Influence of primary homogenization step on microfluidized emulsions formulated with thyme oil and Appyclean 6548

INFLUENCE OF PRIMARY HOMOGENIZATION STEP ON 1 MICROFLUIDIZED EMULSIONS FORMULATED WITH THYME 2 OIL AND APPYCLEAN 6548 3 Luis A. Trujillo-Cayadoa, M Carmen Alfaroa*, Jenifer Santosa, Nuria Caleroa, and José 4 Muñoza 5 a Departamento de Ingeniería Química, Facultad de Química, Universidad de Sevilla, C/ 6 P. García González, 1, E41012, Sevilla, Spain. 7 * Corresponding author. M Carmen Alfaro Rodríguez; Tel.: +34 954 557180; fax: +34 8 954 556447; E-mail address: [email protected] 9 ABSTRACT 10 This contribution deals with the development of emulsions formulated using thyme 11 essential oil and a new biomass-derived surfactant. In addition, this work extends our 12 knowledge concerning the factors that can influence stability and droplet size 13 distributions of microfluidized emulsions, such as the geometry of the rotor-stator used 14 and the homogenization rate in the primary homogenization. Stable thyme oil-in-water 15 emulsions (30 wt%) containing submicron droplets were formed. Interestingly, laser 16 diffraction results reveal that mean droplet sizes are mainly controlled by homogenization 17 rates and polydispersity by the rotor-stator geometry used in the first step of 18 homogenization. In addition, higher droplet sizes for pre-emulsions seem to be a key 19 factor in order to reduce both the degree of recoalescence and the size of the droplets 20 in the second homogenization step. Furthermore, higher droplet sizes in the pre- 21 emulsion favour higher physical stability of the final emulsions. Finally, this research 22 highlights the importance of controlling primary homogenization conditions for the 23 physical stability of microfluidized emulsions that contain natural ingredients. 24 Keywords: Droplet size; emulsion; Microfluidizer; physical stability; thyme oil 25 1. INTRODUCTION 26 There is a growing interest in the use of essential oils, like thyme oil, due to their 27 antimicrobial activity and biocompatible properties. These properties make them widely 28 used in fields such as food, pharmaceutical and cosmetics. In addition, these natural 29 resources have been recognized as GRAS (Generally Recognized As Safe) by U.S. 30 Food and Drug Administration [1], which makes them the most promising natural 31 antimicrobials. Thyme oil, obtained from Thymus vulgaris L, is divided into two different 32 classes: red and white. Red thyme oil is the product of the distillation of dried thyme 33 leaves while white is obtained from red thyme oil re-distillation [2]. The potential 34 application of white thyme oil as a food preservative to replace synthetic chemicals, 35 which are potentially toxic to humans, has been demonstrated [3]. Nevertheless, the 36 major disadvantage of utilizing essential oils is their high volatility and their tendency to 37 oxidise. Emulsion-based systems are a very attractive way to increase their stability by 38 reducing their volatility and conserving their biological characteristics [4]. Oil-in-water 39 emulsions are systems consisting of oil phase dispersed in aqueous phase, usually in 40 the form of droplets. These systems are important vehicles for the delivery of 41 hydrophobic bioactive compounds and have found a wide range of applications in many 42 industries including food, pharmacy, cosmetics and agrochemistry [5–7]. 43 Emulsions need an emulsifier since they are thermodynamically unstable. In recent 44 years, the use of green surfactants has been attracting attention. Appyclean 6548, a new 45 surfactant derived from wheat waste (alkyl poly pentoside), fulfils all the requirements to 46 be considered a green surfactant; namely it is derived from renewable resource and 47 manufactured by environmentally friendly processes. In fact, this emulsifier possesses 48 the ECOCERT certification. 49 In order to produce emulsions with specific physicochemical and functional properties, 50 controlling the droplet size distribution is required. Droplet size distributions are strongly 51 influenced by the emulsification method and conditions used. Emulsions can be 52 developed using low-energy and high-energy approaches. However, the latter are more 53 likely to be used in the food or cosmetic field since their scale-up is easier and the 54 equipment is more readily available. A multitude of homogenizers, such as rotor-stators, 55 ultrasounds, colloid mills or high-pressure valve homogenizers, can be used to prepare 56 these systems. Emulsions can be prepared in two steps: primary and secondary 57 homogenization. The aim of the primary homogenization is to create droplets of 58 dispersed phase such that a coarse emulsion is formed. The goal of the second step 59 (secondary homogenization) is to reduce the size of pre-existing droplets, which usually 60 involves the use of a different homogenizer. There are several studies concerning the 61 influence of homogenization rate and the device used on physical stability, rheology and 62 droplet size distribution for emulsion-based systems [8–13]. One homogenizer used in 63 the second step that has received a lot of attention recently is the Microfluidizer, due to 64 several results that suggest its use provides narrower distributions at smaller droplet 65 sizes [14,15]. Some research has been carried out in order to extend our knowledge 66 about the influence of the number of cycles and homogenization pressure in 67 Microfluidizers [9,16,17]. Furthermore, the comparison between using single- or dual- 68 channel microfluidizers and one or two interaction chambers have been reported [18– 69 20]. The main novelty of this research is the study of the influence of the pre-emulsion 70 properties on the droplet size distribution and physical stability of the final emulsions. For 71 this reason, the present work aims to provide an exhaustive study of the influence of the 72 primary homogenization on microfluidized emulsion properties. On top of that, a further 73 aim of this research was to obtain stable ecological emulsions of thyme oil formulated 74 with an emulsifier obtained from wheat. The results of this study could be useful to design 75 and develop functional oil-in-water emulsions. 76 2. MATERIALS AND METHODS 77 Materials 78 30 wt% oil-in-water emulsions with a surfactant (Appyclean 6548) concentration of 3 wt% 79 were prepared using thyme essential oil as dispersed phase. Thyme oil (Thymus 80 vulgaris) was purchased from Sigma Aldrich. The emulsifier (Appyclean 6548) used was 81 an alkyl poly pentoside provided by Wheatoleo. This non-ionic surfactant is solid and 82 immiscible in water at room temperature. Deionised water obtained from a water 83 purification system was used for the preparation of all samples. Sodium azide (0.1 wt%) 84 was added to the formulation in order to inhibit the growth of microorganisms. All of the 85 chemicals were used as received. 86 Preparation of oil-in-water emulsions 87 The surfactant was dispersed into the oil phase. 5.55 g of Appyclean 6548 was added to 88 55.5 g of Thyme essential oil for emulsion batches of 185 g. Then, the surfactant was 89 melted and dissolved in the essential oil at 70°C in a laboratory oven according to the 90 supplier’s instructions. The continuous phase was prepared by dissolving 0.185 g of 91 sodium azide in the corresponding amount of water. 92 Two different rotor-stator devices (Ultraturrax T50 with a S50NG45F dispersion unit and 93 Silverson L5M equipped with an emulsor mesh screen) were used for the primary 94 homogenization. Pre-emulsions were produced by adding the oil phase at 20 ºC to the 95 continuous phase (deionised water and sodium azide), also at 20 ºC, using a syringe 96 pump at a constant flow rate of 21.67 mL/min during 180 s, and then were homogenized 97 for an additional 30 s. The homogenization speed was fixed at 2000, 4000 or 6000 rpm 98 using Ultraturrax T50 or Silverson L5M, so three batches of six different pre-emulsions 99 were developed. The initial and final temperatures of the emulsion and the amperage 100 were measured during the primary emulsification process in order to determine the 101 power density. 102 The secondary homogenization was performed using a Microfluidizer M110P (interaction 103 chambers F12Y) at 2500 psi (172 bar) for one pass. These conditions were chosen in 104 order to highlight the importance and influence of the primary homogenization for the 105 development of oil-in-water emulsions by microfluidization. The outlet sample tube of the 106 Microfluidizer was cooled with water at 20 °C. The pH values for the continuous phases 107 and the final emulsions were 5.94 and 6.42, respectively 108 Droplet size distributions of emulsions 109 The droplet size distribution and mean droplet sizes were determined using a Malvern 110 Mastersizer X. Volumetric mean diameter (D4,3) was used to compare the droplet sizes 111 of different emulsions: 112 𝐷4,3 =∑𝑛𝑖𝑑𝑖 4𝑁 𝑖=1 ∑𝑛𝑖𝑑𝑖 3 𝑁 𝑖=1 Eq. (1) 113 where di is the droplet diameter, N is the total number of droplets and ni is the number of 114 droplets having a diameter di. Moreover, span was used to study the distribution width of 115 droplet sizes: 116 𝑠𝑝𝑎𝑛 = 𝐷(𝑣,0.9)−𝐷(𝑣,0.1) 𝐷(𝑣,0.5) Eq. (2) 117 where D(v,0.9), D(v,0.5), D(v,0.1) are diameters at 90%, 50% and 10%, respectively, of 118 cumulative volume. The absorption and refraction indexes used for the continuous 119 medium (water) were 0.1 and 1, respectively, whereas the refraction index for the 120 dispersed phase (thyme oil) was 1.50. Droplet size distributions were obtained using a 121 polydisperse analysis. The influence of aging time on droplet size distributions was 122 studied during 20 days after preparation to analyze and quantify coalescence/Ostwald 123 ripening effects. 124 Analysis of emulsion physical stability 125 The physical stability of emulsions whose pre-emulsions were obtained with different 126 emulsification methods was studied and quantified by means of multiple light scattering 127 measurements (Turbiscan Lab Expert) at 25 °C. The results are presented in the form of 128 curves which show intensities of backscattering in reference mode (delta-backscattering, 129 BS% = BSt % - BS0 %) as a function of time and height. The Turbiscan Stability Index 130 (TSI) has been used for the comparison and estimation of the emulsion stability. The 131 value of this parameter was calculated with the special computer program using the 132 equation: 133 𝑇𝑆𝐼 = ∑∑|𝑠𝑐𝑎𝑛𝑖−𝑠𝑐𝑎𝑛𝑖−1| ℎ 𝐻 𝑖 Eq. (3) 134 Where scani is the average backscattering for each time (i) of measurement, scani-1 is 135 the average backscattering for the (i-1) time of measurement and H is the number of 136 scans carried out on the sample. 137 Flow Curves 138 Flow curves for pre-emulsions and emulsions were obtained by means a controlled- 139 stress rheometer, CS Haake-MARS (Thermo), and a sandblasted Z20 coaxial cylinder 140 (Ri = 1 cm, Re/Ri = 1.085) at 24 hours of aging time and they were performed at 25 °C. 141 A step-wise protocol in the 0.1-20 Pa shear stress range was applied. All samples 142 showed Newtonian behaviour which was fitted to Newton’s law. 143 All measurements were done in duplicate and the values shown are the average of the 144 two replicates. 145 3. RESULTS AND DISCUSSION 146 Figure 1 illustrates the power density (Pv) as a function of the residence time for five 147 replicates of the pre-emulsion processed at 2000 rpm using a Silverson L5M rotor-stator 148 device. A decrease in power density values with homogenization time is clearly 149 observed. The energy density (Ev) or mechanical energy input per unit of dispersing 150 volume for all pre-emulsions was calculated using the power density values by means of 151 the following equation [21,22]: 152 𝐸v = ∫Pv(t)𝑑𝑡 = ∫𝑊(𝑁,𝑡) 𝑀𝑐 𝜌𝑐(𝑡)+𝑀0 𝜌0(𝑡) 𝑑𝑡 Eq. (4) 153 Where Ev is the energy density or energy consumption, Pv is the power density, W is the 154 electrical power consumption, MC and M0 are the masses of the continuous and 155 dispersed phases respectively and c and 0 are the densities of the continuous and 156 dispersed phases respectively, N is the homogenization rate and t is the homogenization 157 time. Both densities are time dependent due to the increase in temperature during 158 emulsification. 159 The values of Ev for all pre-emulsions are shown in table 1. It should be noted that, as 160 expected, there is an increment in the energy consumption with homogenization rate for 161 both rotor-stator devices. Interestingly, energy consumption values obtained for the pre- 162 emulsions developed using an Ultraturrax T50 are much higher than those obtained for 163 the samples processed with the Silverson L5M rotor–stator device. This fact is probably 164 due to the differences between the geometries of both rotor-stator devices [23]. The 165 resulting droplet size distributions (DSDs) for these pre-emulsions are shown in figure 2. 166 Higher homogenization rates in both rotor stator devices lead to the production of smaller 167 droplets (Table 1). Pre-emulsions developed at 2000 rpm in both rotor-stator devices 168 show bimodal DSDs. In contrast, monomodal distributions were observed for emulsions 169 processed above 2000 rpm in both homogenizers. This fact reveals the importance of 170 higher homogenization rates/energy input in order to reduce the second peak of the 171 distribution and obtain a monomodal DSD. It is important to note that the use of different 172 types of rotor-stator did not significantly affect volumetric diameter at the same 173 homogenization rate. Although the energy consumption values at the same 174 homogenization rate are different for Silverson L5M and Ultraturrax T50, similar droplet 175 mean diameters are detected. This result suggests the higher efficiency of Silverson 176 L5M. Even so, although the mean droplet sizes are similar, some slight differences in 177 DSD can be observed between the pre-emulsions with SL5M and UT50 at the same 178 homogenization rate. Thyme essential oil-in-water emulsions processed in Silverson 179 L5M tend to possess narrower DSD. In these terms, the existence of little holes in 180 Silverson L5M favours the existence of jets in the velocity profile contrary to UT50, as 181 previously reported by Ozcan-Tas et al., 2011[24]. It may be the reason why the 182 polydispersity of emulsions prepared with SIlverson L5M is lower than those developed 183 using UT50. 184 Figure 3 shows, by way of example, the variation in Backscattering in mode reference 185 (BS%) as a function of the height of the measuring cell with aging time for pre-emulsions 186 developed at 2000 rpm with Silverson L5M. All pre-emulsions follow the same 187 destabilization trend. A marked decrease in BS in the lower zone of the measuring cell 188 is observed during the first 24 hours. This reduction extends over the whole measuring 189 cell with aging time. This fact is a clear indication of a creaming process, such as 190 commonly occurs in dispersed systems such as emulsions [25,26]. This destabilization 191 mechanism is favoured by the low viscosity of these pre-emulsions (3.71–5.25 mPa·s) 192 and the high droplet sizes [27]. In order to improve the stability of these emulsions 193 formulated with thyme oil and Appyclean 6548, a secondary homogenization was carried 194 out. 195 The DSD of the final emulsions is shown in Figure 4. All microfluidized emulsions exhibit 196 two populations of droplets, which is probably due to an excess of energy input 197 (recoalescence). This fact occurs often in emulsions prepared in high pressure 198 homogenizers and Microfluidizers [17,28]. In order to analyse the recoalescence for 199 emulsions studied, the data of the second peak were fitted to a Gaussian model: 200 𝑦 = 𝑦0+ 𝐴𝑒−(𝑥− 𝑥𝑐)2 2𝑤2 Eq. (5) 201 where y0 is the offset, xc the centre of the peak, w the width of the peak and A the 202 amplitude. 203 One factor which should be noted is the low degree of recoalescence observed in 204 microfluidized emulsions whose pre-emulsions were prepared at the lowest 205 homogenization rate (2000 rpm). This fact can be deduced from the area of the second 206 peak (table 2), which is lower for emulsions whose pre-emulsions were prepared at 2000 207 rpm. In addition, microfluidized emulsions processed in UT50 showed a higher degree 208 of recoalescence than those obtained using SL5M, evidence of which can be seen in the 209 larger area or higher median value of the second peak (table 2). These facts support the 210 aforementioned hypothesis concerning over-processing. Moreover, all results suggest 211 that for thyme essential oil-in-water emulsions formulated with Appyclean 6548 the 212 homogenization rate in the primary homogenization and, therefore the droplet size of the 213 pre-emulsion, is a key factor for final emulsions. It seems that bigger droplets in the first 214 step of homogenization could favour the break-up of the droplets inside the Microfluidizer 215 chamber. Accordingly, a lower homogenization rate in pre-emulsion, implying higher 216 droplet sizes, is recommended in order not only to reduce recoalescence in the 217 Microfluidizer but also to obtain final emulsions with better properties. This fact is evident, 218 taking into account the volumetric diameter and span values for final emulsions (figure 219 5). As can be observed in figure 5, there is a noticeable increase in droplet size and 220 polydispersity with homogenization rate applied in the primary homogenization. Hence, 221 although higher droplet size in the first step of homogenization could seem a drawback, 222 results reveal that the break-up of droplets is favoured and recoalescence is less 223 probable. 224 To study the physical stability of emulsions, the backscattering profiles were analysed at 225 different storage times in reference mode allowing a better display of the destabilization 226 processes. Figure 6A and 6B illustrate the variation of delta-backscattering (BS%) 227 versus measuring cell height as a function of aging time for emulsion with pre-emulsion 228 processed using Ultraturrax T50 at 2000 and 6000 rpm. Firstly, it should be noted that 229 the microfluidized emulsions possess higher physical stability than pre-emulsions (see 230 figure 3), as expected [29]. Final emulsions showed a marked drop in backscattering (a 231 peak in BS curves) in the lower zone of the measuring cell (0-5 mm), which is typical of 232 migration of droplets; namely a destabilization mechanism by creaming occurs. All 233 emulsions studied underwent creaming at similar rates (0.048-0.052 mm/day) and 234 intensities, as can be observed in figures 6A and 6B. This instability is clearly related to 235 the low Newtonian viscosity they present (7.78-8.80 mPa·s) [27]. Additionally, there is 236 also a decrease in the BS intensity in the middle zone of the tube, indicative of an 237 increment in mean droplet sizes [6,26]. The decrease in BS intensity in the middle zone 238 is more marked for the emulsion processed at 6000 rpm (see figure 6B). This behaviour 239 may be related to flocculation, coalescence or Ostwald ripening. In order to discern 240 between these destabilization processes as well as quantify the destabilization 241 mechanism that provokes this variation, the evolution over time of droplet size 242 distributions, droplet mean diameters and span was analysed. 243 Figure 7 shows the droplet size distributions with aging time for the microfluidized 244 emulsion whose pre-emulsion was developed using Ultraturrax T50 at 2000 and 6000 245 rpm. This figure illustrates a shift of DSD towards higher droplet sizes and a reduction in 246 polydispersity for both emulsions. It is more clearly observed for 6000 rpm emulsion, 247 which is totally consistent with the results obtained from laser diffraction measurements. 248 This behaviour suggests that the increase in mean diameters observed with the multiple 249 light scattering technique is due to a destabilization mechanism by Ostwald ripening [30]. 250 This result is consistent with works reported by other authors and it is typically observed 251 in essential oil-in-water emulsions [31,32]. All emulsions presented an increase in 252 volumetric mean diameter and a reduction in span, which is shown in figure 8. The 253 variations of Volumetric diameter and span values from day 20 to day 1 were calculated 254 as follows: 255 ∆𝐷4,3 =𝐷4,3 𝑑𝑎𝑦 20− 𝐷4,3 𝑑𝑎𝑦 1 𝐷4,3 𝑑𝑎𝑦 1 Eq. (6) 256 257 ∆𝑠𝑝𝑎𝑛 = 𝑠𝑝𝑎𝑛 𝑑𝑎𝑦 20− 𝑠𝑝𝑎𝑛 𝑑𝑎𝑦 1 𝑠𝑝𝑎𝑛 𝑑𝑎𝑦 1 Eq. (7) 258 A clear influence of both the primary homogenization rate and the rotor-stator device 259 used for the development of the pre-emulsions on the evolution of both the volumetric 260 mean diameter and span of microfluidized emulsions can be observed. In this sense, 261 emulsions developed at higher homogenization rates presented the major variations in 262 both parameters and emulsions processed with SL5M showed smaller variations in 263 droplet sizes and span than their counterparts processed using UT50. These facts are 264 related to the higher polydispersity these emulsions exhibited at 24 hours (see figure 4), 265 i.e. more polydispersity favours destabilization by Ostwald ripening [33]. 266 Turbiscan Stability Index (TSI) values as a function of aging time for all thyme oil 267 microfluidized emulsions are shown in figure 9. This parameter, obtained in the 0-30 mm 268 zone, allows global physical stability (considering all destabilization mechanisms) to be 269 analysed for these dispersed systems. High values of TSI involve poor physical stability. 270 In this way, the influence of the homogenization rate used in the primary homogenization 271 on the final emulsion physical stability is demonstrated. Two trends can be clearly 272 observed, as expected. On the one hand, the use of a lower homogenization rate during 273 the primary homogenization allowed lower TSI values to be obtained. On the other hand, 274 microfluidized emulsion whose pre-emulsion was prepared in Silverson L5M showed 275 better physical stability than those whose pre-emulsion was developed using Ultraturrax 276 T50. These facts are related to the aforementioned differences in polydispersity, i.e. 277 higher polydispersity involves higher TSI values. Hence, the microfluidized emulsion 278 whose primary homogenization was in Silverson L5M at the lowest homogenization rate 279 presented the best global physical stability and the lowest droplet size and polydispersity. 280 This fact demonstrates the necessity for a pre-emulsion with a relatively large droplet 281 size in order to allow the Microfluidizer to play its role in forming finer emulsions. 282 CONCLUSIONS 283 Stable and concentrated thyme oil-in-water emulsions formulated with a biomass- 284 derived surfactant obtained from renewable resources have been developed. No 285 significant differences in droplet sizes for pre-emulsions prepared using Silverson L5M 286 or Ultraturrax T50 at the same homogenization rate were observed. However, the energy 287 consumption for Ultraturrax T50 was much higher than that of Silverson L5M, suggesting 288 the greater efficiency of the latter. Results obtained for these pre-emulsions indicate that 289 mean droplet sizes are mainly controlled by homogenization rates and polydispersity by 290 rotor-stator geometry. All pre-emulsions showed low physical stability involving a 291 creaming destabilization process due to their high mean droplet sizes, high polydispersity 292 and low viscosity. 293 Microfluidized emulsions showed submicron mean droplet sizes but a recoalescence 294 process was observed. This effect was favoured by higher homogenization rates and the 295 use of Ultraturrax T50 in the primary homogenization. Therefore, the droplet size 296 distribution of the pre-emulsions is a key factor that strongly influences the droplet size 297 distribution of the final emulsions. Higher mean droplet sizes for pre-emulsions could 298 favour the break-up of the droplets inside the high-pressure homogenizer chamber, 299 allowing lower droplet size in microfluidized emulsions to be obtained. Hence, the coarse 300 emulsion developed using Silverson L5M at 2000 rpm produced the microfluidized 301 emulsion with the narrowest droplet size distribution and the lowest mean droplet sizes. 302 All microfluidized emulsions presented higher physical stability than those developed 303 only with a rotor-stator device. However, a creaming process and an increase in droplet 304 size were observed in final emulsions. These variations in droplet sizes involved a 305 reduction in span with aging time, which suggests destabilization by Ostwald ripening. 306 0 5 10 15 20 -60 -50 -40 -30 -20 -10 0 Reference 72 h 288 h 360 h 456 h 504 h Backscattering (%) Height (mm) (B) 435 Figure 6. Backscattering versus container height as a function of aging time in reference 436 mode for final emulsions whose pre-emulsion was obtained with Ultraturrax T50 at (A) 437 2000 rpm and (B) 6000 rpm. Temperature: 25 ºC 438 0,1 1 10 0 2 4 6 8 10 12 14 UT50 - 2000rpm - Day 1 UT50 - 2000rpm - Day 11 UT50 - 2000rpm - Day 17 UT50 - 6000rpm - Day 1 UT50 - 6000rpm - Day 11 UT50 - 6000rpm - Day 17 Size (m) Volume (%) 439 Figure 7. Droplet size distributions as a function of aging time for final emulsions whose 440 pre-emulsion was developed using Ultraturrax T50 at 2000 rpm and 6000 rpm. 441 2000 4000 6000 -100 -75 -50 -25 0 25 50 75 100 D4,3 (%) span (%) Homogenization rate (rpm) D4,3 - UT50 D4,3 - SL5M span - UT50 span - SL5M 442 Figure 8. Variation of volumetric mean diameters and span values for final emulsions 443 between day 1 and day 20 as a function of the homogenization rate applied in the primary 444 homogenization. 445 446 447 0 5 10 15 20 25 30 0,0 0,5 1,0 1,5 2,0 2,5 3,0 3,5 4,0 4,5 5,0 Turbiscan Stability Index Aging time (days) UT50 - 2000 rpm UT50 - 4000 rpm UT50 - 6000 rpm SL5M - 2000 rpm SL5M - 4000 rpm SL5M - 6000 rpm 448 Figure 9. Turbiscan Stability Index of microfluidized emulsions as a function of aging 449 time. 450 451