PdBi hydrogels were synthesized through the reduction of palladium acetylacetonate [Pd(acac)₂] and bismuth (III) chloride (BiCl₃) with sodium borohydride (Fig. 1A). The aerogels were obtained via a freeze-drying process of the as-prepared hydrogels (inset in Fig. 1B) [25]. Subsequently, inductively coupled plasma atomic emission spectrometry analysis indicated that the atomic ratio of Pd to Bi in the PdBi aerogels is 100:5 (defined as Pd₁₀₀Bi₅ aerogels). Additionally, pure Pd aerogels and Pd₁₀₀Bi_x (x = 2, 10) were prepared for comparison. This composition is consistent with the molar ratio used in the synthesis process (Table S1). As shown in Fig. 1B and C and Fig. S1, the morphology of the resulting Pd aerogels and Pd₁₀₀Bi₅ aerogels is investigated by scanning electron microscopy and transmission electron microscopy (TEM), revealing 3D architectures composed of interconnected porous and extended nanowire networks [26]. The high-resolution TEM (HR-TEM) result demonstrates that the Pd₁₀₀Bi₅ aerogels have a lattice spacing of 0.237 nm ascribed to the (111) planes of face-centered cubic Pd aerogels (Fig. 1D). Besides, the selected area electron diffraction (SAED) also reveals the polycrystalline nature of Pd₁₀₀Bi₅ aerogels, which are indexed to (111), (200), (220), and (311) of Pd (Fig. S2). The uniformly distributed Pd and Bi elements in the Pd₁₀₀Bi₅ aerogels are further demonstrated by the energy-dispersive x-ray spectroscopy mapping images (Fig. 1E). The x-ray diffraction pattern of Pd and PdBi aerogels reveals the face-centered cubic polycrystalline structure, consistent with the SAED results (Fig. 1F). In contrast to Pd aerogels, the reflections of PdBi aerogels exhibit noticeable shifts toward lower 2θ values as the fraction of Bi increases, which suggests that doping Bi into the Pd lattice can produce tensile strain [27]. Quantitative analysis based on the Bragg formula shows that this tensile strain increases from 0.52% to 1.29% as the Bi content increases (Fig. 1G). Additionally, the (111) lattice stripe spacing is directly measured by HR-TEM, and it is found that the lattice spacing of Pd is 0.232 nm. As the Bi-doping amount increases, the lattice spacing gradually increases, confirming the existence of the strain (Fig. S3). Notably, the continuous increase in this tensile strain suggests that we can precisely control the electronic structure of the Pd lattice by adjusting the amount of Bi doping.

X-ray photoelectron spectroscopy (XPS) was performed to investigate the electronic structure and surface composition. With increasing Bi-doping concentration, a gradual decrease in the binding energy of Pd is observed, accompanied by a concurrent increase in the binding energy of Bi, indicating a clear electron transfer from Pd to Bi atoms (Fig. 2A). More importantly, in the Bi 4f spectra, 2 pairs of peaks are associated with Bi⁰ and Bi³⁺ species. The relative content of Bi³⁺ shows a trend of increasing first, then decreasing, with increasing doping amount. This trend is crucial for understanding the aforementioned charge transfer behavior and its influence on catalytic performance. The Bi³⁺ species possess empty p orbitals, and their increased content importantly enhances the p–d orbital hybridization between Pd and Bi, thereby modulating the electronic interactions within the nanozyme. The XPS results provide direct evidence for the modulation of electronic structure in PdBi aerogels (Fig. 2B).

To verify the presence of orbital interaction between Pd and Bi, the projected density of states (PDOS) was calculated (Fig. 2C). The peaks of the Pd 4d orbitals align well with those of the Bi 6p orbitals, indicating strong p–d orbital hybridization. Based on the orbital symmetry, the d_z² orbital of Pd 4d can directionally hybridize with the p_z orbital of Bi 6p to form σ and σ* orbitals. Meanwhile, the p_x/y orbitals of Bi 6p can directionally interact with the d_xz/d_yz orbitals of Pd 4d, generating π and π* orbitals (Fig. 2D and Fig. S4). Additionally, due to the mismatch in orbital orientation with the Bi 6p orbital, the d_xy and d_x²_−y² orbitals of Pd 4d are considered nonbonding. As shown in Fig. 2E and F, the p–d hybridization center on the orbitals leads to electron transfer from Bi to Pd, which can be verified through XPS spectra and differential charge density analysis. The crucial principle is that the lattice stretching strain caused by Bi doping (Fig. 1G) effectively promotes the occurrence of this p–d hybridization. With the assistance of the strain, the p–d hybridization between Pd and Bi not only induces charge transfer but also adjusts the symmetry and energy level distribution of the Pd d orbitals, thereby shifting the d-band center of Pd toward the Fermi level (Fig. 2G and Fig. S5). It is precisely this strain-induced p–d hybridization that directly regulates the local electronic structure of Pd, laying the foundation for the subsequent cascade orbital hybridization (p–d–σ) in the H₂O₂ activation process.

The ability of Pd₁₀₀Bi_x (x = 2, 5, 10) aerogels to activate H₂O₂ was examined using 3,3′,5,5′-tetramethylbenzidine (TMB) as a substrate, which is quantitatively assessed by measuring the absorbance of the oxidized form of TMB (oxTMB) at 652 nm [28,29]. As expected, Pd₁₀₀Bi_x (x = 2, 5, 10) aerogels have a higher absorbance value than Pd aerogels, demonstrating excellent H₂O₂ activation ability, indicating that the p–d orbital hybridization of Pd₁₀₀Bi_x (x = 2, 5, 10) aerogels plays a critical role in enhancing the H₂O₂ activation ability (Fig. 3A). Among them, Pd₁₀₀Bi_x (x = 2, 10) aerogels are optimized for H₂O₂ activation. The specific activity (SA) is a quantitative measure of the H₂O₂-activation ability of Pd₁₀₀Bi_x (x = 2, 5, 10) aerogels. Pd₁₀₀Bi₅ aerogels demonstrate a high SA value (751.1 U/mg), exceeding those of Pd₁₀₀Bi₁₀ (642.8 U/mg), Pd₁₀₀Bi₂ (161.4 U/mg), and Pd aerogels (107.6 U/mg) (Fig. 3B and Fig. S6). It should be pointed out that the Pd₁₀₀Bi₁₀ aerogels, despite having the largest tensile strain, exhibit a lower H₂O₂ activation performance (SA = 642.8 U/mg) compared with Pd₁₀₀Bi₅ (751.1 U/mg). This indicates that the simple strain effect is not sufficient to optimize the catalytic activity. We believe that the enhanced p–d orbital hybridization due to the moderate Bi doping plays a more crucial role in regulating the electrons, enhancing the H₂O₂ adsorption, and promoting the subsequent cascade hybridization at the core of the Pd d-band center. Besides, the kinetics of Pd and PdBi aerogels are evaluated using the Michaelis–Menten equation (Fig. S7 and Table S2). Pd₁₀₀Bi₅ aerogels exhibit lower K_m values and higher V_max values compared to Pd₁₀₀Bi_x (x = 0, 2, 10), indicating stronger substrate affinity and superior catalytic performance. Then, cyclic voltammetry is performed to assess the catalytic efficiency of H₂O₂ using Pd and Pd₁₀₀Bi₅ aerogel-modified glassy carbon electrodes in an N₂-saturated acetate buffer. As shown in Fig. S8, Pd₁₀₀Bi₅ aerogels exhibit a notable reduction in current, indicating exceptional H₂O₂ reduction activity. To further reveal the mechanism of H₂O₂ activation enhanced by Pd₁₀₀Bi₅ aerogels, the activation energy (E_a) of H₂O₂ activation ability is compared using the Arrhenius equation [30]. The E_a value for the H₂O₂ activation of Pd₁₀₀Bi₅ aerogels (11.51 kJ/mol) is lower than Pd₁₀₀Bi₁₀ (17.50 kJ/mol), Pd₁₀₀Bi₂ (24.91 kJ/mol), and Pd aerogels (30.67 kJ/mol), indicating that Pd₁₀₀Bi₅ aerogels can more easily activate H₂O₂ (Fig. 3C). This value is comparable to those reported for other nanozymes, further confirming the excellent activation ability of Pd₁₀₀Bi₅ aerogels (Table S3). To achieve a more comprehensive understanding of the fundamental mechanism driving the enhanced H₂O₂ activation capability of Pd₁₀₀Bi₅ aerogels, the intermediates were investigated. A standard fluorescent probe, terephthalic acid, can capture •OH, yielding a fluorescent product with an emission peak at 440 nm. Upon the concurrent presence of Pd₁₀₀Bi₅ aerogels, H₂O₂, and terephthalic acid, a conspicuous fluorescence signal emerges, affirming the presence of •OH intermediates (Fig. S9). Electron-paramagnetic-resonance measurements applying 5,5-dimethyl-1-pyrrolidine-N-oxide (DMPO) as the spin-trapping agent were further performed, as shown in Fig. 3D. The Pd₁₀₀Bi₅ aerogels–H₂O₂ system exhibits a more pronounced DMPO–•OH signal than Pd aerogels, indicating that Pd₁₀₀Bi₅ aerogels activate H₂O₂ to generate more •OH. Based on this, to comprehensively determine whether there are other types of ROS in the system, using DMPO and 2,2,6,6-tetramethylpiperidinooxy as traps, respectively, no characteristic signals of superoxide anion radicals (•O₂⁻) and singlet oxygen (¹O₂) were detected, thereby excluding the interference of these 2 common ROS (Fig. S10). In addition, in situ Fourier-transform infrared spectroscopy is used to verify the changes in chemical bonds and functional groups during the activation of H₂O₂ on Pd aerogels and Pd₁₀₀Bi₅ aerogels [31,32]. The surface of the Pd aerogels displays distinct peaks at 1,230 and 1,650 cm⁻¹, which can be attributed to the bending vibrations of •OH and H₂O, respectively. In Pd₁₀₀Bi₅ aerogels, the H₂O bending vibrational peak is importantly enhanced, indicating a higher efficiency of activated H₂O₂. It is noteworthy that, in comparison to Pd aerogels, a bending vibration peak of the O–O bond at 1,125 cm⁻¹ is observed in PdBi aerogels. This suggests that the PdBi aerogels effectively promote the polarized elongation of the O–O bond, thereby facilitating the activation of H₂O₂ for •OH generation (Fig. S11) [33]. To determine the proportion of •OH, a series of scavengers, including tert-butyl alcohol (TBA, •OH scavengers) and dimethylsulfoxide (DMSO, both •OH and metal-oxo species scavengers), are introduced into the H₂O₂ activation process system [34,35].

As depicted in Fig. S12 and Fig. 3E, the inhibitory effects of TBA and DMSO on the activation of H₂O₂ by Pd₁₀₀Bi₅ are similar, indicating that Pd₁₀₀Bi₅ mainly activates H₂O₂ to generate •OH. In contrast to Pd₁₀₀Bi₅, in the Pd system for H₂O₂ activation, the inhibitory effect of DMSO is stronger than that of TBA, suggesting that in the Pd system, H₂O₂ generates •OH and metal-oxo. Further analysis suggests that Pd₁₀₀Bi₅ aerogels decreased metal-oxo selectivity from 23.3% to 1.70%, dramatically increasing the •OH selectivity from 76.7% to 98.3%. Using benzoic acid (BA) as the probe molecule, a specific hydroxylation reaction converts BA into p-hydroxybenzoic acid. We observe that the concentration of •OH in the Pd system is 81.18 μmol/l, while the production of •OH in the Pd₁₀₀Bi₅ system increased to 314.07 μmol/l. Compared with Pd (73.3%), Pd₁₀₀Bi₅ (95.4%) importantly enhances the selectivity of •OH during the activation of H₂O₂ (Fig. S13 and Fig. 3F and G). This result is basically consistent with the results of the quenching experiment with the scavenger. In summary, Pd₁₀₀Bi₅ aerogels can selectively activate H₂O₂ to produce •OH and enhance H₂O₂ activation ability (Fig. 3H). Additionally, the stability of Pd₁₀₀Bi₅ aerogels is assessed. Pd₁₀₀Bi₅ aerogels have excellent tolerance and heat resistance (Figs. S14 and S15). Furthermore, the cyclic stability test demonstrates that after 10 cycles, the Pd₁₀₀Bi₅ aerogels still maintained excellent activity (Fig. S16). To further verify the structural integrity and chemical stability of the Pd₁₀₀Bi₅ after the reaction, we conduct a series of characterization tests on the nanozyme after the reaction. The TEM and x-ray diffraction results show that the Pd₁₀₀Bi₅ aerogel after the reaction still maintains its original porous framework structure and crystal phase composition, which is consistent with the state before the reaction (Fig. S17A to C). The XPS analysis indicated that the surface electronic states of Pd and Bi did not undergo important changes, indicating that the chemical state of the catalyst was maintained during the reaction process (Fig. S17D and E). Additionally, inductively coupled plasma tests on the supernatant after the reaction did not detect any leakage of Pd ions or Bi ions, further ruling out the possibility of the active components leaching out during the reaction. The above results collectively confirm that the Pd₁₀₀Bi₅ aerogel has excellent structural and chemical stability under catalytic reaction conditions (Table S4).

To further clarify the mechanism of H₂O₂ activation and explain the outstanding catalytic performance of Pd₁₀₀Bi₅ aerogels, we propose a cascade hybridization model (p–d–σ), involving site 2 (p-block), site 1 (d-block), and hydrogen peroxide (σ). By precisely controlling the p–d orbital hybridization between sites 1 and 2, specific orbital coupling is expected to be achieved in the subsequent cascade hybridization. Firstly, given that HR-TEM and SAED images confirm (111) as the dominant exposed facet in the Pd and PdBi aerogels, we employ the Pd(111) surface as a representative model to investigate the effect of Bi doping on the electronic structure and adsorption energetics (Fig. S18). Due to the strong orbital hybridization between the d orbitals of Pd and the p orbitals of Bi, the d orbitals of Pd after hybridization have a greater degree of overlap with the bonding and antibonding orbitals when hybridizing with H₂O₂ (Fig. 4B and Fig. S19). Specifically, the energy levels of the 4d_z² and 4d_xz/yz orbitals of Pd atoms in Pd₁₀₀Bi₅ are higher than those in Pd aerogels after p–d hybridization. Therefore, when the 4d orbitals of Pd in Pd₁₀₀Bi₅ interact with the 5σ and 2π* orbitals of H₂O₂, the energy of the antibonding orbitals is higher than that in Pd. This results in an increased electron occupancy of the bonding orbitals, thereby strengthening the adsorption interaction of H₂O₂ on Pd₁₀₀Bi₅ and weakening it on Pd (Fig. 4C and D) [36,37]. Further hybridization between Pd (4d_z², 4d_xy, and 4d_yz) and H₂O₂ (5σ and 2π*) in Pd₁₀₀Bi₅ aerogels is referred to as cascade orbital hybridization, which can achieve highly efficient and selective activation of H₂O₂ into •OH. Furthermore, to verify the role of the cascade orbital hybridization in the selectivity and efficiency of H₂O₂ activation, the lengths of H₂O₂ adsorbed on the surfaces of Pd₁₀₀Bi₅ aerogels and Pd aerogels were also calculated. To be specific, the O–O bonding length of the adsorbent of H₂O₂ (1.508 Å) on PdBi aerogels is longer, while the adsorbent of the O–O bonding length of H₂O₂ (1.489 Å) on Pd is shorter (Fig. 4E). Therefore, compared with Pd aerogels, Pd₁₀₀Bi₅ aerogels are more conducive to the selective generation of •OH from O–O bond breaking in H₂O₂. In addition, as shown in Fig. 4F and Fig. S20, the activation process and energy-change diagram of H₂O₂ on the Pd₁₀₀Bi₅ aerogels and Pd aerogels are presented. Firstly, the free energy of *H₂O₂ on Pd₁₀₀Bi₅ aerogels is significantly lower than that on Pd aerogels, which can promote H₂O₂ adsorption and match well with the PDOS analysis, providing evidence for promoting H₂O₂ activation through cascade orbital hybridization. Besides, *H₂O₂ is cleaved into 2 adsorbed *OH on Pd and Bi sites. Next, *OH is protonated to form *H₂O. Finally, *H₂O is desorbed in succession, and the Pd₁₀₀Bi₅ aerogels return to the initial state. In this reaction, the step from *OH and *H₂O to *OH can be regarded as the rate-determining step. The energy barrier for the rate-determining step of Pd₁₀₀Bi₅ aerogels (1.279 eV) is lower than that of Pd aerogels (1.543 eV), which is attributed to the redistribution of the electronic structure after the hybridization of Pd₁₀₀Bi₅ aerogels and H₂O₂ cascade orbitals. After the occurrence of the cascade orbital hybridization, the d-band center of Pd in the Pd₁₀₀Bi₅ aerogels (−2.42 eV) is lowered relative to the d-band center of pure Pd aerogels (−1.23 eV), which is conducive to the desorption of water molecules on the Pd₁₀₀Bi₅ aerogels (Fig. 4G). In summary, the p–d orbital hybridization interaction between Pd and Bi plays a crucial role in amplifying the cascade orbital hybridization process between Pd and H₂O₂. This interaction promotes the cascade orbital hybridization between Pd and H₂O₂, leading to the redistribution of electrons on the surface of the Pd aerogels, thereby facilitating the cleavage of the O–O bond during H₂O₂ activation and the effective removal of H₂O. As a result, this cascade orbital hybridization importantly enhances the selectivity and efficiency of H₂O₂ activation.

In agriculture, various pesticides, including organophosphorus, organochlorine, carbamates, phenol carboxylic acids, and sulfonamides, are widely used as growth regulators. However, the incomplete degradation of these pesticides leads to residues that pose important threats to food safety. Therefore, simultaneously detecting and differentiating pesticides with diverse structures and properties remains a key scientific issue that urgently needs to be addressed [38–43]. To address this issue, a novel sensing strategy based on the catalytic inhibition effect has been developed. This strategy enables differential recognition of various pesticide types (e.g., organophosphorus, organochlorine, and carbamate) by precisely modulating the interactions between pesticide molecules and the nanozyme’s active sites. Specifically, it shows a gradient inhibition effect on the activation process of H₂O₂ and the generation efficiency of •OH, thereby establishing a quantitative structure–activity relationship between the degree of inhibition and the type and concentration of pesticides.

For instance, organophosphorus pesticides like chlorpyrifos (CPF) strongly adsorb onto the nanozyme surface, importantly suppressing •OH production, while other pesticides (e.g., sulfonamides or phenolic acids) exhibit weaker inhibitory effects. By capturing these varying inhibition signals, the sensor array translates them into distinct response patterns, thereby enabling the detection of different pesticide residues. CPF is employed as a representative model for pesticide molecules. As illustrated in Fig. 5A, the XPS analysis reveals that the peak of Pd 3d shifts importantly with the addition of CPF, indicating an interaction between Pd₁₀₀Bi₅ aerogels and CPF [44,45]. This finding highlights the need for further exploration into the inhibitory effects of CPF on Pd₁₀₀Bi₅ aerogels [27,46,47]. As shown in Fig. 5B, it can be seen that the rate of H₂O₂ activation exhibits a linear relationship with the concentration of Pd₁₀₀Bi₅ aerogels. As CPF increases, the slope of the line through the origin decreases, indicating that CPF reversibly inhibits the ability of Pd₁₀₀Bi₅ aerogels to activate H₂O₂. We additionally determine the initial velocity at various H₂O₂ concentrations in the presence of CPF, and the catalytic kinetics align with the Michaelis–Menten curve equation (Fig. 5C and Table S5). In the presence of the inhibitor CPF, the Michaelis–Menten constant is denoted as K_m^I (inhibited state), while the maximum initial velocity is represented as V_max^I. With the increase of CPF concentration, K_m^I increased and V_max^I decreased (Fig. 5D). Moreover, Lineweaver–Burk plots constructed at various concentrations of CPF crossed in the second quadrant. This suggests that the inhibition is of a mixed type, encompassing both competitive and noncompetitive inhibition (Fig. 5E). K_I and K_I′ represent the inhibition constants of the nanozyme–inhibitor and nanozyme–substrate–inhibitor systems, respectively. K_I and K_I′ are calculated to be 0.557 and 0.783 mM. K_I < K_I′ indicates that the affinity of CPF for Pd₁₀₀Bi₅ is higher than that of the complex formed by Pd₁₀₀Bi₅ and H₂O₂, further suggesting that competitive inhibition plays a dominant role in the mixed inhibition. In conclusion, CPF reduces the affinity of Pd₁₀₀Bi₅ aerogels for the substrate H₂O₂ through a mixed inhibition mode, thereby reducing the activation ability of H₂O₂ (Fig. 5F). To gain deeper insight into the molecular-scale interaction between CPF and the PdBi surface, density functional theory calculations were performed. The CPF molecule is modeled on the Pd₁₀₀Bi₅ surface, with the P=S functional group identified as the primary binding motif. The calculated adsorption energy of CPF on Pd₁₀₀Bi₅ is −0.45 eV, importantly stronger than that of H₂O₂ (−1.55 eV), indicating that CPF preferentially occupies the Pd active sites and competitively inhibits H₂O₂ activation. PDOS further shows pronounced orbital overlap between the Pd 4d orbitals and the S 3p orbitals of CPF near the Fermi level, providing direct evidence for orbital hybridization underlying the competitive inhibition mechanism (Fig. S21). In addition, other typical pesticides, including pentachloronitrobenzene (PCNB), carbaryl (Car), glyphosate (Gly), and metsulfuron-methyl (Met), are also discussed for their inhibition effects on the selective generation of •OH by activated H₂O₂. As shown in Fig. S22, PCNB, Car, Gly, and Met exhibit typical reversible inhibition features. According to the intersection position of the Lineweaver–Burk diagram (Fig. S23), the inhibition types of the above 5 pesticides can be determined (Table S6). Inspired by the results mentioned above, the extent of inhibition of H₂O₂ activation varied among different pesticide molecules, offering a promising approach to distinguish and identify multiple pesticides across diverse categories. Pd, Pd₁₀₀Bi₂, and Pd₁₀₀Bi₅ aerogels are utilized to construct a simple 3-channel sensor array. We employ sensor arrays for cross-reactive discrimination of various pesticides (CPF, PCNB, Car, Gly, and Met). Subsequently, through the application of principal component analysis for transforming the complex training-data matrix, 5 types of pesticides are effectively differentiated, resulting in tightly formed clusters (Fig. 5G). Additionally, the sensing arrays demonstrate the ability to precisely recognize mixtures of CPF and Car in various proportions (Fig. 5H). To provide a more robust and quantitative assessment of sensitivity, we perform conventional univariate calibration using the Pd₁₀₀Bi₅ aerogels that exhibit the highest response to CPF. As shown in Fig. 5I, the response intensity shows a linear relationship with CPF concentration in the range of 0.5 to 2.5 μM. The detection limit (0.23 μM) is calculated as 3σ/slope, where σ is the standard deviation of the blank measurements (n = 10). This detection limit is considerably lower than the maximum residue limits for CPF established by the European Union.

In the actual sample analysis experiment, simulated pesticide spraying is used to analyze pesticide residues on grape surfaces. Extracting solutions containing different types of pesticides as samples to be tested produces unique response signals. As shown in Fig. S24, the 2-dimensional canonical score plot shows that the 5 types of pesticides are separated into 5 distinct regions, clearly apart from each other, demonstrating excellent discriminating ability.

Pd(acac)₂, BiCl₃, PCNB, Gly, CPF, Car, Met, Chipton, and horseradish peroxidase were purchased from Shanghai Aladdin Biochemical Technology Co., Ltd. Sodium acetate (NaAc), terephthalic acid, and TMB were obtained from Sinopharm Chemical Reagent Co., Ltd.

In a typical synthesis of PdBi aerogels, 2 ml of aqueous NaBH₄ (0.05 M) was quickly injected into 30 ml of an aqueous metal precursor solution containing 0.4 ml Pd(acac)₂ (0.1 M) and 0.04 ml BiCl₃ (0.05 M), with stirring at 60 °C. The prepared mixture solution was stirred at 600 rpm for 1 min and changed from faint yellow to black. Finally, the black solution was heated to 60 °C for 2 h to form PdBi aerogels. The black monolith solids were carefully washed with water 3 times and finally dried at −60 °C in the lyophilized chamber. Pd and PdBi aerogels were prepared with Pd and Bi source mole ratios of 100:0, 100:2, and 100:10 using the same process.

The ability of Pd aerogels and PdBi aerogels to activate H₂O₂ was evaluated through a traditional colorimetric reaction, using TMB (a colorimetric agent) as an indicator in the presence of acetic acid–sodium acetate (HAc–NaAc) buffer and an aqueous H₂O₂ solution. Specifically, 5 μl of Pd aerogels or PdBi aerogels (0.05 mg/ml), 100 μl of H₂O₂ (100 mM) water solution, and 100 μl of TMB (1 mM) ethanol solution were sequentially added to 100 μl of HAc–NaAc buffer solution (pH 3.0, 0.1 M). Following a 2-min incubation period, the absorbance at 652 nm of the resulting mixture was measured using a multimode reader.

A colorimetric sensor array based on PdBi aerogels (3 × 6 × 6) was assembled in a 96-well plate. The detailed steps were as follows: HAc–NaAc, TMB, H₂O₂, 3 distinct types of PdBi aerogels, and pesticide standards at various concentrations were added into the 96-well plate. Following a 5-min incubation period, the absorbance of the solution was recorded at 652 nm using a microplate reader. This procedure was conducted 6 times for 6 different antioxidants. As a result, the antioxidants were evaluated using 3 types of PdBi aerogels, generating a 3 × 6 × 6 data matrix.