Aerobic Oxidation of 5-Hydroxymethylfurfural to 2,5-Furandicarboxylic Acid over a Bi-Promoted Pt/Al₂O₃ Catalyst

Abstract

2,5-furandicarboxylic acid (FDCA), a high-value biomass-derived monomer, serves as a crucial building block for sustainable polymers including polyesters, polyamides, and polyurethanes. This study systematically investigated the catalytic oxidation of 5-hydroxymethylfurfural (HMF) to FDCA over Pt/Al₂O₃ and Pt–Bi/Al₂O₃ catalysts. The 5Pt/Al₂O₃ catalyst yielded 60.6% FDCA after 12 h under optimized conditions (80 °C, 0.1 MPa O₂, 1 equiv. Na₂CO₃). Remarkably, Bi-modified 5Pt–1Bi/Al₂O₃ catalyst dramatically enhanced catalytic performance, achieving 94.1% FDCA yield within 6 h under optimized conditions (80 °C, 1.5 MPa O₂, 2 equiv. Na₂CO₃). Comprehensive characterization revealed that the exceptional activity originates from Bi–O–Pt interactions that modulate the electronic structure and oxidation state of Pt active sites, which facilitates the oxidation of intermediate 5-formyl-2-furancarboxylic acid (FFCA) to FDCA, the rate-limiting step of HMF oxidation. This work demonstrates an efficient Bi-promoted Pt catalytic system for FDCA production with significant potential for industrial application.

1. Introduction

Transformation of biomass into furanic platform chemicals represents a crucial strategy for mitigating the depletion of fossil reserves and reducing CO₂ emissions [1,2]. 2,5-furandicarboxylic acid (FDCA) is the sole aromatic chemical among 12 highest-value biomass-derived chemicals identified by the U.S. DOE [3]. It serves as a pivotal monomer for synthesizing biopolymers with broad applications, such as polyamides [4], polyesters [5,6], and polyurethanes [7]. Among others, poly(ethylene 2,5-furandicarboxylate) (PEF) exhibits comparable thermal and mechanical properties to polyethylene terephthalate (PET) while offering superior gas barrier performance and biodegradability, making it particularly promising for packaging applications [5]. Chemocatalytic methods are widely employed for FDCA production via the oxidation of 5-hydroxymethylfurfural (HMF), owing to their operational simplicity and high catalytic efficiency [8,9]. In particular, heterogeneous catalysis systems enable easier separation and achieve higher FDCA yields compared to the homogeneous Co/Mn/Br catalyst system in acetic acid [10,11,12].

Recent studies have focused on improving the oxidation efficiency of non-noble metal catalysts (e.g., Co–Mn, Ce–Mn, and Cu–Mn oxides) [13,14,15,16], process integration with upstream sugar dehydration to HMF [2,17,18], and process intensification, including continuous-flow systems and base-free aqueous conditions [15,19,20,21,22]. However, scaling up most of these systems remains challenging due to complex catalyst preparation procedures and insufficient catalytic activity [12]. Noble metal catalysts represent a promising avenue for the sustainable production of FDCA via oxidation, given their superior efficiency, stability, alkali resistance, and scalability [23]. A representative example is Avantium’s YXY^® technology, which employs an Au-based catalyst to achieve an annual FDCA production capacity of 5000 tons [24]. Extensive studies have been conducted on the application of Pt catalysts in FDCA synthesis, with yields exceeding 95% achieved under optimized conditions [25,26,27]. The reaction intermediates, including 5-hydroxymethyl-2-furancarboxylic acid (HFCA), 2,5-diformylfuran (DFF), and 5-formyl-2-furancarboxylic acid (FFCA), were identified, as illustrated in Scheme 1 [28].

Bimetallic and multi-metallic catalysts often exhibit enhanced activity and selectivity over monometallic systems by promoting electron transfer and reactant adsorption to active sites [29,30,31]. The synergistic effect in Pt–Bi has been demonstrated during HMF oxidation by facilitating the adsorption and dehydrogenation of geminal diol while enhancing resistance to oxygen poisoning [28,32]. Quantitative conversion of HMF (0.1 M) and >99% yield of FDCA were achieved within 2 h using Pt–Bi/TiO₂ (n_Bi/n_Pt = 0.2) at 100 °C under 4 MPa air pressure and optimized basic conditions [33]. Similarly, Díaz et al. [34] achieved an exceptional FDCA yield of 99.7% within 4 h using 9Pt–3Bi/C catalyst (100 °C,1 M Na₂CO₃, 1 MPa O₂), and efficient FDCA recovery via HCl precipitation or isopropanol extraction further enhances process sustainability [35]. Miao et al. [36] reported that the Pt/Ce_0.8Bi_0.2O_2−δ catalyst exhibited exceptional performance, delivering 98% FDCA yield in 30 min at 23 °C under 1 MPa O₂. Beyond batch systems, Lilga et al. [37] employed Pt/Al₂O₃ under basic conditions in a fixed-bed continuous flow reactor for HMF oxidation using O₂ or air. Near-quantitative yields of FDCA were achieved with optimized space velocities, demonstrating industrial potential.

Despite the promising performance of Pt–Bi bimetallic catalysts in HMF oxidation [38,39], studies on Al₂O₃-supported Pt–Bi systems remain limited. γ-alumina is known for its high surface area and excellent rheological properties (e.g., high viscosity) that enable easy shaping and granulation into robust, high-surface-area supports, making it industrially vital for Pt catalysts in oxidation processes. Further catalyst stability and process economic assessments were recommended for scale-up. This work investigates HMF oxidation to FDCA using industrially relevant Pt/Al₂O₃ catalysts, with subsequent Bi modification to form bimetallic sites. Critical parameters such as metal ratio, temperature, oxygen pressure, and alkali dosage were comprehensively assessed and optimized. The changes in structural and property induced by Pt–O–Bi bimetallic formation were characterized through TEM, XRD, XPS, and BET analyses. Moreover, the promotional role of Bi was elucidated by comparative experiments with HMF and FFCA.

2. Results and Discussion

2.1. Physiochemical Properties of 5Pt/Al₂O₃ and 5Pt–nBi/Al₂O₃ Catalysts

The contents of Pt and Bi in the 5Pt/Al₂O₃ and 5Pt–nBi/Al₂O₃ catalysts were confirmed by inductively coupled plasma-atomic emission spectroscopy (ICP-OES) analysis (

Table 1. Elemental composition and pore structure characteristics of different samples.

Samples	d a (nm)	Total Content (wt%) b		DPt c	Pt0/ (Pt2+ + Pt4+) d	Bi/ B2O3 d	SBET e (m2/g)	D e (nm)	V e (cm3/g)
		Pt	Bi
5Pt/Al2O3	1.6	5.4	-	48.1%	30.6%	-	176.7	16.0	0.7
5Pt–0.5Bi/Al2O3	1.6	4.8	0.6	44.2%	32.9%	34.5%	146.7	14.2	0.6
5Pt–1Bi/Al2O3	1.8	4.7	1.2	31.9%	33.0%	31.5%	n.d. f	n.d. f	n.d. f
5Pt–3Bi/Al2O3	1.9	5.9	5.7	3.8%	35.4%	17.1%	153.9	11.6	0.5
5Pt–5Bi/Al2O3	2.3	5.7	7.2	0.6%	17.6%	14.6%	135.3	12.1	0.4
5Pt–1Bi/Al2O3 g	3.3	4.0	1.2	n.d. f	44.0%	54.8%	n.d. f	n.d. f	n.d. f

^a Average diameter of metal nanoparticles measured by HAADF-STEM; ^b Pt and Bi content determined by ICP-OES; ^c Pt dispersion determined by CO pulse chemisorption; ^d Pt and Bi species molar ratios determined by XPS; ^e Catalyst BET specific surface area S_BET, average pore size D, and pore volume V determined by N₂ physisorption; ^f Not determined; ^g regenerated after reaction.

), and the results were consistent with the calculated value in the sample synthesis, confirming the loading of Pt and Bi on the Al₂O₃ by the applied impregnation method.

The crystalline structure of γ-Al₂O₃, 5Pt/Al₂O_3, and 5Pt–nBi/Al₂O₃ was investigated by wide-angle X-ray diffraction (XRD), as shown in Figure 1. The characteristic diffraction peaks observed at 2θ = 39.5° (220), 45.9° (311), and 66.9° (440) were attributed to γ-Al₂O₃ [40]. Characteristics diffraction peaks for metallic Pt (JCPDS Card No. 88-2343) typically appear at 39.7° (111), 46° (200), and 68° (220) [41,42], while metallic Bi (ICPDS No. 85-1329) exhibits peaks at 27° (012), 38° (104), and 39° (110). Bi₂O₃ phases (ICPDS No. 27-0050) are characterized by diffraction peaks at 28° (201), 33° (220), 46° (222), and 55° (421) [43]. The patterns of all catalysts displayed the characteristic diffraction peaks of γ-Al₂O₃, which decreased gradually in intensity with the increased loading of Pt or Bi. However, the diffraction peaks of Pt and Bi were absent in all the patterns, indicating either highly dispersed Pt/Bi particles with a small size that were below the detection limit of XRD, or Pt/Bi existing in their amorphous oxidate form [34].

The morphological features of the catalysts and Pt dispersion were characterized by transmission electron microscopy (TEM), high-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM), as well as the corresponding element mapping, using 5Pt/Al₂O₃ and 5Pt–1Bi/Al₂O₃ as examples. The TEM/STEM images of 5Pt/Al₂O₃ indicated the uniform distribution of Pt nanoparticles on the γ-Al₂O₃ support, which provided a large specific surface area for Pt dispersion (Figure 2a–d). Similar fine metal nanoparticles were also observed for 5Pt–1Bi/Al₂O₃, with both Pt and Bi species uniformly dispersed throughout the support (Figure 2e–i). Metallic Bi or its oxide particles were found to coexist with Pt nanoparticles. The particle size distributions of the metal nanoparticles for different catalysts were shown in Figure 3. As reflected in these distributions, the average nanoparticle diameter increased with higher Bi loading (Table 1), a trend attributed to the surface occupancy by Bi species, which further promoted partial agglomeration. To quantitatively correlate structural changes with metal accessibility, Pt dispersions (D_Pt) were determined by CO pulse adsorption (Table 1). Consistent with the shift in particle size distribution, D_Pt was inversely related to the Bi loading. It progressively dropped from 0.48 (5Pt/Al₂O₃) to 0.32 (5Pt–1Bi/Al₂O₃), and drastically to 0.01 for the 5Pt–5Bi/Al₂O₃ catalyst (Table 1). Considering the absence of the diffraction peaks of Pt in the XRD patterns (i.e., the absence of bulky 3D Pt crystalline), despite the somewhat increased Pt particle size, Pt agglomeration should not be the main cause for the extremely decreased Pt dispersion upon high Bi addition (i.e., 5Pt–3Bi/Al₂O₃ and 5Pt–5Bi/Al₂O₃). Notably, during the preparation of the catalyst (such as impregnation and reduction) or subsequent processing, the reduced state of elemental Bi (Bi⁰) has a relatively high surface mobility. The Bi⁰ tends to selectively adsorb on the low-coordination sites of the Pt particles, such as steps, edges, and corner atoms, physically covering or “poisoning” these adsorption sites that are the most preferred and strongly adsorbed active sites for CO molecules. In addition, the electron transfer between Pt and Bi via Pt–O–Bi causes a slight displacement of the d-band center of Pt, thereby altering the bonding strength between the Pt surface and CO molecules. Therefore, the CO adsorption on Pt was largely inhibited, leading to a low Pt dispersion measured. Moreover, CO rarely adsorbs on the metal oxides; as such, the Pt dispersion is also related to the portion of the oxidates of Pt at varying Bi additions.

The textural properties of 5Pt/Al₂O₃ and 5Pt–nBi/Al₂O₃ were investigated by N₂ sorption (Figure 4). All samples exhibited typical II isotherms with H3-type hysteresis loops, indicating their mesoporous structures. Generally, a slight but logical decrease was observed in the specific surface area S_BET, pore size D, and pore volume V upon the addition of Bi (Table 1), due to the increased occupancy of the inner pores of the catalysts.

The deposition of Bi on the surface not only modulated the morphology and pore structure of the catalyst but also significantly affected the electronic state and catalytic performance of Pt through the interaction between Pt and Bi [34]. Figure 5a,b showed the full XPS spectrum of the 5Pt/Al₂O₃ and 5Pt–1Bi/Al₂O₃ catalyst, respectively, confirming the successful introduction of Pt and Bi. The chemical state of Pt was determined using Pt 4d spectral analysis due to the overlap of the Pt 4f peak with Al 2p of Al₂O₃ (Figure 5c) [44]. When the precursor H₂PtCl₆·6H₂O was impregnated onto the γ-Al₂O₃, Pt existed in the form of Pt⁴⁺, which was then converted to Pt²⁺ after drying, and finally Pt⁰ after H₂ reduction at 350 °C [45]. The Pt 4d_5/2 XPS spectra exhibited three characteristic peaks at 318.3 eV, 315.3 eV, and 312.6 eV, corresponding to Pt⁴⁺, Pt²⁺, and Pt⁰ species [32,44]. After the introduction of Bi, the Pt 4d binding energy and the contents of various Pt species showed nonmonotonic changes with increasing Bi content (Table 1). The asymmetric features of the Bi 4f spectral peaks indicated the simultaneous presence of Bi₂O₃ and metallic Bi monomers, mainly in the form of Bi₂O₃ (Figure 5d). The peaks of 159.6 eV and 164.9 eV were attributed to the Bi 4f_7/2 and Bi 4f_5/2 orbitals of Bi₂O₃ [46], while 157.6 eV and 162.8 eV correspond to the Bi 4f_7/2 and Bi 4f_5/2 orbitals of metallic Bi [32,47]. With increasing Bi incorporation, Bi tended to combine with surface oxygen or atmospheric oxygen to form thermodynamically more stable Bi₂O₃ clusters, resulting in a gradual decrease in the amount of Bi⁰ (Table 1). In the O 1s XPS spectrum, the peaks observed at 531.1 eV and 532.2 eV were attributed to lattice oxygen of Al₂O₃ and surface adsorbed oxygen species, respectively, while the peak at 530.2 eV corresponded to the lattice oxygen of Bi₂O₃ (Figure 5e) [44,47].

Given that the electronegativity of Pt (2.28) is lower than that of O (3.44), the electron density around Pt shifted toward oxygen atoms, resulting in a partial positive charge. The introduction of Bi (2.02), with its lower electronegativity, induced charge transfer from O atoms to Bi, creating a surface with positive charge bias. Specifically, electrons are transferred from Bi atoms to Pt atoms through Pt–O–Bi, leading to increased binding energy for Bi and reduced binding energy for Pt. However, variations in Bi loading alter its chemical state (Bi/Bi₂O₃), subsequently influencing the electronic structure and oxidation state of Pt through both electron transfer and geometric coverage effects. At relatively low Bi loading (5Pt–0.5Bi/Al₂O₃ and 5Pt–1Bi/Al₂O₃), dispersed Bi inhibited Pt oxidation via Pt–O–Bi interaction, leading to increased P⁰ content and decreased Bi/Bi₂O₃ ratios at higher Bi loading. At higher Bi loading (5Pt–3Bi/Al₂O₃), the Bi₂O₃ formation was promoted, and excessive Bi loading introduced additional geometric coverage on Pt species, inhibiting their reduction to Pt⁰. These complex electronic interactions inferred the critical role of Pt–Bi interaction in modulating the catalytic performance.

2.2. Catalytic Oxidation Performance of Pt/Al₂O₃

The effects of temperature, O₂ pressure, and Na₂CO₃ content on HMF oxidation over 5Pt/Al₂O₃ catalysts were exhibited in Figure 6. Product distribution analysis revealed FDCA and FFCA as the primary oxidation products. This is consistent with the fact that oxidation of FFCA to FDCA has been identified as the rate-determining step [28,48]. Reaction temperature significantly affected both product distribution and selectivity. As shown in Figure 6a, the FDCA yield continued decreasing between 80 and 120 °C, and the FFCA concentration demonstrated a sharp decline above 100 °C, suggesting accelerated side reaction kinetics at elevated temperatures. High temperature exerted a pronounced detrimental effect on reaction selectivity by inducing side reactions involving HMF, FFCA, and FDCA, such as condensation, decarboxylation, decarbonylation, and ring opening [10,12].

Pure oxygen provides a higher partial pressure of oxygen, enhancing the solubility of oxygen in alkaline solution and the coverage of active oxygen species on the catalyst surface [49]. The influence of oxygen pressure on the HMF oxidation was elucidated in Figure 6b. The dynamic equilibrium between the active oxygen coverage on the catalyst surface and intermediate adsorption competition played a decisive role in reaction selectivity. Generally, the FDCA yield decreased with the increase in O₂ pressure until reaching the lowest at 1 MPa, and then increased at higher O₂ pressure. At 0.1 MPa O₂, the moderate oxygen concentration matched the low oxygen storage of the catalyst. The density of active oxygen species on the surface was restricted, ensuring that HMF oxidation proceeded smoothly and orderly while preventing the active sites from being covered by O₂ [27]. With the increase in O₂, HMF oxidation exhibited the complexity of the kinetic competition. At 0.1–1 MPa, although oxygen mass transfer was enhanced at higher pressure, excess reactive oxygen species could disrupt the reaction network balance, leading to FFCA accumulation and triggering parallel side reactions (including HMF degradation and FDCA oxidation) [10]. Further increasing the O₂ pressure to 1.5 MPa, the FDCA yield rebounded to 35.5% accompanied by a reduction in C balance, suggesting the promotion of deep oxidation and the occurrence of overoxidation.

Alkaline-mediated HMF oxidation initiates with aldehyde hydration to form a geminal diol intermediate, followed by dehydrogenation to carboxylic acid [50]. OH⁻ serves as a proton acceptor in dehydrogenation while modulating the catalyst’s electronic state, which reduces activation energy and accelerates the oxidation kinetics of hydroxymethyl and aldehyde [33,49,50,51]. OH⁻ also enhances FDCA dissolution, thereby protecting catalyst active sites from coverage and deactivation [52]. Under base-free conditions, Pt-catalyzed oxidation proceeded primarily via hydroxymethyl dehydrogenation, with DFF and FFCA as the main reaction products [28,33,53]. In Figure 6c, limited FDCA was formed (18.9% after 6 h) due to a high proton transfer energy barrier. Optimal Na₂CO₃ content provides sufficient OH⁻ availability, thus enabling efficient dehydrogenation and intermediate desorption while preventing metal leaching and substrate degradation [49,51]. Excessive base, however, promotes side reactions, such as HMF disproportionation, furan ring cleavage [52], intermolecular condensation [33], deep oxidation pathways [10,54]. Under the optimized conditions (80 °C, 0.1 MPa O₂, and 1 equiv. Na₂CO₃), the FDCA yield was 60.6% after 12 h of reaction. Nevertheless, an incomplete carbon balance persisted due to intermediate adsorption and side reactions [27].

2.3. Catalytic Oxidation Performance of 5Pt–nBi/Al₂O₃

To investigate the promoting effects of Bi, HMF oxidation was studied over 5Pt-nBi/Al₂O₃ catalysts and compared with the performance of the 5Pt/Al₂O₃ catalysts at temperatures ranging from 80 to 120 °C (Figure 7). The introduction of Bi enhanced reaction selectivity via Pt–Bi interactions; however, excessive Bi coverage on Pt active sites inhibited the deep oxidation of intermediates. At low Bi content (n(Pt):n(Bi) = 5:0.5), limited oxygen required a higher temperature (100 °C) to achieve a peak FDCA yield of 64.8% (Figure 7b), though selectivity descended at elevated temperatures. The optimal Pt–Bi synergistic effect was achieved at n(Pt):n(Bi) = 5:1, where Bi modulated the electronic state of Pt and provided sufficient reactive oxygen species, ensuring efficient HMF activation while preventing active site blockage caused by excessive adsorption. The highest FDCA yield (67.6%) and overall selectivity (89.2%) were achieved at 80 °C (Figure 7c), demonstrating Bi’s effect on lowering the energy barrier for FFCA oxidation to FDCA. Further increasing Bi promoted Bi₂O₃ formation, partially covering the Pt site, and reducing active oxygen availability. Although overall selectivity reached 91.1%, FDCA yield declined to 45.1% at the optimal temperature of 90 °C (Figure 7d). At the Pt/Bi molar ratio of 5: 5, excessive Bi severely suppressed catalytic activity, yielding only 26.6% FDCA despite 90.5% selectivity (Figure 7e), indicating oxidation stagnation due to Pt site blockage. In general, lower temperature (80–90 °C) favored high selectivity by stabilizing intermediates and optimizing adsorption–desorption kinetics [34]. The 5Pt–1Bi/Al₂O₃ exhibited superior performance at 80 °C, highlighting the importance of precise Bi doping to tailor temperature-responsive catalytic behavior.

To improve HMF oxidation efficiency and FDCA yield, process conditions over the 5Pt–1Bi/Al₂O₃ catalyst at 80 °C were further investigated. HMF oxidation results at different oxygen pressures (0.1–1.5 MPa) were shown in Figure 8a. Under the low pressure of 0.1 MPa, incomplete HMF conversion was observed (58.1% FDCA yield after 6 h), primarily due to oxygen mass transfer limitations leading to insufficient reactive oxygen species. At oxygen pressure over 0.5 MPa, the HMF was completely converted within 6 h, and the selectivity was significantly increased to above 91%, indicating that the system entered the surface reaction control region. This elevated-pressure environment facilitated oxygen dissolution and maintained optimal surface oxygen concentration. Notably, increasing pressure from 1 to 1.5 MPa produced marginal FDCA yield enhancement, suggesting saturation of Bi-mediated oxygen adsorption capacity. Oxygen supply and dehydrogenation rates reached an optimal balance, establishing a stable tandem reaction network. Meanwhile, the competitive adsorption of Bi effectively inhibited the attack of reactive oxygen species on the furan ring of FDCA, avoiding the over-oxidation side reactions. Further pressure increases the risk of disrupting this equilibrium by exceeding the alcohol dehydrogenation rate, potentially causing active site over-oxidation and requiring optimization through reduced oxygen concentration or elevated temperature [51].

On the basis of 1.5 MPa O₂, the effect of Na₂CO₃ content on HMF oxidation was investigated (Figure 8b). HMF oxidation in an alkali-free system followed the intermediate pathway of DFF and exhibited lower catalytic activity (38.5% HMF conversion, 13.0% FDCA yield). Under alkaline conditions, the reaction pathway shifted toward preferential oxidation of the aldehyde group, generating HFCA as an intermediate and ultimately increasing FDCA yield. The system achieved a maximum FDCA yield of 94.1% at the Na₂CO₃ loading of two equivalents relative to theoretical FDCA. At the condition of 4 equiv. Na₂CO₃, the selectivity decreased slightly despite maintaining a comparable FDCA yield (93.9%), suggesting the onset of side reactions induced by excess OH⁻. Consequently, 2 equivalents of Na₂CO₃ were identified as the optimal alkali concentration.

Figure 8c displayed the HMF conversion and oxidation product distribution as a function of reaction time under the optimized conditions. The HMF conversion reached completion within 0.5 h, and the carbon balance stabilized after 2 h. The FDCA yield exhibited a continuous increase with prolonged reaction time, attaining its maximum value at 6 h. For comparison, HMF was oxidized under the same conditions for 6 h using the 5Pt/Al₂O₃ catalyst, which afforded FDCA and FFCA in yields of 56.8% and 22.8%. This result underscores the superior efficacy of the bimetallic Pt–Bi/Al₂O₃ catalyst. Figure 8d illustrated the recycling stability of the 5Pt–1Bi/Al₂O₃ catalyst under optimized reaction conditions (80 °C, 1.5 MPa O₂, 6 h, 2 equiv. Na₂CO₃). The FDCA yield progressively declined to 63.2% by the fourth cycle, accompanied by an increasing FFCA yield and a reduced carbon balance. Nevertheless, the performance was still superior to that of the monometallic Pt/Al₂O₃ catalyst. The decline of catalyst performance may be due to Pt leaching under alkaline conditions, as indicated by ICP analysis revealing a decrease in Pt content to 4.0 wt% after the first cycle, while Bi remained relatively stable at 1.2 wt%. Moreover, sintering of metal nanoparticles during regeneration also contributed to the deactivation. This has been reflected in the increased average metal particle size for the recovered 5Pt–1Bi/Al₂O₃ catalysts, which increased from 1.8 to 3.3 nm after reaction (Figure S6). Simultaneously, the Pt⁰ portion increased from 33.0% to 44.0%, and the Bi/B₂O₃ ratio from 31.5% to 54.8%, due to the partial reduction during the regeneration procedure.

Furthermore, the catalytic performances of 5 wt% Pt/Al₂O₃ and 5Pt–1Bi/Al₂O₃ for HMF and FFCA oxidation were evaluated by constructing time-dependent product profiles under continuous oxygen flow (conditions of 80 °C, 1equiv. Na₂CO₃, and 10 mL/min O₂). As shown in Figure 9a,b, HMF oxidation exhibited typical tandem reaction characteristics, in which HMF was first rapidly converted to HFCA and DFF intermediates, followed by the further oxidation to generate FFCA. The oxidation of FFCA to FDCA represented the rate-determining step, and FDCA generation was positively correlated with the FFCA concentration. The initial carbon balance discrepancy primarily stemmed from the adsorption of intermediates on the catalyst surface. These adsorbed intermediates were gradually consumed as the reaction proceeded and were converted to detectable salts, which stabilize the molar balance [27]. At the late stage of the reaction, FFCA consumption and FDCA generation were essentially conserved, indicating good stability of both under alkaline conditions.

The appropriate amount of Bi in the catalyst significantly enhanced the catalytic performance, which not only accelerated the oxidation efficiency of HMF but also enhanced the deep oxidation ability of the intermediate products. Complete conversion of HMF by 5Pt–1Bi/Al₂O₃ was achieved within just 1 h. The yield of FFCA reached its peak at 0.5 h, and a 94.0% yield of FDCA was obtained within 4 h. The accelerating effect of Bi on the key steps was further confirmed by the oxidation experiments using FFCA as the substrate (Figure 9c,d). 5Pt/Al₂O₃-catalyzed FDCA yield was only 85.6% after 5 h of FFCA oxidation, which was extended to 12 h and then only increased to 89.6% (Figure 9c). In contrast, the 5Pt–1Bi/Al₂O₃ catalyst achieved a 96.5% FDCA yield within 3 h, and near-complete conversion was attained after 12 h. The maximum FDCA yield was nearly 100%, with the carbon balance approaching closure (Figure 9d). This result indicated that the introduction of an appropriate amount of Bi not only dramatically enhanced the reaction rates but also optimized the selectivity of the reaction pathway by significantly shortening the reaction time of tandem oxidation and avoiding the side reactions occurring due to the accumulation of intermediates, which provided an important kinetic basis for the design of high-efficiency catalysts for FDCA synthesis.

3. Experimental

3.1. Materials

5-hydroxymethylfurfural (HMF; Aladdin, Shanghai, China; 99%), 2,5-furandicarboxylic acid (FDCA; Adamas, Basel, Switzerland; ≥99%), 5-hydroxymethyl-2-furancarboxylic acid (HFCA; Aladdin; 98%), 2,5-diformylfuran (DFF; Aladdin; 98%), 5-formyl-2-furancarboxylic acid (FFCA; Aladdin; 98%), sodium carbonate (Na₂CO₃; Macklin, Shanghai, China; >99.9%), chloroplatinic acid hexahydrate (H₂Pt₆Cl₆·6H₂O; Alfa Aesar, Ward Hill, MA, USA; 99.95%); bismuth nitrate pentahydrate (Bi(NO₃)₃·5H₂O, Aladdin; >99.0%), γ-alumina (γ-Al₂O₃; Macklin; 99.99%), phosphoric acid (H₃PO₄; Macklin; for HPLC, 85–90%), nitric acid (HNO₃; Titan, Shanghai, China; AR, 65–68%), high purity O₂, (PuJiang, Shanghai, China; 99.9%), high purity Ar (Air Liquide, Shanghai, China; 99.9%), 10% H₂/Ar (Air Liquide; 99.9%). All chemicals were used as received unless otherwise mentioned.

3.2. Preparation of 5Pt/Al₂O₃ and 5Pt–nBi/Al₂O₃ Catalysts

The Pt/Al₂O₃ catalyst containing 5 wt% of Pt (denoted as 5Pt/Al₂O₃) was synthesized via a wet impregnation method. Typically, 1 g of H₂Pt₆Cl₆·6H₂O was initially dissolved in 50 mL of deionized water to obtain a Pt precursor solution. The Pt content in the solution was determined by ICP analysis to be 0.71 wt%. Subsequently, 7.04 g of the Pt precursor solution was impregnated onto 0.95 g γ-Al₂O₃ under ultrasonic treatment to ensure homogeneous mixing. The suspension mixture was dried in an oven at 60 °C for 12 h, and the obtained solid was ground and then calcined at 450 °C for 3 h under an air atmosphere. Reduction was carried out at 350 °C for 2 h under a 10% H₂/Ar atmosphere to obtain the 5Pt/Al₂O₃.

The 5Pt–nBi/Al₂O₃ catalysts with varying Pt/Bi molar ratios (defined as 5/n; n = 0.5, 1, 3, 5) were synthesized following a similar wet impregnation method. Typically, for the 5Pt–1Bi/Al₂O₃ catalysts, 0.025 g of Bi(NO₃)₃·5H₂O was first dissolved in a small amount of nitric acid. Then, 7.04 g of the Pt precursor solution and 0.94 g γ-Al₂O₃ were added, followed by homogeneous ultrasonic mixing. The mixture was dried at 60 °C for 12 h, and the obtained solid was ground and calcined at 450 °C for 3 h in air. Reduction was carried out at 350 °C for 2 h under a 10% H₂/Ar atmosphere to obtain the 5Pt–1Bi/Al₂O₃ catalyst, with a total mass of 1 g and a Pt mass fraction of 5 wt%. The 5Pt–0.5Bi/Al₂O₃, 5Pt–3Bi/Al₂O₃, and 5Pt–5Bi/Al₂O₃ were prepared similarly by varying the masses of Bi(NO₃)₃·5H₂O and γ-Al₂O₃.

For the reusability test, the spent catalyst was recovered by centrifugation, washing three times with deionized water and ethanol, respectively, and drying at 80 °C for 12 h. The dried catalyst was then ground and subjected to calcination at 450 °C for 3 h (air atmosphere), followed by reduction under 10% H₂/Ar at 350 °C for 2 h.

3.3. Oxidation of HMF to FDCA

The oxidation reaction was conducted in a 50 mL parallel autoclave reactor equipped with a heating block and magnetic stirring system. A standard HMF solution (5 mg/mL) was prepared by dissolving 0.505 g of HMF (99% purity) in 100 mL of Milli-Q water. The reactor was charged with 0.062 g of 5 wt% Pt/Al₂O₃ catalyst, 0.084 g of Na₂CO₃ (>99.5%), and 20 mL of the HMF solution. The system was purged by performing five consecutive pressurization-depressurization cycles with oxygen to displace residual air, followed by pressurization with O₂ to the desired gauge pressure. The reaction was initiated once the solution temperature reached the predetermined set point, controlled by a heating jacket.

The oxidation experiments under continuous oxygen supply at atmospheric pressure were conducted in a 100 mL three-neck round-bottom flask. The reaction system was assembled as follows: 0.062 g of 5 wt% Pt/Al₂O₃ or 5Pt–1Bi/Al₂O₃, 0.084 g of Na₂CO₃, and 20 mL of HMF standard aqueous solution (5 mg/mL) were sequentially added to the flask. The left neck was connected to a glass vacuum adapter, the center neck was equipped with a condenser attached to an exhaust line, and the right neck was fitted with a needle for continuous bubbling of high-purity oxygen into the bottom of the reaction solution.

For the purification of FDCA, the post-reaction mixture was centrifuged, and the supernatant was collected and subsequently acidified with hydrochloric acid, resulting in the formation of a white precipitate. The product was collected by filtration, washed thoroughly with deionized water (three times), and dried under vacuum at 80 °C for 10 h. The purified solid was characterized by nuclear magnetic resonance (1H NMR) spectroscopy in deuterated dimethyl sulfoxide (DMSO-d6; cf. typical results in Figure S1 in Supporting Information).

3.4. Characterization and Analysis

Transmission electron microscopy (TEM) images and high-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM) images, along with energy-dispersive X-ray spectroscopy (EDS), were performed using a JEM-F200 field-emission transmission electron microscope (JEOL, Tokyo, Japan) operated at 200 kV. X-ray diffraction (XRD) patterns were collected on a Bruker D8 ADVANCE diffractometer (Bruker, Berlin, Germany), scanning from 10 to 80° (2θ) at a rate of 2°/min. X-ray photoelectron spectroscopy (XPS) measurements were taken on a Thermo Fisher K-Alpha instrument (Thermo Fisher, Waltham, MA, USA) with Al Kα X-ray radiation. Elemental analysis was quantified by inductively coupled plasma optical emission spectrometry (ICP-OES, Agilent 5800, Santa Clara, CA, USA). CO pulse chemisorption was carried out on a AutoChem III 2930 analyzer (Micromeritics, Norcross, GA, USA) to determine metal dispersion. N₂ physisorption was measured at 77K using a TriStar II 3020 (Micromeritics, USA) to derive the Brunauer–Emmett–Teller (BET) surface area and pore size distribution.

The content of HMF and its derivatives in samples was analyzed by high-performance liquid chromatography (Waters Arc HPLC, Milford, MA, USA) equipped with an ultraviolet-visible spectroscopy detector (UV 2489) on an Aminex HPX-87H column (300 mm × 7.8 mm, Bio-Rad, Hercules, CA, USA). H₃PO₄ aqueous solution (pH = 2) was used as the mobile phase at a flow rate of 0.6 mL/min at 65 °C. The liquid sample was diluted with mobile phase and filtered with a 0.22 µm nylon syringe filter prior to analysis. Specific wavelengths were used for examining different compounds: 320 nm for HMF solution; 290 nm for HFCA, DFF, FFCA, FDCA, and HMF. The quantitative analysis was conducted using the external calibration method. The conversion of HMF and the yield of FDCA/FFCA/HFCA/DFF were calculated according to Equations (1)–(3):

$${\mathrm{Conv}.}_{\mathrm{HMF}}(\%) = (1-{\displaystyle \frac{\mathrm{moles} \mathrm{of} \mathrm{HMF}}{\mathrm{initial} \mathrm{mole} \mathrm{of} \mathrm{HMF}}}) \times 100\%$$

(1)

$${\mathrm{Yield}}_x(\%)={\displaystyle \frac{\mathrm{moles} \mathrm{of} x}{\mathrm{initial} \mathrm{mole} \mathrm{of} \mathrm{HMF}}} \times 100\%$$

(2)

$$\mathrm{C} \mathrm{balance}={\displaystyle \frac{\mathrm{moles} \mathrm{of} \mathrm{C} \mathrm{in} \mathrm{the} \mathrm{products}}{\mathrm{initial} \mathrm{mole} \mathrm{of} \mathrm{C}}} \times 100\%$$

(3)

where x is FDCA, FFCA, HFCA, DFF.

4. Conclusions

In this study, the oxidation of HMF to FDCA was investigated over Pt/Al₂O₃ and 5Pt-nBi/Al₂O₃ catalysts. Comprehensive characterization and experimental analyses revealed that the introduction of Bi species modified the catalyst’s morphology, pore structure, and chemical states, inducing a Bi–O–Pt interaction that altered the electronic state of Pt. This effect enhanced the adsorption and conversion of HMF and the key intermediate FFCA while effectively suppressing undesirable side reactions arising from intermediate accumulation. Under optimized conditions, the 5Pt/Al₂O₃ catalyst achieved an FDCA yield of 60.6% after 12 h, whereas the bimetallic 5Pt–1Bi/Al₂O₃ catalyst significantly improved performance, delivering a 94.1% FDCA yield within 6 h. Additionally, continuous O₂ supply under ambient pressure further validated the superior catalytic performance of the Pt–Bi system, achieving an FDCA yield of 94.0% in 4 h at 80 °C with one equiv. Na₂CO₃. These results highlighted the promising potential of Pt–Bi bimetallic catalysts for efficient FDCA production.