- Email: [email protected]
Effect of preparation conditions on the properties of microspheres prepared using an emulsion-solvent extraction process
chemical engineering research and design 9 0 ( 2 0 1 2 ) 1517–1526
Contents lists available at SciVerse ScienceDirect
Chemical Engineering Research and Design journal homepage: www.elsevier.com/locate/cherd
Effect of preparation conditions on the properties of microspheres prepared using an emulsion-solvent extraction process Harri Heiskanen a,∗ , Peter Denifl b , Päivi Pitkänen b , Markku Hurme a a
Department of Biotechnology and Chemical Technology, Aalto University School of Science and Technology, P.O. Box 6100, FIN-02015 HUT, Finland b Borealis Polymers Oy, P.O. Box 330, FIN-06101 Porvoo, Finland
a b s t r a c t Methylaluminoxane microspheres were prepared using a hydrocarbon-in-perfluorocarbon-emulsion solvent extraction process. The effect of the preparation conditions on the size of the microspheres was investigated. As expected, the size of the microspheres decreased with increasing stirring speed. At low surfactant concentrations the size of the microspheres was independent of the surfactant concentration. However, the size of the microspheres decreased as the surfactant concentration was further increased. The size of the microspheres was not only affected by the surfactant concentration but also by the volume ratio of the dispersed phase to the continuous phase. At a low volume ratio of the phases the effect of the surfactant on the size of the microspheres was larger than the effect of the increased volume ratio of the phases. At high volume ratios of the phases the effect of the volume ratio of phases on the size of the microspheres became more significant than the effect of the surfactant. A slow solidification increased the formation of non-spherical microspheres. © 2012 The Institution of Chemical Engineers. Published by Elsevier B.V. All rights reserved. Keywords: Microsphere; Emulsification; Particle size distribution; Solvent extraction
1.
Introduction
Emulsion-solvent extraction/evaporation processes are widely used to produce microspheres for different applications such as the controlled release of drugs, cosmetics, inks, pigments and chemical reagents. Control of the microsphere size distribution and morphology of the microspheres is important in the microsphere preparation processes, for example for obtaining repeatable and controlled release behavior of a drug. Freitas et al. (2005) and Li et al. (2008) have written recent reviews about the microsphere preparation using emulsion-solvent extraction/evaporation processes. The equilibrium droplet size of the emulsion is the result of the droplet break-up and the coalescence. The droplets break up if the shear forces exceed the resistance forces. The maximum stable droplet size, dmax , in dilute dispersions can be related to the impeller Weber number, We, using the following
∗
equation (Hinze, 1955): dmax = c1 D
c N2 D3
−0.6 = c1 We−0.6
(1)
where c is the density of the continuous phase, N the stirring speed, D the diameter of the stirrer, the interfacial tension and c1 is a coefficient. The maximum droplet diameter correlates linearly with the Sauter mean diameter, d32 (Sprow, 1967). Therefore, Eq. (1) can be written as: d32 = c2 We−0.6 D
(2)
where c2 is a coefficient. Increasing the surfactant concentration reduces the interfacial tension between the phases until the critical micelle
Corresponding author. Fax: +358 9 451 2694. E-mail address: harri.heiskanen@aalto.fi (H. Heiskanen). Received 17 November 2010; Received in revised form 9 January 2012; Accepted 14 February 2012 0263-8762/$ – see front matter © 2012 The Institution of Chemical Engineers. Published by Elsevier B.V. All rights reserved. doi:10.1016/j.cherd.2012.02.008
1518
chemical engineering research and design 9 0 ( 2 0 1 2 ) 1517–1526
(Bahukudumbi et al., 2004; Mu et al., 2005): the size of the microspheres decreases as the Weber number increased in the case of aqueous emulsion systems. In our previous work (Heiskanen et al., 2010) with a non-aqueous emulsion system the Sauter mean diameter of the microspheres decreased as the Weber number was increased at high surfactant concentration. However, at low surfactant concentration the Sauter mean diameter of the microspheres increased as the Weber number was increased. The coalescence of the droplets increased more than the break-up as the mixing was enhanced, therefore the size of the microspheres increased. A surfactant concentration was found in which the Sauter mean diameter of the microspheres kept constant although the Weber number was varied.
Nomenclature ci dmax d32 dx
D ki N Re We
coefficient i maximum stable droplet size (m) Sauter mean diameter (m) diameter corresponding to x vol.% on a relative cumulative droplet diameter distribution curve (m) diameter of stirrer (m) coefficient i stirring speed (1/s) the Reynolds number, Re = DN2 /c the Weber number
Greek letters viscosity of continuous phase (Pa s) c density of continuous phase (kg/m3 ) , c interfacial tension (N/m) volume fraction of disperse phase ˚d
1.2. Effect of surfactant concentration on the microsphere size
concentration is reached. Further increase in the surfactant concentration does not alter the interfacial tension anymore remaining constant. In addition, coalescence can be reduced using a surfactant. As the droplet breaks up a new liquid–liquid interface is created, coalescence is prevented if the surfactant is absorbed to the new liquid–liquid interface fast enough. Eq. (1) depends only on the break-up of the droplets and does not take into account the coalescence of the droplets. Modified correlations between the Sauter mean diameter and the Weber number have been presented for aqueous systems which take into account the volume ratio of phases (Calderbank, 1958; Davies, 1992; Godfrey et al., 1989; Shinnar, 1961; Shinnar and Church, 1960). The correlation presented by Davies (1992) and Godfrey et al. (1989) is written: d32 = k1 (1 + k2 ˚d )We−0.6 D
(3)
where k1 and k2 are coefficients and ˚d is the volume fraction of the dispersed phase. The term (1 + k2 ˚d ) in Eq. (3) reflects the influence of the dispersed phase in reducing the overall level of turbulence in the system. The turbulent damping is thought to be the main reason for increased droplet size as the volume ratio of the phases increases (Leng and Calabrese, 2004). A higher volume ratio of the phases leads to growing collision rates of droplets which results in an increasing number of break-up and coalescence events. As a result of the increased volume ratio of the phases the likelihood of coalescence increases more than the break-up, resulting in an increased droplet size.
1.1.
Effect of stirring speed on the microsphere size
The effect of the stirring speed on the microsphere size has been widely studied using aqueous emulsion systems. In general, an increase in the stirring speed increases the Weber number (Eq. (1)), therefore higher shear forces are created and smaller microspheres are produced (Gabor et al., 1999; Jégat and Taverdet, 2000; Mateovic et al., 2002; Nepal et al., 2007). Gabor et al. (1999) reported that the microsphere size distribution narrowed as the stirring speed increased. Correlations between the Sauter mean diameter of microspheres and the Weber number have been presented
The surfactant affects both the emulsification and the solidification processes during the microsphere preparation. The surfactant decreases the interfacial tension which results in increased film drainage time between approaching droplets, therefore, the coalescence of the droplets is reduced. The surfactant also decreases the mass transfer rate through the interface of the phases (Lee, 2003). As the surfactant concentration increases the interfacial tension decreases and hence the size of the droplets decreases (Eqs. (1) and (2)). The mean size of the microspheres has been reported to decrease almost linearly with increasing surfactant concentration with aqueous emulsion systems (Mu et al., 2005; Rafati et al., 1997). Nepal et al. (2007) also noticed almost linear decrease in the mean size of the microspheres as the surfactant concentration increased from 0.2% to 0.8% (w/v). However, the mean size of the microspheres remained constant as the surfactant concentration was further increased to 1.2% (w/v). The surfactant concentration has been reported to affect the morphology of microspheres and the drug distribution within the microspheres (Yang et al., 2001).
1.3. Effect of volume ratio of phases on the microsphere size According to Eq. (3) the mean droplet size increases as the volume ratio of the phases increases. However, such a clear correlation between the volume ratio of the phases and the size of the microspheres has not been confirmed via experimental data. Gabor et al. (1999) reported that the size of the microspheres decreased when the volume ratio of the phases was increased (dichloromethane-in-water emulsion). A further increase in the volume ratio of the phases increased the size of the microspheres, an explanation for the observed trend was not offered by the authors. Jeffery et al. (1993) reported that the size of the microspheres increased as the volume of the external aqueous phase was increased. The increase in the average microsphere size was attributed to a reduction in agitation due to a decrease in mixing efficiency associated with larger volumes. Ellis and Jacqueir (2009) reported that the size of the microspheres increased (carrageenan-in-oil emulsion) when the volume ratio of the phases was increased. A further increase in the volume ratio of the phases did not have a significant effect on the size of the microspheres. It was assumed that the emulsion became more viscous and hence larger microspheres were produced as the dispersed phase volume increased.
chemical engineering research and design 9 0 ( 2 0 1 2 ) 1517–1526
2.
Aim
Emulsions used in microsphere preparation processes are usually oil-in-water or water-in-oil emulsions but in this study a hydrocarbon-in-perfluorocarbon emulsion was used. The methylaluminoxane droplets were solidified using a solvent extraction method; perfluorocarbon was added to the emulsion to extract the solvent (toluene) from the dispersed phase into the continuous phase. Our previous results (Heiskanen et al., 2010, 2012) showed that the emulsification conditions had a significant effect on the methylaluminoxane microsphere properties. In those experiments a different solidification method was used: the emulsion was added to an excess of perfluorocarbon. The following experiments were performed to investigate how the different microsphere preparation method affects the methylaluminoxane microsphere size and size distribution. The effect of the surfactant concentration, the stirring speed, the volume ratio of the phases, and the solidification rate on the methylaluminoxane microsphere size and the size distribution were all investigated as variables.
3.
Materials and methods
3.1.
Materials
Hexadecafluoro-1,3-dimethylcyclohexane, PFC, (CAS 335-273) was purchased from Fluorochem. 30 wt.% methylaluminoxane in toluene (methylaluminoxane CAS 120144-90-3; toluene CAS 108-88-3) was purchased from Albemarle. 3-Perfluorooctyl-1,2-propenoxide, PFPO, (CAS 38565-53-6; precursor of surfactant) was purchased from Fluorochem. PFC, toluene and PFPO were deoxygenized with argon prior to use. White mineral oil (CAS number: 8042-47-5) was used to prepare a sample for the particle size analysis.
3.2.
Microsphere size analysis and microscopy
Droplets produced in the emulsification step shrink during the solidification because the solvent (toluene) is extracted from the droplet into the continuous phase. Hence, the droplet size is larger than the microsphere size. The microspheres produced were mainly compact but some hollow microspheres were also produced. The compactness of the microspheres was not taken into account when the mean size of the microspheres was analyzed. Microsphere size analysis was performed by dispersing dry MAO microsphere powder (∼20 mg) into white mineral oil. Pictures were taken from this dilute suspension using a light microscope, these images were then analyzed using the CantyVisionClientTM software. Only spherical particles with a diameter of at least 1 m were taken into account for microsphere size distribution. The number of non-spherical microspheres was low and therefore could be discounted when the microsphere size distribution was calculated. At least 3000 spherical microspheres per batch were included during microsphere size distribution analysis. The span was calculated to describe the broadness of the microsphere size distribution:
span =
d90 − d10 d50
(4)
1519
where dx is the diameter of the particle corresponding to x vol.% on a relative cumulative droplet diameter distribution curve. A small span indicates a narrow size distribution. A scanning electron microscope (FEI Quanta 200) was used to determine the microsphere shape and morphology. Microspheres were mounted on metal stubs with double-sided conductive adhesive tape and coated with a thin layer of gold. Coating was performed under inert atmosphere. SEM imaging was performed using an electron beam energy of 3 keV. Some of the microspheres were cut before the SEM images were taken to assess the morphology and compactness of particles.
3.3.
Microsphere preparation
The experiments were performed to study the effect of the microsphere preparation conditions on the methylaluminoxane microsphere morphology, size and size distribution. All experiments were performed under inert conditions (Argon).
3.3.1.
Emulsification step
The emulsification was performed at room temperature (22 ± 1 ◦ C). The hydrocarbon-in-fluorocarbon-emulsion was prepared in a stirred vessel (made from glass; 4 baffles). The diameter of the anchor type stirrer (made from Teflon) was 3.9 cm. The conditions used in the experiments are summarized in Table 1. In the first step 43 ml PFC (the continuous phase) was added into the emulsification vessel. A small amount of toluene was added to the PFC in order to saturate the continuous phase with toluene (while remaining a one liquid phase system) before the emulsification, no solvent (toluene) extraction occurred during the emulsification. The stirring speed was varied from 450 to 600 rpm (Table 1). 2 ml methylaluminoxane (MAO) in toluene, the disperse phase solution, was injected into the emulsification vessel. After addition of the disperse phase solution (∼3 min), the surfactant solution (PFPO in PFC) was added to the emulsion drop wise over a period of approximately 1 min. The surfactant concentration in the continuous phase was varied from 0.01% to 0.15% (v/v) (Table 1). The emulsion was stirred for 15 min.
3.3.1.1. Effect of volume ratio of phases on the microsphere preparation. The volume of the continuous phase was kept constant, 43 ml, while the effect of volume ratio of the phases on the microsphere size and the size distribution was studied. The emulsion was made following the procedure described above with the volume of the dispersed phase varying from 1 to 3 ml. A smaller dispersed phase volume creates fewer droplets, therefore, less surface area is created requiring stabilization during the emulsification if the droplet size distribution remains constant. The surfactant concentration in the continuous phase was varied so that the volume ratio of the surfactant to the dispersed phase was constant.
3.3.2.
Solidification step
The stirring speed during the solidification was the same as it was during the emulsification. PFC (100 ml) was added to the emulsion in order to extract the solvent (toluene) from the dispersed phase into the continuous phase. The flow rate of the PFC added to the emulsion was varied from 15 to 330 ml/min (Table 1). After the PFC addition, mixing was continued for 10 min.
1520
chemical engineering research and design 9 0 ( 2 0 1 2 ) 1517–1526
Table 1 – Summary of the microsphere preparation conditions. Stirring speed (rpm)
Stirring speed
Surfactant concentration
Volume ratio
Solidification rate
3.3.3.
RE09 RE10 RE03 RE08 RE11 RE04 RE05 RE06 RE07 RE03 RE08 RE12 RE13 RE12 RE14 RE03 RE08 RE01 RE02 RE15 RE28 RE29
450 500 550 550 600 550 550 550 550 550 550 550 550 550 550 550 550 550 550 550 550 550
Volume ratio (disp./cont. phase) 0.047 0.047 0.047 0.047 0.047 0.047 0.047 0.047 0.047 0.047 0.047 0.047 0.023 0.047 0.070 0.047 0.047 0.047 0.047 0.047 0.047 0.047
Product isolation
After 30 min without agitation, the microspheres had settled to the top of the PFC. The PFC was siphoned from the vessel and the remaining wet microspheres were dried under argon flow at 50 ◦ C for 1 h. The microsphere yield varied from 0.13 to 0.40 g. The residual toluene content of the dried microspheres was not measured.
4.
Results and discussion
Methylaluminoxane microspheres were successfully produced using the hydrocarbon-in-perfluorocarbon-emulsion solvent extraction process under all tested conditions. The variation between duplicate experiments was small: the average d32 of the experiments RE03 and RE08 was 21.1 ± 1.5 m and the average d32 of the experiments RE02 and RE15 was 22.8 ± 0.7 m (Table 1). The small variation of the replicates indicates that the experiments are repeatable even though the sample for the microsphere size measurement was small and was taken from dry powder. Dry microspheres can segregate by size and therefore taking a representative sample from the dry powder can be difficult.
4.1.
Surfactant in PFC ((v/v)%) 0.09 0.09 0.09 0.09 0.09 0.009 0.019 0.027 0.051 0.09 0.09 0.10 0.05 0.10 0.15 0.09 0.09 0.09 0.09 0.09 0.09 0.09
PFC flow rate (ml/min) 320 320 330 320 320 320 320 320 320 330 320 200 100 200 290 330 320 210 80 80 30 15
Theoretical solidification time (s) 8 8 8 8 8 8 8 8 8 8 8 13 13 13 13 8 8 12 31 31 85 170
d32 (m) 40.5 23.2 22.2 20.0 19.1 35.3 38.5 36.0 38.5 22.2 20.0 19.7 37.8 19.7 25.1 22.2 20.0 21.6 22.3 23.3 24.7 24.7
Span
0.59 0.61 0.59 0.50 0.61 0.61 0.55 0.58 0.68 0.59 0.50 0.63 1.07 0.63 0.56 0.59 0.50 0.52 0.68 0.61 0.84 0.60
the microspheres: from 23.2 to 19.1 m. The microsphere size distribution narrowed significantly as the stirring speed was increased (Fig. 1). A similar trend in the microsphere size distribution was observed by Gabor et al. (1999). The Sauter mean diameter of the microspheres decreased exponentially from 40.5 to 19.1 m with increasing Weber number, from 2400 to 4300 (Fig. 2). The shear forces breaking up droplets increased as the Weber number increased. Therefore, smaller microspheres were produced. In the Weber number calculations (Eq. (2)) the interfacial tension between the emulsion phases was assumed to be 0.0025 N/m which has been measured for a toluene-PFC system (Heiskanen et al., 2010). The addition of a surfactant is known to reduce the interfacial tension, this would indicate that the interfacial tension value used in the calculations was probably too high. The interfacial tension changes the constant in the correlation between the Sauter mean diameter and the Weber number (Eq. (2)) but does not affect the exponent value in the correlation. The theoretical exponent for a non-coalescencing and fully turbulent system is −0.6 (Eq. (2)). These experiments were
Effect of stirring speed on the microsphere size
As expected, the Sauter mean diameter of the microspheres decreased from 40.5 m to 23.2 m as the stirring speed increased from 450 rpm to 500 rpm. The bulk mixing at the slowest stirring speed was not intense enough to disperse the solidifying droplets in the bulk emulsion during the microsphere preparation leading to large droplets. The stability of the studied hydrocarbon-in-fluorocarbon emulsion was poor, partly due to the high density difference between the phases. The bulk mixing was enhanced as the stirring speed was increased from 450 rpm to 500 rpm, reducing the coalescence of the droplets and producing significantly smaller microspheres. Further increase in the stirring speed from 500 rpm to 600 rpm led to a smaller decrease in the size of
Fig. 1 – Particle size distribution (volume) of the spherical microspheres prepared with different stirring speeds (450–600 rpm).
chemical engineering research and design 9 0 ( 2 0 1 2 ) 1517–1526
done at fully turbulent conditions (the Reynolds number varied from 10,000 to 13,300). The exponent value for this system was −0.53 when the slowest stirring speed was not considered (Fig. 2), under the assumption that the bulk mixing was not intense enough to maintain complete dispersion during the solidification. One reason for the deviation from the theoretical value was that the emulsion studied was a coalescencing system. The exponent value was −1.26 (Fig. 2), when all data points were included, which differs significantly from the theoretical exponent value. In our previous study (Heiskanen et al., 2010) with a different test configuration, the exponent value was −0.56 at high surfactant concentration. Possible explanations for the deviation from the theoretical value are that some of the experiments were performed in transient flow and that the emulsion studied was a coalescencing system. Bahukudumbi et al. (2004) reported the same exponent value, −0.53, however, they did not offer any explanation for the deviation from the theoretical value. An exponent value, −0.72, in the correlation between the Sauter mean diameter of the microspheres and the Weber number was presented by Mu et al. (2005), a transient flow was suggested as the reason for the deviation from the theoretical exponent value.
4.2. Effect of surfactant concentration on the microsphere size Increased surfactant concentration is known to reduce the interfacial tension, the coalescence and also the mass transfer. Hence the surfactant concentration affects both the emulsification and the solidification steps during the microsphere preparation process. The Sauter mean diameter of the microspheres did not change when the surfactant concentration was increased from 0.01% to 0.05% (v/v) (Table 1; Fig. 3). The surfactant concentration at that range was not high enough to stabilize the new liquid–liquid interfaces formed during droplet break-up. Therefore, the surfactant addition did not reduce the coalescence of the droplets and the Sauter mean diameter remained almost constant. A further increase in the surfactant concentration from 0.05% to 0.09% (v/v) significantly decreased the Sauter mean diameter from ∼39 m to ∼21 m. The Sauter mean diameter of the microspheres decreased almost inversely as a function of the surfactant
Fig. 2 – The Sauter mean diameter of the spherical microspheres as a function of the Weber number. Two curve fittings: (I) all data points included and (II) the data point achieved using the slowest stirring speed excluded.
1521
Fig. 3 – Particle size distribution (volume) of the spherical microspheres prepared with different surfactant concentrations. The solidification time was ∼8 s in all other experiments but 13 s at the highest surfactant concentration (0.10% (v/v) PFPO in PFC).
concentration in the concentration range of 0.05–0.10% (v/v) (Fig. 4). In several studies a similar linear decrease in the size of the microspheres with increasing surfactant concentration has been observed (Mu et al., 2005; Nepal et al., 2007; Rafati et al., 1997). The surfactant is known to reduce the interfacial tension between the phases (not measured for this system), additionally, the Sauter mean diameter depends inversely on the interfacial tension (Eqs. (1) and (2)). The film drainage rate of the approaching droplets increases with decreasing interfacial tension, leading to a reduction in the coalescence of the droplets and the size of the microspheres decreases. The Sauter mean diameter of the microspheres did not reduce significantly as the surfactant concentration was increased from 0.09% to 0.10% (v/v) indicating that the surfactant concentration was close to the critical micelle concentration. The interfacial tension does not change after the critical micelle concentration is reached although the surfactant concentration is further increased.
Fig. 4 – The Sauter mean diameter of the spherical microspheres as the function of surfactant concentration in the continuous phase. Experiments were performed with different volume ratios of the dispersed to the continuous phase: 0.023, 0.047 and 0.070. The solidification time was varied from ∼8 s to 13 s.
1522
chemical engineering research and design 9 0 ( 2 0 1 2 ) 1517–1526
4.3. Effect of volume ratio of the phases on the microsphere size In theory, the higher the volume ratio of the phases, the larger the interface area created during the emulsification. As such, more surfactant is needed to prevent the coalescence of the droplets. The surfactant concentration in the continuous phase was varied so that the volume ratio of the surfactant to the dispersed phase was constant when the effect of the volume ratio of the phases on the microsphere was studied. The Sauter mean diameter of the microspheres remained almost the same, ∼38 m, because the surfactant concentration was roughly 0.05% (v/v) although the volume ratio of the phases varied from 0.023 to 0.047 (Table 1; Fig. 4). In conclusion, the size of the microspheres was not affected by the volume ratio of the phases at surfactant concentration of 0.05% (v/v). The Sauter mean diameter of the microspheres decreased from 37.8 to 19.7 m when the volume ratio of the phases increased from 0.023 to 0.047 (Table 1; Fig. 4). The surfactant concentration was 0.05% (v/v) at a volume ratio of 0.023% and 0.10% (v/v) at a volume ratio of 0.047. The increased surfactant concentration reduced the interfacial tension and hence smaller microspheres were produced even though the volume ratio of the phases increased. Therefore, at a low volume ratio of the phases (less than 0.047) the size of the microspheres was primarily affected by the surfactant concentration as opposed to the variation in the volume ratio of the phases. The Sauter mean diameter of the microspheres increased from 19.7 to 25.1 m when the volume ratio of the phases was further increased from 0.047 to 0.070 (Table 1; Fig. 4). The surfactant concentration was 0.10% (v/v) at a volume ratio of 0.047% and 0.15% (v/v) at a volume ratio of 0.070. Despite the increased surfactant concentration the size of the microspheres increased as the volume ratio of the phases increased. The effect of the volume ratio of the phases on the size of the microspheres became more significant than the effect of the surfactant at high volume ratio of the phases. According to Eq. (3) the size of the microspheres is expected to increase with the increasing volume ratio of the phases. The main explanation for the increased microsphere size is the turbulence dampening caused by the increased volume ratio of the phases. Additionally, the number of collisions should increase as the volume ratio of the phases increases, thus increasing the size of the microspheres. Gabor et al. (1999) and Heiskanen et al. (2010) noticed a similar trend in the size of the microspheres as the volume ratio of the phases was varied: the size of the microspheres first decreased as the volume ratio of the phases was increased. A further increase in the volume ratio of the phases led to an increase in the size of the microspheres. In our previous experiments (Heiskanen et al., 2010) a different solidification method was used: the hydrocarbon-in-perfluorocarbon emulsion was fed into an excess of perfluorocarbon. No correlation between the Sauter mean diameter of microspheres and the surfactant concentration was seen when the volume ratio of the phases was varied as conflicting with the findings of this study. However, the Sauter mean diameter of the microspheres decreased exponentially as a function of the volume ratio of the surfactant to the dispersed phase although a different volume ratio of the phases was used. The results indicated that the emulsion droplet size distribution changed when the emulsion was fed into the excess of perfluorocarbon, and the size of the microspheres was mainly affected by the volume
Fig. 5 – Particle size distribution (volume) of the spherical microspheres prepared with different volume ratios of the dispersed to the continuous phase. The solidification time was ∼13 s in all experiments. The volume ratio of surfactant to the dispersed phase was kept constant. ratio of the surfactant to the dispersed phase and not by the volume ratio of the phases. The microsphere size distribution was broadest at the lowest volume ratio of the phases due to the low surfactant concentration (Fig. 5; Table 1). The microsphere size distribution narrowed despite the volume ratio of the phases was increasing because the surfactant concentration increased (Fig. 5; Table 1). The volume ratio of the phases had an effect on the microsphere morphology according to the SEM images (Fig. 6). The cracks seen in the SEM images were formed during the SEM-sample preparation and not during the microsphere preparation. The majority of microspheres were produced with all tested volume ratios of the phases were compact spheres. However, the surfaces of the microspheres produced with a high volume ratio of phases (Fig. 6C) were not as smooth as those produced with the low volume ratios of the phases (Fig. 6A and B). The formation of the non-smooth surfaces is probably linked to the higher toluene volume and/or the surfactant concentration. The addition rate of the PFC to the emulsion was set so that the time needed to extract all toluene from the dispersed phase into the continuous phase was constant in all volume ratio experiments. Therefore, the toluene extraction rate (g toluene/s) was the highest at high volume ratios of the phases which could explain the observed change in the surface morphology of the microspheres. The solidification rate had no significant effect on the surface morphology of the microspheres when the volume ratio of the phases was lower, 0.047. Another possible reason for the formation of a non-smooth surface could be the high surfactant concentration (0.15% (v/v)) at the highest volume ratio of the phases. The surfactant concentration had no significant effect on the microsphere morphology when the surfactant concentration was less than 0.10% (v/v) at a volume ratio of phases 0.047. The surface viscosity of the droplets is known to increase with increasing surfactant concentration especially after the critical micelle concentration (Leng and Calabrese, 2004). With the increased surface viscosity, the deformed droplet (deformation caused by shear forces) did not return back to the thermodynamically more favorable spherical shape. Therefore, the microspheres had rough surfaces at high surfactant concentration with a high volume ratio of the phases. A high surfactant concentration is also
chemical engineering research and design 9 0 ( 2 0 1 2 ) 1517–1526
1523
Fig. 6 – SEM-images from microspheres prepared with different volume ratios of the dispersed to the continuous phase: (A) 0.023, (B) 0.047, and (C) 0.070. The solidification time was ∼13 s in all experiments. known to reduce the mass transfer so the droplet viscosity increases slowly and the droplet is exposed to shear forces for longer. Mateovic-Rojnik et al. (2005) noticed that the particle surface roughness increased with decreasing solidification rate.
4.4.
Effect of solidification rate on the microsphere size
During the experiments an excess of PFC was added to the emulsion in order to solidify the droplets. The increased continuous phase volume affected both the surfactant concentration in the continuous phase and the volume ratio of the phases. The surfactant concentration was roughly half of the initial value when all the toluene was extracted from the dispersed phase into the continuous phase. As the surfactant concentration decreases the interfacial tension increases and hence the droplet size increases in theory (see Eqs. (1) and (2)). The number of collisions should reduce as the volume ratio of the phases decreases, the coalescence reduces and the droplet size decreases. In the experiments the volume ratio of the phases decreased from 0.047 to 0.023 during the solvent extraction. The solubility of toluene in PFC is 2.2 wt.% at 22 ◦ C. The time needed to add the volume of PFC required to extract all the toluene from the dispersed phase (solidification time) varied from 8 to 170 s. The Sauter mean diameter of the microspheres increased slightly as the solidification rate decreased (Table 1). More non-spherical microspheres were produced as the solidification rate decreased, as shown in the light microscopy
pictures (Fig. 7A–E). One explanation for the formation of nonspherical microspheres at slow solidification rate was that the solidifying droplets were stretched by the shear forces. The stretched droplets did not return to the thermodynamically more favorable spherical shape due to the high viscosity of the solidifying droplet. Some microspheres had also collided during the solidification process forming agglomerates. A narrow microsphere size distribution was produced in all tested conditions (Fig. 8). The size distribution of the spherical microspheres did not change significantly although the solidification rate was varied. Some deposition of the methylaluminoxane, MAO, on the wall of the preparation vessel was noticed, especially with the two slowest solidification rates. At the height of the deposition (=white ring) the total liquid volume was roughly 80 ml. At that point MAO concentration in the dispersed phase had increased from 30 wt.% to roughly 67 wt.% as a result of the solvent (toluene) extraction. Not all of the solidifying droplets were properly dispersed in the bulk liquid during the microsphere preparation, leading to some of the solidifying droplets separating and accumulating on the wall forming the solid white deposition. Samples (∼1 ml) were taken via needle and syringe from the emulsion during the solidification at three different solidification degrees: ∼91, ∼97 and ∼103%. At the solidification degree 100% all toluene was assumed to be extracted from the dispersed phase into the continuous phase. The time needed to add enough PFC in order to extract all toluene (the solidification time) was 170 s. The samples were injected into
1524
chemical engineering research and design 9 0 ( 2 0 1 2 ) 1517–1526
Fig. 7 – Light microscopy images from the microspheres prepared with different solidification rates; the time needed to extract all toluene from the solidifying droplets: (A) 8 s, (B) 12 s, (C) 31 s, (D) 85 s, and (E) 170 s. Same magnification in all pictures. 10 ml septa vials to investigate whether solid microspheres had formed. The microspheres in all samples seemed to be solid immediately after the sample taking. After 10 min solid microspheres could only be observed in the sample with the highest solidification degree, ∼103%. Some microspheres were seen in the sample with solidification degree ∼97%. The microspheres in the sample with the lowest solidification degree had become liquid. The results indicated that the microspheres have a solid crust when the solidification degree is roughly ∼91% but still contain too much solvent inside the microsphere to remain solid.
5.
Fig. 8 – Particle size distribution (volume) of the spherical microspheres prepared with different solidification rates (the time needed to extract all toluene from the solidifying droplets varied from 8 to 170 s).
Conclusions
The effect of the solidification time, the surfactant concentration, the stirring speed, and the volume ratio of the phases on the methylaluminoxane microsphere size and the size distribution was successfully studied using a hydrocarbonin-fluorocarbon emulsion solvent extraction based microsphere preparation process. The methylaluminoxane droplets
chemical engineering research and design 9 0 ( 2 0 1 2 ) 1517–1526
formed in the first step of the process were solidified by adding an excess of perfluorocarbon to the emulsion. In agreement with the literature, the size of the microspheres was found to decrease with increasing stirring speed. The Sauter mean diameter decreased exponentially from ∼41 to ∼19 m with increasing Weber number (2400–4300). The exponent value in the correlation between the Sauter mean diameter and the Weber number for this system was −0.53 which is close to the theoretical value of −0.6. The experiment performed at the slowest stirring speed was removed from the correlation because the bulk mixing was assumed to be inadequate during that experiment. The microsphere size distribution narrowed significantly as the stirring speed was increased. The size of the microspheres decreased as the surfactant concentration was increased, as expected. Low surfactant concentrations did not stabilize the emulsion enough, meaning the coalescence of the droplets did not reduce and the Sauter mean diameter of the microspheres did not change. However, a further increase in the surfactant concentration significantly decreased the Sauter mean diameter of the microspheres as the interfacial tension and coalescence of the droplets decreased. The Sauter mean diameter of the microspheres decreased inversely as a function of the surfactant concentration, in the concentration range of 0.05–0.10% (v/v). According to the correlation between the Sauter mean diameter and the Weber number, the Sauter mean diameter depends inversely on the interfacial tension which is known to decrease as the surfactant concentration increases. Therefore, the surfactant addition decreased the interfacial tension which resulted in the formation of smaller microspheres. The Sauter mean diameter of the microspheres was not only affected by the surfactant concentration but also by the volume ratio of the phases. At low volume ratios of the phases (from 0.023 to 0.047) the size of the microspheres was found to be primarily affected by the surfactant concentration and not by the volume ratio of the phases. The increased surfactant concentration reduced the interfacial tension and the coalescence of droplets leading to smaller microspheres, although the volume ratio of the phases increased from 0.023 to 0.047. As the volume ratio of the phases was further increased from 0.047 to 0.070, the Sauter mean diameter of the microspheres increased from ∼20 to ∼25 m, even though the surfactant concentration was higher at the volume ratio of 0.070. Therefore, at high volume ratios of the phases the effect of the volume ratio of phases on the size of the microspheres became more significant than the effect of the surfactant concentration. One reason for the increased microspheres size, despite the surfactant concentration increasing, was the dampening of the turbulence due to an increased volume ratio of the phases. Also the number of collisions may be increased as the volume ratio of the phases increased, leading to increased coalescence and larger microspheres. A slow solidification rate increased the formation of nonspherical particles because the solidifying droplets were stretched by the shear forces. At the end of the solidification step the stretched droplets did not return to the thermodynamically more favorable spherical shape because of the high viscosity of the solidifying droplet leading to the formation of non-spherical particles. Some particles also collided during the solidification process forming agglomerates. Deposition of the matrix material on the wall of the preparation vessel was seen at the slowest solidification rates tested, due to solidifying droplets not being properly dispersed in the bulk liquid.
1525
Therefore part of the solidifying droplets separated depositing on the wall.
Acknowledgment Financial support from Borealis Polymers Oy is gratefully acknowledged.
References Bahukudumbi, P., Carson, K.H., Rice-Ficht, A.C., Andrews, M.J., 2004. On the diameter and size distributions of Bovine Serum Albumin (BSA)-based microspheres. J. Microencapsulation 21, 787–803. Calderbank, P.H., 1958. Physical rate processes in industrial fermentation, part I: the interfacial area in gas–liquid contacting with mechanical agitation. Trans. Inst. Chem. Engrs. 36, 443–463. Davies, G.A., 1992. Mixing and coalescence phenomena in liquid–liquid systems. In: Thornton, J.D. (Ed.), Science and Practice of Liquid–Liquid Extraction, vol. I. Clarendon Press, Oxford, UK, p. 245. Ellis, A., Jacqueir, J.C., 2009. Manufacture of food grade -carrageenan microspheres. J. Food Eng. 94, 316–320. Freitas, S., Merkle, H.P., Gander, B., 2005. Microencapsulation by solvent extraction/evaporation: reviewing the state of the art of microsphere preparation process technology. J. Controlled Release 102, 313–332. Gabor, F., Ertl, B., Wirth, M., Mallinger, R., 1999. Ketoprofen-poly(d,l-lactic-co-glycolic acid) microspheres: influence of manufacturing parameters and type of polymer on the release characteristics. J. Microencapsulation 16, 1–12. Godfrey, J.C., Obi, F.I., Reeve, R.N., 1989. Measuring drop size in continuous liquid–liquid mixers. Chem. Eng. Prog. 85, 61–69. Heiskanen, H., Denifl, P., Hurme, M., Pitkänen, P., 2010. Effect of emulsification conditions on the properties of the microspheres prepared by a solvent extraction process. Chem. Eng. Technol. 33, 1635–1644. Heiskanen, H., Denifl, P., Hurme, M., Pitkänen, P., Oksman, M., 2012. Effect of physical properties and emulsification conditions on the microsphere size prepared using a solvent extraction process. J. Disper. Sci. Technol. 33, 1–11. Hinze, J.O., 1955. Fundamentals of the hydrodynamics mechanisms of splitting in dispersion process. AIChE J. 1, 289–295. Jeffery, H., Davis, S.S., O’Hogan, D.T., 1993. The preparation and characterization of poly(lactide-co-glycolide) microparticles. II. The entrapment of a model protein using a (water-in-oil)-in-water emulsion solvent evaporation technique. Pharm. Res. 10, 362–368. Jégat, C., Taverdet, J.L., 2000. Stirring speed influence study on the microencapsulation process and on the drug release from microcapsules. Polym. Bull. 44, 345–351. Lee, Y.-L., 2003. Surfactant effects on mass transfer during the drop-formation and drop falling stages. AIChE J. 49, 1859–1869. Leng, D.E., Calabrese, R.V., 2004. Immiscible liquid–liquid systems. In: Paul, E.L., Atiemo-Obeng, V.A., Kresta, S.M. (Eds.), Handbook of Industrial Mixing: Science and Practice. John Wiley & Sons Inc., NJ, pp. 639–753 (Chapter 12). Li, M., Rouad, O., Poncelet, D., 2008. Review – microencapsulation by solvent evaporation: state of the art for process engineering approaches. Int. J. Pharm. 363, 26–39. Mateovic, T., Kriznar, B., Bogataj, M., Mrhar, A., 2002. The influence of stirring rate on biopharmaceutical properties of Eudragit RS microspheres. J. Microencapsulation 19, 29–36. Mateovic-Rojnik, T., Frlan, R., Bogataj, M., Bukovec, P., Mrhar, A., 2005. Effect of preparation temperature in solvent evaporation Eudragit RS microsphere properties. Chem. Pharm. Bull. 53, 143–146. Mu, Y., Lyddiat, A., Pacek, A.W., 2005. Manufacture by water/oil emulsification of porous agarose beads: effect of processing
1526
chemical engineering research and design 9 0 ( 2 0 1 2 ) 1517–1526
conditions on mean particle size, size distribution and mechanical properties. Chem. Eng. Process. 44, 1157–1166. Nepal, P.R., Chun, M-K., Choi, H-K., 2007. Preparation of floating microspheres for fish farming. Int. J. Pharm. 341, 85–90. Rafati, H., Coombes, A.G.A., Adler, J., Holland, J., Davis, S.S., 1997. Protein-loaded poly(d,l-lactic-co-glycolic acid) microparticles for oral administration: formulation, structural and release characteristics. J. Controlled Release 43, 89–102. Shinnar, R., 1961. On the behavior of liquid dispersions in mixing vessels. J. Fluid Mech. 10, 259–275.
Shinnar, R., Church, J.M., 1960. Statistical theories of turbulence in predicting particle size in agitated dispersions. Ind. Eng. Chem. Res. 52, 253–256. Sprow, F.B., 1967. Distribution of drop sizes produces in turbulent liquid–liquid dispersion. Chem. Eng. Sci. 22, 435–442. Yang, Y-Y., Chung, T-S., Ng, N.P., 2001. Morphology, drug distribution, and in vitro release profiles of biodegradable polymeric microspheres containing protein fabricated by double-emulsion solvent extraction/evaporation method. Biomaterials 22, 231–241.
Contents lists available at SciVerse ScienceDirect
Chemical Engineering Research and Design journal homepage: www.elsevier.com/locate/cherd
Effect of preparation conditions on the properties of microspheres prepared using an emulsion-solvent extraction process Harri Heiskanen a,∗ , Peter Denifl b , Päivi Pitkänen b , Markku Hurme a a
Department of Biotechnology and Chemical Technology, Aalto University School of Science and Technology, P.O. Box 6100, FIN-02015 HUT, Finland b Borealis Polymers Oy, P.O. Box 330, FIN-06101 Porvoo, Finland
a b s t r a c t Methylaluminoxane microspheres were prepared using a hydrocarbon-in-perfluorocarbon-emulsion solvent extraction process. The effect of the preparation conditions on the size of the microspheres was investigated. As expected, the size of the microspheres decreased with increasing stirring speed. At low surfactant concentrations the size of the microspheres was independent of the surfactant concentration. However, the size of the microspheres decreased as the surfactant concentration was further increased. The size of the microspheres was not only affected by the surfactant concentration but also by the volume ratio of the dispersed phase to the continuous phase. At a low volume ratio of the phases the effect of the surfactant on the size of the microspheres was larger than the effect of the increased volume ratio of the phases. At high volume ratios of the phases the effect of the volume ratio of phases on the size of the microspheres became more significant than the effect of the surfactant. A slow solidification increased the formation of non-spherical microspheres. © 2012 The Institution of Chemical Engineers. Published by Elsevier B.V. All rights reserved. Keywords: Microsphere; Emulsification; Particle size distribution; Solvent extraction
1.
Introduction
Emulsion-solvent extraction/evaporation processes are widely used to produce microspheres for different applications such as the controlled release of drugs, cosmetics, inks, pigments and chemical reagents. Control of the microsphere size distribution and morphology of the microspheres is important in the microsphere preparation processes, for example for obtaining repeatable and controlled release behavior of a drug. Freitas et al. (2005) and Li et al. (2008) have written recent reviews about the microsphere preparation using emulsion-solvent extraction/evaporation processes. The equilibrium droplet size of the emulsion is the result of the droplet break-up and the coalescence. The droplets break up if the shear forces exceed the resistance forces. The maximum stable droplet size, dmax , in dilute dispersions can be related to the impeller Weber number, We, using the following
∗
equation (Hinze, 1955): dmax = c1 D
c N2 D3
−0.6 = c1 We−0.6
(1)
where c is the density of the continuous phase, N the stirring speed, D the diameter of the stirrer, the interfacial tension and c1 is a coefficient. The maximum droplet diameter correlates linearly with the Sauter mean diameter, d32 (Sprow, 1967). Therefore, Eq. (1) can be written as: d32 = c2 We−0.6 D
(2)
where c2 is a coefficient. Increasing the surfactant concentration reduces the interfacial tension between the phases until the critical micelle
Corresponding author. Fax: +358 9 451 2694. E-mail address: harri.heiskanen@aalto.fi (H. Heiskanen). Received 17 November 2010; Received in revised form 9 January 2012; Accepted 14 February 2012 0263-8762/$ – see front matter © 2012 The Institution of Chemical Engineers. Published by Elsevier B.V. All rights reserved. doi:10.1016/j.cherd.2012.02.008
1518
chemical engineering research and design 9 0 ( 2 0 1 2 ) 1517–1526
(Bahukudumbi et al., 2004; Mu et al., 2005): the size of the microspheres decreases as the Weber number increased in the case of aqueous emulsion systems. In our previous work (Heiskanen et al., 2010) with a non-aqueous emulsion system the Sauter mean diameter of the microspheres decreased as the Weber number was increased at high surfactant concentration. However, at low surfactant concentration the Sauter mean diameter of the microspheres increased as the Weber number was increased. The coalescence of the droplets increased more than the break-up as the mixing was enhanced, therefore the size of the microspheres increased. A surfactant concentration was found in which the Sauter mean diameter of the microspheres kept constant although the Weber number was varied.
Nomenclature ci dmax d32 dx
D ki N Re We
coefficient i maximum stable droplet size (m) Sauter mean diameter (m) diameter corresponding to x vol.% on a relative cumulative droplet diameter distribution curve (m) diameter of stirrer (m) coefficient i stirring speed (1/s) the Reynolds number, Re = DN2 /c the Weber number
Greek letters viscosity of continuous phase (Pa s) c density of continuous phase (kg/m3 ) , c interfacial tension (N/m) volume fraction of disperse phase ˚d
1.2. Effect of surfactant concentration on the microsphere size
concentration is reached. Further increase in the surfactant concentration does not alter the interfacial tension anymore remaining constant. In addition, coalescence can be reduced using a surfactant. As the droplet breaks up a new liquid–liquid interface is created, coalescence is prevented if the surfactant is absorbed to the new liquid–liquid interface fast enough. Eq. (1) depends only on the break-up of the droplets and does not take into account the coalescence of the droplets. Modified correlations between the Sauter mean diameter and the Weber number have been presented for aqueous systems which take into account the volume ratio of phases (Calderbank, 1958; Davies, 1992; Godfrey et al., 1989; Shinnar, 1961; Shinnar and Church, 1960). The correlation presented by Davies (1992) and Godfrey et al. (1989) is written: d32 = k1 (1 + k2 ˚d )We−0.6 D
(3)
where k1 and k2 are coefficients and ˚d is the volume fraction of the dispersed phase. The term (1 + k2 ˚d ) in Eq. (3) reflects the influence of the dispersed phase in reducing the overall level of turbulence in the system. The turbulent damping is thought to be the main reason for increased droplet size as the volume ratio of the phases increases (Leng and Calabrese, 2004). A higher volume ratio of the phases leads to growing collision rates of droplets which results in an increasing number of break-up and coalescence events. As a result of the increased volume ratio of the phases the likelihood of coalescence increases more than the break-up, resulting in an increased droplet size.
1.1.
Effect of stirring speed on the microsphere size
The effect of the stirring speed on the microsphere size has been widely studied using aqueous emulsion systems. In general, an increase in the stirring speed increases the Weber number (Eq. (1)), therefore higher shear forces are created and smaller microspheres are produced (Gabor et al., 1999; Jégat and Taverdet, 2000; Mateovic et al., 2002; Nepal et al., 2007). Gabor et al. (1999) reported that the microsphere size distribution narrowed as the stirring speed increased. Correlations between the Sauter mean diameter of microspheres and the Weber number have been presented
The surfactant affects both the emulsification and the solidification processes during the microsphere preparation. The surfactant decreases the interfacial tension which results in increased film drainage time between approaching droplets, therefore, the coalescence of the droplets is reduced. The surfactant also decreases the mass transfer rate through the interface of the phases (Lee, 2003). As the surfactant concentration increases the interfacial tension decreases and hence the size of the droplets decreases (Eqs. (1) and (2)). The mean size of the microspheres has been reported to decrease almost linearly with increasing surfactant concentration with aqueous emulsion systems (Mu et al., 2005; Rafati et al., 1997). Nepal et al. (2007) also noticed almost linear decrease in the mean size of the microspheres as the surfactant concentration increased from 0.2% to 0.8% (w/v). However, the mean size of the microspheres remained constant as the surfactant concentration was further increased to 1.2% (w/v). The surfactant concentration has been reported to affect the morphology of microspheres and the drug distribution within the microspheres (Yang et al., 2001).
1.3. Effect of volume ratio of phases on the microsphere size According to Eq. (3) the mean droplet size increases as the volume ratio of the phases increases. However, such a clear correlation between the volume ratio of the phases and the size of the microspheres has not been confirmed via experimental data. Gabor et al. (1999) reported that the size of the microspheres decreased when the volume ratio of the phases was increased (dichloromethane-in-water emulsion). A further increase in the volume ratio of the phases increased the size of the microspheres, an explanation for the observed trend was not offered by the authors. Jeffery et al. (1993) reported that the size of the microspheres increased as the volume of the external aqueous phase was increased. The increase in the average microsphere size was attributed to a reduction in agitation due to a decrease in mixing efficiency associated with larger volumes. Ellis and Jacqueir (2009) reported that the size of the microspheres increased (carrageenan-in-oil emulsion) when the volume ratio of the phases was increased. A further increase in the volume ratio of the phases did not have a significant effect on the size of the microspheres. It was assumed that the emulsion became more viscous and hence larger microspheres were produced as the dispersed phase volume increased.
chemical engineering research and design 9 0 ( 2 0 1 2 ) 1517–1526
2.
Aim
Emulsions used in microsphere preparation processes are usually oil-in-water or water-in-oil emulsions but in this study a hydrocarbon-in-perfluorocarbon emulsion was used. The methylaluminoxane droplets were solidified using a solvent extraction method; perfluorocarbon was added to the emulsion to extract the solvent (toluene) from the dispersed phase into the continuous phase. Our previous results (Heiskanen et al., 2010, 2012) showed that the emulsification conditions had a significant effect on the methylaluminoxane microsphere properties. In those experiments a different solidification method was used: the emulsion was added to an excess of perfluorocarbon. The following experiments were performed to investigate how the different microsphere preparation method affects the methylaluminoxane microsphere size and size distribution. The effect of the surfactant concentration, the stirring speed, the volume ratio of the phases, and the solidification rate on the methylaluminoxane microsphere size and the size distribution were all investigated as variables.
3.
Materials and methods
3.1.
Materials
Hexadecafluoro-1,3-dimethylcyclohexane, PFC, (CAS 335-273) was purchased from Fluorochem. 30 wt.% methylaluminoxane in toluene (methylaluminoxane CAS 120144-90-3; toluene CAS 108-88-3) was purchased from Albemarle. 3-Perfluorooctyl-1,2-propenoxide, PFPO, (CAS 38565-53-6; precursor of surfactant) was purchased from Fluorochem. PFC, toluene and PFPO were deoxygenized with argon prior to use. White mineral oil (CAS number: 8042-47-5) was used to prepare a sample for the particle size analysis.
3.2.
Microsphere size analysis and microscopy
Droplets produced in the emulsification step shrink during the solidification because the solvent (toluene) is extracted from the droplet into the continuous phase. Hence, the droplet size is larger than the microsphere size. The microspheres produced were mainly compact but some hollow microspheres were also produced. The compactness of the microspheres was not taken into account when the mean size of the microspheres was analyzed. Microsphere size analysis was performed by dispersing dry MAO microsphere powder (∼20 mg) into white mineral oil. Pictures were taken from this dilute suspension using a light microscope, these images were then analyzed using the CantyVisionClientTM software. Only spherical particles with a diameter of at least 1 m were taken into account for microsphere size distribution. The number of non-spherical microspheres was low and therefore could be discounted when the microsphere size distribution was calculated. At least 3000 spherical microspheres per batch were included during microsphere size distribution analysis. The span was calculated to describe the broadness of the microsphere size distribution:
span =
d90 − d10 d50
(4)
1519
where dx is the diameter of the particle corresponding to x vol.% on a relative cumulative droplet diameter distribution curve. A small span indicates a narrow size distribution. A scanning electron microscope (FEI Quanta 200) was used to determine the microsphere shape and morphology. Microspheres were mounted on metal stubs with double-sided conductive adhesive tape and coated with a thin layer of gold. Coating was performed under inert atmosphere. SEM imaging was performed using an electron beam energy of 3 keV. Some of the microspheres were cut before the SEM images were taken to assess the morphology and compactness of particles.
3.3.
Microsphere preparation
The experiments were performed to study the effect of the microsphere preparation conditions on the methylaluminoxane microsphere morphology, size and size distribution. All experiments were performed under inert conditions (Argon).
3.3.1.
Emulsification step
The emulsification was performed at room temperature (22 ± 1 ◦ C). The hydrocarbon-in-fluorocarbon-emulsion was prepared in a stirred vessel (made from glass; 4 baffles). The diameter of the anchor type stirrer (made from Teflon) was 3.9 cm. The conditions used in the experiments are summarized in Table 1. In the first step 43 ml PFC (the continuous phase) was added into the emulsification vessel. A small amount of toluene was added to the PFC in order to saturate the continuous phase with toluene (while remaining a one liquid phase system) before the emulsification, no solvent (toluene) extraction occurred during the emulsification. The stirring speed was varied from 450 to 600 rpm (Table 1). 2 ml methylaluminoxane (MAO) in toluene, the disperse phase solution, was injected into the emulsification vessel. After addition of the disperse phase solution (∼3 min), the surfactant solution (PFPO in PFC) was added to the emulsion drop wise over a period of approximately 1 min. The surfactant concentration in the continuous phase was varied from 0.01% to 0.15% (v/v) (Table 1). The emulsion was stirred for 15 min.
3.3.1.1. Effect of volume ratio of phases on the microsphere preparation. The volume of the continuous phase was kept constant, 43 ml, while the effect of volume ratio of the phases on the microsphere size and the size distribution was studied. The emulsion was made following the procedure described above with the volume of the dispersed phase varying from 1 to 3 ml. A smaller dispersed phase volume creates fewer droplets, therefore, less surface area is created requiring stabilization during the emulsification if the droplet size distribution remains constant. The surfactant concentration in the continuous phase was varied so that the volume ratio of the surfactant to the dispersed phase was constant.
3.3.2.
Solidification step
The stirring speed during the solidification was the same as it was during the emulsification. PFC (100 ml) was added to the emulsion in order to extract the solvent (toluene) from the dispersed phase into the continuous phase. The flow rate of the PFC added to the emulsion was varied from 15 to 330 ml/min (Table 1). After the PFC addition, mixing was continued for 10 min.
1520
chemical engineering research and design 9 0 ( 2 0 1 2 ) 1517–1526
Table 1 – Summary of the microsphere preparation conditions. Stirring speed (rpm)
Stirring speed
Surfactant concentration
Volume ratio
Solidification rate
3.3.3.
RE09 RE10 RE03 RE08 RE11 RE04 RE05 RE06 RE07 RE03 RE08 RE12 RE13 RE12 RE14 RE03 RE08 RE01 RE02 RE15 RE28 RE29
450 500 550 550 600 550 550 550 550 550 550 550 550 550 550 550 550 550 550 550 550 550
Volume ratio (disp./cont. phase) 0.047 0.047 0.047 0.047 0.047 0.047 0.047 0.047 0.047 0.047 0.047 0.047 0.023 0.047 0.070 0.047 0.047 0.047 0.047 0.047 0.047 0.047
Product isolation
After 30 min without agitation, the microspheres had settled to the top of the PFC. The PFC was siphoned from the vessel and the remaining wet microspheres were dried under argon flow at 50 ◦ C for 1 h. The microsphere yield varied from 0.13 to 0.40 g. The residual toluene content of the dried microspheres was not measured.
4.
Results and discussion
Methylaluminoxane microspheres were successfully produced using the hydrocarbon-in-perfluorocarbon-emulsion solvent extraction process under all tested conditions. The variation between duplicate experiments was small: the average d32 of the experiments RE03 and RE08 was 21.1 ± 1.5 m and the average d32 of the experiments RE02 and RE15 was 22.8 ± 0.7 m (Table 1). The small variation of the replicates indicates that the experiments are repeatable even though the sample for the microsphere size measurement was small and was taken from dry powder. Dry microspheres can segregate by size and therefore taking a representative sample from the dry powder can be difficult.
4.1.
Surfactant in PFC ((v/v)%) 0.09 0.09 0.09 0.09 0.09 0.009 0.019 0.027 0.051 0.09 0.09 0.10 0.05 0.10 0.15 0.09 0.09 0.09 0.09 0.09 0.09 0.09
PFC flow rate (ml/min) 320 320 330 320 320 320 320 320 320 330 320 200 100 200 290 330 320 210 80 80 30 15
Theoretical solidification time (s) 8 8 8 8 8 8 8 8 8 8 8 13 13 13 13 8 8 12 31 31 85 170
d32 (m) 40.5 23.2 22.2 20.0 19.1 35.3 38.5 36.0 38.5 22.2 20.0 19.7 37.8 19.7 25.1 22.2 20.0 21.6 22.3 23.3 24.7 24.7
Span
0.59 0.61 0.59 0.50 0.61 0.61 0.55 0.58 0.68 0.59 0.50 0.63 1.07 0.63 0.56 0.59 0.50 0.52 0.68 0.61 0.84 0.60
the microspheres: from 23.2 to 19.1 m. The microsphere size distribution narrowed significantly as the stirring speed was increased (Fig. 1). A similar trend in the microsphere size distribution was observed by Gabor et al. (1999). The Sauter mean diameter of the microspheres decreased exponentially from 40.5 to 19.1 m with increasing Weber number, from 2400 to 4300 (Fig. 2). The shear forces breaking up droplets increased as the Weber number increased. Therefore, smaller microspheres were produced. In the Weber number calculations (Eq. (2)) the interfacial tension between the emulsion phases was assumed to be 0.0025 N/m which has been measured for a toluene-PFC system (Heiskanen et al., 2010). The addition of a surfactant is known to reduce the interfacial tension, this would indicate that the interfacial tension value used in the calculations was probably too high. The interfacial tension changes the constant in the correlation between the Sauter mean diameter and the Weber number (Eq. (2)) but does not affect the exponent value in the correlation. The theoretical exponent for a non-coalescencing and fully turbulent system is −0.6 (Eq. (2)). These experiments were
Effect of stirring speed on the microsphere size
As expected, the Sauter mean diameter of the microspheres decreased from 40.5 m to 23.2 m as the stirring speed increased from 450 rpm to 500 rpm. The bulk mixing at the slowest stirring speed was not intense enough to disperse the solidifying droplets in the bulk emulsion during the microsphere preparation leading to large droplets. The stability of the studied hydrocarbon-in-fluorocarbon emulsion was poor, partly due to the high density difference between the phases. The bulk mixing was enhanced as the stirring speed was increased from 450 rpm to 500 rpm, reducing the coalescence of the droplets and producing significantly smaller microspheres. Further increase in the stirring speed from 500 rpm to 600 rpm led to a smaller decrease in the size of
Fig. 1 – Particle size distribution (volume) of the spherical microspheres prepared with different stirring speeds (450–600 rpm).
chemical engineering research and design 9 0 ( 2 0 1 2 ) 1517–1526
done at fully turbulent conditions (the Reynolds number varied from 10,000 to 13,300). The exponent value for this system was −0.53 when the slowest stirring speed was not considered (Fig. 2), under the assumption that the bulk mixing was not intense enough to maintain complete dispersion during the solidification. One reason for the deviation from the theoretical value was that the emulsion studied was a coalescencing system. The exponent value was −1.26 (Fig. 2), when all data points were included, which differs significantly from the theoretical exponent value. In our previous study (Heiskanen et al., 2010) with a different test configuration, the exponent value was −0.56 at high surfactant concentration. Possible explanations for the deviation from the theoretical value are that some of the experiments were performed in transient flow and that the emulsion studied was a coalescencing system. Bahukudumbi et al. (2004) reported the same exponent value, −0.53, however, they did not offer any explanation for the deviation from the theoretical value. An exponent value, −0.72, in the correlation between the Sauter mean diameter of the microspheres and the Weber number was presented by Mu et al. (2005), a transient flow was suggested as the reason for the deviation from the theoretical exponent value.
4.2. Effect of surfactant concentration on the microsphere size Increased surfactant concentration is known to reduce the interfacial tension, the coalescence and also the mass transfer. Hence the surfactant concentration affects both the emulsification and the solidification steps during the microsphere preparation process. The Sauter mean diameter of the microspheres did not change when the surfactant concentration was increased from 0.01% to 0.05% (v/v) (Table 1; Fig. 3). The surfactant concentration at that range was not high enough to stabilize the new liquid–liquid interfaces formed during droplet break-up. Therefore, the surfactant addition did not reduce the coalescence of the droplets and the Sauter mean diameter remained almost constant. A further increase in the surfactant concentration from 0.05% to 0.09% (v/v) significantly decreased the Sauter mean diameter from ∼39 m to ∼21 m. The Sauter mean diameter of the microspheres decreased almost inversely as a function of the surfactant
Fig. 2 – The Sauter mean diameter of the spherical microspheres as a function of the Weber number. Two curve fittings: (I) all data points included and (II) the data point achieved using the slowest stirring speed excluded.
1521
Fig. 3 – Particle size distribution (volume) of the spherical microspheres prepared with different surfactant concentrations. The solidification time was ∼8 s in all other experiments but 13 s at the highest surfactant concentration (0.10% (v/v) PFPO in PFC).
concentration in the concentration range of 0.05–0.10% (v/v) (Fig. 4). In several studies a similar linear decrease in the size of the microspheres with increasing surfactant concentration has been observed (Mu et al., 2005; Nepal et al., 2007; Rafati et al., 1997). The surfactant is known to reduce the interfacial tension between the phases (not measured for this system), additionally, the Sauter mean diameter depends inversely on the interfacial tension (Eqs. (1) and (2)). The film drainage rate of the approaching droplets increases with decreasing interfacial tension, leading to a reduction in the coalescence of the droplets and the size of the microspheres decreases. The Sauter mean diameter of the microspheres did not reduce significantly as the surfactant concentration was increased from 0.09% to 0.10% (v/v) indicating that the surfactant concentration was close to the critical micelle concentration. The interfacial tension does not change after the critical micelle concentration is reached although the surfactant concentration is further increased.
Fig. 4 – The Sauter mean diameter of the spherical microspheres as the function of surfactant concentration in the continuous phase. Experiments were performed with different volume ratios of the dispersed to the continuous phase: 0.023, 0.047 and 0.070. The solidification time was varied from ∼8 s to 13 s.
1522
chemical engineering research and design 9 0 ( 2 0 1 2 ) 1517–1526
4.3. Effect of volume ratio of the phases on the microsphere size In theory, the higher the volume ratio of the phases, the larger the interface area created during the emulsification. As such, more surfactant is needed to prevent the coalescence of the droplets. The surfactant concentration in the continuous phase was varied so that the volume ratio of the surfactant to the dispersed phase was constant when the effect of the volume ratio of the phases on the microsphere was studied. The Sauter mean diameter of the microspheres remained almost the same, ∼38 m, because the surfactant concentration was roughly 0.05% (v/v) although the volume ratio of the phases varied from 0.023 to 0.047 (Table 1; Fig. 4). In conclusion, the size of the microspheres was not affected by the volume ratio of the phases at surfactant concentration of 0.05% (v/v). The Sauter mean diameter of the microspheres decreased from 37.8 to 19.7 m when the volume ratio of the phases increased from 0.023 to 0.047 (Table 1; Fig. 4). The surfactant concentration was 0.05% (v/v) at a volume ratio of 0.023% and 0.10% (v/v) at a volume ratio of 0.047. The increased surfactant concentration reduced the interfacial tension and hence smaller microspheres were produced even though the volume ratio of the phases increased. Therefore, at a low volume ratio of the phases (less than 0.047) the size of the microspheres was primarily affected by the surfactant concentration as opposed to the variation in the volume ratio of the phases. The Sauter mean diameter of the microspheres increased from 19.7 to 25.1 m when the volume ratio of the phases was further increased from 0.047 to 0.070 (Table 1; Fig. 4). The surfactant concentration was 0.10% (v/v) at a volume ratio of 0.047% and 0.15% (v/v) at a volume ratio of 0.070. Despite the increased surfactant concentration the size of the microspheres increased as the volume ratio of the phases increased. The effect of the volume ratio of the phases on the size of the microspheres became more significant than the effect of the surfactant at high volume ratio of the phases. According to Eq. (3) the size of the microspheres is expected to increase with the increasing volume ratio of the phases. The main explanation for the increased microsphere size is the turbulence dampening caused by the increased volume ratio of the phases. Additionally, the number of collisions should increase as the volume ratio of the phases increases, thus increasing the size of the microspheres. Gabor et al. (1999) and Heiskanen et al. (2010) noticed a similar trend in the size of the microspheres as the volume ratio of the phases was varied: the size of the microspheres first decreased as the volume ratio of the phases was increased. A further increase in the volume ratio of the phases led to an increase in the size of the microspheres. In our previous experiments (Heiskanen et al., 2010) a different solidification method was used: the hydrocarbon-in-perfluorocarbon emulsion was fed into an excess of perfluorocarbon. No correlation between the Sauter mean diameter of microspheres and the surfactant concentration was seen when the volume ratio of the phases was varied as conflicting with the findings of this study. However, the Sauter mean diameter of the microspheres decreased exponentially as a function of the volume ratio of the surfactant to the dispersed phase although a different volume ratio of the phases was used. The results indicated that the emulsion droplet size distribution changed when the emulsion was fed into the excess of perfluorocarbon, and the size of the microspheres was mainly affected by the volume
Fig. 5 – Particle size distribution (volume) of the spherical microspheres prepared with different volume ratios of the dispersed to the continuous phase. The solidification time was ∼13 s in all experiments. The volume ratio of surfactant to the dispersed phase was kept constant. ratio of the surfactant to the dispersed phase and not by the volume ratio of the phases. The microsphere size distribution was broadest at the lowest volume ratio of the phases due to the low surfactant concentration (Fig. 5; Table 1). The microsphere size distribution narrowed despite the volume ratio of the phases was increasing because the surfactant concentration increased (Fig. 5; Table 1). The volume ratio of the phases had an effect on the microsphere morphology according to the SEM images (Fig. 6). The cracks seen in the SEM images were formed during the SEM-sample preparation and not during the microsphere preparation. The majority of microspheres were produced with all tested volume ratios of the phases were compact spheres. However, the surfaces of the microspheres produced with a high volume ratio of phases (Fig. 6C) were not as smooth as those produced with the low volume ratios of the phases (Fig. 6A and B). The formation of the non-smooth surfaces is probably linked to the higher toluene volume and/or the surfactant concentration. The addition rate of the PFC to the emulsion was set so that the time needed to extract all toluene from the dispersed phase into the continuous phase was constant in all volume ratio experiments. Therefore, the toluene extraction rate (g toluene/s) was the highest at high volume ratios of the phases which could explain the observed change in the surface morphology of the microspheres. The solidification rate had no significant effect on the surface morphology of the microspheres when the volume ratio of the phases was lower, 0.047. Another possible reason for the formation of a non-smooth surface could be the high surfactant concentration (0.15% (v/v)) at the highest volume ratio of the phases. The surfactant concentration had no significant effect on the microsphere morphology when the surfactant concentration was less than 0.10% (v/v) at a volume ratio of phases 0.047. The surface viscosity of the droplets is known to increase with increasing surfactant concentration especially after the critical micelle concentration (Leng and Calabrese, 2004). With the increased surface viscosity, the deformed droplet (deformation caused by shear forces) did not return back to the thermodynamically more favorable spherical shape. Therefore, the microspheres had rough surfaces at high surfactant concentration with a high volume ratio of the phases. A high surfactant concentration is also
chemical engineering research and design 9 0 ( 2 0 1 2 ) 1517–1526
1523
Fig. 6 – SEM-images from microspheres prepared with different volume ratios of the dispersed to the continuous phase: (A) 0.023, (B) 0.047, and (C) 0.070. The solidification time was ∼13 s in all experiments. known to reduce the mass transfer so the droplet viscosity increases slowly and the droplet is exposed to shear forces for longer. Mateovic-Rojnik et al. (2005) noticed that the particle surface roughness increased with decreasing solidification rate.
4.4.
Effect of solidification rate on the microsphere size
During the experiments an excess of PFC was added to the emulsion in order to solidify the droplets. The increased continuous phase volume affected both the surfactant concentration in the continuous phase and the volume ratio of the phases. The surfactant concentration was roughly half of the initial value when all the toluene was extracted from the dispersed phase into the continuous phase. As the surfactant concentration decreases the interfacial tension increases and hence the droplet size increases in theory (see Eqs. (1) and (2)). The number of collisions should reduce as the volume ratio of the phases decreases, the coalescence reduces and the droplet size decreases. In the experiments the volume ratio of the phases decreased from 0.047 to 0.023 during the solvent extraction. The solubility of toluene in PFC is 2.2 wt.% at 22 ◦ C. The time needed to add the volume of PFC required to extract all the toluene from the dispersed phase (solidification time) varied from 8 to 170 s. The Sauter mean diameter of the microspheres increased slightly as the solidification rate decreased (Table 1). More non-spherical microspheres were produced as the solidification rate decreased, as shown in the light microscopy
pictures (Fig. 7A–E). One explanation for the formation of nonspherical microspheres at slow solidification rate was that the solidifying droplets were stretched by the shear forces. The stretched droplets did not return to the thermodynamically more favorable spherical shape due to the high viscosity of the solidifying droplet. Some microspheres had also collided during the solidification process forming agglomerates. A narrow microsphere size distribution was produced in all tested conditions (Fig. 8). The size distribution of the spherical microspheres did not change significantly although the solidification rate was varied. Some deposition of the methylaluminoxane, MAO, on the wall of the preparation vessel was noticed, especially with the two slowest solidification rates. At the height of the deposition (=white ring) the total liquid volume was roughly 80 ml. At that point MAO concentration in the dispersed phase had increased from 30 wt.% to roughly 67 wt.% as a result of the solvent (toluene) extraction. Not all of the solidifying droplets were properly dispersed in the bulk liquid during the microsphere preparation, leading to some of the solidifying droplets separating and accumulating on the wall forming the solid white deposition. Samples (∼1 ml) were taken via needle and syringe from the emulsion during the solidification at three different solidification degrees: ∼91, ∼97 and ∼103%. At the solidification degree 100% all toluene was assumed to be extracted from the dispersed phase into the continuous phase. The time needed to add enough PFC in order to extract all toluene (the solidification time) was 170 s. The samples were injected into
1524
chemical engineering research and design 9 0 ( 2 0 1 2 ) 1517–1526
Fig. 7 – Light microscopy images from the microspheres prepared with different solidification rates; the time needed to extract all toluene from the solidifying droplets: (A) 8 s, (B) 12 s, (C) 31 s, (D) 85 s, and (E) 170 s. Same magnification in all pictures. 10 ml septa vials to investigate whether solid microspheres had formed. The microspheres in all samples seemed to be solid immediately after the sample taking. After 10 min solid microspheres could only be observed in the sample with the highest solidification degree, ∼103%. Some microspheres were seen in the sample with solidification degree ∼97%. The microspheres in the sample with the lowest solidification degree had become liquid. The results indicated that the microspheres have a solid crust when the solidification degree is roughly ∼91% but still contain too much solvent inside the microsphere to remain solid.
5.
Fig. 8 – Particle size distribution (volume) of the spherical microspheres prepared with different solidification rates (the time needed to extract all toluene from the solidifying droplets varied from 8 to 170 s).
Conclusions
The effect of the solidification time, the surfactant concentration, the stirring speed, and the volume ratio of the phases on the methylaluminoxane microsphere size and the size distribution was successfully studied using a hydrocarbonin-fluorocarbon emulsion solvent extraction based microsphere preparation process. The methylaluminoxane droplets
chemical engineering research and design 9 0 ( 2 0 1 2 ) 1517–1526
formed in the first step of the process were solidified by adding an excess of perfluorocarbon to the emulsion. In agreement with the literature, the size of the microspheres was found to decrease with increasing stirring speed. The Sauter mean diameter decreased exponentially from ∼41 to ∼19 m with increasing Weber number (2400–4300). The exponent value in the correlation between the Sauter mean diameter and the Weber number for this system was −0.53 which is close to the theoretical value of −0.6. The experiment performed at the slowest stirring speed was removed from the correlation because the bulk mixing was assumed to be inadequate during that experiment. The microsphere size distribution narrowed significantly as the stirring speed was increased. The size of the microspheres decreased as the surfactant concentration was increased, as expected. Low surfactant concentrations did not stabilize the emulsion enough, meaning the coalescence of the droplets did not reduce and the Sauter mean diameter of the microspheres did not change. However, a further increase in the surfactant concentration significantly decreased the Sauter mean diameter of the microspheres as the interfacial tension and coalescence of the droplets decreased. The Sauter mean diameter of the microspheres decreased inversely as a function of the surfactant concentration, in the concentration range of 0.05–0.10% (v/v). According to the correlation between the Sauter mean diameter and the Weber number, the Sauter mean diameter depends inversely on the interfacial tension which is known to decrease as the surfactant concentration increases. Therefore, the surfactant addition decreased the interfacial tension which resulted in the formation of smaller microspheres. The Sauter mean diameter of the microspheres was not only affected by the surfactant concentration but also by the volume ratio of the phases. At low volume ratios of the phases (from 0.023 to 0.047) the size of the microspheres was found to be primarily affected by the surfactant concentration and not by the volume ratio of the phases. The increased surfactant concentration reduced the interfacial tension and the coalescence of droplets leading to smaller microspheres, although the volume ratio of the phases increased from 0.023 to 0.047. As the volume ratio of the phases was further increased from 0.047 to 0.070, the Sauter mean diameter of the microspheres increased from ∼20 to ∼25 m, even though the surfactant concentration was higher at the volume ratio of 0.070. Therefore, at high volume ratios of the phases the effect of the volume ratio of phases on the size of the microspheres became more significant than the effect of the surfactant concentration. One reason for the increased microspheres size, despite the surfactant concentration increasing, was the dampening of the turbulence due to an increased volume ratio of the phases. Also the number of collisions may be increased as the volume ratio of the phases increased, leading to increased coalescence and larger microspheres. A slow solidification rate increased the formation of nonspherical particles because the solidifying droplets were stretched by the shear forces. At the end of the solidification step the stretched droplets did not return to the thermodynamically more favorable spherical shape because of the high viscosity of the solidifying droplet leading to the formation of non-spherical particles. Some particles also collided during the solidification process forming agglomerates. Deposition of the matrix material on the wall of the preparation vessel was seen at the slowest solidification rates tested, due to solidifying droplets not being properly dispersed in the bulk liquid.
1525
Therefore part of the solidifying droplets separated depositing on the wall.
Acknowledgment Financial support from Borealis Polymers Oy is gratefully acknowledged.
References Bahukudumbi, P., Carson, K.H., Rice-Ficht, A.C., Andrews, M.J., 2004. On the diameter and size distributions of Bovine Serum Albumin (BSA)-based microspheres. J. Microencapsulation 21, 787–803. Calderbank, P.H., 1958. Physical rate processes in industrial fermentation, part I: the interfacial area in gas–liquid contacting with mechanical agitation. Trans. Inst. Chem. Engrs. 36, 443–463. Davies, G.A., 1992. Mixing and coalescence phenomena in liquid–liquid systems. In: Thornton, J.D. (Ed.), Science and Practice of Liquid–Liquid Extraction, vol. I. Clarendon Press, Oxford, UK, p. 245. Ellis, A., Jacqueir, J.C., 2009. Manufacture of food grade -carrageenan microspheres. J. Food Eng. 94, 316–320. Freitas, S., Merkle, H.P., Gander, B., 2005. Microencapsulation by solvent extraction/evaporation: reviewing the state of the art of microsphere preparation process technology. J. Controlled Release 102, 313–332. Gabor, F., Ertl, B., Wirth, M., Mallinger, R., 1999. Ketoprofen-poly(d,l-lactic-co-glycolic acid) microspheres: influence of manufacturing parameters and type of polymer on the release characteristics. J. Microencapsulation 16, 1–12. Godfrey, J.C., Obi, F.I., Reeve, R.N., 1989. Measuring drop size in continuous liquid–liquid mixers. Chem. Eng. Prog. 85, 61–69. Heiskanen, H., Denifl, P., Hurme, M., Pitkänen, P., 2010. Effect of emulsification conditions on the properties of the microspheres prepared by a solvent extraction process. Chem. Eng. Technol. 33, 1635–1644. Heiskanen, H., Denifl, P., Hurme, M., Pitkänen, P., Oksman, M., 2012. Effect of physical properties and emulsification conditions on the microsphere size prepared using a solvent extraction process. J. Disper. Sci. Technol. 33, 1–11. Hinze, J.O., 1955. Fundamentals of the hydrodynamics mechanisms of splitting in dispersion process. AIChE J. 1, 289–295. Jeffery, H., Davis, S.S., O’Hogan, D.T., 1993. The preparation and characterization of poly(lactide-co-glycolide) microparticles. II. The entrapment of a model protein using a (water-in-oil)-in-water emulsion solvent evaporation technique. Pharm. Res. 10, 362–368. Jégat, C., Taverdet, J.L., 2000. Stirring speed influence study on the microencapsulation process and on the drug release from microcapsules. Polym. Bull. 44, 345–351. Lee, Y.-L., 2003. Surfactant effects on mass transfer during the drop-formation and drop falling stages. AIChE J. 49, 1859–1869. Leng, D.E., Calabrese, R.V., 2004. Immiscible liquid–liquid systems. In: Paul, E.L., Atiemo-Obeng, V.A., Kresta, S.M. (Eds.), Handbook of Industrial Mixing: Science and Practice. John Wiley & Sons Inc., NJ, pp. 639–753 (Chapter 12). Li, M., Rouad, O., Poncelet, D., 2008. Review – microencapsulation by solvent evaporation: state of the art for process engineering approaches. Int. J. Pharm. 363, 26–39. Mateovic, T., Kriznar, B., Bogataj, M., Mrhar, A., 2002. The influence of stirring rate on biopharmaceutical properties of Eudragit RS microspheres. J. Microencapsulation 19, 29–36. Mateovic-Rojnik, T., Frlan, R., Bogataj, M., Bukovec, P., Mrhar, A., 2005. Effect of preparation temperature in solvent evaporation Eudragit RS microsphere properties. Chem. Pharm. Bull. 53, 143–146. Mu, Y., Lyddiat, A., Pacek, A.W., 2005. Manufacture by water/oil emulsification of porous agarose beads: effect of processing
1526
chemical engineering research and design 9 0 ( 2 0 1 2 ) 1517–1526
conditions on mean particle size, size distribution and mechanical properties. Chem. Eng. Process. 44, 1157–1166. Nepal, P.R., Chun, M-K., Choi, H-K., 2007. Preparation of floating microspheres for fish farming. Int. J. Pharm. 341, 85–90. Rafati, H., Coombes, A.G.A., Adler, J., Holland, J., Davis, S.S., 1997. Protein-loaded poly(d,l-lactic-co-glycolic acid) microparticles for oral administration: formulation, structural and release characteristics. J. Controlled Release 43, 89–102. Shinnar, R., 1961. On the behavior of liquid dispersions in mixing vessels. J. Fluid Mech. 10, 259–275.
Shinnar, R., Church, J.M., 1960. Statistical theories of turbulence in predicting particle size in agitated dispersions. Ind. Eng. Chem. Res. 52, 253–256. Sprow, F.B., 1967. Distribution of drop sizes produces in turbulent liquid–liquid dispersion. Chem. Eng. Sci. 22, 435–442. Yang, Y-Y., Chung, T-S., Ng, N.P., 2001. Morphology, drug distribution, and in vitro release profiles of biodegradable polymeric microspheres containing protein fabricated by double-emulsion solvent extraction/evaporation method. Biomaterials 22, 231–241.