Effects of Concentration of Soybean Protein Isolate and Maltose and Oil Phase Volume Fraction on Freeze–Thaw Stability of Pickering Emulsion

There is growing interest in enhancing the freeze–thaw stability of a Pickering emulsion to obtain a better taste in the frozen food field. A Pickering emulsion was prepared using a two-step homogenization method with soybean protein and maltose as raw materials. The outcomes showed that the freeze–thaw stability of the Pickering emulsion increased when prepared with an increase in soybean protein isolate (SPI) and maltose concentration. After three freeze–thaw treatments at 35 mg/mL, the Turbiscan Stability Index (TSI) value of the emulsion was the lowest. At this concentration, the surface hydrophobicity (H0) of the composite particles was 33.6 and the interfacial tension was 44.34 mN/m. Furthermore, the rheological nature of the emulsions proved that the apparent viscosity and viscoelasticity of Pickering emulsions grew with a growing oil phase volume fraction and concentration. The maximum value was reached in the case of the oil phase volume fraction of 50% at a concentration of 35 mg/mL, the apparent viscosity was 18 Pa·s, the storage modulus of the emulsion was 575 Pa, and the loss modulus was 152 Pa. This research is significant for the production of freeze–thaw resistant products, and improvement of protein-stabilized emulsion products with high freeze–thaw stability.


Introduction
Frozen fast food has received wide attention due to consideration of the needs inherent to a fast-paced life and the demand for better quality of products [1]. Freezing is a broadly used method that improves the shelf life of provisions, while the ice crystals composed throughout the freezing process constitute an important environmental stress, causing damage to the structure of provisions and ultimately generating an influence on the quality of products [2,3]. Therefore, it is worth producing emulsion-based products with an enhanced freeze-thaw stability. Pickering emulsion has excellent biocompatibility and a strong anti-aggregation ability. Solid particles adsorbed at the interface between oil and water in a Pickering emulsion can form a thick viscoelastic layer around the surface of oil droplets. During the freezing process, the viscoelastic layer in the Pickering emulsion resist the stress caused by ice expansion, giving it a certain freeze-thaw stability, and thus guaranteeing good product quality after the freezing treatment [4]. At present, the methods used to prepare a Pickering emulsion with high freeze-thaw stability include heating modification of protein, modification of the protein glycosylation, adjustment of ionic strength, and so on. For example, a steady thymol oil-in-water Pickering emulsion was prepared through electrostatic interaction between soluble almond gum and whey protein isolation. The combination of whey protein and soy polysaccharide using heat-induced polymerization was used to strengthen the freeze-thaw stability of the emulsion [5]. The use

Differential Scanning Calorimetry (DSC)
The 204 F1 DSC (NETZSCH-Gerätebau GmbH, Selb, Germany) measurement method was modified as stated by the previous study [23]. A small amount of the Pickering emulsion (10-20 mg) was placed on a dish. The temperature set range was from -50 °C to 40 °C. The cool temperature set range was from 40 ℃ to −50 ℃ at the rate of 5 ℃/min, balanced and then heated to 40°C at the same rate.

Statistical Analysis
The above experiments were carried out at least in triplicate, and the origin 2019 software was adopted for exploration and mapping. SPSS 20 software was adopted for oneway ANOVA and Duncan's test of the experimental data, with p < 0.05 as the great difference, and the outcomes were expressed by mean ± SD.

Particle Size and Potential Change of SPI-maltose
By measuring the complex potential and particle size of SPI-maltose at different concentrations, the properties of SPI-maltose particles were evaluated. Figure 1A shows the particle size (Dz) of SPI-maltose nanoparticles when they were developed at the concentration values of 15-35 mg/mL. The results proved the Dz of the produced composite particle gradually increased, which conformed to the findings of Chen et al.'s research [24]. Within the concentrations of 15-35 mg/mL, particle diameter increased with the increase in the concentrations, and a larger Dz was beneficial to obtain higher freeze-thaw stability for the Pickering emulsion. It was observed that the particle concentration had little impact on the potential, and the potential of the composite particles under different SPImaltose concentrations was in the range of −27 to −29 mV ( Figure 1B).

The H 0 of SPI-maltose
The interaction between two water-insoluble molecules was the key indicator to gauge the working feature of proteins. The H 0 of composite particles gradually increased with the increase in SPI and maltose concentrations ( Figure 2). When the concentration reached 35 mg/mL, the H 0 reached the maximum. Soy protein's conformation was changed by the interaction between SPI and maltose. Glycosylation causes reduced H 0 of the SPI, and heating during the Maillard reaction causes the unfolding of SPI conformation, thereby exposing the internal polar groups and reducing the hydrophobicity on the SPI surface [25]. However, as the concentration of the SPI increases, the hydrophobic amino acid content in the complex increases, which in turn increases the H 0 of the complex particles [26]. the SPI, and heating during the Maillard reaction causes the unfolding of SPI conformation, thereby exposing the internal polar groups and reducing the hydrophobicity on the SPI surface [25]. However, as the concentration of the SPI increases, the hydrophobic amino acid content in the complex increases, which in turn increases the H0 of the complex particles [26].  Figure 2. H0 at different concentration of SPI-maltose. Different letters (a-e) indicate statistically significant differences (p ≤ 0.05), according to ANOVA (one-way) and the Tukey test.

The Interfacial Tension of SPI-Maltose
Interfacial tension was the prerequisite for emulsification and foam generation, and a reduced level was closely relevant to the emulsifying and foaming capabilities. The interfacial tension of SPI-maltose nanoparticles was measured with an interfacial tension meter, and the outcomes were proven in Figure 3. The interfacial tension decreased from 46.62 mN/m to 44.34 mN/m, while the particle concentration rose from 15 mg/mL to 35 mg/mL. The results showed that increasing the concentration of SPI and maltose could reduce the interfacial tension. This was because, with the increase in concentration, the concentration of SPI-maltose molecules increased as well, which increased the possibility of aggregation of SPI-maltose. The enthalpy of aggregation may be lower than that of a single molecule, which made it tend to be adsorbed at the interface, thus reducing the interfacial tension. In addition, the contact probability between SPI-free amino groups and carbonyl groups at the reducing end of maltose increased, which promoted the reaction between SPI and maltose, leading to more SPI expansion and higher SPI flexibility, as well as reducing the interfacial tension of SPI-maltose molecules in oil and water. In addition, the contact probability between the free amino group and the carbonyl group of SPI at the reducing end of maltose increased, promoting the reaction, leading to more SPI unfolding and higher SPI flexibility, and reducing the interfacial tension of SPI-maltose molecules in oil and water. Lower interfacial tension can promote the formation rate of bridges between droplets and shorten the contact time between droplets. The shorter the contact time between droplets, the better the emulsion's stability, as it can better resist the pressure brought by ice crystals and improve the freeze-thaw stability of the emulsion [27].

The Interfacial Tension of SPI-maltose
Interfacial tension was the prerequisite for emulsification and foam generation, and a reduced level was closely relevant to the emulsifying and foaming capabilities. The interfacial tension of SPI-maltose nanoparticles was measured with an interfacial tension meter, and the outcomes were proven in Figure 3. The interfacial tension decreased from 46.62 mN/m to 44.34 mN/m, while the particle concentration rose from 15 mg/mL to 35 mg/mL. The results showed that increasing the concentration of SPI and maltose could reduce the interfacial tension. This was because, with the increase in concentration, the concentration of SPI-maltose molecules increased as well, which increased the possibility of aggregation of SPI-maltose. The enthalpy of aggregation may be lower than that of a single molecule, which made it tend to be adsorbed at the interface, thus reducing the interfacial tension. In addition, the contact probability between SPI-free amino groups and carbonyl groups at the reducing end of maltose increased, which promoted the reaction between SPI and maltose, leading to more SPI expansion and higher SPI flexibility, as well as reducing the interfacial tension of SPI-maltose molecules in oil and water. In addition, the contact probability between the free amino group and the carbonyl group of SPI at the reducing end of maltose increased, promoting the reaction, leading to more SPI unfolding and higher SPI flexibility, and reducing the interfacial tension of SPI-maltose molecules in oil and water. Lower interfacial tension can promote the formation rate of bridges between droplets and shorten the contact time between droplets. The shorter the contact time between droplets, the better the emulsion's stability, as it can better resist the pressure brought by ice crystals and improve the freeze-thaw stability of the emulsion [27].

Pickering Emulsion's Freeze-thaw Stability
Turbiscan stability analyzer is one of the methods that can quickly and reliably measure the stability of the Pickering emulsion. The freeze-thaw stability of the emulsion was evaluated by the change of TSI and backscattered light intensity (ΔBS) (Figure 4).

Pickering Emulsion's Freeze-thaw Stability
Turbiscan stability analyzer is one of the methods that can quickly and reliably measure the stability of the Pickering emulsion. The freeze-thaw stability of the emulsion was evaluated by the change of TSI and backscattered light intensity (∆BS) (Figure 4). Figure 3. Interfacial tension value of SPI-M prepared dispersion. Different letters (a-e) indicate statistically significant differences (p ≤ 0.05), according to ANOVA (one-way) and the Tukey test.

Pickering Emulsion's Freeze-thaw Stability
Turbiscan stability analyzer is one of the methods that can quickly and reliably measure the stability of the Pickering emulsion. The freeze-thaw stability of the emulsion was evaluated by the change of TSI and backscattered light intensity (ΔBS) (Figure 4). Firstly, the slope of the TSI curve of the Pickering emulsion after freeze-thaw cycle was higher than that of the initial emulsion, indicating that freezing would damage the Pickering emulsion and reduce its stability. This is because when the Pickering emulsion was frozen, oil phase or water phase in the Pickering emulsion crystallized to make the droplets gather, which destroyed the film at the oil-water interface of the Pickering emulsion, making the Pickering emulsion easier to agglomerate during the dissolution process, and thus reducing its stability [21]. By increasing the concentration of SPI and maltose,  Firstly, the slope of the TSI curve of the Pickering emulsion after freeze-thaw cycle was higher than that of the initial emulsion, indicating that freezing would damage the Pickering emulsion and reduce its stability. This is because when the Pickering emulsion was frozen, oil phase or water phase in the Pickering emulsion crystallized to make the droplets gather, which destroyed the film at the oil-water interface of the Pickering emulsion, making the Pickering emulsion easier to agglomerate during the dissolution process, and thus reducing its stability [21]. By increasing the concentration of SPI and maltose, the overall TSI of the initial emulsion was reduced from 4.9 to 1.8. The TSI of the prepared Pickering emulsion after three freeze-thaw cycles was 2.9 when the oil volume was 40% and the concentration was 35 mg/mL, which still showed good stability. The reason for this phenomenon was that, at low particle concentration, the emulsion was unstable because the droplet was partially covered by SPI-maltose nanoparticles, and there was almost no excess SPI-maltose particles in the continuous phase. By increasing SPI-maltose concentration, the aggregate of excessive SPI-maltose particles in the continuous phase changed from dense flocs to loose networks, so that the droplets kept a good separation from each other. On the other hand, the lower SPI and maltose concentration might form a thin polymer layer on the surface of the droplets, resulting in the easy aggregation and coalescence of the droplets. As oil phase fraction increased, the overall TSI of the initial emulsion reduced from 4.7 to 0.2. When the concentration was 35 mg/mL and oil phase was 50%, the TSI of freeze-thaw treatment for three times was 0.7. At this time, the Pickering emulsion still had a better stability. This was because when dispersed phase volume fraction increased within a certain range, more oil droplets in the emulsion could be filled into the emulsion layer, and the crosslinked gel structure was tighter, which made the Pickering emulsion system more stable so that it could resist the damage caused by ice crystals during freezing [28]. The above experimental results can be further confirmed by ∆BS ( Figures 5 and 6). ∆BS can evaluate the change of concentration or droplet size in the Pickering emulsion. In Figure 5, it is proved that the Pickering emulsion ∆BS changes with different concentrations of SPI and maltose before and after a freeze-thaw cycle treatment. After freeze-thaw cycle treatment of emulsion, with the increase in freeze-thaw cycle times, ∆BS increased gradually; however, as the concentration increased, ∆BS gradually dropped from −65% to −20% at the bottom of freeze-thaw lotion, while ∆BS decreased from 36% to 10%. The results show that higher particle concentration (35 mg/mL) can improve its ability to stabilize emulsion, and the emulsion droplets are not easy to flocculate or coalesce. The emulsion formed by higher particle concentration still has a certain stability after freeze-thaw cycle treatment. It may be that, with growing concentration, the H 0 of particles increased and the interfacial tension decreased, which is conducive to more effective adsorption of particles at the oil-water interface, so as to improve freeze-thaw stability of emulsions [29,30]. emulsion reduced from 4.7 to 0.2. When the concentration was 35 mg/mL and oil phase was 50%, the TSI of freeze-thaw treatment for three times was 0.7. At this time, the Pickering emulsion still had a better stability. This was because when dispersed phase volume fraction increased within a certain range, more oil droplets in the emulsion could be filled into the emulsion layer, and the crosslinked gel structure was tighter, which made the Pickering emulsion system more stable so that it could resist the damage caused by ice crystals during freezing [28].
The above experimental results can be further confirmed by ΔBS (Figures 5 and 6). ΔBS can evaluate the change of concentration or droplet size in the Pickering emulsion. In Figure 5, it is proved that the Pickering emulsion ΔBS changes with different concentrations of SPI and maltose before and after a freeze-thaw cycle treatment. After freeze-thaw cycle treatment of emulsion, with the increase in freeze-thaw cycle times, ΔBS increased gradually; however, as the concentration increased, ΔBS gradually dropped from −65% to −20% at the bottom of freeze-thaw lotion, while ΔBS decreased from 36% to 10%. The results show that higher particle concentration (35 mg/mL) can improve its ability to stabilize emulsion, and the emulsion droplets are not easy to flocculate or coalesce. The emulsion formed by higher particle concentration still has a certain stability after freeze-thaw cycle treatment. It may be that, with growing concentration, the H0 of particles increased and the interfacial tension decreased, which is conducive to more effective adsorption of particles at the oil-water interface, so as to improve freeze-thaw stability of emulsions [29,30].           To illustrate the potential mechanism of improving the stability of the emulsion by increasing dispersed phase volume fraction, Figure 6 shows the stability of the initial emulsion and the freeze-thaw emulsion with different dispersed phase volume fraction ΔBS. With the dispersed phase volume fraction doubling in the range of 10-50%, the ΔBS gradually decreased from −65% to −20%, while the top ΔBS decreased from about 35% to 10% ( Figure 6). This phenomenon, combined with the rheological results of the emulsion, shows that the interaction between oil droplets affects the viscoelasticity of the emulsion. By increasing the stiffness of the emulsion's gel network structure, the freeze-thaw stability of the emulsion can be improved. Therefore, the volume fraction of oil phase and particle concentration can be appropriately increased to make the emulsion have better freeze-thaw stability [31].

The CLSM of Pickering Emulsion
According to the results of stability analysis, the initial emulsion and frozen processing emulsion under the best conditions of the Pickering emulsion frozen stability were selected. The oil phase indicator was red, and the protein indicator was green. The CLSM (Figure 7) showed that the Pickering emulsion formed under this condition still guaranteed its own stability after one and two freeze-thaw cycles. As stated by the outcomes of protein staining, the protein's fluorescence intensity was weak and a large amount of aggregation was not found, indicating that the protein was in a state of adsorption at the oilwater interface and the stability of emulsion was excellent. The oil phase showed obvious large particles aggregated, and the protein signal was meaningfully enhanced and appeared aggregated when the samples carried out three freeze-thaw cycles. The particle layer formed by irreversibly adsorbed particles on the interface physically blocks the coalescence of droplets and the sharing of gel particles by oil droplets. It can spatially obstruct the agglomeration of droplets through bridging and ensure the stability of the emulsion; thereby, the freeze-thaw stability was enhanced. That is, the freeze-thaw stability of Pickering emulsion may be consequently enhanced by adapting particles' adsorption at the oil-water interface [23].   To illustrate the potential mechanism of improving the stability of the emulsion by increasing dispersed phase volume fraction, Figure 6 shows the stability of the initial emulsion and the freeze-thaw emulsion with different dispersed phase volume fraction ∆BS. With the dispersed phase volume fraction doubling in the range of 10-50%, the ∆BS gradually decreased from −65% to −20%, while the top ∆BS decreased from about 35% to 10% ( Figure 6). This phenomenon, combined with the rheological results of the emulsion, shows that the interaction between oil droplets affects the viscoelasticity of the emulsion. By increasing the stiffness of the emulsion's gel network structure, the freezethaw stability of the emulsion can be improved. Therefore, the volume fraction of oil phase and particle concentration can be appropriately increased to make the emulsion have better freeze-thaw stability [31].

The CLSM of Pickering Emulsion
According to the results of stability analysis, the initial emulsion and frozen processing emulsion under the best conditions of the Pickering emulsion frozen stability were selected. The oil phase indicator was red, and the protein indicator was green. The CLSM (Figure 7) showed that the Pickering emulsion formed under this condition still guaranteed its own stability after one and two freeze-thaw cycles. As stated by the outcomes of protein staining, the protein's fluorescence intensity was weak and a large amount of aggregation was not found, indicating that the protein was in a state of adsorption at the oil-water interface and the stability of emulsion was excellent. The oil phase showed obvious large particles aggregated, and the protein signal was meaningfully enhanced and appeared aggregated when the samples carried out three freeze-thaw cycles. The particle layer formed by irreversibly adsorbed particles on the interface physically blocks the coalescence of droplets and the sharing of gel particles by oil droplets. It can spatially obstruct the agglomeration of droplets through bridging and ensure the stability of the emulsion; thereby, the freeze-thaw stability was enhanced. That is, the freeze-thaw stability of Pickering emulsion may be consequently enhanced by adapting particles' adsorption at the oil-water interface [23].

Rheological Properties of the Pickering Emulsion
In order to clearly show the mechanism of freeze-thaw stability of the Pickering emulsion, the apparent viscosity, G' and G'', of the initial Pickering emulsions was characterized. The apparent viscosity of the emulsion prepared with SPI-maltose at different concentrations and different oil volume fractions was shown in Figure 8. As shear rate increased, the fluidity of the emulsion declined, and the viscosity of the initial emulsion grew with the increased concentration. When the concentration was 35 mg/mL, the apparent viscosity of the Pickering emulsion increased as the volume proportion of oil phase increased. When the particle concentration was 35 mg/mL and the volume of oil phase was 50%, the maximum apparent viscosity was 18 Pa·s. This is because the increase in SPI-maltose concentration and oil volume fraction enhanced the adhesion between the Pickering emulsion droplets, and the spatial network structure between the emulsion is more compact, causing it to show greater apparent viscosity [32,33].

Rheological Properties of the Pickering Emulsion
In order to clearly show the mechanism of freeze-thaw stability of the Pickering emulsion, the apparent viscosity, G and G , of the initial Pickering emulsions was characterized. The apparent viscosity of the emulsion prepared with SPI-maltose at different concentrations and different oil volume fractions was shown in Figure 8. As shear rate increased, the fluidity of the emulsion declined, and the viscosity of the initial emulsion grew with the increased concentration. When the concentration was 35 mg/mL, the apparent viscosity of the Pickering emulsion increased as the volume proportion of oil phase increased. When the particle concentration was 35 mg/mL and the volume of oil phase was 50%, the maximum apparent viscosity was 18 Pa·s. This is because the increase in SPI-maltose concentration and oil volume fraction enhanced the adhesion between the Pickering emulsion droplets, and the spatial network structure between the emulsion is more compact, causing it to show greater apparent viscosity [32,33].
Secondly, the apparent viscosity of the emulsion at different concentrations of SPImaltose and dispersed phase volume fractions under different shear rate γ were determined, and fitted the apparent viscosity of the emulsion according to the power law model. R 2 in the emulsion fitting equation is 0.98-0.995, indicating that the flow curve at this time is more in line with the power law model. The consistency parameter K and flow index in the power law equation were empirical constants related to the properties of liquid. K is a measure of liquid viscosity; the larger the K, the more viscous the emulsion. n value was a measure of the degree of pseudoplasticity. If n value is less than 1, it is a shear thinning pseudoplastic fluid [34]. The greater the deviation of n value from Table 1, the easier the shear thinning is; that is, the greater the degree of pseudoplasticity. It was found from Tables 1 and 2 that the flow index of all samples is n < 1, showing pseudoplastic fluid properties. As the shear rate increased, the apparent viscosity of the Pickering emulsion composed of various components dropped sharply, leading to the shear thinning effect. Firstly, the solid particles in the emulsion were adsorbed on the oil-water interface, and the oil droplets were dispersed in the water. However, the components flow at different speeds, and the emulsion could not keep this state for long. When the emulsion was flowing, every long chain molecule would enter the flow layer with the same flow rate as far as possible. Secondly, the SPI-maltose molecule had a long molecular chain. In the static state, these substances maintained irregular internal order and had high internal resistance to hinder their movement. With the increase in shear rate, the rearrangement of substances led to the decrease in resistance and apparent viscosity of emulsions. With the increase in SPI and maltose concentration, the volume fraction of the oil phase increased, the irregular internal order was maintained, the viscosity increased, and the internal resistance was large, which hindered the flow of emulsion. However, as the SPI-maltose concentration increased and the volume fraction of the oil phase grew, the internal irregular order is maintained, the viscosity increases, and the internal resistance is large, thus hindering the flow of the emulsion. It shows that increasing the concentration of SPI and maltose and increasing the volume ratio of the oil phase are beneficial to enhancing the fluid nature of the Pickering emulsion and improving the freeze-thaw stability of Pickering emulsion [35,36].   Values are expressed as mean ± expanded uncertainty limit (EUL). a-e Various superscripts in the same column stand for great diversities (p < 0.05).
In addition, the G of all emulsion is greater than the G , indicating that a gel network structure guided by spring is formed in the emulsion system. At a frequency in the scope of 0.1-100 rad/s, the G and G of initial emulsions grew in line with the increase in particle concentrations and the dispersed phase volume fraction (Figure 9). When the concentration was 35 mg/mL and the volume fraction of oil phase was 50%, the maximum value of G was 100 Pa and the maximum value of G was 575 Pa. The outcomes showed that the viscoelasticity of the Pickering emulsion could be improved, the damage of ice crystals to the emulsion could be reduced, and the freeze-thaw stability of the Pickering emulsion could be improved by increasing the concentration and dispersed phase volume fraction to a certain extent [37]. The crosslinking of protein molecules increases the penetration resistance of the interface layer during thawing and freezing, thus efficiently controlling the migration of oil droplets and strengthening the freeze-thaw stability of the emulsion gel [2].

Pickering Emulsion's Thermal Properties
The alteration of the crystallization point for the emulsion was also one of the methods used to improve the Pickering emulsion's freeze-thaw stability [38]. To understand the effects of SPI and maltose concentration and oil-water ratio on the temperature of

Pickering Emulsion's Thermal Properties
The alteration of the crystallization point for the emulsion was also one of the methods used to improve the Pickering emulsion's freeze-thaw stability [38]. To understand the effects of SPI and maltose concentration and oil-water ratio on the temperature of crystallization and melting during freezing of SPI-M preparation of the Pickering emulsion, we recorded the DSC thermal spectrum of the Pickering emulsion. Figure 10A shows the DSC thermal spectrum of the Pickering emulsion prepared at different concentrations of SPI and maltose, and Figure 10B is the DSC thermal spectrum of the Pickering emulsion prepared with different oil phase volume fractions.

Pickering Emulsion's Thermal Properties
The alteration of the crystallization point for the emulsion was also one of the methods used to improve the Pickering emulsion's freeze-thaw stability [38]. To understand the effects of SPI and maltose concentration and oil-water ratio on the temperature of crystallization and melting during freezing of SPI-M preparation of the Pickering emulsion, we recorded the DSC thermal spectrum of the Pickering emulsion. Figure 10A shows the DSC thermal spectrum of the Pickering emulsion prepared at different concentrations of SPI and maltose, and Figure 10B is the DSC thermal spectrum of the Pickering emulsion prepared with different oil phase volume fractions.  Figure 10A shows that the crystallization temperature of the Pickering emulsion prepared under different concentrations of SPI and maltose was −12.05 to −14.29 °C. As the concentration grew, the crystallization temperature of the emulsion first declined and then increased. This may be because the increase in SPI and maltose concentration accelerated the reaction speed, increased the particle size of SPI-maltose particles, increased the number of nucleation in the reaction, and lowered the crystallization temperature of the emulsion. In addition, the high concentration of SPI and maltose formed a thick interface film at the interface, which provided great resistance to the phase transition, thus improving the freeze-thaw stability of the emulsion. In addition, the high concentration of SPI and maltose formed a thick interface film at the interface, which provided great resistance to the phase transition, thus improving the freeze-thaw stability of the emulsion. According to Figure 10B, the crystallization temperature of the Pickering emulsion  Figure 10A shows that the crystallization temperature of the Pickering emulsion prepared under different concentrations of SPI and maltose was −12.05 to −14.29 • C. As the concentration grew, the crystallization temperature of the emulsion first declined and then increased. This may be because the increase in SPI and maltose concentration accelerated the reaction speed, increased the particle size of SPI-maltose particles, increased the number of nucleation in the reaction, and lowered the crystallization temperature of the emulsion. In addition, the high concentration of SPI and maltose formed a thick interface film at the interface, which provided great resistance to the phase transition, thus improving the freeze-thaw stability of the emulsion. In addition, the high concentration of SPI and maltose formed a thick interface film at the interface, which provided great resistance to the phase transition, thus improving the freeze-thaw stability of the emulsion. According to Figure 10B, the crystallization temperature of the Pickering emulsion prepared by different oil ratios is from −11.17 to −15.22 • C, which may be due to the crystallization temperature of soybean oil being −15 • C; thus, proper addition of soybean oil could reduce the crystallization temperature of emulsion.

Conclusions
The research showed that the interaction between particles and the highly viscoelastic interface film formed were the key factors for determining the freeze-thaw stability of the Pickering emulsion. The higher SPI-maltose concentration made the particles develop a larger size, higher H 0 , and firmer interior integrity. Gel-like flow behavior was exhibited by all the initial emulsions. Emulsions with higher oil phase volume and particle concentration possessed higher viscosity, lower fluidity and a viscoelastic layer. Freeze-thaw stability of these emulsions changed with the cycles of freeze-thaw, particle concentration, and dispersed phase volume fraction. Compared with the emulsion with lower particle concentration, the emulsion with higher particle concentration can prevent emulsion and coalescence, thus showing better freeze-thaw stability. Higher oil or higher particle concentration phase volume has better freeze-thaw stability than that of lower particle con-centration and lower oil phase volume. This phenomenon may be generated by Pickering steric stability and the network structure of gel-like. These findings are of great significance for the preparation of high Pickering emulsions with the potential for freeze-thaw stability, which can be used in the production of frozen foods.