Effect of Oil–water Surface Area on the Aroma Release Behavior of Mono-dispersed Oil-in-water Emulsions

Abstract

The size of emulsion droplets can influence the taste and texture of food. Therefore, we investigated the effect of droplet diameter of oil-in-water (O/W) emulsions on the aroma release rate from emulsions. O/W emulsions were prepared to droplet sizes of 0.4, 2.0 and 20 µm. The release rates for seven aroma compounds, viz., limonene, ethyl hexanoate, 2-methylpyrazine, benzaldehyde, ethyl benzoate, α-terpineol, and benzyl alcohol, from the O/W emulsions were measured under non-equilibrium. The aroma release rate of 0.4-µm droplets was higher than that of 2.0- and 20-µm emulsion droplets irrespective of their partition coefficients. Aroma compounds in small oil droplets can rapidly diffuse in the water and air phases because a small droplet has a large oil–water surface area. Overall, our findings suggest that small droplet emulsions can be used to release more aroma compounds than large droplet emulsions without changing the emulsion components.

Introduction

Emulsification is widely used to produce processed foods such as mayonnaise, whipped cream, and salad dressing. The palatability of processed foods is important for consumer preferences. Palatability of foods is mainly determined by their aroma, taste, and texture. Therefore, food aroma perception is commonly associated with consumer preference. Aroma compounds are released from foodstuffs into the vapor phase and perceived by chemoreceptors in the olfactory epithelium. Several aroma compounds are lipophilic and retained by lipids in foodstuffs because the lipids present in foodstuffs lower the vapor pressure of certain lipophilic aroma compounds (Seurve, Phillippe, et al., 2007). In general, lipids are more effective in retaining lipophilic aroma compounds in foodstuffs than polysaccharides and proteins (Roberts, Pollein, et al., 2003; Philippe, Seuvre, et al., 2003). In our previous study, we reported that the aroma release behavior of oil-in-water (O/W) emulsions can be predicted using octanol-air partition coefficients of aroma compounds irrespective of the oil to volume ratio in the O/W emulsions (Tamaru, Igura, and Shimoda, 2018; Tamaru, et al., 2019). On the contrary, it is known that the texture of emulsions in the mouth is also influenced by the droplet diameter size and distribution (Golchoobi, Alimi, et al., 2016; Borreani, Hernando, et al., 2017). Moreover, the stability of an emulsion differs with the droplet diameter size and distribution in the emulsion (Davidov-Pardo, and McClements, 2015). In general, smaller diameter droplets in emulsions lead to higher stability and usefulness of the emulsion (Guttoff, Saberi, and McClements, 2015). Above all, the size of dispersed lipid droplets must be controlled properly to develop processed foods in O/W emulsions, with desirable flavor perception for consumers. Therefore, it is important that the release behavior of aroma from emulsions of different droplet sizes be studied.

In O/W emulsions, aroma compounds are transported from oil droplets to the water phase. The aroma compounds are then released from the water phase to the vapor phase. Therefore, the aroma release rate from O/W emulsions might involve the size of oil droplets or the mass transfer path length of aroma compounds and the oil–water surface area in O/W emulsions. Indeed, flavor release behaviors from salad dressing were influenced by the oil droplet size (Charles, Rosselin, et al., 2000). Moreover, more limonene, a lipophilic aroma compound, was released from about 1-µm droplet emulsions than about 12.5-µm droplet emulsions using model emulsions (Karaiskou, Blekas, and Paraskevopoulou, 2008). However, the detailed relationship between droplet surface area in emulsions and flavor release rate has not been shown. Therefore, the aim of this study was to examine the relationship between the droplet size of O/W emulsions and aroma release rate using three droplet size (mean diameter; 0.4, 2, 20 µm) mono-dispersed emulsions and seven aroma compounds. We expect that the results of this study can validate that emulsion droplet size affects food flavor, and thus, contribute to the development of new processed foods.

Materials and Methods

Materials    The aroma compounds examined in this study were limonene (≥97%), 2-methylpyrazine (≥99%), benzaldehyde (≥99%), and ethyl benzoate (≥99%) (Sigma Aldrich Corp., Gillingham, UK); ethyl hexanoate (≥98%) and benzyl alcohol (≥99%) (Nacalai Tesque, Kyoto, Japan); and α-terpineol (≥98%) (Wako Pure Chemical Industries, Ltd., Osaka, Japan). The partition coefficients (octanol-water partition coefficient; logPow, water-air partition coefficient; logPwa, and octanol-air partition coefficient; logPoa) of the seven flavors are shown in Table 1. These volatile compounds were chosen because of wide range of lipophilicity, or logPow. The compounds were dissolved in ethanol (≥99.5%; Nacalai Tesque) to obtain 1 v% stock solutions. Polyoxyethylene sorbitan monooleate (Tween80; Nacalai Tesque) was used as a surfactant and 1,3-butanediol (Sigma Aldrich Corp.) was used as a co-surfactant. Coconut oil (Nacalai Tesque) was used as the dispersed phase. Deionized water served as the continuous phase.

Table 1. Octanol-water, water-air, and octanol-air partition coefficients (logPow, logPwa, and logPoa, respectively) of the various flavor compounds employed herein

Flavor compound	logPow1	logPwa1	logPoa1
2-Methylpyrazine	0.21	4.04	4.25
Limonene	4.38	−0.11	4.27
Ethyl hexanoate	2.83	1.53	4.36
Benzaldehyde	1.48	2.96	4.44
Ethyl benzoate	2.64	2.52	5.16
Benzyl alcohol	1.1	4.86	5.96
α-Terpineol	3.28	3.3	6.58

¹  Calculated using the EPI suite.

Emulsion preparation    Small size (approximately 0.4-µm mean diameter) droplet emulsions were produced by D phase emulsification (Sagitani, Hattori, et al., 1983). Sagitani et al. (1983) used liquid paraffin as the dispersed phase to prepare a nanoemulsion with D phase emulsification. However, liquid paraffin is not used in the food industry, so we used food grade oil as the dispersed phase. Four grams of Tween80, 2 g of 1,3-butanediol, and 2 g of water were mixed. Subsequently, 20 g of liquid coconut oil (25 °C) was mixed with the solution by stirring until a gel structure was formed. Deionized water was used to dilute the gel to 100 g and stirred until the gel was solubilized in water at 300 rpm for 3 h. Medium and large size (approximately 2- and 20-µm mean diameter) droplet emulsions were produced by Shirasu porous glass (SPG) (SPG Technology Co. Ltd., Miyazaki, Japan) membrane emulsification (Shimoda, Miyamae, et al., 2011). Tween80 (0.5 g/L) and 1,3-butanediol (0.25 g/L) were dissolved in deionized water by stirring at 300 rpm for 3 h. Eighty grams of surfactant solutions and 20 g of liquid coconut oil (25 °C) were blended for 2 min at 5 000 rpm using an ultrafast digital homogenizer (AHG-160D; AS ONE, Osaka, Japan) to prepare coarse emulsions. The droplet size of these coarse emulsions was reduced by passing them through an SPG membrane of pore sizes 2.0 and 20.0 µm. With this approach, mono-dispersed emulsions were produced (Shimoda et al., 2011). All emulsions comprised 4 wt% Tween80, 2 wt% 1,3-butanediol, and 20 wt% coconut oil in the final concentration.

The particle size distribution measured using the laser diffraction particle size distribution measurement system (SALD-200V, Shimadzu Corp., Kyoto, Japan) is shown in Fig. 1 and Table 2. The span value was calculated as (90.0% D – 10.0% D) / 50.0% D (Tamaru et al., 2018). In each droplet diameter (0.4, 2, 20 µm), the span values were 0.488, 0.453, and 0.434, respectively. On the other hand, using canola oil, the span value was 0.836 in the 0.4-µm emulsion by D phase emulsification in a preliminary study. Small span values (<0.5) generally represent mono-dispersion. Therefore, we chose liquid coconut oil as the dispersed phase in this study. Appropriate aroma compounds were added to the sample emulsions to prepare emulsions with concentrations of 100 ppm and the prepared emulsions were placed in glass vials (30 mL) without headspace. The glass vials were sealed using a cap with Teflon liner and stored at 25 °C with stirring at 350 rpm for 12 h.

Fig. 1.

The number distribution of each emulsion (mean diameter: ◇, 0.4 µm; △, 2 µm; □, 20 µm).

Table 2. Droplet diameter distribution and the difference in droplet number in the O/W emulsions prepared using different methods

20 µm SPG membrane emulsification		2 µm SPG membrane emulsification		D phase emulsification
Droplet diameter (µm)	Difference value	Droplet diameter (µm)	Difference value	Droplet diameter (µm)	Difference value
16.7	0.137	1.42	0.0309	0.25	0.00229
19.3	0.318	1.64	0.146	0.289	0.0532
22.3	0.291	1.90	0.301	0.334	0.172
25.8	0.196	2.20	0.295	0.386	0.294
29.8	0.0544	2.54	0.206	0.446	0.273
34.5	0.00318	2.94	0.0213	0.516	0.178
				0.596	0.0274

Dynamic headspace analysis    The release rate of volatile aroma compounds was measured using a gas chromatograph (GC-14B; Shimadzu Corp.) equipped with a DB-WAX capillary column (30 m × 0.32 mm i.d., 0.5 µm film thickness; Agilent Technologies Inc., Santa Clara, CA, USA) and coupled with a flame ionization detector (FID). Each emulsion sample (25 g) was placed in a 200-mL flask and maintained at 25 °C. Nitrogen flow (10 mL/s) purged the headspace gas at 25 °C for 10, 60, 120, 240, and 300 s at 250 rpm. The released aroma compounds were collected in a glass tube (95 mm × 3 mm i.d.; Shimadzu Corp.) packed with Tenax TA (0.08 g; GL Science Inc., Tokyo, Japan) as an adsorbent. The glass tube was heated in the injector and the desorbed aroma compounds were injected into the gas chromatograph in split injection mode (1:14). The injector and FID temperatures were 230 °C and 240 °C, respectively. The helium carrier gas flow rate was 35 cm/s. The oven temperature was initially held at 50 °C for 2 min, and then increased at 10 °C/min to 230 °C. The chromatography data were used to calculate the primary approximation expression correlating to the peak area of the eluted aroma compounds and the nitrogen flow time using Microsoft Excel 2010. The aroma compound release rate from O/W emulsions was defined as the slope of the primary approximation expression. Each experiment was repeated at least three times using three different batches from each emulsion, and the mean and standard deviation were calculated.

Calculation of oil–water surface area in O/W emulsion    The oil–water surface area in O/W emulsions was calculated using the equations provided below. First, using the measured droplet size distribution and sphere volume formula, we calculated the number of droplets in emulsions as outlined in Eq. 1:   

where d_i is the measured droplet diameter (µm), x_i is the difference in the droplet diameters, N is the number of droplets, and W is the measured oil droplet weights (g).

Second, the total droplet surface area or oil-water surface area (R, m²) was calculated using Eq. 2 and the sphere surface area formula.

  

We studied the relationship between the measured aroma release rate and surface area in O/W emulsions using these values. We focused on the effectiveness of droplet size on the aroma release rates from O/W emulsions to develop new processed foods.

Results and Discussion

We measured the peak areas of the aroma compounds released from 20 wt% O/W emulsions by 0–300 s nitrogen flow into headspace. Figure 2 shows the release behavior of limonene from O/W emulsions with droplets of different diameters. The other aroma compounds also showed a similar linear behavior (Supplementary Materials). The peak area of each aroma compound from the O/W emulsions increased linearly with the time of nitrogen flow. We calculated the slopes of the primary approximation expression to determine the release rate of droplets of different sizes (Table 3). As shown in Table 3, the aroma compounds with a high octanol-air partition coefficient (logPoa) (>5.0), showed relatively low release rates. However, the release rate of all aroma compounds from 0.4-µm emulsions was approximately 1.4 times higher than that of compounds from 20-µm emulsions. The release rate of seven aroma compounds from 2-µm emulsions was similar to that of compounds from 20-µm emulsions. Decreasing the droplet size in O/W emulsions demonstrated a constant effect on the aroma release rate irrespective of the chemical structure of the aroma compounds.

Fig. 2.

Limonene release behavior from 20 wt% O/W emulsions with droplets of three different diameters (◇, 0.4 µm; △, 2 µm; □, 20 µm) by dynamic headspace analysis.

Table 3. Release rate and octanol-air partition coefficient (logPoa) of each aroma compound from each emulsion with droplets of different diameters

	Droplet diameter
Aroma compound	0.4 µm	2 µm	20 µm	Ratio (0.4 µm/20 µm)	logPoa
2-Methylpyrazine	16.5	10.6	11.8	1.39	4.25
Limonene	32.4	20.4	22.6	1.43	4.27
Ethyl hexanoate	27.7	17.9	18.5	1.49	4.36
Benzaldehyde	31.5	21.7	22.2	1.41	4.44
Ethyl benzoate	3.43	2.51	2.33	1.47	5.16
Benzyl alcohol	1.28	0.99	0.89	1.43	5.96
α-Terpineol	1.12	0.87	0.81	1.38	6.58

Using Eq. 1 and Eq. 2, the number of droplets in the emulsions and oil droplet surface area (m²/25 g emulsion) were calculated, respectively. Power approximation between droplet diameter and oil droplet surface area is shown in Fig. 3. In 25-g emulsions, decreasing droplet diameter from 20 to 2 µm led to a gradual increase in the number of droplets in the emulsions. Contrarily, decreasing the droplet diameter from 2 to 0.4 µm led to a sharp increase in the number of droplets. Therefore, regarding the droplet surface area in 25-g emulsions, similar behavior was shown from 20 to 2 µm and 2 to 0.4 µm. Moreover, the oil-water surface area in 25-g emulsions showed a high correlation with the aroma release rate. The correlation coefficient between the oil–water surface area and aroma release rate was 0.9402 for limonene, 0.9757 for benzaldehyde, 0.9737 for ethyl hexanoate, 0.9327 for 2-methylpyrazine, 0.9997 for ethyl benzoate, 0.9972 for benzyl alcohol, and 0.9999 for α-terpineol (Fig. 4). This suggests that a high oil droplet surface area in the emulsion leads to a high aroma release rate. Moreover, the flavor compounds with high logPoa values (ethyl benzoate, benzyl alcohol, and α-terpineol) had high correlation coefficients between the flavor release rates and total oil–water surface area in O/W emulsions. Flavor compounds with high logPoa values are more lipophilic, so they can retain more in oil droplets than other flavor compounds. Therefore, differences in surface area may easily affect flavor release rate. This relationship between aroma release rate and oil–water surface area in O/W emulsions was mentioned for the first time using the calculation in the present study. The aroma release behavior was affected by the characteristics of emulsions such as oil volume ratio, type of emulsifiers, and droplet diameter (Charles, Rosselin, et al., 2000; Meynier, Lecoq, et al., 2005). In a previous study, decreasing the oil volume significantly increased the aroma release rate from O/W emulsions (Tamaru et al., 2019). On the contrary, in this study, we revealed that oil droplet surface area in O/W emulsions was highly correlated with the aroma release rate. In other words, small droplet emulsions increased the aroma release rate. Karaiskou, Blekas, and Paraskevopoulou (2008) showed a similar tendency of limonene release behavior from O/W emulsions. Moreover, the present study results suggest that regardless of aroma properties, small droplet emulsions release more aroma compounds than large droplet emulsions. In conclusion, decreasing the droplet size in emulsions leads to high aroma release without changing the emulsion compositions. However, foodstuffs contain a number of aroma and nutrient compounds. Future study will focus on the effect of interaction between other compounds in emulsions on aroma release behavior.

Fig. 3.

Relationship between the mean droplet diameter and total droplet surface area in emulsion

Fig. 4.

Relationship between the aroma release rate and oil droplet surface area in 25 g of emulsions (a) for the most abundant flavors (◇, 2-methylpyrazine; □, limonene; △, ethyl hexanoate; ○, benzaldehyde) and (b) for the three least abundant flavors (◇, ethyl benzoate; □, benzyl alcohol; △, α-terpineol).

Conclusions

In this study, we examined the initial release rate of seven different aroma compounds from droplets of different sizes (0.4, 2, and 20 µm) in mono-dispersed O/W emulsions using dynamic headspace extraction under non-equilibrium conditions. The aroma release rate from 0.4-µm emulsions was approximately 1.4 times higher than that from 20-µm emulsions irrespective of the properties of the constituent aroma compounds. The relationship between oil–water surface area and droplet diameter was inversely proportional. A sharp increase in the oil–water surface area of emulsions was observed from 2 to 0.4 µm, with a corresponding increase in the aroma release rate. Moreover, a high correlation between the aroma release rate and oil droplet surface area was observed. Especially, the flavor compounds with high logPoa values showed higher correlation coefficients than the other flavor compounds. We propose that small droplet emulsions release more aroma compounds. Therefore, we can control the aroma release behavior of foods, especially more lipophilic flavor compounds, by decreasing the droplet diameter without changing the emulsion composition. The results of this study may prove useful in the production of processed foods with desirable flavors for consumers.

Acknowledgements    We would like to thank Editage (www. editage.jp) for English language editing.

References

•  Borreani,  J.,  Hernando,  I.,  Salvador,  A., and  Quiles,  A. (2017). New hydrocolloid-based emulsions for replacing fat in panna cottas: a structural and sensory study. J. Sci. Food Agric., 97, 4961-4968
•  Charles,  M.,  Rosselin,  V.,  Beck,  L.,  Sauvageot,  F., and  Guichard,  E. (2000). Flavor release from salad dressings: sensory and physicochemical approaches in relation with the structure. J. Agric. Food Chem., 48, 1810-1816.
•  Davidov-Pardo,  G. and  McClements,  D. J. (2015). Nutraceutical delivery systems: Resveratrol encapsulation in grape seed oil nanoemulsions formed by spontaneous emulsification. Food Chem., 167, 205-212
•  Golchoobi,  L.,  Alimi,  M.,  Shokoohi,  S., and  Yousefi,  H. (2016). Interaction between Nanofibrillated Cellulose with Guar Gum and Carboxy Methyl Cellulose in Low-Fat Mayonnaise. J. Texture Stud., 47, 403-412
•  Guttoff,  M.,  Saberi,  A. H., and  Mcclements,  D. J. (2015). Formation of vitamin D nanoemulsion-based delivery systems by spontaneous emulsification: Factors affecting particle size and stability. Food Chem., 171, 117-122
•  Karaiskou,  S.,  Blekas,  G., and  Paraskevopoulou,  A. (2008). Aroma release from gum Arabic or egg yolk/xanthan-stabilized oil-in-water emulsions. Food Res. Int., 41, 637-645
•  Meynier,  A.,  Lecoq,  C., and  Genet,  C. (2005). Emulsification enhances the retention of esters and aldehydes to a greater extent than changes in the droplet size distribution of the emulsion. Food Chem., 93, 153-159
•  Philippe,  E.,  Seuvre,  A. M.,  Colas,  B.,  Langendorff,  V.,  Schippa,  C., and  Voilley,  A. (2003). Behavior of flavor compounds in model food systems: a thermodynamic study. J. Agric. Food Chem., 51, 1393-1398
•  Roberts,  D. D.,  Pollien,  P., and  Watzke,  B. (2003). Experimental and modeling studies showing the effect of lipid type and level on flavor release from milk-based liquid emulsions. J. Agric. Food Chem., 51, 189-195
•  Sagitani,  H.,  Hattori,  T.,  Nabeta,  K., and  Nagai,  M. (1983). Formation of O/W Emulsion Having fine and Uniform Droplets by the Surfactant (D) Phase Emulsification Method. Nippon Kagaku Kaishi, 10, 1399-1404
•  Seuvre,  A. M.,  Phillippe,  E.,  Rochard,  S., and  Voilley,  A. (2007). Kinetic study of the release aroma compounds in different model food systems. Food Res. Int., 40, 480-492
•  Shimoda,  M.,  Miyamae,  H.,  Nishiyama,  K.,  Yuasa,  T.,  Noma,  S., and  Igura,  N. (2011). Swirl-Flow Membrane Emulsification for High Throughput of Dispersed Phase Flux through Shirasu Porous Glass (SPG) Membrane, J. Chem. Eng. Jpn., 44, 1-6
•  Tamaru,  S.,  Igura,  N., and  Shimoda,  M. (2018). Effectiveness of water-air and octanol-air partition coefficients to predict lipophilic flavor release behavior from O/W emulsions. Food Chem., 239, 712-717
•  Tamaru,  S.,  Ono,  A.,  Igura,  N., and  Shimoda,  M. (2019). High correlation between octanol-air partition coefficient and aroma release rate from O/W emulsions under non-equilibrium. Food Res. Int., 116, 883-887