Recent Progress of Advanced Biofuel 2,5-Dimethylfuran Production from 5-Hydroxymethylfurfural

Abstract

Given the depletion of fossil resources and mounting environmental pressures, the efficient conversion of the biomass-derived platform molecule 5-hydroxymethylfurfural (HMF) into the liquid fuel 2,5-dimethylfuran (DMF) is of critical strategic importance. The recent advances in the hydrogenolysis of HMF to DMF were systematically summarized in this review. The performance advantages and limitations of monometallic catalysts, including noble metals and non-noble metals, were discussed. The bimetallic active centers engineered through synergistic effects to enhance activity were summarized. Crucially, the roles of the physicochemical properties of catalyst supports and the hydrogen donors in governing reaction pathways and efficiency were also analyzed in depth. Finally, future research directions were proposed to address current challenges related to catalyst durability and economic viability.

1. Introduction

In light of increasing fossil fuel consumption and growing environmental concerns, renewable and clean energy have attracted considerable attention as a viable alternative to conventional fossil resources [1,2]. Biomass, as the only renewable natural carbon source, has garnered significant research interest due to its abundance and carbon-neutral characteristics [3,4]. In contrast to fossil fuels, biomass is composed of polar, reactive, and oxygen-containing functional moieties [5,6]. Therefore, the catalytic conversion of biomass into biofuels and high-value chemicals offers a promising route to reduce dependence on fossil resources [7,8,9]. The platform chemical 5-hydroxymethylfurfural (HMF), typically derived from biomass, serves as a crucial bridge between raw biomass and bio-based chemicals [10,11,12,13,14,15].

HMF molecule contains aldehyde, hydroxymethyl, and furan ring functionalities, enabling its further conversion into a wide array of high-value chemicals via hydrogenation, oxidation, amination, etherification, or HDO [16,17]. Among these, the selective hydrogenolysis of HMF to the liquid biofuel 2,5-dimethylfuran (DMF) is of particular significance [18,19,20]. The properties of DMF are similar to gasoline, including a high octane number (119) and energy density (31.5 MJ·L⁻¹), and it also serves as a common intermediate in the chemical industry [21,22]. Moreover, the boiling point of DMF is between that of ethanol and butanol, and readily miscible with gasoline. Its favorable vaporization characteristics help reduce air intake resistance in engines, satisfy cold-start requirements, and perform well in direct-injection and spark-ignition engines [23]. Consequently, the production of DMF is of great importance for advancing the renewability of liquid fuels [24,25].

Two main reaction pathways for the conversion of HMF to DMF were reported (Figure 1). In pathway A, the hydroxymethyl group of HMF undergoes hydrogenolysis and dehydration to form 5-methylfurfural (5-MF), followed by hydrogenation of the aldehyde group to yield 5-methylfurfuryl alcohol (MFA). In pathway B, the aldehyde group of HMF is first hydrogenated to form 2,5-bis(hydroxymethyl)furan (BHMF), after which the hydroxymethyl groups undergo hydrogenolysis to produce MFA as the key intermediate. In both pathways, MFA is subsequently converted to DMF through further hydrogenolysis of its hydroxyl group. Additionally, due to the reactive nature of the furan ring, side reactions such as over-hydrogenation and hydrolytic ring-opening may occur, leading to byproduct production including 2,5-dimethyltetrahydrofuran (DMTHF), 2,5-bis(hydroxymethyl)tetrahydrofuran (BMTHF), 5-methyltetrahydrofurfural alcohol (MTHFA), and 2,5-hexanedione (HD) [26].

In recent years, numerous studies have been devoted to the selective hydrogenation of HMF to DMF. The structure–activity relationship between catalyst supports and active centers, as well as the primary functions and synergistic interactions of catalytic active sites in HMF hydrogenation, were systematically summarized in this review. The influence of hydrogen donors on the hydrogenation process was also discussed. In addressing the key challenges in current research, future directions for catalyst and reaction system design to support the scaled and renewable production of liquid fuels were proposed.

While DMF has garnered the majority of research attention owing to its exceptional octane number (~119) and high energy density (30 MJ/L), the economics of its production require explicit, direct comparison with those of other furanic biofuels. 2-Methylfuran (2-MF) is a related furanic biofuel produced from furfural, a far lower-cost and industrially mature platform molecule. 2-MF benefits from a simpler hydrogenolysis pathway, which requires fewer H₂ equivalents per mole of product. This simplified production route positions 2-MF much closer to commercial viability than DMF. By contrast, DMF production remains burdened by three core cost drivers. First is the high cost of the HMF feedstock. Second is the reliance on noble metals in most high-performing catalytic systems. Third is the stoichiometric consumption of sacrificial hydrogen donors, such as polymethylhydrosiloxane (PMHS). All of these factors introduce cost barriers that are far less pronounced in 2-MF synthesis. The production cost of DMF has also been benchmarked against fossil-derived gasoline. This cost estimate accounts for all key process steps: feedstock pretreatment, catalyst materials, solvent recovery, and product purification. Under this full process accounting, DMF production costs remain at least 1–2 orders of magnitude higher than those of fossil gasoline. This cost disparity cannot be offset by the combustion performance advantages of DMF alone.

Although HMF-to-DMF conversion has not yet been commercialized, several catalyst families are already widely used in related industrial hydrogenation and hydrodeoxygenation processes. These established industrial catalysts include four primary categories: Raney Ni [23], Copper chromite/Cu–Cr oxides such as Cu₂Cr₂O₅ [27], Ni/kieselguhr and Ni/SiO₂ catalysts [27], and supported noble metal catalysts including Pd/C, Pt/C, and Ru/C [23]. These mature catalytic systems serve as valuable industrial benchmarks for HMF-to-DMF process development. However, their direct application to this specific reaction remains limited by multiple critical drawbacks. These limitations include insufficient product selectivity, over-hydrogenation of the furan ring, high noble metal costs, catalyst deactivation in oxygenated biomass-derived media, Cr-related environmental concerns, and the need for harsh sulfiding conditions. Therefore, the development of more selective, stable, and economically viable catalysts remains an essential prerequisite for advancing the practical, large-scale production of DMF from HMF.

Several reviews have discussed the catalytic conversion of HMF into liquid fuels, DMF. For example, Mäki-Arvela et al. provided a comprehensive overview of the catalytic performance for HMF hydrogenolysis in 2021 [19]. Their work emphasized the critical roles of metal and solvent selection, the application of homogeneous acidic co-catalysts, and the operational challenges associated with over-hydrogenation, continuous processing, and catalyst deactivation. Recently, He et al. [28] summarized key advancements in noble and non-noble metal-based heterogeneous catalysts for this filed. Their review focused on the correlation between catalyst structures and reaction mechanisms at active sites, while also prospecting the integration of theoretical calculations and microkinetic modeling. Despite these valuable prior contributions, rapid progress has been made in this field in recent years. The advances in this review were included in non-noble metal catalysts, bimetallic active centers, support engineering, catalytic transfer hydrogenation, and integrated conversion strategies. Accordingly, the interplay effect among metal sites, support properties, and hydrogen donors were emphasized. Through this targeted analysis, we provide a more integrated, actionable framework for future catalyst design and industrial-scale implementation.

2. Metal Active Sites

2.1. Single-Metal Active Site Catalysts

Given the diversity of products arising from the hydrogenation and hydrogenolysis of HMF, achieving highly selective hydrogenolysis to DMF remains a significant challenge. Recent studies on single-metal active site catalysts for the selective hydrogenation of HMF to DMF was summarized (

Table 1. Single-metal active site catalysts for the conversion of HMF to DMF.

Entry	Catalyst and Its Loading a	Solvent	P. b /MPa	Temp. /°C	Time /h	Conv. /%	S. c /%	Y. d %	Productivity e/molDMF h−1 mol−1metal_total	Catalyst Stability f	TOF g/h−1	Ref.
1	Br-Pd/Al2O3, 33.75	THF h	0.5	30	6	100	96	96	5.40	~85 (5 runs)	181.7	[29]
2	Pd/0.15P-TiO2-500, 425.68	THF	1	150	4	94	71	67	71.30	~20 (4 runs)	58.7	[30]
3	Pd-UiO-66, 193.49	THF	1.5	160	3	100	92	92	59.34	~31 (5 runs)	42.66	[6]
4	Pd/NMC, 336.06	FA i	0.5	160	3	99	97	97	108.66	95.1 (5 runs)	150	[31]
5	Pd/MOF-808, 27.67	THF	1	100	3	100	99	99	9.13	~95 (5 runs)	152 j	[32]
6	Ru–N-CMK-1, 40.43	Isopropanol	-	160	8	100	88	88	4.45	~82 (3 runs)	4.70	[33]
7	Ru/ZSM-5, 160.29	Ethanol	1.7	180	3	98	97	95	50.76	76.5 (5 runs)	NA k	[34]
8	Pt/Co2AlO4, 3867.26	Isopropanol	2	180	3	>99	99	99	1276.20	~45 (3 runs)	2.55	[35]
9	Co@NGs-700, 29.47	Ethanol	2	200	6	100	95	95	4.67	~63 (5 runs)	4.73	[36]
10	Co/SBA-15, 8.97	THF	2	150	2	100	96	96	4.30	~85 (5 runs)	NA	[37]
11	Co@BNG, 34.91	THF	1	140	2.5	100	99	99	13.83	~95 (5 runs)	13.85	[38]
12	Co/Sn-Beta, 11.23	THF	1.5	170	14	100	>99	>99	0.79	~58 (4 runs)	0.83	[39]
13	Co/CFA, 1.96	THF	2	140	6	99	98	97	0.32	~90 (5 runs)	NA	[40]
14	HCP-Co, 35.33	THF	2	180	2	100	97	97	17.13	~85 (5 runs)	10.2	[41]
15	Co/N-C, 4.71	THF	0.25	150	4	99	99	98	1.16	~62 (5 runs)	NA	[42]
16	Co/Beta-DA, 5.89	THF	1	150	3	>99	>99	>99	1.94	~40 (4 runs)	1.97	[43]
17	Co@CoO, 9.74	THF	1	130	2	>99	89	89	4.33	NA	4.26	[44]
18	Cu/PBSAC, 9.51	Isopropanol	1	190	6	100	>98	>98	1.55	NA	NA	[22]
19	Cu-CPM, 0.88	Ethanol	1.5	200	3	100	96	96	0.28	~90 (5 runs)	NA	[45]
20	Ni-C3N4/HC, 54.11	THF	1.5	190	4	>99	94	94	12.72	~92 (4 runs)	NA	[46]
21	Ni/ZSM-5, 2.93	THF	0.25	180	7	91	96	87	0.36	~42 (5 runs)	0.40	[47]
22	Ni/WC, 8.16	THF	3	180	3	100	98	98	2.67	~70 (5 runs)	11.3	[48]
23	Ni-SCM-14, 226.15	THF	1	190	2	>99	90	90	11.77	~65 (5 runs)	NA	[49]
24	N-C, 139.60	n-Butanol	4	240	12	100	91	91	10.59	~10 l (5 runs)	~12	[50]
25	CA-Ni/ZrO2, 14.67	THF	0.5	180	6	98	99	97	2.37	~80 (5 runs)	0.79	[51]
26	Ni/HSAG, 5.87	1-Butanol	3	220	4	100	65	65	0.95	NA	1.33	[52]
27	Cu/HSAG, 5.58	1-Butanol	3	180	4	94	98	92	1.28	~90 (2 runs)	6.76	[52]
28	MoNC, 99.86	2-PrOH	2	200	12	100	91	91	7.57	NA	NA	[53]
29	Mo@C-900, 247.03	Ethanol	-	180	10	100	92	92	22.73	~74 (4 runs)	102.83	[54]
30	Fe/TiO2, 106.44	H2O	0.6	50	1.5	98	93	91	64.57	~90 (5 runs)	NA	[55]
31	Co/Mg-Beta, 10.81	THF	2	170	1.5	100	100	100	7.21	~100 (5 runs)	NA	[56]

^a The catalyst loading refers to the molar ratio of HMF to metal; ^b P./MPa is now defined as H₂ pressure (MPa); ^c S./% is defined as selectivity to DMF (%); ^d Y.% is defined as yield to DMF (%); ^e $\mathrm{P}\mathrm{r}\mathrm{o}\mathrm{d}\mathrm{u}\mathrm{c}\mathrm{t}\mathrm{i}\mathrm{v}\mathrm{i}\mathrm{t}\mathrm{y}={\displaystyle \frac{{\mathrm{n}}_{\mathrm{D}\mathrm{M}\mathrm{F}}}{{\mathrm{n}}_{\mathrm{m}\mathrm{e}\mathrm{t}\mathrm{a}\mathrm{l}\_\mathrm{t}\mathrm{o}\mathrm{t}\mathrm{a}\mathrm{l}}\times \mathrm{t}}}$; ^f Product yield after a certain number of recycling experiments under the optimal conditions; ^g TOF number toward DMF formation was calculated based on the number of active sites reported in the literature; ^h THF is tetrahydrofuran; ⁱ FA is formic acid; ^j The TOF number was related to HMF consumption; ^k NA = not available; ^l Reaction time is 2 h.

).

2.1.1. Single Noble Metal Active Site Catalysts

Noble metals such as Pd, Ru, and Pt have been reported for DMF production. Owing to their partially filled d-orbitals, noble metal catalysts readily adsorb reactants and facilitate the formation of highly active intermediates, thereby exhibiting excellent catalytic performance [57]. They also possess high thermal stability and corrosion resistance, which are usually used as common catalysts in early studies [58,59,60]. Pd⁰ exhibits good activity, along with high catalytic performance and excellent recyclability in HDO reactions. To achieve a close integration of metallic and acidic properties on the surface of metal nanoparticles, a Pd–Br bifunctional catalyst featuring halogen-modified noble metal nanoparticles was developed [29]. This catalyst generated electronegative Br atoms through the hydrogenolysis of bromobenzene, which further modified Pd to create strong Brønsted acid sites, enabling efficient conversion of HMF to DMF at room temperature with a yield of 96% [29]. Under the influence of the halogen bromine (Br), the catalyst exhibited an elevated turnover frequency (TOF) of 181.7 h⁻¹. TOF is a core catalytic metric that represents intrinsic catalytic activity, normalized to the number of active sites in a material. This metric reflects how efficiently each active site converts reactants to products per unit time. Valid TOF measurements must be collected under conditions where mass and heat transfer limitations have been rigorously excluded. TOF serves as the most fundamental metric for benchmarking catalysts with different compositions, structures, and active site densities. It enables meaningful, standardized comparisons between distinct catalytic systems. These head-to-head comparisons are essential for establishing reliable structure–performance relationships in heterogeneous catalysis [61]. Furthermore, the synergy between metal sites and acid sites is critical for H₂ activation and C-O bond hydrogenolysis [62,63,64]. For example, hydrogen spillover and reactant adsorption could be enhanced when phosphorus was doped into a TiO₂ support (Figure 2) [30]. The spilled-over active hydrogen was used for a P-doped Pd/P-TiO₂ catalyst to induce Brønsted acid sites, and C-O bond activation was further promoted. A 67% DMF yield could be achieved from HMF by P-doped Pd/P-TiO₂ catalyst at 150 °C and 1 MPa H₂. Although it has a high TOF number (58.7 h⁻¹), its cycle stability is extremely poor.

Pd is generally regarded as the optimal metal for formic acid decomposition, and high activity is exhibited for HMF hydrogenolysis when formic acid (FA) is used as the hydrogen source. For example, a Pd/NMC catalyst (Figure 3) was employed in FA solvent, achieving a complete HMF conversion and 97% DMF yield at 160 °C and 0.5 MPa H₂ with a superior DMF productivity (108.66 mol_DMF h⁻¹ mol⁻¹_{metal_total}), high catalytic stability (95.1% yield after five runs), and minimal Pd leaching confirmed by ICP-OES analysis of the post-reaction solution and TEM imaging showing preserved particle size [31]. This study further revealed that the reactivity of HMF was modulated, and the reaction pathway was also influenced by suppressing ring hydrogenation and acting as a mild reducing agent. Crucially, catalyst stability remains a major unresolved issue for the Pd-catalyzed hydrogenolysis of HMF. The Pd/MOF-808 catalyst achieved a quantitative HMF conversion and >99% DMF yield under mild reaction conditions (100 °C, 1.0 MPa H₂) [32]. Despite this excellent catalytic performance, the activity of the Pd/MOF-808 catalyst was highly sensitive to the nature of the support and the reaction medium. This performance instability was caused by the structural collapse of the MOF support under hydrothermal reaction conditions, which was confirmed by X-ray diffraction (XRD) analysis. MOF-supported and N-doped carbon-supported catalytic systems face persistent challenges during prolonged operation. The most critical of these are hydrothermal instability and pore structural collapse. These stability issues are less pronounced in robust carbon-encapsulated catalyst architectures. Even so, stability limitations remain a central concern for the practical implementation of these catalytic systems.

Ru-based catalysts are less expensive and represent one of the most effective systems for the selective conversion of furanic compounds [65,66]. The good hydrogenation activity, even in acidic hydrothermal environments, is retained for Ru-based catalysts, and they have garnered considerable attention. For example, Ru nanoparticles supported on nitrogen-doped mesoporous carbon could efficiently catalyze the hydrogenolysis of HMF to DMF (88% yield). It was found that the nitrogen-doped carbon support provided suitable basic sites that not only promote the dispersion of Ru nanoparticles but also effectively activate the hydroxyl groups of isopropanol, thereby releasing active hydrogen species [33]. Another study showed that a Ru/ZSM-5 catalyst was employed for DMF production, and 95% DMF yield was observed due to the Brønsted acidity of Ru/ZSM-5 and its affinity toward C-O bonds [34]. Despite this advantage, Ru catalysts supported on oxide supports exhibit a pronounced tendency toward over-hydrogenation. When reaction times are extended beyond the optimal window, substantial amounts of byproducts 2,5-dimethyltetrahydrofuran (DMTHF) is usually generated with Ru-based catalysts [67]. This undesired catalytic behavior is intrinsically linked to the oxophilic nature of Ru. Specifically, this oxophilicity drives a strong, direct interaction between Ru metal and the furan ring of the reaction intermediate. These mechanistic insights indicate that two key strategies are required to direct selectivity exclusively toward DMF: precise control of metal–support interactions and the introduction of selective site-blocking agents.

Pt-based catalysts are similar to Ru-based systems, exhibiting unique catalytic properties and high selectivity for C=O hydrogenation. This is attributed to the large radial extension of their d-orbitals [68]. However, the high activity of Pt often makes the hydrogenation rate difficult to control, leading to over-hydrogenation side reactions, which limit its application in the HDO of HMF. In recent years, single-atom catalysts (SACs) have attracted considerable attention due to their exceptional atom efficiency and unique geometric and electronic structures [69,70,71]. For instance, a single-atom Pt catalyst anchored on a Co₂AlO₄ spinel support was developed. This catalyst achieved a DMF yield of 99% at 180 °C under 2 MPa H₂ for 3 h with extremely high DMF productivity (1276.20 mol_DMF h⁻¹ mol⁻¹_{metal_total}). Unfortunately, it lacks good cyclic stability. In situ studies and theoretical calculations revealed that the activation and adsorption of C=O bonds and the cleavage of the first C-OH bond were promoted by the single-atom Pt sites, while the dissociation of the second C-O bond was accelerated by the adjacent Brønsted acid sites [35].

Single-metal catalysts play a fundamental and critical role in the selective hydrogenation of HMF to DMF, with their performance strongly dependent on the intrinsic properties of the metal center. The excellent adsorption and activation capabilities toward reactants were exhibited by the noble metal catalysts (e.g., Pd, Ru, Pt). It can be attributed to their partially filled d-orbitals, enabling high HMF conversion and DMF selectivity under relatively mild conditions. Through precise tuning of the active center microenvironment, such as constructing metal-acid bifunctional sites, exploiting strong metal–support interactions, or designing atomically dispersed single-atom catalysts, the reaction pathway can be further optimized, and DMF yields were increased to near-quantitative levels. However, the high cost of noble metals and their potential to induce deep hydrogenation remained major barriers to large-scale industrial application.

2.1.2. Single Non-Noble Metal Active Site Catalysts

Searching for low-cost catalysts has been a key research focus for DMF production. The development of low-cost, efficient, and environmentally friendly non-noble metal catalysts to replace noble metal systems is crucial for the economic viability of this process. Among these, Co-based catalysts have attracted considerable attention due to their low cost, high abundance, and good activity for C=O bond hydrogenation. Their catalytic performance can be significantly enhanced through electronic structure modulation. For example, a B,N-codoped multilayer graphene-encapsulated Co-based catalyst (Co@BNG) was prepared for hydrogenolysis of HMF to DMF [38]. It was shown that electron transfer from Co to neighboring B and N atoms was facilitated by the B doping, forming electron-deficient Co centers (Co-N_x/B) and possessing a high TOF number (13.85 h⁻¹). These centers were acted as Lewis acid sites and C=O and C-OH bonds were preferentially activated, accelerating the hydrogenation of HMF to BHMF and its subsequent hydrogenolysis to DMF. The activation energy barrier for the rate-determining step (BHMF → MFA) was significantly reduced for the Co–B₁N₃C model by the theoretical calculations. A 99.2% DMF yield was achieved over Co-N_x/B catalyst at 140 °C under 1 MPa H₂, which also showed an excellent cycling stability. In addition to electronic modulation, metal–metal synergy is another effective strategy for enhancing the performance of Co-based catalysts. For example, supporting Co on Sn-Beta zeolite catalyst (Co/Sn-Beta) was constructed for DMF production [39]. It was found that the HMF conversion efficiency was primarily governed by metallic Co, while the DMF selectivity was significantly enhanced by the framework Sn due to promoting the adsorption and conversion of the key intermediate BHMF. Under optimized conditions (170 °C, 14 h), an exceeding 99% DMF yield was achieved over Co/Sn-Beta. An 89% DMF yield was achieved over a core–shell catalyst (Co@CoO) [44]. There is a fundamental difference between this catalytic system and conventional supported Co catalysts. It stemmed from the fact that the active sites were not located within the metallic Co core, but rather on the oxygen-vacancy-rich CoO shell. The oxygen-vacancy-rich CoO shell had two core catalytic functions. First, it drove both homolytic and heterolytic cleavage of H₂ to generate highly reactive H^δ− species. Second, it promoted the adsorption and activation of HMF via its inherent Lewis acid sites. This integrated bifunctional mechanism bypasses the rate-limiting step of C-OH bond cleavage, which usually constrains the catalytic activity of monometallic Co catalysts. Furthermore, the Co@CoO catalyst exhibited an exceptional operational stability, which maintained its catalytic performance for more than 100 h in a continuous-flow reactor, at a weight hourly space velocity (WHSV) of 26.6 h⁻¹. The corresponding turnover frequency (TOF) of the catalyst was reached 4.26 h⁻¹ under these operating conditions.

Recently, a Co/Mg-Beta catalyst has been developed. This catalyst was synthesized by incorporating Mg directly into the framework of Beta zeolite via a structural reconstruction method [56]. When evaluated for the hydrodeoxygenation of HMF to DMF, Co/Mg-Beta exhibited markedly superior performance. It outperformed both Co/Beta and a co-impregnated CoMg/Beta control catalyst in terms of HMF conversion and DMF selectivity. It also achieved a remarkably high product yield under mild reaction conditions. A comprehensive mechanistic investigation was conducted to explain this enhanced performance. The investigation combined multiple complementary techniques: H₂-TPR, density functional theory (DFT) calculations, kinetic isotope effect studies, and selective poisoning experiments. These studies established that the performance improvement originates from a hydrogen spillover effect mediated by framework Mg species. Specifically, Mg incorporation significantly increases the concentration of active hydrogen on the catalyst surface. It achieves this by facilitating hydrogen spillover from metallic Co nanoparticles to the zeolite support. This spillover hydrogen reservoir then efficiently participates in the hydrogenolysis steps of the reaction. This mechanism provides a new paradigm for the design of non-noble metal HDO catalysts. The study further demonstrated that the Co/Mg-Beta catalyst exhibits reasonable recyclability. The authors also used Aspen Plus simulations to evaluate its industrial application potential. In addition, a single-metal iron-based catalyst has been developed for the first time for this reaction. An Fe/TiO₂ catalyst was constructed via interfacial engineering [55]. In this material, Fe species were anchored onto the surface of a TiO₂ support. Systematic experiments were performed to eliminate internal and external diffusion limitations. This enabled the extraction of reliable intrinsic kinetic data for the catalytic system. The Fe/TiO₂ catalyst exhibited superior activity compared to pure TiO₂ or kaolinite-supported Fe. This enhanced performance was ascribed to a strong synergistic effect between the Fe species and the TiO₂ surface. Kinetic analysis was conducted under mild conditions (50 °C, 0.6 MPa H₂). The results revealed that the Fe/TiO₂ system drives the hydrogenation process more efficiently. It also has an optimized temperature window that avoids the over-hydrogenation typically observed at elevated temperatures. Subsequent investigations extended this interfacial engineering concept. The TiO₂ support was doped with phosphorus to generate Brønsted acidic Fe²⁺-Oᵥ-Ti⁴⁺ structures. These structures are induced by spillover hydrogen species during the reaction. These newly formed acid sites were shown to selectively activate C-O bond. This selective activation shifts the product selectivity toward DMF. Collectively, this work highlights a key design principle for HMF upgrading catalysts. The interplay between metallic Fe sites and support-derived acid functionalities can be precisely tuned to control reaction pathways.

Ni is an ideal transition metal with hydrogenation activity higher than that of Fe and Co but lower than that of noble metals such as Pt and Pd. The over-hydrogenation was avoided due to the moderate activity of Ni-based catalysts. The selectivity could also be controlled by adjusting reaction conditions, making Ni used as a common hydrogenation catalyst. It has been demonstrated that high catalytic activity was exhibited in biomass hydrogenation, hydrogenolysis, and HDO reactions [72,73]. Adding an appropriate amount of Lewis acid sites on the catalyst surface could be enhanced DMF selectivity, which is usually used as a widely adopted strategy. For example, a Ni/ZSM-5 catalyst was prepared via a solid-state grinding method with leveraging of the synergy between metal sites and Lewis acid sites [47]. In total, 91.2% HMF conversion and 96.2% DMF selectivity were achieved over Ni/ZSM-5 under a relatively low H₂ pressure (0.25 MPa). Furthermore, a Ni/TMC catalyst was prepared by self-assembly of Ni with transition metal carbides (TMC) exhibiting high Lewis acidity [48]. It was found that the abundant Lewis acid sites were generated due to the synergy between Ni and TMC, facilitating C=O hydrogenation and C-O bond cleavage.

Beyond the introduction of Lewis acid sites, the construction of strong metal–support interactions has also been explored as a means to enhance catalytic performance. For instance, a catalyst leveraging the synergy between highly dispersed Ni hydrogenation sites and the weak acidity was provided by SCM-14 zeolite, enabling efficient HMF hydrogenolysis (Figure 4), and >99% conversion and 90.3% DMF yield were obtained [49]. Similarly, an ultra-low Ni loading catalyst consisting of Ni-C₃N₄ supported on hydrogenated carbon (HC) was designed and prepared, and a high 94.2% DMF yield was achieved due to Ni-N bonds playing a key role in the activation and cleavage of C-O bonds [46]. Although the reaction time is relatively long (4 h), its cycle stability remains very high (92% yield after four runs). In addition, the strategy of promoting hydrogen spillover via the creation of oxygen vacancies has been applied to Ni-based catalysts. For example, a highly dispersed Ni/ZrO₂ catalyst was found to contain Zr³⁺ species and oxygen vacancies formed during preparation, generating abundant active sites capable of adsorbing and activating C=O/C-O groups. This catalyst ultimately achieved 98.4% HMF conversion and 99.1% DMF selectivity [51]. The collective data on monometallic Ni catalysts reveal a consistent trend: high DMF yields approaching or exceeding 90% are achievable, but only at temperatures generally above 150 °C and hydrogen pressures in the range of 1.5–5.0 MPa. A key observation from previous studies is that large Ni particles tend to promote the hydrogenation of the furan ring, leading to the formation of the undesirable DMTHF [69]. This relationship between structure and selectivity underscores the need for precise synthetic control over metal dispersion when designing Ni catalysts for DMF production.

The more detailed discussion about the role of oxygen vacancies, metal–support interactions, and hydrogen spillover mechanisms has been added to further clarify the reaction mechanism. Combined X-ray photoelectron spectroscopy (XPS) and density functional theory (DFT) analyses have yielded critical insights into the function of oxygen vacancies. These analyses confirmed that oxygen vacancies on the CoO shell of Co@CoO had two key functions. First, they promoted the adsorption and polarization of the hydroxyl group in BHMF. Second, they participated directly in the heterolytic cleavage of H₂ to generate highly active H^δ− species [44]. Similarly, strong metal–support interactions are frequently reported to enhance both catalytic activity and operational stability. However, these interactions must be precisely tuned to avoid performance limitations. Excessively strong interactions can lead to the encapsulation of metal particles by reduced oxide overlayers, which block catalytically active sites. Conversely, insufficient metal–support interactions failed to prevent nanoparticle sintering during prolonged reaction [74]. For Cu-based and Mo-based catalysts, enhancing activity is often observed when the hydrogenation metal and oxyphilic component were mixed [64,75]. This is attributed to hydrogen spillover. In fact, these three mechanisms are deeply interconnected. Oxygen vacancies on the support can serve as anchoring sites that strengthen metal–support interactions. These tuned metal–support interactions, in turn, modulate the directionality and efficiency of hydrogen spillover during the reaction. Hydrogen spillover is another core mechanism phenomenon widely discussed in this field, which describes a process wherein H atoms dissociated on metal nanoparticles migrate onto the support surface. These migrated H atoms can then participate in remote hydrogenation or hydrogenolysis events.

Copper-based catalysts, as oxophilic metals, have garnered attention due to their unique properties and low cost. The intrinsically low activity of Cu results in a slow conversion rate of HMF to DMF, and single-metal Cu systems struggle to achieve complete HDO. However, the d-orbital electron configuration of Cu induces strong repulsion toward carbon atoms, including the carbonyl carbon, which endows Cu-based catalysts with excellent selectivity toward DMF. The C=O bond hydrogenation is promoted and the side reactions (such as ring hydrogenation and decarbonylation) were minimized by this selectivity [28]. Consequently, the outstanding performance in the selective transfer hydrogenation of HMF was exhibited over Cu-based catalysts. For example, without an external hydrogen source, 91.9% HMF conversion and 65.9% DMF yield were detected by a highly dispersed Cu/PBSAC catalyst consisting of small Cu⁰ nanoparticles [22]. Cu⁰ nanoparticles not only catalyze the dehydrogenation of isopropanol but also activate HMF for subsequent hydrogen-free HDO. In another study, a Cu-CPM catalyst was prepared using the inexpensive nitrogen-containing ligand 2-methylimidazole. After calcination at 700 °C, 100% HMF conversion and 95.9% DMF selectivity were achieved in ethanol media [45]. Despite the relatively high reaction temperature (200 °C), it still exhibits excellent cycle stability (90% yield after five runs). Notably, supporting Cu on high-surface-area graphite (HSAG) to prepare Cu/HSAG catalyst, 92% DMF yield was obtained under optimal conditions (180 °C, 4 h) using butanol as both solvent and hydrogen donor [52]. Cu-based catalysts, while generally less active than their Ni and Co counterparts for complete HMF hydrogenolysis to DMF, offer distinct advantages in terms of selectivity control when the reaction pathway is carefully managed. However, compared to noble metal benchmarks, the required hydrogen pressure and reaction time remain quite high, and the DMF yield is highly sensitive to the degree of metal reduction and the surface concentration of Cu⁺/Cu⁰ sites.

In recent years, transition metal carbides and nitrides, particularly molybdenum carbide and nitride, have attracted considerable attention. Doping with C and N atoms modifies their d-band structure, imparting noble metal-like electronic characteristics and positioning them as potential alternatives to Pt-based catalysts for hydrogenation and hydrogenolysis reactions [76]. For instance, 91% selectivity for the hydrogenolysis of HMF to DMF was achieved over a nitrogen-doped Mo single-metal catalyst (MoNC) [53]. β-Mo₂C catalyst was reported to exhibit activity in furfural (FF) hydrogenation, and molybdenum carbide supported on activated carbon or carbon nanotubes served as an excellent catalyst for ethanol decomposition to produce H₂ [77,78,79,80]. Accordingly, a β-Mo₂C catalyst embedded in carbon microspheres under a nitrogen atmosphere was prepared for the synthesis of DMF. Under optimized conditions (180 °C, 10 h), a high DMF yield of 91.6% was achieved in the selective hydrogenolysis of HMF with ethanol as the hydrogen donor [54]. Despite the relatively long reaction time (10 h), it surprisingly exhibits a high TOF number (102.83 h⁻¹). This indicates that excellent catalytic performance can be achieved even under conditions of very low Mo content. Developing advanced molybdenum-based catalyst systems and exploring synergistic effects with other metals to further enhance low-temperature activity and DMF selectivity should be the focus of future research. Remarkable DMF yield (91%) was achieved using a “metal-free catalyst” (N-C), which exhibited a high TOF value (12 h⁻¹) and high DMF productivity (10.59 mol_DMF h⁻¹ mol_{metal_total}) [50]. Intriguingly, investigations into N-doped carbon materials have revealed that trace Fe impurities, rather than the carbon matrix itself, can serve as the active sites for hydrodeoxygenation, an observation that calls for rigorous control experiments and comprehensive elemental analysis before claims of “metal-free” catalysis are substantiated. These findings collectively expand the design space beyond conventional metal-on-support formulations and open new avenues for the rational development of earth-abundant and cost-effective catalytic materials.

The reaction pathway preference over monometallic catalysts is governed by two intrinsic properties of the metal: its oxophilicity and electronic structure. Two primary reaction pathways have been identified for HMF-to-DMF conversion: Path A (via 5-methylfuran, 5-MF) and Path B (via 2,5-bis(hydroxymethyl)furan, BHMF). Noble metals such as Pd and Ru predominantly follow Path B [29,32,33]. These metals exhibit strong hydrogenation activity but weak oxophilicity. In this pathway, hydride species preferentially attack the carbonyl carbon of HMF to yield BHMF. Subsequent C-OH bond cleavage of BHMF requires either elevated temperatures (≥150 °C) or cooperation with acid sites on the catalyst support. In contrast, Cu-based catalysts favor Path A through a Cu⁰/Cu⁺ synergistic mechanism. Cu⁺ Lewis acid sites located at the metal–support interface activated the hydroxyl group of HMF. This activation enables direct C-OH scission to form 5-MF [81,82]. The 5-MF intermediate is then rapidly hydrogenated to DMF on adjacent Cu⁰ sites. This sequential mechanism bypasses the thermodynamic bottleneck associated with Path B. Ni and Co catalysts occupy an intermediate reactivity regime between noble metals and Cu-based systems. Their stronger H₂ activation capability can drive the reaction toward undesired ring-hydrogenated byproducts. However, when Ni or Co is well-dispersed on acidic or reducible supports, both Path A and Path B may operate concurrently. The dominant reaction route under these conditions is determined by two key factors: the support’s acid–base character and the size of the metal particles.

A critical yet often underappreciated metric exists for evaluating the practical viability of HMF-to-DMF catalytic systems: catalyst loading. This metric is defined here as the molar ratio of HMF to total metal (n_HMF/n_M). A high n_HMF/n_M ratio corresponds to a low metal inventory per unit of substrate processed. This low metal loading directly translates to two key benefits: reduced catalyst costs and, in many cases, attenuated metal leaching and sintering during prolonged operation. Examination of the data compiled in Table 1 reveals a striking divergence in this parameter across reported catalytic systems. This variability has profound implications for the industrial translation of these technologies. On the one hand, several noble metal catalysts achieve outstanding DMF yields at remarkably low metal loadings. For instance, Pd/0.15P-TiO₂-500 (Entry 2) delivered a 67% DMF yield at an n_HMF/n_M ratio of 425.68. Similarly, Pd/NMC (Entry 4) provided a 97% DMF yield at an n_HMF/n_M ratio of 336.06. These high-performance, low-loading systems demonstrate a key principle. When the active metal is atomically dispersed or optimally interfaced with a functional support, the intrinsic activity per metal atom can be sufficiently high to render noble metal use economically justifiable. Such ultra-low-loading noble metal catalysts represent a compelling direction for further development. They combine two critical advantages: the high activity and mild operating conditions characteristic of noble metals, and a material cost profile approaching that of base metal systems.

On the other hand, a number of noble metal catalysts with otherwise respectable performance operate at substantially higher metal inventories. For example, Br-Pd/Al₂O₃ (Entry 1) and Pd/MOF-808 (Entry 5) exhibited only 33.75 and 27.67 n_HMF/n_M ratios, respectively. At such metal-intensive loadings, the capital cost associated with the noble metal component becomes prohibitive for large-scale deployment, regardless of the system’s catalytic efficiency. A parallel concern emerges when evaluating non-noble metal systems. Several non-noble metal catalysts require extremely high metal loadings to achieve competitive DMF yields. Ni/HSAG (Entry 27) and Cu/HSAG (Entry 28) operated at n_HMF/n_M ratios of merely 5.87 and 5.58, respectively. Meanwhile, CA-Ni/ZrO₂ (Entry 26) required an n_HMF/n_M ratio of 14.67 to attain its 97% DMF yield. Although the base metals themselves are inexpensive, such excessive loadings carry critical drawbacks. They inevitably lead to low atom economy, increased susceptibility to metal agglomeration and leaching, and greater challenges in catalyst separation and recycling. All of these limitations erode the economic advantages that initially motivate the shift away from noble metals.

Conversely, non-noble metal catalysts with moderate to high n_HMF/n_M ratio have demonstrated exceptional performance. Key examples include Mo@C-900 (Entry 30, ratio of 247.03) and Ni-SCM-14 (Entry 24, ratio of 226.15). These systems demonstrate that base metals can deliver high DMF yields without resorting to excessive metal inventories, provided that the active phase is judiciously dispersed and stabilized.

Collectively, these observations define a clear strategic imperative for future research in this field. First, the field should prioritize the development of ultra-low-loading noble metal catalysts, ideally approaching the single-atom limit, that maximize the return on noble metal investment. Second, parallel advancement is required for non-noble metal catalysts with moderate, optimized metal contents. These base metal systems must preserve high atom efficiency while maintaining the activity and durability required for continuous industrial operation.

In contrast to noble metals, non-noble metal catalysts represented by Co, Ni, Cu, and Mo are abundant and low-cost, making them highly promising alternatives. Studies have shown that strategies such as support engineering (e.g., using zeolites [39,43], nitrogen-doped carbon [36,38], or constructing oxygen vacancies [44]), introducing Lewis acid sites [37,47,48], or forming specific metal–nitrogen coordination structures (e.g., Co-N_x [43]) can significantly enhance the selectivity of catalysts for C=O hydrogenation and C-O bond hydrogenolysis, enabling excellent DMF yields even under relatively harsh conditions [39,43]. Nevertheless, non-noble metal catalysts often face challenges related to their intrinsically lower activity, the need for more severe reaction conditions (higher temperatures and pressures), and stability issues such as sintering or leaching of active components during long-term operation.

In summary, it is necessary to clarify the decisive influence of the intrinsic activity of the metal, the microenvironment of the active center (acid–base property, coordination structure), and the metal–support interaction on the selective hydrogenolysis and deoxygenation performance of HMF in single-metal catalyst systems. However, whether considering the cost and selectivity control challenges of noble metals or the activity and stability limitations of non-noble metals, the inherent constraints of single-metal active centers in simultaneously meeting the multiple objectives of high activity, high selectivity, high stability, and economic viability are evident. This situation naturally leads to an advanced strategy in catalyst design: the construction of bimetallic or alloy catalyst systems through the introduction of a second metal component. This approach aims to leverage electronic and geometric effects between the metals to synergistically regulate reactant adsorption, H₂ activation, and the energy barriers for key C-O bond cleavage. It holds promise for integrating the advantages of single-metal components under milder conditions while overcoming their respective drawbacks, thereby opening new avenues for the efficient and economical conversion of HMF to DMF.

The development of monometallic catalysts for HMF-to-DMF conversion follows a clear, deliberate trajectory. This trajectory is defined by a distinct shift away from noble metals (Pd, Pt, Ru) and toward earth-abundant alternative metals (Cu, Ni, Co). Noble metal catalysts retain key performance advantages for this reaction. They continue to deliver the highest turnover frequencies (TOFs) and can operate under the mildest reaction conditions. Despite these performance benefits, their practical industrial deployment is increasingly regarded as economically unsustainable. The central challenge for non-noble monometallic catalysts remains an intrinsic trade-off between catalytic activity and product selectivity. Elevated temperatures are required to activate the active sites of Cu or Ni catalysts. These high temperatures frequently promote two major undesired side reactions: over-hydrogenation of the furan ring, and C-O bond scission at non-target positions. Future research efforts should focus on the rational modulation of key properties of the metal catalyst. These core properties include particle size, morphology, and electronic state. This modulation will be implemented through advanced synthetic strategies, such as the controlled engineering of strong metal–support interactions. The core objective of these efforts is to narrow the performance gap with noble metal benchmarks. Critically, this goal must be achieved without sacrificing product selectivity.

2.2. Bimetallic Metal Active Site Catalysts

Designing bimetallic catalysts effectively addresses the problems of low activity and poor selectivity often encountered with monometallic systems, while also significantly enhancing HMF conversion and DMF yield. The superior activity of bimetallic catalysts stems from strong metal–metal interactions and the high dispersion of nanoparticles, enabling efficient conversion under milder conditions. Moreover, bimetallic systems offer additional tunability, such as optimizing performance through the geometric configuration of reactant molecules on the surface of alloy nanoparticles, thereby improving catalyst stability [83,84,85]. Importantly, the synergistic effects and alloying of bimetallic catalysts can reduce the usage of noble metals, enhancing process economics. Recent advances in the application of bimetallic active site catalysts for the selective hydrogenation of HMF to DMF are summarized in

Table 2. Bimetallic active site catalysts for the conversion of HMF to DMF.

Entry	Catalyst	Solvent	P. /MPa	Temp. /°C	Time /h	Conv. /%	S. /%	Y. %	Productivity a/mmolDMF h−1 g−1	Catalyst Stability	TOF/h−1	Ref.
1	Ru-Ir/C	THF	1	120	1	100	99	99	22.00	NA	291.46	[86]
2	ZrO2-Co/Al2O3	THF	2	150	6	100	97	97	6.41	~45 (5 runs)	NA	[87]
3	Cu10Zr0.39Ox	Isopropanol	-	180	3	100	98	98	5.44	~85 (5 runs)	NA	[81]
4	Cu2ZnAl	Isopropanol	-	180	4	99	92	92	9.19	NA	NA	[88]
5	CZ-ag-70	Isopropanol	-	190	10	100	82	82	1.30	~80 (5 runs)	1.12	[89]
6	Ni1.52Fe0.36BOx	Ethanol	1	160	1	100	99	99	16.50	~15 b (5 runs)	1.30	[90]
7	Ir-MoS2	THF	2.5	160	6	>99	97	97	4.27	~87 (5 runs)	22.20	[91]
8	PtCo-1.2	Butanol	1	50	2	>99	100	100	8.70	96 (5 runs)	9733	[74]
9	PtCo@C	Butanol	1	120	2	>98	100	100	191.94	NA	296.60	[74]
10	CuPd@DAB-1	THF	1.5	130	10	92	82	75	2.50	NA	94.3	[92]
11	CoMo1@NC	n-Propanol	2	170	8	95	77	73	0.88	~70 (5 runs)	28.7	[93]
12	Ni1.5Co1	THF	0.2	100	4	100	80	80	5.56	NA	0.86	[94]
13	Ru-Ni/TiO2	1,4-Dioxane	3	160	4	100	71	71	9.38	77 (3 runs)	8.97	[95]
14	Co1Pt0.013Al	2-Propanol	2	160	2	100	87	87	172.47	~65 (4 runs)	1.74	[96]
15	Cu1Co4	2-Propanol	-	170	12	100	99	99	1.96	91 (5 runs)	NA	[97]
16	Co2Ni1@NC	THF	1	150	3	100	99	99	22.00	~35 (5 runs)	7.89	[98]
17	10Cu3Pd/BCN	THF	1.5	180	4	100	99	99	2.47	~85 (5 runs)	1.37	[99]
18	PdCo/MoCx	THF	2	180	8	99	97	97	1.21	~30 (5 runs)	2.09	[75]
19	Co-FeOx/NC	THF	2	180	6	100	99	99	13.08	~50 (5 runs)	NA	[100]
20	NiFe(C)-500	1,4-Dioxane	3	220	1	96	74	71	37.53	NA	22.48	[101]
21	Co2Ni1@NC	Ethanol	2	220	4	100	93	93	9.22	~90 (5 runs)	3.7	[102]
22	5Pt5Fe/C	1,4-Dioxane	3	180	4	100	99	99	4.91	~75 (5 runs)	5.06	[103]
23	1.5Ni-1.5Fe-1.0Al	Isopropanol	4	200	6	100	93	93	2.33	NA	54.9	[104]
24	Ru-Co/AC	THF	1	200	1.5	99	99	98	46.00	~80 (4 runs)	89.86	[105]
25	MA-NiCo@ZrO2/NBC	THF + FA	-	200	2	99	97	96	4.80	~95 (5 runs)	NA	[106]
26	Pt@OM-SnO2	THF	1	120	4.5	100	99	99	4.40	NA	97.5	[20]
27	20Co-CoOx-10FeNiCo/γ-Al2O3-500	THF	2	190	4	100	100	100	4.95	~99 (5 runs)	0.02	[107]
28	CAA-2	THF	1.5	180	6	100	93	93	2.07	~50 (5 runs)	1.24	[108]
29	Ni-ZnO/AC	1-Propanol	1	160	3	100	99	99	5.23	~90 (5 runs)	1.54	[109]
30	Ni-280/Fe-N-C-800	THF	4	240	12	100	96	96	<0.01	~75 (3runs)	0.01	[110]

^a $Productivity={\displaystyle \frac{n_{DMF}}{{mass}_{catalyst}\times t}}$; ^b The reaction time is 130 °C.

.

Weak interactions between the support and metal nanoparticles can lead to metal leaching, a primary cause of the poor stability exhibited by monometallic noble metal catalysts. Incorporating a second metal represents an effective strategy to overcome this limitation [58]. The introduction of an appropriate second metal into noble metals (e.g., Ru, Pt, and Pd) is crucial, as bimetallic catalysts allow catalytic performance to be tuned through geometric and electronic effects [28,111,112]. Ru is among the most effective metals for the hydrogenation and hydrogenolysis of biomass-derived oxygenates, largely due to its ability to substantially lower the energy barrier of reaction pathways in protic solvents [113]. Its activity can be further enhanced and its properties tuned to avoid side reactions such as over-hydrogenation through modification with other metals [95]. For instance, Ir-based catalysts have been shown to modulate the product distribution in the hydrogenation of FF and HMF [114]. Based on this, a carbon-supported Ru-Ir alloy nanoparticle catalyst (Ru-Ir/C) was developed [86]. Leveraging the synergistic effects between Ru and Ir, the substrate reduction was accelerated over Ru-Ir/C, and a DMF yield of 99% was achieved within 1 h. The TOF number can reach as high as 291.46 h⁻¹. Similarly, it has been demonstrated that the addition of Ni can modify Ru catalysts, and the resulting Ru-Ni/TiO₂ bimetallic catalyst exhibited enhanced surface acidity and improved performance due to changes in the electronic properties, particle size, and support interactions of Ru [95]. Chemical reduction and hydrogen reduction are common methods for preparing hydrogenation catalysts [115]. A two-step reduction method (NaBH₄ followed by H₂) was employed for the sequential reduction in Ru and Co to prepare a Ru-Co/AC bimetallic catalyst [105]. The introduction of Co not only promoted the dispersion of Ru nanoparticles on activated carbon (AC) but also facilitated electron transfer from Co to Ru, thereby increasing catalytic activity. Under optimal condition (200 °C, 1 MPa), a DMF yield of 97.9% and an HMF conversion of 98.7% were achieved over Ru-Co/AC within 1.5 h. This catalyst is highly versatile, featuring a high TOF number (89.86 h⁻¹), excellent cycling performance, and short reaction times (1 h). It is considered an excellent bimetallic site catalyst.

Pt-Co catalysts, known for their excellent performance in FF hydrogenation, have also been applied to the hydrogenation of HMF. For instance, a catalyst prepared by adjusting nanoscale Pt on a Co/Al₂O₃ support was investigated [96]. It was found that the coexistence of metallic Pt⁰/Co⁰ and oxidized PtO₂/CoO_x species acted as the active component, and the catalyst acidity was influenced by the Co/Pt ratio. Under complete HMF conversion, a DMF selectivity of 85% was obtained over the Co₁Pt_0.013Al catalyst with a low Pt loading. Additionally, an ultrasmall (1.2 nm) PtCo-1.2 bimetallic alloy nanoparticle catalyst (Figure 5) was synthesized [74]. It stood as a landmark example: it achieved a TOF of 9733 h⁻¹ and productivity of 191.94 mmol_DMFh⁻¹g⁻¹ at mild 50 °C, a value that places it among the most intrinsically active catalysts ever reported for HMF hydrogenolysis. Notably, it possessed excellent cycle stability. The size of bimetallic nanoparticles played a decisive role in catalytic performance for the hydrogenolysis of HMF to DMF. Miniaturizing bimetallic alloy nanoparticles to sizes below the 3 nm threshold has been shown to yield distinct catalytic properties. These properties differed significantly from those of both single-atom catalysts and larger nanoparticles [74]. However, conventional synthesis methods faced critical limitations in producing well-defined ultrasmall alloy nanoparticles. Common methods include impregnation and nanocluster chemistry. These approaches often yield nanoparticles with broad size and composition distributions. This heterogeneity made it challenging to establish reliable structure–activity relationships. To address this critical limitation, A synthesis strategy based on thermodynamic control has been developed. This strategy utilized nanoscale compartments formed within microemulsions. It enabled the precise synthesis of sub-3 nm bimetallic alloy nanoparticles. Importantly, both the size and composition of these nanoparticles could be independently tuned [74]. This novel synthetic approach enables systematic investigation of the catalytic properties of ultrasmall bimetallic nanoparticles. It provides a robust experimental platform for elucidating intrinsic size effects in HMF hydrogenolysis.

Remarkably, a near 100% DMF selectivity was achieved in butanol solvent under mild condition (120 °C, 1 MPa H₂, 2 h). Notably, a catalyst featuring Pt₁Sn₁ intermetallic species coupled with ordered mesoporous SnO₂ (OM-SnO₂) was constructed [20]. Impressively, a DMF yield of up to 99% was achieved over the Pt@OM-SnO₂ catalyst, which can be attributed to the Pt₁Sn₁ nanoparticles and the defect-rich OM-SnO₂ serving as highly active species. It has strong catalytic performance and a high TOF number (97.5 h⁻¹). The combination of noble and non-noble metals represents a pragmatic strategy to dilute noble metal content while simultaneously harnessing synergistic electronic and geometric effects. The combination of noble and non-noble metals represents a pragmatic strategy to dilute noble metal content while simultaneously harnessing synergistic electronic and geometric effects.

Pd-based bimetallic catalysts have also provided promising performance. For example, supporting atomically dispersed Cu-Pd bimetallic alloy nanospheres on a nitrogen-rich porous organic polymer (POP) was prepared (Figure 6) [92]. Due to its high TOF number (94.3 h⁻¹), it has certain prospects for industrial applications. The stable Cu^0/2+ and Pd^0/2+ active surface species were maintained for bimetallic nanospheres, achieving a DMF selectivity of 81.5% under optimized condition (130 °C, 1.5 MPa H₂,10 h) in THF solvent. Similarly, a Cu-Pd bimetallic catalyst supported on boron carbonitride (CuPd/BCN) was developed [99]. The strong interactions between the alloy and the BCN support catalyst were exhibited, enabling complete HMF conversion and a DMF selectivity of 99% with an excellent cycle stability (85% yield after five runs). Furthermore, the strong synergistic effects between Pd and Co were also shown. For example, a highly dispersed Pd-Co bimetallic catalyst supported on molybdenum carbide (PdCo/MoC_x) was designed [75]. It was confirmed that the HDO reaction was facilitated due to abundant oxygen vacancies and the presence of low-valence Pd and Co species. Meanwhile, the synergy between the bimetallic sites alters the reaction pathway and promotes aldehyde group hydrogenation, significantly enhancing reaction efficiency and achieving a DMF yield exceeding 97% under mild conditions. These cases illustrate a broader design principle: the noble metal component functions primarily as a potent H₂ dissociation center, while the adjacent non-noble metal and metal oxide sites facilitate the adsorption and polarization of the C-OH bond, thereby enabling efficient bifunctional catalysis.

The addition of metals or metal oxides such as Ni, Cu, Ag and Mo can enhance the activity of Co-based catalysts. For example, a series of ZrO₂-modified Co nanocatalysts was synthesized [87]. It was found that the addition of ZrO₂ stabilized CoO species, enhancing HMF adsorption and heterolytic H₂ dissociation, generating highly active H^δ− species. The excellent performance and stability were shown for 2ZrO₂-Co/Al₂O₃ catalyst, achieving a DMF yield of 97.3%. In addition, Lu et al. [93] revealed a synergistic catalytic effect between Co and Mo sites, wherein Mo sites primarily governed substrate adsorption while Co sites dominated H₂ adsorption and activation.

The alloying of Ni and Co not only prevents the formation of large Ni particles but also promotes selective cleavage of C=O bonds to enhance hydrogenation, making NiCo alloys ideal metal sites [116,117,118,119]. Recently, a cost-effective and magnetically separable bimetallic Ni_xCo_y catalyst was developed, wherein the hydrogenation sites of Ni and Co cooperate with Lewis acid sites in NiO and CoO to promote the conversion of HMF to DMF [94]. Furthermore, the HDO activity was enhanced, and over-hydrogenation and ring-opening of the furan ring were suppressed due to the electronic effect of Ni-Co nanoalloys [102]. Beyond preformed nanoparticles, nanoparticle catalysts formed in situ represent another important class of catalytic materials. These catalysts are generated during the catalyst preparation process under actual reaction conditions. For example, the Co₂Ni₁@N₁C-700 catalyst was developed [98], where CoNi alloy nanoparticles were confined within N-doped porous carbon. The nanoparticles were formed in situ during the carbonization process of the precursor. This in situ formation strategy offers several key advantages. First, the N-doped carbon matrix provides effective spatial confinement. This confinement limits nanoparticle sintering and leaching during the reaction. Second, in situ formation from the Co₂Ni₁@N₁C precursor ensures intimate contact between the alloy nanoparticles and the N-doped carbon support. Density functional theory (DFT) calculations provide mechanistic insights into this effect. They demonstrated that this intimate contact enhanced the adsorption of both H₂ and HMF. It simultaneously lowers the adsorption energies of H* intermediates and DMF products. This electronic modulation facilitates two critical reaction steps: C=O hydrogenation and C-OH hydrogenolysis. Both steps are essential for selective DMF formation. Under optimized reaction conditions, this catalyst achieved a 98.9% DMF yield at 150 °C. This performance validates the effectiveness of in situ-formed nanoscale alloy particles confined in N-doped carbon for HMF hydrogenolysis. In addition, A mechanical activation (MA) strategy was employed to construct an efficient and stable MA-NiCo@ZrO₂/NBC/NBC composite [106]. Using formic acid (FA) as the hydrogen source at 200 °C for 2 h, this catalyst achieved 99.3% HMF conversion and 96.7% DMF yield. Importantly, the formation of NiCo alloys and the rapid electron transfer at metal-acid bifunctional sites enhanced preferential C=O adsorption while preserving the integrity of the furan ring during C-O hydrogenolysis. Beyond CoNi alloys, the synergistic interaction between CoNi alloys and Co-N_x species also markedly improves catalytic performance. It was found that Co-Nₓ species can act as Lewis acid sites, facilitating C=O activation and subsequent C-OH cleavage, thereby enhancing DMF selectivity [98].

Cu-based catalysts are often regarded as potential substitutes for noble metals in catalytic transfer hydrogenation (CTH) reactions [120]. Zr single-atom species were successfully introduced into CuO_x, creating strong positively charged atomic Lewis acid sites [81]. These sites not only modulated and increased the content and stability of Cu⁺ species through electronic interactions but also enhanced catalyst acidity. Using isopropanol as the hydrogen donor, this single-atom Zr-doped CuO_x catalyst (Cu₁₀Zr_0.39O_x) exhibited outstanding performance, achieving a DMF yield of 97.7%. Similarly, the CZ-ag-70 catalyst, which is rich in Cu^δ+-$V_{Zr}^{+}$ interfacial sites, favored the formation of the dehydroxylated product DMF [89]. Hence, this catalyst achieved 100% HMF conversion and 82.2% DMF selectivity. In another study, the catalytic transfer hydrogenation performance of a CuCoO_x catalyst using 2-propanol (IPA) as the hydrogen donor was investigated [97]. At 170 °C, 100% HMF conversion was reached with a DMF yield of 99%. The study also identified the hydrogenolysis of BHMF to 5-MFA as the rate-determining step. Similarly, a Cu_xZnAl catalyst using isopropanol as the hydrogen donor also performed well [88]. This catalyst not only achieved good dispersion of Cu species but also attained an optimal ratio of Cu²⁺, Cu⁰, and Cu⁺. The synergistic effect between Cu⁰ and Cu⁺ enabled efficient hydrogenation/hydrogenolysis of HMF, yielding 91.7% DMF after 4 h at 180 °C. Additionally, Ag can improve the dispersion of Cu on the surface of catalysts. Lakshmi et al. [108] found that the activity of Cu-Ag catalysts depends on the silver content. The catalyst with 2 wt% Ag and 10% Cu exhibited the best performance, achieving complete HMF conversion and a DMF yield of 93% at 180 °C.

Fe is abundant, inexpensive, and environmentally benign; However, the hydrogenation/hydrogenolysis activity of monometallic Fe catalysts was very limited. Therefore, Fe is typically employed as an oxophilic metal additive in the catalytic conversion of furan derivatives. The incorporation of Fe promotes the formation of electron-rich metal sites and strong acid sites within the catalyst. For instance, an amorphous Ni-Fe bimetallic boride catalyst was constructed via a simple chemical reduction method [90]. This catalyst exhibited a DMF formation rate 1.7–16.5 times higher than that of state-of-the-art nickel-based catalysts. It can be attributed to the incorporation of Fe facilitating hydrogen activation and C-O bond cleavage, thereby accelerating the hydrogenolysis process. Fe species can also form alloys with Ni [121,122]. The presence of Fe-containing surface phases helps maintain a high degree of Ni atom dispersion. For example, a TiO₂-supported bimetallic NiFe alloy catalyst was developed [101]. Its specific surface structure promotes HMF adsorption via the carbonyl group while preventing hydrogenation of the aromatic furan ring, thus preserving high DMF selectivity. Another study reached a similar conclusion, introducing Fe into Ni-Al resulted in the formation of an oxophilic FeNi alloy [104]. This modification influences the interactions between components, the formation of Lewis acid sites, H₂ adsorption/activation, and the activation of oxygen-containing groups in reaction intermediates, thereby affecting DMF selectivity and the structural stability of the Ni-Fe-Al catalyst. Furthermore, Fe can serve as an acidic site for HMF adsorption and activation in the form of FeO_x. For example, a stable and efficient Co-FeO_x/NC (NC = nitrogen−doped carbon) catalyst was prepared [100]. Due to electron transfer from Co to FeO_x, an electron-deficient state was exhibited for the interfacial Co exhibits an electron-deficient state, enhancing hydrogen activation and spillover to the interfacial FeO_x. The combination of efficient hydrogen spillover and interfacial acidity significantly improved the selectivity for the selective hydrogenolysis of HMF to DMF, achieving a yield as high as 99.9%. Fe can also form synergistic bimetallic catalysts with noble metals, exhibiting excellent performance (e.g., a DMF yield of 99.6%) [103]. It was reported that the interaction between Pt and Fe species establishes a balanced synergy between metal and acid sites, thereby enhancing HDO activity. Recently, a NiFe alloy catalyst supported on nitrogen-doped carbon has been reported [110]. This catalyst was prepared via a two-step pyrolysis method. It achieves 96.3% DMF selectivity at complete HMF conversion. The active site structure of this catalyst was precisely elucidated using extended X-ray absorption fine structure (EXAFS) spectroscopy. EXAFS analysis revealed that the key catalytic species is an Fe-N₄-stabilized NiFe alloy nanoparticle. Complementary characterization and control experiments provided further mechanistic insights. Hydrogen temperature-programmed reduction (H₂-TPR) measurements and control experiments over model catalysts demonstrated two core functions of the Fe-N₄ coordination framework. First, it promotes the uniform dispersion of the NiFe alloy nanoparticles. Second, it electronically modulates the active metal sites to enhance deoxygenation selectivity. The catalyst was further modified with a small amount of Ru to form a ternary system. In this modified catalyst, DMF selectivity remained high (88.2% vs. 96.3% for the unmodified material). Notably, the cycling stability was significantly improved. Only a minor decrease in activity was observed after three consecutive reaction runs. This stability enhancement was attributed to well-dispersed Ru clusters. These clusters inhibit the aggregation of NiFe alloy nanoparticles during the reaction. This interpretation was supported by post-reaction transmission electron microscopy (TEM) analysis. Collectively, these findings underscore the critical role of M-Nₓ coordination in non-noble HDO catalysts. This coordination environment defines both the catalytic activity and long-term durability of these materials. Fully non-noble bimetallic systems have attracted intense research interest in this field. This interest stems from two key advantages: their inherent economic viability and the rich phase space of possible compositions and alloy structures available for optimization.

A common mechanistic motif has emerged consistently across these studies. This motif centers on the electronic modulation effect of a secondary metal (e.g., Fe, Zn, Sn) on the primary hydrogenation metal (e.g., Ni, Co). This electronic tuning adjusts the adsorption strength of the furan ring and aldehyde/hydroxyl functional groups on the catalyst surface, which directly governs reaction selectivity. Furthermore, many of these bimetallic systems offer an additional key advantage. They enable the in situ generation of interfacial acid–base pairs during the reaction. These acid–base pairs facilitate the dehydration of BHMF to MFA, a critical intermediate in the HMF-to-DMF conversion pathway. Despite these encouraging performance results, critical knowledge gaps remain for non-noble bimetallic catalysts. Most notably, their long-term operational stability remains inadequately characterized. Systematic studies are therefore urgently needed to guide the rational optimization of these promising materials. These studies must correlate three core factors with catalytic performance: alloy composition, surface segregation phenomena, and catalyst deactivation mechanisms.

Bimetallic catalysts enable rational manipulation of reaction pathway selectivity. This control is achieved through spatial and electronic interplay between two core components: a hydrogenation-active metal and an oxophilic promoter (e.g., Fe, Mo, Zn, Co, Re). However, the actual dominant reaction pathway is highly system-dependent. It cannot be generalized across all bimetallic catalyst compositions. In Ni-Fe bimetallic systems, both Path A and Path B have been observed to operate concurrently. For example, on NiFe/TiO₂ catalysts, HMF is preferentially adsorbed via the carbonyl group on metallic Ni sites. This adsorption behavior is consistent with a Path B reaction bias. In this system, Fe-containing surface phases serve two primary functions: they maintain Ni nanoparticle dispersion, and they suppress undesired furan ring hydrogenation. Notably, they do not exclusively promote direct C-OH bond scission [101]. In contrast, other studies have identified Path B as the sole primary reaction route for Ni-Fe catalysts. These Ni-Fe systems generate electron-rich metal sites and strongly acidic sites. These sites collectively facilitate two key reaction steps: hydrogen activation and C-O bond cleavage. This synergistic effect thereby promotes the efficient conversion of HMF to DMF [90]. In contrast, noble metal–cobalt bimetallic systems exhibit distinct reaction pathway preferences. This is observed for both Pd-Co and Pt-Co catalyst formulations. For the Pd-Co/MoCₓ catalyst, Path A (via 5-MF) is the predominant reaction route. Two key factors drive this pathway selectivity. First, the high concentration of oxygen vacancies in the support facilitates C-O bond cleavage. Second, the synergistic interaction between bimetallic sites redirects the reaction pathway. This interaction promotes selective hydrogenation of the aldehyde group, thereby significantly enhancing overall reaction efficiency [75]. In contrast, Path B is dominant for the PtCo-1.2 catalyst [74]. Stoichiometric alloying was achieved through high-temperature synthesis. This process produced homogeneous bimetallic alloy nanoparticles with precisely controlled size (<3 nm) and composition. These well-defined nanoparticles delivered exceptional catalytic yields. On the other hand, the Co₁Pt_0.013Al catalyst follows Path B exclusively [96]. Characterization confirmed the coexistence of two distinct phases in this material: metallic Pt⁰/Co⁰ and oxygen-containing PtO₂/CoOₓ species. This catalyst exhibits remarkable temperature-dependent functional diversity. By adjusting the reaction temperature, the selective production of different products was achieved. BHMF was obtained at 40 °C, while DMF was produced at 160 °C. Therefore, the general design principle for bimetallic catalysts should be further refined. Pairing a hydrogenation-active metal with an oxophilic promoter can accelerate the rate-limiting C-OH cleavage step. This acceleration effect holds true regardless of the dominant reaction pathway. Furthermore, the optimal metal-promoter pairing must be tailored to the specific C-OH cleavage site that constitutes the kinetic bottleneck in each target catalytic system.

Although one-pot conversion strategies bypass the energy-intensive isolation of HMF, current systems still fall short of practical efficiency benchmarks. 2D MOF-based bifunctional catalysts (Pd/NUS-SO₃H) [123] achieved a ~90% DMF yield from fructose over 6 h. This performance is enabled by the catalyst’s ultrathin nanosheet structure, which integrates Brønsted acid sites, Lewis acid sites, and Pd hydrogenation sites within a confined space. This confined spatial design directly suppresses humin formation during the reaction. However, this system has two key limitations. First, it relies on the noble metal Pd. Second, it is primarily effective for high-cost fructose feedstocks. When it was extended to glucose, the system required a prolonged 12 h reaction time and delivered only a ~73% DMF yield. This performance gap underscores the inherent difficulty of balancing the additional glucose isomerization step with hydrogenolysis reaction kinetics. A fully noble-metal-free alternative has also been reported. This system combines P-UiO-66 (a solid acid catalyst) with Ni-Co@NC (a hydrogenation/hydrogenolysis catalyst), and delivered >80% DMF yield from glucose in 8 h, with a comparable timeframe for fructose feedstocks [124]. Although this system is commendable for eliminating noble metals, its space–time yield remained markedly inferior to that of optimized two-step processes. In these established two-step processes, purified HMF undergoes hydrogenolysis at substantially higher reaction rates. A fundamental conflict persists across all reported one-pot systems. The dehydration step requires elevated temperature and the absence of H₂ to suppress humin formation. In contrast, the hydrogenolysis step requires sufficient H₂ pressure and moderate temperatures to achieve high selectivity. The compromised operating conditions required to reconcile these conflicting demands invariably extend overall reaction times or reduce product selectivity. Furthermore, the durability of these catalysts under the harsh hydrothermal and reducing environments of one-pot conversion has only been evaluated in short-term recycling tests. Collectively, these studies establish the conceptual viability of one-pot DMF synthesis. At the same time, they reveal that current state-of-the-art systems are constrained by three key limitations: long reaction time, a narrow substrate scope, and unproven long-term stability. These gaps highlight an urgent need for more robust, efficient, and earth-abundant catalytic systems for this reaction.

Bimetallic catalysts have demonstrated significant advantages in the selective hydrogenation of HMF to DMF and have emerged as an important research direction. Compared to monometallic systems, bimetallic catalysts leverage metal–metal synergy to achieve superior activity, selectivity, and stability. They enable efficient conversion under milder conditions while reducing the use of noble metals, aligning with the dual requirements of green chemistry and economic viability [83,84,85]. Their enhanced performance primarily arises from the combined effects of geometric and electronic modulation, as well as strengthened metal–support interactions. These factors collectively promote reactant adsorption, hydrogen activation, and the selective cleavage of key C-O bonds [28,111,112]. Although bimetallic catalysts exhibit high activity, good recyclability, and strong resistance to poisoning and sintering, their preparation is often complex [125,126]. Precise control over the metal ratio, dispersion, and degree of alloying is required to achieve optimal active site configurations [28]. Furthermore, the long-term operational stability, regenerability, and adaptability to practical reaction systems still warrant systematic investigation. In summary, the rational design and integration of multiple metal components in bimetallic catalysts provide a powerful platform for achieving highly efficient and selective conversion of HMF. This approach not only advances the optimization of synthetic pathways for furan-based fuels and chemicals but also aligns closely with the goals of renewable resource utilization and sustainable development. Future research should focus on the precise construction of active interfaces, elucidation of the dynamic evolution of metal sites during reaction, and the development of low-cost metal combinations to facilitate the practical industrial application of such catalysts.

Bimetallic catalysts have emerged as the most versatile platform for HMF hydrodeoxygenation (HDO). They uniquely address the conflicting demands of three core performance metrics: high catalytic activity, product selectivity, and material cost. The performance advantages of these catalysts stem from the synergistic interplay between two key components. The first is a hydrogenation-active primary metal, such as Cu, Ni, or Co. The second is an oxophilic promoter that facilitates C-O bond scission, with common examples including Ru, Pd, Fe, and Zn. This synergistic effect has proven highly effective for directing the reaction selectively toward DMF, while simultaneously suppressing undesired furan ring hydrogenation. Operando spectroscopic techniques and density functional theory (DFT) calculations will be indispensable in this research endeavor. Moreover, the integration of bimetallic particles with acidic supports represents an exciting research direction. This material design enables true one-pot conversion of saccharides to DMF. It also bridges rational catalyst design with industrial process intensification.

3. Catalyst Support

The choice of support plays a critical role in the catalytic performance for the selective hydrogenation of HMF to DMF. Various support materials optimize the catalytic process through their unique physicochemical properties, including activated carbon (AC), molecular sieves [127], zeolites [62,128], metal oxides [129,130,131], carbon nanotubes [132], and graphene-based materials [36,38]. A key function of the support lies in its ability to modulate the electronic properties of the active metal centers. For instance, when Co catalyst was encapsulated in nitrogen-doped graphene (Co@NGs), strong electronic interactions occurred at the interface between the support and metallic Co. This led to electron transfer from Co to the doped N, generating electron-deficient metal centers. Such a decrease in electron density has been shown to favor the selective hydrogenation of HMF, achieving a DMF selectivity of up to 95% [36]. A B,N-codoped graphene-encapsulated Co catalyst (Co@BNG) induced the formation of electron-deficient Co-Nₓ/B species [38]. These species acted as Lewis acid centers, facilitating the activation of C=O bonds and subsequent cleavage of C-OH bonds.

The structural characteristics of the support also govern the dispersion, accessibility, and mass transfer efficiency of the active sites. SBA-15, with its high surface area and well-ordered mesoporous structure, provided an excellent environment for the dispersion of Co species while ensuring efficient diffusion of both reactants and products [37]. This forms the basis for achieving high activity (complete HMF conversion) and high selectivity (96% DMF yield). Similarly, DMF production was also facilitated over dealuminated Beta zeolite (Beta-DA) due to its relatively open channel architecture [43].

In addition, certain supports can create synergistic catalytic sites either intrinsically or through interactions with the metal. Highly dispersed metallic Co worked in concert with adjacent Lewis acidic Co²⁺ sites to jointly catalyze the hydrogenation and hydrogenolysis steps for the Co/SBA-15 catalyst [37]. The adsorption and conversion of the key intermediate BHMF were enhanced by framework Sn species of the Co/Sn-Beta catalyst, thereby promoting the selective formation of DMF [39]. For metal oxide supports such as Co@CoO, surface oxygen vacancies (O_v) enable efficient heterolytic and homolytic cleavage of H₂, generating highly active H^δ− [45,133,134] species [44,135,136]. These species directly participate in and accelerate the hydrogenolysis reaction.

Metal–organic frameworks (MOFs) are widely employed as supports for heterogeneous catalysts due to their exceptionally high porosity, regularly arranged metal nodes, and tunable structures [133,137]. So they are frequently used as supports for various catalysts. For example, a Pd-UiO-66 catalyst with controlled ligand vacancy content was developed. By tailoring the support structure to enhance hydrogenolysis activity, a higher proportion of Pd⁰ species and Brønsted acid sites were generated. Remarkably, the DMF yield was increased by approximately threefold, achieving a selectivity of 92.2% at 160 °C and 1.5 MPa H₂ [6]. Furthermore, Pd supported on the Zr-based MOF-808 (Pd/MOF-808) was also designed, and it could enable efficient hydrogenolysis of HMF to DMF [32]. A yield as high as 99% was detected under mild conditions (100 °C) and without any additives. It revealed that the Pd²⁺ species not only enhanced catalytic performance but also bound to the oxygen atoms of reactants and stabilized active hydrogen, thereby significantly improving the hydrogenolysis capability of the catalyst. This work also revealed a strong metal–support interaction between Pd²⁺ species and pyridinic nitrogen atoms.

Zeolitic imidazolate frameworks (ZIFs), a subclass of MOFs, exhibit high stability, high specific surface area, and nitrogen-rich composition, making them ideal templates for the preparation of nitrogen-doped carbon catalysts [134]. A nitrogen-doped carbon catalyst (Co/N-C) with a hexagonal structure via solvothermal crystallization was developed, utilizing coordination between Co atoms and benzimidazole ligands [43]. Notably, Co nanoparticles are highly dispersed within the porous N-C framework, forming Co-N_x active species and achieving excellent catalytic activity. Under relatively mild conditions (150 °C, 0.25 MPa H₂), the Co_0.25/N-C-500 catalyst with a specific Co loading delivered 99.9% HMF conversion and 98.5% DMF selectivity.

Beyond conventional supports, the use of waste materials as catalyst supports offers a novel approach. This strategy not only addresses waste disposal but also provides new support options for the green synthesis of DMF. Coal fly ash (CFA), a major solid waste generated by coal-fired power plants, occupies substantial land and causes environmental pollution when stockpiled. Its comprehensive utilization offers both cost advantages and environmental benefits [138,139,140]. Recently, CFA-supported Co used as a catalyst was first reported for the HDO of HMF to DMF [40]. The pore structure and specific surface area of the CFA were increased by acid modification treatment, enhancing the activity of the Co/CFA catalyst. Under the optimal Co loading, this catalyst achieved 99.0% HMF conversion and 98.1% DMF selectivity.

Carbon materials are commonly used as supports for metal nanoparticles, but some studies have employed them directly as catalysts [141]. Recently, a simple nitrogen-doped carbon catalyst was shown to achieve HDO of HMF to DMF with a yield as high as 91%. C-O bond cleavage through heteroatom doping was promoted over this carbon-based catalyst. The role of metal impurities in “quasi-metal-free” catalysts was explored for the first time. Characterization, experimental results, and calculations indicated that trace Fe-Nₓ sites formed during pyrolysis may play a significant role in key steps of the HDO reaction, such as H₂ activation and -OH deoxygenation [50].

In summary, the support is far from an inert carrier in the conversion of HMF to DMF; rather, it constitutes a critical component of the efficient catalytic system alongside the active metal centers through mechanisms such as electronic modulation, structural optimization, and the provision of synergistic active sites. Rational design of support properties represents an effective strategy for steering reaction pathways and enhancing catalyst activity and selectivity.

Future design strategies should have two clear objectives. First, it should enhance the diffusion of polar reactive intermediates. Second, it should confine metal nanoparticles within well-defined microenvironments. The key challenge in this field remains the hydrothermal stability of these engineered supports. This stability must be maintained under harsh operating conditions, including aqueous-phase environments, high temperature, and the reaction conditions relevant to integrated biomass conversion processes.

4. Hydrogen Donor

In the catalytic hydrogenation of HMF to DMF, the choice of hydrogen donor is a critical factor determining the reaction pathway, catalytic efficiency, and process sustainability. Different hydrogen sources not only influence the mode of hydrogen supply and reaction thermodynamics but also profoundly modulate the rates and selectivities of elementary steps such as hydrogenation, hydrogenolysis, and dehydration through their microscopic interactions with the catalyst active centers, thereby governing the final product distribution and yield. Recent studies on the effects of hydrogen donors are summarized in

Table 3. Effects of hydrogen donor on the conversion of HMF to DMF.

Entry	Catalyst	Solvent	P. /MPa	Temp. /°C	Time /h	Conv. /%	S. /%	Y. %	Ref.
1	Ru-N-CMK-1	Isopropanol	-	160	8	100	88	88	[33]
2	Pt/Co2AlO4	Isopropanol	2	180	3	>99	99	99	[35]
3	Cu/PBSAC	Isopropanol	1	190	6	100	>98	>98	[22]
4	Cu10Zr0.39Ox	Isopropanol	-	180	3	100	98	98	[81]
5	Cu2ZnAl	Isopropanol	-	180	4	99	92	92	[88]
6	CZ-ag-70	Isopropanol	-	190	10	100	82	82	[89]
7	1.5Ni-1.5Fe-1.0Al	Isopropanol	4	200	6	100	93	93	[104]
8	Ru/ZSM-5	Ethanol	1.7	180	3	98	97	95	[34]
9	Co@NGs-700	Ethanol	2	200	6	100	95	95	[36]
10	Cu-CPM	Ethanol	1.5	200	3	100	96	96	[45]
11	Mo@C-900	Ethanol	-	180	10	100	92	92	[54]
12	Ni1.52Fe0.36BOx	Ethanol	1	160	1	100	99	99	[90]
13	Co2Ni1@NC	Ethanol	2	220	4	100	93	93	[102]
14	MA-NiCo@ZrO2/NBC	THF + FA	-	200	2	99	97	97	[106]
15	Pd/NMC	FA	0.5	160	3	99	97	97	[31]
16	PdCl2	PMHS a + ethanol	-	25	0.5	100	90	90	[142]
17	Pd/MIL-53(Al)-P	PMHS a	-	25	2.5	100	99	99	[143]

^a PMHS is Polymethylhydrosiloxane.

.

To overcome the limitations associated with the direct use of H₂, CTH using liquid organic compounds as hydrogen donors has emerged as a research frontier. Alcohols are currently regarded as the most promising hydrogen donors and are also frequently employed as reaction solvents. Compared to H₂, the use of alcohols as hydrogen donors offers advantages such as milder reaction conditions, enhanced safety, and operational simplicity. Lower alcohols (e.g., isopropanol and ethanol) and FA represent the two most representative classes of hydrogen donors. These low-carbon alcohols can dictate the catalytic transfer hydrogenation pathway through their hydrogen-donating ability and can act as surface ligands on metals via strong dissociative adsorption, thereby modulating the transient states of reaction intermediates through steric effects [144].

Isopropanol (IPA) is commonly employed as a reaction medium for the hydrogenation of furan-based compounds [145,146], owing to its favorable hydrogen-donating ability, low toxicity, and ease of separation (Figure 7). For example, the influence of various hydrogen donors was investigated, and it was found that the DMF yields obtained with isopropanol and 2-butanol were significantly higher than those with methanol, ethanol, n-propanol, and n-butanol [33]. Secondary alcohol exhibited higher activity than primary alcohols for the selective synthesis of DMF from HMF via catalytic transfer hydrogenation, which may be attributed to their lower reduction potential. Furthermore, the strong proton-donating ability of isopropanol plays a crucial role in promoting the HDO of HMF. Huang et al. discovered a hydrogen-shuttling mechanism for isopropanol, in which protons assist in the removal of hydroxyl groups, thereby lowering the activation barrier and achieving exceptionally high DMF selectivity (99.5%) [109]. Interestingly, a positive linear correlation between solvent polarity (ET(30)) and DMF selectivity was also observed [147], which was attributed to the enhanced solubility of substrates in more polar solvents, leading to more complete reactions.

In addition, ethanol has attracted considerable attention as a green solvent due to its high hydrogen-donating tendency, low dielectric constant, and low boiling point. For instance, it demonstrated that the choice of hydrogen donor influenced product distribution [34]; DMF selectivity was significantly higher when ethanol was used as the solvent compared to water or methanol. This phenomenon was attributed to the lower solubility of hydrogen and the enhanced etherification effect promoted by lower alcohols [148]. Moreover, specific catalysts can enhance H₂ release through ethanol decomposition and hydrogenolysis. This was exemplified by the Mo@C-900 catalyst embedded in carbon microspheres, which leverages the synergy between nanoparticles and adjacent defective carbon to facilitate the selective hydrogenation of HMF, achieving a DMF yield as high as 91.6% [54]. In summary, the use of isopropanol and renewable ethanol, both possessing excellent hydrogen-donating capabilities, provides an effective and economical route for the production of DMF via the selective hydrogenolysis of HMF.

In contrast to the alcohols discussed above, FA, an acidic solvent, can also serve as a hydrogen donor. It is a readily available hydrogen donor material that is easy to store and handle [149]. FA can be derived from biomass oxidation or acid hydrolysis and has attracted considerable attention due to its high hydrogen content (4.4 wt%), liquid state at room temperature, and ability to catalytically decompose into H₂ and CO₂ under mild conditions [150]. Zhang’s group discussed the advantages of FA as a hydrogen source relative to methanol, ethanol, and isopropanol in the catalytic transfer hydrogenation of HMF [106]. It was found that while methanol, ethanol, and isopropanol all resulted in high HMF conversion, DMF yields were low, a phenomenon attributed to acetal formation in methanol and ethanol systems [151,152] and etherification in the isopropanol system [153]. By selecting FA, a green and cost-effective hydrogen donor, a DMF yield of 96.7% was achieved after 2 h at 200 °C. Moreover, FA plays multiple roles in the hydrogenolysis of HMF to DMF: (i) suppressing ring hydrogenation to enhance DMF selectivity; (ii) serving as a mild hydrogen source; (iii) generating formate esters as intermediates; and (iv) forming protonated intermediates that lower the activation barrier for C-O bond cleavage. Functions (ii) through (iv) collectively contribute to the enhancement of catalytic activity [31].

Beyond alcohols and FA, silane compounds have emerged as a novel class of hydrogen donors in the HDO of HMF to DMF, offering distinct advantages. Among them, polymethylhydrosiloxane (PMHS) has attracted considerable attention due to its non-toxic, inexpensive, safe, and air- and moisture-stable characteristics. As a byproduct of the silicone industry, PMHS has been widely applied in the reduction in amides, carbonyl compounds, and other functional groups. The electronegativity difference between Si and H in silanes imparts a partial negative charge to the hydrogen atom, enabling the generation of H⁻ species through catalyst activation under mild conditions for use as a hydrogen donor [142,154,155,156] using commercial PdCl₂ as the catalyst and PMHS as the hydrogen donor was studied, achieving a DMF yield of 89.7% from HMF HDO within 0.5 h at room temperature (298 K). A synergistic effect among the hydrogen donor PMHS, in situ generated Pd⁰ species, and the acidic HCl generated during the reaction was found to promote the efficient formation of DMF. Notably, this system required no external hydrogen supply or pre-reduction treatment and can also convert biomass-derived sugars to DMF in a one-pot reaction with relatively high efficiency. Li’s group further applied the hydride (H⁻) from PMHS in a direct conversion system for sugars catalyzed by hydrophobic palladium nanoparticles [143]. Under mild conditions (110–130 °C), this system achieved efficient conversion of various sugars to furan-based biofuels (DMF and 2-methylfuran) with yields exceeding 95%. The study demonstrated that PMHS, as a green hydrogen donor, does not interfere with the upstream reactions for the in situ formation of furanic aldehydes/alcohols from sugars and selectively promotes the subsequent HDO steps. The catalytic system exhibited excellent stability, with a polydimethylsiloxane-coated Pd/MIL-53(Al) catalyst showing negligible deactivation or Pd leaching over at least five cycles, achieving 100% HMF conversion and 99% DMF yield. Thus, PMHS represents an inexpensive, green, and readily accessible hydrogen donor, offering a new route for the efficient conversion of HMF and sugars to DMF under mild conditions.

The safety and toxicity profiles of solvents and hydrogen donors used in HMF-to-DMF conversion present critical concerns. These risks are seldom addressed in studies focused on catalyst development. Tetrahydrofuran (THF) is the most prevalent solvent in reported catalytic systems for this reaction. It is highly flammable, with a flash point of −17 °C, and prone to forming explosive peroxides upon exposure to air. These properties create a severe hazard during distillation-based solvent recovery. Formic acid is an attractive renewable hydrogen donor, but it carries notable safety risks. It is corrosive and can decompose exothermically to carbon monoxide (CO) under specific reaction conditions. These hazards necessitate rigorous process control for industrial use. Polymethylhydrosiloxane (PMHS) is non-toxic and non-flammable in its native form. However, it generates stoichiometric quantities of siloxane byproducts during the reaction. The separation and safe disposal of these byproducts pose substantial practical challenges for process scale-up. In contrast, ethanol and isopropanol offer inherently safer hazard profiles. Ethanol benefits from inherent renewability and well-established industrial handling infrastructure. Despite these advantages, both solvents require explosion-proof process equipment for safe large-scale operation. There is a clear divergence between two classes of solvents: those that deliver optimal laboratory-scale yields, and those that are compatible with safe, scalable industrial operation. This gap underscores the critical need for systematic solvent-catalyst co-design. This co-design process must be guided by both catalytic performance metrics and inherent safety considerations.

Although the primary focus of catalyst development has been on activity and selectivity, the environmental footprint of the overall process warrants equal scrutiny. The majority of high-performing catalytic systems were summarized in this review operate in organic solvents such as THF, n-butanol, and 1,4-dioxane, which are derived from petrochemical feedstocks and carry inherent toxicity and volatility concerns. From a life-cycle perspective, the energy input associated with elevated temperatures (130–200 °C) and moderate to high H₂ pressures (1.0–3.0 MPa) represents a substantial contributor to the global warming potential of DMF production, particularly when H₂ is sourced from steam methane reforming without carbon capture. Moreover, the environmental cost of noble metals, encompassing energy-intensive mining, ore processing, and geopolitical supply risks, must be weighed against the marginal gains in catalytic efficiency. In this context, the recent shift toward non-noble metal catalysts and catalytic transfer hydrogenation using renewable hydrogen donors (e.g., formic acid, isopropanol, and polymethylhydrosiloxane) represents not merely an economic imperative but an environmental one.

5. Current Challenges and Future Prospects

In summary, the conversion of the platform chemical HMF to DMF offers a promising route to reduce dependence on fossil resources. Supported monometallic and bimetallic catalysts have enabled the selective hydrogenation of HMF to DMF under relatively mild conditions, achieving significant advances in both selectivity and yield. Notably, the research focus for the catalytic synthesis of DMF has shifted from noble metal-based catalysts to non-noble metal systems. However, the industrial production of DMF has not yet been realized, and the development of highly effective, low-cost catalysts remains a critical objective. Current catalytic systems still face challenges such as harsh reaction conditions involving high temperatures, high pressures, and the use of hazardous organic solvents. Non-noble metal catalysts also suffer from issues including relatively low catalytic efficiency and limited cycling stability in some cases. As for the future industrial production of DMF, the most challenging aspect to industrialization lies in the HMF supply chain itself. Currently, there are two commercial grades of HMF available on the market. The first is an aqueous solution, which is low-cost but highly impure, containing water and other organic impurities. The second is high-purity 5-HMF, which remains costly at present. This price discrepancy arises for a clear reason: this heat-sensitive and highly polar product must be separated from aqueous reaction mixtures through an extraction process with high yield losses, which keeps its price prohibitively high [157]. This high feedstock cost renders DMF produced via this pathway far outside the economic viability range for any bulk fuel application. Until scalable, energy-efficient separation technologies are developed to bridge the purity–cost gap, the catalytic performance of HDO catalysts will remain of secondary relevance to industrial viability, no matter how impressive the performance is. The second key barrier concerns the absence of clearly defined techno-economic performance targets for HDO catalysts. Unlike well-established industrial hydrogenation processes, the present studies on HMF-to-DMF conversion lack consensus on the minimum catalyst productivity, stability, and cost thresholds required for commercial feasibility. To address this gap, we propose that the field should work toward the following actionable benchmarks:

• TOF is reached 100 h⁻¹ under mild conditions (≤150 °C, ≤10 bar H₂ or equivalent CTH conditions), to ensure reasonable reactor productivity. This target has already been achieved by several noble metal catalysts, but it is rarely met by non-noble metal systems.
• Stable operation is at least 1000 h with less than 10% activity loss. This lifetime corresponds to a catalyst service life that justifies the upfront material investment.

To meet these proposed targets, catalyst systems will require continued innovation. Accordingly, we propose the following key research directions:

• Develop efficient and stable non-noble metal catalytic systems under mild conditions. Although non-noble metal catalysts offer significant cost advantages, they often require high reaction temperatures and hydrogen pressures, and their long-term operational stability remains insufficient. Future efforts should focus on precisely constructing active centers with high intrinsic activity through electronic structure modulation (e.g., heteroatom doping, alloying) and nanostructure design, aiming to achieve efficient conversion under low-temperature and ambient-pressure hydrogen conditions.
• Deepen understanding of reaction mechanisms and advance the application of characterization techniques. The HDO of HMF involves complex reaction pathways with multiple competing processes. In situ characterization techniques combined with theoretical calculations are needed to elucidate, at the molecular and atomic levels, the structural evolution of active centers, the transformation pathways of key intermediates, and the mechanisms governing side-reaction suppression. Such insights can provide theoretical guidance for the rational design of catalysts. Given the complexity and heterogeneity of the reaction system, more precise characterization is essential to determine reaction mechanisms and pathways.
• Developing green solvent systems and sustainable reaction processes. Conventional reaction systems often rely on toxic organic solvents, which are inconsistent with the principles of green chemistry. Efforts should be directed toward developing bio-based solvents, aqueous-phase systems, or solvent-free systems, and toward systematically investigating how solvent properties influence reaction selectivity. In addition, continuous-flow reaction technologies should be explored to enhance mass and heat transfer efficiency while reducing energy consumption and solvent usage.
• Advance integrated processes for the one-pot synthesis of DMF directly from biomass feedstocks. Current DMF production typically relies on two-step processes involving multiple separation and purification steps, which limit overall yields. Future efforts should focus on designing multifunctional catalysts that combine dehydration and HDO functionalities, enabling the sequential completion of hydrolysis, dehydration, and HDO steps in a single reactor using feedstocks such as sugars, cellulose, or even raw biomass. Compatible product separation technologies should also be developed to enhance process economics. In particular, the development of multifunctional catalysts capable of one-pot conversion from crude sugar streams directly to DMF, together with compatible product recovery technologies, offers the most viable path toward economically sustainable DMF production.

The findings summarized in this review carry significant practical implications for sustainable biofuel development. DMF exhibits a unique combination of three desirable fuel properties: high energy density, high octane number, and low water solubility. These properties collectively make it a more attractive gasoline substitute than ethanol. Three core design principles have been elucidated in this review: bifunctional metal–oxophilic promoter synergy, support engineering, and metal–support electronic interactions. These principles provide a rational, evidence-based framework for developing industrially viable catalysts for HMF-to-DMF conversion. The field is converging toward earth-abundant, non-noble metal catalyst formulations that rival the performance of noble metal systems. This transition promises two key benefits: significant reductions in raw material cost and decreased supply chain vulnerabilities. The rapid progress in cost-competitive, structurally stabilized catalyst systems documented herein provides strong grounds for cautious optimism. The catalytic hydrogenolysis of HMF to DMF can transition from a well-studied academic reaction to a cornerstone technology of the future bio-based economy.