Sodium carboxymethylcellulose–sodium caseinate complex hydrocolloid enhances the oil barrier properties of batter‐coated deep‐fried pork

Abstract

BACKGROUND

The rising prevalence of chronic diseases has driven consumer demand for low‐fat deep‐fried foods. Understanding the oil barrier mechanism of batter‐coated deep‐fried pork is essential for developing healthier products.

RESULTS

This study evaluated the effects of sodium carboxymethylcellulose (CN), sodium caseinate (SC), and their complex (CN‐SC) on batter properties and the quality of batter‐coated deep‐fried pork. CN‐SC (0.6%) increased batter viscosity, which consequently enhanced the batter pick‐up rate to 56.76%. This higher pick‐up rate led to a more uniform crust formation during frying. As a result, compared to the control, fat content decreased by 5.75% in the crust and 4.40% in the meat, while moisture content increased by 16.27% and 17.35%, respectively. The higher moisture content, in turn, reduced chewiness by up to 45%. Furthermore, the CN‐SC complex increased bound water by 1.5% and immobilized water by 4.2%, while reducing free water by 5.8%. This reduced the evaporable water during frying, thereby suppressing pore formation and the subsequent water–oil displacement reaction. Consistently, microstructural observation showed fewer holes and cracks in the crust, confirming that the CN‐SC‐induced three‐dimensional network effectively locked water and blocked oil penetration.

CONCLUSIONS

The CN‐SC complex outperforms single hydrocolloid in reducing oil penetration. The innovation lies in its dual functionality: enhancing batter viscosity and shifting free water to bound/immobilized water, thereby suppressing water–oil displacement reaction. Consequently, 0.6% CN‐SC achieves a ‘low fat–juicy’ balance, solving the dryness issue of low‐fat fried foods. © 2026 Society of Chemical Industry.

INTRODUCTION

Deep‐frying, as a traditional and widely used thermal processing method for food, is capable of imparting distinct flavor, texture, and taste to food. ¹ Among the many deep‐fried delicacies, batter‐coated deep‐fried pork is undoubtedly a representative traditional cuisine due to its unique taste and rich nutritional value. However, the frying process usually involves high oil penetration, and excessive consumption of deep‐fried foods can lead to obesity, cancer, and cardiovascular disease. ² At present, due to the growing demand for a healthy diet, it is particularly urgent for the food industry to produce deep‐fried foods with low fat content. ³ The development of low‐fat‐content deep‐fried foods is not only a need of the market but also a direction of innovation in the food industry.

Numerous studies have shown that innovations in processing technology and optimization of formulations serve as effective means to reduce the fat content of deep‐fried foods.4, 5, 6 Oil absorption during frying is primarily governed by two key mechanisms: the water replacement mechanism and the cooling phase effect. ⁷ The water replacement mechanism, also known as water–oil displacement, explains that the evaporation of water inside the food during frying creates pores. ⁸ These pores act as channels for oil penetration. The cooling phase effect means that when the food is removed from the hot oil after frying, the oil adsorbed on the surface of the food enters the food through the pores due to the ‘vacuum effect’. ⁹ In response to both theories, batter is an effective way to reduce oil penetration. ⁸ The batter acts as a protective layer on the outside of deep‐fried foods and consists mainly of a starch matrix, as well as hydrocolloids, bulking agents, salt, etc. ² While prior research has explored batter formulations to mitigate oil uptake, ¹⁰ most studies focus on single hydrocolloids or physical modifications, neglecting the potential of complex hydrocolloid systems. Physical modification of the batter formulation refers to the technical means of optimizing the properties of batter by physical means such as mechanical treatment or formulation adjustment.11, 12

The viscosity of batter allows it to form a uniform coating on the surface of the ingredients, preventing direct oil contact with the food surface, thereby enhancing the oil barrier performance. ¹³ Secondly, a high‐viscosity batter is more stable and less prone to detachment during the frying process, with a complete barrier layer to further prevent oil penetration. ¹⁴ Moreover, such high‐viscosity batter can better lock water, reduce water evaporation, and indirectly reduce oil penetration. ²

Hydrocolloids, acting as a type of thickener that can regulate viscosity, have extensive applications in the food industry. ¹⁵ Their hydroxyl‐rich structures enhance water binding, forming viscous colloidal solutions,16, 17 which will likely help batter to adhere better to the surface of a fried substance, forming a uniform and dense coating to block oil penetration.

Batter viscosity is a critical factor in oil barrier performance, yet the combined effect of sodium carboxymethylcellulose (CN) and sodium caseinate (SC) (two typical hydrocolloids with thickening mechanisms), remains unexplored in batter‐coated deep‐fried pork applications. This study bridges this gap by systematically evaluating their interaction effects on batter properties and oil penetration. CN is a chemically modified cellulose ether compound made from natural cellulose with high molecular weight and long‐chain structure, ¹⁸ which exhibits remarkable high‐viscosity properties as well as accessibility and safety in applications. ¹⁹ SC is the sodium salt of casein, produced by the reaction of casein with sodium hydroxide, ¹⁹ and its high‐molecular‐weight protein structure and amphiphilicity give it high‐viscosity features, and it exhibits high nutritive properties and strong thermal stability in applications. ¹⁵ Furthermore, CN and SC are both high‐viscosity ionic colloids, and their combined use in batter systems could exploit unique hydrogen bonding ²⁰ and steric hindrance effects to form a denser, more stable network. ²¹

The study reported here innovatively investigated the synergistic effect of CN and SC in batter. Based on the assumption that it may form a dense three‐dimensional network, the study focused on the effect of this system on batter viscosity and thermal stability, as well as the modulation of the quality, oil distribution, and microstructure of batter‐coated deep‐fried pork.

MATERIALS AND METHODS

Materials

Fresh pork tenderloin (1000 ± 10 g) and eggs (700 g ± 5 g) were sourced from a local market (Hefei, Anhui, China). Wheat starch and wheat gluten were obtained from Xinxiang Liangrun Whole Grain Foods Co. Ltd (Xinxiang, Henan, China). Baking powder was sourced from Huahaishunda Food Co. Ltd (Cangzhou, China). CN and SC were purchased from Xinrui Biotechnology Co. Ltd (Zhengzhou, China). Soybean oil was provided by Yihai Kerry Arawana Holdings Co. Ltd (Wuhan, China). Seasonings were all sourced from local supermarkets (Hefei, Anhui, China). All chemical reagents were of analytical grade.

Preparation and characterization of batter

Preparation of batter

The basic batter formula consisted of wheat starch (45.85%), wheat gluten (4.15%), whole egg liquid (20%), and deionized water (30%). Baking powder (0.2%) and salt (0.5%) were added based on the total weight of wheat starch and wheat gluten. On this basis, CN, SC, or their complex colloid (CN:SC = 1:1) were added at levels of 0.2%, 0.4%, and 0.6% (relative to the total weight of wheat starch and wheat gluten). The resulting batters were named as follows: B‐CN‐0.2, B‐CN‐0.4, B‐CN‐0.6 (CN only); B‐SC‐0.2, B‐SC‐0.4, B‐SC‐0.6 (SC only); B‐CN‐SC‐0.2, B‐CN‐SC‐0.4, B‐CN‐SC‐0.6 (CN:SC = 1:1). The basic batter without added hydrocolloids served as the control. All mixtures were stirred for 3 min using a handheld electric mixer (EGK518, Hefei Rongshida Electronic Appliance Group Co. Ltd, China) to obtain a homogeneous batter.

Batter pick‐up rate

In accordance with a previously reported methodology, ¹⁷ the batter pick‐up rate of the samples was determined. The tenderloin was cut into 4 × 1 × 1 cm³ strips. The cut pieces of meat were submerged in the batters for 10 s, then lifted out and placed on a steel wire support, allowing excess batter to naturally drip off for 15 s. The coating operation was then repeated to obtain batter‐coated deep‐fried pork with its exterior uniformly and stably covered by the batter. To prevent water evaporation during the process, the batter container was sealed with cling film. The above steps were repeated until ten sets of samples from that batch were processed. The weight of the batter before and after coating, as well as the weight of the sample prior to coating, was determined. The batter pick‐up rate was then calculated using the following formula:

$$\text{Batter pick}-\mathrm{up}\text{rate}\left(\%\right)=\frac{\text{Weight of batter before coating}-\text{weight of batter after coating}}{\text{Weight of sample before coating}}$$ (1)

Apparent viscosity

The apparent viscosity of the batter was determined using a rotational rheometer (DHR‐1, TA Instruments, USA). With slight modifications to the referenced method, ²² the batter was applied to a flat plate with a gap of 1 mm and a diameter of 40 mm. A layer of silicone oil was applied around the sample surface, and the shear rate was set to range from 0.1 to 10 s⁻¹.

Thermal properties

The thermal properties were determined using a differential scanning calorimeter (Q200, TA Instruments, USA). Before conducting thermal characterization, the batter samples were subjected to the following freeze‐drying process: the homogeneously prepared batters were dispensed into freeze‐drying dishes, pre‐frozen, and subsequently transferred to a freeze dryer (Biocool Experimental Instrument, Beijing, China) for 24 h freeze‐drying and sieved through a 100‐mesh sieve. Referring to a previously reported method ²³ with slight modifications, 3 mg of lyophilized batter was added to 9 μL of deionized water and mixed homogeneously, sealed in an aluminium crucible, and equilibrated overnight at 4 °C, using the empty crucible as a reference. The sample was heated at a rate of 10 °C min⁻¹ from 20 to 100 °C. The initial temperature (T ₀), peak temperature (T _P), end temperature (T _C), and enthalpy (ΔH) were recorded.

Frying process and characterization of batter‐coated deep‐fried pork

Frying process

First, the tenderloin was cut into 4 × 1 × 1 cm³ strips. A marinade was then prepared by precisely weighing the following ingredients based on the tenderloin's weight: salt (1.5%), monosodium glutamate (0.3%), pepper powder (0.5%), chili powder (0.5%), Sichuan pepper powder (0.5%), ginger powder (0.5%), compound phosphate (0.8%), soy sauce (1%), and cooking wine (2%). Subsequently, the cut tenderloin was completely immersed in the marinade and marinated at 4 °C for 1 h to fully absorb the flavor of the seasonings. The marinated tenderloins were submerged in the corresponding prepared batters for 10 s, then lifted out and placed on a steel wire support, allowing excess batter to naturally drip off for 15 s. The coating operation was then repeated to obtain batter‐coated deep‐fried pork with its exterior uniformly and stably covered by the batter. Two identical fryers were used to control the temperature conditions of 170 and 180 °C, respectively, and each fryer (CookerBene, Jiaxing Yingjiao Household Products Co. Ltd, China) was filled with 1.2 L of soybean oil. In each batch of the experiment, each group of samples was cooked independently, and the number of the batter‐coated deep‐fried pork was one per frying. The groups of samples were processed in a randomized order, and uniform frying conditions were maintained in the same equipment. Initially, frying was at a temperature of 170 °C for 2 min, followed by a second frying at a temperature of 180 °C for 35 s. The oil was replaced with fresh oil after frying in each group to ensure consistency of experimental conditions. The batter‐coated deep‐fried pork prepared with different compositions of batter was designed as follows: F‐CN (F‐CN‐0.2, F‐CN‐0.4, F‐CN‐0.6) represents samples of the batter‐coated deep‐fried pork made by deep‐frying using batter containing 0.2%, 0.4%, and 0.6% CN; F‐SC (F‐SC‐0.2, F‐SC‐0.4, F‐SC‐0.6) represents samples of the batter‐coated deep‐fried pork made by deep‐frying using batter containing 0.2%, 0.4%, and 0.6% SC; F‐CN‐SC (F‐CN‐SC‐0.2, F‐CN‐SC‐0.4, F‐CN‐SC‐0.6) represents samples of the batter‐coated deep‐fried pork made by deep‐frying after hanging the batter using CN and SC compounded at 1:1 (total additions of 0.2%, 0.4%, and 0.6%, respectively).

Textural properties

Referring to a previously reported method ⁶ with modifications, the texture profile analysis of the batter‐coated deep‐fried pork was conducted using a texture analyzer (TA‐XT Plus, Stable Micro Systems, USA). A P50 probe was selected for the analysis. The pre‐test, test, and post‐test speeds were set at 1.00, 5.00, and 5.00 mm s⁻¹, respectively, with a compression ratio of 40%.

Color

Color measurement was performed using a colorimeter (ZE7700, Shanghai Shouli Industrial Co. Ltd, Japan) on the intact, undamaged batter surface of the batter‐coated deep‐fried pork to assess the ideal golden yellow color after frying. The colorimeter was calibrated using a white plate prior to measurement. The illuminant was D65. The diameter of the measuring area was 6 mm with a 10° observer. Each group of samples was measured at three different sites. The L*, a*, and b* values of the batter‐coated deep‐fried pork were measured. The L* value indicates brightness, the a* value indicates the red–green value, and the b* value indicates the yellow–blue value. The total color difference of the batter‐coated deep‐fried pork crust was calculated using the following formula ²⁴ :

$$\begin{array}{c}∆E=\sqrt{{\left(L_0-L^{*}\right)}^2-{\left(a_0-a^{*}\right)}^2-{\left(b_0-b^{*}\right)}^2}\end{array}$$ (2)

In this context, L ₀, a ₀, and b ₀ represent the color values of the control sample.

Moisture content

Firstly, the crust was separated from the core of each batter‐coated deep‐fried pork sample. Then, 2 g of both the crust and the core were separately placed into a moisture analyzer (MB27ZH, OHAUS Instruments Co. Ltd, Changzhou, China) and dried until a constant weight was achieved. The values of moisture content were recorded. Three replicate experiments were performed.

Fat content

The fat content of the samples was determined by the Soxhlet extraction method (SOX406, Hanon Advanced Technology Group Co. Ltd, Jinan, China). Then 1 g each of the dried crust and core was taken, wrapped in filter paper, and placed into a Soxhlet extraction thimble, followed by the addition of petroleum ether. Extraction was conducted at 60 °C for 4 h. The fat content of the samples was calculated based on Eqn (3). ²⁵ Three replicate experiments were performed.

$$\begin{array}{c}\mathrm{Fat}\text{content}\left(\%\right)=\frac{W_{\text{Sample}1-\text{Filter Paper}}-W_{\text{Sample}2-\text{Filter Paper}}}{W_{\text{Sample}1-\text{Filter Paper}}-W_{\text{Filter Paper}}}\times 100\%\end{array}$$ (3)

Water state

According to a method previously reported, ¹⁰ the moisture status of the batter‐coated deep‐fried pork was determined using a low‐field nuclear magnetic resonance (LF‐NMR) imaging analyzer (NMI20‐015V‐I, Niumag Electric Co., Suzhou, China). Two grams each of crust and core were taken into sample vials. The samples were then subjected to LF‐NMR CPMG sequence measurements, followed by T2 inversion analysis.

Thermogravimetric analysis of crust

The thermogravimetric characteristics of the batter‐coated deep‐fried pork crust were determined using a thermogravimetric analyzer (TGA 8000, PerkinElmer Instrument Co. Ltd, USA). Approximately 5 mg of the fried crust was placed in an alumina crucible and tested under a nitrogen atmosphere. The heating rate was 10 °C min⁻¹, and the temperature range for the test was from 30 to 700 °C. ²⁶

Scanning electron microscopy

The crust of each batter‐coated deep‐fried pork sample was peeled off. The samples were placed in a freeze‐drying dish for pre‐freezing and subsequently transferred to a freeze dryer (Biocool Experimental Instrument, Beijing, China) for freeze‐drying for 24 h to remove moisture. The dried crusts were then degreased using the Soxhlet extraction method with reference to the method described in the Moisture content section. After the treated samples were sprayed with a gold coating on the surface, the microstructure was observed using a scanning electron microscope (EM30+, COXEM, Daejeon, Korea) at ×200 magnification. ²⁷

Confocal laser scanning microscopy

Following a previously reported method ²⁸ with some modifications, 0.1 g of crust was added to deionized water (1 g) along with 10 μL of 0.1% Nile Red. After vortexing for 10 min, the mixture was stained in the dark at 4 °C for 1 h. The sample was then observed using a confocal laser scanning microscope (FV1000, Olympus, Tokyo, Japan) at a magnification of ×10 with an excitation wavelength of 543 nm.

Statistical analysis

Three independent batches of batter and batter‐coated deep‐fried pork were prepared, and the indicators were tested in triplicate as technical replicates of each batch. The test data are shown as mean ± standard error (SE). In the experimental design, different additions of CN, SC, and CN‐SC complex colloids were categorized as fixed effects. Meanwhile, the experimental batches were categorized as random effects to account for possible random differences between batches. The differences between samples were determined by one‐way analysis of variance and Tukey's test (P < 0.05). The experimental results were graphed using Origin 2024 (Origin Lab, Northampton, MA, USA).

RESULTS AND DISCUSSION

Characteristics of batter

Batter pick‐up rate

Batter pick‐up rate is critical for deep‐fried battered food products, as this indicator directly affects the yield and quality of the product. ²⁹ As shown in Fig. 1(A), the types and amounts of hydrocolloids significantly affect the batter pick‐up rate of the batter‐coated deep‐fried pork (P < 0.05). The batter pick‐up rate increased significantly with an increase of the amounts of hydrocolloids (P < 0.05). The B‐CN‐SC‐0.6 group reached the peak (56.76%), which was 24.76% higher than that of the control group. The viscosity‐increasing effect of CN was significantly greater than that of SC (P < 0.05). When the addition amount was 0.6%, the batter pick‐up rate of the CN group was 9.38% higher than that of the SC group, which was due to the interaction of hydrocolloids with water molecules, such as the formation of hydrogen bonds, and this close binding with water molecules could reduce the fluidity of the batter system, resulting in an increase in the viscosity of the batter. ³⁰ Previous study has shown that the greater the viscosity of batter, the greater the batter pick‐up rate. ³¹ The significant synergistic effect was found only at the level of 0.6% CN‐SC (P < 0.05). The batter pick‐up rate of 0.6% CN‐SC was 4.75% and 14.13% higher than that of the same amount of CN and SC single groups, respectively. It is speculated that the entanglement of the two types of colloidal molecular chains formed a denser spatial network structure.

Figure 1.

Pick‐up rate (A) and apparent viscosity (B) of batters with different contents of hydrocolloids. The values are given as mean ± SE. Different letters indicate significant differences (P < 0.05).

Apparent viscosity

The effect of hydrocolloid types (CN, SC, and CN‐SC) and addition amount (0.2%, 0.4%, 0.6%) on the rheological properties of the batters is shown in Fig. 1(B). Rheological analysis showed that all batter systems exhibited typical shear‐thinning behavior, and their apparent viscosity decreased significantly with an increase of shear rate (P < 0.05), which was attributed to the directional arrangement of hydrophilic colloidal molecular chains under shear force. ³² The system exhibited the highest viscosity with 0.4% CN‐SC and reached peak batter pick‐up rate with 0.6% CN‐SC, indicating the need for balanced optimization between viscosity and pick‐up efficiency. In practice, a significant synergistic effect existed at the 0.6% CN‐SC level, and the batter pick‐up rate and rheological properties were favorable for batter hanging. This not only ensured a high paste retention (>50%), but also possibly created a homogeneous and dense frying crust for the subsequent frying of the batter‐coated deep‐fried pork, thus potentially reducing oil penetration. All batter systems in this study exhibited shear‐thinning behavior, which was consistent with the rheological properties of hydroxypropylmethylcellulose and xanthan gum batter systems as previously reported. ³³

Thermal properties

The effects of hydrocolloids on the thermodynamic parameters of the batters are presented in Table 1. Thermal stability (which refers to the ability of a material to maintain its structural integrity at high temperatures) and thermal decomposition (degradation of a substance due to the breaking of chemical bonds at high temperatures) are key parameters affecting the quality of fried foods. It was shown that hydrocolloids delayed the onset of thermal decomposition (T₀ and T _p increase) while decreasing the enthalpy (ΔH) by increasing the gelatinization temperature and prolonging the paste time of the batter. These results suggest that the incorporation of hydrocolloids enhanced the thermal stability of the protein gel network in the batter through the formation of covalent bonds, thus allowing the batter to maintain its structural stability for a longer period of time at high temperatures. ³⁴ For batters with high thermal stability, their oil absorption after frying is usually reduced. ¹¹ This is because better stability can suppress the formation of excessive holes and cracks during frying, making the structure relatively dense, while reducing the contact between the internal structure and lipophilic groups at high temperatures. In contrast, thermally unstable batters could increase oil absorption when applied to deep‐fried foods.

Table 1.

Thermal properties of batters with different contents of hydrocolloids

	T 0 (°C)	T P (°C)	T C (°C)	ΔH (J g−1)
Control	61.61 ± 0.46a	67.20 ± 0.07a	75.89 ± 0.13ab	2.07 ± 0.15a
B‐CN‐0.2	61.88 ± 0.14a	66.71 ± 0.12b	74.98 ± 1.26abc	1.99 ± 0.25ab
B‐CN‐0.4	62.11 ± 0.10a	67.00 ± 0.02ab	76.27 ± 0.27a	1.75 ± 0.04abc
B‐CN‐0.6	62.13 ± 0.04a	67.10 ± 0.01a	74.79 ± 0.57abc	1.72 ± 0.08abc
B‐SC‐0.2	62.10 ± 0.18a	66.66 ± 0.27b	74.31 ± 0.14abc	1.97 ± 0.10ab
B‐SC‐0.4	61.85 ± 0.04a	66.70 ± 0.07b	73.23 ± 0.53c	1.77 ± 0.11abc
B‐SC‐0.6	62.12 ± 0.19a	66.71 ± 0.13b	73.84 ± 0.05bc	1.71 ± 0.12abc
B‐CN‐SC‐0.2	61.84 ± 0.08a	66.92 ± 0.01ab	74.57 ± 0.63abc	1.65 ± 0.12bc
B‐CN‐SC‐0.4	62.27 ± 0.02a	67.25 ± 0.06a	75.45 ± 0.39ab	1.49 ± 0.03c
B‐CN‐SC‐0.6	62.02 ± 0.03a	66.70 ± 0.08b	74.76 ± 0.78abc	1.48 ± 0.07c

Data are presented as mean ± SE. Different letters within the same column represent statistically significant differences (P < 0.05). T ₀: initial temperature; T _P: peak temperature; T _C: end temperature; ΔH: enthalpy. Control: without hydrocolloids; B: batter; CN: sodium carboxymethylcellulose; SC: sodium caseinate; 0.2, 0.4, 0.6: percentage of hydrocolloid addition.

Characteristics of batter‐coated deep‐fried pork

Textural properties

The textural properties of the samples with hydrocolloids added to the batter are presented in Table 2. The hardness of the batter‐coated deep‐fried pork was significantly reduced by the addition of CN, SC, and complex colloids compared to the control (P < 0.05). This decrease in hardness is mainly due to the fact that the hydrocolloid locks water through the hydrogen bond network to improve the water content of the product. The increase of water content can effectively weaken the mechanical strength of the food matrix. ³⁵ It is worth noting that although parameters such as springiness, cohesiveness, and resilience did not show significant changes, all the samples with hydrocolloid addition showed a decrease in chewiness (about 20–45%), suggesting that the product was easier to chew and swallow, significantly improving the palatability.

Table 2.

Textural properties of batter‐coated deep‐fried pork with different batter ingredients

	Hardness (N)	Springiness	Cohesiveness	Chewiness (N)	Resilience
Control	19.19 ± 0.20a	0.78 ± 0.06b	0.69 ± 0.03a	10.28 ± 1.23a	0.40 ± 0.03a
F‐CN‐0.2	15.72 ± 0.74b	0.94 ± 0.06a	0.75 ± 0.03a	11.10 ± 1.29a	0.49 ± 0.05a
F‐CN‐0.4	13.15 ± 0.47c	0.85 ± 0.04ab	0.69 ± 0.02a	7.67 ± 0.25b	0.41 ± 0.01a
F‐CN‐0.6	13.67 ± 0.30c	0.81 ± 0.03ab	0.69 ± 0.02a	7.63 ± 0.42bc	0.41 ± 0.04a
F‐SC‐0.2	10.33 ± 0.43d	0.78 ± 0.02b	0.66 ± 0.01a	5.37 ± 0.41d	0.39 ± 0.01a
F‐SC‐0.4	13.97 ± 0.11c	0.79 ± 0.01b	0.68 ± 0.02a	7.51 ± 0.40bc	0.39 ± 0.03a
F‐SC‐0.6	13.50 ± 0.06c	0.83 ± 0.03ab	0.72 ± 0.01a	8.05 ± 0.39b	0.45 ± 0.00a
F‐CN‐SC‐0.2	11.07 ± 0.02d	0.80 ± 0.03ab	0.70 ± 0.02a	6.22 ± 0.39bcd	0.42 ± 0.02a
F‐CN‐SC‐0.4	10.15 ± 0.47d	0.81 ± 0.03ab	0.70 ± 0.02a	5.74 ± 0.37cd	0.42 ± 0.03a
F‐CN‐SC‐0.6	13.68 ± 0.19c	0.79 ± 0.02b	0.67 ± 0.02a	7.28 ± 0.45bcd	0.40 ± 0.03a

Data are presented as mean ± SE. Different letters within the same column represent statistically significant differences (P < 0.05). Control: without hydrocolloids; B: batter; CN: sodium carboxymethylcellulose; SC: sodium caseinate; 0.2, 0.4, 0.6: percentage of hydrocolloid addition.

Unlike the findings of Shahzad et al., ³⁶ who reported that most hydrocolloids increased the hardness of sweet potato starch gel, the present study showed that CN, SC, and CN‐SC significantly reduced the hardness of batter‐coated deep‐fried pork. This difference may be attributed to the complexity of the research systems: Shahzad et al. investigated a simple starch gel system, whereas the present study involved a starch–wheat gluten–hydrocolloid multicomponent system. Moreover, the hydrocolloids locked a large amount of water through hydrogen bonding, and the increased water content weakened the mechanical strength of the food matrix, thereby reducing hardness.

This texture improvement effect is closely related to the multifunctional properties of the hydrocolloid ³⁷ : its thickening effect regulates the rheological behavior of the batter, the water retention effect maintains the juicy feeling of the product, and the gelation effect optimizes the three‐dimensional network structure of the final product. ³⁸

Color

The color of deep‐fried foods is an important indicator of consumer appeal. Golden‐colored deep‐fried foods are often considered to be attractive. The main color‐rendering mechanism in fried foods is the Maillard reaction, and the moisture content of the samples is also influential. ³⁹ It is also affected by the type and amount of amino acids and reducing sugars in the food, as well as the frying temperature and time. ⁴⁰ Figure 2(A) shows the color values for all batter‐coated deep‐fried pork. ΔE represents the degree of difference in color between two samples. It is generally accepted that when ΔE is greater than 3, it means that a color difference can be observed. ⁴¹ The moisture content of a sample significantly affects the change in L* value. Hydrocolloids can effectively lock water and increase the water content of fried food, which in turn enhances the light reflection of the surface, ultimately leading to an increase in the L* value. ⁴² Deep‐fried foods with an attractive appearance usually correspond to a higher brightness (L*) and yellowness (b*), and a lower redness (a*). ¹² Thus, the color profile of the CN‐SC group was more in line with consumer preferences. This may be attributed to the formation of a more uniform heat conduction network through the synergistic effect of the two colloidal molecules, making the Maillard reaction proceed at a more controllable rate. These findings provide a new idea for the development of fried food with an ideal golden color so as to obtain an end product with an attractive appearance and uniform color.

Figure 2.

Color values (A), fat content (B), and moisture content (C) of batter‐coated deep‐fried pork with different batter compositions. The values are given as mean ± SE. Different letters indicate significant differences (P < 0.05).

Moisture content

The moisture content of the crust and core meat of batter‐coated deep‐fried pork is shown in Fig. 2(B). The experimental data showed that the moisture content of the crust and core meat of the samples with added CN, SC, and CN‐SC (24.38–42.30% and 38.80–56.84%, respectively) was significantly higher than that of the control group (23.29% and 36.34%) (P < 0.05). The improvement of water‐holding capacity was mainly due to: first, polar groups (such as hydroxyl and carboxyl groups) on the surface of hydrocolloids form a stable hydration layer through hydrogen bonding ⁴³ ; second, the colloidal molecular network structure delayed the water migration rate during frying; third, as shown by the above rheological analysis, the increase in viscosity effectively inhibited water evaporation. Water–oil displacement is a fundamental mechanism governing oil absorption during deep‐frying. Specifically, when food is immersed in hot oil, the rapid evaporation of internal water creates pores and voids within the food matrix. Once the food is removed from the oil, the cooling phase generates negative pressure, drawing surface oil into these voids through capillary action. ⁴⁴ This process directly leads to oil absorption. As shown in Fig. 2(B),(C), the inverse relationship between moisture content and fat content in this study clearly verifies the occurrence of water–oil displacement (Fig. 2(B),(C)). The CN‐SC group showed the optimal water retention performance, and after adding 0.6% complex colloid, the moisture content of the crust and meat increased by 16.27% and 17.35%, respectively, compared with the control group, which was closely related to its denser network structure and higher water‐holding capacity. Based on the data of moisture content and fat content, it can be concluded that the higher the moisture content, the lower the fat content, and vice versa. These findings provide a new perspective for understanding the mechanism of water–oil interaction during frying: by adjusting the type and amounts of hydrocolloids, a high‐water‐holding and low‐fat system can be established. In other studies, pectin, hydroxypropylmethylcellulose, and guar gum improved the surface moisture response of 3D‐printed deep‐fried samples, reflecting the enhanced water‐retention capacity. This is consistent with the results of CN, SC, and complex colloids to enhance the moisture content of the crust and core of batter‐coated deep‐fried pork in this study, confirming the positive effect of hydrocolloids on the water retention of deep‐fried systems. ⁴⁵

Fat content

Soxhlet extraction is a commonly used method for measuring fat in foods. The addition of hydrocolloids significantly reduced the fat content of batter‐coated deep‐fried pork (Fig. 2(C)), which showed a significant negative correlation with the increase of the batter pick‐up rate. The 0.6% CN‐SC sample showed the optimal fat reduction effect, and the fat content of crust and core meat decreased by 5.75% and 4.40%, respectively, compared with the control group, which was significantly higher than that of the single colloid group (P < 0.05). This fat reduction was mainly due to: first, a more complete batter coating layer is formed with a higher batter pick‐up rate, which effectively blocks oil penetration; second, hydrocolloids reduce the formation of pores during frying by improving the gluten network structure ⁴⁶ ; third, colloidal molecules fix water through hydrogen bonds, which produces a vapor barrier effect during frying. The synergistic effect of the complex colloid is particularly prominent, which can achieve a more significant fat reduction effect by simultaneously optimizing the pick‐up performance and microstructure. The 0.6% CN‐SC complex colloid reduced the fat content of the crust by 5.75%, showing a more pronounced impact compared to the reported effects (surface oil and penetration oil reduction <3%) of single hydrocolloid. ³¹ This difference may be attributed to the ability of the complex colloid to simultaneously optimize batter pick‐up rate, viscosity, and microstructure, whereas single hydrocolloid often improves only one of these aspects.

Water state

LF‐NMR is a nondestructive testing technique that does not require disruption of the sample structure to obtain moisture information, as determined by the transverse relaxation time (T2). The peak areas represent the relative amount of water at different stages. For the crust of batter‐coated deep‐fried pork (Fig. 3(A),(C)), the first peak (A₂₁) (0.01–1 ms), the second peak (A₂₂) (1–100 ms), and the third peak (A₂₃) (100–1000 ms) represent bound water, immobilized water, and free water, respectively. For the core meat of batter‐coated deep‐fried pork (Fig. 3(B),(D)), the first peak (A₂₁) (0.01–10 ms), the second peak (A₂₂) (10–100 ms), and the third peak (A₂₃) (>100 ms) represent bound water, immobilized water, and free water, respectively. ³¹ The incorporation of hydrocolloids had a significant effect on the relaxation time. The longer relaxation time indicates that the water molecules interact weakly with the surrounding macromolecules. The bound water in the core meat of batter‐coated deep‐fried pork would be tightly bound to myofibrillar proteins through hydrogen bonding, and this strong interaction resulted in a weaker relaxation time. The results showed that 0.4% CN‐SC significantly changed the existing state of water molecules: in the crust, the peak areas of bound water (A₂₁) and fixed water (A₂₂) increased by 1.55% and 4.21%, respectively, while the peak areas of free water (A₂₃) decreased by 5.77% (P < 0.05). Free water was converted into water that was not easy to flow, which changed the texture of the food. Bound water combines with protein, starch, and other macromolecules to make the structure of batter‐coated deep‐fried pork more compact, further preventing the penetration of oil.

Figure 3.

Batter moisture distribution (A), core moisture distribution (B), batter moisture peak area ratio (C), and core moisture peak area ratio (D) of batter‐coated deep‐fried pork with different batter compositions. The values are given as mean ± SE. Different letters indicate significant differences (P < 0.05).

Thermogravimetric characteristics of crust

Thermogravimetric analysis was used to study the mass dynamics of batter‐coated deep‐fried pork crust as a function of temperature (Fig. 4(A),(B)). The temperature corresponding to the peak in the differential thermogravimetry (DTG) curve gives an indication of the thermal stability of a substance. ⁴⁷ The thermal decomposition curves of batter‐coated deep‐fried pork crust with different hydrocolloid types and addition amounts had similar trends, and the weight loss was divided into three stages. ⁴⁸ The initial phase occurs between 0 and 200 °C and is dominated by water loss from the crust. Compared with the control, the temperature corresponding to the first‐stage peak in the DTG curve after the addition of hydrocolloid was significantly higher than that of the control, indicating that the structure of the crust was significantly denser, thus reducing the water loss in this stage. During this phase, the main substance loss is free water. The complex colloidal group also resulted in less water loss at this stage due to the conversion of free water into immobilized water.

Figure 4.

Thermogravimetric (TG) curves of crusts of batter‐coated deep‐fried pork with different batter ingredients (A) and the first derivatives of the TG (DTG) curves (B). Each TG curve represents mass versus temperature, and each DTG curve represents rate of change of mass versus temperature.

This was followed by the 200–350 °C stage, which was dominated by starch decomposition and protein denaturation. The thermogravimetry curves of the control had a significant downward trend, indicating that significant thermal decomposition and mass loss of components such as starch and protein occurred in the samples in this temperature range. Starch undergoes glycosidic bond breakage, and protein spatial structure disruption decomposes into small molecules such as amino acids, resulting in mass loss. In contrast, the samples with the addition of CN or SC showed a more moderate decrease in mass. This is because CN and SC molecules may combine with starch chains through hydrogen bonding to form a stable composite structure, hindering the depolymerization of starch molecules and glycosidic bond breaking at high temperatures, and at the same time, the colloidal network may encapsulate starch granules, reducing their direct contact with the thermal environment and slowing down the rate of thermal decomposition. SC, as a protein colloid, can be crosslinked with other proteins in the matrix through hydrophobic interactions or disulfide bonds to enhance the binding force between protein molecules and increase the thermal denaturation temperature, ⁴⁹ whereas polysaccharide colloids, such as CN, could hinder the unfolding and aggregation of protein molecules at high temperatures through the spatial site‐blocking effect to inhibit their further decomposition. ⁵⁰ The composite colloid samples also showed a similar mild tendency, and it was hypothesized that the synergistic effect of CN and SC could enhance the stability of the above network structure and further inhibit the thermal decomposition of starch and protein.

The final stage occurs in the range of 350–450 °C and is attributed to the degradation of the starch–lipid complex. The thermal decomposition temperature increased as the amount of hydrocolloid added increased, which can also be seen in the thermal properties (Table 1). It is suggested that colloids may stabilize the structure of starch–lipid complexes and retard their high‐temperature degradation by binding to lipids.

In summary, hydrocolloids reduce water loss by strengthening the crust structure and inhibit thermal decomposition by interacting with starch, proteins, and lipids, thus delaying the destruction of the matrix structure.

Microstructure of crust

Oil absorption is the result of oil entering through the pores and cracks of the crust and displacing evaporated water. This mechanism suggests that a dense crust acts as a barrier to prevent water migration and oil absorption. ⁵¹ The microstructure of batter‐coated deep‐fried pork crust is shown in Fig. 5(A). The microstructure characterization showed that the addition of hydrocolloids significantly improved the compactness of the crust. The control group had larger and more holes and a rough surface. With the addition of hydrocolloids, the surface of the crust became denser and more uniform, and the number of holes and cracks was relatively small. With an increase of CN and SC addition amounts, it was evident that the crust became denser. Among them, the surface of the fried crust of the 0.4% and 0.6% complex colloid groups had visually fewer holes. This may be due to the reinforcing effect of the three‐dimensional mesh formed by the hydrocolloids in the batter on the structural integrity. This three‐dimensional mesh structure also slows down the rate of water loss during high‐temperature deep‐frying, making the surface of the crust more uniform when the batter is deep‐fried, thus avoiding the rapid loss of localized water and the creation of holes. ⁵² The dense crust blocked the penetration of oil, resulting in a lower fat content of batter‐coated deep‐fried pork, which can also be seen in Fig. 2(C). It was previously found that the addition of hydrocolloids to 3D‐printed fried meat analogues resulted in smoother surfaces and reduced roughness, ⁵³ which is consistent with the results of the present study, where hydrocolloids (CN and SC) improved the densification of the batter crust and reduced the number of holes, both suggesting that hydrocolloids reduce water loss and oil penetration during deep‐frying by optimizing the microstructure.

Figure 5.

Microstructure (A) and oil distribution (B) of batter‐coated deep‐fried pork crust with different batter compositions.

Oil distribution of crust

The molecular mechanism of hydrocolloids regulating the oil distribution of batter‐coated deep‐fried pork was systematically revealed using confocal laser scanning microscopy observation and analysis (Fig. 5(B)). The oil in the fried crust was colored with Nile Red, and the greater intensity of the red fluorescence indicates a higher fat content. ⁵⁴ The experimental results showed that among all the samples, the fat content of the control was the highest, with many larger oil droplets attached to the surface. As shown in the results discussed above (Figs 1(B) and 5(A)), hydrocolloids improved the viscosity of the batter, formed a denser three‐dimensional network structure, and formed a dense crust of batter‐coated deep‐fried pork during the frying process, preventing moisture from migrating outward, thus significantly reducing the fat content, making the oil droplets smaller and more dispersed, especially in the 0.6% CN‐SC group. This is consistent with the fat content results shown in Fig. 2(C).

CONCLUSIONS

The binary‐hydrocolloid system CN‐SC exhibited an impact in reducing oil penetration in batter‐coated deep‐fried pork, outperforming single‐hydrocolloid systems. CN and SC synergistically increased batter viscosity, achieving a pick‐up rate of 56.76%, which was 24.76% higher than that of the control group. Further, they promoted the conversion of free water to bound and immobilized water, with bound water increased by 1.55% and immobilized water by 4.21%, and free water reduced by 5.77%, thereby reducing evaporable water and suppressing the water–oil displacement reaction. The 0.6% CN‐SC sample achieved ideal performance: fat decreased by 5.75% in the crust and 4.40% in the meat, while moisture increased by 16.27% and 17.35%, respectively.

CONFLICT OF INTEREST

The authors have no conflicts of interest or personal relationships that might have influenced the work related to this study.

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.