Hemp Seed Oil Extraction and Stable Emulsion Formulation with Hemp Protein Isolates

Abstract

Industrial hemp (Cannabis sativa L.) is traditionally processed for its high fibre content in the textile industry, but in recent years, it has come to constitute a new raw material in the food industry. Hemp seeds, but also the seed meal, are rich in protein (25%) and oil (30%), the latter consisting of 80–90% of unsaturated fatty acids; they represent a rich and balanced nutritional source to replace the classic animal sources, and they are used in the food industry to supply new food trends. In this work, the hemp oil extraction process was studied, taking advantage of the supercritical CO₂ and ultrasounds, and comparing it with the exhaustive Soxhlet technique. The residual cake from extraction is a protein-rich waste that can be used for food formulations. From this perspective, the hemp oil extracted was used to formulate emulsions with the consistency of vegetable drinks, enriched with standard hemp protein isolate and stabilized with the addition of 0.05% w/w of thickening polysaccharides (Gellan gum). The formulated emulsion is stable, and this can encourage the process improvement and the use of the waste from hemp seed extraction for the valorisation of by-products and waste to obtain complete food products with high nutritional value.

1. Introduction

The hemp plant (Cannabis sativa L.) is widely cultivated and used in industrial production because of its multi-functionality. It is possible to obtain isolated proteins, fibres for paper and textile, oil from seeds and flours either for food or for cosmetics and pharmaceutical industries [1,2,3].

Particularly, the oil from hemp seeds has a high nutritional profile because of polyunsaturated essential fatty acids (EFAs) contained in this oil, which are linoleic acid (C18:2ω6) and α-linolenic acid (C18:2ω3) contained in the hemp seed oil at a ratio of 3:1 [4]. In addition, hemp seed oil has a rare acid, γ-linolenic acid, that is advantageous for hormonal balance. Linolenic acid also a positive impact on rheumatoid arthritis, atopic dermatitis, and allergies and has anti-inflammatory, antihypertension, anti-vasoconstrictive, and anti-cancer effects [5]. Additionally, the oil is suitable for both human nutrition and the preparation of several oils and body creams, thanks to its good absorption through the skin. Tocopherols contained in hemp oil are also used to reduce the risk of cardiovascular diseases, cancers and skin-ageing alterations within the blood vessels [6,7], and the well-balanced quantity of linolenic acid and α-linolenic acid gives hemp seed oil unique healthy properties [5].

Hemp oil is generally produced by solvent extraction or by cold pressing. The first type of extraction uses solvents, and for this reason is dangerous for the environment and can contain toxic residue. Thus, in the last few years, cold pressing has come to be preferred despite the lower yield of extraction [8]. The cold-pressing extraction can recover up to 65% of the oil in the seeds, and 35% of the available oil remains in the seed cake, but a high chlorophyll pigment content is also present in the oil [9]. The pigments modify the oil colour from dark to light green and promote the oxidation of lipids, which leads to the degradation of hemp oil quality, requiring its storage in dark or matt packaging [9,10]. Therefore, other extraction techniques are investigated by researchers such as supercritical extraction with CO₂ (SFE-CO₂), which can give an oil of very high quality also in the presence of an ultrasound pre-treatment [8,10]. SFE-CO₂ is becoming an important technology to obtain oil with a high purification degree for nutritional purposes or pharmaceutical applications. The solvent used is nontoxic and easy to separate from the extract, and moreover, it is tunable; the use of ultrasounds, coupled with this green technique, can improve the yield of oil obtained because they can disrupt the oil bodies [8]. The ultrasonic pretreatment was also evaluated to extract oil with SFE-CO₂ from hemp seeds and other types of seeds [8,11,12,13]. The use of ultrasonic conditions reduces the extraction time [12,13,14], but when ultrasounds are used, they generate heat, the amount of which is directly proportional to the ultrasonic power used.

It should be pointed out that In recent years, food industries have also focused their attention on the use of hemp proteins because hemp proteins consist of albumin and edestin, and these are an excellent source of digestible amino acids [15,16,17,18]. The amount of protein inside the hemp seeds is around 25%, and it is possible to extract them as hemp protein isolates after extracting the oil from the seeds and after performing particular steps to isolate them [15,17].

This type of protein can be used for different foods, such as baked goods [19], food foams, or emulsions [15,17,20,21] because they are also easily digested [20,22]. Unfortunately, hemp proteins have low solubility at neutral pH, making their use limited in systems such as foams or emulsions [20].

Tang and co-workers [21] studied the stability of hemp protein emulsions to find a trend with pH and compare their functional properties with those of soy protein isolate. Other literature works have looked for a correlation between the ways of isolation of hemp proteins, their solubility and the stability of the emulsions [17,21,23]. It is also possible to stabilize emulsions prepared with hemp protein isolate using nanoparticles [20] in a complex of hemp globulin (HG) with sodium caseinate (SC) via a pH-cycling method, even if the obtained stable emulsion has a solid-like consistency, or by adding pectins [24]. With the aim of obtaining hemp seed oil emulsion from hemp seeds, two nonionic surfactants, tween and span, were also used in the literature [25]. The use of polysaccharides in addition to proteins is suitable for stabilizing emulsions because of the stabilizing effect of the interaction between the two biopolymers but also because of the thickening effect induced by polysaccharides [22,26].

Due to the interesting hemp seed oil properties and high-quality profile of the hemp proteins, which are also characterized by interesting interfacial activity [15], this work aims at preparing stable oil-in-water emulsions based on hemp oil and commercial hemp protein isolates, trying also to investigate the possibility of maximizing the yield of oil extraction from seeds and enhancing the waste. Oil extraction by SFE-CO₂ was also investigated, combining the extraction with ultrasound pretreatment and comparing the lipidic profile with the oil extracted by cold pressing and by classical Soxhlet extraction.

2. Material and Methods

2.1. Raw Material and Samples Pre-Treatment

Hemp seed (Cannabis sativa L.) samples with 11.50 ± 0.28% w/w moisture and water activity of 0.385 ± 0.004 were collected from experimental cultivation of hemp Futura 75 carried out at Spezzano Sila (Cosenza, Italy).

The hemp seed oil was obtained from grounded seeds. A quantity of 30 g of hemp seeds was ground for 5 min in a stainless-steel blender to obtain hemp seed particles. After this first step, the hemp seed particles were put in a beaker without any solvent according to the literature [8] and pretreated before extraction with an ultrasound bath Bandelin Sonorex RK 102 H (BANDELIN electronic GmbH & Co., Berlin, Germany) or directly processed by SFE-CO₂ or Soxhlet extractor. The ultrasound bath works at a fixed frequency of 35 kHz and has a power of 120/480 W. The device is equipped with a thermostat suitable to change the bath temperature (range: 20–80 °C).

Different sonication times (10, 20, and 40 min) were used to investigate the effect of the ultrasounds pretreatment on the hemp oil obtained by SFE-CO₂ to increase the yield of extraction without altering the lipid profile. The temperature during ultrasound pretreatment increases, and for this reason during sonication, the temperature was maintained at the desired level, between 20 and 30 °C, by controlling the ultrasound bath. The sample temperature was also monitored by a thermocouple during the sonication time.

Table 1. Hempseed samples processed by Soxhlet and SFE-CO₂ at different conditions.

Sample ID	Pre-Treatment Type	Type of Extraction
HS_SFE	-	SFE-CO2
HS_SFE _10U	10 min ultrasounds	SFE-CO2
HS_SFE _20U	20 min ultrasounds	SFE-CO2
HS _SFE_40U	40 min ultrasounds	SFE-CO2
HS_S	-	Soxhlet extraction

shows the samples investigated and the sample identifiers.

2.2. Soxhlet Extraction

30 ± 1 g of hemp seeds and 250 mL of n-hexane of analytical grade (Sigma-Aldrich Co., Milano, Italy) were used. Ground hemp seeds were placed into a thimble filter and located in the Soxhlet extractor. N-hexane was added, and the system was heated until boiling. Reflux was kept for 8 h. A mixture of n-hexane rich in hemp seed oil was collected [8].

To remove water, anhydrous sodium sulphate was added to the mixture of oil and n-hexane before evaporation [27]. After the water removal, the oil was isolated thanks to a Rotavapor at 50 °C (Heidolph G3, Hei-VAP Value) [8]. In the end, the yield was calculated after weighting. Extractions were done in triplicate and the maximum yield was computed according to the following equation:

$$Yiel{d}_{max}(\%)=\frac{mass of oil extracted \left[g\right]}{mass of hemp seed feed \left[g\right]}\times 100$$

(1)

Determination was done in triplicate.

2.3. Supercritical CO₂ Extraction (CO₂-SFE)

25 ± 2 g of grounded hemp seeds were put in the vessel extractor. Supercritical CO₂ extraction experiments were carried out in a laboratory-scale plant (Spe-ed SFE, Applied Separations, Allentown (PA), USA) using a temperature of 40 °C and a pressure of 300 bar according to Da Porto and coworkers [8]. The used carbon dioxide (purity >99.99%) is supplied by SIAD Spa (Bergamo, Italy).

The SFE procedure is composed of alternated static and dynamic extraction steps. After an initial static phase of 30 min, to promote intimate contact between the fluid and the matrix, the extraction was carried out for 4 h and 30 min. During the total extraction time, two steps were alternated: one of 15 min in dynamic and another one of 15 min in static conditions. Extracts obtained from hemp seeds were collected in a volumetric flask and weighed during extraction. The yield max was calculated according to Equation (1).

2.4. Determination of Fatty Acids Compositions

2.4.1. Oil Sample Preparation

The oil samples to be analysed were diluted to a ratio of 1:10 with chloroform (CHCl₃) saturated with NaCl. Solvents (CHCl₃ HPLC grade) and NaCl were purchased from Sigma Aldrich Fluka (Milano, Italy).

2.4.2. Mass Spectrometry

Each sample was directly spotted three times on 384-well insert Opt-TOFTM stainless steel MALDI plates (AB SCIEX. Darmstadt. Germany). Mass spectrometric analyses were performed using a 5800 MALDI-TOF-TOF Analyzer (AB SCIEX. Darmstadt. Germany) equipped with an Nd: YLF Laser with λ = 345-nm wavelength of <500 ps pulse length and p to 1000 Hz repetition rate, in reflectron positive mode with a mass accuracy of 5 ppm. Mass spectra were acquired automatically in the positive reflector mode between 200 and 2000 with fixed laser intensity. Spectra with signal-to-noise below 200 were discarded automatically by the instrument. The operation parameter was optimized for the mass region of interest. Laser intensity was adjusted manually to avoid detector saturation. At least 4000 laser shots are typically accumulated with a laser pulse rate of 400 Hz in the MS mode. After acquisition, spectra were handled using Data Explorer version 4.11 (AB Sciex). All data presented in this work are averages of three replicates. LD MS spectra were evaluated against a computationally generated database of lipids [28] by entering a list of precursor ion m/z values restricted to 10 ppm of mass tolerance, and to commonly occurring acyl chains in edible oil, Chain positions and double bond regiochemistry and geometry was not specified.

2.5. Emulsion

2.5.1. Emulsions Preparation

In this study, gellan gum (G) (Sigma Aldrich, Hamburg, Germania) was used as a stabilizing agent with hemp protein isolate (H). The H used is a commercial product purchased from Bulk Powders^® (Brunel Way, Colchester, UK) and its composition, as described in the datasheet supplied by the manufacturer, is 47% w/w of protein, 15.3% w/w of carbohydrate, 12.7% w/w of fat and 20% w/w of fibre.

All the emulsions were prepared with distilled MilliQ ultrapure water (W) (Millipore, Burlington, VT, USA), hemp seed oil and H. The hemp oil used is obtained by SFE-CO₂ without ultrasound pretreatment because of the light green colour of the other extracts obtained with sonication. The H was mixed into a beaker at 35 °C with water for 2 h with a magnetic stirrer (AREX Heating Magnetic Stirrer, Velp Scientifica, Usmate (MB), Italia), and then the solution was centrifuged at 2900 rpm (centrifuge 5810, Eppendorf, Hamburg Germany) for 30 min to remove the fiber part. After centrifugation, NaCl and CaCl₂ were added and the solution was mixed for 30 min by a magnetic stirrer (AREX Heating Magnetic Stirrer, Velp Scientifica, Italia) [22]. All solutions were prepared by considering the real protein content in raw materials, then calculating the protein content purity, and then using, for each sample, the protein isolate amount suitable for giving the correct protein content in the investigated samples reported in the label sample in

Table 2. Emulsions formulation.

Emulsion ID	H (% w/w)	Hemp Seed Oil (% w/w)	W (% w/w)	G (% w/w)	NaCl (% w/w)	CaCl2 (% w/w)
H_1.5	3.20	5	91.69	-	0.1	0.01
H_2	4.26	5	90.63	-	0.1	0.01
H_3	6.38	5	88.51	-	0.1	0.01
H_2_G	4.26	5	90.58	0.05	0.1	0.01

(i.e., H_2 means that 2% w/w of hemp protein is present in the formulation).

When emulsions with stabilizing agents were prepared, the right quantity of stabilizing agent was added to bidistilled MilliQ ultrapure water (Millipore, USA), and the solution was kept at 90 °C for 10 min [15] to ensure the solubilization by magnetic stirrer with temperature control (AREX Heating Magnetic Stirrer, Velp Scientifica, Italia); then the H was added.

After obtaining the aqueous solution, the production of emulsions followed a typical direct homogenization method. Two steps of homogenization were carried out [29]: In the first step, the two phases were mixed at high shear rates with a rotor-stator device (UT T50, IKA, Königswinter, Germany, tool S50N G45F) at 4000 rpm for 120 s, whereas the second step was carried out with gentle mixing with a magnetic stirrer (AREX Heating Magnetic Stirrer, Velp Scientifica, Italia) at room temperature to complete the emulsion stabilization [30]. The pH of samples was not varied, and nor were reverse methods.

Following [20], the dispersion of one emulsion droplet into the water and hemp oil was analysed, and it was found that the emulsions are of O/W type. After emulsion preparation, the samples were transferred into a beaker and covered with parafilm to prevent drying for 3 h, and after this time, the characterization was performed [31].

2.5.2. Emulsion Morphology and Determination of Droplet Size

The morphology of the emulsion droplets was observed with a contrast phase microscope (MX5000, Meiji, Chikumazawa, Japan) equipped with a phase contrast 20× lens. The samples were prepared according to [31,32].

Particle size distribution (PSD) was obtained with the statistical analysis of data measured thanks to the contrast phase optical microscope (MX5300H, MEIJI, Japan, magnification 20× for all samples). The statistical analysis of the DSD (Droplets size distribution) was obtained with the image processing software Dhs Particle Analysis (DHS image database, Greifenstein, Germany), which by the selective colouring of the particles allows tracing their average diameter (d_s) and the standard deviation of the log-normal distribution (σ_s) [31,32] through an analysis of the number of pixels necessary for the selective colouring of the single particle.

The drop size distribution can be well described by a lognormal model:

$$f\left(d\right)=\frac{1}{d·{\sigma }_{ln}·\sqrt{2\pi }}exp\left[\frac{-{\left(\ln \left(d\right)-d_{ln}\right)}^2}{2{\sigma }_{ln}^2}\right]$$

(2)

where $d_{ln}$ and ${\sigma }_{ln}$ are, respectively, the mean and the standard deviation of the normal model [31,32], which allow the evaluation of the mean diameter $d_s$ and the variance ${\sigma }_s^2$:

$$d_s=e^{d_{ln}+{\sigma }_{ln}^2/2}$$

(3)

$${\sigma }_s^2=e^{2d_{ln}+{\sigma }_{ln}^2\left(e^{{\sigma }_{ln}^2}-1\right)}$$

(4)

${\sigma }_s$, is the standard deviation and it is commonly considered an index of polydispersity [31,32].

2.5.3. ζ-Potential Measurements

The ζ-potential measurements were performed with a zeta potential analyser (Zetasizer Nano ZS, Malvern Instrument, Malvern, UK) using electrophoretic dynamic light scattering. A static electrical field was applied by electrodes to the emulsion sample in a cell in order to move charged oil drops towards the oppositely charged electrode. The ζ-potential was calculated from the Smoluchowski equation. The absolute zeta potential gives information about the stability of the emulsions; in particular, for absolute zeta potential greater than 30 mV, the emulsion can be considered stable [33]. For the testing, the emulsions were diluted to a droplet concentration of about 0.001 wt% using buffer solutions of the appropriate pH according to Noshad and coworkers [34].

2.5.4. Rheological Characterization

The stable emulsion was characterized by rheological investigation with a rotational rheometer ARES (Rheometric Scientific, New Castle, DE 19720, USA) adopting parallel plates geometry (diameter = 50 mm) thermostated with a Peltier system (±0.1 °C). Flow curves were performed at 25 °C in triplicate, and the data are shown as mean and standard deviation.

3. Results and Discussion

3.1. Comparison of SFE-CO₂ Extraction and Solvent Extraction

Before analysing the ultrasound effect on the yield of extraction, the yield max and the composition of hemp seed oil obtained with supercritical CO₂ are compared. It was found that 35.33 ± 2.00 g hemp oil/100 g of hemp seed and 14.77 ± 0.82 g hemp oil/100 g of hemp seed feed of oil was extracted by soxhlet and supercritical CO₂ method, respectively.

Figure 1 shows the yield obtained for the samples processed at 300 bar and 40 °C to evaluate the effect of pretreatment with ultrasounds.

From the obtained trend, it is possible to point out that the penetration of CO₂ inside the tissue is favoured by ultrasound pretreatment because it facilitates the oil extraction, as evidenced by the higher slope of the three pretreated samples compared with the untreated one. In particular, the slope increases increase the sonication time even if the amount of the extracted hemp oil does not increase in a relevant way.

Specifically speaking, a longer ultrasound pretreatment from 10 to 40 min causes extraction yields of 19.85 ± 0.02, 20.04 ± 0.03 and 20.92 ± 0.02 (% w/w) compared with the yield obtained without pretreatment, but it is possible to observe a slight greenish colour due to the probable presence of chlorophyll [10].

The increase in the penetration rate of supercritical CO₂ solvent into the tissue is evidenced by the differences in the slopes of the curve with and without ultrasound pretreatment. Moreover, it is possible to observe that the penetration velocity of the supercritical fluid increases with increasing the ultrasound time from 10 to 20 min because of the higher yield at the same extraction time. According to Ivanovs and Blumberga [35], ultrasound pretreatment decreases the extraction time and solvent consumption, also giving a higher penetration of chosen solvent into the cellular material and an enhanced release of cell content, even if it is possible to highlight that the quantity of oil extracted as the pretreatment time increases remains unchanged contrary to Da Porto and coworkers [8].

The quality of the extracted oil is fundamentally related to the extraction technique employed. Hemp seed oil is generally extracted by cold pressing, avoiding the use of organic solvents if intended for human consumption. Solvent extraction takes a long time, and the presence of toxic residues in the final product is a disadvantage, but it is in general used to know the maximum yield of extraction [10]. A green alternative with relatively low extraction times could be optimal because the oil colour is yellow without sonication pretreatment, contrary to cold pressing, and light green using the sonication. Moreover, SFE is considered a clean technology (green process/eco-friendly) that allows the use of low-environmental-impact solvents such as carbon dioxide that allow you to work at low temperatures and extraction pressures [10].

3.2. LD TOF MS Analysis of Hemp Seed Oil

Considering that the high nutritional quality of hempseed oil is essentially determined by its fatty acid composition, the application of MS-based chemical component profiling offers significant opportunities to obtain detailed information that can be directly correlated to oil quality [36].

A typical distribution of the major fatty acids in industrial hemp seed oil is as follows [37]: palmitic acid (C16:0; 6.66–6.98%), stearic acid (C18:0; 2.08–2.82%), oleic acid (C18:l; 9.38–13.00%), linoleic acid (C18:2; 55.56–56.58 %), α-linolenic acid (C18:3; 14.69–17.27%) γ-linolenic acid (C18:3; 2.56–4.49%). The fatty acid composition of the hemp seed oils, determined in accordance with [38,39] obtained by Soxhlet and supercritical CO₂ extraction with and without ultrasound pretreatment is reported in

Table 3. Fatty acid composition of hempseed oil, cultivar Futura 75, extracted by supercritical CO₂ from untreated and ultrasound pre-treated hemp seeds and Soxhlet. Each data represents the mean value of three replicates ± standard deviation.

	HS_SFE	HS_SFE_10U	HS_SFE_20U	HS_SFE_40U	HS_S
Monounsaturated of	11.74 ± 0.04	11.8 ± 0.2	11.71 ± 0.07	11.54 ± 0.04	11.5 ± 0.2
Oleic acid (C18:1)	11.74 ± 0.04	11.8 ± 0.2	11.71 ± 0.07	11.54 ± 0.04	11.5 ± 0.2
PUFA sum of	81.71	80.67	81.46	81.29	81.77
Linoleic acid (C18:2)	59.88 ± 0.09	59.37 ± 0.05	59.6 ± 0.3	59.79 ± 0.08	59.86 ± 0.08
alfa-linolenic gamma-linolenic (C18:3)	21.8 ± 0.2	21.3 ± 0.2	21.8 ± 0.2	21.51 ± 0.06	21.91 ± 0.01
ω-6/ω-3 ratio	2.74	2.79	2.73	2.78	2.73
Saturated	7.18 ± 0.03	7.5 ± 0.1	7.18 ± 0.04	7.28 ± 0.07	7.28 ± 0.05
PUFA sum/saturated	11.39	10.7	11.35	11.16	11.24

.

No significant differences were found when the hemp seed oil was extracted by supercritical fluid extraction using CO₂ as a fluid or by Soxhlet. The main components were oleic acid (O) (18:1), linolenic acid (L) (18:2) and linolenic (Ln) (18:3) (Table 3). The determined polyunsaturated fatty acids (PUFA) were 81% of total fatty acids, while the monounsaturated and saturated fatty acids amounted to 11.4–11.7% and 7–7.5%, respectively. The ratio of polyunsaturated to saturated components was 11–11.5%. The determined omega 6/omega 3 ratio for all samples (2.7–2.9%) is slightly different from the commonly accepted value (3–3.3%) because this approach does not allow for distinguishing α-linolenic acid from γ-linolenic acid. This composition agrees with results reported in Canada [40,41] and for hemp cultivated in Europe [10].

3.3. Droplet Size Distribution of Emulsions

The emulsifying properties of the vegetable proteins were investigated, varying the protein concentration in the emulsions. Drop size distributions (DSD) were obtained by image analysis of emulsion microphotographs; the images were captured for the emulsions after 3 h storage time and are reported in Figure 2.

As it is possible to observe in the microphotographs, the emulsions show a broad distribution of oil droplets. The average droplet diameter (d_s) and standard deviation (σ_s) of emulsions were evaluated from the lognormal model to investigate the effects of the increased hemp protein isolate fraction in the emulsions as well as the gellan gum effect, which allows for stabilizing the sample with a final quantity of 2% w/w of hemp protein.

Experimental data show the mean droplet diameter and the standard deviation for all the tested emulsions. It is possible to point out that both mean diameter and standard deviation increase by increasing the quantity of the protein from 1.5 to higher values in the emulsions. It is possible to observe that emulsions with only protein isolate have an average droplet diameter that goes from 2.1 μm at 1.5% w/w up to 2.8 μm at the highest protein concentration. The value found at the lowest protein level is compatible with the mean droplet diameter found by Dapčević-Hadnađev and coworkers [17], who prepared emulsions with protein isolates from hemp seed meal obtained using two different isolation techniques [17].

The emulsions with only protein show instability phenomena after about 6 h from the preparation. For this reason, it was necessary to add a stabilizing agent. The gellan gum in the quantity of 0.05% w/w was found to be suitable for obtaining a stable emulsion, but as seen in Figure 3, the mean diameter and the standard deviation increase after the addition of the gellan gum as compared with the same emulsion without it. The increase in d_s after the addition of a polysaccharide had already been observed in the literature [31] and was attributed to the thickening effect of the polysaccharide and to the variations in the adsorption behaviour of proteins at the interface O/W. These phenomena can make it difficult to form small droplets of oil. The polydispersity has also a similar trend after the gellan gum addition.

3.4. Rheology and ζ-Potential Emulsions Measurements

The emulsions prepared using only vegetable proteins as stabilizing agents were all unstable, indicating that the interfacial properties controlled by vegetable proteins are not sufficient to stabilize the whole system. The instability phenomena are not only due to the interface but also to the difference in density between the mixed phases (oily and aqueous), which leads to obvious phenomena of phase separation, such as the creaming of large drops, or sedimentation. On the contrary, the emulsion with gellan gum was also visually monitored to verify the formation of any layers, and it was observed that no stratification had occurred within one week. Therefore, the visual observation evidence that the polysaccharide, in the quantity of 0.05% w/w, can avoid destabilization phenomena. In light of these results, the viscosity of the sample H_2_G was measured because this is an important measure for understanding the structural organization and the network formation. The viscosity is also an important commercial parameter that influences the texture and the sensory properties [17]. A previous work of literature shows that emulsions prepared with only hemp protein isolate have a Newtonian behaviour [17] while, as Figure 4 shows, the emulsion prepared with gellan shows a non-Newtonian behaviour. The shear thinning behaviour can be due to the stabilizing agent added to control the instability phenomena. In fact, as reported by Herrera [22], emulsions obtained with gellan gum show high viscosity, probably caused by the presence of Ca²⁺ that causes carboxylate–cation2⁺–carboxylate interactions increasing the gel structure [22]. The same result was found by Feng and coworkers [24] using pectin as a stabilizing agent. Pectin as gellan gum probably forms a three-dimensional network, and this network gives a high-viscosity and shear-thinning behaviour to the emulsion [24].

The result found is also confirmed by the charge analysis of the emulsions carried out at the different protein quantities as shown in Figure 5. Before measuring the zeta-potential, pH was measured, obtaining a pH of 6.8 ± 0.1 for all the emulsions.

As is possible to observe, all the samples have a negative charge in accordance with the H isoelectric point [15]. For the studied emulsions, it is worth noting that the ζ-potential seems to slightly decrease with increasing protein quantity, but only after the addition of gellan gum, the ζ-potential absolute value at 2% w/w of H increase from −33 ± 2 to –26.5 ± 2.6 mV. As is well known, absolute zeta potential greater than 30 mV indicates a stable system. The electrostatic repulsion between the polysaccharide and protein particles, coupled with the tridimensional network formation, enhances the stability and the result obtained thanks to the polysaccharide addition, which was also obtained for oil-in-water hemp emulsions with another polysaccharide, pectin, as a stabilizing agent by Feng et al. [24].

This formulation can give the possibility of creating food-grade emulsions, capable of valorising both the oil and the hemp protein isolate. Interesting applications could be in creating high nutritional food thanks to the good amino acid profile of the isolate proteins [17,25,42]; even if the sample H_2_G has a non-Newtonian behaviour with high viscosity, which does not yet make it in line with the rheological characteristics of other systems based on vegetable proteins as vegetable drinks, it is possible in any case to use the formulation for other food applications (i.e., desserts or soft foods).

4. Conclusions

In this work, the possibility of extracting oil from hemp seeds using supercritical CO₂ was investigated. It is noted how the lipid profile remains optimal even in the presence of pretreatment with ultrasounds, which increases the extraction yield even if a slight green colouring is observable. The oil extracted by the supercritical process was then used to obtain oil-in-water emulsions, by adding isolated hemp proteins. A commercial isolate was used to give a standard residue for the formulation of the emulsions. The emulsions obtained were characterized, and the 2% w/w protein emulsions were stabilized by adding 0.05% w/w of gellan gum. Although the emulsion obtained was stable, the sample showed a relatively high viscosity, which contributed to the stability of the system but which requires optimizing for eventual marketing as a beverage. The product can instead be used as a base for obtaining slightly structured products for people with special dietary needs (e.g., dysphagia).