**4.1 Mass transfer**

In SAS process, mass transfer occurs between a droplet of organic solvent and a compressed antisolvent. In miscible conditions, above mixture critical point, there is no obvious way to define the interface between the two fluids. Dukhin et al. has evidenced the transient existence of droplets at conditions slightly above the mixture critical point, due to the existence of a dynamic interfacial tension, so a description of mass transfer from a droplet even in miscible conditions seems reasonable (Dukhin et al., 2003).

On the other hand, in the SAS process the solution is generally dilute and the equilibrium compositions of the binary and ternary mixtures are not significantly different. Accordingly to this, the solid present in the solution is not likely to affect the rates of mass transfer of CO2 and solvent to and from the droplet respectively.

Mass transfer depend on the densities differences between solvent and antisolvent, viscosity, diffusivity, droplet or particle diameter and solvent flow rate. Chong et al. developed a mathematical model form mass transfer between a droplet of organic solvent and a compressed antisolvent in complete miscibility in SAS process. Calculations using Peng-Robinson equation of state showed that droplets swell upon interdiffusion when the

Particles Formation Using Supercritical Fluids 469

should be performed at completely developed supercritical conditions (when surface tension vanishes before jet break up occurs). If spherical microparticles are the target, the process conditions in which jet break-up produces micrometric droplets are the right ones; the increase in concentration of the starting solution will increase the average diameter of

Among organic and inorganic compounds that have been processed with SAS process, polymers have remarkable interest and significance. Yeo and Kiran (Yeo & Kiran, 2005) and Tomasko et al. (Tomasko et al., 2003) presented extensive reviews of the supercritical processing of polymers. Because most of polymers are not soluble in supercritical fluids, this antisolvent process is especially suitable for their recrystallization or precipitation in form of microparticles. The polymer is firstly dissolved in a liquid organic solvent and a supercritical fluid is employed as an antisolvent for the polymer. Polymers in form of small particles are useful for several applications like stationary phases in chromatography, adsorbents and catalyst supports, as well as drug delivery systems (Dixon et al., 1993). The polymers must fulfil several requisites: its biocompatibility, non toxicity, providing a suitable medium for preserving the properties and activity of the active substance and easy

It is particularly important for polymer processing with supercritical processes is the glass transition and the melting point temperature depressions induced by the supercritical fluid. In particular, the dissolution of SC-CO2 into the polymer can reduce the glass transition temperature of amorphous polymers (Tomasko et al., 2003), an effect that is caused by intermolecular interactions between the dissolved CO2 and the polymer. The melting point

A number of RESS processes for the encapsulation of particles with polymer (polylactic acid (PLA), polyethylene glycol (PEG), Eudragit) or composite particle formation for the

However, the potential application of RESS for particle coating or encapsulation is limited because the solubility of polymers in SC-CO2 is generally very poor (O'Neill et al., 1998) Compared to RESS, the SAS process offers much more flexibility in terms of choosing suitable solvents. Furthermore, SAS has advantages over RESS because SAS is usually operated under mild conditions compared with those of RESS, which is associated to relatively high temperature and high pressure. Therefore RESS is also less attractive from the perspectives of safety and cost. The SAS process has been carried out for many particles

In order to obtain polymer-drug composites several researches have been carried out at our laboratory. Ethyl cellulose (EC) is a biocompatible and non biodegradable polymer. Ethyl cellulose is commonly used as drug carrier in controlled delivery systems. For instance, ethyl cellulose microcapsules has been used as a drug-delivery device for protecting folic acid from release and degradation in the undesirable environmental conditions of the stomach, whilst allowing its release in the intestinal tract to make it available for absorption. In the same way, ethyl cellulose and antibiotic microcapsules have been developed to use as

At University of Cádiz, ethyl cellulose microparticles were successfully precipitated from dichloromethane (DCM) by SAS process (Gordillo et al., 2008) and particles were reduced from 50-100 to 3-5 µm (Figure 5). The concentration was the factor that had the greatest

depression caused by the dissolution of CO2 is less noticeable in magnitude.

controlled release of drugs have been reported as it was referenced before.

precipitation and polymeric encapsulation of particles of active ingredients.

drug delivery protecting antibiotic of conditions of the stomach.

the particles, but, also their polydispersity.

to process with the selected precipitation technique.

**4.2 Polymer and biopolymers** 

solvent is denser than the antisolvent and shrink when the antisolvent is denser (Chong et al., 2009b).

Some authors have modelled the behavior of an organic solvent droplet, considering different local mass transfer both at subcritical (Werling & Debenedetti, 1999) and supercritical (Werling & Debenedetti, 2000) conditions. In this case, the droplet is considered to be stagnant. Therefore, the only convective motion considered is that induced by the diffusion. At subcritical conditions, calculations showed that there is an initial period of droplet swelling, due to the diffusion of CO2 into the organic solvent. Droplet lifetime decreases as the pressure increases, and increases sharply at near-critical conditions, because diffusivities tend to zero near the critical point. At miscible conditions, mass transfer is much faster than at subcritical conditions. Droplet diameter increases if the density of the organic solvent is higher than that of the CO2, and vice versa. However, this ideal and local approach is often not enough to interpret the results. Elvassore et al. developed a model based on the mass transfer simulations of Werling and Debenedetti. This model included the solute in mass transfer calculations (Elvassore et al., 2004).

Pérez de Diego et al. and Martin et al. developed both models for the evaporation of dichloromethane (Pérez de Diego et al., 2006) and ethanol (Martín et al., 2007) droplets respectively which accounted for the higher mass transfer coefficients due to the convective motion of CO2, explaining the change in particle morphology.

Shekunov et al proposed a simplified approach based on the calculation of different characteristic times (diffusion, jet break up...). They studied the phenomena of turbulent dispersion and micromixing in supercritical carbon dioxide using paracetamol as a model drug compound. They tried to describe the effect of mass-transfer on the particle size and morphology and suggested that particle growth is the time-limiting step (Shekunov et al., 1999).

Then, in order to describe the mass transfer, the drop size distribution, nucleation and particle growth during the drying of the drops as well as the fluid dynamics of the dispersed liquid must be known. Other works also carried out a complete modelling. In these cases, the overall process is modelled, taking into account thermodynamic, hydrodynamic, crystallization and mass transfer aspects (Cardoso et al., 2008; Martin & Cocero, 2004; Lora et al., 2000; Reverchon et al., 2010).

At the University of Cádiz, Tenorio et al, by determining the thermodynamic properties of the phases involved in the process, and applying empirical equations (operations with dimensionless numbers), have estimated the different disintegration regimes of the jet when an N-methyl-pyrrolidone (NMP)-ampicillin solution was injected into the CO2-pressurized chamber. The application of the empirical hydrodynamics model proved the existence of significant mechanisms that stabilize the liquid jet, and it showed that there were limiting hydrodynamic conditions that had to be overcome to drive the process toward the formation of uniform spherical nanoparticles and the achievement of higher yields (Tenorio et al., 2009)

Reverchon et al., in some recent papers (Reverchon et al., 2007, 2008b, 2008c, 2011) studied the link between SAS morphologies and the relative position of the SAS operating point with respect the mixture critical point of the solvent-CO2 mixtures. It was proposed several mechanisms and their interactions to elucidate the different morphologies and dimensions of precipitates. From a practical point of view, the knowledge of the competing mechanisms allows to select the dimensions of the precipitated particles. If nanoparticles are the objective of the process, low concentrations of the liquid solution are preferable and SAS operation

solvent is denser than the antisolvent and shrink when the antisolvent is denser (Chong et

Some authors have modelled the behavior of an organic solvent droplet, considering different local mass transfer both at subcritical (Werling & Debenedetti, 1999) and supercritical (Werling & Debenedetti, 2000) conditions. In this case, the droplet is considered to be stagnant. Therefore, the only convective motion considered is that induced by the diffusion. At subcritical conditions, calculations showed that there is an initial period of droplet swelling, due to the diffusion of CO2 into the organic solvent. Droplet lifetime decreases as the pressure increases, and increases sharply at near-critical conditions, because diffusivities tend to zero near the critical point. At miscible conditions, mass transfer is much faster than at subcritical conditions. Droplet diameter increases if the density of the organic solvent is higher than that of the CO2, and vice versa. However, this ideal and local approach is often not enough to interpret the results. Elvassore et al. developed a model based on the mass transfer simulations of Werling and Debenedetti. This model included

Pérez de Diego et al. and Martin et al. developed both models for the evaporation of dichloromethane (Pérez de Diego et al., 2006) and ethanol (Martín et al., 2007) droplets respectively which accounted for the higher mass transfer coefficients due to the convective

Shekunov et al proposed a simplified approach based on the calculation of different characteristic times (diffusion, jet break up...). They studied the phenomena of turbulent dispersion and micromixing in supercritical carbon dioxide using paracetamol as a model drug compound. They tried to describe the effect of mass-transfer on the particle size and morphology and suggested that particle growth is the time-limiting step (Shekunov et al.,

Then, in order to describe the mass transfer, the drop size distribution, nucleation and particle growth during the drying of the drops as well as the fluid dynamics of the dispersed liquid must be known. Other works also carried out a complete modelling. In these cases, the overall process is modelled, taking into account thermodynamic, hydrodynamic, crystallization and mass transfer aspects (Cardoso et al., 2008; Martin & Cocero, 2004; Lora

At the University of Cádiz, Tenorio et al, by determining the thermodynamic properties of the phases involved in the process, and applying empirical equations (operations with dimensionless numbers), have estimated the different disintegration regimes of the jet when an N-methyl-pyrrolidone (NMP)-ampicillin solution was injected into the CO2-pressurized chamber. The application of the empirical hydrodynamics model proved the existence of significant mechanisms that stabilize the liquid jet, and it showed that there were limiting hydrodynamic conditions that had to be overcome to drive the process toward the formation of uniform spherical nanoparticles and the achievement of higher yields (Tenorio

Reverchon et al., in some recent papers (Reverchon et al., 2007, 2008b, 2008c, 2011) studied the link between SAS morphologies and the relative position of the SAS operating point with respect the mixture critical point of the solvent-CO2 mixtures. It was proposed several mechanisms and their interactions to elucidate the different morphologies and dimensions of precipitates. From a practical point of view, the knowledge of the competing mechanisms allows to select the dimensions of the precipitated particles. If nanoparticles are the objective of the process, low concentrations of the liquid solution are preferable and SAS operation

the solute in mass transfer calculations (Elvassore et al., 2004).

motion of CO2, explaining the change in particle morphology.

al., 2009b).

1999).

et al., 2009)

et al., 2000; Reverchon et al., 2010).

should be performed at completely developed supercritical conditions (when surface tension vanishes before jet break up occurs). If spherical microparticles are the target, the process conditions in which jet break-up produces micrometric droplets are the right ones; the increase in concentration of the starting solution will increase the average diameter of the particles, but, also their polydispersity.
