Quality Evaluation Focusing on Tissue Fractal Dimension and Chemical Changes for Frozen Tilapia with Treatment by Tangerine Peel Extract

This work aimed to establish an effective approach to evaluate the quality of frozen fish, focusing on changes in fish tissue structure and chemical composition during storage. Fresh tilapia samples were treated by coating with tangerine peel (TP) extract and then stored at −4, −8 and −18 °C, respectively, for 40 days. The frozen fish tissues were analyzed for structural and chemical changes. Fractal dimension, which quantifies the porous structure formed in the tissue samples, texture properties including hardness and springiness, and moisture content and water activity all decreased during the storage, while the extents of lipid oxidation, measured as peroxide value and thiobarbituric acid concentration, and protein degradation, monitored with total volatile basic nitrogen and trichloroacetic acid soluble peptides, increased. The change rates of these parameters decreased with decreasing the storage temperature and by applying TP extract. A model was developed for predicting fractal dimension, which indicated the quality of preserved tilapia and thus can be used to predict the shelf life under different storage temperatures. The results demonstrated that TP extract could extend the shelf life of frozen tilapia by 35–45% by inhibiting changes in tissue structure, moisture loss, lipid oxidation and protein degradation during frozen storage.

Nile Tilapia (Oreochromis Niloticus) is an extensively farmed fish. Because of its excellent biological characteristics and edible value, it becomes a popular food material around the world 1 . However, like other aquatic products, tilapia is highly perishable 2 . Generally, the most common and effective method to maintain its freshness is frozen storage 3 . Low temperatures can delay the deterioration of fish, but it also inevitably brings negative impacts on fish tissue 4 . During frozen storage, composition changes and moisture loss result in a honeycomb microstructure in the fish tissue. Such porous structure is inherently irregular and difficult to describe 5 .
The fractal dimension is a geometry parameter characterized as the space-filling capacity of a space structure at a certain scale 5 , which can be used to describe the porous structure in some food materials including frozen tilapia 6 . Fractal dimension analysis was also used as a quantitative indicator reflecting the surface roughness of frozen tofu 7 . In addition, Kerdpiboona et al. established a fractal dimension model for describing the relationship between microstructural and physical properties of dried foods 8 .
In this study, the preservation effects of tangerine peel (TP, Citri reticulatae pericarpium) 9 on the tissue structure and qualities of frozen tilapia were evaluated using fractal dimension along with measurements of changes in hardness and springiness, as well as chemical compositions due to moisture loss, lipid oxidation and protein decomposition. The results demonstrated the applicability and accuracy of fractal dimension analysis for predicting qualities and shelf life of frozen tilapia.
where d is the fractal dimension of the divided image and a is a constant. By plotting log N(ε) vs. log ε, the fractal dimension d can be obtained from the slope of the regression line (see Fig. S1 in Supplemental Materials).
Texture profile analysis. Texture profile analysis (TPA) was performed under the conditions described in a similar study 13 . A 6 mm cylindrical probe was used with 5 g trigger load and 5 mm/s test speed. Twice compressions with 5 mm depth were performed on the same point of each sample and separated by a 5 s interval. The TPA parameters including hardness and springiness were obtained from the measured force (N), area (Ns) and distance (mm) between peak heights.
Moisture content. The moisture content (MC) was estimated from the weights before and after drying to a constant weight in a 105 °C oven 14 . Water activity (A w ) was measured with minced tilapia tissue in a portable A w instrument (Hygro Palm, Rotronic Company, Switzerland) 15 .
Lipid oxidation. Peroxide value (PV) was measured using the method reported by Lea 16 . Fifty grams of tilapia were minced and extracted with a mixture of 50 mL distilled water, 100 mL methanol, and 100 mL chloroform. After dissolving 1 g liquid extract in a mixture of 10 mL chloroform and 15 mL acetic acid, 1 mL saturated potassium iodide was added and the mixture was kept in the dark for 10 min. Then, 30 mL distilled water and 1 mL starch solution (1% w/v) were added, and the solution was titrated with 0.01 mol equiv/L Na 2 S 2 O 3 until colorless. The PV was calculated from the dosage of Na 2 S 2 O 3. Thiobarbituric acid (TBA) was analyzed as described by Li et al. with some modifications 17 . One gram of tilapia was minced, dissolved in 5 mL of 4% v/v 1-butanol aqueous solution, pipetted into a dry stoppered test tube, mixed with 5 ml TBA reagent (200 mg 2-TBA dissolved in 100 mL 1-butanol), and incubated in a water bath at 95 °C for 120 min. After cooling, the absorbance at 530 nm of the mixture was measured to calculate the TBA. studied. These three temperatures are representative in aquatic products storage 21 . Figure 1 shows the microscopic images of micro-sections of tilapia tissues after 20 days and 40 days at − 4 °C, − 8 °C and − 18 °C. Compared to the initial samples, there were increasing void spaces (white) inside the frozon tissue (red) during storage 22 . Initially, tilapia tissues had only a few small void spaces, which grew into larger irregular pores over time 23 , which could be attributed to fluids migration caused by the ice crystallization and muscle fiber disarrangement caused by protein decomposition 6 . The increases in the void spaces were faster at higher temperatures. Interestingly, TP extract coating significantly hindered the structural change and was able to maintain tilapia tissue with fewer and smaller void spaces, especially at − 18 °C.
Changes in tilapia tissue structure can be quantitatively analyzed by measuring the fractal dimension, which is sensitive enough to distinguish tiny differences in the structure, size, or area fraction 24,25 . Figure 2 shows the fractal dimension values of tilapia samples at various frozen temperatures and storage times. In general, fractal dimension decreased with increasing the temperature and storage time, and the decrease was significantly slower for samples treated with TP extract.
Changes in the tissue structure of tilapia should also affect its texture properties 7 . As shown in Fig. 3a,b, both hardness and springiness of frozen tilapia tissues also decreased over time, in similar trends to those for fractal dimension. Clearly, the increasingly porous structure, as quantified by fractal dimension, formed during frozen storage resulted in the declines in hardness and springiness. As expected, TP extract treatment slowed down the declines in these texture properties.
Changes in moisture content during frozen storage. Variations in tissue structure and texture of tilapia tissue are greatly affected by the moisture content and chemical compositions 26 . During frozen storage, the crystallized tissue fluids greatly destroyed the structure of the sample tissue 27 , while the loss of moisture due to sublimation created the honeycomb microstructure with large void spaces in the muscle tissue 28 . Figure 3c,d shows changes in the moisture content (MC) and water activity in frozen tilapia tissues. After 40 days, MC in the control group reduced 25.3%, 22.3% and 16.4% at − 4 °C, − 8 °C and − 18 °C, respectively, from the initial value of 80.18 ± 3.15%. In contrast, tilapia tissues treated with the TP extract lost only 19.0%, 17.2% and 12.0%, respectively. Similar decreasing trends in A w were also observed (Fig. 4b). The classical Brunauer-Emmet-Teller model has been used to describe the close relationship between fractal dimension and moisture content 29,30 . Based on   Changes in chemical composition. Protein and lipid are two major organic components in fish tissue 32 . A chemical composition analysis showed that fresh tilapia tissue contained 14.70 ± 0.63% protein and 2.94 ± 1.19% lipid. Changes in protein and lipid contents in frozen tilapia also occurred during storage, which would reduce the water-retaining capability of the tissue and result in faster moisture loss. It can also directly disarrange muscle fibers and change tissue structure and texture 33,34 . Lipid oxidation. Changes in the lipid content usually can be quantified by monitoring the degree of lipid oxidation, which is a significant process of chemical deterioration in the refrigerated aquatic products 35 . In this process, free radicals "stolen" electrons from the polyunsaturated fatty acids which contain multiple double bonds in cell membranes. It impacts water-retaining capability of tissue and results in the variations in the tissue structure 36,37 . PV and TBA are two common parameters for determining the degree of lipid oxidation 38 . In general, both PV and TBA in frozen tilapia increased with the storage time (Fig. 4a,b). The tilapia tissue had the initial PV of ~0.76 meq/kg, which increased to 3.69, 2.35, and 1.48 meq/kg in the control samples after 40 days at − 4 °C, − 8 °C and − 18 °C, respectively (Fig. 4a). In contrast, the TP extract treated tilapia tissues had significantly lower PV of 2.29, 1.34 and 1.06 meq/kg at the respective storage temperatures. Similar trends were also observed with TBA, which increased from the initial 0.58 mg MDA/kg to 2.29, 1.31, and 0.93 mg MDA/kg for the control and from 0.51 MDA/kg to 1.39, 0.88, and 0.74 mg MDA/kg for the treated samples after 40 days at − 4 °C, − 8 °C and − 18 °C, respectively (Fig. 4b). Since lipid oxidation is mainly associated with unsaturated fats exposed to oxygen or air, TP extract as a protective barrier on the sample surface can slow down lipid oxidation by deterring the contact with oxygen.
Protein decomposition. TVB-N and TCA soluble peptides are parameters useful for monitoring protein decomposition 31 . Both TVB-N and TCA soluble peptides in frozen tilapia increased significantly during frozen storage (Fig. 4c,d). As expected, the increases were slower at lower temperatures and with TP extract treatment. The muscle fiber in tilapia tissue composed mainly proteins, and therefore protein decomposition would greatly affect tissue morphology and texture 39 . Accordingly, changes in TVB-N and TCA soluble peptides would also affect the fractal dimension.

Fractal dimension as the key parameter for shelf-life prediction. Pearson correlation coefficient,
which is widely used to determine the degree of linear dependence between two variables 6 , was calculated to determine correlations between fractal dimension and various texture and chemical parameters during frozen storage of tilapia (see Fig. S2 in Supplemental Materials). Table 1 shows that all have a Pearson correlation coefficient of ± 0.95. This result indicated that fractal dimension can be employed as a reliable quality index to reveal variations in chemical and texture parameters quantitatively.
The shelf-life of the tilapia samples was predicted by first-order kinetics model with using fractal dimension of tilapia tissues as the key quality factor 40 . The model can be expressed as Eq. 2.
where t is the storage time (d), d 0 is the initial value of fractal dimension, d t is the fractal dimension value at storage time t, and k T is the rate constant at storage temperature T.
Additionally, the relationship between the rate constant k T and the storage temperature T can be shown by the Arrhenius equation as follows 41 : T a 0 where k 0 is the pre-exponential factor (d −1 ), E a is the activation energy (kJ/mol), and R is the gas constant (8.314 J/ mol K). T, k T , k 0 , and E a are constants associated with the physical nature of the reaction system.
Incorporating Eq. 2 with Eq. 3, a global equation can be formulated: With using the data of fractal dimension in the Fig. 2, activation energy E a of the control samples was calculated as − 1.104 × 10 13 kJ/mol, and pre-exponential factor k 0 was 9883.975 d −1 . Meanwhile, E a of treated samples was − 2.343 × 10 13 kJ/mol, and k 0 was 10179.648 d −1 . Accordingly, the first-order kinetics equation of fractal dimension was obtained as follows: tT 13 2 In addition, the maximum of fractal dimension in spoiled tilapia samples were 1.801 ± 0.025 6 . So the results of the shelf-life were obtained as listed in Table 2. The data was harmonious with similar researches 6,40,41 , which indicated that fractal dimension employed as a quality index was reliable. Additionally, the TP extracts treatment prolonged the shelf-life of the frozen tilapia about 40%. The results can confirm and quantify the effects of TP extracts on freshness of preserved fish reported in a previous study 9 .  Effects of TP extract. The result of GC-MS analysis of TP extracts was listed in Table S1. Limonene (68.44% w/w) and γ -terpinene (18.39% w/w) made up 85% of total TP extract. Limonene has been proved as an effective agent to inhibit the activity of spoilage bacteria, and its molecule containing multiple double bonds, which can be an effective barrier against lipid oxidation as an alternative of polyunsaturated fatty acids to offer electronics 42 . γ -terpinene was also recorded as an important preservative agent because of its antibacterial and antioxidant effects to some extent 2 . Additionally, many other compounds was known to be able to play positive roles to maintain the chemical properties of preserved materials 9,[43][44][45][46] . Therefore, as shown in Fig. 4, TP extract showed good preservation effect for frozen-stored tilapia with significant inhibition effects on of chemical degradation. The effects can accordingly reduce the variation in tissue structure and resulting fractal dimension of tilapia.

Conclusion
In this study, the frozen fish tissues with TP extracts treatment were analyzed for structural and chemical changes. During frozen storage, fractal dimension, which quantifies the porous structure formed in the tissue samples, as well as texture properties including hardness and springiness, and moisture content and water activity all decreased during the storage, while the extents of lipid oxidation, measured as peroxide value and thiobarbituric acid concentration, and protein degradation, monitored with total volatile basic nitrogen and trichloroacetic acid soluble peptides increased. The change rates of these parameters decreased with decreasing the storage temperature and by applying TP extract. With excellent correlations among each other, fractal dimension can reveal the variations of chemical parameters as well as texture parameters accurately. A shelf life prediction with using fractal dimension as the quality index demonstrated that TP extract could extend the shelf life of frozen tilapia by 35-45%.