Study on the Quality Characteristics of Preprocessed, Cooked and Frozen‐Reheated Tofu Balls

ABSTRACT

To elucidate the characteristics and evolution of eating quality in tofu balls during preprocessing, cooking, and frozen reheating, volatile compounds and textural properties were systematically analyzed at each stage. Molecular sensory science, texture analysis, and low‐field nuclear magnetic resonance were applied. In total, 43 volatile compounds were identified, comprising 14 alcohols, 9 aldehydes, 5 ketones, and 9 esters, as well as 1 acid, 1 furan, and 1 pyrazine. Preprocessed tofu balls were mainly characterized by a beany flavor, primarily attributed to 1‐octen‐3‐ol, 3‐heptanol, hexanol, heptanal, hexanal, and 2‐pentylfuran. After cooking and frozen reheating, the aroma profile shifted toward beany, sweet‐fruity, and cereal‐like notes, largely influenced by (E)‐2‐octenal, ethyl butyrate, and 3‐methylbutanol. Following frozen reheating, hardness, gumminess, and chewiness increased significantly. The proportion of bound water increased markedly, and water‐holding capacity reached a maximum. In addition, cooked and reheated samples exhibited a desirable brownish‐yellow color. This study provides a theoretical basis for developing preprocessed plant‐based products that retain traditional tofu flavor while meeting contemporary nutritional requirements. It also offers practical guidance for improving processing technology and the market competitiveness of plant‐based meat alternatives.

1. Introduction

Traditional fresh tofu is recognized as an affordable, nutritious, and tender food and is widely consumed in daily diets. In recent years, growing interest in plant‐based meat alternatives has drawn considerable attention to tofu balls, which are primarily prepared from tofu. However, their performance, particularly as ready‐to‐cook ingredients, is generally considered inferior to that of traditional meatballs. After freezing and reheating, tofu balls often exhibit poor water retention, a coarse and crumbly texture, and a weak aroma, failing to replicate the texture and mouthfeel of conventional tofu balls. These deficiencies significantly reduce consumer sensory acceptance and pose a major challenge for the production of prepared frozen tofu balls (Zheng et al. 2020).

Currently, commercially available prepared tofu balls are mainly produced from tofu and soy protein isolate, supplemented with vegetables and high‐gluten flour. Their production typically includes mixing of raw materials, chopping, shaping, steaming, and freezing. However, the structural cohesiveness of crushed tofu is inherently limited. The addition of soy protein isolate effectively promotes intermolecular cross‐linking among components, thereby improving final product quality. Nevertheless, freezing can induce denaturation of soy proteins, disrupt protein–water hydration, and reduce system viscosity. Concurrently, protein subunits dissociate and reassociate, leading to exposure of internal hydrophobic groups. These changes ultimately result in a coarse and uneven texture in the final product (Chen et al. 2022). Such effects not only compromise water‐holding capacity (WHC) and textural properties but also weaken the gel strength and binding capacity of the soy protein matrix. Therefore, the sensory texture of reheated protein‐based products is adversely affected. Improving the freeze–thaw stability of soy protein is thus critical for maintaining textural consistency in tofu balls during preprocessing, cooking, and frozen reheating (M. Zhang et al. 2024).

In addition, volatile compounds in tofu and soy protein mainly include aldehydes, alcohols, ketones, furans, acids, and esters. These compounds exist in both free and bound forms (J. Zhang et al. 2021). Previous studies have shown that volatile compounds can bind to soy protein isolate, and these interactions are influenced by temperature‐induced changes in protein conformation (B. Zhang et al. 2022). Freezing weakens the covalent binding capacity of proteins toward aldehydes such as hexanal and octanal, and hydrogen bonding is also reduced. Therefore, non‐covalent interactions decrease, promoting the release of bound aldehydes (Bi et al. 2022). Furthermore, changes in protein conformation and aggregation state during freezing can induce rearrangement of flavor‐binding sites, thereby altering the overall flavor profile (Boland et al. 2006). These effects readily lead to the intensification of undesirable odors, such as beany notes, after reheating. The recovery of the pre‐freezing flavor profile is hindered, which ultimately reduces sensory acceptance. Therefore, the stability of soy protein under frozen conditions is essential for maintaining flavor consistency in tofu balls throughout preprocessing, cooking, and frozen reheating.

Accordingly, a composite matrix was developed using tofu and soy protein isolate as the primary ingredients, supplemented with vegetables, high‐gluten flour, vital wheat gluten, and transglutaminase to achieve desirable elasticity and chewiness. Molecular sensory science, flavoromics, texture analysis, and low‐field nuclear magnetic resonance (LF‐NMR) were applied to systematically examine changes in volatile compounds, moisture migration, textural properties, and sensory attributes across preprocessing, cooking, and frozen reheating stages. This study aims to clarify the stability of flavor and eating quality after frozen storage and reheating. The resulting insights provide a theoretical basis for developing ready‐to‐cook plant‐based products that preserve traditional tofu flavor while meeting modern nutritional requirements and offer practical guidance for improving processing technology and the market competitiveness of plant‐based meat alternatives.

2. Materials and Methods

2.1. Raw Materials and Chemicals

Tofu, spinach, shiitake mushrooms, and carrots were purchased from local markets (Jinzhou, China). Soy protein isolate was obtained from Shandong Yu Wang Co. Ltd. (Shandong, China); vital wheat gluten from Henan Midaner Trading Co. Ltd. (Henan, China); and transglutaminase from Henan Yuwanbang Food Technology Co. Ltd.(Henan, China). High‐gluten flour was provided by Haitian Flavouring & Food Co. Ltd. (Foshan, China). 2‐Methyl‐3‐heptanone (99.5%), used as an internal standard, was purchased from Macklin Biochemical Technology Co. Ltd. (Shanghai, China).

2.2. Sample Preparation

Process flow chart for tofu balls preparation is shown in Figure 1.

FIGURE 1.

Process flow chart for tofu balls preparation.

Preprocessed tofu balls (TB‐PP): First, 100 g of tofu, 85 g of soy protein isolate, and 15 g of vegetables were separately ground using a food processor. In parallel, 15 g of high‐gluten flour, 3 g of vital wheat gluten, and 0.5 g of transglutaminase were weighed and thoroughly mixed to obtain a dry blend. The ground ingredients were then sequentially transferred to a vacuum cutter‐mixer and homogenized. The resulting mixture was removed, manually shaped into 3‐cm spherical forms, placed in containers, and designated as TB‐PP.

Cooked Tofu Balls (TB‐C): The shaped tofu balls were steamed in a cooking pot for 1 h. They were then removed, allowed to cool, and reserved for subsequent analyses to obtain the cooked samples.

Frozen and Reheated Tofu Balls (TB‐FR): The cooled tofu balls were quickly transferred into a freezer to lower their core temperature to −18°C. The frozen tofu balls were directly put into a pot and steamed for 20 min to obtain frozen–reheated samples.

2.3. Sensory Evaluation

A sensory evaluation panel consisted of 30 nonsmoking members (15 males and 15 females) aged 20–30 years. All panelists were recruited from the College of Food Science and Engineering at Bohai University. The study was reviewed and approved by the Bohai University IRB under approval #BHU202605004, and informed consent was obtained from each subject prior to their participation in the study panel selection, training, and performance monitoring were conducted in accordance with ISO 8586: 2012. For evaluation, 100 g of each sample was placed in a clean, odorless, food‐grade container. Aroma and taste characteristics were assessed at room temperature. Panelists were instructed to repeatedly smell and describe the aroma and taste attributes within a specified time period. Through panel discussion, beany, beany odor, sweet‐fruity, steamed, and cereal‐like aromas were identified as representative of the overall aroma profile of tofu balls. A five‐point intensity scale was applied: 0 (not detectable), 1–4 (gradually increasing intensity), and 5 (strongest intensity). Mean scores for each sensory attribute from three replicates were calculated and used to generate a heatmap. The sensory descriptors and corresponding definitions applied in aroma evaluation are presented in Table 1.

TABLE 1.

Tofu balls flavor lexicon in descriptive analysis (Keast and Lau 2006; Lozano et al. 2007).

Flavor dictionary	Definition	Reference standard
Cooked beans	Roasted soybean aroma	A 0.01% (w/w) solution of (E,E)‐2,4‐heptadienal prepared in triacetin
Beany odor	Raw bean aroma	Fifty grams of soybeans heated in 400 mL of water for 15 min
Sweet‐fruity aroma	Sweet banana aroma	A 2% (v/v) solution containing isoamyl acetate and ethyl butyrate prepared in glycerol
Steamed	Steamed soybean aroma	Water boiled in an iron pot
Cereal‐like aroma	Cooked oatmeal aroma	Fifty grams of oatmeal combined with 500 mL of water

2.4. Identification and Quantitative Analysis of Volatile Components

Volatile compounds in the samples were analyzed using the GC–IMS FlavourSpec system manufactured by G.A.S. (Dortmund, Germany). For each analysis, 5 g of sample was accurately weighed into a 20 mL headspace vial and incubated at 40°C for 15 min. Volatile compounds were then extracted by automatic headspace sampling with an injection volume of 500 µL. The operating temperatures were set at 45°C for the IMS drift tube, 85°C for the injector needle, and 60°C for the chromatographic column (MXT‐5, 15 m × 0.53 mm, 1 µm). High‐purity nitrogen (99.999%) was used as the carrier gas. The flow rate program was set as follows: 2 mL/min for 0–2 min, increased from 2 to 10 mL/min over 2–10 min, further increased from 10 to 100 mL/min over 10–20 min, and finally increased from 100 to 150 mL/min over 20–25 min. The total analysis time was 25 min. Qualitative identification was carried out by matching the obtained spectra with the GC–IMS library database using the built‐in software. Linear retention indices (RIs) were calculated using a C₄–C₉ n‐ketone series as external reference standards.

For quantitative analysis, 5.00 g of sample was placed into a 20 mL headspace vial, and 5 µL of 2‐methyl‐3‐heptanone solution (0.8 mg/mL) was added as an internal standard. Quantification was performed using the internal standard method according to the following formula (Yu et al. 2023):

$$\mathrm{ms}=\mathrm{mi}\times \frac{A_{\mathrm{s}}}{A_{\mathrm{i}}}$$ (1)

where, ms is the content of the target volatile compound (mg/kg), mi is the amount of internal standard added (mg/kg), and A _s and A _i are the chromatographic peak areas of the target compound and the internal standard, respectively.

2.5. Calculation of OAV

Odor activity value (OAV) of each compound was calculated using the following equation (Ye et al. 2024):

$$\mathrm{OAV}=\frac{C}{\mathrm{OT}}$$ (2)

where, C is the concentration of the compound in the sample (mg/kg), and OT is its odor threshold in water (mg/kg) (Gemert 2011). Compounds with an OAV ≥ 1 were defined as aroma‐active, indicating perceptible olfactory contribution to the overall flavor profile. The relative contribution of each compound to overall flavor was considered proportional to its OAV.

2.6. Determination of Textural Properties

Samples were trimmed into uniform cubes measuring 2 cm × 2 cm × 2 cm. Hardness, gumminess, and chewiness were measured using a texture analyzer (TA‐XT Plus, Stable Microsy Stems, UK) equipped with a P/50 cylindrical probe. The test conditions were set as follows: pretest speed of 5.00 mm/s, test speed of 1.00 mm/s, and posttest speed of 5 mm/s. The trigger force was 5 g, and the target deformation was 10.0 mm. A two‐bite compression test was conducted, with each compression lasting 5 s.

2.7. Determination of WHC

A 5.0 g sample was accurately weighed into a 50 mL centrifuge tube pre‐lined with absorbent cotton at the bottom. The tube was centrifuged at 2000 × g for 10 min. After centrifugation, the sample was removed, and any attached absorbent cotton was carefully detached from the surface. The sample mass was recorded as W ₁. The sample was then comminuted and dried at 105°C for 4 h until a constant weight was reached, recorded as W ₂. WHC was calculated as follows:

$$\mathrm{WHC}\left(\%\right)=\frac{W_2}{W_1}\times 100$$ (3)

2.8. LF‐NMR Analysis

Moisture distribution and state in tofu balls were analyzed using LF‐NMR (Niumag PQ‐001, Niumag Electric Corporation, Shanghai, China). A 5 g sample was weighed and vertically positioned at the center of the magnetic field for measurement at 32°C. The free induction decay (FID) sequence was first applied to determine the central frequency and 90° pulse width (P1). Subsequently, the Carr–Purcell–Meiboom–Gill (CPMG) sequence was used to measure transverse relaxation time (T2). The parameters were set as follows: main frequency (SF1) = 21 MHz; offset frequency (O1) = 147.60225 kHz; 90° pulse width (P1) = 11.52 µs; 180° pulse width (P2) = 20.56 µs; number of data points (TD) = 250,024; sampling frequency (SW) = 100 and 250 kHz; repetition time (TW) = 1000 ms; echo time (TE) = 0.200 ms; number of echoes (NECH) = 5000; and number of scans (NS) = 4.

2.9. Color Measurement

Color parameters L ^* (lightness), a ^* (red‐green value), and b ^* (yellow‐blue value) of tofu balls were measured using a CR‐400 colorimeter (Minolta, Japan). The total color difference (ΔE) between samples at different processing stages was calculated using the following equation:

$$\mathrm{\Delta }E=\sqrt{{\left(\mathrm{\Delta }{L}^{\ast }\right)}^2{\left(\mathrm{\Delta }{a}^{\ast }\right)}^2{\left(\mathrm{\Delta }{b}^{\ast }\right)}^2}$$ (4)

where, ∆L ^*, ∆a ^*, and ∆b ^* represent the differences in the respective color parameters between two samples.

2.10. Statistical Analysis

All experiments were conducted in triplicate. Data are expressed as mean ± standard deviation. Statistical analysis was performed using SPSS software. Significant differences among sample means (p < 0.05) were evaluated by one‐way analysis of variance (ANOVA), followed by Duncan's multiple range test. Figures were generated using Origin 2024 and TBtools v.1.068. GC–IMS data were processed and analyzed using Laboratory Analytical Viewer (LAV) software.

3. Results and Discussion

3.1. Aroma Sensory Evaluation Results of Tofu Balls

As shown in the aroma profile heatmap (Figure 2), TB‐PP exhibited a low overall odor intensity, which was dominated by beany odor, resulting in reduced sensory acceptance. After cooking and frozen reheating, the overall odor intensity increased markedly. Beany, sweet‐fruity, steamed, and cereal‐like aromas were significantly enhanced, with beany, steamed, and cereal‐like notes becoming dominant in the aroma profiles of both cooked and reheated tofu balls. After frozen storage and reheating, the tofu balls largely restored the aroma characteristics of freshly cooked samples, indicating favorable stability of flavor sensory attributes.

FIGURE 2.

Aroma sensory heatmap of tofu balls during preprocessing, cooking, and frozen reheating. Color intensity was normalized on a scale from 1.00 (minimum) to 4.50 (maximum), indicating sensory intensity from weakest to strongest. Higher sensory intensity is represented by larger and darker circles.

Volatile compounds in soy‐based foods mainly comprise aldehydes, ketones, alcohols, acids, and esters. These compounds are generated through enzymatic catalysis of polyunsaturated fatty acids, such as linoleic acid and linolenic acid in soybeans, by lipoxygenase (LOX) and fatty acid hydroperoxide lyase (J. Zhou et al. 2024). Compounds including hexanal, hexanol, (E)‐2‐hexenal, 1‐octen‐3‐ol, pentanol, acetic acid, benzaldehyde, and (E,E)‐2,4‐decadienal are commonly associated with beany odors, as they contribute grassy, green, mushroom‐like, and pungent sensory notes (Lv et al. 2011). In contrast, compounds such as 3‐methylbutanal, (E)‐2‐octenal, nonanal, and (E)‐2‐nonenal, which impart chocolate, cucumber, citrus, and sweet‐fruity aromas, are typically classified as non‐beany aroma‐active compounds. Previous studies have demonstrated that volatile compounds in soy can bind to soy protein isolate (Kühn et al. 2006). During thermal processing, increased energy input induces conformational changes in proteins, including denaturation and aggregation. These changes modify the binding and release behavior of volatile compounds on protein matrices, thereby altering the concentration of free volatiles in the system (Sun et al. 2022). Therefore, the overall odor intensity increases after both initial cooking and frozen reheating. Freezing further weakens binding interactions between soy proteins and volatile compounds. Subsequent reheating promotes the conversion of bound flavor components into the free state, resulting in a more intense aroma in frozen–reheated tofu balls. In addition, supplementary ingredients such as spinach and carrots contain considerable amounts of limonene, (E,E)‐2,4‐decadienal, nonanal, and heptanal, which mainly contribute sweet‐fruity, floral, and citrus‐like aromas. These flavor compounds can interact hydrophobically with soy proteins, which limits their volatility in preprocessed samples. Thus, the aroma of TB‐PP is dominated by beany notes. Structural modifications of soy proteins induced by freezing and thermal treatment reduce available binding sites and weaken affinity for volatile compounds on the protein surface (Duppeti et al. 2023), promoting the release of previously bound volatiles. Therefore, both cooked and frozen–reheated samples exhibited higher overall aroma intensity. Moreover, thermal processing promotes fatty acid oxidation and glyceride hydrolysis in soy protein and carrots that can subsequently form ethyl acetate and ethyl butyrate (Solomando et al. 2021). These compounds typically exhibit low sensory thresholds and contribute pleasant fruity aromas to cooked samples.

3.2. Analysis of Volatile Compounds in Tofu Balls During Preprocessing, Cooking, and Frozen Reheating

In this study, volatile compounds in tofu balls at the preprocessing, cooking, and frozen‐reheating stages were qualitatively and quantitatively analyzed using GC‐IMS. As shown in Figure 3A, the reactant ion peak (RIP) was generated by ionization of water molecules in air by tritium radiation and was used as a reference representing the total population of ionizable ions. To reduce measurement variability caused by pressure and temperature fluctuations during IMS analysis, all ion drift times (Dt) in the GC‐IMS spectra were normalized to the RIP. By matching RI and Dt of signal peaks with the IMS database, a total of 47 volatile compounds were detected, of which 43 were identified from the IMS library. Among these, compounds such as 1‐octen‐3‐ol (A1), 3‐heptanol (A2), pentanol (A9), heptanal (A18), 3‐octanone (A25), and ethyl acetate (A35) were detected in both monomeric and dimeric forms. Although these ions exhibited similar retention times, differences in Dt and peak intensities were observed, which were mainly attributed to differences in chemical properties and concentrations of the volatile compounds.

FIGURE 3.

Results of HS‐GC‐IMS analysis of volatile compounds (A) Comparison of chromatograms. (B) Fingerprint profile of volatile compounds. (C) Classification and concentration changes of flavor substances, where I, II, III, IV, V, and VI represent alcohols, aldehydes, ketones, esters, other compounds, and undetected compounds, respectively.

In HS‐GC‐IMS analysis, preliminary separation was achieved by gas chromatography based on differences in interactions between volatile compounds and the stationary phase of the chromatographic column, resulting in different retention times for neutral molecules. The separated molecules then entered the IMS region and were ionized. When the ionization source was not saturated, prolonged reaction time increased the number of product ions. As reactant ions in the charge reservoir were depleted, the concentration of product ions no longer maintained a linear relationship with neutral molecule concentration. Under these conditions, monomeric cluster ions further reacted with neutral molecules to form protonated dimer cluster ions. Driven by periodic ion pulses under ambient pressure, cluster ions with different masses, collision cross‐sections, and charge numbers sequentially entered a linear drift electric field, acquired different drift velocities, and underwent secondary separation, which was detected as distinct Dt and peak intensities (Ma et al. 2024).

In the ion mobility difference spectra, regions with high compound concentrations are displayed in red, whereas regions with low concentrations are shown in blue (He et al. 2023). For tofu balls, the color intensity of spots corresponding to volatile compounds gradually shifted toward red across the preprocessing, cooking, and frozen‐reheating stages, indicating a progressive increase in volatile compound concentrations. This trend was consistent with the sensory evaluation results, which showed gradual enhancement of aroma intensity during processing.

Compared with two‐dimensional GC‐IMS spectra, fingerprint plots enable clearer visualization and comparison of aroma compound profiles among samples. The fingerprint plot shown in Figure 3B was generated by selecting characteristic ion peaks from Figure 3A using the GalleryPlot plugin in LAV software. As shown in Figure 3B and Table A1, 43 volatile compounds were identified in pretreated tofu balls, including 14 alcohols, 9 aldehydes, 5 ketones, and 9 esters, as well as 1 acid, 1 furan, and 1 pyrazine. These compounds reflect aroma components released from the food matrix into the air and contributing to overall flavor perception.

In soy‐based foods, compounds such as 1‐octen‐3‐ol, 3‐octanol, pentanol, hexanol, 3‐heptanol, 3‐methyl‐3‐buten‐1‐ol, (E)‐2‐heptenal, heptanal, pentanal, 2‐methyl‐2‐pentenal, (E)‐2‐pentenal, 3‐octanone, acetic acid, and 2‐pentylfuran are commonly classified as beany odor substances because they are associated with undesirable sensory attributes, including mushroom‐like, grassy, fat oxidation, earthy, and musty notes (Shi et al. 2022). In contrast, desirable aromas such as sweetness, caramel, creaminess, and fruitiness are mainly contributed by compounds including (E)‐2‐octenal, 2‐pentanone, 2,3‐pentanedione, 2‐butanone, and ethylpyrazine. These flavor components are primarily generated through enzymatic catalysis of polyunsaturated fatty acids, such as linoleic acid and linolenic acid in soybeans, by LOX and fatty acid hydroperoxide lyase (C. Zhang et al. 2020). Malty aromas contributed by butanol and 3‐methylbutanol are associated with degradation of branched‐chain amino acids, such as threonine and leucine, in soybeans (Feng et al. 2012). Sweet‐fruity, creamy, and caramel‐like aromas are mainly imparted by esters, including ethyl hexanoate, ethyl valerate, isobutyl isobutyrate, methyl butyrate, isobutyl acetate, ethyl acetate, ethyl butyrate, and γ‐butyrolactone. These esters are generated through enzymatic breakdown of hydroperoxides in soybeans, spinach, and carrots, mediated by hydroperoxide lyase, alcohol dehydrogenase, and acyltransferase. Phenylacetaldehyde, which has a hyacinth‐like aroma, is mainly derived from phenylalanine present in soybeans and carrots (Deng et al. 2025). A citrus aroma is imparted by linalool, which is produced in carrots through enzymatic conversion of geranyl pyrophosphate catalyzed by linalool synthase (P. Zhou et al. 2023). α‐Pinene, characterized by a pine‐like odor, is mainly synthesized in carrots through the mevalonic acid pathway.

As shown in Figure 3C, the overall composition of volatile compounds in the cooked and frozen–reheated samples remained consistent with that of the TB‐PP. In contrast, the concentrations of alcohols, aldehydes, ketones, and esters increased significantly. Notably, the highest aroma intensity was observed in tofu balls after frozen storage and reheating. This increase can be attributed to two main factors. First, Maillard reactions between abundant reducing sugars and free amino acids in the raw materials were initiated under high‐temperature conditions (Qiu et al. 2024). At the same time, thermal processing disrupted the cellular structures of spinach and carrots, promoting the release of a substantial amount of flavor precursors. These processes jointly facilitated the formation of volatile compounds. Second, heating weakened the binding interactions between volatile compounds and soy protein. This binding affinity was further modified by freezing, which caused a considerable proportion of bound volatiles to transition into the free state. Therefore, the overall concentration of free volatile compounds in the system increased.

3.3. Identification of Characteristic Volatile Compounds in Tofu Balls During Preprocessing, Cooking, and Frozen Reheating

To identify key volatile compounds responsible for differences among tofu balls subjected to preprocessing, cooking, and frozen reheating, partial least squares‐discriminant analysis (PLS‐DA) was performed. In this model, the 43 volatile compounds identified by GC‐IMS were used as variables, and processing stages were treated as categorical variables. As shown in the PCA score plot (Figure 4A), the first principal component (PC1) explained 76.94% of the total variance, and the second principal component (PC2) accounted for 13.20%. The cumulative contribution rate reached 90.14%, indicating good model fit and satisfactory discriminative capability. As shown in Figure 4B, 21 volatile compounds with significant differences were screened based on variable importance in projection (VIP) ≥ 1. These included beany odor compounds such as 1‐octen‐3‐ol, 3‐heptanol, hexanol, heptanal, 2‐methyl‐2‐pentenal, hexanal, 3‐octanone, and 2‐pentylfuran; cooked‐bean components represented by (E)‐2‐octenal; sweet‐creamy aroma compounds including 2,3‐butanediol and 2‐butanone; cereal‐like aroma contributors such as pentanol, propanol, and 3‐methylbutanol; and sweet‐fruity esters including isobutyl isobutyrate, methyl butyrate, and ethyl butyrate.

FIGURE 4.

PLS‐DA analysis of volatile compounds in preprocessed, cooked, and frozen–reheated tofu balls. (A) PCA score plot of volatile compounds. (B) Variable importance in projection (VIP) values of volatile compounds.

Previous studies have demonstrated that volatile compounds interact with soy proteins, polysaccharides, and other constituents in raw materials. Changes in their physicochemical states during different processing stages can affect the flavor expression of aroma‐active compounds (Yao and Zhao 2015). Therefore, OAVs of volatile compounds identified in tofu balls during preprocessing, cooking, and frozen reheating were further analyzed. As shown in Table 2, 37 aroma compounds exhibited OAV ≥ 1. The total OAV of beany odor compounds, including 1‐octen‐3‐ol, 3‐octanol, 2‐hexen‐1‐ol, pentanol, hexanol, 3‐heptanol, (E)‐2‐heptenal, heptanal, pentanal, 2‐methyl‐2‐pentenal, hexanal, 3‐octanone, and 2‐pentylfuran, reached 1369.19, 1506.43, and 2052.79 for the preprocessed, cooked, and frozen–reheated samples, respectively. The OAV of the cooked‐bean compound (E)‐2‐octenal was 36.66, 45.43, and 69.80 at the three stages. For the sweet‐creamy aroma compound 2,3‐pentanedione, OAVs of 6.02, 7.87, and 11.12 were obtained. The combined OAV of cereal‐like aroma compounds, 1‐butanol and 3‐methylbutanol, reached 84.51, 87.16, and 103.46, respectively. Compared with TB‐PP, the total OAV of beany odor compounds increased by 10.03% after cooking and by 49.93% after frozen reheating. The OAV of cooked‐bean compounds increased by 23.92% and 90.40%, respectively. Increases of 21.72% and 84.72% were observed for sweet‐creamy aroma compounds, whereas the total OAV of cereal‐like aroma compounds increased by 3.13% and 22.42%. The OAV of sweet‐fruity aroma compounds increased by 13.43% and 31.29% across the same stages. After cooking and frozen reheating, contributions of beany, sweet, and cereal‐like aroma compounds to the overall flavor profile increased more significant than those of beany off‐odor components. These positive aroma compounds exhibited high overall contribution levels and were identified as key odor‐active constituents. Consistent with sensory evaluation results, lower perceived intensity of beany off‐odor was observed in cooked and reheated tofu balls, and the aroma profile was dominated by beany, sweet, and cereal‐like notes. Although steamed aroma showed relatively high sensory intensity, corresponding volatile compounds were not identified. This may be attributed to incomplete capture during headspace sampling or limitations in instrumental detection and identification.

TABLE 2.

Odor activity values (OAVs) of major aroma‐active compounds in tofu balls during preprocessing, cooking, and frozen reheating.

No.	Compound	Odorant description	OT (mg/kg)	TB‐PP	TB‐C	TB‐FR
	Alcohol
A 1 (M)	1‐Octen‐3‐ol (M)	Mushroom‐like	0.0015	473.5351	667.2785	825.8923
A 1 (D)	1‐Octen‐3‐ol (D)	Mushroom‐like	0.0015	182.8102	205.8570	319.5451
A 2 (M)	3‐Heptanol (M)	Grassy	0.24	4.2904	5.0934	5.7214
A 2(D)	3‐Heptanol (D)	Grassy	0.24	3.9725	3.2217	4.4757
A 3	Linalool	Citrusy	0.00022	673.0150	361.9441	283.7112
A 4	3‐octanol	Mushroom‐like	0.078	4.1679	3.3540	3.2219
A 5	2‐Hexanol	Grassy	1.5082	1.2686	1.4242	1.4812
A 6	2,3‐Butanediol	Creamy, sweet	100	0.0057	0.0031	0.0034
A 7	2‐Hexen‐1‐ol	Grassy, green	0.005	49.1476	48.4467	62.9647
A 8	Hexanol	Grassy, slightly fatty	0.0056	109.7267	89.0176	120.3827
A 9 (M)	Pentanol (M)	Grassy	0.1502	4.5220	4.8691	5.7708
A 9(D)	Pentanol (D)	Grassy	0.1502	2.9728	2.5448	3.6538
A 10	Propanol	Alcoholic	8.5056	0.0158	0.0257	0.0075
A 11	Butanol	Sweet, wine‐like	0.4592	1.0177	1.4893	1.9109
A 12	Borneol	Camphoraceous, minty	0.18	1.1653	1.3129	1.5841
A 13	3‐Methylbutanol	Malty, wine‐like	0.004	83.4916	155.6676	251.5458
A 14	3‐Methyl‐3‐buten‐1‐ol	Grassy	0.5471	0.4304	0.4953	0.4432
	Aldehyde
A 15	(E)‐2‐octenal	Sweet	0.003	36.6566	95.4336	159.8026
A 16	Phenylacetaldehyde	Floral	0.0063	22.4496	7.5114	7.3896
A 17	(E)‐2‐Heptenal	Grassy	0.04	4.5518	8.8808	13.0347
A 18(M)	Heptanal (M)	Grassy	0.0028	34.1266	74.1827	159.1884
A 18(D)	Heptanal (D)	Grassy	0.0028	96.7415	163.5269	221.5228
A 19	2‐Methyl‐2‐pentenal	Grassy	0.29	2.3034	1.3001	0.9327
A 20	Pentanal	Grassy	0.012	3.0294	6.1216	19.1120
A 21	(E)‐2‐Pentenal	Grassy	0.98	0.1892	0.1871	0.2702
A 22	Butanal	Malty	0.002	106.4293	138.8537	152.9162
A 23	Hexanal	Intensely grassy	0.005	113.1484	158.4550	193.6364
A 24	Methional	Sulfurous	0.00045	215.6507	75.6952	90.9119
	Ketone
A 25(M)	3‐Octanone (M)	Mushroom‐like	0.0214	11.2520	12.6174	13.0988
A 25(D)	3‐Octanone (D)	Mushroom‐like	0.0214	39.7208	32.8002	39.4857
A 26	2‐Pentanone	Sweet	1.38	0.1090	0.1549	0.2539
A 27	2,3‐Pentanedione	Creamy, caramel‐like	0.03	6.0223	7.8678	11.1215
A 28	2‐Butanone	Sweet	35.4002	0.0107	0.0139	0.0220
	Ester
A 29	Hexyl acetate	Tropical fruity	0.115	2.1184	2.7476	3.6662
A 30	Ethyl pentanoate	Apple‐like, sweet fruity	0.0058	62.5592	78.5257	93.9086
A 31	Isobutyl isobutyrate	Fruity	3.9	0.1342	0.1783	0.2218
A 32	Methyl methacrylate	Pungent, plastic‐like	0.05	0.9763	1.3860	3.0090
A 33	Methyl butyrate	Apple, pineapple	0.8	0.2319	0.3432	0.5613
A 34	Isobutyl acetate	Banana‐like	0.00002	12,438.2465	13,366.8300	14,835.7424
A 35(M)	Ethyl acetate (M)	Fruity	0.005	31.0871	54.2154	59.3152
A 35(D)	Ethyl acetate (D)	Fruity	0.005	18.7608	22.6880	25.9050
A 36	Ethyl butyrate	Sweet fruity	0.0009	1036.2141	1888.4089	2821.8299
A 37	γ‐Butyrolactone	Creamy, caramel‐like	1	0.2275	0.1034	0.0830
	Other compounds
A 38	Acetic acid	Pungent, sour	99	0.0023	0.0032	0.0031
A 39	Ethylpyrazine	Nutty, roasted coffee/bread‐like	4	0.0175	0.0047	0.0056
A 40	2‐Pentylfuran	Beany odor	0.0058	125.3956	100.1938	142.3886
A 41	1,4‐Cineole	Cooling, medicinal	1	0.1529	0.0765	0.0634
A 42	Alpha‐pinene	Piney, resinous	0.041	11.6928	10.0781	11.5407
A 43	Ortho‐guaiacol	Smoky, burnt wood	0.0016	47.4780	38.8464	48.7017

Key differential compounds responsible for flavor changes in tofu balls across preprocessing, cooking, and frozen reheating were further identified, as shown in Figure 5, by selecting volatile compounds with both VIP ≥ 1 and OAV ≥ 1. Thirteen key volatile compounds were identified, including six alcohols, four aldehydes, one ketone, one ester, and one furan. Key beany odor compounds identified across the three stages are listed in Table 3 and include 1‐octen‐3‐ol, 3‐heptanol, 2‐hexanol, hexanol, pentanol, heptanal, 2‐methyl‐2‐pentenal, hexanal, 3‐octanone, and 2‐pentylfuran. The total OAV of these compounds reached 1186.04, 1501.67, and 2033.83 for preprocessed, cooked, and frozen–reheated samples, respectively. The key contributor to cereal‐like aroma was 3‐methylbutanol, with OAVs of 83.49, 155.67, and 251.55. The key cooked‐bean component was (E)‐2‐octenal, with OAVs of 36.67, 95.43, and 159.80. Ethyl butyrate was identified as the principal contributor to sweet‐fruity aroma, with OAVs of 1036.21, 1888.41, and 2821.83 at the corresponding stages. Proportional changes in key aroma categories are shown in Figure 6. Beany off‐odor compounds accounted for the highest proportion, 50.63%, among key components in TB‐PP. Their proportion decreased to 41.24% and 38.61% after cooking and frozen reheating, respectively. In contrast, cereal‐like aroma compounds accounted for 3.56%, 4.28%, and 4.78% across the three stages. Cooked‐bean compounds increased to 1.56%, 2.62%, and 3.03%, whereas sweet‐fruity aroma compounds showed a significant increase, reaching 44.24%, 51.86%, and 53.58%, respectively. Therefore, compounds such as 1‐octen‐3‐ol, 3‐heptanol, 2‐hexanol, hexanol, pentanol, heptanal, 2‐methyl‐2‐pentenal, hexanal, 3‐octanone, and 2‐pentylfuran were identified as key contributors to beany off‐flavor in TB‐PP. In contrast, (E)‐2‐octenal, ethyl butyrate, and 3‐methylbutanol were identified as major contributors to beany, sweet‐fruity, and cereal‐like aromas in cooked and frozen–reheated tofu balls.

FIGURE 5.

Key aroma‐active compounds in tofu balls during preprocessing, cooking, and frozen reheating screened using dual criteria of VIP ≥ 1 and OAV ≥ 1.

TABLE 3.

Key aroma‐active compounds in tofu balls during preprocessing, cooking, and frozen reheating screened using dual criteria of VIP ≥ 1 and OAV ≥ 1.

No.	Compound	TB‐PP	TB‐C	TB‐FR
	Alcohol
A 1 (M)	1‐Octen‐3‐ol (M)	473.5351	667.2785	825.8923
A 1 (D)	1‐Octen‐3‐ol (D)	182.8102	205.8570	319.5451
A 2 (M)	3‐Heptanol (M)	4.2904	5.0934	5.7214
A 5	2‐Hexanol	1.2686	1.4242	1.4812
A8	Hexanol	109.7267	89.0176	120.3827
A 9(D)	Pentanol (D)	2.9728	2.5448	3.6538
A 13	3‐Methylbutanol	83.4916	155.6676	251.5458
	Aldehyde
A 15	(E)‐2‐Octenal	36.6566	95.4336	159.8026
A 18(M)	Heptanal (M)	34.1266	74.1827	159.1884
A 18(D)	Heptanal (D)	96.7415	163.5269	221.5228
A 19	2‐Methyl‐2‐pentenal	2.3034	1.3001	0.9327
A 23	Hexanal	113.1484	158.4550	193.6364
	Ketone
A 25(D)	3‐Octanone (D)	39.7208	32.8002	39.4857
	Esters
A36	Ethyl butyrate	1036.2141	1888.4089	2821.8299
	Furan
A40	2‐Pentylfuran	125.3956	100.1938	142.3886

FIGURE 6.

Proportions of key aroma compounds in tofu balls during preprocessing, cooking, and frozen reheating screened using criteria of VIP ≥ 1 and OAV ≥ 1.

3.4. Analysis of Physical Properties of Tofu Balls During Preprocessing, Cooking, and Frozen Reheating

Stability of textural properties is critical for achieving desirable eating quality in tofu balls after cooking and frozen reheating. Therefore, key textural parameters, including hardness, gumminess, and chewiness, were evaluated. As shown in Table 4, the hardness, gumminess, and chewiness of TB‐PP were 251.14, 85.35, and 28.67 N, respectively. After cooking, these parameters increased significantly to 1315.10, 581.45, and 314.70 N. This notable increase in firmness can be attributed to incomplete protein denaturation during preprocessing, which limits the formation of a continuous gel network. In contrast, thermal treatment during cooking induces protein denaturation and aggregation, resulting in the development of a more rigid and cohesive structure (Wang et al. 2025). After frozen storage and reheating, hardness, gumminess, and chewiness further increased to 1978.19, 1065.55, and 707.44 N, respectively, indicating continued enhancement of textural properties. Under freezing conditions, sulfhydryl groups from different polypeptide chains are brought into closer proximity, promoting covalent cross‐linking (Hong et al. 2021). Subsequent reheating further facilitates disulfide bond formation and strengthens hydrophobic interactions among protein molecules (Peyrano et al. 2022), yielding a more compact protein network and firmer texture.

TABLE 4.

Textural properties of tofu balls during preprocessing, cooking, and frozen reheating.

Sample	Hardness (N)	Gumminess (N)	Chewiness (N)
TB‐PP	251.14 ± 17.38c	85.35 ± 6.75c	28.67 ± 3.51c
TB‐C	1315.10 ± 107.70b	581.45 ± 45.57b	314.70 ± 29.68b
TB‐FR	1978.19 ± 188.75a	1065.55 ± 114.74a	707.44 ± 79.50a

Note: Different lowercase letters indicate a statistically significant difference (p < 0.05).

WHC, reflecting water retention ability and the compactness of the gel network, was used as an indicator of processing performance. As shown in Table 5, WHC increased progressively during preprocessing, cooking, and frozen reheating. Specifically, WHC increased from 65.46% at the preprocessed stage to 73.44% after cooking and further to 78.58% after frozen storage and reheating.

TABLE 5.

WHC of tofu balls during preprocessing, cooking, and frozen reheating.

Sample	WHC (%)
TB‐PP	65.46 ± 5.78b
TB‐C	73.44 ± 4.11ab
TB‐FR	78.58 ± 0.36a

Note: Different lowercase letters indicate a statistically significant difference (p < 0.05).

Moisture distribution during different processing stages was further characterized using LF‐NMR based on transverse relaxation time (T ₂). According to established classifications, T ₂ values of 0–0.5 ms correspond to strongly bound water, 0.5–10 ms to weakly bound water, 10–100 ms to immobilized water, and values exceeding 100 ms to free water (Hu et al. 2025).

As shown in Table A2 and Figure 7, four T ₂ relaxation peaks were observed in TB‐PP. Strongly bound water exhibited a relaxation time of 0.37 ms with a proportion of 1.28%. Weakly bound water showed a relaxation time of 2.41 ms and accounted for 5.27%. Immobilized water dominated the moisture distribution, with a relaxation time of 41.50 ms and a proportion of 89.13%. Free water exhibited a relaxation time of 444.57 ms and accounted for 4.32%. These results indicate that moisture was mainly present as immobilized water, reflecting a relatively intact structure with limited free water.

FIGURE 7.

Moisture distribution in tofu balls during preprocessing, cooking, and frozen reheating.

After cooking, three T ₂ relaxation peaks were observed. Strongly bound water exhibited a relaxation time of 0.71 ms and accounted for 3.17%. Weakly bound water showed a relaxation time of 16.83 ms with a proportion of 18.85%, representing an almost four‐fold increase compared with the preprocessed stage. Immobilized water exhibited a relaxation time of 109.70 ms and accounted for 77.98%, whereas the free water peak was no longer detected. Heating induces protein denaturation and network rearrangement in soy protein, characterized by reduced α‐helix content and increased β‐sheet structures (Ge et al. 2021). This structural transition promotes formation of a more compact gel network, explaining the increased proportion of bound water after cooking.

Following frozen storage and reheating, moisture distribution was further altered, with only two T ₂ relaxation peaks observed. Strongly bound water exhibited a relaxation time of 0.77 ms and accounted for 3.11%, whereas weakly bound water exhibited a relaxation time of 102.34 ms and accounted for 96.89%. Peaks corresponding to immobilized and free water were no longer detected. Protein denaturation during freezing causes free water to crystallize as ice, detaching it from the macromolecular protein matrix. Subsequently, a fraction of bound water associated with protein side chains also freezes (Lee et al. 2022), leading to irreversible changes in protein–water interactions. As ice crystals grow, proteins become increasingly concentrated. Upon reheating, structural reorganization occurs, resulting in a more homogeneous distribution of bound water. Although extensive protein denaturation and aggregation trap water within a more rigid network, these changes indicate that juiciness and a tender mouthfeel can be retained in tofu balls after freezing and reheating.

3.5. Colorimetric Analysis of Tofu Balls During Preprocessing, Cooking, and Frozen Reheating

As shown in Figure 8, the surface color of tofu balls gradually deepened during preprocessing, cooking, and frozen reheating, ultimately developing a brownish‐yellow appearance. Colorimetric analysis results are presented in Table 6. For TB‐PP, L ^*, a ^*, and b ^* values were 71.75, 4.39, and 26.17, respectively. After cooking, the L ^* value decreased significantly to 64.96, whereas a ^* and b ^* values increased to 7.14 and 31.77, indicating color darkening accompanied by increased redness and yellowness. During cooking, elevated temperature intensifies nonenzymatic browning, primarily through Maillard reactions between amino acids and reducing sugars, resulting in formation of brown‐colored compounds (Xu et al. 2016). Therefore, a marked decrease in L ^* was observed. Compared with cooked samples, frozen–reheated tofu balls exhibited further decrease in L ^*, a ^*, and b ^* values to 63.62, 6.37, and 31.62, respectively, indicating additional darkening, reduced redness, and a slight decrease in yellowness. During freezing, protein denaturation and unfolding expose pigment groups (Gao et al. 2023), while moisture reduction concentrates these pigments. Subsequent reheating accelerates pigment migration and redistribution, leading to a lower L ^* value, darker color, and an overall desirable visual appearance.

FIGURE 8.

Surface color of tofu balls during preprocessing, cooking, and frozen reheating.

TABLE 6.

Colorimetric parameters of tofu balls during preprocessing, cooking, and frozen reheating.

Parameter	TB‐PP	TB‐C	TB‐FR
L *	71.75 ± 0.91a	64.96 ± 3.11b	63.62 ± 2.15b
a *	4.39 ± 0.58a	7.14 ± 1.32b	6.37 ± 0.25b
b *	26.17 ± 0.44a	31.77 ± 1.71b	31.62 ± 0.39b
△E*	33.26 ± 0.93a	42.41 ± 3.64b	43.18 ± 1.42b

Note: Different lowercase letters indicate a statistically significant difference (p < 0.05).

3.6. Analysis of Key Factors Affecting the Quality Variation of Tofu Balls

To further identify the key factors responsible for differences in eating quality among preprocessed, cooked, and frozen–reheated tofu balls, principal component analysis (PCA) was conducted. As illustrated in Figure 9, the first principal component (PC1) explained 68.9% of the variance, while the second principal component (PC2) accounted for 20.0%. The cumulative contribution rate reached 88.9%, indicating that these two components adequately captured the contributions of various indicators to the eating quality of tofu balls. Specifically, TB‐PP was positioned on the opposite side of the vertical axis relative to TB‐C and TB‐FR, suggesting that significant changes in eating quality occurred after cooking and frozen reheating. The closer clustering of sample points for TB‐C and TB‐FR indicated that the magnitude of change diminished during these subsequent stages. Along PC1, relatively high positive loadings were observed for cooked beans, beany odor, sweet‐fruity aroma, steamed, cereal‐like aroma, 1‐octen‐3‐ol, 3‐methylbutanol, (E)‐2‐octenal, heptanal, hexanal, ethyl butyrate, hardness, gumminess, chewiness, WHC, and b ^* value, whereas L ^* value exhibited a high negative loading. Along PC2, high positive loadings were displayed by 3‐heptanol, hexanol, pentanol, 3‐octanone, and 2‐pentylfuran, while a ^* value showed a high negative loading. A larger absolute loading value indicated a stronger correlation between that indicator and the principal component. Therefore, the aforementioned indicators were regarded as key characteristic measures for distinguishing the eating quality of preprocessed, cooked, and frozen‐reheated tofu balls.

FIGURE 9.

PCA plots for various indicators of preprocessed, cooked, and frozen‐reheated tofu balls.

4. Conclusions

In this study, changes in key quality attributes of tofu balls were systematically investigated across preprocessing, cooking, and frozen reheating stages. A total of 43 volatile compounds were identified, comprising 14 alcohols, 9 aldehydes, 5 ketones, and 9 esters, as well as 1 acid, 1 furan, and 1 pyrazine. The flavor profile of TB‐PP was dominated by beany odors, with major contributors identified as 1‐octen‐3‐ol, 3‐heptanol, 2‐hexanol, hexanol, pentanol, heptanal, 2‐methyl‐2‐pentenal, hexanal, 3‐octanone, and 2‐pentylfuran. In contrast, cooked and frozen–reheated tofu balls exhibited dominant beany, sweet‐fruity, and cereal‐like aromas, with (E)‐2‐octenal, ethyl butyrate, and 3‐methylbutanol identified as key contributors. After frozen reheating, textural properties, including hardness and gumminess, were substantially improved, with an increase in WHC reaching a maximum value of 78.58%. Higher a ^* and b ^* values were observed in cooked and frozen–reheated tofu balls, resulting in a brownish‐yellow appearance. Cooking was identified as the primary stage responsible for color development, whereas frozen reheating had no significant additional effect on final color. Overall, tofu balls subjected to frozen storage and reheating effectively preserved their original flavor characteristics while exhibiting improved eating quality. This study provides a theoretical basis for optimizing tofu ball processing, mitigating flavor defects, and improving overall product quality. It also offers practical guidance for addressing quality limitations of traditional tofu balls and supporting their industrial‐scale production.

Author Contributions

Bowen Zhang: conceptualization, methodology, data curation, formal analysis, software. Zhengqi Hao: conceptualization, writing – review and editing, formal analysis, visualization. Xiaoxi Yu: software, investigation. Xin Wen: software, investigation. Fusheng Song: investigation. He Liu: conceptualization, resources, supervision, writing – review and editing, funding acquisition.

Conflicts of Interest

The authors declare no conflicts of interest.

Ethics Statement

Ethical approval was not required for this research.

TABLE A1.

Identification of volatile compounds in tofu balls during preprocessing, cooking, and frozen‐reheating.

No.	Compound	CAS#	Formula	MWa	RIb	Rt (s)c	Dt (a.u.)d	Identification method
	Alcohol
A 1 (M)	1‐Octen‐3‐ol (M)	3391‐86‐4	C8H16O	128.2	994.1	580.726	1.15681	RI, DT
A 1 (D)	1‐Octen‐3‐ol (D)	3391‐86‐4	C8H16O	128.2	994.6	581.782	1.59494	RI, DT
A 2 (M)	3‐Heptanol (M)	589‐82‐2	C7H16O	116.2	887.7	381.428	1.32782	RI, DT
A 2 (D)	3‐Heptanol (D)	589‐82‐2	C7H16O	116.2	881.7	373.343	1.64555	RI, DT
A 3	Linalool	78‐70‐6	C10H18O	154.3	1115.3	813.095	1.21322	RI, DT
A 4	3‐Octanol	589‐98‐0	C8H18O	130.2	984.6	559.31	1.40226	RI, DT
A 5	2‐Hexanol	626‐93‐7	C6H14O	102.2	956.7	500.69	1.28697	RI, DT
A 6	2,3‐Butanediol	513‐85‐9	C4H10O2	90.1	819.7	298.59	1.36414	RI, DT
A 7	2‐Hexen‐1‐ol	2305‐21‐7	C6H12O	100.2	863.3	349.423	1.17894	RI, DT
A 8	Hexanol	111‐27‐3	C6H14O	102.2	899.8	399.621	1.63011	RI, DT
A 9 (M)	Pentanol (M)	71‐4‐10	C5H12O	88.1	776.5	255.247	1.25452	RI, DT
A 9(D)	Pentanol (D)	71‐41‐0	C5H12O	88.1	779.4	258.04	1.51769	RI, DT
A 10	Propanol	71‐23‐8	C3H8O	60.1	558	128.753	1.11777	RI, DT
A 11	Butanol	71‐36‐3	C4H10O	74.1	690.1	184.095	1.18239	RI, DT
A 12	Borneol	507‐70‐0	C10H18O	154.3	1151.8	899.177	1.21087	RI, DT
A 13	3‐Methylbutanol	123‐51‐3	C5H12O	88.1	743.4	225.174	1.24642	RI, DT
A 14	3‐Methyl‐3‐buten‐1‐ol	763‐32‐6	C5H10O	86.1	730	214.048	1.16342	RI, DT
	Aldehyde
A 15	(E)‐2‐Octenal	2548‐87‐0	C8H14O	126.2	1070.9	719.619	1.33285	RI, DT
A 16	Phenylacetaldehyde	122‐78‐1	C8H8O	120.2	1057.1	692.686	1.25661	RI, DT
A 17	(E)‐2‐Heptenal	18829‐55‐5	C7H12O	112.2	967.2	521.915	1.25156	RI, DT
A 18(M)	Heptanal (M)	111‐71‐7	C7H14O	114.2	911.3	418.15	1.69457	RI, DT
A 18(D)	Heptanal (D)	111‐71‐7	C7H14O	114.2	912.3	419.835	1.34053	RI, DT
A 19	2‐Methyl‐2‐pentenal	623‐36‐9	C6H10O	98.1	821.9	300.952	1.15896	RI, DT
A 20	Pentanal	110‐62‐3	C5H10O	86.1	716.5	203.398	1.42428	RI, DT
A 21	(E)‐2‐Pentenal	1576‐87‐0	C5H8O	84.1	766.8	246.032	1.10532	RI, DT
A 22	Butanal	123‐72‐8	C4H8O	72.1	597.7	143.301	1.10354	RI, DT
A 23	Hexanal	66‐25‐1	C6H12O	100.2	813.2	291.709	1.26284	RI, DT
A 24	Methional	3268‐49‐3	C4H8OS	104.2	914.9	424.214	1.09088	RI, DT
	Ketone
A 25(M)	3‐Octanone (M)	106‐68‐3	C8H16O	128.2	997.3	587.592	1.31033	RI, DT
A 25(D)	3‐Octanone (D)	106‐68‐3	C8H16O	128.2	996.2	585.479	1.71185	RI, DT
A 26	2‐Pentanone	107‐87‐9	C5H10O	86.1	687.8	182.669	1.40234	RI, DT
A 27	2,3‐Pentanedione	600‐14‐6	C5H8O2	100.1	631.1	156.804	1.29266	RI, DT
A 28	2‐Butanone	78‐93‐3	C4H8O	72.1	623.9	153.761	1.2482	RI, DT
	Ester
A 29	Hexyl acetate	142‐92‐7	C8H16O2	144.2	1020.4	626.144	1.41307	RI, DT
A 30	Ethyl pentanoate	539‐82‐2	C7H14O2	130.2	902.4	403.663	1.26427	RI, DT
A 31	Isobutyl isobutyrate	540‐42‐1	C7H14O2	130.2	852.1	335.597	1.27094	RI, DT
A 32	Methyl methacrylate	80‐62‐6	C5H8O2	100.1	717.3	204.064	1.38871	RI, DT
A 33	Methyl butyrate	592‐84‐7	C5H10O2	102.1	717.1	203.874	1.20077	RI, DT
A 34	Isobutyl acetate	97‐62‐1	C6H12O2	116.2	738.4	220.99	1.18713	RI, DT
A 35(M)	Ethyl acetate (M)	141‐78‐6	C4H8O2	88.1	645.5	162.985	1.09939	RI, DT
A 35(D)	Ethyl acetate (D)	141‐78‐6	C4H8O2	88.1	646.1	163.27	1.33535	RI, DT
A 36	Ethyl butyrate	105‐54‐4	C6H12O2	116.2	807.5	285.795	1.5606	RI, DT
A 37	γ‐Butyrolactone	96‐48‐0	C4H6O2	86.1	929.3	449.145	1.08453	RI, DT
	Other compounds
A 38	Acetic acid	64‐19‐7	C2H4O2	60.1	621.8	152.906	1.1569	RI, DT
A 39	Ethylpyrazine	13925‐00‐3	C6H8N2	108.1	934.4	458.241	1.51301	RI, DT
A 40	2‐Pentylfuran	3777‐69‐3	C9H14O	138.2	1004.1	598.682	1.25128	RI, DT
A 41	1,4‐Cineole	470‐67‐7	C10H18O	154.3	1020.7	626.672	1.30939	RI, DT
A 42	Alpha‐pinene	80‐56‐8	C10H16	136.2	934.5	458.578	1.28515	RI, DT
A 43	Ortho‐guaiacol	90‐05‐1	C7H8O2	124.1	1085.5	749.194	1.11704	RI, DT

Note: (a) Molecular weight, (b) linear retention index obtained using n‑ketones C₄–C₉ as external standards, (c) GC column retention time, and (d) drift time.

TABLE A2.

Water state and distribution in tofu balls during preprocessing, cooking, and frozen‐reheating.

Sample remarks	TB‐PP	TB‐C	TB‐FR
T 2b(ms)	0.37 ± 0.23b	0.71 ± 0.15a	0.77 ± 0.05a
A 2b	217.18 ± 18.28b	566.07 ± 30.06a	552.19 ± 24.58a
M 2b(%)	1.28 ± 1.06b	3.17 ± 0.16a	3.11 ± 0.12a
T 21(ms)	2.41 ± 0.67c	16.83 ± 0.00a	102.34 ± 0.00b
A 21	890.28 ± 92.10c	3366.87 ± 33.42a	17,188.04 ± 1.63b
M 21(%)	5.27 ± 0.55c	18.85 ± 0.11a	96.89 ± 0.16b
T 22(ms)	41.50 ± 0b	109.69 ± 0a	0
A 22	15,066.543 ± 206.49a	13,924.89 ± 14.15b	0
M 22(%)	89.13 ± 1.44a	77.98 ± 0.24b	0
T 23(ms)	444.57 ± 77.64a	0	0
A 23	731.41 ± 160.19a	0	0
T 23(%)	4.3 ± 0.93a	0	0

Note: Different lowercase letters indicate a statistically significant difference (p < 0.05).

Data Availability Statement

The data that this study will be made available upon request.