Synthesis of Glyceric Acid by Mixed-Metal Oxide-Supported AuPt Alloy Catalyst in Mild Conditions

Abstract

Thermal valorization of surplus biomass-derived feedstocks such as glycerol into high-value chemicals represents a sustainable strategy for biomass utilization and decarbonization of chemical manufacturing. However, conventional glycerol conversion processes are often restricted to low-value C1 products owing to rapid C–C bond cleavage during thermo-oxidation. Herein, we report highly efficient Au-Pt bimetallic alloy catalysts supported on mixed-oxide catalysts that enable the selective oxidation of glycerol under ambient conditions in the absence of a base. The synergistic interaction between Au and Pt promotes preferential oxidation of the terminal hydroxyl groups while preserving the C3 backbone, thereby affording the desirable C3 product, glyceric acid. The single-factor experiments and response surface analysis demonstrated that the Au-Pt bimetallic alloy catalysts supported on the mixed oxide MgO-Al₂O₃ exhibited a glycerol conversion of up to 82.0% and a glyceric acid selectivity of 62.1% under favorable reaction conditions. Kinetic studies further indicated that the activation energy of this catalyst in the reaction system is 32.7 kJ/mol.

1. Introduction

Biodiesel has emerged as a renewable alternative to fossil fuels [1], with glycerol (GLY) generated as a major byproduct during its production. The effective valorization of surplus GLY is therefore critical for improving the economic viability of biodiesel processes [2,3,4,5]. Among various derivatives, glyceric acid (GLA) is particularly attractive due to its substantially higher market value and wide range of applications in cosmetics, pharmaceuticals, and the food industry [6,7]. Traditionally, GLA has been produced by biological fermentation, yet this route suffers from long reaction cycles, complex separation steps, and low conversion efficiency. As a result, heterogeneous catalysis has drawn increasing attention as a more efficient and scalable strategy for GLY oxidation [8,9].

Supported noble metal nanoparticles (e.g., Au, Pd, Pt) are widely employed in biomass oxidation [10,11,12]; however, they often face critical challenges, including rapid deactivation and the requirement of alkaline conditions to achieve high activity and selectivity [13]. Mechanistic studies suggest that GLY oxidation on metal surfaces can proceed via two main pathways. Initially, the primary hydroxyl group is oxidized to form glyceraldehyde, which can subsequently be converted to GLA or undergo oxidative C–C bond cleavage to yield secondary products such as glycolic and formic acids. The latter pathway is more favorable under alkaline conditions [14,15]. These mechanistic insights guide the rational design of catalysts capable of operating under base-free conditions.

Pt-based catalysts have demonstrated potential for base-free GLY oxidation, yet their performance remains limited. For instance, Pt/C exhibits only 19.4% GLA selectivity at 37% conversion under acidic conditions, whereas Pt/MWCNT achieves 70.1% conversion with 69.8% selectivity at 333 K [16]. These findings highlight the importance of support effects, particularly acid–base properties, which strongly influence the activity and stability of GLY oxidation catalysts. Mixed oxides such as Mg-Al hydrotalcite-derived materials provide abundant basic sites and high thermal stability, and further modification (e.g., incorporation of Ce or Mn) has been shown to improve redox properties and catalytic activity in related oxidation reactions [17,18]. In addition to support properties, metal alloying is also a key factor governing catalytic activity and selectivity. Alloying noble metals is particularly effective in enhancing stability and mitigating deactivation [19,20]. Cherni et al. demonstrated high activity in the GLY oxidation reaction using N-doped titanium dioxide-based catalyst supports. The highest conversion (92% after 6 h) and the highest selectivity to glyceric acid (79.9%) were achieved using AuPt [21]. Ke et al. prepared a Au-Pt/Mn_xO_yC_z catalyst supported on Mn-modified carbon, exhibiting a complete glycerol conversion and a glyceric acid selectivity of 57.3% after a 2 h reaction [22].

Motivated by these insights, the present work explores Au–Pt bimetallic alloy catalysts supported on mixed oxides (MgO-Al₂O₃ and ZnO-Al₂O₃). This study aims to elucidate the role of support composition and metal ratio on GLY oxidation under mild, base-free conditions and to develop an efficient catalyst for the selective production of GLA. Moreover, optimal reaction conditions were determined through a response surface design. Subsequently, kinetic analysis was conducted to investigate the influence of reaction parameters on oxidation.

2. Results

2.1. Texture and Morphology of Catalysts

2.1.1. XRD Analysis

The XRD patterns of all synthesized samples are shown in Figure 1a. For the MgO-Al₂O₃-supported catalysts, diffraction peaks are observed at 2θ = 34.7°, 39.9°, and 61.3°, which can be attributed to the characteristic reflections of γ-Al₂O₃ (JCPDS No. 04-0877) [23,24]. Additional peaks at 2θ = 42.9° and 62.3° correspond to the typical diffraction features of MgO (JCPDS No. 45-0946) [25]. In the case of ZnO-Al₂O₃-supported samples, similar γ-Al₂O₃-related peaks are detected, along with additional reflections at 2θ = 31.73°, 34.36°, 36.21°, and 62.3°, matching well with the characteristic peaks of ZnO (JCPDS No. 80-0074). These results confirm the successful formation of MgAl and ZnAl mixed oxides. Notably, no distinct diffraction peaks corresponding to Au or Pt species (JCPDS No. 87-0647) are detected in any of the supported catalysts through the wet chemical reduction method, suggesting that both Au and Pt are highly dispersed on the mixed-oxide supports, an observation consistent with the TEM results. However, for the ⁰Au_1.5Pt_1.5/MgO-Al₂O₃ catalyst reduced under hydrogen, in addition to the support-related peaks, clear diffraction signals appear at 2θ = 38.2° and 44.4°, which correspond to metallic Au (JCPDS No. 04-0784). This indicates that hydrogen reduction leads to the formation of larger Au particles compared to those generated via the conventional wet chemical reduction method.

2.1.2. BET Analysis

The N₂ adsorption–desorption isotherms of the catalysts are presented in Figure 1b and Figure S1, and the corresponding textural properties are summarized in Table S1. The pristine MgO-Al₂O₃ support exhibited a typical type IV isotherm with a H4-type hysteresis loop [26,27,28], characteristic of mesoporous materials. It possessed a BET surface area of 107.9 m²·g⁻¹, a pore volume of 0.21 cm³·g⁻¹, and an average pore size of 7.6 nm. Upon loading with Au, Pt, or Au-Pt nanoparticles, both the surface area and pore structure undergo remarkable changes. Specifically, the BET surface areas of Au₃/MgO-Al₂O₃, Pt₃/MgO-Al₂O₃, Au₁Pt₂/MgO-Al₂O₃, Au_1.5Pt_1.5/MgO-Al₂O₃, and Au₂Pt₁/MgO-Al₂O₃ decreased to 9.9, 25.6, 69.3, 87.2, and 13.5 m²·g⁻¹, respectively. This decrease could be attributed to partial coverage of the support surface and the blockage of mesopores by metal nanoparticles. Notably, Au_1.5Pt_1.5/MgO-Al₂O₃ retained the highest surface area (87.2 m²·g⁻¹) and pore volume (0.33 cm³·g⁻¹) among the supported catalysts, indicating that the formation of AuPt alloy nanoparticles enhances the dispersion and stability of the metal phase. The bimetallic Au-Pt catalysts, particularly Au_1.5Pt_1.5/MgO-Al₂O₃, retained significantly higher BET surface areas and pore volumes compared to their monometallic counterparts (Au₃/MgO-Al₂O₃ and Pt₃/MgO-Al₂O₃). This confirms that the formation of Au-Pt alloy nanoparticles effectively inhibits the sintering of the metal phase and promotes its high dispersion on the support. The general increase in the average pore size suggests that the metal nanoparticles primarily block the smaller mesopores of the support. The Au_1.5Pt_1.5/MgO-Al₂O₃ sample exhibited a similar isotherm shape to MgO-Al₂O₃, with a moderate surface area of 54.5 m²·g⁻¹ and an enlarged pore size of 17.6 nm, reflecting the influence of the MgO-Al₂O₃ support on textural properties. For the ⁰Au_1.5Pt_1.5/MgO-Al₂O₃ catalyst reduced in H₂, the BET surface area dramatically decreased to 5.5 m²·g⁻¹ due to nanoparticle sintering and severe pore collapse during prolonged high-temperature reduction.

2.1.3. XPS Analysis

The surface chemical states of Au and Pt in the Au_1.5Pt_1.5/MgO-Al₂O₃ catalyst were analyzed by XPS before and after the glycerol oxidation reaction (Figure 1c,d and

Table 1. Loading and surface elemental composition of catalysts.

	Binding Energy (eV)								Ratio a		Au Loading b	Pt Loading b
	Au0		Auδ+		Pt0		Pt2+		$\frac{{\mathbf{Pt}}^0}{{\mathbf{Pt}}^0+{\mathbf{Pt}}^{2+}}$	$\frac{{\mathbf{Au}}^0}{{\mathbf{Au}}^0+{\mathbf{Au}}^{\mathsf{\delta }+}}$
	4f7/2	4f5/2	4f7/2	4f5/2	4f7/2	4f5/2	4f7/2	4f5/2
Fresh Au1.5Pt1.5/MgO-Al2O3	83.6	87.1	85.4	88.9	71.1	74.5	72.8	76.2	0.84	0.87	1.37	1.41
Spent Au1.5Pt1.5/MgO-Al2O3	83.3	86.8	85.1	88.6	71.4	74.8	73.1	76.5	0.81	0.86	1.35	1.38

^a Detected via XPS (model Escalab 250Xi, Thermo Fisher Scientific, MA, USA). ^b Detected via ICP analysis (model Optima 7300DV, PerkinElmer Inc., MA, USA.

). The spectra reveal the coexistence of metallic (Au⁰, Pt⁰) and oxidized (Au^δ+, Pt²⁺) species. For Au 4f, the binding energies at approximately 83.6 eV (4f_7/2) and 87.1 eV (4f_5/2) correspond to Au⁰, while the peaks at 85.4 eV and 88.9 eV are assigned to Au^δ+ species [29]. In the Pt 4f region, the peaks at 71.1 eV and 74.5 eV are attributed to Pt⁰ 4f_7/2 and Pt⁰ 4f_5/2, whereas the peaks at 72.8 eV and 76.2 eV are assigned to Pt²⁺ 4f_5/2 and 4f_7/2, respectively. Compared with the binding energy values reported for monometallic Au and Pt, all Au and Pt peaks in the Au-Pt alloy exhibit distinct shifts. This shift can be attributed to electronic transfer during alloy formation, whereby Au tends to gain electrons while Pt loses electrons. After the glycerol oxidation reaction, the relative proportions of Au⁰ and Pt⁰ decreased slightly, suggesting that partial oxidation of the surface metals occurred during catalysis [30,31]. These observations confirm the successful formation of Au-Pt alloy nanoparticles and reveal strong electronic interactions between Au and Pt. Such electron transfer not only modifies the surface electronic structure but also contributes to the enhanced catalytic activity of the alloy in base-free glycerol oxidation. Moreover, the ICP test results before and after the reaction for the Au_1.5Pt_1.5/MgO-Al₂O₃ catalyst indicate that the loading amounts of Au and Pt slightly decreased post-reaction (Table 1), suggesting that the MgO-Al₂O₃ support exhibits strong anchoring effects on the metallic nanoparticles.

2.1.4. SEM-EDS and TEM Analysis

To gain detailed insights into the surface morphology, particle dispersion, and lattice structure, the catalysts were characterized by SEM, TEM, and HRTEM (Figure 2 and Figure S2). The SEM image of the Au_1.5Pt_1.5/MgO-Al₂O₃ catalyst reveals a well-defined nanoflower-like structure composed of interconnected thin nanosheets, which are densely and orderly stacked, forming a porous framework. Such morphology is characteristic of mixed oxides derived from layered double hydroxides after thermal decomposition (Figure 2a). EDS analysis confirms that Au and Pt nanoparticles are uniformly distributed across the catalyst’s surface (Figure S2). TEM images of Au₃/MgO-Al₂O₃, Pt₃/MgO-Al₂O₃, Au₁Pt₂/MgO-Al₂O₃, Au₂Pt₁/MgO-Al₂O₃, and Au_1.5Pt_1.5/MgO-Al₂O₃ (Figure 2b and Figure S3) reveal that the metal nanoparticles are irregularly shaped and mainly composed of smaller clusters with sizes ranging from 3 to 6 nm. The corresponding average particle sizes are calculated to be 4.49, 3.26, 4.67, 3.64, and 4.52 nm, respectively, based on particle size distribution analysis. HRTEM images (Figure 2c) display clear lattice fringes with interplanar spacings of 2.26–2.28 Å, corresponding to the (111) planes of a face-centered cubic (FCC) structure. These values lie between those of pure Au (2.35 Å) and pure Pt (2.26 Å), indicating the formation of a homogeneously alloyed AuPt phase rather than separate Au and Pt nanoparticles. Furthermore, the SAED pattern (Figure 2d) exhibits multiple concentric diffraction rings corresponding to the (111), (200), (220), and (311) planes of an FCC lattice, consistent with a polycrystalline AuPt alloy. The absence of ring splitting or extra diffraction features supports the presence of a single-phase bimetallic structure. These results collectively confirm the successful synthesis of well-dispersed, crystalline Au-Pt alloy nanoparticles supported on MgO–Al₂O₃ [32].

2.1.5. Support Basicity

Potentiometric titration was employed to determined the support basicity. Figure S4 presents the titration curves for the pristine MgO-Al₂O₃ and ZnO-Al₂O₃ supports. The MgO-Al₂O₃ support exhibited a higher base strength (E_max = 448.01 mV) than ZnO-Al₂O₃ (E_max = 380.21 mV). Furthermore, the density of basic sites was greater for MgO-Al₂O₃ (0.58 mmol/g) compared to ZnO-Al₂O₃ (0.51 mmol/g) [33].

2.2. Catalytic Screening Experiments

Catalytic performance tests were conducted on seven different supported catalysts, and the results are summarized in

Table 2. Catalytic screening experiments of various catalysts in base-free condition.

Catalysts	GLY Conversion (%)	Product Selectivity (%)
		GLA	DHA	GLD	TTA	GCA	OA	AA	COx
Au3/MgO-Al2O3	14.0 ± 1.3	15.1 ± 1.7	56.5 ± 2.7	0.1 ± 0.01	0.4 ± 0.01	6.2 ± 1.0	2.3 ± 0.1	11.1 ± 1.4	8.3 ± 1.1
Pt3/MgO-Al2O3	43.0 ± 1.0	55.1 ± 1.4	11.6 ± 1.5	0.1 ± 0.03	1.6 ± 0.07	1.5 ± 0.2	2.2 ± 1.0	15.8 ± 1.7	12.1 ± 1.0
Au1Pt2/MgO-Al2O3	46.0 ± 1.2	50.1 ± 1.6	11.2 ± 1.8	0.1 ± 0.01	2.3 ± 0.3	1.2 ± 0.01	5.7 ± 1.0	19.8 ± 1.5	9.6 ± 1.0
Au1.5Pt1.5/MgO-Al2O3	50.0 ± 2.0	61.8 ± 2.3	19.0 ± 2.7	0.2 ± 0.02	1.0 ± 0.05	1.4 ± 0.04	1.5 ± 0.2	7.7 ± 1.2	7.4 ± 1.1
Au2Pt1/MgO-Al2O3	34.0 ± 2.2	40.6 ± 1.5	31.9 ± 1.6	0.2 ± 0.03	1.9 ± 0.1	1.4 ± 0.05	4.5 ± 1.0	8.5 ± 1.0	11.0 ± 2.0
Au1.5Pt1.5/ZnO-Al2O3	14.2 ± 1.3	48.7 ± 1.1	28.3 ± 1.9	0.1 ± 0.01	0.8 ± 0.04	6.2 ± 0.9	5.1 ± 1.20	0.5 ± 0.01	10.3 ± 1.0
0Au1.5Pt1.5/MgO-Al2O3	37.0 ± 1.9	33.0 ± 1.6	29.3 ± 2.1	0.2 ± 0.01	2.6 ± 0.3	2.1 ± 0.1	6.8 ± 1.1	20.0 ± 2.0	6.0 ± 0.8
MgO-Al2O3	3.9 ± 1.5	36.6 ± 1.9	19.8 ± 1.7	0.2 ± 0.03	0.4 ± 0.01	11.6 ± 1.1	1.0 ± 0.1	13.5 ± 1.2	16.9 ± 1.7
Reaction conditions: 30 mL 0.1 M GLY, 30 °C, reaction time = 8 h, GLY/(AuxPty) = 100 mol/mol, ${\mathrm{P}}_{{\mathrm{O}}_2}$ = 0.5 MPa, stirring speed = 500 rpm.

⁰Au_1.5Pt_1.5/MgO-Al₂O₃ was prepared using the H₂ reduction method; all other catalysts were prepared using the wet chemical reduction method.

. Under identical reaction conditions, all catalysts containing metal nanoparticles, including both monometallic Au or Pt and bimetallic Au-Pt alloys, exhibited significantly higher activity than the bare MgO-Al₂O₃ support. This indicates that the incorporation of noble metals plays a crucial role in promoting GLY oxidation to GLA under base-free conditions. A comparison between Au₃/MgO-Al₂O₃ and Pt₃/MgO-Al₂O₃ reveals GLY conversions of 14.0% and 43.0%, respectively, which highlights the superior catalytic activity of Pt-based catalysts over Au-based ones in base-free environments [34]. For the bimetallic catalysts with different Au/Pt mass ratios, all displayed higher activity than their monometallic counterparts. Notably, Au_1.5Pt_1.5/MgO-Al₂O₃ achieved the highest GLY conversion of 50.0%, demonstrating a pronounced synergistic effect between Au and Pt. Furthermore, when comparing catalysts prepared by different reduction methods, smaller bimetallic nanoparticles were found to correlate with markedly enhanced catalytic activity. In addition, the influence of support composition was also evident. Au_1.5Pt_1.5/MgO-Al₂O₃ exhibited higher activity than Au_1.5Pt_1.5/ZnO-Al₂O₃, which can be attributed to the stronger basicity of the MgO-Al₂O₃ support relative to ZnO-Al₂O₃.

Combined with the literature, our findings demonstrate that Au nanoparticles selectively oxidize the C3 primary hydroxyl group and effectively suppress C–C bond cleavage, whereas Pt excels at activating O₂ and promoting intermediate conversion [35]. The incorporation of Pt markedly accelerates the oxidation of glyceraldehyde to glyceric acid. This synergistic effect within the Au–Pt bimetallic system results in a substantially higher glyceric acid selectivity compared to its monometallic counterparts [19,35,36].

2.3. Optimization of Reaction Conditions

To investigate the influence of different reaction parameters on GLY oxidation, single-factor experiments were performed using the Au_1.5Pt_1.5/MgO-Al₂O₃ catalyst. The effects of reaction temperature, reaction time, catalyst dosage, and oxygen pressure on GLY conversion and product distribution were systematically examined. As shown in Figure 3a, GLY conversion increased continuously with rising temperature in the range of 20–50 °C. At 30 °C, a conversion of 70.3% and a GLA selectivity of 57.0% were achieved, accompanied by the formation of other byproducts. Since the highest GLA selectivity was obtained at 30 °C, this temperature was selected for further optimization. The influence of reaction time was then studied between 12 and 18 h (Figure 3b). GLY conversion increased with prolonged reaction time, but the improvement became marginal beyond 14 h. Meanwhile, GLA selectivity decreased from 62.1% to 47.3% with increasing reaction time. Therefore, 14 h was chosen as the optimal reaction time. The effect of catalyst dosage was subsequently evaluated by varying the GLY/(Au+Pt) ratio from 100 to 1000 mol·mol⁻¹ (Figure 3c). GLY conversion decreased sharply from 70.3% to 12.0% as the catalyst dosage decreased, while GLA selectivity also declined. This behavior indicates that the number of available active sites significantly influences catalytic performance. Finally, the effect of oxygen pressure was investigated (Figure 3d). GLY conversion increased from 49.0% to 74.0% as the O₂ pressure rose from 0.25 to 1.0 MPa. Interestingly, GLA selectivity improved when the pressure was below 0.5 MPa but declined at higher pressures. Throughout the reaction, carbon balance was maintained, with other C₃ and C₂ products forming alongside GLA. Among the byproducts, dihydroxyacetone (DHA) was the most abundant, further supporting the proposed mechanism of GLY oxidation under base-free conditions (Scheme S1).

2.4. Response Surface Analysis Experiment

2.4.1. RSM Experiments and Predicted Model

Reaction temperature ($X_1$), reaction time ($X_2$), GLY/(Au+Pt) molar ratio ($X_3$), and O₂ pressure ($X_4$) were identified as key factors influencing the conversion of GLY to GLA. Response surface methodology (RSM) based on a Box–Behnken design (BBD) was applied to optimize GLA yield [37,38,39]. The regression model obtained from the statistical design is expressed in coded units as follows:

$$\begin{array}{ll}\mathrm{Y}\left(\%\right)=20.10-& 0.64X_1+1.20X_2-12.28X_3+1.72X_4-0.08X_1{X}_2-2.25X_1{X}_3\\ & +0.10X_1{X}_4-0.80X_2{X}_3+0.18X_2{X}_4-2.72 X_3{X}_4-1.37X_1^2\\ & -2.00X_2^2+0.71X_3^2-3.18X_4^2\end{array}$$

(1)

In this equation, Y represents the GLA yield. Predicted yields calculated from the model were compared with experimental values, and the results demonstrated a good linear correlation (Figure S5), indicating that the BBD model adequately describes the relationship between GLA yield and the reaction parameters [40]. In the regression equation, a negative coefficient in the interaction term indicates an antagonistic effect, meaning that increasing these variables decreases the GLA yield [41,42,43]. Conversely, a positive coefficient reflects a synergistic effect. Among the studied variables, the GLY/(Au+Pt) molar ratio ($X_3$) exhibited the largest linear coefficient (12.28), suggesting that catalyst dosage had the most pronounced influence on GLA yield, followed by O₂ pressure ($X_4$).

Analysis of variance (ANOVA) confirmed the validity of the model. The regression model showed a p-value < 0.0001 and an F-value of 17.45, signifying strong statistical significance [39,43,44,45]. Furthermore, the quadratic term of $X_3$ exhibited a p-value < 0.05, indicating a significant effect on GLA yield. The coefficient of determination (R² = 0.9548), adjusted R² (0.9016), and predicted R² (0.8329) were in good agreement, with the difference between adjusted and predicted R² being within ±0.2, further confirming that the model can accurately describe the experimental data.

2.4.2. Response Surface Analysis

The interaction effects of the variables on GLA yield were analyzed using two- and three-dimensional response surface plots (Figure S6). Figure S6a,b illustrate the combined influence of reaction temperature (X₁) and time (X₂) when X₃ = 300 and X₄ = 0.5 MPa. The results show that GLA yield increases with rising temperature at moderate levels but decreases slightly at higher temperatures, likely due to partial deactivation of the Au_1.5Pt_1.5/MgO-Al₂O₃ catalyst. The elliptical contour indicates a significant interaction between temperature and time. Figure S6c,d show the interaction between temperature (X₁) and catalyst dosage (X₃) at X₂ = 14 h and X₄ = 0.5 MPa, where higher catalyst dosage leads to improved GLA yield. Figure S6e,f illustrate the effect of temperature (X₁) and O₂ pressure (X₄), showing that GLA yield first increases and then decreases as O₂ pressure rises. Figure S6g,h and Figure S6i,j present the interactions of time (X₂) with catalyst dosage (X₃) and O₂ pressure (X₄), respectively. Figure S6k,l show the combined effect of X₃ and X₄, which had the most significant influence on GLA yield among all two-variable interactions, as confirmed by a p-value < 0.05 (Table S2).

2.4.3. Verification of Response Surface Model

The purpose of the RSM analysis was to determine the optimal conditions for the selective oxidation of GLY to GLA. The model predicted the following optimal conditions—36 °C, 15 h, GLY/(Au+Pt) = 100 mol·mol⁻¹, and O₂ pressure = 0.68 MPa—under which the GLA yield was expected to reach 40.3%. Experimental validation under these conditions gave a GLY conversion of 72.0% and a GLA yield of 41.7%. The deviation between experimental and predicted values was less than 5% (Table S3), confirming that the BBD model is reliable for predicting the selective oxidation of GLY under base-free conditions. Compared with other studies (Table S4), the GLY conversion rate and GLA selectivity achieved under alkali-free conditions in this study hold practical significance.

2.5. Kinetic Study

A power-law kinetic model was employed to investigate the effects of reaction parameters on the catalytic oxidation of GLY to GLA. The Au_1.5Pt_1.5/MgO-Al₂O₃ catalyst, which exhibited the highest activity in the screening experiments, was selected for kinetic analysis. To ensure differential conditions, the reaction time was fixed at 2 h, and the initial reaction rate was measured at GLY conversions below 20%. The reaction rate can be expressed by the power-law equation (Equation (2)). The temperature dependence of the rate constant follows the Arrhenius law (Equation (3)). Substituting Equation (3) into Equation (2) and taking the natural logarithm yield the linearized form (Equation (4)). This linear form allows for the determination of kinetic parameters by multiple linear regression. Based on the initial reaction rate experimental data (Table S5), the fitting results are given in Equation (5), with a correlation coefficient of R² = 0.9605; the apparent activation energy was determined to be 32.7 kJ·mol⁻¹. In the range of 0.25 to 0.5 MPa, the reaction orders with respect to catalyst dosage and oxygen pressure were 0.43 and 0.28, respectively, indicating that the amount of catalyst used has a greater effect on glycerol oxidation than oxygen pressure. A significant transition in kinetic behavior was observed at higher O₂ pressures. When the O₂ pressure exceeds 0.5 MPa, the reaction rate shows no significant change, likely due to saturation of oxygen adsorption sites. Further increases in pressure effectively produce a zero-order dependence on O₂.

$${-r}_A={k\left(T\right)m}_{Cat}^a{\mathrm{P}}_{{\mathrm{O}}_2}^b$$

(2)

$$\mathrm{k}\left(\mathrm{T}\right)=\mathrm{A}exp(-E_a/\mathrm{R}\mathrm{T})$$

(3)

$$\ln \left({-r}_A\right)=\mathrm{l}\mathrm{n}\mathrm{A}+\left(-{\displaystyle \frac{E_a}{\mathrm{R}\mathrm{T}}}\right)+\mathrm{a}\mathrm{l}\mathrm{n}{m}_{Cat}+\mathrm{b}\mathrm{l}\mathrm{n}{\mathrm{P}}_{{\mathrm{O}}_2}$$

(4)

$$-r_A=533.6exp(-32719/\mathrm{R}\mathrm{T})m_{Cat}^{0.43}{\mathrm{P}}_{{\mathrm{O}}_2}^{0.28}$$

(5)

$m_{cat}$: the catalyst dosage; ${\mathrm{P}}_{{\mathrm{O}}_2}$: the oxygen partial pressure; a and b: the apparent reaction orders; A: the pre-exponential factor; $E_a$: the apparent activation energy; R: the universal gas constant; T: the absolute temperature.

Since the Au_1.5Pt_1.5/MgO-Al₂O₃ catalyst was prepared by sol-immobilization of Au-Pt species on the support, any mass transfer limitations would be expected to arise primarily from external diffusion. To evaluate this effect, the oxidation of GLY was conducted at different stirring speeds. When the stirring rate was increased to 700 rpm, the GLY conversion profiles overlapped at various reaction times, indicating the absence of external diffusion limitations. Therefore, a stirring speed of 500 rpm was considered sufficient to eliminate external mass transfer effects (Figure S7). Taken together, these results indicate that the oxidation of glycerol to glyceric acid over Au_1.5Pt_1.5/MgO-Al₂O₃ proceeds under kinetic control, with negligible influence from mass transfer limitations under the studied conditions.

3. Materials and Methods

3.1. Materials

Glycerol (GLY), glyceric acid (GLA), acetic acid (AA), oxalic acid (OA), tartronic acid (TTA), glycolic acid (GA), 1,3-dihydroxyacetone (DHA), and NaBH₄, hexamethylenetetramine (HMT) were purchased from Sinopharm Chemical Reagent Co., Ltd. (Shanghai, China). HAuCl₄·3H₂O, H₂PtCl₆·6H₂O, Mg(NO₃)₂·6H₂O, Al(NO₃)₃·9H₂O, and Zn(NO₃)₂·6H₂O were purchased from Macklin (Shanghai, China). Deionized water was used in all of the experiments. All chemicals were used without further purification.

3.2. Catalyst Preparation

Different types of mixed-oxide supports were prepared by the sedimentation–hydrothermal method. Firstly, Mg(NO₃)₂·6H₂O, Al(NO₃)₃·9H₂O, and Zn(NO₃)₂·6H₂O were used as precursors to prepare different oxide supports. The specific preparation method was as follows: using hexamethylenetetramine as a precipitator, Mg(NO₃)₂·6H₂O, Al(NO₃)₃·9H₂O, and Zn(NO₃)₂·6H₂O were reacted in a polytetrafluoroethylene autoclave heated at 140 °C for 24 h. The materials recovered by filtration were cleaned with water, dried in the air, and calcined for 4 h at 500 °C, finally. The resulting materials were named MgO-Al₂O₃ and ZnO-Al₂O₃.

Au and Pt monometallic and bimetallic nanoparticle catalysts with varying Au:Pt mass ratios were synthesized on MgO-Al₂O₃ and ZnO-Al₂O₃ supports via a wet chemical reduction method. Specific quantities of HAuCl₄·4H₂O and H₂PtCl₆·6H₂O were mixed with the supports and stirred for 30 min, followed by dropwise addition of an excess of NaBH₄ (molar ratio of NaBH₄/AuPt = 10:1). After vacuum drying at 40 °C, catalysts designated as Au₃/MgO-Al₂O₃, Pt₃/MgO-Al₂O₃, Au₁Pt₂/MgO-Al₂O₃, Au₂Pt₁/MgO-Al₂O₃, Au_1.5Pt_1.5/MgO-Al₂O₃, and Au_1.5Pt_1.5/ZnO-Al₂O₃ were obtained. Additionally, a bimetallic catalyst with Au:Pt = 1.5:1.5 was prepared via H₂ reduction on MgO-Al₂O₃, denoted as ⁰Au_1.5Pt_1.5/MgO-Al₂O₃. The catalytic activity of this H₂-reduced catalyst (⁰Au_1.5Pt_1.5/MgO-Al₂O₃) was subsequently compared with that of the wet chemically reduced Au_1.5Pt_1.5/MgO-Al₂O₃ catalysts.

3.3. Catalyst Characterization

The morphologies of MgO-Al₂O₃-supported Au NP, Pt NP, and AuPt alloy NP catalysts (Au₃/MgO-Al₂O₃, Pt₃/MgO-Al₂O₃, Au₁Pt₂/MgO-Al₂O₃, Au₂Pt₁/MgO-Al₂O₃, Au_1.5Pt_1.5/MgO-Al₂O₃) were observed by transmission electron microscopy (TEM, JSM-7800F, JEOL, Tokyo, Japan). The samples dispersed in anhydrous ethanol solvent were dropped onto copper grids and dried overnight in a desiccator. The sample powders were recorded on an X-ray diffractometer (XRD, D8 Advance, Bruker AXS GmbH, Karlsruhe, Germany) using Cu Kα radiation (k = 0.154 nm) with a Ni filter. The measurements were conducted with steps of 7°/min and a 2θ range of 10–80°. The morphology and chemical analyses of Au_1.5Pt_1.5/MgO-Al₂O₃ were studied using scanning electron micrographs coupled with energy-dispersive spectroscopy (SEM, HiVac Apreo S, Thermo Fisher Scientific, Waltham, MA, USA). The samples were deposited on an aluminum sample holder and sputtered with platinum before taking measurements. N₂ adsorption–desorption isotherms were measured at −196 °C with physical adsorption apparatus (Micromeritics TriStar II 3020 analyzer, Norcross, GA, USA). The samples were previously degassed at 120 °C for 180 min under vacuum. The specific surface areas (S_BET) and average pore sizes of the samples were determined by BET and BJH methods, respectively. The binding energy and surface valence of the catalysts were recorded on an X-ray photoemission spectrometer (XPS, model Escalab 250Xi, Thermo Fisher Scientific, MA, USA). The bulk atomic ratio was analyzed using an ICP-AES ( model Optima 7300DV, PerkinElmer Inc., MA, USA).

3.4. Catalysis and Analysis Procedures

The catalytic oxidation of GLY was carried out in a miniature high-pressure reactor (YZQR-100, Yanzheng, Shanghai, China), with a heater, electronically controlled mechanical stirrer, and thermocouple. By adding 30 mL GLY (0.1 M) and a certain amount of Au_1.5Pt_1.5/MgO-Al₂O₃ (GLY/Au+Pt = 100 mol/mol), while maintaining a stirring speed of 500 rpm, 0.5 MPa oxygen was passed in, and the reaction was performed at 30 °C. After the reaction, the reactor was cooled in the air for sampling, and the catalyst was filtered and recovered. The product was analyzed on an Agilent high-performance liquid chromatograph equipped with a differential refractive index detector (RID), UV-vis detector (Agilent 1120 HPLC, Agilent Technologies, Santa Clara, CA, USA), and Carbomix H-NP column (7.8 × 300 mm). And sulfuric acid (2.5 mM) was used as a mobile phase at a flow rate of 0.6 mL/min. The concentration of product was analyzed by an external standard method.

3.5. Experimental Design and Mathematical Model

The alkalinity of the support was determined by potentiometric titration using an electrochemical workstation (CHI660E, Shanghai CH Instruments, China) equipped with a pH composite electrode. In the experiment, 35 mg samples of MgO-Al₂O₃ and ZnO-Al₂O₃ were weighed separately. Each sample was added to 50 mL of acetonitrile solution containing 0.1 M NaCl and ultrasonically dispersed for 10 min to form a uniform suspension. Continuous N₂ was bubbled into the solution to remove dissolved oxygen. Titrant (0.01 M HCl in acetonitrile) was added at 40s intervals, with the potential difference recorded until a stable measurement value and cumulative titrant volume were achieved. Basic strength was defined as the maximum potential difference (E_max) reached at the start of titration. The total number of basic active sites is calculated using Equation (6):

$$\mathrm{n} (\mathrm{mmol}/\mathrm{g}) = {\displaystyle \frac{{\mathrm{c}}_{\mathrm{HCL}}\times {\mathrm{V}}_{\mathrm{e}}}{{\mathrm{m}}_{\mathrm{sample}}}}$$

(6)

RSM was used to optimize the key factors affecting the yield of GLY to GLA catalyzed by Au_1.5Pt_1.5/MgO-Al₂O₃. The BBD model in Design-Expert 13 software was used to evaluate the synergistic effect of reaction temperature (X₁), reaction time (X₂) GLY/(Au+Pt) molar ratio (X₃), and O₂ pressure (X₄). Based on the BBD model, 29 groups of experiments were conducted. The research factor-level design for optimizing the yield of GLA is shown in Table S5.

Analysis of variance (ANOVA) was used to evaluate the reliability of the empirical model and to measure the interaction between variables. The response surface model was used to predict the relationship between each variable at the response and experiment levels. This paper illustrates the response surface 3D and 2D interaction diagrams to derive optimal conditions. The coefficient of determination (R²) was used to represent the fitting quality of the polynomial model, and its statistical significance was determined by the F test.

4. Conclusions

In this work, composite oxides were synthesized as catalyst supports, and Au, Pt, and Au-Pt alloy catalysts were prepared via a wet chemical reduction method. These catalysts were employed for the selective oxidation of GLY to GLA. Compared with the monometallic Au or Pt catalysts, the Au-Pt alloy catalysts exhibited superior catalytic performance. Among them, the Au_1.5Pt_1.5/MgO-Al₂O₃ catalyst demonstrated the highest activity. The optimal reaction conditions, determined by response surface methodology, were 36 °C, 0.68 MPa O₂ pressure, 15 h reaction time, and GLY/(Au+Pt) = 100 mol·mol⁻¹. Under these conditions, the GLY conversion reached 72%, with a glyceric acid selectivity of 57.9%. Kinetic analysis revealed that the apparent activation energy of the Au_1.5Pt_1.5/MgO-Al₂O₃ catalyst in this reaction system was 32.7 kJ·mol⁻¹. Overall, the Au-Pt alloy catalysts showed markedly enhanced catalytic activity compared with monometallic systems. Importantly, the reaction proceeded under extremely mild conditions, highlighting the great potential of this catalytic system for future industrial-scale production of GLA.