The sustained use of petroleum and coal resources causes the continuous depletion of fossil materials and exacerbates the greenhouse effect. Hence, both industry and academic research groups are seeking fossil fuel alternatives in an attempt to minimize fossil fuel consumption and CO₂ emissions. Biomass-derived resources have gained significant attention as promising alternatives for fossil materials because biomass resources can be converted into value-added chemicals and biofuels, are relatively affordable and abundant, and eventually close the carbon cycle in ecosystems [1]. Therefore, it is extremely important to develop efficient, environmentally friendly and economically viable technologies to convert biomass resources into industrially viable bulk chemicals at large scale.

Glucose is the most abundant unit in cellulosic biomass and the most widely distributed monosaccharide in nature. Because glucose can be used to produce gluconic acid, which has widespread applications in the pharmaceuticals, detergents, concretes, biodegradable polymers, and food industries, its catalytic oxidation has been extensively researched in the past decade [2, 3]. Furthermore, gluconic acid can be oxidized into glucuronic acid and its derivatives, which are important chemicals for the production of nylon, plastics, and food additives [4]. Currently, gluconic acid is mainly manufactured via the aspergillus fermentation of glucose [5]. However, this process is unsuited for continuous large-scale production because it generates biological wastes and the enzymes are deactivated at low pH [6]. Furthermore, this is a multistep, time-consuming, cumbersome and expensive process [5]. Photochemical, electrochemical, or catalytical processes that are performed in the presence of noble metal-based heterogeneous materials are alternative methods for the biochemical synthesis of gluconic acid. The photocatalytical oxidation of glucose can be carried out under UV and visible light [7, 8] and using Pt- [9], Cr- [10], Cu- [11] and Ag- [12] based carbon or metal oxide materials as active catalysts. The electrocatalytic oxidation of glucose can be conducted in glucose fuel cells that use Pt/C [13], Au/C [14], Pd/C [15], Cu₂O/TiO₂ [16], or MnO₂ [17] as active catalysts. The noble metal-based heterogeneous catalytic oxidation of glucose is more efficient and environmentally friendly [18]. Various catalysts have been used for the oxidation of glucose to gluconic acid, including Au/C [19], Au/Al₂O₃ [20], Au/TiO₂ [21], Pd/C [22], Pd/Al₂O₃ [23], Pt/C [24, 25], Pt-TiO₂ [26], CuO [27], and FeCl₃ [28]. Of these, Au-based catalysts are the most popular. The use of Pt or Pd is limited owing to the leaching of toxic promoters in the reaction mixtures. Moreover, these catalysts deactivate easily owing to oxygen poisoning and the unfavorable product adsorption on their active centers. Au-based catalysts are less sensitive to oxygen poisoning than Pt- and Pd-based ones and are more tolerant to a wide range of pH values [29].

Recently, glucose oxidation technologies have developed rapidly and have replaced heterogeneous catalysis. Ramachandran et al. [4] summarized the production and use of microbial gluconic acid. The patented processes and their advantages and disadvantages for the production of gluconic acid production were reviewed in 2007 [30, 31]. In 2015, Chatterjee et al. [1] reviewed the chemical conversion methods of carbohydrates and briefly summarized the oxidation of glucose to gluconic acid. Ennaert et al. [32] reviewed the catalytic conversion of biomass using zeolites, and analyzed the performances of zeolites with respect to Lewis acid, Brönsted acid, or multifunctionality. However, the oxidation of glucose to gluconic acid was not mentioned in their paper. In 2017, Saeed et al. [33] summarized the emerging methods for the synthesis of glucaric acid. Typically, glucaric acid is as a byproduct of the oxidation of glucose to gluconic acid. In 2018, Zhang et al. [34] reviewed the oxidation of cellobiose to gluconic acid and Diamond et al. [2] reviewed the catalytic conversion of low-cost sugars into market chemicals via a continuous flow, trickle bed process. In 2016, Zhang et al. [29] reviewed the techniques for converting C₅ and C₆ sugars into chemical intermediates. The oxidation of glucose to gluconic acid over Pd- and Au-based catalysts is further discussed in detail herein. However, more catalytic systems have been proposed in the past five years. The selectivity and long-term stability of catalysts needed to be improved for the catalytic processes to be competitive with the existing fermentation processes. Therefore, the ultimate goal of researchers would be to seek inexpensive metal catalysts that present comparable catalytic activity for the oxidation of glucose with noble metals, as this would enhance the overall economics of the gluconic acid manufacturing process. Hence, it would be necessary to critically revise the most recent progress in this field, summarize the structure-function relationship of various catalysts, provide potential solutions for the unsolved problems, and offer possible insights into the development of the glucose oxidation process. The biocatalytic, photocatalytic, and electrocatalytic glucose oxidation methods will be only briefly introduced herein.

In the past decades, various highly active Au-based catalysts have been widely used for the aerobic oxidation of glucose [35]. When Au was supported on various types of carbons, including activated carbon, carbon nanofibers, carbon sheets, and carbon nanotubes, it displayed good catalytic performances for the aerobic oxidation of glucose [36, 37]. Moreover, when Au was supported on various metal oxides, e.g. Al₂O₃, CeO₂, TiO₂, ZrO₂, MgO, and Fe₃O₄, it presented good catalytic activity and selectivity [38-46]. The representative types of supported Au catalysts are discussed in details in the following section.

Carbonaceous materials have been considered to be suitable supports for anchoring metal nanoparticles (NPs) owing to their high porosity and the multifunctionality of surface groups [47]. Carbon materials exhibit excellent properties, such as (1) good intrinsic catalytic activity before and after doping with heteroatoms (N, O, B, P, and S) owing to high electron conductivity, which favors their use as metal-free catalysts [48]; (2) high chemical and thermal stability in both acidic and basic media [49]; (3) availability and facile synthesis from abundant biomass, such as glucose [47]; (4) widely tunable pore size, pore connectivity, pore size distribution, and electronic properties, which can be adjusted by changing the surface functional groups or heteroatoms [49]; (5) facile metal phase recovery via burning; (6) lower costs compared to other types of supports [50]; and (7) high hydrophobicity and weak chemical interactions with the metal species, which could facilitate the formation of active sites and prevent the active metal species from leaching [51]. The above-mentioned advantages of carbon supports render such solid catalysts to be most promising candidates for future industrial uses [52]. Several representative carbon-supported Au catalysts with good performance for the oxidation of glucose are further reviewed and discussed below.

Carbon-supported Au clusters synthesized by capping Au particles with various ligands have been reported as efficient catalysts for the oxidation of glucose [48, 53]. Zheng et al. [48] synthesized activated carbon-supported rod-like Au nanoclusters by capping Au particles with conjugated delocalized pπ electron-mediated ligands (Au₂₅SPh/AC). They established an effective strategy and built models to transform homo-Au nanoclusters into alloy Au nanoclusters. The catalytic activity of Au₂₅SPh/AC (turnover frequency (TOF) of 13.5 s^‒1, Table 1, #1) was higher than that of Au₂₅SC₆H₁₃/AC, which is its aliphatic thiolate-protected Au analogue, (TOF of 9.56 s^‒1, Table 1, #2) and those of the commercial Pd/AC and Pd-Bi/AC catalysts [54] for the oxidation of glucose to gluconic acid. Density functional theory (DFT) calculations indicated that the conjugated delocalized pπ electrons mediated the electronic properties of Au kernels, and thus, improved the catalytic performance of Au₂₅SPh/AC [48]. Recyclability studies indicated that the catalytic activity of Au₂₅SPh/AC decreased after the third recycle. The transmission electron microscopy (TEM) images of the used catalysts revealed that the phosphine ligands detached from the Au nanoclusters owing to the presence of bases in the reaction media. Hence, the uncovered Au nanoclusters aggregated and the catalytic activity of Au₂₅SPh/AC decreased. Li et al. [53] used their experimental results to propose a catalytic mechanism for the oxidation of glucose over Au nanorods. First, the base facilitated the release of the uncovered active Au sites via the detachment of the phosphine ligands from the clusters. Second, the glucose molecules were adsorbed and activated on the exposed Au sites, and O₂ was activated on the surface of the activated carbon rather than that of the ligand-protected Au nanoclusters [53]. The activation of O₂ was determined to be the rate-determining step because the activation rate of O₂ on activated carbon was very low [53]. Liu et al. [54] reported a different ligand abscission trend for the Au(PET)/AC catalyst. They determined that the partial removal of the surface 2-phenylethanethiolate (PET) ligands could expose more active sites for oxidation reactions. Au clusters with PET ligands were anchored on active carbon via simple impregnation. The Au/AC-120 catalyst (TOF of 1.51 s^‒1, Table 1, #3) exhibited higher catalytic activity than Au/AC (TOF of 1.43 s^‒1), Pd/AC (TOF of 0.23 s^‒1), and Pd-Bi/AC (TOF of 0.36 s^‒1). In this case, the active carbon support exhibited no catalytic activity, which indicated that the catalytic activity was exclusively attributed to the Au clusters. The TEM images of Au/AC-120 confirmed that the Au clusters, with average size of 1.6 nm, were highly dispersed on the surface of the activated carbon (Fig. 1(a) and (b)). To improve the catalytic activity of Au/AC-120, Liu et al. [54] claimed that more active sites were formed via the removal of the surface PET ligand. The recyclability of the catalysts was also studied, and it was reported that after seven cycles, the catalytic activity decreased by less than 6%; moreover the PET ligands were still present on the surface of the Au clusters.

Particle size plays an important role for the catalytic performance of catalysts [35]. Delidovich et al. [35] reported that the catalytic activity of Au/Al₂O₃ increased as the size of the Au particles decreased. Solmi et al. [25] analyzed the catalytic activity of poly(vinyl alcohol) (PVA)-protected AuBi NPs supported on activated carbon (Au_PVA/AC). The TEM images of Au_PVA/AC revealed that its particle size was approximately 7.3 nm (Fig. 2(a) and (b)). The X-ray photoelectron spectroscopy (XPS) profile of Au_PVA/AC confirmed the presence of metallic Au and Bi as well as that of Bi³⁺ ions in its structure. The Au_PVA/AC catalyst was active for the direct oxidation of glucose in the presence of NaOH (TOF of 0.024 s^‒1, Table 1, #4), and it was reported that the only primary product of the reaction was gluconic acid, which was subsequently oxidized into glucaric acid or other lighter acids. Moreover, Solmi et al. [25] determined that the particle size and experimental parameters significantly affected gluconic acid selectivity. The smaller the Au particles, the more active and less selective the catalysts. Light acids were formed not only via Au-catalyzed reactions, but also via the non-catalytic thermal decomposition of glucose in the presence of bases. Moreover, it was reported that organic residue deposition and NP agglomeration and sintering were the main reasons for the deactivation of the catalyst [25]. Qi et al. [55] synthesized Au/CMK-3, which is an ordered mesoporous carbon (OMC)-supported Au catalyst, and evaluated its performance for the oxidation of glucose to gluconic acid in the absence of bases (TOF of 4.92 s^‒1, Table 1, #5), and reported that glucose conversion was higher when the mesopore volume of the catalyst was larger. The typical P6mm symmetry of the CMK-3 support and uniform dispersion of the Au species could be distinctly identified in its TEM images (Fig. 2(c) and (d)). They also synthesized other supported Au catalysts, such as Au/SBA-15 (TOF of 4.76 s^‒1, V of 0.62 mL g^‒1, Table 1, #6), Au/CNTs (TOF of 2.4 s^‒1, V of 0.38 mL g^‒1, Table 1, #7), Au/graphene (TOF of 1.75 s^‒1, V of 0.11 mL g^‒1, Table 1, #8), Au/AC (TOF of 0.61 s^‒1, Table 1, #9), Au/ZrO₂ (TOF of 0.35 s^‒1, Table 1, #10). The catalytic activity of Au/CMK-3 (2.98 nm, V of 1.17 mL g^‒1) was much higher than those of the other materials. The authors noticed that glucose conversion increased and gluconic acid selectivity decreased as the reaction temperature and oxygen pressure increased. The deactivation of the catalyst was mainly attributed to the adsorption of carboxylic acids on its surface. The Au/CMK-3 catalyst could be regenerated via simple calcination and NaOH treatment methods. This conclusion was in agreement with the information reported by Megias-Sayago et al. [19]. However, Qi et al. [55] also claimed that calcination could not regenerate the catalyst completely, because they noticed that the Au NPs were larger than the as-prepared catalysts ones owing to their aggregation.

Ding et al. [56] and Sun et al. [57] synthesized Au-based OMCs and used them for the oxidation of glucose to gluconic acid. Compared to traditional active carbons, OMCs, presented larger surface area and more uniform mesopore size. Ma et al. [52] prepared a series of Au-based OMC catalysts via nano-replication and colloidal Au deposition. The particle size of the catalysts was tuned in the range of 3.2–7.6 nm via the addition of different amount of boric acid to the reaction mixture during the preparation of the carbon supports (Fig. 3(a)-(c)). When the amount of doped B increased, the hydrophilicity of the surface increased, which resulted in the facile infiltration of the Au precursor and higher Au loading of the catalyst. The Au/C-OMC-MSU catalyst with 5.4 nm mesopores (TOF of 4.31 s^‒1, Table 1, #11, and Fig. 3(d)) presented better catalytic performance than those reported by other researchers, and that was ascribed to the 2D hexagonal meso-structure that featured 5.4 nm channels, highly dispersed 3.3 nm Au NPs, and abundant active oxygen species. This structure could not be easily blocked by metal particles. Moreover, Ma et al. [52] concluded that the more facile the desorption of the surface-active oxygen species was, the higher the catalytic activity for the oxidation of glucose was. The schematic model of the active sites on the surface of the Au/OMC catalysts for the oxidation of glucose is depicted in Fig. 3(e).

Megias-Sayago et al. [49] investigated the effect of the catalyst synthesis conditions on particle size, and determined that the size of the colloidal Au particles was dependent on the stabilizing/reducing agent-to-Au ratio [49]. At constant PVA/Au ratio, increasing the NaBH₄/Au ratio could decrease the average size of the Au particles [49]. The maximum conversion for the Col X/AuC X catalyst was obtained using 9 nm Au particles (TOF of 1.5 × 10^‒2 s^‒1, Table 1, #12) [49]. Megias-Sayago et al. [19] immobilized Au colloids on carbon (AuC.I) and tested the obtained catalyst for the oxidation of glucose in the absence of bases (TOF of 1.2 ×10^‒2 s^‒1, Table 1, #13). Possible methods for the deactivation/reactivation of the AuC.I were also discussed. The changes in the morphology of the Au particles was the most important reason because the oxidation of glucose occurred at the active Au sites. The agglomeration of the Au particles, metal leaching, active sites blocking, and support modification could cause changes in the state of Au. The inductively coupled plasma analysis results indicated that 15% of the initial Au loading leached during the first run.

The effects of the catalyst preparation method and nature of the support on the physical and chemical properties of Au-based catalysts have been widely studied in the literature [58]. However, the effects of the heteroatoms and surface functional groups on the properties of the catalysts was rarely investigated. Lama et al. [58] synthesized various salt-templated porous carbon supports and functionalized their surface with N₂, O₂, or H₂. The surface atomic structure of the carbonaceous supports was studied. The asprepared hydrophobic catalysts (Fig. 4) presented smaller Au NPs (3–5 nm), and therefore the catalytic activity of the Au-C catalyst (TOF of 1.5 s^‒1, Table 1, #14) was high. The O₂- and N₂-doped hydrophilic supports presented larger Au particles and lower catalytic activity than the H₂-doped ones. In addition, the transportation of the substrate to the active metal center was slower owing to the stronger adsorption of the substrate on the hydrophilic support. Recycle testing indicated that during the second cycle the catalytic activity of both Au-C and Au-C-H remained nearly unchanged.

In addition to carbon-supported catalysts, metal oxide-supported Au catalysts, where metal oxides, such as Al₂O₃ [59], TiO₂ [60], ZrO₂ [61], and CeO₂ [62] were used as supports have also been intensively studied. Delidovich et al. [35] prepared Au/C and Au/Al₂O₃ catalysts and compared their catalytic performance for the oxidation of glucose in the presence of bases. When the size of the metal particles of the Au/Al₂O₃ catalyst ranged from 1 to 5 nm, the recorded TOF values were extremely high. Moreover, the authors carried out kinetics studies in a wide range of glucose/Au molar ratios and observed different trends. At high glucose/Au ratios, the catalytic performance of Au/Al₂O₃ was higher than that of Au/C. In the presence of excess glucose, the reaction rate and overall reaction kinetics were controlled by the O₂ transfer at the gas/liquid/solid interface and internal O₂ diffusion. By contrast, the oxidation rate of glucose over the Au/C catalyst exceeded that over the Au/Al₂O₃ catalyst at lower glucose/Au ratios, because the reaction rate was determined by the dissolution of O₂ in the liquid phase. This was attributed to the facile transfer of O₂ to the active sites of the hydrophobic carbon support. The models for the oxidation of glucose over Au/Al₂O₃ and Au/C catalysts are illustrated in Fig. 5(a) and (b).

Saliger et al. [63] also studied the kinetics of the oxidation of glucose over the Au/Al₂O₃ catalyst in the presence of H₂O₂ and under mild alkaline conditions (pH 9) at 40 ℃ (TOF of 27.2 s^‒1, Table 1, #16). H₂O₂ is an oxidant that can be used as an alternative to O₂ to avoid the limiting mass transfer on the oxidation rate attributed to the low solubility of O₂ in water [64]. When H₂O₂ was used as the oxidant instead of O₂ the yield of gluconic acid was 10 times higher. Moreover, the activation energy was 48 kJ mol⁻¹ and the reaction order was 0.5 with respect to glucose [63]. Ono et al. [65] tracked the concentration of dissolved O₂ and determined that the effective oxidants were the intermediate peroxo and superoxo species generated by the decomposition of H₂O₂ on the Au catalyst, and not H₂O₂ itself. These results were in agreement with those reported by Pruesse et al. [66, 67], and Cao et al. [68]. Pruesse et al. [66] prepared 250 μm monodisperse γ-Al₂O₃ beads using the sol-gel method. The egg-shell Au/Al₂O₃ catalyst was synthesized using the deposition-precipitation method using urea as the precipitation agent. These catalysts were tested for the oxidation of glucose using H₂O₂ as the oxidant. The authors calculated the effectiveness factors of the Au/Al₂O₃ catalyst and concluded that its catalytic activity was higher when H₂O₂ was used as the oxidant (TOF of 0.37 s^‒1, Table 1, #17) than when O₂ was the oxidant (TOF of 0.28 s^‒1, Table 1, #18). The Au/Al₂O₃ catalyst presented long-term stability for the oxidation of glucose.

In the absence of bases, most catalysts presented poor activity and gluconic acid selectivity. However, Au/Al₂O₃ exhibited good activity for the oxidation of glucose when H₂O₂ was used as the oxidant and the reaction was microwave assisted. Rautiainen et al. [20] reported that the TOF values of the Au/Al₂O₃ and Au/MgAl₂O₄ catalysts were 3.58 and 2.78 s^‒1, respectively, (Table 1, #19 and 20) when the reactions were microwave assisted. The TEM images of the catalysts revealed that the particle size of Au/MgAl₂O₄ was 3.8 nm. When the reaction was performed using an oil bath, the conversion was only 62%, compared with the conversion of 83%, which was reached for the microwave-assisted reaction. Liao et al. [69] speculated that microwave irradiation could heat the reactants rapidly and uniformly, because aqueous glucose solutions present good microwave absorbing capacity. However, product selectivity was not affected by the heating method, because catalysts with low microwave absorbing capacity do not overheat [70]. After four cycles, the Au/MgAl₂O₄ catalyst exhibited only minor particle growth and its activity was almost unchanged. Rautiainen et al. [20] also studied the decomposition of H₂O₂ and reported that Au presented superior activity for this reaction. Hence, the oxidation activity of the Au/MgAl₂O₄ catalyst was closely related to the H₂O₂ decomposition rate [20].

Xin et al. [59] reported that the Au/Al₂O₃ catalyst that was prepared using a wetness Mb-mediated bioreduction method, exhibited high activity (TOF of 5.58 s^‒1, Table 1, #21). This synthesis method involved no loss of Au or waste formation; moreover, it required only a small setup and short preparation time. The as-prepared Au/Al₂O₃ catalyst particles with the Au loading of 1 wt% presented the average diameter of 4 nm. Furthermore, the catalyst maintained its activity after eight reaction cycles.

In addition to Al₂O₃, TiO₂, CeO₂, and ZrO₂ have been frequently studied for the oxidation of glucose. All these supports are affordable and present porous structures, acid-base properties, high surface areas, and chemical stability [60]. Some research groups studied the effect of the supports on the activity of different metal oxide-supported Au catalysts, which were used for the oxidation of glucose. Benko et al. [62] synthesized SiO₂- (TOF of 0.8 s^‒1), TiO₂- (TOF of 0.65 s^‒1), and CeO₂- (TOF of 0.52 s^‒1) supported Au catalysts via the deposition of colloidal Au and tested them for the oxidation of glucose (pH = 9) to elucidate the role of the support during the catalytic process. Their results indicated that SiO₂-supported Au that featured large particles was more active during the liquid phase reaction, which indicated that the high dispersion and nature of the support were more important than the particle size. Signoretto et al. [71] prepared CeO₂-, TiO₂-, and ZrO₂-supported Au catalysts (Table 1, #22, 23, and 24, respectively) and demonstrated that the CeO₂-supported Au catalyst was the most active of the three analyzed catalyst owing to its Au NPs being the smallest. In addition, CeO₂ could store and deliver O₂ to change the oxidation state from 3+ to 4+ [71]. Au/TiO₂ did not present catalytic activity owing to the metal agglomerates. The TEM and SEM analysis data together with the experimental results indicated that the particle size depended on the synthesis method and the nature of the support. Author et al. [71] compared the colloidal and deposition-precipitation methods and concluded that the use of PVA was more favorable for the preparation of Au NPs than the use of polyvinylpyrrolidone (PVP). For this method, the surfactant was only partially removed, so the remaining surfactant could stabilize the sample and prevent the aggregation of the NPs. This conclusion was in agreement with the results obtained by Liu et al. [54]. No leaching, sintering, or coating was observed on the surface of Au/CeO₂, and therefore no activity loss was observed for this catalyst. Ishida et al. [41] reported different results on the effects of the supports on the catalytic activity. They prepared ZrO₂- (3.7 nm), CeO₂- (4.0 nm), and TiO₂- (2.9 nm) supported Au catalysts using the deposition-precipitation method. The TOF values of the Au/ZrO₂ catalyst at 50 ℃ were 56 and 45 s^‒1 at pH 9 and 9.5, respectively. Therefore, the authors demonstrated that the nature of the support played the major role for gas-phase oxidation reactions, whereas the size of the Au particles was more important for liquid-phase reactions. Ishimoto et al. [72] investigated the reaction mechanism of Au/ZrO₂- and Au/SnO₂-catalyzed glucose oxidation using the theoretical DFT method. They demonstrated that the catalytic activity of Au/SnO₂ was higher than that of Au/ZrO₂, and these results were attributed to the positive charge of the Au ions that were obtained via the charge transfer from the Au species to the support.

The effect of the support was also studied in the absence of base promoters. Wang et al. [73] synthesized CeO₂- (TOF of 0.014 s^‒1, Table 1, #25), ZrO₂- (TOF of 0.017 s^‒1, Table 1, #26), and TiO₂- (TOF of 0.018 s^‒1, Table 1, #27) supported Au catalysts using the position-precipitation method. The nature of the support determined the electronic state and dispersion of the Au particles. The metal-support interactions affected the size of the Au NPs. Supports with higher Fermi levels presented stronger electronic interaction with the Au NPs, and thus, their agglomeration tendency was lower [74]. Wang et al. [73] concluded that the density of the Au NPs on the support surface could be maintained at low level to limit the sintering rate and increase the activity and stability of the catalysts. Megias-Sayago et al. [75] also observed that the catalytic activity of Au/CeO₂ was inversely proportional to the Ce loading. They studied the Au/Al₂O₃ (TOF of 0.054 s^‒1, Table 1, #28), Au/CeO₂-Al₂O₃ (TOF of 0.026 s^‒1, Table 1, #29), Au/CeO₂ (TOF of 0.017 s^‒1, Table 1, #30), and Au/CeO₂-ZrO₂ (TOF of 0.025 s^‒1, Table 1, #31) catalysts, which featured particles that ranged in size from 3.5 to 5.5 nm, for the base-free oxidation of glucose and determined that higher Ce loadings were associated with higher TOF values. In addition, the high Lewis acidity of the support decreased the gluconic acid selectivity and increased the lactic acid selectivity. Hence, further investigations on the support type, acidity, and stability are required. Zhuge et al. [76] synthesized a hybrid Ca-Al layered double hydroxide (LDH) via an in situ growth method and immobilized Au on it (Au/HAP-LDH). The synthesized catalyst was studied for the base-free oxidation of glucose, and it was determined that the Au/HAP-LDH catalyst presented high gluconic acid yield (> 98%), whereas its TOF was approximately 5.62 s^‒1 (Table 1, #32). These results were ascribed to the abundant number of surface Au (Au⁰/Au^δ+) and basic sites and confirmed the important effect of the composition of the support on the activity of the catalyst. Conversely, the hybrid 3D microstructure of the support enhanced the immobilization of the active Au species and the stability of the support.

Guo et al. [21] carried out a kinetics study to investigate the catalytic activity of metallic Au with different charge states. They synthesized TiO₂-supported ultrasmall Au clusters (1.2–1.7 nm) via the simple incipient wetness protocol using anthranilic acid as the stabilizing agent. Their results indicated that the Au^δ+ species were inactive for the oxidation of glucose in the absence of bases. The XPS profile of the catalyst revealed that Au⁰ (TOF of 0.53 s^‒1, Table 1, #35) exhibited higher catalytic performance than the commercial Pd-Bi/C (pH 9, TOF of 0.36 s^‒1) catalyst. The apparent activation energy of Au/TiO₂ was 47 kJ mol⁻¹, which was similar to that of unsupported Au colloids [21]. It was concluded that the oxidation of glucose occurred at the surface of the Au species rather than at the support/Au clusters interface (Fig. 6).

In addition to the particle size and the nature and composition of the support, the additive and ligands used for the preparation of catalysts could affect their catalytic activity for the oxidation of glucose. Wojcieszak et al. [77] anchored Au NPs on CeO₂ supports via the soft chemical reduction method using hydrazine. The as-prepared Au/CeO₂ catalysts did not present gluconic acid selectivity. However, the selectivity of the catalyst increased after it was modified with Cs (Table 1, #36), because CsOH increased the basicity of the catalyst. Moreover, An et al. [78] prepared Cs-modified Au/HPA catalysts and tested their performance for the oxidation of cellobiose to gluconic acid and other products. The results indicated that the gluconic acid selectivity was closely related to their Cs content. At low Cs content, the high acidity of the catalyst hampered the adsorption of gluconic acid, and thus, led to low conversion. Cao et al. [68] prepared an Au/TiO₂ catalyst and tested it for the base-free oxidation of glucose. The best catalytic activity (TOF of 0.082 s^‒1, Table 1, #37) was obtained for the catalyst synthesized using the sol-immobilization method. The authors claimed that the catalytic performance depended on both the Au particle size and the amount of stabilizing PVA ligand. For this method, the quantity of remaining ligand was more significant than the particle size. Karra et al. [79] tested the performance of Au NPs as electrocatalysts for the oxidation of glucose in neutral solutions. The results of the shape effect study indicated that the irregular Au NPs were more active than the spherical ones owing to the higher surface density of the incipient Au oxide which acted as fast redox mediator.

The possible size-dependent kinetic behavior during the oxidation of glucose over Au NPs of different sizes was attributed to different mechanisms. Therefore, the mechanisms of the glucose oxidation reaction over Au-based catalysts either in the presence or absence of bases are discussed in this section. Under alkaline conditions, the widely accepted mechanism was demonstrated using both experimental data and theoretical calculation as follows. The base contributed to the cleavage of the C–H bonds of the –CHO groups during the oxidation of glucose (Fig. 7) [72]. (1) The OH⁻ ions were adsorbed on the Au active sites. (2) Glucose was adsorbed on the –OH groups on the surface of Au. (3) The –CHO group of glucose interacted with the OH⁻ ions in the alkaline solution. (4) Protons were transferred from the –CHO groups to the OH⁻ ions and water was formed and released in the process. (5) The –OH groups were transferred from the surface of Au and formed gluconic acid. (6) The Au active sites were recovered. This mechanism was also supported by Pasta et al. [80] who proposed that ionic species presented a significant effect on the oxidation of glucose.

In the absence of bases the oxidation of glucose proceeded via a carbonyl conjugated radical mechanism, including the formation of H₂O₂ which activated O₂ [81]. The electron-rich Au species activated O₂ via nucleophilic attacks and formed hydroxyperoxides [81]. Many researchers, including Qi et al. [55], Saliger et al. [63], Pruesse et al. [66, 67], reported that H₂O₂ was generated during the reaction and claimed that the formed H₂O₂ oxidized glucose and then decomposed to O₂. H₂O₂ could also produce OH^‒ on the surface of Au via a series of chain reactions [55]. The carbonyl groups were converted to carboxyl groups by O₂ over the basic sites on the surface of the catalyst (Fig. 8). The mechanism of base free glucose oxidation reaction was determined to be as follows. (1) Glucose molecules coordinated on the basic sites on the surface of the catalyst via carbonyl groups. (2) The adsorbed glucose molecule reacted with water to form the RCH(OH)₂ intermediate [82-84]. (3) The RCH(OH)₂ intermediate coordinated to the Au species via Au–H bonds and formed an unstable hydride. (4) The Au sites activated the O₂ in the solution via a double linear model and formed Au–O–O–Au species [85, 86]. (5) The hydride adsorbed on the surface of Au and basic sites formed Au–H and gluconic acid via the elimination of hydrogen. (6) Water formed and the active Au sites were recovered. This mechanism was in accordance with the key reaction steps proposed by Comotti et al. [87] and the DFT calculation results reported in the literature [83].

Au supported on either carbon or metal oxides is highly active and selective for the oxidation of glucose in the presence and absence of bases owing to its unique size, shape-dependent, catalytical, optical, and electronic properties, being different from those of other metals. These physical and chemical properties could be tuned easily by changing the preparation methods and conditions. The most common factors that have been reported to affect the catalytic performance of supported Au catalysts could be summarized as follows. (1) The presence of bases. Bases could not only improved catalytic activity by promoting the cleavage of the C–H bonds of the –CHO groups during the oxidation of glucose, but could also reduce the leaching and aggregation of Au and improve the stability of the catalysts. (2) The specific size of Au NPs and the dispersion of Au NPs on supports. (3) The interactions between Au NPs and those between Au NPs and supports. (4) The structure and characteristics of the supports. (5) The nature of the ligands and coverage of the Au active sites. Particle size played the most important role in establishing the activity and selectivity of Au catalysts for the glucose oxidation reaction. Other factors could directly affect particle sizes and indirectly affect the catalytic performance of Au catalysts, e.g., the catalyst synthesis method could tune the metal particle size. Impregnation, co-precipitation, deposition-precipitation and colloidal methods are the most reported techniques that have been used to prepare heterogeneous catalysts. Dry impregnation and co-precipitation usually produce less active large Au particles (> 30 nm) and low Au dispersion owing to the Cl^‒ ions-promoted sintering that occurs during the heat treatment. The deposition-precipitation method could pass on hydrogen for reduction, but presents poor reproducibility and Au capture efficiency.

Supported Au catalysts for the oxidation of glucose have been restricted. Au leaching, sintering, poisoning, and covering lead to activity loss during recycling tests. Therefore, it is still necessary to further analyze the mechanism of catalyst deactivation in depth and comprehensively using theoretical calculation.

Noble metals have been reported to be highly active for the oxidation of glucose. Bi- or Pb-loaded Pd-based catalysts for the oxidation of glucose have been reported to be the best alternatives to the conventional fermentation process [88]. However, Bi could cause product contamination, and thus, further purification would be required. Furthermore, the high cost of noble metals is the primary drawback for the commercialization of these catalysts. Hence, researchers have focused on studying the synthesis methods, support effects, and reaction conditions to improve the activity and stability of Pd-based catalysts and avoid the use of Bi.

Haynes et al. [22] selected carbonaceous materials (carbon black) as supports to anchor Pd particles to avoid internal diffusional limitations during the oxidation of glucose. They compared the urea-assisted deposition and precipitation reduction methods to elucidate the strategy for controlling the size of the Pd particles. The authors determined that the catalysts prepared using the precipitation reduction method at 573 K (particle size of 7 nm) exhibited higher catalytic activity than that prepared via homogeneous deposition (Table 2, #1 and 2). A balance could exist between the O coverage and the accessibility of glucose (Fig. 9). The catalytic activity of Pd-based catalysts that featured particles smaller than 7 nm for the oxidation of glucose appeared to be less efficient. Parmon et al. [24] also claimed that Pd/C catalysts with small particle size (3 nm) deactivated more rapidly and were less performant. According to the well-accepted oxidative dehydrogenation mechanism, glucose first adsorbed on the Au active sites to abstract H atoms. However, the strength of the H atom adsorption could be weakened by the presence of electronegative O atoms. Hence, the catalyst with the smallest size was less active owing to the high O coverage of the surface. Conversely, it would be more difficult to bind glucose molecules to the Au active sited and abstract H atoms for the largest catalyst particles. Recycling experiments revealed that the loss of activity was not caused by Pd leaching, but the coverage of the Pd particles by O layers [24]. Liu et al. [89] used Pd/C as the catalyst; they achieved the complete conversion of glucose and reported the gluconic acid selectivity of 98% after 2 h at room temperature. Moreover, the authors reported that the 10.9 nm Pd/C particles presented the highest TOF (0.24 s^‒1, Table 2, #3). Characterization results indicated that the Pd NPs were anchored on the surface of the carbon support via O-containing functional groups. The catalyst presented good stability after four cycles and the Pd NPs exhibited only minor aggregation. Liu et al. [89] also compared different bases and determined that the nature of the base affected product selectivity significantly. The gluconic acid yield in the presence of the analyzed bases decreased as follows: KHCO₃ > Na₂CO₃ > NaOH > K₂HPO₄ > pyridine. Pyridine possesses lone electron pairs, which could block the active sites of the Pd/C catalyst. Pd leaching, which was caused by the coordination of gluconic acid with Pd active sites, was observed during the first two cycles [23].

Liu et al. [89] carried out kinetic studies over the Pd/C catalyst. They determined that the initial reaction rate was proportional to the concentration of glucose at concentrations below 0.18 μmol L^‒1. Hence the reaction order should be 1 and the absorption of glucose on the active sites was the rate-determining step. When the concentration of glucose reached 0.2 μmol L^‒1, the active sites of the Pd/C catalysts were saturated with glucose molecules. The formation rate of gluconic acid was lower than the consumption rate of glucose, which suggested that the calculated glucose conversion was inaccurate. Pd NPs were grown on six carbon materials and the obtained catalysts were tested for the electrocatalytic oxidation of glucose in alkaline medium [15]. It was reported that the performance of Pt immobilized on carboxylated multi-walled carbon nanotubes (MWCNTs) was better than that of Pt immobilized on amine- and hydroxyl-modified MWCNTs owing to the carboxy functionalities and uniform distribution of Pd on the surface of the catalyst.

Liang et al. [90] reported the scalable synthesis of Al₂O₃-supported Pd NPs via atomic layer deposition (ALD) (Table 2, #4). The size of the Pd particles could be tuned using the Pd precursor content and the duration of the ALD coating process. They reported that the activity of the Pd/Al₂O₃ catalyst was comparable to that of the commercial Pd catalyst and attributed that to the strong metal-support interaction and small particle size [88]. In addition, the Pd/Al₂O₃ catalyst presented negligible Pd leaching during the vigorously stirred reaction. Hence, the ALD technique appeared to present significant advantages for the preparation of noble metal NPs, such as precise particle size control, environmental friendliness, and cost effectiveness. Furthermore, the organic component of the Pd precursor could prevent the deactivation of the catalyst [90]. Doluda et al. [91] anchored Pd NPs on a typical nanoporous hypercrosslinked polystyrene polymer. The glucose conversion and gluconic acid selectivity of Pd/PdO NPs 1.7 nm in size were 93.6% and 99.6%, respectively, when NaHCO₃ was used as the alkalizing agent. This was ascribed to the small Pd/PdO NPs presenting low activation energy for the oxidation of glucose to gluconic acid. Moreover, this catalyst was stable for at least five cycles.

In addition to Pd-based catalysts, Pt-based ones present high efficiency for the oxidation of glucose. To better understand the effects of active species on the catalytic performance of Pt- and Pd-based catalysts, Pd/C (TOF of 1.3 s^‒1, Table 2, #5) and Pt/C (TOF of 0.05 s^‒1, Table 2, #6) were synthesized using the same method and were subsequently tested for glucose oxidation reactions under the same conditions [24]. The activity and selectivity of the Pt/C catalyst were lower than those of Pd/C. Delidovich et al. [24] noticed the negative effect of the particle size on the performance of the Pd/C catalyst when the particles were smaller than 3 nm, which was attributed to the oxidative deactivation of the surface Pt particles. The TOF values of the Pd/C catalysts increased as the particle size increased from 3 to 6 nm owing to the delay in the deactivation of the catalyst. The oxidative deactivation was diminished when the reaction was conducted under oxygen-diffusion control [24]. XPS and TEM analyses revealed the absence of any likeness of the relationship between the catalytic activity, metal dispersion, and electronic state of the Pt/C and Pd/C catalysts.

Lee et al. [92] investigated the oxidation of glucose using activated carbon-, SiO₂-, and Al₂O₃-supported Pt catalysts. They determined that the effectiveness of the catalysts for the production of gluconic acid decreased as follows: Pt/SiO₂ > Pt/Al₂O₃ > Pt/C (TOF of 0.8 × 10^‒2 s^‒1, Table 2, #7). Tathod et al. [93] immobilized Pt on γ-Al₂O₃ (acidic support) and hydrotalcite (HT, basic support that featured OH^‒ and HCO₃^‒ ions and double-layered structure) using the impregnation method and evaluated the prepared catalyst for the oxidation of glucose in the absence of bases. The gluconic acid yields of 76%, 82%, and 56% were achieved over the Pt/HT (TOF of 0.25 s^‒1, Table 2, #8), Pt/Al₂O₃ + Na₂CO₃ (TOF of 5.28 s^‒1, Table 2, #9), and Pt/Al₂O₃ + HT (TOF of 5.49 s^‒1, Table 2, #10) systems, respectively. The results indicated that HT, which is a basic support, could generate and stabilize more electron-rich and highly dispersed metallic particles. Koklin et al. [94] investigated the oxidation of glucose in a flow reactor over a Pt/Sibunit catalyst in the presence of H₂, He, or air (Table 2, #11). The highly selective conversion of glucose into gluconic acid could be performed at low temperature depending on the reaction atmosphere in the presence of H₂ and air, respectively.

Pt-based photocatalysts have been reported to be valuable, environmentally friendly, and efficient catalysts for the oxidation of glucose to gluconic acid. Bellardita et al. [9] immobilized Pt on TiO₂ and the obtained catalyst was tested for the photocatalytic oxidation of glucose under aerated and nonaerated conditions. They determined that product distribution depended on the physical and structural properties of the material, such as the crystalline phase, number of the Lewis acid sites, and surface hydroxylation. Pt supported on electrospun carbon fibers was also reported to be an effective electrochemical catalysts for the oxidation of glucose [13]. Pt-CF39 presented higher catalytic activity, sensitivity, and detection limit than other Pt-CF catalysts owing to its high electrochemically real surface area, small diameter, and high curvature.

Supports and synthesis methods present significant effects on the catalytic activity of Pt- and Pd-based catalysts for the oxidation of glucose. Particle size should be preferably controlled in the range of 3–7 nm, to avoid the rapid deactivation of the catalyst and NP aggregation. The types of ligands and the method used to remove the ligands have also been claimed to be important factors that could affect the catalytic performance of Pt- and Pd-based catalysts. A compromise between the adsorption of glucose and oxygen coverage should be considered when synthesizing these catalysts.

Bimetallic catalysts have been reported to be more active and selective for oxidation reactions than their monometallic counterparts. This has been ascribed to the synergistic effects induced by the geometrical or electronic interactions between two metals [95]. Various researchers have reported the use of bimetallic catalysts for the oxidation of glucose. The presence of synergistic effects between the two metal species, rather than their physical mixture, has been observed and demonstrated using various characterization techniques.

Shi et al. [96] synthesized Pt-Cu/TiO₂ bimetallic catalysts that featured round Pt-Cu nanoclusters (2.8 nm). The prepared catalyst facilitated the oxidation of glucose under basic conditions, and its conversion and gluconic acid selectivity were 100% and 29%, respectively (Table 2, #12). The authors proposed that strong metal-support interactions existed between Pt and TiO₂. Jin et al. [26, 97] synthesized TiO₂-supported bimetallic PtCu and tested it for the oxidation of glucose as a green approach for the synthesis of gluconic acid that did not involve the use of bleach enhancers or mineral acids. The authors used a one-pot synthesis method and generated smaller metal particles (4–5 nm, Fig. 10) than those obtained using traditional two-step methods (10 nm). TiO₂-supported Pt₁Cu₃ alloy presented extraordinary catalytic activity (TOF of 0.98 s^‒1, Table 2, #13) which was much higher than those of monometallic Pt and Cu catalysts (TOF of 0–0.15 s^‒1) at 45 ℃ and 0.1 MPa O₂ in the presence of NaOH. They also reported that the glucose conversion and gluconic acid selectivity of the catalysts increased as their Cu content increased. Jin et al. [97] later analyzed the use of TiO₂-supported PtPd alloy catalyst for the oxidation of glucose, and reported that its glucose conversion and gluconic acid selectivity were 100% and 69%, respectively (Table 2, #14). The PtPd/TiO₂ catalyst exhibited synergistic activity compared to the monometallic Pt and Pd catalysts. The structure of the alloy was confirmed using TEM, SEM, and ultraviolet-visible (UV-Vis) analyses. Jin et al. [97] reported that glucose inhibited the oxidation of gluconic acid, which was further oxidized only after the entire amount of glucose was consumed.

Cao et al. [98] prepared Au-Pd/TiO₂ (TOF of 0.028 s^‒1) and Au-Pt/TiO₂ (TOF of 0.048 s^‒1, Table 2, #15) catalysts using the sol immobilization method. These catalysts displayed good performance for the oxidation of glucose under base-free conditions. The authors concluded that the interaction between the PVA ligand and substrate affected the performance of the catalysts during the reaction. This has also been proposed by Pratti et al. [99]. In addition, Pratti et al. [99] noticed that the use of a second metal could decrease the surface Au content, and thus, could increase the activity of the catalysts. Derrien et al. [100] prepared Au-Pt and Au-Pd catalysts using different methods, e.g. incipient-wetness impregnation and deposition-precipitation. The performance of the Au-Pt catalysts was superior to that of Au-Pd catalysts under the same operating conditions. The glucose conversion and gluconic acid selectivity of the Au₁-Pt₁/ZrO₂ catalyst at 100 ℃, and 40 bar air were 100% and 30%, respectively. The long-term stability of the catalyst was tested in a continuous reactor. Xin et al. [101] prepared Au-Pd/Al₂O₃ catalysts using a bio-reductive approach and evaluated their performances for the oxidation of glucose with H₂O₂ at atmospheric pressure. The Au_0.8-Pd_0.2/Al₂O₃ catalyst exhibited the TOF of 8.51 s^‒1 at 323 K when H₂O₂ was used as the oxidant (Table 2, #16). The catalytic activity was highly related to the Au/Pd ratio and decreased as follows: Au_0.8Pd_0.2/Al₂O₃ > Au_0.5Pd_0.5/Al₂O₃ > Au_0.2Pd_0.8/Al₂O₃ > Au/Al₂O₃ > Pd/Al₂O₃. The diffuse reflectance UV–Vis spectroscopy analysis revealed strong mutual interactions between the Au and Pd NPs. The TEM images of the catalysts displayed the changes in particle size and demonstrated the presence of Au-Pd bimetallic NPs in the structure of the catalysts [102]. Recycling tests indicated that the catalytic activity was not significantly decreased after 17 runs.

Witonska et al. [103] introduced Te to SiO₂-supported Pd catalysts and evaluated their performance for the oxidation of glucose to gluconic acid under mild conditions. Bimetallic 5%Pd-5%Te/SiO₂ (TOF of 48.06 s^‒1, Table 2, #17) and 5%Pd-5%Bi/SiO₂ (TOF of 34.2 s^‒1, Table 2, #18) catalysts were more active for the oxidation of glucose than Pd/SiO₂ (TOF of 0.79 s^‒1, Table 2, #19) and Pd/Al₂O₃ (TOF of 0.16 s^‒1) owing to the presence of the PdTe intermetallic phase on the surface of the support as illustrated in the X-ray diffraction and XPS profiles of the catalysts. The formation of the intermetallic phase with firm connection between Pd and Te led to changes in the binding energy between the adsorbed particle and active metal center, and thus, led to the improvement of the catalytic activity and stability.

It could be challenging to elucidate the factors that affect catalytic performance of bimetallic catalysts for the oxidation of glucose. However, the presence of synergistic effects between metals and the metal-support interactions, rather than the physical mixing of the metals, are necessary and important for highly active bimetallic catalysts.

In addition to Au, Pt, and Pt, other metals, such as Cu [11, 16, 27, 104], Ce [104], Ti [7, 8, 10, 16, 105-107], Mn [17, 108], Fe [28, 109], Ag [12, 110], and Cr [10] have been reported as being active for the oxidation of glucose to gluconic acid.

Amaniampong et al. [27] investigated the effect of the surface lattice O of CuO on its performance for the oxidation of glucose. They determined that the surface lattice O activated the formyl C–H bonds and oxidized glucose to gluconic acid. The gluconic acid yield of 86.8% was achieved at 150 ℃ after 30 min. However, the used CuO catalyst was deactivated owing to the loss of O from its surface. Later, Amaniampong et al. [104] synthesized porous CuO-CeO₂ catalysts with nanosphere shape for the conversion of glucose to gluconic acid in the absence of O₂. The authors further confirmed that function of the surface lattice O was to catalyze the aldehyde functional group of glucose to gluconic acid. Conversely, the Lewis acid sites catalyzed the isomerization of glucose to fructose. Cu₂O supported on helical TiO₂ nanotubes was also reported as electrode for the oxidation of glucose [16]. Cu₂O was electrochemically deposited on the TiO₂ nanotubes in a three-electrode cell (Fig. 11). The obtained catalyst exhibited high sensitivity (14.6 μA cm^‒2 mM^‒1) and great reproducibility [16]. However, the effect of the morphology of Cu₂O supported on TiO₂ nanotubes for the oxidation of glucose to gluconic acid should be further investigated.

Photocatalytic oxidation is an effective method for converting glucose into value-added products [7, 8, 105, 107]. Fabrao et al. [107] achieved the glucose conversion of 78% during the photoelectrocatalytic oxidation of glucose over Ti/TiO₂ under the potential of 1.5 V and under UV irradiation. The authors observed that the target reaction mainly occurred in the photogenerated cavities on the surface of the semiconductor rather than at the –OH groups on its surface. The visible light-assisted photooxidation of glucose over TiO₂ was also studied [8]. The glucose conversion of 42% and low gluconic acid selectivity of 7% were achieved, which were lower than those obtained during the UV light-assisted reaction. Davia et al. [12] determined that the ligand (glucose)-to-metal (Ti) charge transfer effect (Fig. 12) played a major role during the visible light driven reaction. A Ti-supported Ag catalyst was also studied for the visible light-driven photocatalytic oxidation of glucose, and the glucose conversion and gluconic acid selectivity of 6% and 18%, respectively, were achieved [12]. These results indicated that the presence of Ag promoted the mineralization reaction and the formation of CO₂ over Ag/TiO₂ under UV light irradiation. The authors also concluded that the surface coverage of the catalyst determined its catalytic activity. Ag and Au NPs were reported as active metals that could be used to modify glassy carbon electrodes and were analyzed for the electrooxidation of glucose in alkaline media [14]. The reaction rate was controlled by the Ag content of the electrodes.

Colmenares et al. [105, 106] synthesized zeolite-supported TiO₂ using the sol-gel method under ultrasound irradiation. This catalyst reached the glucose conversion and carboxylic acid selectivity of 49.2% and 29.5%, respectively, at 30 ℃ when acetonitrile was used as the solvent; these values were higher than those obtained when TiO₂/SiO₂ and unsupported TiO₂ were used as catalysts. TiO₂ could be used to anchor heteropolyacids (HPAs) to obtain HPA/TiO₂ catalysts for the photocatalytic conversion of glucose (glucose conversion of 85% and carboxylic acid selectivity of 34%) [7]. The performance of the HPA/TiO₂ catalyst exceeded that of the TiO₂ powder owing to the HPAs changing the acidity of the TiO₂ surface. This was confirmed by using the HPAs as catalysts, which could only induce the isomerization of glucose.

Mn-based catalysts have also been reported to present good catalytic activity for the oxidation of glucose [17]. A MnO₂-loaded tubular Ti electrode was prepared using the sol-gel method and was used as the anode in an electrocatalytic reactor. The glucose conversion, gluconic acid selectivity, and current density of 98%, 43%, and 4 mA cm⁻², respectively, were reached within 19 min at 30 ℃. When the current density increased, the catalytic activity increased; however, the gluconic acid selectivity decreased to 15% as more glucaric acid formed. Sen et al. [108] supported MnO₂ on 2D gum acacia sheets and evaluated the performance of the obtained catalyst for the oxidation of glucose. The kinetics and mechanism of the reaction were studied and the results revealed that this was a first order reaction with respect to MnO₂ and complex order reaction with respect to glucose.

Zhang et al. [28] investigated the oxidation of glucose using concentrated FeCl₃ solution (40 wt%) as the oxidant and achieved the gluconic acid yield of 52.3% within 2 h at 110 ℃. The concentration of glucose affected the gluconic acid selectivity significantly. The highest acetic acid yield was reached when the concentration of glucose was the lowest [28]. Rinsant et al. [109] reported the oxidation of glucose over FeSO₄ catalyst when H₂O₂ was used as the oxidant, and the glucose conversion and gluconic acid selectivity were determined to be 76% and 94.7%, respectively. Low frequency ultrasounds could improve the reaction rate significantly (glucose conversion and gluconic acid selectivity were determined to be 99% and 98%, respectively) by generating more HO• radicals via a sono-Fenton process [109].

Photocatalysts and electrocatalysts were briefly introduced in this section. TiO₂ and CuO have been the most studied substances for these processes. Even though the catalytic activity of TiO₂ and CuO was lower than that of Au- and Pd-based catalysts, these processes brought new insight into the use of greener methods for the production of gluconic acid and conversion of biomass.

The progress and potential significance of supported metal catalysts for the oxidation of glucose to gluconic acid in the presence and absence of bases have been critically reviewed in this paper. This study summarizes and discusses the synthesis and characterization of catalysts, their structure-function relationships, and also their durability. The reaction mechanisms in the presence and absence of bases were discussed separately using experimental and DFT calculation results. Of the various catalyst candidates used for the oxidation of glucose to gluconic acid, carbon-supported Au has been recognized as being the most promising. In addition, bimetallic catalysts presented great potential owing to the affordability of the metal additives. The catalytic performance of supported catalysts could be affected by several factors, e.g., the size of the Au NPs, nature of the support, synthesis method, types of ligands, and presence of bases. Of these factors, the size of the Au NPs has been widely accepted as being the most significant one, because many researchers reported that catalytic activity was inversely proportional to the size of the Au NPs. The nature of the support and synthesis method determined the structure and particle size of the catalysts, and thus affected their catalytic performance indirectly. Based on the thorough review and summarization of the recent progress, we concluded that the following concerns should still be addressed.

(1) Base-free processes are highly favorable for the direct, selective, and one-pot oxidation of glucose to gluconic acid, because the addition of extra base and use of strong acids to react with the salt to obtain gluconic acid could be avoided. Base-free processes could prevent metal NP sintering and leaching. Moreover, these processes could decrease the post-treatment cost of the products, environmental pollution, and equipment corrosion. Hence, the development of catalysts that are highly active and stable under acidic conditions is an attractive strategy.

(2) H₂O₂ could be used an oxidant instead of O₂ to avoid the effect of the rate-limiting mass transfer on the oxidation process, which is attributed to the low solubility of O₂ in water. Several studies revealed that higher gluconic acid yields were obtained when H₂O₂ was used instead of O₂ as the oxidant. Consequently, it would be desirable to further investigate the glucose oxidation mechanism over the H₂O₂-Au system.

(3) Catalyst deactivation could be caused by the following: hydrothermal particle sintering, active site poisoning or adsorption of organic species, and metal leaching. It has been reported that decreasing particle density on the catalyst surface could minimize particle sintering. During the synthesis of catalysts, the surfactants should be partially removed. The remaining surfactant could protect the active metal against sintering and poisoning by the organic species in the reaction mixture. The development of effective methods for controlling particle morphology to obtain highly active and stable catalyst is highly desirable.

(4) The mechanisms of the glucose oxidation reaction were different in the presence and absence of bases over different types of catalyst systems. Hence, further exploration via both experimental and theoretical approaches is still required to elucidate the reaction mechanism comprehensively.

(5) Kinetic studies have guided catalyst design and process development by clarifying the relationship between catalyst morphology and catalytic activity. The rate-limiting factors were different at different glucose/metal ratios. Therefore, kinetic studies would be necessary for any newly developed systems.

(6) The reaction conditions significantly affect product distribution. At high temperature, glucose is converted into more than 20 different compounds. The oxidation of glucose involves several consecutive-parallel reactions, including those accompanied by the cleavage of the C–C bonds. Glucose reforming from H₂ and CO₂ takes place during the catalytic oxidation of glucose in aqueous solution at 200 ℃. The interaction between glucose and the H₂ formed during the oxidation of glucose resulted in the formation of sorbitol even under inert atmosphere.

(7) Investigations into the replacement of noble metals with inexpensive metals should be conducted. The high prices of noble metals limit their large-scale industrial use. Hence, the development of catalysts for the oxidation of glucose that would be based on inexpensive metals and would present good activity and gluconic acid selectivity is highly desired.

W. Y. drafted the manuscript, D. Z., Y. S., Z. Z., Y. D., and Y. D. conducted the literature survey, Y. L. and M. L. organized the figures, Y. Z., J. S., and X. J. provided comments.