Supercritical antisolvent micronization of cyclodextrins

Available online at www.sciencedirect.com

(2008) 239–246www.elsevier.com/locate/powtec

Powder Technology 183

Supercritical antisolvent micronization of cyclodextrins

Iolanda De Marco, Ernesto Reverchon ⁎

Università degli studi di Salerno, Dipartimento di Ingegneria Chimica e Alimentare, Via Ponte Don Melillo, Fisciano (SA), Italy

Received 29 December 2006; received in revised form 25 May 2007; accepted 23 July 2007Available online 9 August 2007

Abstract

Micronization of α- and β-cyclodextrins solubilized in dimethylsulfoxide (DMSO) has been successfully performed using the SupercriticalAntiSolvent (SAS) precipitation. We obtained sub-microparticles, microparticles and expanded microparticles of both cyclodextrins, ranging fromabout 0.1 to 11 μm, varying the concentration of the liquid solution from 5 to 200 mg/mL, process temperature (40–60 °C) and pressure (90–180 bar). Particularly, we observed for both materials that, increasing the concentration of the liquid solution, decreasing the pressure or increasingthe temperature, the mean particle size increased and the particle size distribution enlarged.

We also tried to relate the morphologies obtained to the position of the process operating point with respect to the mixture critical point (MCP)of the ternary system cyclodextrin–DMSO–CO2.© 2007 Elsevier B.V. All rights reserved.

Keywords: Supercritical AntiSolvent process; Cyclodextrins; Vapor–liquid equilibria; Micronization

1. Introduction

Supercritical AntiSolvent precipitation (SAS) has been usedto micronize several kinds of compounds and some reviewshave been published, in which the results obtained on differentcompounds have been illustrated [1–5]. The results can be quitedifferent, depending on the process mode (batch or semi-continuous), on the nature of the material and on the high-pressure vapor liquid equilibria (VLEs) characterizing theternary system solvent–solute–supercritical antisolvent. Crys-tals, nanoparticles, spherical sub-micro and microparticles withmean diameters ranging from 0.1 μm to several tenths ofmicrons and empty shells are the morphologies that have beenmost frequently observed [1–5].

Some studies on the interplay between spray formation, masstransfer and SAS performance have been proposed [4,6,7].Other aspects of the SAS process have been relatively lessstudied, such as the VLEs of the ternary system involved in theprocess and the corresponding morphologies and dimensions ofthe precipitated powders [8,9]. The phase behavior of a binarymixture containing supercritical CO2 and an organic solvent can

⁎ Corresponding author. Fax: +39 89 964057.E-mail address: [email protected] (E. Reverchon).

0032-5910/$ - see front matter © 2007 Elsevier B.V. All rights reserved.doi:10.1016/j.powtec.2007.07.038

be modified by the addition of a low volatility third component.When a mixture of two components solvent + solute is bettersoluble in a supercritical solvent than each of the pure com-ponents alone, the so-called cosolvency effect can occur. If aternary system shows poorer solubility compared to the binarysystems antisolvent+solvent and antisolvent+solute, it is callednon-cosolvency (antisolvent) system [10,11].

In some recent papers [12–14], our research group attemptedto relate VLEs to the morphology of SAS precipitated materials.For example, we discussed the connection between theobserved morphologies and the position of the processoperating point with respect to the mixture critical point(MCP); i.e., the pressure, in an isothermal pressure vs. carbondioxide molar fraction diagram, at which a supercritical phasecan exist. In particular, we studied the modifications of thebinary system dimethylsulfoxide (DMSO)–CO2 induced by thepresence of a third compound [12–14]. In the case of SASprecipitation of yttrium acetate from DMSO [12], VLE behaviorseems to be substantially the same as the corresponding binarysystem and differs only for the movement of the MCP towardshigher pressures. In the case of cefonicid, cefoperazone andcefuroxime (some cephalosporinic antibiotics) precipitation[13,14], a substantial modification of the binary VLE behaviourwas observed.

Table 1SAS experiments performed on α-cyclodextrin at different operating conditions,organized by the SAS process parameters considered.

# P [bar] T [°C] c a [mg/mL] SEM results

12 150 40 10 Sub-microparticles11 3015 504 100 Microparticles6 150 Microparticles and balloons7 2004 150 40 100 Microparticles13 5014 608 90 40 100 Crystals9 120 Microparticles4 15010 180a c=mgCD/mLorganic solvent.

240 I. De Marco, E. Reverchon / Powder Technology 183 (2008) 239–246

Cyclodextrins (CDs) are cyclic oligosaccharides, consistingof glucopyranose units, which are linked by 1,4 glycosidicbonds and have a toroidal or a truncated cone shape. They arewater-soluble since all the hydroxyl groups of the glucopyr-anose units are located on the outside surface of the rings.However, the internal cavity of the doughnut-shaped moleculeis relatively non-polar. This unique characteristic of cyclodex-trins allows a guest molecule with an appropriate size to beincluded fully or partially in the hydrophobic cavity. During theguest-CD complex formation no covalent bonds are establishedor broken. The formation of inclusion complexes has been usedto improve the solubility, stability and bioavailability of a widevariety of pharmaceuticals, such as poorly water-soluble drugs[15,16]. The most common cyclodextrins are α-, β- and γ-cyclodextrins, which consist of six, seven and eight glucopyr-anose units, respectively. Among them, α-cyclodextrin iswidely used in food, pharmaceutical and perfumery industries;it is used to stabilize medications, rid of undesirable odour andtaste. β-cyclodextrin is used in food formulations for flavorprotection or flavor delivery, in the pharmaceutical field toimprove the solubility of the drugs in water, in the agricultural

Table 2SAS experiments performed on β-cyclodextrin at different operating conditions,organized by the SAS process parameters considered

# P [bar] T [°C] c a [mg/mL] SEM results

9 150 40 30 Sub-microparticles1 505 75 Microparticles2 1003 150 Microparticles and expanded microparticles4 2002 150 40 100 Microparticles14 5013 606 90 40 100 Microparticles7 1202 1508 180a c=mgCD/mLorganic solvent.

industry to form complexes with herbicides, insecticides,fungicides, repellents, pheromones and growth regulators.

The micronization of a complex CD-drug can be useful toenhance the interaction between the pharmaceutical compoundand the CD. Micronized CDs can be effective in improving theaerodynamic performance of respiratory active pharmaceutical

Fig. 1. SEM images taken at the same enlargement of α-CD precipitated fromDMSO at 150 bar, 40 °C at different concentrations of the liquid solution:a) 10 mg/mL; b) 100 mg/mL; c) 150 mg/mL.

241I. De Marco, E. Reverchon / Powder Technology 183 (2008) 239–246

ingredient particles [17]. The use of CDs as hygroscopic growthinhibitors in drug formulations for lung delivery has beenproposed [18]. The reduction of moisture sorption may improvethe stability of inhalable formulations towards hygroscopicgrowth during storage and in the lung region. The reduction ofthe particle growth favors the deposition of the formulation inthe deep region of the lungs and drug bioavailability [18].

SCFs have been applied in some cases to the preparation ofdrug/CDs complexes. The most used process has been theimpregnation of CDs particles using supercritical CO2 (SC-CO2),or SC-CO2 modified with organic solvents [19–29]. The key ideaof this process is to dissolve the drug of interest in supercriticalCO2 (in some cases, a cosolvent is used to increase the solubilityof the drug in the supercritical fluid), followed by permeation ofthe supercritical solution into the pores of the carrier and byprecipitation of the drug inside the pores.

In a few works, drug/CD complexes preparation has beenattempted by Supercritical AntiSolvent precipitation (SAS)based processes. In some cases [30–32], a systematic study onthe effect of the process parameters on the morphology andparticle size distribution of the powders is missing.

Mammucari et al. [33] performed the precipitation ofhydroxypropyl- and methyl-β-CDs alone and the coprecipita-tion of one of these CDs with naproxen; but, they obtainedhighly agglomerated particles, probably due to the selectedprocess conditions. Rodier et al. [34] used a three-step processbased on supercritical fluids, but, they obtained large crystals ofeflucimibe with γ-CD. In all cases, the complex formation wasonly partly successful.

To our knowledge, a successful study on the supercriticalantisolvent precipitation to produce micrometric CDs alone (forexample, for pulmonary delivery) or to preliminary find theprocess conditions to perform the proper complex formationand micronization, has not been performed.

Therefore, the scope of this work is to evaluate the effect ofthe SAS process parameters, such as concentration of the liquid

Fig. 2. Particle size distributions of α-CD powders precipitated at 150 bar and 40 °C.terms of number of particles; b) the cumulative distributions are evaluated in terms

solution, pressure and temperature on micronization of α- andβ-cyclodextrins from DMSO.

We will try to relate the different morphologies, the trend ofthe mean particle size and particle size distributions to the VLEsof the systems involved.

2. Experimental apparatuses, materials, procedures andmethods

2.1. Apparatus

The SAS laboratory apparatus used in this work consists ofan HPLC pump equipped with a pulse dampener (Gilson, mod.805) used to deliver the liquid solution, and a diaphragm high-pressure pump (Milton Roy, mod. Milroyal B) used to deliversupercritical CO2. A cylindrical vessel with an internal volumeof 500 mL is used as the precipitation chamber. The liquidmixture is delivered to the precipitator through a 200 μmstainless steel nozzle. Nozzles with other diameters are alsoavailable. A second collection chamber located downstream theprecipitator is used to recover the liquid solvent. Furtherinformation and a schematic representation of the apparatushave been given elsewhere [1,13,35].

2.2. Materials and methods

α- and β-cyclodextrins with purities of 99.9% and DMSOwith purity of 99% were purchased from Sigma-Aldrich. CO2

(purity 99%) was purchased from SON (Naples, Italy). Theapproximate saturation concentrations of cyclodextrins inDMSO were measured at room temperature, by adding theorganic solvent drop to drop in a certain quantity of solid, untilthe solute was dissolved in DMSO; both the CDs are verysoluble in DMSO, their solubilities are approximately equal to750 mg/mL. The untreated materials were irregular crystalswith a particle size ranging between 20 and 100 μm, in the case

Effect of concentration in the liquid solution. a) the distributions are evaluated inof volume.

Fig. 3. SEM images of β-CD precipitated from DMSO at 150 bar, 40 °C atdifferent concentrations of the liquid solution: a) 50 mg/mL; b) 100 mg/mL;c) 200 mg/mL.

242 I. De Marco, E. Reverchon / Powder Technology 183 (2008) 239–246

of α-cyclodextrin, and between 30 and 150 μm, in the case ofβ-cyclodextrin. All materials were used as received.

Samples of the precipitated powder were observed using aScanning Electron Microscope (SEM) (Assing, mod. LEO420). SEM samples were covered with 250 Å of gold using asputter coater (Agar, mod. 108A). The particle size (PS) and theparticle size distributions (PSDs) were measured using an imageanalysis performed with Sigma Scan Pro software (JandelScientific), an image processing program to count, measure andanalyze digital images; about 1000 particles were considered ineach PSD calculation. The number of particles was chosen inagreement with the criteria used in the image analyses ofpharmaceutical powders [36]: 500 to 1500 measured particlesrepresent a good compromise between the time spent for theanalysis and the accuracy of results.

2.3. Experimental procedures

A SAS experiment begins by delivering supercritical CO2 tothe precipitation chamber until the desired pressure is reached.Antisolvent steady flow is established; then, pure solvent is sentthrough the injector to the pressurized chamber with the aim ofobtaining steady state composition conditions during the soluteprecipitation. At this point, the flow of the organic solvent isstopped and the liquid solution formed by the organic solventand the CD is delivered through the nozzle. Once injected thefixed quantity of organic solution, the pump for liquids isstopped. However, supercritical CO2 continues to flow to washthe chamber from the residual content of liquid solubilized inthe supercritical antisolvent. If the final purge with pure CO2 isnot performed, solvent condenses during the depressurizationand can solubilize or modify the powder. More details havebeen given elsewhere [37].

3. Results and discussion

The range of operating conditions used in this work wasinitially selected on the basis of previous experiences on the SASprocess [12,35]. The pressure was fixed at 150 bar and thetemperature at 40 °C, and the first set of experiments was per-formed, varying the concentration of the CD in the liquid solution(see Tables 1 and 2). The liquid solution flow rate was fixed at1.2 mL/min DMSO, the ratio between CO2 flow rate and liquidflow rate (R) was set equal to 30 on amass basis and theCO2molarfraction was equal to 0.98. In each experiment, 20 mL of organicsolution were injected; therefore, a quantity of powder rangingbetween 400 mg and 4 g of CD was produced, depending on theconcentration of the solution. Series of experiments wereperformed varying one parameter at a time: the solute concentra-tion in the organic solution, the pressure and the temperature.

3.1. Effect of liquid solution concentration

The first set of experiments was performed on α-CD, varyingthe concentration from 10 to 200 mg/mL DMSO, at a fixedpressure of 150 bar and a temperature of 40 °C. A summary ofthese experiments is reported in the first part of Table 1.

SEM analyses of the powders precipitated in these experi-ments show that α-CD precipitated in the form of sub-microparticles for concentrations up to 50 mg/mL in DMSO(see Fig. 1a), in the form of microparticles for a concentration of100 mg/mL (Fig. 1b), and of microparticles and micrometricballoons at concentrations higher than 150 mg/mL (Fig. 1c).

Fig. 5. Cumulative particle size distributions in terms of volume of particles ofα-CD powders obtained at 40 °C and 100 mg/mL. Effect of pressure.

Fig. 4. Particle size distributions of β-CD powders precipitated at 150 bar and 40 °C. Effect of concentration in the liquid solution. a) the distributions are evaluated interms of number of particles; b) the cumulative distributions are evaluated in terms of volume.

243I. De Marco, E. Reverchon / Powder Technology 183 (2008) 239–246

These balloons are smooth empty shells with some holes on thesurface.

From the elaboration of SEM images information, weobtained the PSDs of α-CD powders, as reported in Fig. 2.PSDs were calculated in terms of particle number percentage(Fig. 2a) and of volume cumulative percentage (Fig. 2b).

By increasing the concentration of α-CD in DMSO from 10to 200 mg/mL, the mode increases (from 0.25 μm at 10 mg/mLto 0.90 μm at 200 mg/mL) and particle size distributionsenlarge, showing particles up to a maximum diameter of about8 μm. To represent the distributions in Fig. 2a at all theconcentrations in a single graph, it has been necessary to breakthe abscissa axis, because the distributions at 150 and 200 mg/mL show long tails that end at about 8 μm. In Fig. 2b, werepresented the percentages in volume of particles that are mostrepresentative in the pharmaceutical field, because they give anaccount of the quantity (mass) of drug in a fixed range ofdimensions. In this case, to represent all the curves in a singlegraph, we used a logarithmic scale on the abscissa.

We evaluated the effect of concentration on particlesmorphology also in the case of β-CD, again fixing the pressureat 150 bar and temperature at 40 °C and varying theconcentration of the liquid solution from 30 to 200 mg/mL inDMSO (as it is possible to see in the first part of Table 2).

β-CD precipitated in form of sub-microparticles, in corre-spondence of liquid concentrations up to 50 mg/mL (Fig. 3a). At75 and 100 mg/mL, sub-microparticles and microparticlesranging from 0.2 to 3 μm (Fig. 3b) were obtained. At 150 and200 mg/mL, microparticles and expanded microparticles(ranging from 5 to 20 μm) were obtained (Fig. 3c). Theexpanded microparticles obtained in the case of the β-CD aredifferent from the balloons obtained for the α-CD (the onesshown in Fig. 1c), because they presented a porous internalstructure. The images reported in Fig. 3 have not been taken atthe same enlargement; but, they are, in any case, representativeof the different morphologies obtained. PSDs were calculated in

terms of particle number percentage (Fig. 4a) and of volumecumulative percentage (Fig. 4b).

In some previous SAS works [13,35,38,39], it was observedthat the most common effect of the concentration of the liquidsolution is that it produces an increase of the mean PS and anenlargement of the PSD. However, for some materials [13,14],we observed that an increase of concentration produces avariation in the morphology of precipitates: sub-microparticleswere obtained at lower concentrations; whereas, balloons(empty shells with a smooth surface) were obtained at higherconcentrations. To explain the results obtained, we tried to relatethe operating point with the VLEs of the ternary systemsinvolved; but, in the literature, only binary data for the systemDMSO–CO2 are available.

Operating at different concentrations, the process operatingpoint in a pressure vs. carbon dioxide molar fraction diagram is

Fig. 6. Cumulative particle size distributions in terms of volume of particles ofβ-CD powders obtained at 40 °C and 100 mg/mL. Effect of pressure.

244 I. De Marco, E. Reverchon / Powder Technology 183 (2008) 239–246

always the same, but VLEs curves may change. Indeed,increasing the concentration, the affinity between the solventand the supercritical carbon dioxide can be reduced and the twophases region can enlarge. At lower concentrations (10 mg/mL),the operating point can be well above the MCP of the system,therefore α-CD precipitates from the supercritical phase in formof sub-microparticles. Indeed, at these conditions, probably nosurface tension operates in the fluid phase and, therefore, theprecipitation of solids is obtained from a homogeneoussupercritical phase [40].

At 75 and 100 mg/mL, the operating point lies in theproximity of the transition between supercritical and subcriticalconditions and microparticles are produced.

At higher concentrations, the operating point lies in the onephase subcritical region and expanded microparticles areproduced. Indeed, the presence of a non-zero surface tension

Fig. 7. Cumulative particle size distributions in terms of volume percentage of a) α-C

produces the formation of liquid droplets that are the geometricalbasis of the observed microparticles. When droplets areexpanded by the slower transfer of the gas inside the liquid,balloons are obtained [41,42].

3.2. Effect of pressure and temperature

Once evaluated the effect of the concentration in the liquidsolution, the temperature was set at 40 °C, the concentration ofthe organic solution at 100 mg/mL and the pressure was variedfrom 90 to 180 bar.

In the case of α-CD, as it can be observed from Table 1,different morphologies were obtained. In particular, at 90 bar,α-CD precipitated in form of crystals, at 120 bar, we obtainedmicrometric particles, spherical in shape but with a partly roughsurface. Increasing the pressure at 150 and 180 bar, micrometricand sub-micrometric particles were obtained.

In Fig. 5, the particle size distributions of the powdersobtained varying the pressure from 120 to 180 bar arerepresented. The curves related to 150 and 180 bar are similar:the mean particle size lies in the range 0.3–0.35 μm and thelarger particles have a diameter of 2 μm. At 120 bar, the meanparticle size increases (about 1 μm) and the largest particlesreach the dimension of 4 μm in diameter.

The effect of pressure was evaluated also for β-CD. Inparticular, at 90 bar, microparticles were obtained with a meandiameter equal to 4.7 μm, whereas from 120 to 180 bar,microparticles with the mode of the distributions varying from0.37 μm at 120 bar to 0.23 μm at 180 bar were obtained. Thecumulative distributions are reported in Fig. 6, in terms ofvolume.

The last sets of experiments have been performed on α- andβ-CDs at 150 bar, 100 mg/mL DMSO and varying thetemperatures from 40 to 60 °C. For both the materials, weobserved that, increasing the temperature, the mean particle sizeincreased and the PSD enlarged. The effect of temperature on α-and β-CDs powders has been shown in Fig. 7.

D; b) β-CD powders precipitated at 150 bar, 100 mg/mL. Effect of temperature.

245I. De Marco, E. Reverchon / Powder Technology 183 (2008) 239–246

The results obtained at different pressures and temperaturescan be explained in an analogous manner with respect to thoseobtained varying the concentration of the liquid solution. In apressure vs. carbon dioxide mole fraction diagram, theoperating point changes but the VLEs are always the same, ifthe concentration of the liquid solution has been fixed. In thecase of α-CD, at 90 bar, the operating point lies in the twophases region and the material precipitates from a liquid reachphase, producing crystals. At 120 bar, we obtained particles,spherical in shape but with a partly rough surface. It is hypo-thesizable that the operating point lies near the MCP and theprecipitated particles represent a transition morphology fromcrystals to microparticles. At 150 and 180 bar, the precipita-tion occurred from a supercritical phase and the particles arespherical.

In the case of β-CD, the explanation is similar; the onlydifference of behavior is in the test performed at 90 bar. In thiscase, the particles are spherical but the mean diameter is welllarger than the mean diameter of the particles obtained at higherpressures. Probably, the operating point lies in the subcriticalone phase region.

The effect of the temperature may be explained consideringhow the VLEs changed with temperature. Indeed, increasing thetemperature, the MCP shifts towards higher pressures. There-fore, an operating point that, at 40 °C, lies in the supercriticalregion may lie in the subcritical region with respect to the VLEsof the same system at higher temperatures.

4. Conclusions and perspectives

The α- and β-CDs have been successfully micronized andwe demonstrated that, varying the operating parameters, it ispossible to control the particle size and the particle sizedistribution of the powders. Moreover, different morphologieshave been obtained: crystals, sub-microparticles, microparticlesand expanded microparticles. An attempt at linking the differentmorphologies with the VLE behavior of the binary CO2–organic solvent system has been made. The major interest inusing CDs in pharmaceutical applications is to produce drug-CD complexes to improve the dissolution rate of poorly solubledrugs. From this point of view, the results obtained in this workform the basis for the next step: drug-CD complexes processingby SAS.

Acknowledgments

The authors acknowledge Dr. Daniela Perri for her help inperforming the experiments reported in this work. The financialsupport of MiUR (Italian Ministry of Scientific Research) isalso acknowledged.

References

[1] E. Reverchon, Supercritical antisolvent precipitation of micro- andnanoparticles, J. Supercrit. Fluids 15 (1) (1999) 1–21.

[2] J. Jung, M. Perrut, Particle design using supercritical fluids: literature andpatent survey, J. Supercrit. Fluids 20 (3) (2001) 179–219.

[3] Y. Hakuta, H. Hayashi, K. Arai, Fine particle formation using supercriticalfluids, Curr. Opin. Solid State Mater. Sci. 7 (4-5) (2003) 341–351.

[4] A. Shariati, C.J. Peters, Recent developments in particle design usingsupercritical fluids, Curr. Opin. Solid State Mater. Sci. 7 (4-5) (2003)371–383.

[5] E. Reverchon, R. Adami, Nanoparticles and supercritical fluids,J. Supercrit. Fluids 37 (2006) 1–22.

[6] C.S. Lengsfeld, J.P. Delplanque, V.H. Barocas, T.W. Randolph, Mecha-nism governing microparticle morphology during precipitation by acompressed antisolvent: atomization vs nucleation and growth, J. Phys.Chem., B 104 (12) (2000) 2725–2735.

[7] D.J. Dixon, K.P. Johnston, R.A. Bodmeier, Polymeric materials formed byprecipitation with a compressed fluid antisolvent, AIChE J. 39 (1993)127–139.

[8] Y. Perez, H. Pellikaan, F.E. Wubbolts, G.J. Witkamp, P.J. Jansens,Production of L-PLA microparticles below and above the mixture criticalpressure of the system DCM–CO2, in: I. Kikic, M. Perrut (Eds.),Proceedings of the 9th Meeting on Supercritical Fluids, Trieste (Italy),2004, MD29.pdf.

[9] A. Weber, L.V. Yelash, T. Kraska, Effect of the phase behaviour of thesolvent–antisolvent systems on the gas–antisolvent–crystallisation ofparacetamol, J. Supercrit. Fluids 33 (2) (2005) 107–113.

[10] A.L. Scheidgen, G.M. Schneider, Complex phase equilibrium phenomenain fluid ternary mixtures up to 100 MPa: cosolvency, holes, windows andislands — review and results, Fluid Phase Equilib. 194-197 (2002)1009–1028.

[11] F. Becker, P. Richter, Non-aqueous ternary mixtures with “island”miscibility gaps, Fluid Phase Equilib. 49 (1989) 157–166.

[12] E. Reverchon, G. Caputo, I. De Marco, The role of phase behavior andatomization in the supercritical antisolvent precipitation, Ind. Eng. Chem.Res. 42 (25) (2003) 6406–6414.

[13] E. Reverchon, I. De Marco, Supercritical antisolvent micronization ofCefonicid: thermodynamic interpretation of results, J. Supercrit. Fluids 31(2) (2004) 207–215.

[14] E. Reverchon, I. De Marco, Supercritical antisolvent precipitation ofcephalosporins, Powder Technol. 164 (2006) 139–146.

[15] K.A. Connors, J.A. Mollica, Theoretical analysis of comparative studies ofcomplex formation, J. Pharm. Sci. 55 (1966) 772–780.

[16] C.M. Fernandes, F.J.B. Veiga, The cyclodextrins as modelling agents ofdrug controlled release, J. Incl. Phenom. Macrocycl. Chem. 44 (2002)79–85.

[17] J.M.C. Leite Pinto, H.M. Cabral Marques, Beclomethasone/cyclodextrininclusion complex for dry powder inhalation, STP Pharma Sci. 9 (3)(1999) 253–256.

[18] A. Clark, M.C. Kuo, C. Lalor, Phospholipids, cyclodextrins, starch, andcellulose as hygroscopic growth inhibitors in dry powders for pulmonarydrug delivery, Inhale Therapeutic Systems, Inc., USA, WO, 2000, p. 46.

[19] S. Junco, T. Casimiro, N. Ribeiro, M. Nunes Da Ponte, H.C. Marques,Optimization of supercritical carbon dioxide systems for complexation ofnaproxen: beta cyclodextrin, J. Incl. Phenom. Macrocycl. Chem. 44 (2002)69–73.

[20] S. Junco, T. Casimiro, N. Ribeiro, M. Nunes Da Ponte, H.C. Marques,A comparative study of naproxen–beta cyclodextrin complexes preparedby conventional methods and using supercritical carbon dioxide, J. Incl.Phenom. Macrocycl. Chem. 44 (2002) 117–121.

[21] M. Türk, G. Upper, M. Steurenthaler, Kh. Hussein, M.A. Wahl, Complexformation of ibuprofen and β-cyclodextrin by controlled particledeposition (CPD) using SC-CO2, J. Supercrit. Fluids 39 (2007) 435–443.

[22] Y. Tozuka, T. Fujito, K. Moribe, K. Yamamoto, Ibuprofen–cyclodextrininclusion complex formation using supercritical carbon dioxide, J. Incl.Phenom. Macrocycl. Chem. 56 (2006) 33–37.

[23] M. Charoenchaitrakool, F. Dehghani, N.R. Foster, Utilization ofsupercritical carbon dioxide for complex formation of ibuprofen andmethyl-β-cyclodextrin, Int. J. Pharm. 239 (2002) 103–112.

[24] T. Van Hees, V. Barillaro, G. Piel, P. Bertholet, S. Henry De Hassonville,B. Evrard, L. Delattre, Application of supercritical carbon dioxide for thepreparation of drug-cyclodextrin inclusion compounds, J. Incl. Phenom.Macrocycl. Chem. 44 (2002) 271–274.

246 I. De Marco, E. Reverchon / Powder Technology 183 (2008) 239–246

[25] A.H. Al-Marzouqi, I. Shehatta, B. Jobe, A. Dowaidar, Phase solubility andinclusion complex of itraconazole with β-cyclodextrin using supercriticalcarbon dioxide, J. Pharm. Sci. 95 (2) (2006) 292–304.

[26] S. Lai, E. Locci, A. Piras, S. Porcedda, A. Lai, B. Marongiu, Imazalil/cyclomaltoheptaose (β-cyclodextrin) inclusion complex: preparation bysupercritical carbon dioxide and 13C CPMAS and 1H NMR character-ization, Carbohydr. Res. 338 (2003) 2227–2232.

[27] B. Wang, J. He, D. Sun, R. Zhang, B. Han, Utilization of supercriticalcarbon dioxide for the preparation of 3-hydroxyflavone and β-cyclodextrincomplex, J. Incl. Phenom. Macrocycl. Chem. 55 (2006) 37–40.

[28] T. Toropainen, S. Velaga, T. Heikkilä, L. Matilainen, P. Jarho, J. Carlfors,V.-P. Lehto, T. Järvinen, K. Järvinen, Preparation of budesonide/γ-cyclodextrin complexes in supercritical fluids with a novel SEDS method,J. Pharm. Sci. 95 (10) (2006) 2235–2245.

[29] F. Cristini, M. Delalonde, C. Joussot-Dubien, B. Bataille, Elaboration ofibuprofen microcomposites in supercritical CO2, Proceed. 6th Internat.Symp. Supercrit Fluids, Versailles, France, 2003, PMd39.pdf.

[30] I. Fabing, F. Leboeuf, J. Jung, M. Perrut, WO Patent 02/32462, EP 1 330266, 2002.

[31] N. Foster, R. Mammucari, F. Dehghani, Dense gas antisolvent processingof cyclodextrin from aqueous solutions, Proceed. of 9th Meeting onSupercrit. Fluids, Trieste, Italy, 2004, Md12.pdf.

[32] S.-J. Hwang, J.-S. Kim, M.-S. Kim, H.J. Park, S. Lee, Inclusion complexof simvastatin with hydroxypropyl-b-cyclodextrin using supercriticalantisolvent (SAS) process: in vitro and in vivo evaluation, Proceedingsof 8th International Symposium on Supercritical Fluids, Kioto, Japan,2006, PA-1-46.pdf.

[33] R. Mammucari, F. Dehghani, N.R. Foster, Dense gas processing of micron-sized drug formulations incorporating hydroxypropylated and methylatedbeta-cyclodextrin, Pharm. Res. 23 (2) (2006) 429–437.

[34] E. Rodier, H. Lochard, M. Sauceau, J.-J. Letourneau, B. Freiss, J. Fages, Athree step supercritical process to improve the dissolution rate ofEflucimibe, Eur. J. Pharm. Sci. 26 (2005) 184–193.

[35] E. Reverchon, I. De Marco, G. Della Porta, Rifampicin microparticlesproduction by supercritical antisolvent precipitation, Int. J. Pharm. 243(1-2) (2002) 83–91.

[36] M.-N. Pons, H. Vivier, V. Delcour, J.-R. Authelin, L. Pailleres-Hubert,Morphological analysis of pharmaceutical powders, Powder Technol. 128(2-3) (2002) 276–286.

[37] E. Reverchon, G. Della Porta, A. Di Trolio, S. Pace, Supercriticalantisolvent precipitation of nanoparticles of superconductor precursors,Ind. Eng. Chem. Res. 37 (1998) 952–958.

[38] W.Z. He, Q.L. Suo, Z.H. Jiang, S.A., H.L. Hong, Precipitation ofephedrine by SEDS process using a specially designed prefilmingatomizer, J. Supercrit. Fluids 31 (1) (2004) 101–110.

[39] E. Reverchon, I. De Marco, G. Caputo, G. Della Porta, Pilot scalemicronization of amoxicillin by Supercritical Antisolvent precipitation,J. Supercrit. Fluids 26 (1) (2003) 1–7.

[40] C.S. Lengsfeld, J.P. Delplanque, V.H. Barocas, T.W. Randolph, Mecha-nism governing microparticle morphology during precipitation by acompressed antisolvent: atomization vs nucleation and growth, J. Phys.Chem., B 104 (2000) 2725–2735.

[41] J.O. Werling, P.G. Debenedetti, Numerical modelling of mass transfer inthe supercritical antisolvent process, J. Supercrit. Fluids 16 (2) (1999)167–181.

[42] J.O. Werling, P.G. Debenedetti, Numerical modelling of mass transfer inthe supercritical antisolvent process: miscible conditions, J. Supercrit.Fluids 18 (1) (2000) 11–24.