Pea Protein for Hempseed Oil Nanoemulsion Stabilization

In this paper, we present the possibility of using pea protein isolates as a stabilizer for hempseed oil (HSO)-based water/oil emulsions in conjunction with lecithin as a co-surfactant. A Box-Behnken design was employed to build polynomial models for optimization of the ultrasonication process to prepare the emulsions. The stability of the system was verified by droplet size measurements using dynamic light scattering (DLS) as well as centrifugation and thermal challenge tests. The z-ave droplet diameters of optimized emulsion were 209 and 207 nm after preparation and 1 week storage, respectively. The concentration of free Linoleic acid (C18:2; n-6) was used for calculation of entrapment efficiency in prepared nanoemulsions. At optimum conditions of the process, up to 98.63% ± 1.95 of entrapment was achieved. FTIR analysis and rheological tests were also performed to evaluate the quality of oil and emulsion, and to verify the close-to-water like behavior of the prepared samples compared to the viscous nature of the original oil. Obtained results confirmed the high impact of lecithin and pea protein concentrations on the emulsion droplet size and homogeneity confirmed by microscopic imaging. The presented results are the first steps towards using hempseed oil-based emulsions as a potential food additive carrier, such as flavor. Furthermore, the good stability of the prepared nanoemulsion gives opportunities for potential use in biomedical and cosmetic applications.


Introduction
Emulsion is a thermodynamically unstable state of the mixed phases which tend to rapidly separate. Two major categories of emulsion systems frequently used in food industry can be distinguished as: oil-in-water (O/W) and water-in-oil (W/O). There are many factors affecting the emulsion stability such as concentration, preparation techniques, presence of surfactants (stabilizers), temperature, storage conditions, etc. For the development of novel food and biomedical and pharmacological products,

Results and Discussion
Using the optimum point values for various variables affecting the emulsion process can significantly improve the quality and properties of nanoemulsion. With an efficient optimization of these variables including concentration of protein and lecithin as well as ultra-sound process time, a HSO nanoemulsion with reasonable stability can be obtained. In this research, based on preliminary studies regarding factors affecting the properties of HSO including nanoemulsion particle size (PS), zeta potential (ZP) and polidispersity index (pdi), three independent variables of pea protein (0.2-1.4%), lecithin (0.0-6.0%) and time of ultra-sound process (0-20 min) were selected. Based on the Box-Behnken design, 15 runs of the experiment with two replications and three trials of the center points were tested. The results for PS, ZP and pdi responses are presented in Table 1.

Effect of Process Condition on Nano-Emulsion Properties
Applying an experimental design for process condition, response variables of hydrodynamic diameter z-ave (PS) and zeta potential (ZP) were recorded from 215-806 (nm) and −12.5-−27.3 (mV) with the pdi from 0.233 to 0.787 (Table 1). Using regression analysis and ANOVA and considering the significant probability (p-value, F-ratio) effects of nanoemulsion process variables on nanoemulsion properties, quadratic model equations with R 2 from 92.93-94.34% were developed based on regression coefficients values to predict the responses of PS (nm), ZP(mV) and pdi (y 1 , y 2 and y 3 ) to independent variables of protein concentration (%), lecithin concentration (%), and time of process (min) (x 1 , x 2 and x 3 ). y 1 = 316.314 − 172.921x 3 (1) It was found that the time of the ultrasound process had the most significant effect (p < 0.05) on all the response variables, especially on the particle size of emulsion. However, using different percentages of protein and lecithin showed no linear or quadratic effect (p > 0.05) on this response. Interaction effects of the process condition on all the emulsion properties were also found to be insignificant (p > 0.05). For better understanding the relation and effects of process factors on response variables, 3D graphical methodology are presented in Figure 1.
It was found that the time of the ultrasound process had the most significant effect (p < 0.05) on all the response variables, especially on the particle size of emulsion. However, using different percentages of protein and lecithin showed no linear or quadratic effect (p > 0.05) on this response. Interaction effects of the process condition on all the emulsion properties were also found to be insignificant (p > 0.05). For better understanding the relation and effects of process factors on response variables, 3D graphical methodology are presented in Figure 1.

Optimization Procedure
Using graphical and numerical optimization for desired approaches, the optimum condition for the nanoemulsion process based on minimum hydrodynamic diameter z-ave (nm), polydispersity index (pdi) and maximum of zeta potential (mV), was developed. The overall optimal condition resulted in the highest amounts of zeta potential (−27.3 mV), and the smallest values for particle size and pdi (209 nm and 0.239, respectively) were predicted to be achieved at combined levels of 0.4% pea protein, 5.0% lecithin and 18.0 min ultra-sound with composite desirability of 0.985. The optimal levels of parameters were verified for the experimental responses as predicted by the mathematical

Optimization Procedure
Using graphical and numerical optimization for desired approaches, the optimum condition for the nanoemulsion process based on minimum hydrodynamic diameter z-ave (nm), polydispersity index (pdi) and maximum of zeta potential (mV), was developed. The overall optimal condition resulted in the highest amounts of zeta potential (−27.3 mV), and the smallest values for particle size and pdi (209 nm and 0.239, respectively) were predicted to be achieved at combined levels of 0.4% pea protein, 5.0% lecithin and 18.0 min ultra-sound with composite desirability of 0.985. The optimal levels of parameters were verified for the experimental responses as predicted by the mathematical model presented in Table 2. It can be seen that there was no significant difference between experimental and predicted values (p > 0.05) for all of the responses (error < 5.00%).

Optimized Emulsion Properties
Using the optimum conditions for preparing HSO emulsions resulted in 98.63% ± 1.95 and 92.72% ± 2.21 entrapment efficiency (based on the linoleic acid released in emulsion medium) for EMOP and EMO samples, respectively, indicating a significant effect (p < 0.05) of pea protein on HSO entrapment. One of the protein characteristics is surface hydrophobicity, resulting from hydrophobic groups present on the protein surface. These groups are partially denaturated proteins which provide greater potential for adsorption of oil at the oil/water interface [24].

Particle Size and Zeta Potential Results
Particle and droplet size is one of the crucial parameters of emulsion stability. The stability of nanoemulsion system might be improved by the application of ultrasounds [25]. To increase the stability of HSO base emulsion, the ultrasounds process was used for preparation of the emulsion after the ingredients were mixed by a homogenizer. Then the emulsion droplet size was evaluated by DLS. Figure 2 shows the droplet size distribution of the optimized samples with (EMOP) and without (EMO) pea protein, obtained by DLS. Both the homogenization process and pea protein had significant impacts on final emulsion droplet size. Smaller droplet z-ave (209 nm) were observed in EMOP than EMO z-ave (309 nm). To make smaller droplets and increase the stability, the interfacial tension between the droplet must be decreased. A dding protein molecules such as pea protein to an emulsion system can result in decreasing interfacial tension, leading to lower Laplace pressure and smaller droplets. However, with a high concentration of protein, the interfacial elasticity of droplets decreases, resulting in less stability [26].  Table 2. It can be seen that there was no significant difference between experimental and predicted values (p > 0.05) for all of the responses (error < 5.00%).

Optimized Emulsion Properties
Using the optimum conditions for preparing HSO emulsions resulted in 98.63% ± 1.95 and 92.72% ± 2.21 entrapment efficiency (based on the linoleic acid released in emulsion medium) for EMOP and EMO samples, respectively, indicating a significant effect (p < 0.05) of pea protein on HSO entrapment. One of the protein characteristics is surface hydrophobicity, resulting from hydrophobic groups present on the protein surface. These groups are partially denaturated proteins which provide greater potential for adsorption of oil at the oil/water interface [24].

Particle Size and Zeta Potential Results
Particle and droplet size is one of the crucial parameters of emulsion stability. The stability of nanoemulsion system might be improved by the application of ultrasounds [25]. To increase the stability of HSO base emulsion, the ultrasounds process was used for preparation of the emulsion after the ingredients were mixed by a homogenizer. Then the emulsion droplet size was evaluated by DLS. Figure 2 shows the droplet size distribution of the optimized samples with (EMOP) and without (EMO) pea protein, obtained by DLS. Both the homogenization process and pea protein had significant impacts on final emulsion droplet size. Smaller droplet z-ave (209 nm) were observed in EMOP than EMO z-ave (309 nm). To make smaller droplets and increase the stability, the interfacial tension between the droplet must be decreased. Adding protein molecules such as pea protein to an emulsion system can result in decreasing interfacial tension, leading to lower Laplace pressure and smaller droplets. However, with a high concentration of protein, the interfacial elasticity of droplets decreases, resulting in less stability [26]. In both samples, no significant differences (p > 0.05) were observed in ZP results. Preliminary stability tests of the emulsions stored in a cold room at 4 • C were performed after 1 week. DLS results ( Figure 2) showed that the maximum peak of droplet size distribution decreased from 265 to 250 nm and from 543 to 272 nm for EMOP and EMO, respectively. A decreasing droplet size during storage of the nanoemulsion has already been observed [27,28]. Badolato et al. [29] suggested that the less stable emulsion system possibly loses larger droplets (according to investigated storage conditions). Hence, detailed stability tests were also performed in the current study (see Section 2.4.).

FTIR Results
Fourier transform infrared spectroscopy (FTIR) is one of the fastest spectroscopic methods for basic oil quality determination [30]. Figure 3 presents typical spectra obtained by FTIR for pure HSO from Canadian manufactured and two optimized emulsions (EMOP and EMO). There is no significant differences (p > 0.05) in spectra between samples with and without pea protein. Siano et al. [31] performed more detailed studies using Attenuated Total Reflectance-Fourier Transform Infrared (ATR-FTIR) of seeds and oils from edible fedora cultivar hemp (Cannabis sativa L.). They pointed out that signals registered around 1700 cm −1 belonged to a carbonyl band. Moreover, they indicated characteristic regions of Cannabis sativa L. at FTIR spectra for moisture, lipid, protein and carbohydrate. Our studies confirmed the presence of the strong signals at around 1700 cm −1 . Furthermore, the signal was weaker in prepared emulsions. A broad signal occurred in the emulsions from 3200-3300 cm −1 . It was noticed that some of the signals (peaks maxima) from HSO were relatively lower in the investigated emulsion. It might be caused by the low oil and high water concentrations, as well as the addition of surfactants in the emulsions. In both samples, no significant differences (p > 0.05) were observed in ZP results. Preliminary stability tests of the emulsions stored in a cold room at 4 °C were performed after 1 week. DLS results ( Figure 2) showed that the maximum peak of droplet size distribution decreased from 265 to 250 nm and from 543 to 272 nm for EMOP and EMO, respectively. A decreasing droplet size during storage of the nanoemulsion has already been observed [27,28]. Badolato et al. [29] suggested that the less stable emulsion system possibly loses larger droplets (according to investigated storage conditions). Hence, detailed stability tests were also performed in the current study (see Section 2.4.).

FTIR Results
Fourier transform infrared spectroscopy (FTIR) is one of the fastest spectroscopic methods for basic oil quality determination [30]. Figure 3 presents typical spectra obtained by FTIR for pure HSO from Canadian manufactured and two optimized emulsions (EMOP and EMO). There is no significant differences (p > 0.05) in spectra between samples with and without pea protein. Siano et al. [31] performed more detailed studies using Attenuated Total Reflectance-Fourier Transform Infrared (ATR-FTIR) of seeds and oils from edible fedora cultivar hemp (Cannabis sativa L.). They pointed out that signals registered around 1700 cm −1 belonged to a carbonyl band. Moreover, they indicated characteristic regions of Cannabis sativa L. at FTIR spectra for moisture, lipid, protein and carbohydrate. Our studies confirmed the presence of the strong signals at around 1700 cm −1 . Furthermore, the signal was weaker in prepared emulsions. A broad signal occurred in the emulsions from 3200-3300 cm −1 . It was noticed that some of the signals (peaks maxima) from HSO were relatively lower in the investigated emulsion. It might be caused by the low oil and high water concentrations, as well as the addition of surfactants in the emulsions.

Rheological and Color Tests
Simple rheological tests on shear viscosity were performed to compare the behavior of prepared emulsion systems under continuous shear conditions. Results from rheological tests can provide valuable insight into the molecular interactions within the emulsion system, which may influence the stability as well as sensory properties. In conclusion, hempseed oil emulsions were showed to have water-like behaviors. Both EMOP and EMO emulsion systems had similar but slightly higher viscosity than water (Figure 4). The water-like behavior of hempseed oil emulsions was not

Rheological and Color Tests
Simple rheological tests on shear viscosity were performed to compare the behavior of prepared emulsion systems under continuous shear conditions. Results from rheological tests can provide valuable insight into the molecular interactions within the emulsion system, which may influence the stability as well as sensory properties. In conclusion, hempseed oil emulsions were showed to have water-like behaviors. Both EMOP and EMO emulsion systems had similar but slightly higher viscosity than water (Figure 4). The water-like behavior of hempseed oil emulsions was not significantly influenced by the addition of pea protein as a stabilizer. Besides, power law model (Ostwald de Waele model) was chosen to quantify the dependency of shear stress on shear rate by the power law indexes (n) because of its simplicity as well as the good fit to the data. Both EMO (n = 0.9598) and EMOP (n = 0.9583) showed near Newtonian behavior. These results indicated that the viscosity of samples were virtually independent from applied shear rates, which were in accordance with results reported by Demetriades et al. [32]. With low viscosity and near Newtonian behaviors, both EMO and EMOP can be more easily utilized in food production processes than other non-Newtonian fluids.
significantly influenced by the addition of pea protein as a stabilizer. Besides, power law model (Ostwald de Waele model) was chosen to quantify the dependency of shear stress on shear rate by the power law indexes (n) because of its simplicity as well as the good fit to the data. Both EMO (n = 0.9598) and EMOP (n = 0.9583) showed near Newtonian behavior. These results indicated that the viscosity of samples were virtually independent from applied shear rates, which were in accordance with results reported by Demetriades et al. [32]. With low viscosity and near Newtonian behaviors, both EMO and EMOP can be more easily utilized in food production processes than other non-Newtonian fluids. Moreover, it was shown that EMOP had slightly higher viscosities than EMO given the same volume fraction of the oil phase ( Figure 4). The higher shear viscosity caused by additional pea protein can be accounted for by bridging flocculation of dispersing droplets. Other studies also confirmed that flocculated emulsions had higher viscosity. Flocculated droplets in emulsion systems are able to trap parts of the continuous phase, exhibiting a higher effective volume fraction of the dispersion phase. Therefore, with the same actual volume fraction phase, EMOP had slightly higher viscosities.
Simple CIE L*a*b* analysis for the evaluation of color properties of the prepared emulsions showed more lightness, slightly less green tonality and more yellow shifting for EMO (L/a/b:76.07/−2.28/10.31) than EMOP emulsion (L/a/b: 75.21/−2.44/9.94).

Stability of the Emulsion
For the optimized samples (EMOP and EMO), detailed stability tests were performed. The time and temperature dependence were investigated and verified by DLS. Alvarado et al. [33] suggested that DLS is favorable for small droplet size and mostly for oil-in-water emulsion systems and might be used for possible coalescence evaluation. In our studies, the prepared emulsion systems have been investigated using the procedure described in p. 2.3.1 in different periods of time. For the comparison of the results, the concentration of the samples was kept the same for all measurements.
The results presented in Table 3 showed that after 1 week of storage, emulsion droplet size slightly decreased at room temperature and 50 °C. High emulsion droplet size stability at 50 °C suggested that a higher temperature has a preserving impact on nanoemulsion. An increasing droplet size and unpleasant smell were noticed in the samples stored at 37 °C (no detailed experiments were conducted). The homogeneity of the samples was evaluated using an optical microscope in visible and fluorescence mode. No coalescence of the droplets were observed. Moreover, a weak signal from the fluorescent part of oil was observed ( Figure 5C), but no detailed studies were performed. Moreover, it was shown that EMOP had slightly higher viscosities than EMO given the same volume fraction of the oil phase ( Figure 4). The higher shear viscosity caused by additional pea protein can be accounted for by bridging flocculation of dispersing droplets. Other studies also confirmed that flocculated emulsions had higher viscosity. Flocculated droplets in emulsion systems are able to trap parts of the continuous phase, exhibiting a higher effective volume fraction of the dispersion phase. Therefore, with the same actual volume fraction phase, EMOP had slightly higher viscosities.
Simple CIE L*a*b* analysis for the evaluation of color properties of the prepared emulsions showed more lightness, slightly less green tonality and more yellow shifting for EMO (L/a/b:76.07/−2.28/10.31) than EMOP emulsion (L/a/b: 75.21/−2.44/9.94).

Stability of the Emulsion
For the optimized samples (EMOP and EMO), detailed stability tests were performed. The time and temperature dependence were investigated and verified by DLS. A lvarado et al. [33] suggested that DLS is favorable for small droplet size and mostly for oil-in-water emulsion systems and might be used for possible coalescence evaluation. In our studies, the prepared emulsion systems have been investigated using the procedure described in Section 2.3.1 in different periods of time. For the comparison of the results, the concentration of the samples was kept the same for all measurements.
The results presented in Table 3 showed that after 1 week of storage, emulsion droplet size slightly decreased at room temperature and 50 • C. High emulsion droplet size stability at 50 • C suggested that a higher temperature has a preserving impact on nanoemulsion. A n increasing droplet size and unpleasant smell were noticed in the samples stored at 37 • C (no detailed experiments were conducted). The homogeneity of the samples was evaluated using an optical microscope in visible and fluorescence mode. No coalescence of the droplets were observed. Moreover, a weak signal from the fluorescent part of oil was observed ( Figure 5C), but no detailed studies were performed.   Furthermore, emulsion stability was monitored under gravitational and centrifugal forces. The emulsion with the protein (EMOP) showed better stability in both tests. Under the gravitational forces, sediments only occurred in the emulsion without the protein (EMO) after 2 weeks of storage ( Figure 5). Generally, all emulsions are thermodynamically unstable. If they remain long enough, they will be separated into several layers, consisting of the top layer of oil droplets (cream) and the bottom water layer (serum). In order to test this phenomenon, centrifugal forces were used to accelerate this process. Under 5000 rpm, EMOP showed little sediments and no creaming layer while the EMO showed more sediment and a slight creaming layer ( Figure 5). For the 7500 rpm test, EMOP still showed fewer sediments and no creaming layer. In contrast, the EMO had more sediments and an obvious creaming layer ( Figure 5). Under 10,000 and 12,500 rpm, both EMOP and EMO showed a similar amount of sediments properly as they were both completely precipitated. However, EMO had a greater creaming layer compared to the EMOP ( Figure 5). A ll the results showed that the EMOP had better stability compared to the EMO, which may result from more viscoelastic films formed by additional proteins.
Thermal resistance was tested based on three temperatures, which were room temperature, 37 • C and 50 • C. Particle size, zeta potential and entrapment efficiency were selected as three key factors to determine their thermal resistance. The temperatures were chosen according to their possible use in food, cosmetics or biomedical applications considering shelf-stable storage in room temperatures and possible temperature abuse in hot climates. The particle size and zeta potential of both EMO and EMOP were constant after 1 week of storage under room temperature and 50 • C, but significantly changed under 37 • C (Table 3). This was probably because 37 • C is more suitable for bacteria growth which might break down the emulsion system. The unpleasant smells of the samples after 1 week of storage under 37 • C confirmed this. A ll the EMOP samples also showed higher values for entrapment efficiency compared to EMO samples, suggesting less hempseed oil was released during 1 week of storage under three different temperatures ( Table 3). The viscoelastic films formed by the protein can protect the oil phase in the particle and maximize the stability of the emulsion system.
The color analysis results for EMO and EMOP emulsion stored at room temperature, 37 • C and 50 • C showed no significant changes (p < 0.05) after 1 week of storage.

Materials
Unrefined cold-pressed hempseed oil (HSO) was purchased from Manitoba Harvest Hemp Foods (Winnipeg, MB, Canada). Refined lecithin was purchased from Alfa Aesar Co. Inc (Mississauga, ON, Canada). The pea protein was kindly given by Daiya Foods (Vancouver, BC, Canada). Other chemicals and solvents were of analytical or chromatography grade.

Sample Preparation
A two-step process was employed to prepare the O/W nanoemulsions. The pre-emulsions were prepared by a 1200W Polytron PCU-2-110 homogenizer (Brinkmann Instruments, Inc., Westbury, NY, USA). The preparing process was described as follows: 5 mL HSO was mixed with 45 mL distilled water. Then the different concentrations of lecithin and pea protein (shown in Table 1) were added into this solution and homogenized for 10 min at a speed setting of half of the full speed (rated at 10,000 rpm). Within 15 min, these pre-emulsions were ultrasonicated using a 200W UP 200ST (Hielscher Ultrasonics, Teltow, Germany) for a particular time determined by the experimental design using 40% of full power in continuous mode. To control the temperature during the ultrasonication process, all samples were kept in an ice-bath. A ll the nanoemulsions were stored at 4 • C for further tests.

Droplet Size and Zeta Potential Determination
The hydrodynamic diameters (d H ) of the emulsion droplets were obtained with dynamic light scattering (DLS) technique using Zetasizer Nano-ZSP (Malvern Panalytical Ltd., Malvern, UK). The measurements were performed at a detector angle of 173 • and temperature of 25 • C. The average values from three measurements of particle size distribution were presented. Zeta potential was determined using Universal 'Dip' Cell Kit with Zetasizer Nano-ZSP (Malvern Panalytical Ltd., Malvern, UK) at 25 • C. A ll samples were diluted to the concentration of 1 mg/mL and all measurements were in triplicate, as well.

Spectroscopic investigations FTIR
Infrared spectra were obtained using PerkinElmer Spectrum™ 100 FT-IR spectrometer (PerkinElmer Inc., Waltham, MA, USA). The experiments were performed using an HATR sampling accessory dedicated for liquid samples. Spectra were recorded in the rage from 370-7800 cm −1 .

Rheological Tests
Rheological tests were carried out using Modular Compact Rheometer MCR302 (Anton Paar GmbH, Graz, Austria). Measurements were performed at 25 • C using measuring plate PP50 (D:50 mm) and inset I-PP50/SS (D: 50mm; stainless steel cat. no. 16222). A fter the plate reached the adjusted temperature, 2 mL of samples were pipetted on the plate surface. The predefined templates from RheoCompass software (Anton Paar GmbH, Graz, Austria) for low-viscous samples were used for the test. The results from three replications were evaluated with RheoCompass software (Anton Paar GmbH, Graz, Austria). The GraphPad Prism v. 7.04 was used to treat the data.

Colorimetry
For color evaluation, LabScan XE spectrophotometer (Hunter Associates Laboratory, Inc., Reston, VA, USA) equipped with EasyMatch QC software was applied. Before examinations, 5 mL of the sample was inserted into the transparent plastic small Petri dish-like plate. Whole measurements were repeated three times and average values ± SD were recorded.

Fatty Acid Analysis (GC Method)
Fatty acid profile of the emulsions was determined according to the standard AOAC procedure [34] with some modification using a Shimadzu GC-17A Gas Chromatograph system (Shimadzu, Scientific Instruments, Inc., Columbia, MD, USA) equipped with a flame ionization detector (FID) and Shimadzu Class-VP Software. A fused silica capillary column (Omegawax™ 320, 30 m × 0.32 mm ID × 0.25 µm film thickness) was used for the separation. The initial temperature of the column was set at 165 • C for 10 min followed by increasing to 200 • C with a rate of 1.5 • C/min. Individual fatty acids were determined by comparison to the retention times of a mixture of fatty acid methyl ester (FAMEs) standard.

Entrapment Efficiency (%)
Preliminary experiments were performed in order to find a reliable marker for determination of the entrapment or incorporation efficiency of HSO within the emulsion structure. It was found that linoleic acid (C18:2; n-6) as the most abundant fatty acid in HSO [17] could be considered as a marker for tracking the entrapment efficiency. For this purpose, the prepared samples were centrifuged at 12,000 rpm at 15 • C using Legend X1R centrifuge (Thermo Fisher Scientific, Inc., Waltham, MA, USA) for 15 min to sediment the nanoemulsion particles from the aqueous phase. When the emulsion suspension was centrifuged, the un-bond HSO was floated. Following centrifugation, the supernatant was collected and the fatty acid percentages including linoleic acid were determined according to the method described in 3.3.5. The entrapment efficiency (%) of linoleic acid (LA) concentration was determined by comparing the amount of LA incorporated into the emulsion structure with the initial amounts of LA in hempseed oil wherein the amounts of LA are based on the respective chromatogram peak area.

Stability Tests
The stabilities of emulsions with or without proteins (EMO and EMOP) were tested using two methods: simple and accelerated methods under gravitational and centrifugal forces, respectively. Both methods were followed by visual observation based on sedimentation, creaming or flocculation. The tested emulsions for the simple method were placed in 50-mL sealed transparent centrifuge tubes and kept for 14 days at room temperature (22 • C). The tested emulsions for the accelerated method were placed in 15-mL sealed transparent centrifuge tubes and diluted 10 times for visual proposes. Then, all the samples were centrifuged for 10 min under 5000, 7500, 10,000, and 12,500 rpm. A ll the accelerated tests were conducted with Sorvall Legend X1R Centrifuge Series (Thermo Fisher Scientific, Inc., Waltham, MA, USA).

Thermal Resistance Studies
Heat resistance of emulsions with or without proteins (EMO and EMOP) were tested under three temperatures including room temperature, 37 • C and 50 • C. A n amount of 5 mL of these emulsions were incubated at Sanyo MIR-553 (Sanyo, Japan) incubator for 1 week under the mentioned temperatures. Then, all the samples were tested for their particle size (nm), zeta potential (mV) and entrapment efficiency (%) as described before. A ll the tests were performed in triplicate.

Microscopic and Macroscopic Investigations
The images of the emulsions were taken from Axio Observer.z1 (Zeiss, Oberkochen, Germany). The imaging was performed in fluorescence and light mode with different magnifications (10×/0.25 air, 40×/1.3 oil, 100×/1.45 oil). The images were processed using ZEN imaging software (Zeiss, Oberkochen, Germany). The stability and homogeneity of the samples were evaluated by eye observation during storage. To present the inhomogeneity, phase separation and easily distinguish color variation between samples, simple pictures were taken using Samsung J7's built-in camera.

Experimental Design and Statistical Analysis
In the present study, a Box-Behnken design was employed to build polynomial models for process optimization using Minitab software version 18.0. The design consisted of three replications at center points to evaluate the pure error. To assess the main, interaction and quadratic effects of the process parameters including protein concentration (0.2-1.4%), surfactant concentration (0-6%) and sonication time (0-20) on response variables, a non-linear model was generated as follows: y = β 0 + β 1 x 1 + β 2 x 2 + β 3 x 3 + β 4 x 1 x 2 + β 5 x 2 x 3 + β 6 x 1 x 3 + β 7 x 12 + β 8 x 22 + β 9 x 32 (5) in which y is the response variables associated with the combination of each factor-level; β 0 -β 9 are the coefficients of the regression for respective variables computed from the experimental values of y; and x 1 , x 2 , and x 3 represent the independent variables. The experiments were randomized to decrease the unexplained effects of responses variability due to unimportant factors. The numerical and graphical optimization were performed for finding the optimal condition of the nanoemulsion process for desirable values of response variables including particle size (nm), polydispersity index (pdi) and Zeta potential. For a better conception of the significant effects (p < 0.05), graphical optimization was performed by 3D plotting of the reduced response models. The mathematical model was verified by carrying out the experiment in the given optimal conditions and data analysis using T-test.
To define the significance of the difference between the experimental runs, data were subjected to one way of ANOVA. A ll of the test analyses were performed in triplicate and reported as the mean ± SD. The significant difference (p < 0.05) among means was determined by Tukey's test The experiment design matrix generated by the software is shown in Table 1.

Conclusions
Performed examinations on hempseed oil nanoemulsion proved the impact of pea protein on emulsion stability improvement. A pplying a Box-Behnken design of the experiment for process conditions, followed by process optimization for preparing the nanoemulsion with pea protein resulted in obtaining more than 98.00% encapsulation efficiency with the smallest average particle size and pdi values (209 nm and 0.239) and desired zeta potential (−27.3 mV). The presented results showed that the prepared emulsion had a water-like rheological behavior. Both EMOP and EMO nanoemulsions had similar but slightly higher viscosity than water. This indicated that the water-like behavior of hempseed oil emulsions was not influenced by adding pea protein as a stabilizer. Based on the results of the stability tests, the EMOP showed better results for particle size, zeta potential and entrapment efficiency after 1 week of storage and centrifugal accelerating tests. This was caused by the fact that the protein can form viscoelastic films on the outer part of the droplets, protecting the oil phase in the particle and maximizing the stability of the emulsion system. In addition, obtained preliminary tests showed that the storage temperature of 37 • C might result in microbial growth. Further detailed investigations are needed for the possible use of hempseed oil nanoemulsion in food, cosmetic and biomedical applications.