Study on the Mechanism of Phenylacetaldehyde Formation in a Chinese Water Chestnut-Based Medium during the Steaming Process

The white pulp of the Chinese water chestnut (CWC) is crisp and sweet with delicious flavours and is an important ingredient in many Chinese dishes. Phenylacetaldehyde is a characteristic flavoured substance produced in the steaming and cooking process of CWC. The steaming process and conditions were simulated to construct three Maillard reaction systems which consisted of glucose and phenylalanine, and of both alone. The simulation results showed that glucose and phenylalanine were the reaction substrates for the formation of phenylacetaldehyde. The intermediate α-dicarbonyl compounds (α-DCs) and the final products of the simulated system were detected by solid-phase microextraction (SPME) and gas chromatography–mass spectrometry (GC-MS) methods. Through the above methods the formation mechanism of phenylacetaldehyde is clarified; under the conditions of the steaming process, glucose is caramelized to produce Methylglyoxal (MGO), 2,3-Butanedione (BD), Glyoxal (GO) and other α-DCs. α-DCs and phenylalanine undergo a Strecker degradation reaction to generate phenylacetaldehyde. The optimal ratio of the amount of substance of glucose to phenylalanine for Maillard reaction is 1:4. The results can provide scientific reference for the regulation of flavour substances and the evaluation of flavour quality in the steaming process of fruits and vegetables.


Introduction
The Chinese water chestnut (CWC) is a characteristic agricultural product in China, especially in Guangxi. The annual output of CWC from China exceeds one million tons, and the commodity trade accounts for more than 99% of the world. This is mainly produced in Hezhou and Guilin of Guangxi. CWC is very nutritious yet low in calories. A 100 g raw fresh CWC sample provides about 97 kcal energy. It consists of 73% water, 24% carbohydrate, 5% sugars, 3% dietary fiber, 1.4% protein, 0.1% lipid and a small amount of minerals and vitamins [1]. The white pulp is crisp and sweet with delicious flavours both raw and cooked, and is an important ingredient of many Chinese dishes. Li et al. [2] found through the research of electronic nose, SPME-GC-MS and ultra-performance liquid chromatography-tandem mass spectrometry (UPLC-MS/MS) technology that the characteristic flavour substances produced by CWC in the steaming and cooking process are mainly phenylacetaldehyde, nonanal, decanal, trans-2-octenal, etc. Among them, phenylacetaldehyde has a strong scent of hosta flower [3], which is the main source of the sweet smell of CWC after steaming and cooking. The formation mechanism of flavour substances refers to the processes of cracking, polymerization, condensation, oxidation, dehydration and other reactions between substrates, and flavour substances are finally formed through flavour substance for Strecker aldehyde was detected. Currently, the main methods for the detection of α-DCs include high pressure liquid chromatography-ultraviolet spectrometry (HPLC-UV) [23], HPLC-MS [24], GC-MS [25], etc. HPLC-UV can detect α-DCs introduced into chromophore by derivation, but an extensive derivation time will affect the accuracy of detection. HPLC-MS is an effective method to separate and analyse complex organic mixtures, but short-chain α-DCs cannot be detected, and by-products may be produced during the determination of samples with high sugar content, affecting the accuracy of the detection results [26]. GC-MS has the advantages of high sensitivity, wide detection range, etc. It is used to detect volatile compounds, which can make up for the shortcoming of HPLC-MS in detecting short-chain α-DCs [27].
Compared with fresh CWC, the flavour of CWC becomes more intense after steaming, which may be because the substances contained in CWC crack or react with each other at high temperature to produce flavour substances [2]. The substances involved in the reaction can be preliminarily determined by comparing the content changes before and after steaming. Our team used the detection method in reference [28] to detect the substances contained in fresh CWC and steamed CWC by HPLC-MS/MS. According to the changes in its peak area, the contents of phenylalanine and glucose in steamed CWC were significantly reduced, which may be the possible substrate of the characteristic flavour substance phenylacetaldehyde. Our previous research showed that the peak area of Lphenylalanine decreased from 1,360,000.00 to 1,233,333.33 after the CWC was steamed, the peak area of Phe-Phe decreased from 1,366,666.67 to 1,173,333.33, the peak area of N-Acetyl-L-phenylalanine decreased from 1,836,666.67 to 1,306,666.67, the peak area of D-(+)-Phenylalanine decreased from 3,620,000.00 to 3,346,666.67, the peak area of glucosamine decreased from 537,000.00 to 491,666.67, the peak area of N-Acetyl-D-glucosamine decreased from 711,666.67 to 650,666.67, the peak area of D-Glucose 6-phosphate decreased from 2,136,666.67 to 1,406,666.67, and the peak area of Glucose-1-phosphate decreased from 2,156,666.67 to 1,426,666.67. As can be seen from the detection data, the content decreased greatly. According to reference [29], phenylacetaldehyde is generated from glucose and phenylalanine by the Maillard reaction. In this study, glucose and phenylalanine were selected as possible substrates for the characteristic flavour substance phenylacetaldehyde, and three simulation systems were established to confirm the reaction substrates for phenylacetaldehyde. We simulated the contents of glucose and phenylalanine in CWC and its steaming processing conditions, captured the small molecules formed in the reaction process with OPD, detected α-DCs and the final products with SPME-GC-MS, and determined the formation substrate and molecular transfer pathway of phenylacetaldehyde to clarify the formation mechanism of the flavour substance phenylacetaldehyde. The results could provide a scientific reference for study on the controlled release of flavour substances and the evaluation and regulation of flavour quality in the process of CWC steaming, and could have important significance for improving the flavour quality of processed CWC products.

Construction of Maillard Reaction Simulation System
Referring to the data on the content of glucose and phenylalanine in CWC detected by the research group using HPLC with GB 5009.8-2016 and GB 5009.124-2016, respectively, we built three simulation systems.
System 1: 2.408/100 g glucose and 0.011/100 g phenylalanine were weighed and dissolved in a 0.2 mol/L buffer solution of disodium hydrogen phosphate and sodium dihydrogen phosphate at pH 5.4, and then fixed at a volume of 100 mL.
System 2: 2.408/100 g glucose was weighed and dissolved in the buffer solution above. System 3: 0.011/100 g phenylalanine was weighed and dissolved in the buffer solution above.
Accurately, 2.408 g of the glucose and 0.011 g of the phenylalanine were weighed and dissolved in 0.2 mol/L at pH 5.4 disodium hydrogen phosphate and sodium dihydrogen phosphate buffer solution, respectively, and the volume was fixed to 100 mL to obtain two Maillard reaction control simulation systems (2 and 3).
Then, 5 mL solution of the above three simulated systems was removed with a pipette gun and placed in three 20 mL headspace bottles; the internal standard TMV was dissolved in methanol to make the final volume concentration of the solution 0.2 µL/mL, and the solution was accurately absorbed and added to the solution in the headspace bottle with 1 µL; an ordinary steamer and an induction cooker were used for steaming. The output power of the induction cooker was 800 W. After the water boiled, the headspace bottle was put into the steamer for steaming and was timed. The steaming times were 0, 10, 20 and 30 min. The steam environment in the steamer at this time was 100 • C. After the steaming, the test sample was obtained.

SPME-GC-MS Detection
The steaming flask was placed in a constant temperature water bath at 80 • C and stirred with 100 rpm magnetic force. The 50/30 µm DVB/CAR/PDMS Gray extraction head, aged at 280 • C for 35 min, was passed through the sealing pad of the flask for solidphase microextraction (SPME) for 40 min. Three parallel experiments were conducted.
MS conditions: Elector ionization (EI) mode of electron bombardment; ionization energy was 70 eV; ion source temperature was 230 • C; interface temperature was 280 • C; full-scan monitoring mode; mass scan range was 30-500 m/z.
Qualitative method [30]: We manually analysed the mass spectrum information of the sample, matched it with the Library Mainlib standard library and retained the positive and negative matching degrees of more than 800, or one volatile substance more than 900.
Quantitative method [30]: A semi-quantitative method was used in this study. We calculated the ratio of the peak area of the volatile to TMV to obtain its absolute concentration (assuming that the absolute correction factor of each volatile was 1.0) [31]. The calculation formula of the volatile matter concentration in the sample was as follows: where V 2 is the peak area of volatile substance, V 1 is the peak area of the internal standard, 0.2 is the concentration of the internal standard (µL/mL), 0.92 is the density of the internal standard (mg/ µL), 1 is the volume of the internal standard (µL), and m is the mass of the sample 5 g.

Confirmation of Simulation System and Reaction Substrate
The SPME-GC-MS detection results of the reaction products of the three simulation systems were compared with the main aldehyde flavour substances in CWC in the literature [2] to confirm the experimental simulation system and the reaction substrate for the formation of phenylacetaldehyde.

Determination of the Optimum Proportion of Reaction Substrate
With reference to method 2.2, the molar ratios of glucose to phenylalanine in the experimental simulation system were made into 1:1, 1:2, 1:3, 1:4, 1:5, respectively, and steamed for 30 min. The reaction products of the simulation system were tested with reference to method 2.3, and the optimal reaction ratio was determined based on the yield of phenylacetaldehyde.

Determination of α-DCs
With reference to the literature [32], according to the quality of glucose, 0.7236 g OPD was added to the experimental simulation system and simulation system 2, respectively. The experiment was carried out with reference to method 2.2. OPD captured the α-DCs, the intermediate products of the simulation system, which should not be directly detected by GC-MS, to generate stable quinoxalines ( Figure 1). Then, quinoxalines were detected according to the method in Section 2.3, and the quinoxalines were analysed according to the structure of quinoxalines in Figure 1. α-DCs were obtained.
where V2 is the peak area of volatile substance, V1 is the peak area of the internal standard, 0.2 is the concentration of the internal standard (μL/mL), 0.92 is the density of the internal standard (mg/ μL), 1 is the volume of the internal standard (μL), and m is the mass of the sample 5 g.

Confirmation of Simulation System and Reaction Substrate
The SPME-GC-MS detection results of the reaction products of the three simulation systems were compared with the main aldehyde flavour substances in CWC in the literature [2] to confirm the experimental simulation system and the reaction substrate for the formation of phenylacetaldehyde.

Determination of the Optimum Proportion of Reaction Substrate
With reference to method 2.2, the molar ratios of glucose to phenylalanine in the experimental simulation system were made into 1:1, 1:2, 1:3, 1:4, 1:5, respectively, and steamed for 30 min. The reaction products of the simulation system were tested with reference to method 2.3, and the optimal reaction ratio was determined based on the yield of phenylacetaldehyde.

Determination of α-DCs
With reference to the literature [32], according to the quality of glucose, 0.7236 g OPD was added to the experimental simulation system and simulation system 2, respectively. The experiment was carried out with reference to method 2.2. OPD captured the α-DCs, the intermediate products of the simulation system, which should not be directly detected by GC-MS, to generate stable quinoxalines ( Figure 1). Then, quinoxalines were detected according to the method in Section 2.3, and the quinoxalines were analysed according to the structure of quinoxalines in Figure 1. α-DCs were obtained.

Reaction Substrate for the Formation of Phenylacetaldehyde
The SPME-GC-MS results of the reaction products of the three simulated systems are shown in Table 1. It can be seen from Table 1 that the main aldehydes produced by simulation system 1 are consistent with the aldehydes produced by CWC cooking in the literature [2], and only simulation system 1 generates phenylacetaldehyde, the characteristic flavour substance of CWC. Therefore, simulation system 1 was selected as the experimental simulation system to study the formation mechanism of phenylacetaldehyde in the steaming process of CWC. It can also be found from Table 1 that the relative content of phenylacetaldehyde is the highest among the aldehydes produced in simulation system 1 and increases with the increase in steaming time ( Figure 2). This indicates that phenylacetaldehyde is formed by a Maillard reaction of glucose and phenylalanine, which are the reaction substrates for the formation of phenylacetaldehyde. Phenylacetaldehyde was formed at 0 min in Figure 2, which indicates that the Maillard reaction between glucose and phenylalanine could take place when solid-phase microextraction is carried out at 80 °C.

Reaction Substrate for the Formation of Phenylacetaldehyde
The SPME-GC-MS results of the reaction products of the three simulated systems are shown in Table 1. It can be seen from Table 1 that the main aldehydes produced by simulation system 1 are consistent with the aldehydes produced by CWC cooking in the literature [2], and only simulation system 1 generates phenylacetaldehyde, the characteristic flavour substance of CWC. Therefore, simulation system 1 was selected as the experimental simulation system to study the formation mechanism of phenylacetaldehyde in the steaming process of CWC. It can also be found from Table 1 that the relative content of phenylacetaldehyde is the highest among the aldehydes produced in simulation system 1 and increases with the increase in steaming time ( Figure 2). This indicates that phenylacetaldehyde is formed by a Maillard reaction of glucose and phenylalanine, which are the reaction substrates for the formation of phenylacetaldehyde. Phenylacetaldehyde was formed at 0 min in Figure 2, which indicates that the Maillard reaction between glucose and phenylalanine could take place when solid-phase microextraction is carried out at 80 • C. With reference to the literature [32], according to the quality of glucose, 0.7236 g O was added to the experimental simulation system and simulation system 2, respective The experiment was carried out with reference to method 2.2. OPD captured the α-D the intermediate products of the simulation system, which should not be directly detec by GC-MS, to generate stable quinoxalines ( Figure 1). Then, quinoxalines were detec according to the method in Section 2.3, and the quinoxalines were analysed according the structure of quinoxalines in Figure 1. α-DCs were obtained.

Reaction Substrate for the Formation of Phenylacetaldehyde
The SPME-GC-MS results of the reaction products of the three simulated systems shown in Table 1. It can be seen from Table 1 that the main aldehydes produced by sim lation system 1 are consistent with the aldehydes produced by CWC cooking in the lit ature [2], and only simulation system 1 generates phenylacetaldehyde, the characteris flavour substance of CWC. Therefore, simulation system 1 was selected as the expe mental simulation system to study the formation mechanism of phenylacetaldehyde the steaming process of CWC. It can also be found from Table 1 that the relative cont of phenylacetaldehyde is the highest among the aldehydes produced in simulation syst 1 and increases with the increase in steaming time ( Figure 2). This indicates that pheny cetaldehyde is formed by a Maillard reaction of glucose and phenylalanine, which are reaction substrates for the formation of phenylacetaldehyde. Phenylacetaldehyde w formed at 0 min in Figure 2, which indicates that the Maillard reaction between gluc and phenylalanine could take place when solid-phase microextraction is carried out at °C.

The Optimal Molar Ratio of the Substrate to the Reaction
According to simulation system 1, the relative content of phenylacetaldehyde generated by the reaction of glucose and phenylalanine with different molar ratios of substances is shown in Figure 3. With the increase in the molar ratio of phenylalanine, the relative content of phenylacetaldehyde increased. When the molar ratio of glucose to phenylalanine was 1:4, the relative content of phenylacetaldehyde was close to the maximum, and when the molar ratio of the two substances was 1:5, the relative content of phenylacetaldehyde increased slowly. Therefore, the molar ratio of glucose to phenylalanine is 1:4, which is the best proportion of substances to generate phenylacetaldehyde. According to 2.2, the molar ratio of glucose to phenylalanine is 200.67:1, so the amount of glucose in simulation system 1 is greatly excessive. Determining the optimal ratio of the two substrates involves determining how much glucose is involved in the Maillard reaction with phenylalanine. According to the results, about 2.75 mg of glucose was involved in the Maillard reaction with phenylalanine in the simulated system 1, the glucose was greatly excessive, and the α-DCs mainly came from the decomposition of glucose. The optimal reaction ratio of glucose and phenylalanine can provide a reference for how many molecules of α-DCs are produced on average per glucose molecule.
where V2 is the peak area of volatile substance, V1 is the peak area of the internal standa 0.2 is the concentration of the internal standard (μL/mL), 0.92 is the density of the inter standard (mg/ μL), 1 is the volume of the internal standard (μL), and m is the mass of sample 5 g.

Confirmation of Simulation System and Reaction Substrate
The SPME-GC-MS detection results of the reaction products of the three simulat systems were compared with the main aldehyde flavour substances in CWC in the lite ture [2] to confirm the experimental simulation system and the reaction substrate for formation of phenylacetaldehyde.

Determination of the Optimum Proportion of Reaction Substrate
With reference to method 2.2, the molar ratios of glucose to phenylalanine in the perimental simulation system were made into 1:1, 1:2, 1:3, 1:4, 1:5, respectively, a steamed for 30 min. The reaction products of the simulation system were tested with r erence to method 2.3, and the optimal reaction ratio was determined based on the yield phenylacetaldehyde.

Determination of α-DCs
With reference to the literature [32], according to the quality of glucose, 0.7236 g O was added to the experimental simulation system and simulation system 2, respective The experiment was carried out with reference to method 2.2. OPD captured the α-D the intermediate products of the simulation system, which should not be directly detec by GC-MS, to generate stable quinoxalines ( Figure 1). Then, quinoxalines were detec according to the method in Section 2.3, and the quinoxalines were analysed according the structure of quinoxalines in Figure 1. α-DCs were obtained.

Reaction Substrate for the Formation of Phenylacetaldehyde
The SPME-GC-MS results of the reaction products of the three simulated systems shown in Table 1. It can be seen from Table 1 that the main aldehydes produced by sim lation system 1 are consistent with the aldehydes produced by CWC cooking in the lit ature [2], and only simulation system 1 generates phenylacetaldehyde, the characteris flavour substance of CWC. Therefore, simulation system 1 was selected as the expe mental simulation system to study the formation mechanism of phenylacetaldehyde the steaming process of CWC. It can also be found from Table 1 that the relative cont of phenylacetaldehyde is the highest among the aldehydes produced in simulation syst 1 and increases with the increase in steaming time ( Figure 2). This indicates that pheny cetaldehyde is formed by a Maillard reaction of glucose and phenylalanine, which are reaction substrates for the formation of phenylacetaldehyde. Phenylacetaldehyde w formed at 0 min in Figure 2, which indicates that the Maillard reaction between gluc and phenylalanine could take place when solid-phase microextraction is carried out at °C.

The Main α-DCs Formed
As can be seen from Table 2, α-DCs in the experimental simulation system included MGO, BD, GO, PD and OP, and the sum of their relative contents increased with the increase in the steaming time, which was consistent with the trend of the relative content of phenylacetaldehyde increasing with the increase in the steaming time ( Figure 2). When α-DCs were captured by OPD in the experimental simulation system, no phenylacetaldehyde was detected, indicating that the intermediate product of phenylacetaldehyde was α-DCs. Table 3 shows that the main α-DCs in simulation system 2 were MGO, BD, GO and PD. It can be seen that the two simulation systems generate relatively high contents of MGO and BD from 0 min, namely, 40 min extraction at 80 • C, indicating that MGO and BD mainly come from the caramelization reaction of glucose by comparing the reactants of the two simulation systems and the products in Tables 2 and 3. The relative contents of MGO and BD in the experimental simulation system were slightly higher than that in simulation system 2, indicating that only a small amount of MGO and BD came from the Maillard reaction between glucose and phenylalanine. The relative content of GO ranked third among the intermediates, and the relative content of GO in the two simulation systems was similar, indicating that GO mainly comes from the caramelization reaction of glucose. GO is produced after 20 min of steaming, while MGO and BD are produced at 0 min of steaming, indicating that MGO and BD are more likely to be produced in a caramelization reaction. The relative content of PD in the simulation system was about 50% of that in the experimental simulation system, indicating that both the caramelization and Maillard reactions can generate PD, but the relative content of PD in the two simulation systems was small, and the contribution of PD to the generation of the characteristic flavour substance phenylacetaldehyde was small. OP only existed in the experimental simulation system and was derived from the Maillard reaction between glucose and phenylalanine, but its relative content was small, and it did not contribute much to the formation of the characteristic flavour substance phenylacetaldehyde. Table 2. Species and relative content of α-DCs in experimental simulation system (µg/g).

Steamed
Time  Therefore, the main α-DCs in the experimental simulation system were MGO, BD and GO, which mainly came from the caramelization reaction of glucose. Their relative content was high, and they contributed the most to the formation of the characteristic flavour substance phenylacetaldehyde.

The Formation Mechanism of Major α-DCs
According to the analysis results in Section 3.3, the formation mechanism of MGO, BD and GO can be analysed. The 3-DG generated by the enolization of glucose under acidic conditions and G generated by oxidation retains the complete hexose skeleton. These long-chain α-dicarbonyl substances are easily broken to form small-molecule α-DCs through enolization and oxidation reactions.
(1) MGO is mainly produced by the caramelization of glucose, and 3-DG is generated by the enolation of glucose through 1,2-enolation ( Figure 4) [33]. The reverse aldol condensation of 3-DG at C3-C4 directly produces MGO [18,34] (Figure 5). Under the condition of excessive glucose in CWC, this is the main pathway.    Therefore, the main α-DCs in the experimental simulation system were MGO, BD and GO, which mainly came from the caramelization reaction of glucose. Their relative content was high, and they contributed the most to the formation of the characteristic flavour substance phenylacetaldehyde.

The Formation Mechanism of Major α-DCs
According to the analysis results in Section 3.3, the formation mechanism of MGO, BD and GO can be analysed. The 3-DG generated by the enolization of glucose under acidic conditions and G generated by oxidation retains the complete hexose skeleton. These long-chain α-dicarbonyl substances are easily broken to form small-molecule α-DCs through enolization and oxidation reactions.
(1) MGO is mainly produced by the caramelization of glucose, and 3-DG is generated by the enolation of glucose through 1,2-enolation ( Figure 4) [33]. The reverse aldol condensation of 3-DG at C3-C4 directly produces MGO [18,34] (Figure 5). Under the condition of excessive glucose in CWC, this is the main pathway.  Therefore, the main α-DCs in the experimental simulation system were MG and GO, which mainly came from the caramelization reaction of glucose. Their r content was high, and they contributed the most to the formation of the characteris vour substance phenylacetaldehyde.

The Formation Mechanism of Major α-DCs
According to the analysis results in Section 3.3, the formation mechanism of BD and GO can be analysed. The 3-DG generated by the enolization of glucose acidic conditions and G generated by oxidation retains the complete hexose ske These long-chain α-dicarbonyl substances are easily broken to form small-molec DCs through enolization and oxidation reactions.
(1) MGO is mainly produced by the caramelization of glucose, and 3-DG is gen by the enolation of glucose through 1,2-enolation ( Figure 4) [33]. The reverse aldo densation of 3-DG at C3-C4 directly produces MGO [18,34] (Figure 5). Under the con of excessive glucose in CWC, this is the main pathway.   Therefore, the main α-DCs in the experimental simulatio and GO, which mainly came from the caramelization reaction content was high, and they contributed the most to the formatio vour substance phenylacetaldehyde.

The Formation Mechanism of Major α-DCs
According to the analysis results in Section 3.3, the forma BD and GO can be analysed. The 3-DG generated by the eno acidic conditions and G generated by oxidation retains the c These long-chain α-dicarbonyl substances are easily broken to DCs through enolization and oxidation reactions.

Glyoxal (GO)
Foods 2023, 12, 498  Therefore, the main α-DCs in the experim and GO, which mainly came from the carame content was high, and they contributed the mo vour substance phenylacetaldehyde.

The Formation Mechanism of Major α-D
According to the analysis results in Sectio BD and GO can be analysed. The 3-DG gene acidic conditions and G generated by oxidat These long-chain α-dicarbonyl substances are DCs through enolization and oxidation reactio (1) MGO is mainly produced by the caram by the enolation of glucose through 1,2-enola densation of 3-DG at C3-C4 directly produces M of excessive glucose in CWC, this is the main p   Therefore, the main α-DC and GO, which mainly came f content was high, and they con vour substance phenylacetalde

The Formation Mechanism
According to the analysis BD and GO can be analysed. acidic conditions and G gener These long-chain α-dicarbonyl DCs through enolization and o (1) MGO is mainly produc by the enolation of glucose thr densation of 3-DG at C3-C4 dire of excessive glucose in CWC, th    Table 3. Species and relative content of α-DCs in simulation system 2 (µg/g).

Methylglyoxal (MGO)
, 12, 498 8 of 13  Therefore, the main α-DCs in the experimental simulation system were MGO, BD and GO, which mainly came from the caramelization reaction of glucose. Their relative content was high, and they contributed the most to the formation of the characteristic flavour substance phenylacetaldehyde.

The Formation Mechanism of Major α-DCs
According to the analysis results in Section 3.3, the formation mechanism of MGO, BD and GO can be analysed. The 3-DG generated by the enolization of glucose under acidic conditions and G generated by oxidation retains the complete hexose skeleton. These long-chain α-dicarbonyl substances are easily broken to form small-molecule α-DCs through enolization and oxidation reactions.
(1) MGO is mainly produced by the caramelization of glucose, and 3-DG is generated

2,3-Butanedione (BD)
Foods 2023, 12, 498 8 of 13  Therefore, the main α-DCs in the experimental simulation system were MGO, BD and GO, which mainly came from the caramelization reaction of glucose. Their relative content was high, and they contributed the most to the formation of the characteristic flavour substance phenylacetaldehyde.

The Formation Mechanism of Major α-DCs
According to the analysis results in Section 3.3, the formation mechanism of MGO, BD and GO can be analysed. The 3-DG generated by the enolization of glucose under acidic conditions and G generated by oxidation retains the complete hexose skeleton. These long-chain α-dicarbonyl substances are easily broken to form small-molecule α-DCs through enolization and oxidation reactions.
(1) MGO is mainly produced by the caramelization of glucose, and 3-DG is generated

2,3-Pentanedione (PD)
Foods 2023, 12, 498  Therefore, the main α-DCs in the experimental simulation system w and GO, which mainly came from the caramelization reaction of glucose content was high, and they contributed the most to the formation of the ch vour substance phenylacetaldehyde.

The Formation Mechanism of Major α-DCs
According to the analysis results in Section 3.3, the formation mech BD and GO can be analysed. The 3-DG generated by the enolization of acidic conditions and G generated by oxidation retains the complete h These long-chain α-dicarbonyl substances are easily broken to form sm DCs through enolization and oxidation reactions.
(1) MGO is mainly produced by the caramelization of glucose, and 3-D

Glyoxal (GO)
Foods 2023, 12, 498  Therefore, the main α-DCs in the experimental and GO, which mainly came from the caramelization content was high, and they contributed the most to the vour substance phenylacetaldehyde.

The Formation Mechanism of Major α-DCs
According to the analysis results in Section 3.3, t BD and GO can be analysed. The 3-DG generated b acidic conditions and G generated by oxidation reta These long-chain α-dicarbonyl substances are easily DCs through enolization and oxidation reactions.
(1) MGO is mainly produced by the caramelizatio   Therefore, the main α-DCs in the experimental simulation system were MGO, BD and GO, which mainly came from the caramelization reaction of glucose. Their relative content was high, and they contributed the most to the formation of the characteristic flavour substance phenylacetaldehyde.

The Formation Mechanism of Major α-DCs
According to the analysis results in Section 3.3, the formation mechanism of MGO, BD and GO can be analysed. The 3-DG generated by the enolization of glucose under acidic conditions and G generated by oxidation retains the complete hexose skeleton. These longchain α-dicarbonyl substances are easily broken to form small-molecule α-DCs through enolization and oxidation reactions.
(1) MGO is mainly produced by the caramelization of glucose, and 3-DG is generated by the enolation of glucose through 1,2-enolation ( Figure 4) [33]. The reverse aldol condensation of 3-DG at C3-C4 directly produces MGO [18,34] (Figure 5). Under the condition of excessive glucose in CWC, this is the main pathway. where V2 is the peak area of volatile substance, V1 is the peak area of the internal standa 0.2 is the concentration of the internal standard (μL/mL), 0.92 is the density of the inter standard (mg/ μL), 1 is the volume of the internal standard (μL), and m is the mass of sample 5 g.

Confirmation of Simulation System and Reaction Substrate
The SPME-GC-MS detection results of the reaction products of the three simulat systems were compared with the main aldehyde flavour substances in CWC in the lite ture [2] to confirm the experimental simulation system and the reaction substrate for formation of phenylacetaldehyde.

Determination of the Optimum Proportion of Reaction Substrate
With reference to method 2.2, the molar ratios of glucose to phenylalanine in the perimental simulation system were made into 1:1, 1:2, 1:3, 1:4, 1:5, respectively, a steamed for 30 min. The reaction products of the simulation system were tested with r erence to method 2.3, and the optimal reaction ratio was determined based on the yield phenylacetaldehyde.

Determination of α-DCs
With reference to the literature [32], according to the quality of glucose, 0.7236 g O was added to the experimental simulation system and simulation system 2, respective The experiment was carried out with reference to method 2.2. OPD captured the α-D the intermediate products of the simulation system, which should not be directly detec by GC-MS, to generate stable quinoxalines (Figure 1). Then, quinoxalines were detec according to the method in Section 2.3, and the quinoxalines were analysed according the structure of quinoxalines in Figure 1. α-DCs were obtained.

Reaction Substrate for the Formation of Phenylacetaldehyde
The SPME-GC-MS results of the reaction products of the three simulated systems shown in Table 1. It can be seen from Table 1 that the main aldehydes produced by sim lation system 1 are consistent with the aldehydes produced by CWC cooking in the lit With reference to the literature [32], according to the quality of glucose, 0.7236 g O was added to the experimental simulation system and simulation system 2, respective The experiment was carried out with reference to method 2.2. OPD captured the α-D the intermediate products of the simulation system, which should not be directly detec by GC-MS, to generate stable quinoxalines (Figure 1). Then, quinoxalines were detec according to the method in Section 2.3, and the quinoxalines were analysed according the structure of quinoxalines in Figure 1. α-DCs were obtained.

Reaction Substrate for the Formation of Phenylacetaldehyde
The SPME-GC-MS results of the reaction products of the three simulated systems shown in Table 1. It can be seen from Table 1 that the main aldehydes produced by sim lation system 1 are consistent with the aldehydes produced by CWC cooking in the lit ature [2], and only simulation system 1 generates phenylacetaldehyde, the characteris flavour substance of CWC. Therefore, simulation system 1 was selected as the expe mental simulation system to study the formation mechanism of phenylacetaldehyde the steaming process of CWC. It can also be found from Table 1 that the relative cont of phenylacetaldehyde is the highest among the aldehydes produced in simulation syst 1 and increases with the increase in steaming time ( Figure 2). This indicates that pheny cetaldehyde is formed by a Maillard reaction of glucose and phenylalanine, which are reaction substrates for the formation of phenylacetaldehyde. Phenylacetaldehyde w formed at 0 min in Figure 2, which indicates that the Maillard reaction between gluc and phenylalanine could take place when solid-phase microextraction is carried out at °C. (2) There are three GO formation mechanisms: 1 G produced by glucose oxidation can form GO through C2-C3 cleavage or reverse aldol condensation [35][36][37] (Figure 6). Another product, 2,3,4-trihydroxybutyral, can undergo C2-C3 reverse aldol condensation and an oxidation reaction to produce GO (Figure 7). systems were compared with the main aldehyde flavour substances in CWC in the lite ture [2] to confirm the experimental simulation system and the reaction substrate for formation of phenylacetaldehyde.

Determination of the Optimum Proportion of Reaction Substrate
With reference to method 2.2, the molar ratios of glucose to phenylalanine in the perimental simulation system were made into 1:1, 1:2, 1:3, 1:4, 1:5, respectively, a steamed for 30 min. The reaction products of the simulation system were tested with r erence to method 2.3, and the optimal reaction ratio was determined based on the yield phenylacetaldehyde.

Determination of α-DCs
With reference to the literature [32], according to the quality of glucose, 0.7236 g O was added to the experimental simulation system and simulation system 2, respective The experiment was carried out with reference to method 2.2. OPD captured the α-D the intermediate products of the simulation system, which should not be directly detec by GC-MS, to generate stable quinoxalines (Figure 1). Then, quinoxalines were detec according to the method in Section 2.3, and the quinoxalines were analysed according the structure of quinoxalines in Figure 1. α-DCs were obtained.

Reaction Substrate for the Formation of Phenylacetaldehyde
The SPME-GC-MS results of the reaction products of the three simulated systems shown in Table 1. It can be seen from Table 1 that the main aldehydes produced by sim lation system 1 are consistent with the aldehydes produced by CWC cooking in the lit ature [2], and only simulation system 1 generates phenylacetaldehyde, the characteris flavour substance of CWC. Therefore, simulation system 1 was selected as the expe mental simulation system to study the formation mechanism of phenylacetaldehyde the steaming process of CWC. It can also be found from Table 1 that the relative cont of phenylacetaldehyde is the highest among the aldehydes produced in simulation syst 1 and increases with the increase in steaming time (Figure 2). This indicates that pheny cetaldehyde is formed by a Maillard reaction of glucose and phenylalanine, which are reaction substrates for the formation of phenylacetaldehyde. Phenylacetaldehyde w formed at 0 min in Figure 2, which indicates that the Maillard reaction between gluc and phenylalanine could take place when solid-phase microextraction is carried out at °C. where V2 is the peak area of volatile substance, V1 is the peak area of the internal standa 0.2 is the concentration of the internal standard (μL/mL), 0.92 is the density of the inter standard (mg/ μL), 1 is the volume of the internal standard (μL), and m is the mass of sample 5 g.

Confirmation of Simulation System and Reaction Substrate
The SPME-GC-MS detection results of the reaction products of the three simulat systems were compared with the main aldehyde flavour substances in CWC in the lite ture [2] to confirm the experimental simulation system and the reaction substrate for formation of phenylacetaldehyde.

Determination of the Optimum Proportion of Reaction Substrate
With reference to method 2.2, the molar ratios of glucose to phenylalanine in the perimental simulation system were made into 1:1, 1:2, 1:3, 1:4, 1:5, respectively, a steamed for 30 min. The reaction products of the simulation system were tested with r erence to method 2.3, and the optimal reaction ratio was determined based on the yield phenylacetaldehyde.

Determination of α-DCs
With reference to the literature [32], according to the quality of glucose, 0.7236 g O was added to the experimental simulation system and simulation system 2, respective The experiment was carried out with reference to method 2.2. OPD captured the α-D the intermediate products of the simulation system, which should not be directly detec by GC-MS, to generate stable quinoxalines ( Figure 1). Then, quinoxalines were detec according to the method in Section 2.3, and the quinoxalines were analysed according the structure of quinoxalines in Figure 1. α-DCs were obtained.

Reaction Substrate for the Formation of Phenylacetaldehyde
The SPME-GC-MS results of the reaction products of the three simulated systems shown in Table 1. It can be seen from Table 1 that the main aldehydes produced by sim lation system 1 are consistent with the aldehydes produced by CWC cooking in the lit ature [2], and only simulation system 1 generates phenylacetaldehyde, the characteris flavour substance of CWC. Therefore, simulation system 1 was selected as the expe mental simulation system to study the formation mechanism of phenylacetaldehyde the steaming process of CWC. It can also be found from Table 1 that the relative cont of phenylacetaldehyde is the highest among the aldehydes produced in simulation syst 1 and increases with the increase in steaming time ( Figure 2). This indicates that pheny cetaldehyde is formed by a Maillard reaction of glucose and phenylalanine, which are reaction substrates for the formation of phenylacetaldehyde. Phenylacetaldehyde w formed at 0 min in Figure 2, which indicates that the Maillard reaction between gluc and phenylalanine could take place when solid-phase microextraction is carried out at °C.  where V2 is the peak area of volatile substance, V1 is the peak area of the internal standa 0.2 is the concentration of the internal standard (μL/mL), 0.92 is the density of the inter standard (mg/ μL), 1 is the volume of the internal standard (μL), and m is the mass of sample 5 g.

Confirmation of Simulation System and Reaction Substrate
The SPME-GC-MS detection results of the reaction products of the three simulat systems were compared with the main aldehyde flavour substances in CWC in the lite ture [2] to confirm the experimental simulation system and the reaction substrate for formation of phenylacetaldehyde.

Determination of the Optimum Proportion of Reaction Substrate
With reference to method 2.2, the molar ratios of glucose to phenylalanine in the perimental simulation system were made into 1:1, 1:2, 1:3, 1:4, 1:5, respectively, a steamed for 30 min. The reaction products of the simulation system were tested with r erence to method 2.3, and the optimal reaction ratio was determined based on the yield phenylacetaldehyde.

Determination of α-DCs
With reference to the literature [32], according to the quality of glucose, 0.7236 g O was added to the experimental simulation system and simulation system 2, respective The experiment was carried out with reference to method 2.2. OPD captured the α-D the intermediate products of the simulation system, which should not be directly detec by GC-MS, to generate stable quinoxalines ( Figure 1). Then, quinoxalines were detec according to the method in Section 2.3, and the quinoxalines were analysed according the structure of quinoxalines in Figure 1. α-DCs were obtained.

Reaction Substrate for the Formation of Phenylacetaldehyde
The SPME-GC-MS results of the reaction products of the three simulated systems shown in Table 1. It can be seen from Table 1 that the main aldehydes produced by sim lation system 1 are consistent with the aldehydes produced by CWC cooking in the lit ature [2], and only simulation system 1 generates phenylacetaldehyde, the characteris flavour substance of CWC. Therefore, simulation system 1 was selected as the expe mental simulation system to study the formation mechanism of phenylacetaldehyde the steaming process of CWC. It can also be found from Table 1 that the relative cont of phenylacetaldehyde is the highest among the aldehydes produced in simulation syst 1 and increases with the increase in steaming time (Figure 2). This indicates that pheny cetaldehyde is formed by a Maillard reaction of glucose and phenylalanine, which are reaction substrates for the formation of phenylacetaldehyde. Phenylacetaldehyde w formed at 0 min in Figure 2, which indicates that the Maillard reaction between gluc and phenylalanine could take place when solid-phase microextraction is carried out at °C. 3 GO can be generated through reverse aldol condensation in C3-C4 and oxidation reactions of glucose (Figures 7 and 9). GO was produced only after 30 min of steaming, indicating that the glucose in CWC could not easily produce G, but was more likely to produce 3-DG, and the generated 3-DG mainly produced MGO. where V2 is the peak area of volatile substance, V1 is the peak area of the internal standa 0.2 is the concentration of the internal standard (μL/mL), 0.92 is the density of the inter standard (mg/ μL), 1 is the volume of the internal standard (μL), and m is the mass of sample 5 g.

Confirmation of Simulation System and Reaction Substrate
The SPME-GC-MS detection results of the reaction products of the three simulat systems were compared with the main aldehyde flavour substances in CWC in the lite ture [2] to confirm the experimental simulation system and the reaction substrate for formation of phenylacetaldehyde.

Determination of the Optimum Proportion of Reaction Substrate
With reference to method 2.2, the molar ratios of glucose to phenylalanine in the perimental simulation system were made into 1:1, 1:2, 1:3, 1:4, 1:5, respectively, a steamed for 30 min. The reaction products of the simulation system were tested with r erence to method 2.3, and the optimal reaction ratio was determined based on the yield phenylacetaldehyde.

Determination of α-DCs
With reference to the literature [32], according to the quality of glucose, 0.7236 g O was added to the experimental simulation system and simulation system 2, respective The experiment was carried out with reference to method 2.2. OPD captured the α-D the intermediate products of the simulation system, which should not be directly detec by GC-MS, to generate stable quinoxalines ( Figure 1). Then, quinoxalines were detec according to the method in Section 2.3, and the quinoxalines were analysed according the structure of quinoxalines in Figure 1. α-DCs were obtained.

Reaction Substrate for the Formation of Phenylacetaldehyde
The SPME-GC-MS results of the reaction products of the three simulated systems shown in Table 1. It can be seen from Table 1 that the main aldehydes produced by sim lation system 1 are consistent with the aldehydes produced by CWC cooking in the lit ature [2], and only simulation system 1 generates phenylacetaldehyde, the characteris flavour substance of CWC. Therefore, simulation system 1 was selected as the expe mental simulation system to study the formation mechanism of phenylacetaldehyde the steaming process of CWC. It can also be found from Table 1 that the relative cont of phenylacetaldehyde is the highest among the aldehydes produced in simulation syst (3) There are two mechanisms of BD formation: under the condition of excessive glucose, the main method of BD formation is the oxidation and elimination of 2,3,4trihydroxybutyral produced by the reverse aldol condensation of glucose [13]. In addition, glucose can be polymerized into oligosaccharides, which produce Amadori products and then undergo a "peeling-off" mechanism to generate 1,4-dideoxy-2,3-hexadiketolose, which is then subjected to reverse aldol condensation to generate BD ( Figure 10) [38,39].
With reference to the literature [32], according to the quality of glucose, 0.7236 g O was added to the experimental simulation system and simulation system 2, respective The experiment was carried out with reference to method 2.2. OPD captured the α-D the intermediate products of the simulation system, which should not be directly detec by GC-MS, to generate stable quinoxalines ( Figure 1). Then, quinoxalines were detec according to the method in Section 2.3, and the quinoxalines were analysed according the structure of quinoxalines in Figure 1. α-DCs were obtained.

Reaction Substrate for the Formation of Phenylacetaldehyde
The SPME-GC-MS results of the reaction products of the three simulated systems shown in Table 1. It can be seen from Table 1 that the main aldehydes produced by sim lation system 1 are consistent with the aldehydes produced by CWC cooking in the lit ature [2], and only simulation system 1 generates phenylacetaldehyde, the characteris flavour substance of CWC. Therefore, simulation system 1 was selected as the expe mental simulation system to study the formation mechanism of phenylacetaldehyde the steaming process of CWC. It can also be found from Table 1 that the relative cont of phenylacetaldehyde is the highest among the aldehydes produced in simulation syst 1 and increases with the increase in steaming time ( Figure 2). This indicates that pheny cetaldehyde is formed by a Maillard reaction of glucose and phenylalanine, which are reaction substrates for the formation of phenylacetaldehyde. Phenylacetaldehyde w formed at 0 min in Figure 2, which indicates that the Maillard reaction between gluc and phenylalanine could take place when solid-phase microextraction is carried out at °C. Figure 10. The glucose produces BD by reverse aldol condensation and "peeling-off" mechanism.

α-DCs React with Phenylalanine to Form Phenylacetaldehyde
In this study, the reaction mechanism of α-DCs with phenylalanine to produce phenylacetaldehyde is shown in Figure 11 [20]. The amino group of L-phenylalanine first underwent a nucleophilic addition reaction with the carbonyl group of α-DCs and removed one molecule of water to form imine I. After intramolecular rearrangement and decarboxylation, one molecule of CO 2 was lost to form intermediate II. Then, α-aminoketones and the target product phenylacetaldehyde were obtained after hydrolysis.

Confirmation of Simulation System and Reaction Substrate
The SPME-GC-MS detection results of the reaction products of the three simulat systems were compared with the main aldehyde flavour substances in CWC in the lite ture [2] to confirm the experimental simulation system and the reaction substrate for formation of phenylacetaldehyde.

Determination of the Optimum Proportion of Reaction Substrate
With reference to method 2.2, the molar ratios of glucose to phenylalanine in the perimental simulation system were made into 1:1, 1:2, 1:3, 1:4, 1:5, respectively, a steamed for 30 min. The reaction products of the simulation system were tested with r erence to method 2.3, and the optimal reaction ratio was determined based on the yield phenylacetaldehyde.

Determination of α-DCs
With reference to the literature [32], according to the quality of glucose, 0.7236 g O was added to the experimental simulation system and simulation system 2, respective The experiment was carried out with reference to method 2.2. OPD captured the α-D the intermediate products of the simulation system, which should not be directly detec by GC-MS, to generate stable quinoxalines ( Figure 1). Then, quinoxalines were detec according to the method in Section 2.3, and the quinoxalines were analysed according the structure of quinoxalines in Figure 1. α-DCs were obtained.

Reaction Substrate for the Formation of Phenylacetaldehyde
The SPME-GC-MS results of the reaction products of the three simulated systems shown in Table 1. It can be seen from Table 1 that the main aldehydes produced by sim lation system 1 are consistent with the aldehydes produced by CWC cooking in the lit ature [2], and only simulation system 1 generates phenylacetaldehyde, the characteris flavour substance of CWC. Therefore, simulation system 1 was selected as the expe mental simulation system to study the formation mechanism of phenylacetaldehyde the steaming process of CWC. It can also be found from Table 1 that the relative cont of phenylacetaldehyde is the highest among the aldehydes produced in simulation syst 1 and increases with the increase in steaming time ( Figure 2). This indicates that pheny cetaldehyde is formed by a Maillard reaction of glucose and phenylalanine, which are reaction substrates for the formation of phenylacetaldehyde. Phenylacetaldehyde w formed at 0 min in Figure 2, which indicates that the Maillard reaction between gluc and phenylalanine could take place when solid-phase microextraction is carried out at °C. In this study, the formation mechanism of phenylacetaldehyde in the Maillard reaction system is shown in Figure 12. where V2 is the peak area of volatile substance, V1 is the peak area of the internal standa 0.2 is the concentration of the internal standard (μL/mL), 0.92 is the density of the inter standard (mg/ μL), 1 is the volume of the internal standard (μL), and m is the mass of sample 5 g.

Confirmation of Simulation System and Reaction Substrate
The SPME-GC-MS detection results of the reaction products of the three simulat systems were compared with the main aldehyde flavour substances in CWC in the lite ture [2] to confirm the experimental simulation system and the reaction substrate for formation of phenylacetaldehyde.

Determination of the Optimum Proportion of Reaction Substrate
With reference to method 2.2, the molar ratios of glucose to phenylalanine in the perimental simulation system were made into 1:1, 1:2, 1:3, 1:4, 1:5, respectively, a steamed for 30 min. The reaction products of the simulation system were tested with r erence to method 2.3, and the optimal reaction ratio was determined based on the yield phenylacetaldehyde.

Determination of α-DCs
With reference to the literature [32], according to the quality of glucose, 0.7236 g O was added to the experimental simulation system and simulation system 2, respective The experiment was carried out with reference to method 2.2. OPD captured the α-D the intermediate products of the simulation system, which should not be directly detec by GC-MS, to generate stable quinoxalines ( Figure 1). Then, quinoxalines were detec according to the method in Section 2.3, and the quinoxalines were analysed according the structure of quinoxalines in Figure 1. α-DCs were obtained.

Reaction Substrate for the Formation of Phenylacetaldehyde
The SPME-GC-MS results of the reaction products of the three simulated systems shown in Table 1. It can be seen from Table 1 that the main aldehydes produced by sim lation system 1 are consistent with the aldehydes produced by CWC cooking in the lit ature [2], and only simulation system 1 generates phenylacetaldehyde, the characteris flavour substance of CWC. Therefore, simulation system 1 was selected as the expe mental simulation system to study the formation mechanism of phenylacetaldehyde the steaming process of CWC. It can also be found from Table 1 that the relative cont of phenylacetaldehyde is the highest among the aldehydes produced in simulation syst 1 and increases with the increase in steaming time ( Figure 2). This indicates that pheny cetaldehyde is formed by a Maillard reaction of glucose and phenylalanine, which are reaction substrates for the formation of phenylacetaldehyde. Phenylacetaldehyde w formed at 0 min in Figure 2, which indicates that the Maillard reaction between gluc and phenylalanine could take place when solid-phase microextraction is carried out at °C.

Conclusions
In the process of CWC steaming, glucose and phenylalanine in CWC undergo the Maillard reaction to produce the characteristic flavour substance phenylacetaldehyde. The formation mechanism was as follows: under the condition of steaming, glucose caramelized, mainly including the formation of 3-DG by enolization and the formation of G by oxidation. These long-chain α-DCs were converted into small-molecule α-DCs such as MGO, BD and GO by reverse aldol condensation and an oxidation reaction. Among them, MGO had the highest content and was the main product; then, α-DCs and phenylalanine underwent the Strecker degradation reaction to produce phenylacetaldehyde. The optimal molar ratio of glucose and phenylalanine to the Maillard reaction was 1:4. Glucose in CWC is greatly excessive, and phenylalanine can participate in the Maillard reaction completely.