In order to restrict the experimental variables to the formulation, the behavior of the droplet size was evaluated as a function of homogenization time under constant shear stress. The results presented in Table 2 indicate that for up to 60 s, the homogenization time is insufficient to efficiently reduce droplet size in the emulsions. However, when the homogenization time increased from 30 to 90 s, oil droplet size reduced significantly from 1.38 to 1.23 µm and after 90 s, droplet size remained unaltered. Droplet size distribution was bimodal, but presented one predominant peak, for all the tested homogenization times (Fig. 2.a). The span was another parameter that indicated the polydispersity of the oil droplets formed. In this time range, the span changed from 20.03 to 4.45%. The high shear mixer was not able to produce a narrow range of droplets size at the homogenization times of 30 and 60 s, as the emulsion did not stay enough time in the homogenization zone to be efficiently disrupted (McClements, 2004). No significant differences in size and span were observed for homogenization times above 90 s (Table 2).
Table 2
Oil droplet size of emulsions (D3,2) and span, as a function of homogenization time, under fixed emulsification conditions: oil fraction (Φ) = 0.08, 20.000rpm. All solutions were at neutral pH (pH ≈ 6.4–6.8) and with 12% of surfactant (Capsul).
Homogenization time (s)
|
D3,2 (µm)
|
Span
|
30
|
1.38 ± 0.02a
|
20.03 ± 11.36a
|
60
|
1.34 ± 0.04ab
|
9.57 ± 0.34ab
|
90
|
1.23 ± 0.04b
|
4.45 ± 0.53b
|
120
|
1.22 ± 0.05b
|
4.95 ± 0.85ab
|
240
|
1.21 ± 0.07b
|
8.4 ± 3.48ab
|
Values are expressed as mean ± standard deviation. In each column, different letters correspond to statistically different values (p < 0.05).
The reduction of droplet size with the increase in the homogenization time or rotation speed has been extensively reported in the literature (Kori, Mahesar, Sherazi, Khatri, Laghari & Panhwar, 2021; McClements, 2004; Walstra, 1983); however, it is limited and dependent on the nature of the surfactant, the concentration of the ingredients used, as well as the power density of the mixer, as all these parameters influence hydrodynamic conditions. Tcholakova, Denkov, Sidzhakova, Ivanov & Campbell (2003) also observed that for other surfactants, surfactant:oil ratio and homogenizer type, a steady droplet size distribution value is reached after some time. This value is dependent on surfactant concentration, since the size of the droplets decreases and therefore, the interfacial area increases as homogenization proceeds. Greater concentrations tended to reach the steady size value earlier (Tcholakova, Denkov, Sidzhakova, Ivanov & Campbell, 2003).
However, even though there is sufficient surfactant in solution, the mechanism of disruption provided by the homogenizer implies that in ensuring enough time is spent within the homogenization zone, the emulsion will be sufficiently disrupted (Walstra, 1983). In this study, a high-speed mixer, Ultra-Turrax IKA T18 basic (Wilmington, USA) was used. This type of rotor-stator homogenizer is widely used for preparing emulsions with low or intermediate viscosities (McClements, 2004). Several works in the literature that focused on the microencapsulation of oils and flavors use this device, and vary the homogenization times and rotation speed. Xiao, Kang, Hou, Niu & Kou (2019) investigated the effects of OSA-modified starch and MD on the release properties of essential oils. The authors evaluated the homogenization times of the emulsions from 10 to 50 min, using a high-speed mixer at 10,000 rpm as the homogenizer/disperser. Wang, Liu, Wen, Li, Wang & Ni (2017) studied the microencapsulation of chili seed oil by spray drying with starch sodium octenylsuccinate and MD, and prepared the emulsions by stirring at 10,000 rpm for 3 min using a disperser (T18 digital Ultra Turrax, IKA). Carneiro et al. (2013) studied flaxseed oil microencapsulation by spray drying different combinations of wall materials (maltodextrin (MD), whey protein concentrate, gum Arabic and two chemically n-octenyl succinic anhydride (OSAN)-modified starches: Capsul TA® and Hi-Cap 100TM). Emulsions were formed using an Ultra-Turrax homogenizer MA-102 (Marconi, Piracicaba, Brazil) operating at 18,000 rpm for 5 min. De Melo Ramos, Silveira Júnior & Prata (2019) prepared emulsions for the microencapsulation process by vacuum spray drying, the combination of OSA-modified starch and MD as wall materials and orange essential oil as active compound; with stirring at 20,000 rpm for 90 s using the Ultra-Turrax IKA T18 basic homogenizer (Wilmington, USA) to form the emulsions. Thereby, we have not observed a standard binomial homogenization time to rotation speed ratio, even for studies using similar wall materials and homogenization devices. Thus, a study was carried out to evaluate the influence of homogenization time on droplet size distribution, using a constant rotation speed of 20,000 rpm. Results show that homogenization for 120 s presented a sufficiently small droplet size. Therefore, this was defined as the best condition for the following steps.
3.1 INFLUENCE OF THE SURFACTANT AND OIL CONTENT ON THE DROPLET SIZE DISTRIBUTION AND STABILITY OF THE EMULSION
Under a given set of homogenization conditions, the emulsion droplet size distribution will depend on the relative composition of the emulsion (oil, water, and surfactant); and there is a maximum interfacial area that can be completely covered by the surfactant. The surfactant acts in two situations: firstly, it helps to reduce the interfacial tension, which allows for oil droplet diameter reduction. Then, it forms a film around the oil droplets for stabilization.
The size for all tested formulations ranged from 0.89 ± 0.02 to 2.08 ± 0.30 µm (Table 3), with bimodal size distributions (Fig. 2.a-c), but tending to have one predominant peak. The span values indicated a high polydispersity degree in the droplet size distribution for all the emulsions. Span values for bimodal or multimodal emulsions are not representative for comparison purposes between droplet size distribution curves (Silva, Costa, Gomes, Bargas, Cunha & Meireles, 2018).
Table 3
Characterization of emulsions regarding oil droplet size, span and kinetic stability (evaluated by the optical scanning instrument Turbiscan ASC).
Formulation
|
d-Limonene oil
(% w/w)
|
MD
(% w/w)
|
MS Capsul
(% w/w)
|
D [3,2] (µm)
|
Span
|
48 h emulsion
stability (phase separation)
|
C10
|
8
|
28
|
4
|
1.45 ± 0.04a
|
3.21 ± 0.13c
|
No
|
C20
|
8
|
24
|
8
|
1.00 ± 0.03b
|
4.38 ± 0.48bc
|
No
|
C30
|
8
|
20
|
12
|
0.98 ± 0.03bc
|
7.71 ± 2.00b
|
No
|
C40
|
8
|
16
|
16
|
0.96 ± 0.02bc
|
13.33 ± 1.65a
|
No
|
C60
|
8
|
8
|
24
|
0.89 ± 0.02c
|
15.30 ± 0.50a
|
No
|
O30–1
|
12
|
21
|
7
|
1.71 ± 0.01b
|
11.57 ± 0.07ab
|
No
|
O30–2
|
12
|
7
|
21
|
0.93 ± 0.03c
|
14.57 ± 2.05a
|
No
|
O40–1
|
16
|
18
|
6
|
1.79 ± 0.03ab
|
1.98 ± 0.44c
|
No
|
O40–2
|
16
|
6
|
18
|
0.94 ± 0.03c
|
3.95 ± 0.67c
|
No
|
O50–1
|
20
|
15
|
5
|
2.08 ± 0.30a
|
1.66 ± 0.08c
|
Yes
|
O50–2
|
20
|
5
|
15
|
1.04 ± 0.04c
|
2.79 ± 0.30c
|
No
|
Values are expressed as mean ± standard deviation. In each column, different letters correspond to statistically different values (p < 0.05).
The formulations containing the same oil content and different modified starch concentrations (10–60% in relation to the total solid concentration) presented similar distribution profiles. The exception was the formulation containing the lowest surfactant concentration (C10 sample) (Fig. 2), which presented a peak with a greater mean diameter (1.45 um), while all others were 1 µm.
The droplet size was reduced by increasing the concentration of MS Capsul. The influence of surfactant concentration on droplet size can be divided into two regions, as already observed in the literature for other surfactants (McClements, 2004; Tcholakova, Denkov, Ivanov & Campbell, 2006), and can be seen in Fig. 3.a. In the region with high surfactant/oil ratio (surfactant rich regime), the mean droplet size is not dependent on the surfactant concentration, but rather, is mainly controlled by the processes of droplet disruption (Taisne, Walstra & Cabane, 1996; Tcholakova, Denkov & Danner, 2004). However, when surfactant concentration decreased from 8–4% (surfactant poor regime), the mean droplet size increased from about 0.95 µm to 1.5 µm.
Other authors reported similar qualitative results for different concentrations and types of surfactants (protein, non-ionic and anionic surfactants; in high and low concentrations of electrolytes). Regarding proteins, they explained that adsorption on the oil surface is practically independent of any external factor (hydrodynamic conditions during emulsification, initial protein concentration and oil fraction) and that the protein quickly fills the monolayer (Tcholakova, Denkov & Danner, 2004). Similar to proteins, MS Capsul (OSA starch) has amphiphilic properties, which confers emulsifying properties to this hydrophobically modified starch derivative (Sweedman, Tizzotti, Schäfer & Gilbert, 2013; Trubiano, 1995). However, the mode of adsorption of MS Capsul at an oil-water interface is different from that of proteins. The surface activity of the MS Capsul hydrocolloid is related to its non-polar character of chemical groups attached to the hydrophilic polysaccharide backbone (Dickinson, 2009; Wang, Li, Chen, Xie, Yu & Li, 2011). In the case of proteins, the distribution pattern of hydrophobic and hydrophilic segments on the surface is complex, as well as the structural rigidity of the molecule, causing limitations in adsorption and orientation (Damodaran & Park, 2017). Although proteins contain both hydrophobic and hydrophilic groups in their primary structure, there is no clearly defined hydrophilic head or hydrophobic tail (Damodaran & Park, 2017). Most of the non-polar (hydrophobic) groups are inserted inside the protein, and the polar (hydrophilic) groups are present on the outside (Damodaran & Park, 2017; Dalgleish, 2006; Jayasundera, Adhikari, Aldred & Ghandi, 2009).
Next, as C20 and C60 presented small droplet size and opposite wall material formulations (MD:MS 3:1 and 1:3, respectively), they were chosen as the base formulations for the oil saturation study.
In this second study, the surfactant to oil (MS / Oil) ratio varied from 0.4 to 3.0 g surfactant/g oil (Fig. 3.b). The results are shown in Table 3. The increase in surfactant concentration favored the reduction in oil droplet size, although the reduction was more sensitive at lower concentrations of surfactant considering the amount of oil, for all the oil saturated formulations (C30-1, C30-2, C40-1, C40-2, C50-2 and C50-1). Emulsions with MS concentrations below 7% (O30-1, O40-1 and O50-1) presented larger oil droplet sizes (Table 3). The reasons for this increase in droplet size is because of the increase in coalescence due to the increased content of the oil phase, and the insufficient concentration of surfactants to form a physical layer around the oil drops (McClements, 2004).
In the work performed by Hadnađev, Dokic, Krstonosic & Hadnađev (2013), the increase in oil concentration resulted in an increased coalescence of the droplets. The authors concluded that higher oil concentrations require more surfactant for surface coverage of oil droplets, and to maintain the average droplet size.
The value of 0.5 g of surfactant/g oil was considered as a critical value because there is not enough of surfactant to continue to reducing oil droplet size; and droplet sizes were kept greater than 1 µm (Fig. 3.b). However, the increase of the surfactant to oil ratio up to the maximum adopted in this study (3 g of surfactant / g oil) led to a reduction from 1 µm to 0.9 µm (Fig. 3.b). Similar results were obtained in studies that used the same continuous phase (MD and MS Capsul) and homogenization device (rotor-stator Ultra-Turrax) for emulsion preparation. De Melo Ramos, Silveira Júnior & Prata (2019) obtained droplet mean diameters of 0.913 µm in their fresh emulsions, using the same solids concentration (40% total solids w/w and 4 g of surfactant / g oil ratio) and stirring speed (20,000 rpm) as this study.
We also observed that oil droplet surface area was largely influenced by surfactant concentration (Fig. 3.c). Indeed, the formulations with a concentration above 8% of MS presented higher oil droplet surface area compared to the other formulations (Table 3).
The oil droplet size diameter was mainly influenced by the concentration of MS Capsul. The formulations with the highest concentration of surfactant were the ones that obtained the smallest droplets sizes (Table 3). The results obtained in our study are similar to those obtained by Paulo, Alvim, Reineccius & Prata (2020), who reported the same reduction in oil droplet size by increasing the concentration of MS Capsul, which increased the stability of the emulsions through the contribution of surfactant viscosity.
Reineccius (1991) and Drusch, Serfert, Scampicchio, Hansberg & Schwarz (2007) also reported smaller oil droplet sizes for emulsions prepared from an OSA modified starch, compared to emulsions prepared from gum Arabic. Carneiro et al. (2013) verified that the droplet sizes are not affected only by the viscosity of the emulsion, but also by the intrinsic emulsifying properties of each type of wall material.
Studies on the reduction of emulsion oil droplet size demonstrated better results in emulsion stability, retention of active material during spray drying and prevent air inclusion in the particles because they are more efficiently enclosed and embedded within the wall matrix of microparticles (Drusch, Serfert & Schwarz, 2006; Jafari, Assadpoor, He & Bhandari, 2008; McClements, 2004; McClements, Decker & Weiss, 2007;Soottitantawat, Yoshii, Furuta, Ohkawara & Linko, 2003).
3.1.1. EMULSION STABILITY
A visual observation method, the creaming index, was also applied to evaluate kinetic stability of the studied formulations. The samples remained visually stable for 48 h, without phase separation. Only sample O50-1 presented a creaming layer during the study. After, this phenomenon was confirmed by the BS profile over time (Fig. 4.b).
In most of the cases, emulsions did not generally undergo only one instability phenomenon, but several at the same time. The stability analysis using Turbiscan allows for a macroscopic visualization of the kinetic stability of concentrated dispersions, making it possible to discriminate various destabilization mechanisms.
The stability study using Turbiscan revealed that most of the emulsions (O50-2, C20, C30, O40-1 and C10) presented tendencies towards flocculation/coalescence phenomenon, but only in the initial stage, and resulted in products with similar profiles of backscattering (BS), as shown in the Fig. 4. In addition, BS slightly decreased during the 48 h (except for C10), due to the global increase of emulsion oil droplet sizes. Despite this, these destabilization mechanisms did not promote phase separation of the samples, either by creaming or sedimentation, and there was no change in the BS along the height of the sample. Therefore, these samples proved to be stable during the studied period. Depending on the initial size of the droplets (smaller or larger than the wavelength, 880 nm), the level of backscattering will increase or decrease, as the backscattering is inversely proportional to the size of the droplets. Only in the sample C10 was there an increase in the level of BS at 24 and 48 h.
However, as previously mentioned, a low concentration of modified starch and high concentration of oil, corresponding to the formulation O50-1, favored system destabilization during a period of 24 h, and a separate layer of creaming was observed (Fig. 4.b). This behavior was due to the low density of the oil phase in relation to the water phase (very common in food emulsions), so a net gravitational force acts upon them (Lakkis, 2007; McClements, 2004), and the droplets tend to move upwards, and form a layer of creaming (McClements, 2007). The 24 h period for phase separation can be justified by the high oil concentration, as it reduces the speed of creaming (Tadros, Izquierdo, Esquena & Solans, 2004). Thus, a minimal ratio of 0.25 w/w surfactant:oil is needed to allow emulsion stability.
The creaming behavior demonstrated that the emulsion suffered droplet aggregation. The decrease of the BS profile with increasing observation time may be related to greater oil droplet size, thus it is an indicative of the occurrence of droplet aggregation (Gomes & Kurozawa, 2020).
All O/W emulsions remained stabilized and no phase separation was observed under a period of 2 hours. It implies that they are suitable for the microencapsulation process of spray drying. This was expected, because OSA modified starches possess excellent emulsification properties. One of the advantages in using modified starch as the stabilizer in emulsions is due to the presence of both lipophilic and hydrophilic groups, which are responsible for the starch molecules to be attracted to the oil-water interface and form a viscoelastic film around the oil droplets (Friberg & Larsson, 1997). The formation of the interfacial film prevents oil droplets from coalescing and forming larger droplets through steric repulsion (Friberg & Larsson, 1997; McClements, 2004). Compared to proteins or surfactants, the modified starch has a slightly low interfacial activity, and therefore, a large excess must be added to ensure that every oil droplet surface is adequately coated (McClements, 2004; Paulo, Alvim, Reineccius & Prata, 2020). Additionally, Drusch et al. (2007) highlighted that OSA starch molecules have high mobility, and thus, rapidly stabilize the new developed oil–water interface after homogenization. Modified starches can be used at high infeed solid levels for spray drying microencapsulation (compared to gum Arabic), and may provide excellent emulsion stability (Risch & Reineccius, 1988; Trubiano & Lacourse, 1988).
The ratio limit between surfactant mass to surface area above 0.4 g/µm2 was determined to keep the emulsion stable (Fig. 3.c), in which lower oil droplet sizes around 1 µm were observed.
3.2. RHEOLOGICAL BEHAVIOR
For the rheological and following studies of this work, 6 samples were selected: C10, C20, C30, O40-1, O50-1 and O50-2. The criterion used to select these formulations were based on the observed influences of the wall material and oil concentrations on stability and oil droplet size (Table 3) of the evaluated emulsions.
Feed emulsion viscosity was determined through steady-shear flow curves (Fig. 5). Overall, the emulsions exhibited Newtonian flow behavior (n ≈ 1), in which the sample viscosities were constant with shear rate. Only the formulation, that presented characteristics of a pseudoplastic fluid (n < 1) was C20; therefore, the viscosity is inversely proportional to the shear rate.
The formulations containing a high concentration of oil (O40-1, O50-1 and O50-2) presented the lowest viscosity values. On the other hand, for formulations with a high concentration of wall matrix and low concentration of d-limonene, there was an increase in emulsion viscosities. Similar behavior was observed by Bae & Lee (2008), Tonon, Grosso & Hubinger (2011) and Frascareli, Silva, Tonon & Hubinger (2012).
Analyzing Fig. 5 and Table 3, an inverse relationship between emulsion viscosity and oil droplet size is observed. The same behavior was also observed by Hong, Kim & Lee (2018), who investigated the effects of Hydrophilic-Lipophilic Balance (HLB) value on the O/W emulsion stability and rheological behavior of mixed nonionic surfactants, Span/Tween. In addition, the authors noted that the smaller the oil droplet size, the greater the surface area of the oil droplets in contact to the rheometer’s geometry, thus leading to an increase of viscosity.
3.3. IMPACT OF THE ATOMIZATION PROCESS ON THE DROPLET SIZE DISTRIBUTION
The impact of emulsion formulation (C10, C20, C30, O40-1, O50-1 and O50-2) was evaluated from three microscopic spray characteristics: atomized droplet sizes, drop size distributions and droplet velocities.
3.3.1. DROPLET SIZE
Tables 4 and 5 show the variation of arithmetic diameter (D10) and Sauter Mean Diameter (D3,2), respectively, for each emulsion formulation at the measurement positions of 5 cm, 10 cm, and 20 cm downstream from the atomizing nozzle tip along its axis.
Table 4
Variation of D10 (µm) for different formulations and measurement positions.
Formulation
|
D10 – 5 cm
|
D 10 − 10 cm
|
D 10 − 20 cm
|
C 10
|
40.95 ± 0.75bcG
|
55.8 ± 1.80bABC
|
58.1 ± 0bA
|
C 20
|
30.25 ± 0.64cH
|
41.2 ± 3.54cG
|
47.8 ± 1.56cG
|
C 30
|
29.3 ± 0.71cH
|
40.45 ± 0.49cG
|
44.25 ± 1.91cFG
|
O 40–1
|
48.2 ± 0.42baEFG
|
63.8 ± 3.39baBCD
|
69.35 ± 1.34aAB
|
O 50–1
|
59.4 ± 1.84aCD
|
71.25 ± 0.49aAB
|
73.1 ± 0.28aAB
|
O 50–2
|
44.75 ± 7.42bG
|
66.85 ± 2.62abCDE
|
75.05 ± 1.77aDEF
|
Values are expressed as mean ± standard deviation. Lower case and capital different letters correspond to statistically different values (p < 0.05) to each column and line, respectively.
Table 5
Variation of D3,2 (µm) for different formulations and measurement positions.
Formulation
|
D3,2 − 5cm
|
D3,2 − 10 cm
|
D3,2 − 20 cm
|
C 10
|
149.4 ± 1.13bA
|
146.0 ± 0.40bDEFG
|
148. 05 ± 021aG
|
C 20
|
149.1 ± 10bcFG
|
149.75 ± 0.64aFG
|
147.50 ± 0.57aFG
|
C 30
|
145.95 ± 0.35dBCDE
|
146.0 ± 0bBC
|
146.65 ± 0.15abCDEFG
|
O 40–1
|
150.35 ± 0.64abAB
|
145.65 ± 1.48bFG
|
147. 10 ± 0.14aDEFG
|
O 50–1
|
146.8 ± 0.14dcEFG
|
145.4 ± 0.57bG
|
145.15 ± 0.21cG
|
O 50–2
|
152.6 ± 0.71aBCD
|
147.05 ± 0.35abFG
|
145.40 ± 0.57bcBCDEF
|
Values are expressed as mean ± standard deviation Lower case and capital different letters correspond to statistically different values (p < 0.05) to each column and line, respectively.
The results presented in Table 4 indicate that droplet size diameter (D10) increases with distance between the drops and the atomizer nozzle.
Formulations C10, C20, and C30 presented similar behavior at three measurement positions (Figs. 6.a, 6.b, and 6.c). Also, an increase in the concentration of MS Capsul reduced droplet diameter, as MS has emulsifying capacity.
The formulations O40-1 and O50-1 (which had a high oil concentration and low MS Capsul concentration) showed a significant increase in droplet diameter for the three measurement positions. O50-2, however, presented droplet diameters similar to the C10 formulation, at the distance of 5 cm. However, between the tested formulations, it was the one which showed a considerable increase of droplet diameter at further distances. Formulations with high oil concentrations and low viscosity, presented the highest diameters (D10) for the atomization process, with two channel atomizers.
Kleinhans, Hornfischer, Gaukel, & Schuchmann (2013) studied the influence of viscosity ratio and initial oil drop size on the oil drop breakup during effervescent atomization. The authors verified that emulsions with high oil concentrations presented larger oil droplets for atomization with effervescent atomizers.
The viscosity and the interfacial tension are the emulsion feeding properties that oppose the disintegration of the jet into drops (Lefebvre, 1989). Thus, the decrease of these properties favors disintegration, and reduces the droplet diameter and improves atomization efficiency (Nigra Júnior, Krieger Filho & Santos, 2019). By our results, we could not observe a significant effect of viscosity on the reduction of droplet size.
According to Jafari, Assadpoor, He & Bhandari (2008), this behavior can be related to the increase of oil concentration, to droplet coalescence due to the increase of the collision frequency, or to droplet breakdown processes due to the alteration of emulsion viscosity.
Evaluating the D3,2 results (Table 5), a great variation in droplet size between formulations at the distance of 5 cm is observed. C20 and C30 presented an increase of droplet diameter, as the size of C10, O40-1, O50-1, and O50-2 decreased; from a distance of 5 cm to 10 cm. Moreover, O50-2, O40-1, and C10 had the greatest variations of D3,2 over the three measurements. It was not possible to determine a pattern for all evaluated emulsions. In future work, an analysis by means of the Shadowgraph optical method could be performed, to evaluate the flow field through the shadow projection of the atomized drops.
3.3.2. DROP SIZE DISTRIBUTION
Figures 6 and 7 show the Probability Density Functions (PDF) of the diameter and velocities of the droplets, respectively, for the following test conditions: measurement positions of 5 cm, 10 cm, and 20 cm downstream from the atomizing nozzle tip along its axis, and atomizing pressure of 200 kPa. The axial measurement positions were defined based on the dimensions of the spray dryer chamber. Each PDF curve was constructed based on the method of acquiring drops over a certain time period. The characteristics of single drops were detected and validated by the PDPA technique over multiple atomization cycles for a period of approximately 1 minute. This duration was enough to obtain statistical convergence of the data.
During the atomization process, a wide range of droplet sizes were found. To visualize how these droplet sizes are distributed at the measurement point, drop size distributions were determined, which are represented by Probability Density Functions (PDF) of droplet diameters (Fig. 6).
The PDF obtained for the different formulations at the different measurement points indicate a monomodal behavior in diameter distribution (Fig. 6). Table 6 shows the peaks most likely to find droplets of a given diameter (i.e. the emulsion drop size distributions are centered on these droplet diameters).
Table 6
Diameter values of the predominant peak of droplet size distribution curves for different formulations and measurement positions.
Formulation
|
Diameter in axial measurement points (µm)
|
Standard deviation
|
|
5 cm
|
10 cm
|
20 cm
|
|
C10
|
12.0
|
14.1
|
13.0
|
1.02
|
C 20
|
11.8
|
11.5
|
12.9
|
0.75
|
C 30
|
13.6
|
12.5
|
15.1
|
1.30
|
O 40 − 1
|
8.6
|
12.6
|
13.3
|
2.53
|
O 50 − 1
|
8.5
|
12.1
|
14.6
|
3.06
|
O 50 − 2
|
6.7
|
12.5
|
16.9
|
5.12
|
For the air atomization experiments at the measuring position of 5 cm downstream from the atomizer nozzle, the O40-1, O50-1 and O50-2 formulations (high oil concentration and low wall material concentration) present a higher probability of finding smaller droplets below 10 µm; compared to the other emulsion formulations. On the other hand, formulations C10, C20, and C30 (low oil concentration and high wall material concentration) present a higher probability of finding droplets between 10 µm and 40 µm. The O50-1 formulation has a higher probability of finding larger droplets above 40 µm.
In the measurement positions of 10 cm and 20 cm, an increase in the size of droplet diameters for all formulations, and a flattening of the droplet size distribution curves and a consequent reduction of the PDF values (Table 6) is observed. This result may be related to the shape of the larger droplets, which are often not spherical and the PDI technique cannot detect these drops.
The experimental results demonstrated that the O50-2, O40-1 and O50-1 emulsions showed an increase in droplet diameter, during the atomization process in the air between the three axial measurement points (Table 6). This behavior can be justified by an intense coalescence of the droplets for formulations with higher oil content (16% and 20%) and low wall material concentration (Table 3). Chesters (1991) explains that coalescence is greatly favored by a low dispersed phase viscosity and by high disperse phase fractions.
Taboada, Heiden-Hecht, Brückner-Gühmann, Karbstein, Drusch & Gaukel (2021) investigated the changes in oil droplet size in whey protein–stabilized emulsions, during the atomization and the subsequent drying step of a spray drying process. The authors observed a destabilization by coalescence of the oil droplets for emulsions with a high oil content of 30% wt., directly after breakup during atomization.
During the spray drying process, in which the emulsion is dried into a powder, the integrity of the interfacial layer surrounding the droplets is disrupted, which leads to coalescence within the drying chamber. The coalescence stability can be improved by adding relatively high concentrations of protein or carbohydrates to the system prior to drying (Hebishy, Buffa, Guamis, Blasco-Moreno & Trujillo, 2015; Taboada, Heiden-Hecht, Brückner-Gühmann, Karbstein, Drusch & Gaukel, 2021; Tcholakova, Denkov, Ivanov & Campbell, 2006; Young, Sarda & Rosenberg, 1993). This is possible because these molecules form a thick interfacial membrane around the droplets that is less prone to disruption during the dehydration process (McClements, 2004). In this work, the C10, C20 and C30 formulations, which have a higher concentration of wall material and a lower concentration of oil (8%) (Table 3), remained more stable against the coalescence destabilization process. Furthermore, the C10, C20 and C30 emulsion formulations showed a similar behavior in droplet size distributions and approximate emulsion droplet diameter values at the three measurement positions evaluated (Fig. 6).
Sijs, Kooij, Holterman, van de Zande, & Bonn (2021) investigated four different methods for measuring droplet size distributions: the Image Analysis VisiSizer technique, a stroboscopic imaging method developed in-house; the phase Doppler particle analysis (PDPA); and laser diffraction (Malvern Spraytec). The authors observed that the larger the droplet size, the greater the differences between the results obtained by the different methods. In addition, their measurements confirmed how the limitations of the PDPA can influence the results. The presence of an internal structure within the droplets, such as air inclusions, caused by surfactants or emulsions, can be confused with smaller droplets (Sijs, Kooij, Holterman, van de Zande, & Bonn, 2021).
3.3.3. DROPLET VELOCITIES
Figure 7 shows the PDFs of the axial drop velocities for the six evaluated emulsions, and their respective measurement positions during the atomization process.
All PDFs presented single-modal velocity distributions. Comparing the three measurement positions, the formulations taken at 5 cm from the atomizer nozzle outlet (Fig. 7.a) presented velocity distributions with a high probability of finding drops with higher velocities. As they move away from the nozzle, the velocity of these droplets is considerably reduced at the measurement positions of 10 cm and 20 cm (Fig. 7.b-c). The behavior of drop velocity distributions at 10 cm and 20 cm was similar. Furthermore, it was observed that the C10, O 50 − 1 and C30 emulsions presented lower velocities than the C20, O40-1 and O50-2 emulsions. The C10 formulation presented a higher probability of finding drops with lower velocity than the other formulations, and its velocity distribution is centered at 5.39 m/s, while the O 50 − 2 formulation presented a higher probability of finding drops with higher velocity, and its velocity distribution is centered at 24.65 m/s; at 5 cm from the atomizing nozzle.
These results showed that viscosity had no significant effect on the speed distribution of the emulsions since the C20 formulation, which is more viscous, presented a centered velocity of 21.09 m/s, a result close to that of the O 50 − 2 emulsion, which has lower viscosity. Evaluating the drop diameter distributions (Fig. 6) with the drop velocity distributions (Fig. 7) of the tested emulsions, it is not possible to infer any correlation between these results.
3.4. INFLUENCE OF FORMULATION COMPOSITION ON MICROPARTICLE PROPERTIES OBTAINED BY SPRAY DRYING
3.4.1. MOISTURE AND WATER ACTIVITY
The powder moisture content (Table 7) varied from 0.19–14.38%. Higher oil concentrations (16% and 20% in the formulations O40-1, O50-1, and O50-2, Table 3) resulted in higher moisture content for the microcapsules. The high concentration of oil may have saturated the feed emulsion, thus, the asset could not be completely encapsulated and remained on the particles’ surface, which also limited the diffusion of water through the wall during spray drying. Moreover, as D-limonene is a volatile compound it could be volatilized during the determination of the analysis and contribute to an overestimation of moisture content.
Table 7
Powder moisture and water activity (aw) of d-limonene oil microencapsulated.
Formulation
|
Powder Moisture (%)
|
Aw
|
C10
|
1.39 ± 0.071c
|
0.026 ± 0.014b
|
C20
|
0.19 ± 0.10 c
|
0.026 ± 0.004b
|
C30
|
0.78 ± 0.01c
|
0.034 ± 0.009b
|
O40–1
|
12.56 ± 0.23b
|
0.036 ± 0.004ab
|
O50–1
|
13.67 ± 1.09ab
|
0.032 ± 0.008b
|
O50–2
|
14.38 ± 0.3a
|
0.050 ± 0.007a
|
Values are expressed as mean ± standard deviation. In each column, different letters correspond to statistically different values (p < 0.05).
All samples with 8% oil (C10, C20 and C30) (Table 3) had low moisture content. This may be related to a high maltodextrin fraction in those samples. Previous works observed that the increase in maltodextrin concentration resulted in a lower moisture content for the obtained powders (Quek, Chok & Swedlund, 2007; Rodríguez-Díaz, Tonon & Hubinger, 2014).
Water activity (aw) of the microparticles (Table 7) were also low, ranging from 0.026 to 0.05. These values promote positive effects on the preservation of microparticles against microorganisms. Moreover, as aw > 0.2 for all the samples tested, it also prevents the oxidation of the encapsulated oil.
There was no significant difference in aw between the samples with different proportions of wall material and oil, except for sample O50-2 (20% oil and wall material in the proportion of 5/15 (MD / MS). Similar results were reported by Carneiro et al. (2013), for particles prepared with MS Capsul and MD.
3.4.2. PARTICLE SIZE DISTRIBUTION AND MICROSTRUCTURE
As shown in Fig. 8, all samples exhibit a bimodal distribution in their particle size. Particle mean diameter (Table 8) varied from 17.72 to 29.9 µm, and the particle size distribution ranged from 1µm to 416 µm. Note that, formulations with a high oil concentration and low modified starch concentration (O40-1 and O50-1) displayed smaller particle size despite having a larger oil droplet size. This can be explained by the low viscosity of these emulsions (Fig. 5), due to the reduction in the concentration of wall material. According to previous studies, viscosity has a great influence on both the droplet size formed during atomization, and on the size of the dry particles produced by spray drying.
Table 8
Powder size (D4,3), particle polydispersity (PPD), encapsulation efficiency of microparticles and D3,2 of the reconstituted emulsion.
Formulations
|
Particle size (D4,3)
(µm)
|
PPD
|
Encapsulation efficiency (%)
|
D3,2 of reconstituted emulsion
(µm)
|
C10
|
26.76 ± 1.60 ab
|
2.70 ± 0.23 ab
|
78.90 ± 3.02a
|
2.04 ± 0.13a
|
C20
|
29.35 ± 2.01a
|
2.89 ± 0.22 ab
|
84.27 ± 0.00a
|
0.93 ± 0.01c
|
C30
|
29.90 ± 2.41a
|
3.08 ± 0.43 a
|
84.77 ± 0.00a
|
0.89 ± 0.01c
|
O40-1
|
20.71 ± 2.32 bc
|
2.71 ± 0.17ab
|
63.73 + 5.10b
|
1.15 ± 0.06b
|
O50-1
|
17.72 ± 1.16 c
|
2.40 ± 0.15b
|
51.63 ± 1.38c
|
1.15 ± 0.02b
|
O50-2
|
24.84 ± 7.52 ab
|
2.69 ± 0.42 ab
|
81.53 ± 1.39a
|
0.92 ± 0.01c
|
Values are expressed as mean ± standard deviation. In each column, different letters correspond to statistically different values (p < 0.05).
In this work, emulsions with a smaller oil droplet size (C20, C30 and O50-2) (Table 3) presented large particle size and high encapsulation efficiency (Table 8). Similar results were also reported for spray dried microparticles containing fish, moringa, flaxseed, and soya oils (Carneiro, Tonon, Grosso & Hubinger, 2013; Chang, Tan, Tan, Nehdi, Sbihi & Tan, 2020; Hogan, McNamee, O’Riordan & O’Sullivan, 2001; Premi & Sharma, 2017). Larger particle sizes protect the active ingredient inside the carrier agent more efficiently than smaller ones, and thus, has a positive effect in protection by increasing the shell around the core material of the particle (Soottitantawat, Bigeard, Yoshii, Furuta, Ohkawara & Linko, 2005).
Frascareli et al. (2012) found that the presence of larger particles (over 40 µm) can be associated with an onset of agglomeration, due to the formation of irreversible bonding bridges (bridging flocculation). These authors indicated that the presence of free oil on the surface of the particles may be related to the formation of these bonds.
SEM images of microparticles produced from different mixtures are shown in Fig. 9. The presence of spherical and rough particles, in a wide variety of sizes, can be observed for all dry emulsions. The particles have a rigid, porous appearance without breakage, which may be related to the high drying temperature that the emulsions were subjected to around 180°C (Oakley, 1997).
In addition, the microparticles showed no cracks. It is known that the production of intact microcapsules is vital to protect the oil from oxidation reactions and uncontrolled leakage of oil droplets onto the surface of the particles (Shamaei, Seiiedlou, Aghbashlo, Tsotsas & Kharaghani, 2017).
Particles were cut to observe their internal microstructure (Fig. 9.a2, 9.b2 and 9.c2). Analyzing the internal microstructure of the microparticles with the highest oil concentration (Fig. 9.b2 and 9.c2), larger amounts of oil droplets were distributed on the walls of the particles, in relation to the microparticles that had 12% less oil load (sample C20, Fig. 9.a2).
In drying experiments, there are changes in the size and shape of particles (Alamilla-Beltrán, Chanona-Pérez, Jiménez-Aparicio & Gutiérrez-López, 2005). The morphological and size differences between the powders obtained from different mixtures result from the drying rate that each formulation undergoes during dehydration. At high temperatures, the crusting speed is very fast; therefore, for formulations with a high oil content and low concentrations of wall material, the oil may not be completely encapsulated and will remain at the droplets surface (Figs. 9.a, 9.b, and 9.c), limiting the diffusion of water.
Powders with a high oil content showed a greater tendency to agglomerate (mainly formulation C 50 − 1), which can be explained by the increase in free surface oil (Taboada, Heiden-Hecht, Brückner-Gühmann, Karbstein, Drusch & Gaukel, 2021). The C50-1 formulation showed a greater impregnation of the powder on the wall of the dryer chamber, and consequently, presented a lower yield at the end of the drying process. The samples that showed the best process yields were C20, C10, and C30, respectively. Among the dry samples with a high oil content, the C50-2 formulation stands out, due to it having the highest concentration of MS Capsul (1:3 MD/MS). There was a positive impact on the drying process performance for formulations with the highest wall material concentrations.
3.4.3. ENCAPSULATION EFFICIENCY
A negative and linear correlation between encapsulation efficiency of the microparticles (Table 8) and oil droplet size in the emulsion was observed (Fig. 10), as reported by several authors (Danviriyakul, McClements, Decker, Nawar & Chinachoti, 2002; Gomes & Kurozawa, 2020; Jafari, Assadpoor, He & Bhandari, 2008; Jafari, He & Bhandari, 2007; Liu, Furuta, Yoshii, Linko & Coumans, 2000; Risch & Reineccius, 1988; Soottitantawat, Bigeard, Yoshii, Furuta, Ohkawara & Linko, 2005; Soottitantawat, Yoshii, Furuta, Ohkawara & Linko, 2003).
Above 0.75 of the surfactant to oil ratio, encapsulation efficiency above 80% was obtained. However, there was no significant difference between the C10, C20, C30 (with 8% oil) formulations and the O50-2 (with 20% oil) formulation. In addition, despite O50-2 presenting a high oil load, the high concentration of surfactant (15% MS) favored encapsulation efficiency.
For formulations containing a high oil load (16% and 20%) and low surfactant concentrations (C40-1 and C50-1), there was a reduction in the encapsulation efficiency compared to other samples. When the surfactant:oil ratio values were below 0.4 w/w, there was a considerable decrease in encapsulation efficiency. These results indicate a loss of free oil during the spray drying process (McNamee, O’Riorda & O’Sullivan, 1998).
The results suggest a tendency for samples with a higher viscosity to improve oil retention, due to the increase in the wall material concentration. Similar results were observed by Mehran, Masoum, & Memarzadeh (2020), who evaluated microencapsulation optimization conditions in spearmint essential oil by spray drying, using a blend of inulin and gum Arabic as wall material. Hogan et al. (2001) and Najaf Najafi, Kadkhodaee, & Mortazavi (2011) observed that flavor retention of the spray-dried microcapsules is dependent on the type and percentage of coating material, emulsion total solids concentration and spray-dryer atomizer nozzle size.
The surfactant (MS Capsul) in adequate concentrations allowed significant improvement in oil retention and process efficiency when applied to a formulation with a high oil load.
3.4.4. CORRELATION BETWEEN THE RECONSTITUTED EMULSION FROM THE MICROENCAPSULATED POWDER AND THE INITIAL EMULSION
Samples C20, C30 and O50-2 showed no significant difference in relation to the average oil droplet size of the reconstituted emulsion and the initial emulsion (Table 9), and is related to a high surfactant to surface area ratio. We observed significant difference for D10 between initial and reconstituted emulsion for all the samples, except C20 (Table 9). Thus, the hydrodynamic conditions of the process (i.e., initial homogenization or atomization followed by reconstitution) governed that parameter (Tcholakova, Denkov, Sidzhakova, Ivanov & Campbell, 2003). Also, there was a tendency of increase of D10 when comparing initial and reconstituted emulsions C10, C20, and C30 (Table 9). For D90, however, we observed an inverse tendency when compared to D10. C10, C20 and C30 had a decrease of D90 when reconstituted, while O40-1 and O50-2 presented an opposite trend (Table 9). Thus, the oil concentration of the samples had a great influence on the behavior of D90, when comparing initial and reconstituted emulsions. Mode and percentage of volume of the predominant peak are also presented in the Supplementary Table S.1.
In addition, they presented similar behavior in droplet size distribution, between the initial and reconstituted emulsion (Fig. 11.b, 11.c and 11.f). The C20, C30 and O50-2 formulations (Table 8) also presented the highest encapsulation efficiencies, with values above 81.53%. These results are interesting, because smaller oil droplet sizes after spray drying (i.e. < 1 µm) lead to increased stability of the oil phase in the powder, and high encapsulation efficiency; as reported by Hogan et al. (2001), Jafari et al. (2008) and Taboada et al (2021)
Table 9
Correlation between initial and reconstituted emulsion oil droplet diameters.
Formulation
|
Initial Emulsion
|
Reconstituted Emulsion
|
D 10 (µm)
|
D 90 (µm)
|
D3,2 (µm)
|
D 10 (µm)
|
D 90 (µm)
|
D3,2 (µm)
|
C10
|
0.76 ± 0.03cC
|
11.54 ± 2.69cC
|
1.45 ± 0.04bcBC
|
0.83 ± 0.01aB
|
5.19 ± 1.37bcDE
|
2.04 ± 0.13aA
|
C20
|
0.62 ± 0.01dDEF
|
25.21 ± 3.04aA
|
1.00 ± 0.03dD
|
0.67 ± 0.02cD
|
2.29 ± 0.16cE
|
0.93 ± 0.01dD
|
C30
|
0.60 ± 0.00dEF
|
16.59 ± 0.31bB
|
0.98 ± 0.03dD
|
0.73 ± 0.00bC
|
7.33 ± 1.03bCD
|
0.89 ± 0.01dD
|
O40–1
|
0.83 ± 0.03bB
|
5.17 ± 0.23dDE
|
1.79 ± 0.03abAB
|
0.65 ± 0.01cDE
|
6.32 ± 0.11bcDE
|
1.15 ± 0.06cdCD
|
O50–1
|
1.02 ± 0.02aA
|
6.86 ± 0.55dCDE
|
2.08 ± 0.30aA
|
0.63 ± 0.01cDEF
|
3.73 ± 0.32bcDE
|
1.15 ± 0.02dCD
|
O50–2
|
0.58 ± 0.02dF
|
3.63 ± 0.26dDE
|
1.04 ± 0.04dD
|
0.85 ± 0.03aB
|
16.47 ± 3.36aB
|
0.92 ± 0.01dD
|
Values are expressed as mean ± standard deviation. Lower case and capital different letters correspond to statistically different values (p < 0.05) to each column and line, respectively.
The reconstituted emulsions with high oil concentrations and 3:1 MD:MS Capsul proportion (O40-1 and O50-1), showed a significant reduction in the oil droplet size compared to initial emulsion (Table 9). As can be seen in Fig. 11.d and 11.e, the oil droplet size distribution of these emulsions (O40-1 and O50-1) changed from bimodal to monomodal after drying. This change in oil droplet size distribution for reconstituted emulsions can be explained by the reduction in the oil droplet diameter. This reduction of the size could be associated to the high pressure and conformation of nozzle in the process of atomization, thus promoting the breakdown of the droplets.
Park, Choi, & Lee (2019) analyzed colloidal characteristics of a bimodal distribution emulsion system using bulk rheological and numerical approaches, they observed that the viscosity of the mixture for bimodal systems (small drops mixed with large drops) was lower than those for the monodisperse system. The reduced viscosity was related to an increase in droplet deformability and a decrease in shear stress at the droplet surface. As mentioned earlier in this work, an inverse relationship between emulsion viscosity and oil droplet size was observed.
The droplet size of reconstituted emulsions (C20, C30, O40-1, O50-1 and O50-2) decreased due to the atomization process of applying 200 kPa of pressure, which may have broken the oil droplets (Consoli, Hubinger & Dragosavac, 2020).
The reconstituted C10 emulsion, which contains a high concentration of MD (28% w/w) and a low concentration of MS Capsul (4% w/w), showed a significant increase in relation to the initial emulsion. This result may be correlated with the low surfactant concentration, which led to the formation of larger droplets in the initial emulsion. Moreover, the coalescence of the oil droplets after spray-drying or reconstitution is facilitated by the low MS content at the surface (surfactant poor regime) (Polavarapu, Oliver, Ajlouni & Augustin, 2011).
The backscattering (BS) profile obtained from the turbidimetric measurements were performed to evaluate reconstituted emulsion stability. BS profiles versus height of the measuring cell were obtained from the freshly reconstituted sample after 2, 24 and 48 h (Fig. 12). BS values for emulsions C10 and O50-1 decreased slightly over time (Fig. 12.a and 12.e). This behavior can be attributed to a slight coalescence of oil droplets. However, the emulsions in general proved to be sufficiently stable during the analyzed time, due to little change in the BS values as a function of the height of the measuring cell.
The BS profiles obtained for the initial emulsions and from the reconstituted emulsions (C20, C30, O40-1 and O50-2) displayed tendencies for flocculation/coalescence phenomenon, but only at the initial stage, and resulted in products with similar backscattering profiles (BS); as shown in Fig. 4.a, 4.d-f, 12.b-d and 12.f, respectively. The C10 and O50-1 emulsions (Fig. 4.b, 4.g, 12.a and 12.e) showed different profiles when comparing between the initial and reconstituted emulsions. The initial C10 formulation emulsion had an increase in the level of BS at 24 and 48 h, while the reconstituted emulsion only decreased slightly over time. For the O50-1 formulation, there was a destabilization in the initial emulsion over a period of 24 h, followed by the formation of a cream layer (Fig. 4.b). In regards to the reconstituted emulsion, there was a slight decrease over time, with no phase separation. This behavior can be explained by the smaller oil droplets in the reconstituted emulsion, which positively impacts stability.