL g⁻¹ h⁻¹, H₂/Ar was used as a carry gas simultaneously. In the quantitative analysis, we use diphenyl as an internal standard to determine the C balance, and the C balance was above 96%. The liquid products were analyzed by using the gas chromatography equipped with CBX-WAX columns to analyze aromatics, the conversion of toluene and production selectivity was calculated as follows:

$${Conv}.(\%)=\frac{\sum {n}_{i}}{{n}_{{toulene}}^{{in}}}\times 100$$

(1)

$${Sel}(\%)=\frac{{n}_{i}}{\sum {n}_{i}}\times 100$$

(2)

Where n_i is the molar amount at the outlet of the individual aromatic product _i.

Catalyst characterization

The XRD analysis was performed using a Philips PW1050/81 diffractometer equipped with a CuKα radiation source (λ = 1.5418 Å). The instrument was operated in Bragg-Brentano geometry at 40 kV and 30 mA.

The specific surface area was calculated using the Brunauer-Emmett-Teller (BET) method. The textural properties of the samples, including specific surface area and pore size distribution, were characterized by N₂ physisorption using a MicrotracBEL BELSORP-mini II analyzer. Prior to analysis, approximately 0.2 g of each catalyst was subjected to degassing at 300 °C for 2 h.

The microstructural and elemental composition of the catalysts were characterized by aberration-corrected HAADF-STEM and STEM-EDX mapping, respectively, using an FEI Cubed Titan G2 60-300 microscope at an accelerating voltage of 300 kV. Morphological analysis was performed by scanning electron microscopy (SEM) on a Thermo Fisher Scientific Apreo S instrument. The SEM specifications were as follows: electron gun: field emission (FEG); accelerating voltage: 30–200 kV.

CO₂ and NH₃ temperature-programmed desorption (TPD) experiments were conducted using a Micromeritics ASAP 2920 system. Approximately 0.1 g of catalyst, loaded in a quartz reactor, was first pretreated at 500 °C under Ar flow for 3 h and subsequently cooled to 50 °C. The sample was then saturated with a flow of either 10 vol% CO₂/He or 10 vol% NH₃/He (50 mL min⁻¹) for 1 h at this temperature. Subsequently, the flow was switched to pure He for 1 h at 50 °C to purge physically adsorbed molecules. After stabilizing the baseline, the temperature was ramped to 700 °C at a rate of 10 °C min⁻¹ under He flow, and the desorption signal was monitored by a thermal conductivity detector (TCD).

In-situ DRIFT spectra of surface species formed from the CO₂/¹³CO₂ + H₂/D₂ reaction. Prior to the experiment, the samples were heated to 450 °C for 2 h under an Ar atmosphere to remove any water contained in the cell and on the catalyst surface. In situ DRIFT spectra of surface species over ZZO from CO₂ + H₂ and subsequently switched to D₂. Reaction conditions: ZZO, 0.1 MPa, 320 °C, 30 mL min⁻¹ CO₂ + H₂, D₂ 10 mL min⁻¹. In situ DRIFT spectra of surface species over ZZO & 4Si/M-22, 30 ml/min CO₂ + 10 mL min⁻¹ H₂ +toluene, subsequently switched to ¹³CO₂ was injected into the cell, and infrared spectrum information of the catalyst surface and species was collected.

FTIR spectra of pyridine was performed for both M-22 and nSi/M-22. The sample (about 30 mg) was pressed into a wafer, and then treated under vacuum at 400 °C for 1 h. After cooling to 50 °C, pyridine was introduced to the transmission cell until saturation. Then, the signal vs. the reference signal was recorded at 150 °C until the signal unchanged.

Magic angle spinning (MAS) nuclear magnetic resonance (NMR) spectra of ¹H and ²⁷Al were collected on a Bruker Avance III 600 MHz Wide Bore spectrometer operating at a magnetic field of 14.1 T, 4 mm dural-channel probe; the rotate speed of samples is 10.0 kHz; The chemical shifts for ²⁷Al MAS NMR spectra were referenced to the aqueous solution of Al(NO₃)₃ (0 ppm), respectively. The chemical shifts for ¹H MAS NMR spectra were referenced to tetramethylsilane (TMS). Quantitative ¹H MAS NMR measurement was performed by comparing the signal area of the zeolite sample under study with that of adamantine as an external standard.

Quasi in situ¹³C CP MAS NMR experiment and Quasi in situ 2D ¹³C-¹H HETCOR MAS NMR experiment are carried out as follows: the basic experimental device is a home-built high-pressure fixed-bed catalytic reactor. Standard reaction conditions: 3 MPa, 320 °C, 10 mL min⁻¹ H₂ + 60 mL min^{−1 13}CO₂. The reactor was quickly immersed into liquid nitrogen to quench the reaction, so that highly reactive intermediates on the catalysts can be capture. Subsequently, the reactor was sealed, after waiting for recovery to room temperature, and further transferred into 4 mm rotor for further NMR experiment.

In situ SVUV-PIMS experiments were carried out beamline BL03U at the National Synchrotron Radiation Laboratory, Hefei, P.R. China. This technique is employed for the detection of oxygen-containing intermediates during the CO₂ hydrogenation process. The photoionization efficiency spectra were collected for 300 s in the energy range of 10.5-11.3 eV with a step of 0.05 eV. Reaction conditions: the ZnZrO catalyst (0.5 g); reaction gas (H₂/CO₂ = 3/1), at 300 °C, 0.8 MPa, GHSV = 12,000 mL g_cat⁻¹ h⁻¹.