MICROPARTICLE PRODUCTION FOR DRUG CONTROLLED

RELEASE BY SUPERCRITICAL ASSISTED ATOMIZATION

Alessandra Antonacci

Unione Europea UNIVERSITÀ DEGLI

STUDI DI SALERNO

Fondo sociale europeo Programma Operativo Nazionale 2000/2006

“Ricerca Scientifica, Sviluppo Tecnologico, Alta Formazione” Regioni dell’Obiettivo 1 – Misura III.4

“Formazione superiore ed universitaria”

Department of Chemical and Food Engineering

Ph.D. Course in Chemical Engineering (V Cycle-New Series)

MICROPARTICLE PRODUCTION FOR DRUG CONTROLLED RELEASE BY

SUPERCRITICAL ASSISTED ATOMIZATION

Supervisor Ph.D. student Prof. Ernesto Reverchon Alessandra Antonacci Scientific Referees Dr. Francois Cansell Prof. Libero Sesti Osséo Ph.D. Course Coordinator Prof. Paolo Ciambelli

Papers produced during this work

International journals: 1) E. Reverchon, G. Della Porta, A. Spada, A. Antonacci “Griseofulvin micronization and dissolution rate improvement by supercritical assisted atomization”, Journal of Pharmacy and Pharmacology, 56, pp. 1379-1387, 2004. 2) E. Reverchon, A. Antonacci “Cyclodextrins Micrometric Powders Obtained by Supercritical Fluid Processing”, Biotechnology & Bioengineering, 94, pp. 753 - 761, 2006.

3) E. Reverchon, A. Antonacci “Chitosan Microparticles Production by Supercritical Fluid Processing”, Industrial & Engineering Chemistry Research, 45, pp. 5722-5728, 2006.

4) E. Reverchon, A. Antonacci “Polymer Microparticles Production by Supercritical Assisted Atomization”, The Journal of Supercritical Fluids, 39, pp. 444-452, 2007. 5) E. Reverchon, A. Antonacci “Drug-polymer microparticles produced by Supercritical Assisted Atomization”, Biotechnology & Bioengineering, 2006. In Press.

Proceedings of International Conferences: 6) E. Reverchon, A. Antonacci “Cyclodextrins Micronization by Supercritical Assisted Atomization”, Proceedings of the 10th European Meeting on Supercritical Fluids. Strasbourg/Colmar (France), 12-14 December, 2005.

7) A. Antonacci, R. Adami, G. Della Porta, E. Reverchon Supercritical Assisted Atomization of Corticosteroids. Proceedings of the 8th Conference on Supercritical Fluids and Their Applications. Ischia (Italy), 28-31 May 2006. 8) E. Reverchon, A. Antonacci Supercritical CO2 assisted micronization of chitosan. Proceedings of the 17th International Congress of Chemical and Process Engineering. Prague (Czech Republic), 27-31 August 2006.

INDEX

ABSTRACT................................................................................................ XI

INTRODUCTION.........................................................................................1

STATE OF THE ART ..................................................................................5 II.1 POLYMER/DRUG MICROPARTICLES

PRODUCTION.....................................................................5 II.1.1 Traditional Processes For Microcapsules and

microspheres production .......................................................5 II.1.2 Supercritical Fluids based Processes For

microcapsules and microspheres production.......................10 II.2 MICROPARTICLES OF DRUG/CICLODEXTRIN

INCLUSION COMPLEXES ..............................................15 II.2.1 Conventional Processes for Inclusion Complexes

Production ...........................................................................19 II.2.2 Paste Method.......................................................................20 II.2.3 Supercritical Fluids based Processes for inclusion

complexes production .........................................................22

AIM OF THE Ph.D. THESIS.....................................................................25

SUPERCRITICAL ASSISTED ATOMIZATION................. ..................27 IV.1 EXPERIMENTAL APPARATUS......................................27 IV.2 PROCESS DESCRIPTION ................................................31 IV.2.1 Process parameters ..............................................................33

ANALYTICAL METHODS.......................................................................37 V.1 SCANNING ELECTRON MICROSCOPE (SEM)............37 V.2 PARTICLE SIZE DISTRIBUTIONS.................................38 V.3 DIFFERENTIAL SCANNING CALORIMETER

(DSC) ..................................................................................39 V.4 X-RAY................................................................................40 V.5 UV-VISIBLE SPECTROSCOPY.......................................41 V.6 ENERGY DISPERSIVE MICROANALYSIS ...................42

RESULTS PART I: SAA CARRIERS PROCESSING ...........................45

II

VI.1 SYNTHETIC POLYMERS MICROPARTICLES PRODUTION: PMMA AND PLLA.................................. 46

VI.1.1 Results and Discussion....................................................... 46 VI.2 NATURAL POLYMER MICROPARTICLES

PRODUCTION: CHITOSAN............................................ 56 VI.2.1 Results and Discussion....................................................... 58 VI.3 CYCLODEXTRIN MICROPARTICLES

PRODUCTION .................................................................. 68 VI.3.1 Results And Discussion...................................................... 69

RESULTS PART II: COPRECIPITATION ........................................... 83 VII.1 PMMA-MEPA ................................................................... 83 VII.1.1 Medroxyprogesterone acetate (MEPA).............................. 83 VII.1.2 PMMA and MEPA coprecipitation.................................... 84 VII.2 CHITOSAN-AMPICILLIN ............................................... 90 VII.2.1 Ampicillin........................................................................... 90 VII.2.2 Chitosan-Ampicillin coprecipitation .................................. 91

CONCLUSIONS AND FUTURE DEVELOPMENTS ......................... 127

REFERENCES ......................................................................................... 129

INDEX OF FIGURES

Figure II.1: The conformation of the glucose units in the cyclodextrin places the hydrophilic hydroxyl groups at the top and bottom of the three dimensional ring and the hydrophobic glycosidic groups on the interior. ...16

Figure II.2: Forming an inclusion complex involves multiple interactions between active, solvent and cyclodextrin. .....................................................17

Figure II.3: Type of phase solubility diagram. ..............................................19

Figure IV.4: Schematic representation of the SAA apparatus: 1) CO2 cylinder; 2) liquid solution; 3) N2 cylinder; 4) cooling bath; 5) heating bath; 6) high pressure pumps; 7) dampener; 8) heat exchanger; 9) saturator; 10) precipitator; 11) condenser...........................................................................27

Figure IV.5: Photograph of the helicoidal flux conveyor and the collection filter in the first configuration of the precipitator. ........................................29

Figure IV.6: Second configuration of the precipitator: (a) Head; (b) central body; (c) cylindrical filter; (d) collection vessel; (e) gas conveyor. .............30

Figure IV.7: Third configuration of the precipitator: a) central body, b) filter. ..............................................................................................................31

Figure IV.8: (a) pneumatic atomization; (b) decompressive atomization.....32

Figure IV.9: Schematic representation of microsphere production by SAA process...........................................................................................................33

Figure IV.10: Effect of the process parameter on the miscibility hole. The continuous line refers to the binary system DMSO-CO2; the dash line refers to the ternary system DMSO-CO2-solute. .....................................................34

Figure V.1: Schematic diagram of a Scanning Electron Microscope. ..........36

Figure V.2: Schematic functioning of laser diffractrometer. ........................37

Figure V.3: Schematic representation of DSC furnace. ................................38

IV

Figure V.4: Schematic diagram of an X-ray equipment............................... 39

Figure V.5: Functioning of Cary 50 Scan.................................................... 40

Figure V.6: Generation of X-ray by an excited atom................................... 41

Figure V.7: Cutway of a Si X-ray Detector.................................................. 42

Figure VI.1: SEM images of PMMA precipitated by SAA from acetone at different mass flow ratios (R), at Csol=50 mg/mL and 60°C in the precipitator................................................................................................... 46

Figure VI.2: PSDs of micronized PMMA at different R. Calculations in terms of particle volume percentages..................................................................... 47

Figure VI.3: SEM images of PLLA precipitated by SAA from DCM at different mass flow ratios (R), at Csol=20 mg/mL and 57°C in the precipitator................................................................................................... 48

Figure VI.4: PSDs of micronized PLLA at different R. Calculations in terms of particle volume percentages..................................................................... 49

Figure VI.5. Vapour-liquid equilibria for the system CO2-acetone: (○) at 30°C; (▲) at 35°C; (■) at 40°C [40]. (▬) Vapour-liquid equilibrium curve at 80°C using SRK EoS. Points A, B, C and D are referred to R-values of 0.9, 1.0, 1.2 and 1.6, respectively................................................................. 50

Figure VI.6: SEM image of PLLA precipitated by SAA from DCM at a precipitator temperature of 67°C................................................................. 51

Figure VI.7: SEM images of micronized PMMA obtained at different Csol values, at R=1.2 and 60°C in the precipitator............................................. 52

Figure VI.8: PSDs of micronized PMMA at different Csol values. Calculations in terms of particle volume percentages. ................................ 53

Figure VI.9: SEM image of chitosan microparticles precipitated by SAA from 1% acetic acid aqueous solution at R=1.8, Csol = 5 mg/mL and Tp=110°C..................................................................................................... 56

Figure VI.10: SEM images of chitosan microparticles precipitated by SAA from 1% acetic acid aqueous solution at precipitation temperatures of 87 °C (a), 95°C (b) and 135°C (c) (R=1.8, Csol = 5 mg/mL)................................ 59

Figure VI.11: SEM images of chitosan microparticles precipitated by SAA from 1% acetic acid aqueous solution at concentrations and precipitation temperatures of 1 mg/mL and 95°C (a), 5 mg/mL and 95°C (b) and 10 mg/mL and 106°C (c), respectively.............................................................. 60

Figure VI.12: PSDs in terms of number of particles percentages of micronized chitosan at C = 1 mg/mL and Tp = 95°C (□), C = 5 mg/mL and Tp = 95°C (●) and C = 10 mg/mL and Tp = 106°C (∆)................................ 61

V

Figure VI.13: PSDs in terms of particle volume percentages of micronized chitosan at C = 1 mg/mL and Tp = 95°C (□), C = 5 mg/mL and Tp = 95°C (●) and C = 10 mg/mL and Tp = 106°C (∆). .................................................62

Figure VI.14: X-Ray diffraction patterns of untreated and SAA processed CH at different temperatures in the precipitator.................................................63

Figure VI.15: DSC curves of untreated and SAA processed CH at different temperatures in the precipitator (Tp).............................................................64

Figure VI.16: SEM images of α-CD microparticles precipitated by SAA at temperatures of 108 °C (a), 118 °C (b) and 186 °C (c) in the precipitator (Csol =50 mg/mL)...........................................................................................68

Figure VI.17: SEM images of α-CD microparticles precipitated by SAA at solute concentrations of 20 mg/mL (a), 50 mg/mL (b) and 130 mg/mL (c) in the liquid solution (T=118°C in the precipitator).........................................70

Figure VI.18: Histograms in terms of number of particles percentages of micronized α-CD at C = 20mg/mL. Comparison between PSDs obtained using LS (■) and SEM (□) analysis. Expanded X-axis below 3 µm. .............71

Figure VI.19: PSDs in terms of number of particles and volume percentages of micronized α-CD at C = 20mg/mL. Comparison between PSDs obtained using LS (—) and SEM (– –) analysis...........................................................72

Figure VI.20: SEM image at magnification of 3K reporting a landscape of α-CD microparticles precipitated at C=130 mg/mL. .......................................73

Figure VI.21: PSDs in terms of particle volume percentages of micronized α-CD at different C values, obtained by LS. .....................................................75

Figure VI.22: SEM images of HP-β-CD microparticles precipitated by SAA at solute concentrations of 20 mg/mL (a), 50 mg/mL (b) and 100 mg/mL (c) in the liquid solution (T=118°C in the precipitator).....................................76

Figure VI.23: PSDs in terms of particle volume percentages of micronized HP-β-CD at different C values, obtained by LS............................................77

Figure VI.24: X-Ray diffraction patterns (a) and DSC thermograms (b) of untreated and SAA processed α-CD and HP-β-CD......................................78

Figure VII.1: SEM image of MEPA microparticles produced by SAA (adapted from: Williams, 2002)....................................................................81

Figure VII.2: SEM image of PMMA and MEPA coprecipitates...................82

Figure VII.3: PSD of PMMA, MEPA and PMMA/MEPA SAA coprecipitated particles. Calculations in terms of number of particles percentages............83

VI

Figure VII.4: PSD of PMMA, MEPA and PMMA/MEPA coprecipitated particles. Calculations in terms of particle volume percentages. ................ 83

Figure VII.5: DSC thermograms of: a) raw MEPA; b) physical mixture of PMMA microparticles/MEPA (ratio 4/1); c) SAA coprecipitated PMMA/MEPA microparticles; d) PMMA microparticles............................ 84

Figure VII.6: Drug release rate: (∆) raw MEPA; (ο) SAA coprecipitated PMMA/MEPA microparticles. Drug release rates from coprecipitates are reported up to 12 h (left) and to 96 h (right)................................................ 86

Figure VII.7: SEM image of ampicillin microparticles precipitated at Tpr=95°C from 1% acid acetic aqueous solution. ....................................... 88

Figure VII.8: SEM images of CH/ampicillin microparticles coprecipitated by SAA at Tpr=95°C from 1% acid acetic aqueous solution, using polymer/drug ratios of 1:1 (a), 2:1 (b), 5:1 (c) and 8:1 (d)............................................... 89

Figure VII.9: PSDs in terms of number of particles and volume percentages of CH/Amp microparticles coprecipitated by SAA at polymer/drug ratios of 1:1 (∆), 2:1 (ο), 5:1 (□) and 8:1 (����). .......................................................... 90

Figure VII.10: X-Ray diffraction patterns of: a) raw Amp; b) raw CH; c) SAA produced Amp microparticles; d) SAA produced CH microparticles; e) physical mixture of SAA produced CH microparticles and raw Amp in the ratio 1:1; f) physical mixture of SAA produced CH and Amp microparticles in the ratio 1:1; g) SAA coprecipitated microparticles in the ratio 1:1....... 92

Figure VII.11: X-Ray diffraction patterns of: a) raw Amp; b) raw CH; c) SAA produced Amp microparticles; d) SAA produced CH microparticles; e) physical mixture of SAA produced CH microparticles and raw Amp in the ratio 5:1; f) physical mixture of SAA produced CH and Amp microparticles in the ratio 5:1; g) SAA coprecipitated microparticles in the ratio 5:1....... 94

Figure VII.12: X-Ray diffraction patterns of SAA coprecipitated microparticles in the ratio 1:1, 2:1, 5:1 and 8:1.......................................... 95

Figure VII.13: DSC thermograms of: a) raw Amp; b) SAA produced CH microparticles; c) physical mixture of SAA produced CH microparticles and raw Amp in the ratio 1:1; d) SAA coprecipitated microparticles in the ratio 1:1................................................................................................................. 96

Figure VII.14: DSC thermograms of: a) raw Amp; b) SAA produced CH microparticles; c) physical mixture of SAA produced CH microparticles and raw Amp in the ratio 5:1; d) SAA coprecipitated microparticles in the ratio 5:1................................................................................................................. 97

Figure VII.15: DSC thermograms of SAA coprecipitated microparticles in the ratio 1:1, 2:1, 5:1 and 8:1. DSC thermogram of raw Amp is reported as reference.......................................................................................................98

VII

Figure VII.16: Ampicillin and chitosan chemical formulae..........................99

Figure VII.17: EDX microanalysis of physical mixture of SAA produced CH microparticles and raw Amp: SEM image of the analysed area (SEM); Oxygen map (O); Sulphur map (S)..............................................................100

Figure VII.18: EDX microanalysis of physical mixture of SAA produced CH microparticles and SAA produced ampicillin microparticles: SEM image of the analysed area (SEM); Oxygen map (O); Sulphur map (S)....................101

Figure VII.19: EDX microanalysis of SAA coprecipitated CH/Amp microparticles. Landscape of: SEM image of the analysed area (SEM); Oxygen map (O); Sulphur map(S)...............................................................102

Figure VII.20: EDX microanalysis of SAA coprecipitated CH/Amp microparticles. Particulars of: SEM image of the analysed areas (SEM); Oxygen maps (O); Sulphur maps (S)...........................................................102

Figure VII.21: In vitro Amp release profiles from: physical mixture of raw CH and raw Amp in the ratio 5:1 (a); physical mixture of SAA produced CH microparticles and raw Amp in the ratio 5:1 (b); SAA coprecipitated CH/Amp microparticles in the ratio 5:1 (c). Drug release rates from coprecipitates are reported up to 120 min (left) and to 96 h (right). ..........104

Figure VII.22: In vitro Amp release profiles from SAA coprecipitated CH/Amp microparticles in the ratio 1:1 (a), 5:1 (b) and 8:1 (c). Drug release rates from coprecipitates are reported up to 24 h (left) and to 96 h (right). .........................................................................................................105

Figure VII.23: Time dependence of normalized gel layer thickness in presence of drug (adapted from Harland et al., 1988). ..............................108

Figure VII.24: Schematic illustration (cross-section view) o a swellable matrix tablet during radial drug release.....................................................109

Figure VII.25: Photograph of the upper base of a HPMC cylindrical matrix containing 60% w/w of buflomedil pyridoxalphosphate, placed in between two transparent discs after one hour of swelling–release (adapted from Colombo et al., 2000a)................................................................................110

Figure VII.26: Schematic representation of swellable tablet containing CH/Amp coprecipitate during axial drug release. ......................................111

Figure VII.27: Definition of main parameters in swellable-soluble matrix tablet............................................................................................................114

Figure VII.28: (a) Scheme of the matrix for mathematical analysis, with (b) symmetry planes in axial and radial direction for the water and drug concentration profiles, (c) “sequential layer” structure for numerical analysis........................................................................................................116

VIII

Figure VII.29: Exponential fit (─) of k1 and k2 values vs. r and linear fit (─) of m values (●) vs. r.................................................................................... 120

Figure VII.30: Fitting curves of experimental Amp release data calculated using k1, k2 and m as functions of r (Eq. 13)............................................... 121

Figure VII.31: Relaxation/Fickian ratio (R/F) as a function of the fractional Amp release obtained by coprecipitates at different CH/Amp ratios (r).... 122

IX

INDEX OF TABLES

Table VI.1: Effect of the chitosan concentration in the liquid solution and of the temperature in the precipitator on the morphology of CH particles precipitated by SAA from 1% acetic acid aqueous solution..........................62

Table VI.2: Number of particles PSDs data on of micronized α-CD at different concentration values. Comparison between LS and SEM analysis........................................................................................................................78

Table VI.3: Volumetric PSDs data of micronized α-CD at different concentration values. Comparison between LS and SEM analysis...............78

Table VI.4: Number of particles PSDs data on of micronized HP-β-CD at different concentration values. Comparison between LS and SEM analysis........................................................................................................................80

Table VI.5: Volumetric PSDs data of micronized HP-β-CD at different concentration values. Comparison between LS and SEM analysis...............81

Table VII.1: Volumetric PSDs percentages of CH/Amp microparticles coprecipitated by SAA at polymer/drug ratios of 1:1, 2:1, 5:1 and 8:1........96

Table VII.2: Time values related to the 50% (t50) and 90% (t90) of Amp released from physical mixtures and SAA coprecipitates at ratio 5:1.........110

Table VII.3: Time values related to the 50% (t50) and 90% (t90) of Amp released from physical mixtures and SAA coprecipitates at different CH : Amp ratios...................................................................................................111

Table VII.4: k1, k2 and m values calculated as pure constants using Eq. 5 for Amp release kinetics from SAA coprecipitates at different CH/Amp ratios (r). ...............................................................................................................126

Table VII.5: k1, k2 and m values calculated as functions of r using Eq 13 for Amp release kinetics from SAA coprecipitates............................................129

X

ABSTRACT

Controlled and modified release formulations offer an effective way to optimise the bioavailability and the resulting blood concentration-time profiles of drugs that otherwise suffer from such limitation. Controlled release of drugs presents many advantages, such as the reduction in frequency dosing, blood level fluctuations and adverse side effects. Among the different formulations of drug controlled release systems, microparticulate drug delivery offers several advantages over the conventional dosage forms, such as higher efficacy and flexibility of administration. Indeed, according to the pharmaceutical target, different kinds of formulations can be produced using microparticles.

Several processes are currently studied to design composite particles and SCFs are promising technologies able to overcome many drawbacks presented by conventional methods. Among SCFs processes, Supercritical Assisted Atomization (SAA) proved to be a very efficient and versatile process for the micronization of pure compounds. The aim of this thesis is to broaden the field of SAA applications, demonstrating that this supercritical based technique can be applied not only to particle size reduction of pure compounds, but also to the production of composite microparticles for the drug controlled release.

An initial investigation about the SAA processability of different kinds of carriers for drug release has been performed. Synthetic (PMMA and PLLA) and natural (chitosan) polymers and cyclodextrins have been successfully processed, obtaining well-defined spherical microparticles at the optimal process conditions, for all individual compounds. The precipitation temperature has revealed to be a key parameter for polymer processability: a successfully micronization is possible if the temperature in the precipitator is far from the polymer glass transition temperature, but high enough to allow a fast solvent evaporation.

Afterwards, Medroxyprogesterone acetate (MEPA) and Ampicillin trihydrate (Amp) have been selected as model drugs for the first attempts of SAA coprecipitation and have been processed with PMMA and chitosan (CH), respectively. SAA composite microparticles have been characterized by X-ray, DSC, EDX and UV-vis analysis

XII

In the case of PMMA/MEPA coprecipitation, good results have been obtained, with the production of spherical and well defined microparticles with a uniform morphology. Analyses have confirmed that MEPA and PMMA form a solid solution, leading to a prolonged release of the drug.

SAA CH/Amp coprecipitation been successful as well, producing also in this case well-defined spherical microparticles, with a uniform morphology and a narrow distribution. X-ray, DSC and EDX and drug release analysis confirmed that the drug is entrapped in an amorphous solid state into the polymeric matrix. Moreover, CH avoids the degradation of Amp.

A prolonged release has been obtained for SAA coprecipitates with respect to raw drug and physical mixtures of CH and Amp. The polymer/drug ratio revealed to be a controlling parameter for drug release; in particular, it allows the modulation of drug release rate from coprecipitates. CH/Amp system exhibits a swelling-controlled behavior, with drug release kinetics depending on front movements inside the tablet. Drug release depends on both relaxation and diffusive mechanisms. Indeed, the empirical binomial equation proposed by Peppas and Sahlin (1989), fairly well describe Amp release rate if both the relaxational and the diffusional parameters are function of the polymer/drug ratio.

In conclusion, SAA has shown to be an efficient technique for the production of composite microparticles formed by a solid solution of an active agent and a carrier, resulting in thermal stabilization and controlled release of the drug.

Chapter I INTRODUCTION

The optimisation and development of new products are two important

aspects of industrial research. The pharmaceutical industry is one of the most active in this field. In the past, research concentrated on the synthesis of new pharmacological molecules. Nowadays, several active principles have been successfully selected, but they cannot be used since limited water solubility strongly reduces their bioavailability. Therefore, now the pharmaceutical research is aimed at the engineering not only of the production processes, but also of the characteristics of the products.

Controlled and modified release formulation technologies offer an effective way to optimise the bioavailability and the resulting blood concentration-time profiles of drugs that otherwise suffer from such limitation. The controlled or modified release refers to both delayed- and extended-release systems for drug administration as well as drug delivery systems designed specifically to modify the release of poorly water-soluble drugs. Indeed, some drugs are inherently long lasting and require only once-a-day dosing to sustain adequate blood levels and the desired therapeutic effect. These drugs are formulated in the conventional manner in immediate-release dosage forms. Many other drugs are not inherently long lasting and require multiple daily dosing to achieve the desired therapeutic results. This is often inconvenient for the patient and can result in missed doses and patient non-compliance with the therapeutic regimen. Extended release formulations, for example, are commonly taken only once or twice a day compared with counterpart conventional forms that may need to be taken three to four times a day to achieve the same therapeutic effect.

The controlled release of drugs presents many advantages over conventional forms:

� The reduction in drug blood level fluctuations, since a controlled drug-blood level is possible.

� The reduction in frequency dosing, since extended release products frequently deliver more than a single dose of medication and thus they may be taken less often than conventional forms.

Chapter I

2

� The enhanced patient convenience and compliance, deriving from less frequent dose administrations. As a result, the patient is less apt to neglect taking a dose and there is also greater patient and caregiver convenience with daytime and nighttime medication administrations.

� A reduction in adverse side effects as there are fewer drug blood level peaks outside of the drug therapeutic range and into the toxic range.

� A reduction in the overall health care costs. Although the initial cost of extended-release dosage forms may be greater than for conventional dosage forms, the overall cost of treatment may be lower due to enhanced therapeutic benefits, fewer side-effects and less time required of health care personnel to dispense and administer drugs and monitor patients.

However, controlled release delivery systems may present some disadvantages, such as the loss of flexibility in adjusting the drug dose and/or dosage regimen and an increased risk of sudden and total release or “dose dumping” due to technological failure of the dosage unit.

Another frequent limitation of a number of drugs is their low solubility in water which affects their bioavailability to the organism and the development of both immediate-release and modified-release dosage forms. Indeed, the challenge for poorly soluble drugs is to enhance the rate of dissolution to minimize variations and maintain a well-dispersed system that allows the drug to be absorbed. One of the approaches used to overcome dissolution limitation is the formulation of the active drug in an amorphous form. A parallel strategy is a reduction in particle size through the preparation of nano- and micro-particles (Mosharraf&Nystrom, 1995; Reverchon et al., 2004). A chemical approach to facilitate this result is the incorporation of Cyclodextrin derivatives as solubilizing agents into modified-release systems and the water-soluble polymer/drug co-formulations (Szejtli, 1994; Uekama et al, 1998).

The main method used to obtain controlled release is the incorporation of biologically active agents within polymers (i.e., biopolymers). In particular, of extreme interest are microcapsules and microspheres. The former are microparticles formed by an active agent core wrapped in a polymeric shell; the latter are polymer microparticles in which the active agent is uniformly dispersed.

Composite microparticles offer many advantages with respect to other delivery systems. For example, in oral dosage forms the administered dose of the drug is subdivided into small units spread over a large area of the gastrointestinal tract, with resulting enhanced absorption due to the lower localized drug concentration. Moreover, microparticles avoid the inconvenient surgical insertion of large implants and allow pulmonary and

INTRODUCTION

3

lung administrations. Microparticles of size less than 250 µm, ideally less than 125 µm are suitable for this purpose (Tice et al., 1991).

Several techniques can be used to prepare polymeric microparticles for controlled drug delivery, for instance: solvent evaporation, emulsion polymerisation, jet milling and spray drying. However, these conventional processes have some drawbacks, such as high temperatures and limited size control during spray drying, or the manufacturing complexity of current emulsion techniques. Moreover, the organic solvents used to dissolve polymers are difficult to remove and several further operations are necessary, therefore, to reduce solvent residues below safety limits.

In the last two decades, supercritical fluid (SCF) technology has been proposed as a valid alternative to conventional techniques for the production of micro- and nano-sized particles. SCFs take advantage of some specific properties of gases at supercritical conditions. They are characterized by a continuous adjustable solvent power/selectivity obtained varying pressure and temperature; their diffusivities can be about two orders of magnitude larger than those of liquids. As a result, SCF based processes can show very fast mass transfer and performances that cannot be obtained by conventional solvents and they are characterized by solventless or organic solvent-reduced operations.

Among the many possible SCFs, carbon dioxide is the most widely used. It has readily accessible critical points (Tc=31.1 °C and Pc=79.9 bar) and as a process solvent offers the additional benefits of being non-toxic, non-flammable, environmentally acceptable, inexpensive, and can be used at a mild critical temperature suitable for processing thermally labile compounds.

The SCF techniques for the preparation of microparticles can be classified into three groups according to the solvating behaviour of the SCF, which can act as a solvent, as an antisolvent and as a solute.

The Rapid Expansion of Supercritical Solutions (RESS), uses the SCF as a solvent. This technique consists of saturating a supercritical fluid with a solid substance and then depressurising it through a nozzle into a low-pressure chamber. As a consequence of the large reduction of density, an extremely rapid supersaturation of the solution is obtained and small particles are produced (Matson et al., 1987; Reverchon et al., 1995; Jung&Perrut, 2001).

A second kind of processes uses the SCF as an antisolvent fluid and has been proposed in literature under several names, some of which are trademarks (SEDS, ASES, GAS, SAS). In this work, these processes will be called Supercritical Antisolvent precipitation (SAS). It is similar to the conventional crystallization induced by a liquid antisolvent. The supersaturation of the solution, formed by the solute and a liquid solvent, occurs due to the solubilization of the solvent in the SCF, causing the precipitation of the solute as micro- and nano-particles (Bleich et al., 1994; Subramaniam et al., 1997; Reverchon and Della Porta, 1999).

Chapter I

4

In the third kind of technologies the SCF acts as a solute. Indeed, in the Particles from Gas Saturated Solution (PGSS) technique the product to be micronized is liquefied by heating and addition of supercritical CO2; then the gas-liquid solution is sprayed in a low-pressure vessel, thus obtaining microparticles (Weidner et al., 1995; Sencar-Bozic et al., 1997; Kerc et al., 1999).

The latter group also includes some supercritical based atomization techniques. Among these, we have Carbon Dioxide-Assisted Nebulization with a Bubble Dryer (CAN-BD) which is based on the mixing of a liquid solution and SC-CO2 in a near-zero volume tee connection and subsequent atomization through a capillary tube. It has been successfully tested for several materials (Sievers and Karst, 1997; Sievers et al., 2000; Sellers et al., 2001).

The Supercritical Assisted Atomization (SAA) is similar to CAN-BD in some aspects, but substantially differs in the solubilization of CO2 (Reverchon, 2002a; Reverchon, 2003). Indeed, SAA is based on the solubilization of controlled quantities of SC-CO2 in liquid solutions (formed by organic solvents or water and a solid solute) using a saturator that contains high surface packings and ensures long residence times. Therefore, a near-equilibrium solution is formed that is subsequently atomized through a nozzle. In this case, SC-CO2 acts both as a co-solute being partially miscible with the solution to be treated, as well as a pneumatic agent to atomize the solution in fine droplets. The liquid solvent, meanwhile, acts as a carrier for the product to be treated.

This technique offers several advantages, such as the absence of chemical or thermal degradation and the possibility of using both organic and inorganic solvents, thus managing to process both water-soluble as well as non water-soluble compounds. Moreover, SAA provides a good particle size and particle size distribution control and microparticles often ranging between 0.5 and 5 µm can be produced. These characteristics, joined to the quality of a green-process and not aggressive process regarding the substances treated, make the SAA technique able to successfully micronize a wide range of compounds.

To date, this process has been successfully tested on a number of pharmaceutical compounds, catalyst and superconductor precursors obtaining micrometric and submicrometric particles of controlled size and distribution (Reverchon et al., 2002; Reverchon, 2002a; Reverchon, 2002b; Reverchon&Della Porta, 2003a, 2003b, 2003c; Reverchon et al., 2004). But, SAA process as yet has never been applied to the precipitation of composite materials, such as microspheres and microcapsules. Therefore, the aim of this work is to propose the SAA as a valid alternative technique to conventional ones for the production of microparticles to be used in controlled release delivery systems.

Chapter II STATE OF THE ART

Several processes have been proposed to obtain microparticles for

controlled drug release. Different kinds of formulations are required depending on the final aim. For example, polymer/drug co-formulations can be used either for extended and delayed release or to enhance the solubility of poorly water-soluble compounds. Inclusion complexes, such as drug/cyclodextrin complexes, can achieve these performances too. Obviously, different techniques can be used to produce polymer/drug formulations or inclusion complexes. Therefore, in this work, the state of the art is divided into two parts of composite microparticles production, respectively regarding the polymer/drug formulations and the drug/cyclodextrin complexes.

II.1 POLYMER/DRUG MICROPARTICLES PRODUCTION

II.1.1 Traditional Processes For Microcapsules and microspheres

production

Although, a number of microencapsulation techniques have been developed and reported to date, the choice of the technique depends on the nature of the polymer, the drug, the intended use, and the duration of the therapy (Jalil&Nixon, 1990; Tice et al., 1991; Wu, 1995; Lewis, 1990; Arshady, 1991). The microencapsulation method employed must include the following requirements (Fong, 1988):

� The stability and biological activity of the drug should not be adversely affected during the encapsulation process or in the final microsphere product.

� The yield of the microspheres having the required size range (up to 250 µm) and the drug encapsulation efficiency should be high.

Chapter II

6

� The microsphere quality and the drug release profile should be reproducible within specified limits.

� The microspheres should be produced as a free flowing powder and should not exhibit aggregation or adherence.

Several techniques can be used to prepare polymeric microparticles but many of them are based on the solvent evaporation, thus presenting some analogies in the process. Particularly, all these processes practically consist of two main steps. The first one is the preparation of an emulsion of the active compound in a polymer solution; the second one is the removal of the solvents and the microparticle formation. Both the ways of preparing the emulsion and the following phase of particle formation present several variations (Jain, 2000; O’Donnell et al., 1997).

II.1.1.1 Solvent evaporation and solvent extraction process

Single emulsion process. This process involves oil-in-water (O/W) emulsification. The polymer is first dissolved in a water immiscible, volatile organic solvent. The drug is then added to the polymer solution to produce a solution or dispersion of the drug particles (particle size of the drug added should be < 20 µm). This polymer-solvent-drug solution/dispersion is, then, emulsified (with appropriate stirring and temperature conditions) in a larger volume of water in presence of an emulsifier (like poly(vinyl alcohol) (PVA)) to yield an o/w emulsion. The emulsion is subjected to solvent removal by either evaporation or extraction process to harden the oil droplets (Arshady, 1991). In the former case, the emulsion is maintained at reduced pressure or at atmospheric pressure, with a low stir rate to enable the volatile solvent to evaporate. In the latter case, the emulsion is transferred into a large quantity of water (with or without surfactant) or other quench medium, into which the solvent associated with the oil droplets is diffused out.

The solid microspheres obtained are, then, washed and collected by filtration, sieving, or centrifugation. They are dried under appropriate conditions or lyophilized. It should be noted that for the solvent evaporation process is similar to the extraction method, in the sense that the solvent must first diffuse out into the external aqueous dispersion medium before it could be removed from the system by evaporation (Wu, 1995; Arshady, 1991).

The rate of solvent removal by the extraction method depends on the temperature of quench water or other medium, ratio of emulsion volume to quench water/medium volume and the solubility characteristics of the polymer, the solvent, and the dispersion medium. The rate of solvent removal by evaporation method strongly influences the characteristics of the final microspheres and it depends on the temperature, pressure, and the solubility parameters of the polymer, the solvent, and the dispersion medium. Very rapid solvent evaporation may cause local explosion inside the droplets and lead to formation of porous structures on the microsphere

STATE OF THE ART

7

surface. The solvent removal by extraction method is faster (generally <30 min) than the evaporation process and hence the microspheres made by the former method are more porous in comparison to those made from the latter method under similar conditions (Arshady, 1991).

One of the disadvantages of the o/w emulsification method is the poor encapsulation efficiencies of moderately water-soluble and water-soluble drugs. The drug could diffuse out or partition from the dispersed oil phase into the aqueous continuous phase and microcrystalline fragments of the hydrophilic drugs can deposited on the microsphere surface and dispersed in the polymer matrix (Wada et al., 1988; Cavalier et al., 1986). This would result in poor trapping of the drug and initial rapid release of the drug (burst effect). The o/w emulsification process is widely used to encapsulate lipid-soluble drugs like steroids (Jalil&Nixon, 1990).

In order to increase the encapsulation of the water-soluble drugs, an oil-in-oil (o/o) emulsification method was developed (Tsai et al., 1986). A water-miscible organic solvent is employed to solubilize both the drug and the polymer. This solution is dispersed into an oil, in presence of an oil soluble surfactant, to yield the o/o emulsion. Microspheres are, at the end, obtained by evaporation or extraction of the organic solvent from the dispersed oil droplets. The formation of the microspheres is affected by a number of factors. The main variables that influence the microencapsulation process and the final microsphere product are:

• the nature and solubility of the drug being encapsulated; • the polymer concentration, composition, and molecular weight; • the drug/polymer ratio; • the organic solvent used; • the concentration and nature of the emulsifier used; • the temperature and stirring/agitation speed of the emulsification

process; • the viscosities and volume ratio of the dispersed and continuous

phases (Lewis, 1990).

Double (multiple) emulsion process. It is a water-in-oil-in-water (W/O/W) method and is best suited to encapsulate water-soluble drugs like peptides, proteins, and vaccines, unlike the o/w method which is ideal for water-insoluble drugs (Jalil&Nixon, 1990; Wu, 1995; Lewis, 1990). A buffered or plain aqueous solution of the drug (sometimes containing a viscosity enhancer and/or stabilizing protein like gelatin) is added to an organic phase consisting of the polymer and an organic solvent with vigorous stirring to form the first microfine w/o emulsion. This emulsion is gently added with stirring into a large-volume water containing an emulsifier to form the w/o/w emulsion. The emulsion is, then, subjected to solvent removal by either evaporation or extraction process. The solid microspheres so obtained are then washed and collected by filtration, sieving, or

Chapter II

8

centrifugation. These are then dried under appropriate conditions or are lyophilized to give the final free flowing microsphere product.

Multiple emulsions of the W/O/O or W/O/O/O type. Iwata and McGinity (1992) developed a multiple emulsion of the W/O/O/O type. Multiphase microspheres of either PLA or PLGA containing water-oil (W/O) emulsions were prepared by a multiple emulsion solvent evaporation technique. Acetonitrile was used as the solvent for the polymer, and light mineral oil comprised the continuous phase for the encapsulation procedure. Scanning electron microscopy of transverse cross sections of the multiphase microspheres revealed cavities in which the W/O emulsion resided. This suggested that the multi-phase microspheres of the W/O/O/O type belonged to the class of reservoir type drug delivery devices.

This type of multiple emulsion system allows the encapsulation of a primary water-in-oil emulsion within a polymeric microsphere. The oil in the primary emulsion prevents contact between the internalized protein and the polymer/ solvent systems. The isolation of the protein from the polymer/solvent system prevents possible denaturation of protein by the polymer or the solvent. Likewise, the possibility of polymeric degradation due reactive proteins or drug compounds is also limited.

O’Donnell and coworkers (1995) prepared multi-phase microspheres of poly(DL-lactic-co-glycolic acid) by a multiple emulsion potentiometric dispersion technique. Water soluble compounds were dissolved in the aqueous phase (W) and emulsified in soybean oil (O) to form a stable emulsion. This primary emulsion was dispersed in a solution of PLGA and acetonitrile (O) to form a W/O/O emulsion. The W/O/O emulsion was then dispersed in a hardening solution of light mineral oil (O) using a potentiometric dispersion technique to produce microspheres of the W/O/O/O type with a very narrow and selective size distribution.

II.1.1.2 Phase separation (coacervation)

This process consists of decreasing the solubility of the encapsulating polymer by addition of a third component to the polymer solution in an organic solution. The process yields to phase separation into two liquid phases: the polymer containing coacervate phase and the supernatant phase depleted in polymer. The drug, which is dispersed in the polymer solution, is surrounded by the coacervate. Therefore, the coacervation process includes the following three steps:

• phase separation of the coating polymer solution, • adsorption of the coacervate around the drug particles, • solidification of the microspheres (Edelman et al., 1993).

The polymer is first dissolved in an organic solution. The water-soluble drugs, like peptides and proteins, are dissolved in water and dispersed in the

STATE OF THE ART

9

polymer solution (W/O emulsion). Hydrophobic drugs are either dispersed in the polymer solution. An organic antisolvent is added to the polymer-drug-solvent system with stirring which gradually extracts the polymer solvent. As a result, the polymer is subjected to phase separation and it forms very soft coacervate droplets (size controlled by stirring) entrapping the drug. Then this system is transferred to a large quantity of another organic antisolvent to harden the microdroplets and form the final microspheres that are collected by washing, sieving, filtration, or centrifugation, and are finally dried (Wu, 1995; Edelman et al., 1993).

This process, unlike the O/W emulsification method, is suitable to encapsulate both water-soluble as well as water-insoluble drugs, since is a non-aqueous method. However, the coacervation process is mainly used to encapsulate water-soluble drugs like peptides, proteins, and vaccines.

The addition rate of the first antisolvent is an important process parameter since should be such that the polymer solvent is extracted slowly, to let the polymer have sufficient time to deposit and coat on the drug particle surface during the coacervation process. The concentration of the polymer used is important as well, since too higher concentrations would result in rapid phase separation and not-uniform coating of the polymer on the drug particles. Due to absence of any emulsion stabilizer in the coacervation process, agglomeration is a frequent problem in this method. The coacervate droplets are extremely sticky and adhere to each other before the complete phase separation or the hardening stages of this method. Adjusting the stirring rate, temperature, or the addition of an additive is known to rectify this problem (Wu, 1995).

Compared to the solvent evaporation/extraction process, the requirement of solvents for the polymer is less stringent since the solvent need not be immiscible with water and the boiling point can be higher than that of water. DCM, acetonitrile, ethyl acetate, and toluene have been used in this process (Edelman et al., 1993; Mandal et al., 1995). The antisolvent affect both the phase separation and the hardening stages of the coacervation process. The antisolvent should not dissolve the polymer or the drug and should be miscible with the polymer solvent.

The second antisolvent should be relatively volatile and should easily remove the first viscous antisolvent by washing. Some of the oils used as the first antisolvent are silicone oil, vegetable oils, light liquid paraffin, low molecular weight liquid polybutadiene, and low molecular weight liquid methacrylic polymers. Examples of the second antisolvent include aliphatic hydrocarbons like hexane, heptane, and petroleum ether.

In the phase separation method the phase equilibrium is never reached. Therefore, the formulation and process variables significantly affect the kinetics of the entire process and ultimately the characteristics of the final microspheres. The preparation conditions substantially affect the

Chapter II

10

morphology and porosity of microspheres. Stability of the primary emulsion is a prerequisite the successful encapsulation of multiple emulsions.

II.1.1.3 Spray drying

The spray drying method is very rapid, convenient, easy to scale-up and is less dependent on the solubility parameter of the drug and the polymer (Takada et al., 1995; Wagenaar&Muller, 1994).

Both the polymer and the drug are dissolved in the same solvent and processed by spray drying. For example, microspheres were produced by suspending a water-soluble drug (theophylline) or dissolving a water-insoluble drug (progesterone) in a PLA/DCM solution and then spray-drying the solution to produce particles smaller than 5 µm (Bodmeier&McGinity, 1988). Due to incompatibility of the hydrophilic drug and PLA, needle-shaped crystals grew on the micro-sphere surface, while the progesterone-PLA solution gave smooth particles. The nature of the solvent used, temperature of the solvent evaporation, and presence of PLA microspheres during the spray-drying process affected the polymorphic form of progesterone.

Some of the major problems encountered with this technique are the poor control of particle size and particle size distribution and the formation of fibers due to insufficient force available to break up the polymer solution. An efficient dispersion of the filament into polymer droplets was dependent on the type of polymer and the viscosity of the spray solution. Other groups have also reported successful preparation of PLGA and PLA particles using the spray drying technique (Wagenaar&Muller, 1994; Tamber et al., 1996; Lee et al., 1996). A solution of the polymer, DCM, and piroxicam was spray-dried to yield microspheres that were hollow. The microspheres ranged between 1 and 15 µm, with a high drug encapsulation efficiency of 99.0%.

One of the major disadvantages is a significant loss of the product that may occur during spray-drying, due to adhesion of the microparticles to the inside wall of the spray-drier apparatus; moreover, agglomeration of the microparticles can take place (Takada et al., 1995) and thermal degradation can occur due to the high process temperatures.

II.1.2 Supercritical Fluids based Processes For microcapsules and

microspheres production

As discussed in the previous sections, biodegradable polymer/drug microparticles have been successfully prepared by double-emulsion and

STATE OF THE ART

11

phase-separation methods. However, these conventional processes are expensive and with a high environmental impact. The coacervation method, indeed, tends to produce agglomerated particles, with a consequent difficulty in mass production. The double-emulsion method on the other hand requires many steps, rigid control of the temperature and viscosity of the inner w/o emulsion, and is difficult to encapsulate higher amounts of hydrophilic drugs (Jalil&Nixon, 1990; Takada et al., 1995). Hence, emulsification, solvent evaporation methods need a large use of organic solvents; consequently they require many operations at the end of the process to remove them and to achieve residual solvent values below the safety limits. Furthermore, limited particle size and particle size distribution control is possible.

SCF particle technology offer potential advantages related to the efficient extraction of solvents and impurities and the plasticization of polymers. An efficient control over particles precipitation is also possible by changing the mass transfer conditions and the phase behaviour of the ternary mixture solvent-antisolvent-solute. The techniques for the preparation of polymer particles (Reverchon et al., 2003) using SCFs can be divided in three main categories in which the SCF is used as:

� a solvent, for example the Rapid Expansion of Supercritical Solutions (RESS),

� an antisolvent, for example the Supercritical Antisolvent precipitation (SAS),

� a solute, for example the Particle from Gas Saturated Solution (PGSS).

II.1.2.1 RESS-based techniques

The RESS process consists of saturating a supercritical fluid with a substrate; then, depressurising this solution through a heated nozzle into a low pressure chamber, an extremely rapid nucleation of the substrate occur and very small particles (or fibres, or films) precipitate.

Ribero Dos Santos et al. (2002) and Benoit et al. (2000) recently proposed the production of microcapsules solubilizing the coating material in SC-CO2 and gradually reducing the temperature and/or the pressure to precipitate the coating material onto drug particles dispersed in the SC-phase. This technique can be classified as a RESS process because the SCF acts as a solvent; but it differs from other RESS processes because the depressurization step is slow. Moreover, supersaturation can be obtained changing simultaneously pressure and temperature.

The primary advantages of this process are the absence of organic solvents, the low temperature and the absence of any physical and chemical transformation of the active substance. For this reason, it can be applied to proteins processing avoiding their inactivation. The major drawback of this process is the low solubility of several polymers in SC-CO2. For this reason,

Chapter II

12

preferred coating materials are lipids since they show a significant solubility in SC-CO2. Moreover, the control of particle size and morphology of the capsules is difficult.

The use of the previous process has been also proposed by Dulieu et al. (2001) for the production of microcapsules composed of a protein and a coating lipid to be used for injection delivery. Lockemann et al. (2002) patented a process and an apparatus in which an active substance and a coating material are dissolved in separate streams of SCFs and sprayed through a nozzle arranged to produce the interpenetration of the two spray cones. The supersaturation of the active substance has to occur faster than that of the coating material. According to the authors, submicrometric-particles with a mean diameter of 400 nm are produced using β-carotene as the core material around which a block copolymer based on maleic anhydride and C20 –C24 olefins forms a uniform envelope.

The peculiarity of the RESS-based processes is the possibility to inhibit the growing step of particles thus obtaining very small particles. However, in some cases, coalescence occurs and the final size of the particles increases.

This technique is simple and attractive; nevertheless, industrial applications are limited since few solids show acceptable solubility in SC-CO2 and a use of CO2 up to 10,000 kg per kg of product can be obtained (Bertucco, 1999). Moreover, the produced particles are very difficult to be collected.

II.1.2.2 SAS-based techniques

In the SAS process the drug, that is not soluble in the SCF, is solubilized in a liquid solvent that is completely miscible with the SCF. The solution formed by the drug and the liquid solvent is sprayed into a high-pressure vessel containing the SCF. The supersaturation of the solution, due to the miscibility of the SCF with the liquid solvent, causes the precipitation of the solute as micro- and nano-particles. SAS-based precipitation is a very rapid process if compared with the traditional liquid antisolvent crystallization. For this reason, amorphous spherical particles are a common habit of SAS precipitates. However, in some cases, when the growing rate of the substance is of the same order of magnitude of the precipitation rate, the particles can form crystalline structures.

The main advantages of the SAS process are the wide range of materials that can be treated, and the control of particle size and distribution (Jung&Perrut, 2001; Reverchon, 1999). The drawbacks are the large amount of antisolvent that is required (up to about 600 kg of CO2 per kg of product), and the difficulty to process aqueous solutions, because of the very low solubility of water in SC-CO2 at the ordinary process conditions. Moreover, the solvent residue in the powder can be larger than the quantity allowed by regulatory limits and post-treatments can be necessary.

STATE OF THE ART

13

A pioneering work for microspheres production was proposed by Fisher and Muller (1988). More recently, Richard et al. (2001) patented a SAS-based process for the production of microcapsules for pulmonary administration. The coating polymer is dissolved in an organic solvent and the active substance is suspended in the solution. SC-CO2 acts as the antisolvent producing a solution with the liquid solvent and induces the precipitation of the polymer on the suspended particles of the active substance. In this way microparticles of insulin coated with Polylactide-co-glicolide (PLGA) have been obtained with an average diameter of 3 µm and insulin loading up to 90% by weight.

Ghaderi et al. (2000) used supercritical antisolvent process to entrap hydrocortisone in poly(D,L-lactide-co-glycolide) (DL-PLG). In this case, compressed nitrogen was added to SC-CO2 with the aim of improving the atomization rate of the polymer solution. However, the obtained particles were irregular and agglomerated and the entrapping efficiency was relatively low. Foster et al. (2003) proposed a modified antisolvent process to generate a protein–sugar co-precipitate from an aqueous solution. In this composite, sugar can have the capacity to protect the protein against denaturation during manufacturing and storage. These authors used lysozyme and lactose as protein and sugar, respectively. To apply the SC-antisolvent technology to an aqueous solution, ethanol was added to SC-CO2 as a co-solvent. At the best process conditions, nanospheres were obtained with a lysozyme content of 92.5 %wt.

Other examples of successful formation of microcapsule are the co-precipitation of Budesonide into poly(L L-lactic acid) in an amorphous form (2002), the formation of a solid solution of theophylline in ethylcellulose (2003) and the batch antisolvent precipitation process used by Pallado et al. (2001) to prepare nanospheres composed of calcitonin or insulin coated with a polysaccharide. Elvassore et al. (2001) tested the encapsulation of insulin with PLA and PEG/PLA using a batch antisolvent process. Microspheres in the range 400–600 nm were obtained with drug loading up to 90%. The release of insulin from the PLA/PEG nanospheres was constant over a period of 2 months. Perrut (2001) developed an antisolvent technique for the preparation of microcapsule of an antibiotic coated with ethyl cellulose. In this application the drug was suspended in gaseous CO2 that acted as an antisolvent for the coating polymer that was dissolved in the liquid solvent. According to the authors, the precipitation of the polymer occurs on the surface of the suspended crystals. Shekunov et al. (2003) proposed an interesting technique based on the use of SC-CO2 as an extraction medium of the solvents that remains entrapped in the droplets formed at the end f an emulsion polymerization. Drug loaded PLGA nanospheres or hollow microcapsules were obtained.

Chapter II

14

II.1.2.3 PGSS-based techniques

The PGSS process consists of atomizing a solution containing a compressed gas or supercritical fluid; indeed, the solubilities of compressed gases in liquids and solids like polymers are usually high. By decompression in a low-pressure vessel, the product can be treated, either in form of a solute or of a suspension. CO2 either swells the polymer or melts it at a temperature largely below its melting/glass transition temperature. Recently, Mandel et al. (2002) reported a PGSS-based technique for the manufacture of controlled-release pharmaceuticals. The polymer and the drug are mixed with SC-CO2 to form a homogeneous, gas saturated suspension. After the rapid expansion of the suspension, simultaneous precipitation of the polymer and the drug occurs in the form of a powder. This process has been tested with PLGA as the coating polymer and some enzymes as the active substances. The formation of microspheres infused with the drug is claimed.

STATE OF THE ART

15

II.2 MICROPARTICLES OF DRUG/CICLODEXTRIN

INCLUSION COMPLEXES

Cyclodextrins (CDs) are bucket-shaped oligosaccharides produced from starch. As a result of their molecular structure and shape, they possess a unique ability to act as molecular containers by entrapping guest molecules in their internal cavity, altering physical, chemical, and biological properties of guest molecules. The resulting inclusion complexes offer a number of potential advantages in pharmaceutical formulations.

The α−, β− and γ−cyclodextrins are widely used natural cyclodextrins, consisting of six, seven, and eight D-glucopyranose residues, respectively, linked by α-1,4 glycosidic bonds into a macrocycle.

The most common pharmaceutical application of cyclodextrins is to enhance the solubility, stability, and bioavailability of drug molecules. The cyclodextrins are very interesting products for several reasons:

� They are seminatural products, produced from a renewable natural material (starch) by a relatively simple enzymic conversion.

� They are produced in thousands of tons per year amounts by environmentally friendly technologies.

� Their initially high prices have dropped to levels where they become acceptable for most industrial purposes.

� Through their inclusion complex forming ability, important properties of the complexed substances can be modified significantly. This unprecedented “molecular encapsulation” is already widely utilized in many industrial products, technologies, and analytical methods.

� Any of their toxic effect is of secondary character and can be eliminated by selecting the appropriate CD type or derivative or mode of application. Therefore, CDs can be consumed by humans as ingredients of drugs, foods, or cosmetics.

The most stable three dimensional structure of cyclodextrins is a toroid with the larger and smaller openings respectively presenting hydroxyl groups to the external environment and mostly hydrophobic functionality lining the interior of the cavity (Figure II.1). It is this unique configuration that gives cyclodextrins their interesting properties and creates the thermodynamic driving force needed to form host-guest complexes with apolar molecules and functional groups.

Chapter II

16

Figure II.1: The conformation of the glucose units in the cyclodextrin places the hydrophilic hydroxyl groups at the top and bottom of the three dimensional ring and the hydrophobic glycosidic groups on the interior.

The ability of a cyclodextrin to form an inclusion complex with a guest molecule is a function of two key factors. The first is steric and depends on the relative size of the cyclodextrin, on the size of the guest molecule or certain key functional groups within the guest. If the guest is of the wrong size, it will not fit properly into the cyclodextrin cavity. The second critical factor is thermodynamic interactions between the different components of the system (cyclodextrin, guest, solvent). For a complex to form, there must be a favourable net energetic driving force that pulls the guest into the cyclodextrin. While the height of the cyclodextrin cavity is the same for all three types, the number of glucose units determines the internal diameter of the cavity and its volume. For example, α-cyclodextrin can typically complex low molecular weight molecules or compounds with aliphatic side chains, β-cyclodextrin will complex aromatics and heterocycles and γ-cyclodextrin can accommodate larger molecules such as macrocycles and steroids.

Figure II.2 shows an example of inclusion complex formation. In an aqueous solution, the slightly apolar cyclodextrin cavity is occupied by water molecules which are energetically unfavoured (polar-apolar interaction), and therefore can be readily substituted by appropriate “guest molecules” which are less polar than water. The dissolved cyclodextrin is the “host” molecule, and the “driving force” of the complex formation is the substitution of the high enthalpy water molecules by an appropriate “guest” molecule.

STATE OF THE ART

17

Figure II.2: Forming an inclusion complex involves multiple interactions between active, solvent and cyclodextrin.

Cyclodextrin molecules contain one or more entrapped “guest” molecules. Most frequently the ‘host : guest’ ratio is 1:1. This is the essence of “molecular encapsulation”. This is the simplest and most frequent case. However 2:1, 1:2, 2:2, or even more complicated associations, and higher order equilibria exist, almost always simultaneously.

The formed inclusion complexes can be isolated as stable crystalline substances. Upon dissolving these complexes, an equilibrium is established between dissociated and associated species, and this is expressed by the complex stability constant Ka. The association of the CD and guest (D) molecules, and the dissociation of the formed CD/guest complex is governed by a thermodynamic equilibrium:

DCDDCD ⋅⇔+

]][[][

1:1 DCD

DCDK

⋅=

As a result of their unique ability to form inclusion complexes, cyclodextrins provide a number of benefits in pharmaceutical formulations. Many of these applications have been well-studied and a significant amount of information exists in the technical literature (Connors,&Mollica,1966; Khan, 1975; Blanco et al., 1991; Fernandes&Veiga, 2002). However, it is only recently that cyclodextrins have started to become commercially significant as production improvements have made them more economically available in large scale and formulators and regulatory agencies become more familiar with their properties.

A very interesting advantage of CDs application is the control of solubility of drugs. The solubility of a guest compound can be changed upon complexation with a cyclodextrin either to increase or to decrease the solubility. When a guest compound is complexed by a cyclodextrin, the

Chapter II

18

guest in the cavity of the cyclodextrin is essentially surrounded by the molecule of cyclodextrin. The hydrophobic groups of the guest in contact with the solvent in the free state interact with the atoms of the cavity of the cyclodextrin instead. The outer surface of the cyclodextrin interacts with the solvent. As a result, this outer surface of the cyclodextrin contributes to the solubility of the complex and not the portion of the guest interacting with the cavity of the cyclodextrin (Moyano et al., 1995; Hedges, 1998).

In many cases the contribution of the outer surface to the solubility of the complex is not sufficient to obtain the desired solubility properties. Modification of the hydroxyl groups of the outer surface of the cyclodextrin can alter the solubility properties markedly. Substitution with a neutral group, such as a hydroxypropyl group, or an ionic group, such as a carboxymethyl, tertiary amine, or quaternary amine group, increases the solubility of the modified cyclodextrin to 60% or greater in water. Modification with aliphatic groups, such as hexyl groups, or nearly complete substitution with smaller groups, such as acetyl groups, results in decreased, non-detectable solubility in water and increased solubility in organic solvents.

One of the most important applications of cyclodextrins in pharmaceutical fields, therefore, is to enhance aqueous solubility of drugs through inclusion complexation. Moreover, to improve solubility, cyclodextrins also prevent crystallization of active ingredients by complexing individual drug molecules so that they can no longer self-assemble into a crystal structure.

The solubilization ability of cyclodextrins can be quantitatively evaluated by the phase solubility method developed by Higuchi and Connors (1965). The phase solubility diagrams, i.e., plots of solubility of guest as a function of cyclodextrin concentration, are generally classified as either type A (a soluble complex is formed) or type B (a complex with definite solubility is formed), as shown in Figure II.3.

Figure II.3: Type of phase solubility diagram.

STATE OF THE ART

19

The type A can be further classified in subtypes AL, AP, and AN, where the guest solubility of the first type increases linearly with cyclodextrin concentration while those of the second and third types deviate positively and negatively, respectively, from the straight line. The complex formation with a 1:1 stoichiometry gives the AL-type diagram, whereas the higher order complex formation in which more than one cyclodextrin molecules are involved in the complexation gives the AP-type. The interaction mechanism for the AN-type is complicated, because of a significant contribution of solute-solvent interaction to the complexation.

In the case of the BS-type, the initial ascending portion of the solubility change is followed by a plateau region and, then, a decrease in the solubility at higher cyclodextrin concentrations, accompanying a microcrystalline precipitation of the complex. The BI-type diagram is indicative of the formation of insoluble complexes in water. The stability constant and stoichiometry of complexes are determined by analysing quantitatively the phase solubility diagram. The solid cyclodextrin complexes can be prepared by referring the B-type solubility diagram.

II.2.1 Conventional Processes for Inclusion Complexes Production

The most commonly used methods are the coprecipitation, slurry, paste, and dry mixing methods. Each method mainly differs from the other in the amount of water used.

Water is important in the formation of complexes. In addition to being a driving force for the hydrophobic interaction of the guest with the cavity of the cyclodextrin, water is a medium for dissolution of both the cyclodextrin and the guest. Complexation is a molecular phenomenon where one molecule of guest and one molecule of cyclodextrin come into contact with each other to associate and form a complex. In some cases, water is required to maintain the integrity of the complex. Water present in the crystals of the complex can form a bridge between the hydroxyl groups of adjacent molecules of cyclodextrin to form a cage to assist in trapping the guest molecule (Hedges, 1998).

II.2.1.1 Coprecipitation Method

The coprecipitation method is the most widely used method in the laboratory. The guest compound is added to a solution of cyclodextrin while stirring. Conditions are selected to exceed the solubility of the complex, that, then, is collected as a precipitate by filtration or centrifugation.

In the case of β-cyclodextrin, the least soluble of the cyclodextrins, the solution is heated to about 60 °C to dissolve the β-cyclodextrin before

Chapter II

20

adding the guest. The solution of cyclodextrin is allowed to cool to ambient temperature as it is stirred with the guest to form the complex. The collected complex is, then, dried.

This method is essentially a laboratory method. It is performed with readily available equipment, beaker, stirrer, and heat source. It has the advantage that the complex forming can be easily seen by the disappearance of the guest.

Because of the large amount of water used, this method is not frequently used for large-scale formation of complexes because of the size of tanks required and wastewater disposal considerations. This method is used to prove feasibility of complexation of a particular guest, to obtain small quantities of complex to test functioning of the complex in applications, and to characterize the complex to determine the properties of the complex to be able to scale-up complexation using other methods (Szente&Strattan, 1988).

II.2.1.2 Slurry Method

In this method, cyclodextrins need not to be completely dissolved to form complexes. Cyclodextrin is suspended in water up to a 40-45% w/w concentration. The guest can have an effect upon the viscosity of the slurry, and concentrations are adjusted to allow mixing of the cyclodextrin and guest. As the solubilized cyclodextrin complexes, the complex precipitates; therefore more cyclodextrin dissolves to form complex. Heating can be used if desired and is compatible with the guest. The amount of time needed to complete the complexation depends on the particular guest and the vigorousness of the stirring. The mixing time needed is determined by experimentation and comparison of the characteristics of the resulting complex with the characteristics of the complex made using the coprecipitation method. The complex is generally collected by filtration and dried if a dry complex is required (Szente&Strattan, 1988; Hedges, 1998).

II.2.2 Paste Method

The paste method uses a minimum amount of water, 20-30% w/w. Cyclodextrin, water, and the guest are added to a mixing device and mixed. Because of the high viscosity, this method is usually not performed in the laboratory. The mixing time varies with the guest to be complexed, amount of water, and mixing device used. A mixing device providing high shear generally completes the complexation in a shorter amount of time than a low-shear mixer. Mixing time is determined experimentally by comparing the properties and performance of the complex formed with those of the complex made using the coprecipitation method. In most cases, a mixing time of about 30 min is sufficient, depending upon the mixer selected and the guest. The complex is, then, dried without any further treatment.

STATE OF THE ART

21

In some cases, this method can be carried out in two steps: a mixing step followed by a holding step to allow completion of the complexation reaction. The guest and cyclodextrin are mixed with a minimum amount of water. The mixture is placed in a container or is left in the mixer if it can be closed and sealed. The mixture is, then, held while heating to complete the complexation reaction. Temperature, time, the amount of water, and the particular guest control the rate at which the complexation reaction is completed (Hedges, 1998).

II.2.2.1 Dry Mixing Method

The dry mixing method involves mixing the cyclodextrin with the guest with no added water. This is generally not an efficient method of making complexes since mixing times can range from hours to days. There are some exceptions, such as lemon oil, where complexation is completed in a few minutes. In these cases, the guest might also be serving as a solvent for the cyclodextrin (Szente&Strattan, 1988; Hedges, 1998).

II.2.2.2 Drying of Complexes

When drying complexes, it is desirable to remove water as quickly as possible, especially when a volatile guest has been complexed. In the presence of water, even with the small amounts of water present in the filter cake or paste, there is equilibrium between the insoluble complex and soluble complex. The small amount of complex will dissociate, resulting in some free guest. With volatile guests, this will result in some evaporative loss of the guest so that the time for drying and reduction of the amount of water supporting solubilization of the guest should be as short as possible.

Several types of dryer, such as a tray dryer, spray dryer, fluid bed dryer, or vacuum-dryer, can be used. Dryers with heated surfaces can have some local hot spots. In addition to the usual problems of charring and coating of these surfaces, the heat can also cause dissociation of the complex. Use of hot air ovens avoids these surface concerns. The circulating air also helps to remove water vapour more quickly than diffusion. A dryer that physically mixes the complex is also preferred. Dissolution and decomplexation of the complex can occur while the hot complex remains stationary while the water slowly moves from the interior of the mass of complex to the surface to be removed. Constant stirring of the complex provides a more rapid means of removing water to keep the complex intact.

The properties of the complex also drive the choice of the dryer. Soluble complexes are most frequently dried using a spray dryer. Some complexes will partially melt or solubilize when exposed to heat. A glass or very hard complex requiring milling can result from drying soluble complexes unless the water is removed rapidly. Use of a vacuum-dryer and lower

Chapter II

22

temperatures, such as freeze-drying, generally results in a dry powdery complex.

Drying temperatures in ovens are generally around 100 °C, the boiling temperature of water. Some adjustment of the temperature might be needed for volatile guests. If the boiling temperature of the guest is exceeded, the guest will evaporate. For a volatile guest, a drying temperature from 3 to 5 °C below the boiling temperature of the guest is usually optimal. If the temperature is too low, volatile guest will also be lost due to the free guest resulting from the equilibrium established by the complex with the water present.

Factors such as temperature, amount of water, mixing time, and drying conditions must be established for the equipment and each guest and cyclodextrin used to optimize the process (Szente&Strattan, 1988; Hedges, 1998).

II.2.3 Supercritical Fluids based Processes for inclusion complexes

production

SCFs have recently been applied to cyclodextrin/drug formation and therefore there are few works about these techniques in literature.

Fabing and coworkers (2002) patented a SAS-based process to precipitate very fine particles containing at least an active principle inserted in a cyclodextrin. The method consists of forming a solution of the active principle in a first liquid solvent and of a product formed by the host molecules in a second liquid solvent; further the resulting solutions are put in contact with the SCF that plays as antisolvent and causes the precipitation of the host molecules dissolved therein. Using this techniques, the authors reported the complexation of celecoxib with hydroxypropyl- and methyl-β-CD in acetone. The complexed microparticles showed enhanced dissolution rate in water compared to that of the raw drug.

The potential of the SAS process for pharmaceutical formulations containing hydrophilic CDs from organic solution was also investigated by other authors (Mammucari et al., 2002; Foster et al, 2004). Submicrometric particles of CDs such as β-CD, methyl-β-CD and hydroxypropyl-β-CD were precipitated from dimethylsulfoxide (DMSO), ethanol and DMSO/ethanol (1/1 volume ratio) solutions using CO2 as an antisolvent. Mixtures of a hydrophobic drugs such as naproxen and CDs were dissolved in an organic solvent and coprecipitated by SAS successfully. Microspheres in the range of 200-500 nm were formed whereas the dissolution rate of the manufactured product was comparable with the physical mixture prepared by co-evaporation technique. The SAS precipitation from organic solutions

STATE OF THE ART

23

can be particularly advantageous, when compared to the co-evaporation method, for manufacturing CD systems containing labile components and where control over particle size is important (Mammucari et al., 2002).

Foster and co-workers (2004) proposed the processing of β-cyclodextrins (β-CD) from acqueous solutions by a kind of SAS process using CO2 modified with ethanol as antisolvent. Indeed, the solubility of β-CD in low viscosity organic solvents such as methanol, ethanol and acetone is limited, but their high solubility in DMSO suffered the drawbacks of the presence of residual solvent due to the strong intermolecular interaction between DMSO and β-CD. In order to eliminate the residual solvent problem, the authors proposed the SAS technique for the manufacturing of β-CD complexes from aqueous solutions, thus allowing the processing of high polar active compounds such as amino acids and polypeptides.

Marongiu and coworkers (2002) proposed the use of the SCF as solvent to enhance the complexation efficiency of CDs. Indeed, the SCF capability to swell polymeric matrixes is due to their high diffusivity and low viscosity, induced the authors to hypothesize that inclusion of some class of compounds in oligomers like β-CD in a supercritical medium is possible. The authors included some monoterpenic compounds in β-CD using sc-CO2 as solvent and compared this preparation with that carried out in water using classical methods. The complexes were prepared adjusting pressure and temperature conditions allowing the solubility of the guest molecule. After a sufficient time of contact, an abrupt depressurization caused the host/guest complex precipitation. All the preparation carried out under SC-CO2 showed increasing of the guest/host ratio respect to the preparation in water.

Cristini et al. (2002) precipitated ibuprofen/β-CD microcomposite particles by using a RESS impregnation technique. The supercritical solution of ibuprofen was injected in a stirred vessel, at atmospheric pressure, where a known amount of β-CD was previously introduced. The jet power and agitation allowed the formation of a homogenous mixture of the two components. The ibuprofen/β-CD microparticles showed enhanced release kinetic respect to ibuprofen/lactose microparticles and raw material.

Chapter II

24

Chapter III AIM OF THE Ph.D. THESIS Several processes are currently being studied to design composite

particles and SCFs are promising technologies able to overcome many drawbacks present in conventional methods. Among these SCF processes, to date, Supercritical Assisted Atomization (SAA) has proved to be a very efficient and versatile process for the micronization of pure compounds (Reverchon, 2002a, 2002b, 2003; Reverchon et al., 2002, 2003a, 2004; Reverchon&Della Porta, 2003a, 2003b, 2003c; Reverchon&Spada, 2004). However, the field of its applications could be further broadened. Indeed, SAA working principle allows the production of composite microparticles, since multi-components solutions can be processed. The aim of this thesis, therefore, is to investigate this possibility and, in particular, the application of SAA to microparticles production for drug controlled release.

The drug release mechanism and release rate depend on various factors, such as the carrier chemistry (polymer composition, molecular weight, etc.), the formulation parameters and the physical characteristics of the resulting particles (size, morphology, crystallinity, etc.). Hence, the study of the effects of process parameters on particle size, particle size distribution and particle morphology will be carried out. The influence of the carrier/drug/solvent ratio on release times and the interactions between the carrier and the drug will also be investigated.

This work will be developed in the following steps: • The preliminary processing of the carrier and of the drug separately,

in order to investigate the effect of process parameters on the micronization of pure compounds;

• The coprecipitation of the drug/carrier composite microparticles and the correlated study of process parameters;

• The characterization of the co-precipitates and study of the influence of different process conditions on microparticle release time, to tailor the microparticles to their final application;

• The general interpretation of the results, both from the point of view of processing and of the mechanism of drug release.

Chapter III

26

Microparticle characterization and drug release analyses will require the use and the mastery of several analytical methods, such as scanning electron microscopy, UV-vis spectrometry, differential scanning calorimetry, X-Ray diffractometry and SEM-EDX microanalysis.

Chapter IV SUPERCRITICAL ASSISTED

ATOMIZATION

IV.1 EXPERIMENTAL APPARATUS

The laboratory apparatus used for SAA studies is schematically reported in Figure IV.4. It consists of three feed lines delivering supercritical CO2, liquid solution and warm N2, respectively. Three vessels are the major capacities of the plant: saturator , precipitator and condenser.

2040

2040

20

2040

1

4

6

6

7

5

2

9

3

8

10

11

OUT

Figure IV.4: Schematic representation of the SAA apparatus: 1) CO2 cylinder; 2) liquid solution; 3) N2 cylinder; 4) cooling bath; 5) heating bath; 6) high pressure pumps; 7) dampener; 8) heat exchanger; 9) saturator; 10) precipitator; 11) condenser.

Chapter IV

28

Liquid CO2 is delivered by high-pressure pump to a heated bath and, then, to the saturator where the solubilization into the liquid solution takes place.

The liquid solution is taken from a graduated glass vessel, pressurized by a high-pressure pump, heated and sent to the saturator.

A controlled flow rate of heated N2 is also delivered into the precipitator to induce the evaporation of the liquid solvent. N2 is taken from a cylinder, heated in an electric heat exchanger and, then, sent to the precipitator.

The saturator (Sa) is a high-pressure vessel (I.V. 25 cm3) loaded with stainless steel perforated saddles; it provides a large contacting surface and a residence time long enough (5-6 min) to allow the dissolution of SC-CO2 in the liquid solution up to the saturation conditions at process pressure and temperature. Two heating bands cover the external surface, assuring uniformity of temperature inside the saturator and stability of the operating conditions.

The obtained solid-liquid-gas mixture, at the exit of the saturator, is sent into a thin wall stainless steel injector (I.D. 80 µm) to produce a spray of liquid droplets in the precipitator.

The precipitator is a stainless steel vessel operating at near atmospheric conditions. It is electrically heated using thin band heaters connected to a temperature controller. The powder precipitated into the precipitator is, then, collected at the bottom of the chamber on a stainless steel sintered filter.

Once out of the precipitator, the gases are discharged into a cooled condenser to separate the liquid solvent from N2 and CO2 by condensation.

Calibrated thermocouples, manometers, check valves, high pressure tubing and connections complete the apparatus.

During this Ph.D. thesis, the three different configurations of the precipitation vessel have been developed driven by the necessity of improving the performance of the powder collection devices.

In the first configuration, the precipitator consisted of four main parts: the head, the central body, the helicoidal flux conveyor and the collection filter.

� The head was supplied of three threaded holes: the central one to lodge the injection system, the two lateral ones to connect the nitrogen feed to the precipitator and to a manometer measuring the precipitator pressure. The head is remained unchanged in all the precipitator configurations.

� The precipitation chamber was cylindrical with an internal volume of 3 dm3. The powder was collected at the bottom of the precipitator on a stainless steel sintered plane disc.

� To reduce particle dispersion onto the vessel walls, a flux conveyor was introduced into the precipitator chamber to force the mixture formed by CO2, nitrogen and liquid solvent vapours to an ordinate motion inside the precipitator. This subsystem consisted of a

SUPERCRITICAL ASSISTED ATOMIZATION

29

helicoidal stainless steel device occupying all the section of the precipitator.

A photograph showing the helicoidal flux conveyor and the collection filter is reported in Figure IV.5.

Figure IV.5: Photograph of the helicoidal flux conveyor and the collection filter in the first configuration of the precipitator.

This kind of configuration often caused the failure of experiments. When large amount of precipitated powder covered the entire filter surface, the gaseous flux was hindered to flow out the chamber and lead to a pressure increase in the precipitator; if pressure exceeded 4-5 bar, the plane filter broke causing the loss of the micronized powder and the consequent substitution of this plant part for further experiments. Moreover, the pressures raising produced an increase in boiling temperature of liquid solution, thus causing solvent condensation in the chamber. Therefore, a different geometry of the collection device was developed to overcome these drawbacks.

The basic element of the second plant configuration was the filtering system. It was planned in cylindrical shape to impose radial direction, instead of the axial one as for the plane frit, on the gaseous stream. The gaseous stream flowed from the inner (lower radius) to the outer surface (larger radius) of the cylinder. The filter was connected to a vessel placed at the bottom, in which microparticles fell after depositing onto the filter. The new cylindrical system (mean pores diameter, 0.5 µm) assured the overcoming of process limits due to plane geometry: the large leaking surface (267 cm2) and the gases outlet in radial direction avoided the filter obstruction and allowed the collection of large amount of powder without the risk of filter breaking. Two flanges have been soldered at the filter to connect it to the central part and to the collection vessel of the precipitation chamber. Also in this configuration, the central body of the collection vessel had a cylindrical geometry. However, in this case, it was connected not only to the filter, but also to a gas conveyor, projected to guide exhaust gas out of the cylindrical filter, by flanges. The maximum pressure applicable to the

Chapter IV

30

gas conveyor was 38 bars. Heating bands were installed on the central body on the gas conveyor walls to provide uniform and faster heating. The upper and the lower section heating bands were connected to two different temperature controllers, since these two sections were exposed to different temperature gradients due to the chamber geometry. The parts constituting the second precipitator configuration are reported in Figure IV.6.

Figure IV.6: Second configuration of the precipitator: (a) Head; (b) central body; (c) cylindrical filter; (d) collection vessel; (e) gas conveyor.

Although some drawbacks of plane geometry filter were overcome, this kind of configuration was not optimal. Two main problems were due to the complex geometry of the chamber:

� Quite long time was needed to assemble the entire precipitation apparatus, and its encumbrance and heaviness resulted in not easy handling.

� Not uniform temperature profiles were established in the chamber, especially between the upper and the lower part of the apparatus, with consequent problems of temperature control and reproducibility of experiments.

Therefore, a further solution for the collection device was developed, showed in Figure IV.7. The filter has cylindrical geometry again, but, in this case, the gaseous stream flows from the outer (larger radius) to the inner (smaller radius) surface. The filter is connected to the central body of the chamber by a flange; when the apparatus is assembled, the entire filter is collocated into the inner section of the precipitator. The heating bands were installed on the precipitator chamber external walls and connected to a unique temperature controller. This configuration is simpler and more efficient than previous ones:

SUPERCRITICAL ASSISTED ATOMIZATION

31

� Only three parts constitutes the chamber (the head, the central body and the filter) resulting in easy and fast handling.

� No risk of filter breaking occurs due to its cylindrical geometry. � Uniform temperature profiles are obtained; indeed, in this case, the

filter is placed not at the bottom, but entirely inside the precipitator.

(a)

(b)

Figure IV.7: Third configuration of the precipitator: a) central body, b) filter.

To allow plant flexibility, the sintered filters, produced by cold isostatic pressing, have been projected with different pore size: 0.5, 1 and 3 µm as mean diameter. The smallest pore size filter is suitable to collect microparticles with small size but, necessarily, in small amounts; indeed, large amounts of powder can cause the obstruction of filter pores with the consequent pressure raising. This problem can be avoided using larger pore size filters; however, they presents as drawback the loss of a little amount of smaller microparticles that remain entrapped into the pores. Therefore, the selection of the adequate pore size filter is driven by the amount of powder to be collected and the size of precipitated microparticles.

IV.2 PROCESS DESCRIPTION

Atomization assisted by an inert gas is generally used in spray drying of solutions, but the innovative aspect of SAA process is the solubilization of supercritical carbon dioxide in a liquid solution and the subsequent

Chapter IV

32

atomisation of the gas-solid-liquid mixture. Indeed, inert gases are released slowly from the liquid phase and their contribution to a two steps atomization is not relevant (Huimin, 1999). Gases at supercritical conditions are released from the liquid phase by a faster process (Reverchon 2003); therefore, SC-CO2 can improve the atomization process, contributing to a two-step atomization.

The SAA process is based on the use of a packed saturator characterized by high specific surface and large residence times that can vary from 4 to 6 minutes at the commonly adopted process conditions. Up to now, ternary solute/solvent/SC-CO2 systems have been processed. In the saturator, SC-CO2 dissolves in the liquid solution (formed by an organic solvent or water and a solid solute) before the atomization in a thin wall injector, forming the ternary solution.

SAA takes advantage of two mechanisms of atomisation: pneumatic atomisation and decompressive atomisation (Figure IV.8). The pneumatic atomisation takes place when a high-speed fluid pushes the solution into the injector. Droplet dimensions depend on saturator pressure, injector diameter and geometry and some parameters, as flows, viscosities and surface tensions of the two fluids. This is also the principle of the generic gas atomization process (Lavernia&Wu, 1996). The decompressive atomization is, instead, caused by the mixing between CO2 and the solvent, and represents the key aspect of SAA process. Indeed, the efficiency of the atomization is due to the fast release of CO2. The rapid depressurization (occurring in passing from the high-pressure saturator into the law-pressure precipitator) makes the CO2 to move from the liquid solution in which was dissolved into the gaseous state. I.e., its abrupt moving away from inside to outside of droplets enhances their fragmentation. Then, the secondary droplets are rapidly dried by warm N2 causing the micrometric and submicrometric particles precipitation.

(a) (b)

Figure IV.8: (a) pneumatic atomization; (b) decompressive atomization

In the case of coprecipitates, we expect that mechanism of particles formation will be able to produce composite microspheres formed by a carrier and an active agent. Obviously, no more a ternary system is involved,

SUPERCRITICAL ASSISTED ATOMIZATION

33

but a quaternary one, formed by the drug, the carrier, the solvent and SC-CO2. The carrier and the drug will be dissolved in an opportune solvent able to solubilize both the compounds. The ternary solution will be sent into the saturator and will mix with the SC-CO2, thus forming the quaternary system: the droplets at the exit of the injector will be formed by both the solutes dissolved in the solvent and CO2. Once the solvent evaporated, the precipitated microparticles will be formed by the active compound uniformly dispersed in the carrier matrix. Indeed, the two components dissolved in the liquid droplets will remain entrapped in the solid particles, thus generating the desired microspheres (Figure IV.9).

drug polymer

solvent

Ternary solution

SC-CO2

Quaternary systems

microsphere

Pneumatic atomization

Decompressive atomization

drug polymersolvent

Ternary solution

SC-CO2

Quaternary systems

microsphere

drug polymersolvent

Ternary solution

SC-CO2

Quaternary systems

microsphere

Pneumatic atomization

Decompressive atomization

drug polymer

Quaternary system

Figure IV.9: Schematic representation of microsphere production by SAA process

IV.2.1 Process parameters

Relevant process parameters influencing the efficiency of the atomization process, the particle size (PS) and particle size distribution (PSD) are:

• the mass flow ratio (R) between CO2 and liquid solution, • the operating pressure (Pmix) and temperature (Tmix) in the saturator, • the solute concentration in the liquid solution (Csol), • the temperature in the precipitator (Tpr).

Moreover, in the case of coprecipitates, other important parameters should be considered:

• the carrier/drug/solvent ratio in the liquid solution, • the transition temperature (Tg) if a polymer is used as carrier.

Chapter IV

34

Pmix, Tmix, R and Csol influence the thermodynamics of the process. Although vapour-liquid equilibrium (VLE) data are available for many solvent/CO2 binary systems, quantitative data at high pressure on multiple (ternary and quaternary) systems is still missing and difficult to determine. For this reason, we usually consider the binary solvent/CO2 system involved at fixed process conditions (Pmix and Tmix) as the starting point to choose the operating conditions that assure the complete system miscibility. An empirical study of the effects of R and Csol on VLE is, then, necessary to optimise the process conditions. The presence of the third and fourth component could significantly modify the VLE respect to the binary one. This modification often consists of the shifting of the mixture critical point towards higher pressures and it is the more pronounced the higher is Csol (Figure IV.9).

When Pmix and Tmix are fixed, the mass flow ratio R corresponds to a given operating point into the VLE diagram of the system considered. A change of the molar fractions of system components produces a modification of R, and if R increases, the operating point shifts towards a richer in CO2

phase (Figure IV.10). Since an efficient atomization process occurs when the operating point falls in the liquid phase, the choice of opportune R-values is fundamental to the optimisation of the process.

0.0 0.2 0.4 0.6 0.8 1.00

50

100

150

200

Pre

ssure

, bar

xCO

2

RR

TT

CC

LL

GG

SCSC

0.0 0.2 0.4 0.6 0.8 1.00

50

100

150

200

Pre

ssure

, bar

xCO

2

RR

TT

CC

0.0 0.2 0.4 0.6 0.8 1.00

50

100

150

200

Pre

ssure

, bar

xCO

2

RRRR

TTTT

CCCC

LL

GG

SCSC

Figure IV.10: Effect of the process parameter on the miscibility hole. The continuous line refers to the binary system DMSO-CO2; the dash line refers to the ternary system DMSO-CO2-solute.

R and Csol not only influence the thermodynamic system behaviour, but also PS and PSD of precipitated particles. Previous results showed that the higher R the lower PS and narrow PSD resulted (Reverchon&Spada, 2004). Indeed, as R increases a greater amount of CO2 is dissolved in the liquid

SUPERCRITICAL ASSISTED ATOMIZATION

35

solution and an enhanced decompressive atomization occurs. On the other hand, usually the higher Csol the higher the PS and the larger PSD are obtained (Reverchon&Della Porta, 2003a, 2003b, 2003c). Indeed, high Csol may result in an increased solution viscosity and a non-negligible ebullioscope raising, thus influencing both the atomization and evaporation processes.

In the case of coprecipitates, where the solutes are two (carrier and drug), the study of Csol effect is more complex since several carrier/drug/solvent ratios are possible. Hence, an investigation about the carrier/drug/solvent ratio is necessary to individuate which one provides the desired release times and good controlled drug release.

The temperature in the precipitator is another important parameter to be monitored, since it can affect particle morphology. Indeed, Tpr should be high enough to allow the solvent evaporation from the droplets, but it should be lower than the degradation temperature of the active agent. Moreover, it could influence the solid state of precipitated microparticles since high temperature could favour their crystallization process.

When a polymeric material is used as carrier, the effect that Tpr has on the solid state of the polymer should be considered. Indeed, at temperature higher than the polymer Tg the polymer is in the rubbery state and sticky particles could result.

Chapter IV

36

Chapter V ANALYTICAL METHODS

V.1 SCANNING ELECTRON MICROSCOPE (SEM)

Morphological characteristics of SAA processed powders are analysed by a Scanning Electron Microscope (SEM, mod. 420, LEO, Germany). Powders are dispersed on a carbon tab previously stuck to an aluminum stub (Agar Scientific, UK). Samples are coated with gold-palladium (layer thickness 250Å) using a sputter coater (mod. 108A, Agar Scientific, UK).

The Scanning Electron Microscope uses electrons rather than light to form an image. There are many advantages to using the SEM instead of a light microscope. The SEM has a large depth of field, which allows a large amount of the sample to be in focus at one time. The SEM also produces images of high resolution, which means that closely spaced features can be examined at a high magnification. Preparation of the samples is relatively easy since most SEMs only require the sample to be conductive.

A scanning electron microscope is mainly formed by an electron gun, the condensing lenses, an objective lens and an electron detector, as schematically represented in Figure V.1.

Figure V.1: Schematic diagram of a Scanning Electron Microscope.

Chapter V

38

V.2 PARTICLE SIZE DISTRIBUTIONS

In this work, particle size analysis is performed on SAA produced microparticles using two techniques: image analysis on SEM pictures and laser scattering (LS)

In the image analysis, PSDs are measured from SEM images using the Sigma Scan Pro Software (rel. 5.0, Jandel Scientific, Erkrath, Germany). About 1000 particle diameters are considered in each PSD calculation. Histograms representing PSDs in terms of particles number are calculated using Microcal Origin Software (rel. 6.0, Microcal Software Inc., Northampton, USA); then, they are converted in volumetric distributions and plotted in a cumulative form.

Laser scattering analysis is performed using a Malvern Mastersizer S laser diffractometer (Alfatest s.r.l., Rome, Italy). Size range detectable with this instrument is 0.05 - 900µm. The powders are suspended in a suitable dispersant, then, sonicated for some minutes before analysis.

Laser diffraction particle size analysis is based on the scientific phenomenon that particles in a laser beam scatter laser light at angles that are inversely proportional to the size of the particles: i.e., large particles scatter at small forward angles, whereas small particles scatter light at wider angles. The primary key to consistent and accurate particle size measurement is the ability to present a well-dispersed, homogenous sample to the laser beam of the system. By the use of Fourier and Reverse Fourier optics, this scattering is imaged to an array of detectors at the focal plane of the optics. There is a direct relationship between the distribution of the scattered light energy on these detectors and the particle size distribution which gives rise to it. The Mastersizer S software uses Mie theory to obtain an optimal analysis of this light energy distribution to arrive at the size distribution of the particles.

Figure V.2: Schematic functioning of laser diffractrometer.

ANALYTICAL METHODS

39

V.3 DIFFERENTIAL SCANNING CALORIMETER (DSC)

Differential Scanning Calorimeter (DSC) (DSC mod. TC11, Mettler, USA) is used to investigate microparticles solid state, to calculate their crystalline degree and to obtain information about polymers transition temperature (Tg) and structure.

Differential Scanning Calorimetry is a thermal analysis technique which is used to measure the temperatures and heat flows associated with transitions in materials as a function of time and temperature. Such measurements provide qualitative and quantitative information about physical and chemical changes that involve endothermic and exothermic processes, or changes in heat capacity.

The DSC measurements are based on the heat flow exchanges (dH/dt) between the temperature control system of the apparatus and the sample being analysed, as well as the temperature control system and the reference sample. Each sample is put into a pan and subsequently placed inside a cell with a continuous nitrogen flow circulating, in order to ensure a constant heat transfer coefficient. The DSC furnace made of pure silver is heated by means of a coaxial heater coil, while cooling is provided by a cold nitrogen gas flowing through the gap between the furnace and the infrared reflector. To avoid the forming of ice in the furnace during intensive cooling, a dry purge gas is used (usually 50 mL/min). The gas enters the furnace body from the bottom, where it is heated to the cell temperature and led into the sample chamber. It finally escapes through the hole in the centre of the lid. The DSC cell is also equipped with a temperature sensor placed on a disk in direct thermal contact with the heater plate of the silver furnace.

The phase transitions and chemical reactions are then shown on a thermogram in the form of endothermic or exothermic peaks, while the glass transition as an endothermic shift of the baseline.

Figure V.3: Schematic representation of DSC furnace.

Chapter V

40

V.4 X-RAY

Diffraction patterns of SAA precipitated powders were obtained using an X-ray diffractometer (mod. D8 Discover, Bruker, USA) with a Cu sealed tube source.

X-ray diffraction is a versatile, non-destructive analytical technique for identification and quantitative determination of the various crystalline compounds, known as 'phases', present in solid materials and powders.

A crystal structure is a regular three-dimensional distribution (cubic, rhombic, etc.) of atoms in space. These are arranged so that they form a series of parallel planes separated from one another by a distance d, which varies according to the nature of the material. For any crystal, planes exist in a number of different orientations - each with its own specific d-spacing.

The diffraction is regulated by the Bragg's Law:

θλ sin2dn =

When a monochromatic X-ray beam with wavelength lambda is projected onto a crystalline material at an angle theta, diffraction occurs only when the distance travelled by the rays reflected from successive planes differs by a complete number n of wavelengths.

By varying the angle theta, the Bragg's Law conditions are satisfied by different d-spacings in polycrystalline materials. Plotting the angular positions and intensities of the resultant diffracted peaks of radiation produces a pattern that is characteristic of the sample. Where a mixture of different phases is present, the resultant diffractogram is formed by addition of the individual patterns.

A schematic diagram of an X-ray equipment is reported in Figure V.4.

Figure V.4: Schematic diagram of an X-ray equipment

Samples analyzed in this work have been placed in the holder and flattened with a glass slide to assure a good surface texture. The measuring conditions are as follows: Ni-filtered CuKα radiation, λ=1.54 Å, 2θ angle

ANALYTICAL METHODS

41

ranging between 10° and 70° with a scan rate of 3 seconds/step and a step size of 0.2°.

V.5 UV-VISIBLE SPECTROSCOPY

UV-vis spectrophotometer (Cary 50 Scan, Varian) was used to analyze drug release of SAA coprecipitated powders.

To obtain absorption information, a sample is placed in the spectrophotometer and ultraviolet and/or visible light at a certain wavelength (or range of wavelengths) is shone through the sample. The spectrophotometer measures how much of the light is absorbed by the sample. Indeed, when white light passes through or is reflected by a coloured substance, a characteristic portion of the mixed wavelengths is absorbed.

The absorption of UV or visible radiation corresponds to the excitation of outer electrons of molecules. When an atom or molecule absorbs energy, electrons are promoted from their ground state to an excited state. When sample molecules are exposed to light having an energy that matches a possible electronic transition within the molecule, some of the light energy will be absorbed as the electron is promoted to a higher energy orbital. The spectrometer records the wavelengths at which absorption occurs, together with the degree of absorption at each wavelength.

How the spectrophotometer works is shown in Figure V.5: a beam of light produced by a visible and/or UV light source is separated into its component wavelengths by a prism or diffraction grating. Then, the beam passes through a small transparent container (cuvette) containing the sample to be analyzed. The spectrometer records the wavelengths at which absorption occurs, together with the degree of absorption at each wavelength.

Sample

Xenon lamp

Light beam

Prism

Figure V.5: Functioning of Cary 50 Scan

Chapter V

42

The resulting spectrum is presented as a graph of absorbance (A) versus wavelength. By an opportune previous calibration, this instrument reads a compound concentration in a solution, allowing the monitoring drug release times from a carrier matrix.

V.6 ENERGY DISPERSIVE MICROANALYSIS

SAA coprecipitated powders are characterized by microanalysis (INCA Energy 350, Oxford Instruments) to investigate their chemical structure.

Energy Dispersive X-Ray Spectroscopy (EDS/EDX), or Energy Dispersive Microanalysis, is the measurement of X-rays emitted during electron bombardment in an electron microscope (SEM or TEM) to determine the chemical composition of materials on the micro- and nano- scale. Indeed, an electron microscope uses a focused electron beam to interact with the atoms in a sample. As the electron beam displaces electrons in the sample, detection equipment converts the electrons scattered by the electron beam into a microscopic image. Another phenomenon occurs in this interaction: the generation of characteristic X-rays.

Figure V.6: Generation of X-ray by an excited atom.

In order to return the atom to its normal state, an electron from an outer atomic shell “drops” into the vacancy in the inner shell. This drop results in the loss of a specific amount of energy, namely, the difference in energy between the vacant shell and the shell contributing the electron. This energy is given up in the form of electromagnetic radiation X-rays. Since energy levels in all elements are different, element-specific, or characteristic, X-rays are generated.

Energy dispersive X-ray microanalysis uses detection equipment to measure the energy values of the characteristic X-rays generated within the electron microscope. Using semiconductor material, such as Silicium, to

ANALYTICAL METHODS

43

detect the X-rays and a multichannel analyzer, an X-ray microanalysis system converts X-ray energy into an electronic “count.”

Figure V.7: Cutway of a Si X-ray Detector.

The accumulation of these energy counts creates a spectrum representing the chemical analysis of the sample.

Therefore, while the electron microscope produces an image of the sample topography, energy dispersive X-ray microanalysis reveals what elements are present in the sample (qualitative analysis). The rate of detection of these characteristic X-rays is used to measure the amounts of elements present (quantitative analysis).

If the electron beam is rastered over an area of the sample then EDS systems can also acquire X-ray maps showing spatial variation of elements in the sample.

Chapter V

44

Chapter VI RESULTS PART I: SAA

CARRIERS PROCESSING Two approaches are commonly used to produce microparticles for drug

controlled release: the first is the formation of polymer particles and, then, their impregnation with the active ingredients; the second is the direct production of composite particles formed by the polymer and the active compound. In every case, polymer particles formation is a crucial step for the preparation of drug loaded polymer particles. The dimension, the shape and the structure of the polymeric matrix play a relevant role, since they influence the way of administration to the patient and the drug release rate in the organism.

Therefore, the work developed during this Ph.D has been divided into two sections:

� The study of SAA processability of different kinds of carriers for drug release;

� The coprecipitation of some carriers previously investigated with model drugs.

Different types of polymers have been processed by SAA, both synthetic and natural ones, to investigate the effect of the process parameters on their structure and the possible microparticles production. Poly(methyl methacrylate) (PMMA) and poly-L-lactide (PLLA) have been selected as model biocompatible synthetic polymers, since they are largely used in pharmaceutical applications; whereas, chitosan has been selected as natural polymer, because of the attention as promising renewable material received in the last years.

The potential of SAA in the production of cyclodextrin microparticles has been investigated as well, as the first stage to the direct production of micronized drug/CD complexes. In particular α-cyclodextrin (α-CD) and hydroxypropyl-β-cyclodextrin (HP-β-CD), two of the most employed CDs in pharmaceutical field, have been processed.

Chapter VI

46

VI.1 SYNTHETIC POLYMERS MICROPARTICLES

PRODUTION: PMMA AND PLLA

PMMA is one of the most hydrophobic non-biodegradable polymers used for drug delivery systems. It has the advantage of being essentially inert and does not undergo any chemical change in vivo. PMMA microspheres have successfully been used as substrate for the immobilization of enzymes (Bulmus et al., 1997), for the design of microfluidic devices, for metal deposition (Henry and McCarley, 2001) and for immobilization of proteins and DNA (Waddell et al., 2000; Henry et al., 2000; Fixe et al., 2004).

Poly-L-lactic acid (PLLA) is a biocompatible, biodegradable, and immunologically inert synthetic polymer. PLLA microparticles are widely used as drug delivery systems (Lewis, 1990; Jalil and Nixon, 1990; Wake et al., 1998; Flick-Smith et al., 2002; Liggins and Burt, 2004). A new trend in medicine is the use of polymer microparticles as dermal fillers; for example, injectable PLLA microparticles, mixed with water, can be injected to increase the thickness of the skin (Köse et al., 2002).

The effect of some relevant process parameters, like mass flow ratio between liquid solution and CO2, temperature in the precipitator and solute concentration, on PMMA and PLLA particle size and particle size distribution (PSD) has been studied. Polymer microparticles have been characterized by scanning electron microscopy, particle size measurement and differential scanning calorimetry.

VI.1.1 Results and Discussion

Acetone and dichloromethane (DCM) have been selected as liquid solvents for PMMA and PLLA respectively, since PMMA solubility in acetone is larger than 300 mg/mL and PLLA solubility in DCM is larger than 100 mg/mL.

Untreated PMMA and PLLA used in this work were formed by irregularly shaped particles ranging between 100 and 500 µm and by pellets of some millimeters, respectively.

VI.1.1.1 Selection of the saturator operating parameters

The solubilization of SC-CO2 in the liquid solution inside the saturator is one of the key steps controlling the efficiency of the SAA process. CO2 solubility in liquids depends on the liquid solvent chosen, on temperature and pressure in the saturator, since it is related to high pressure Vapor Liquid Equilibria (VLEs) characteristic of the selected system liquid-CO2.

RESULTS PART I: SAA CARRIERS PROCESSING

47

Data on high pressure VLEs for the binary systems acetone/CO2 and DCM/CO2 are available in literature only in a limited range of temperature (Vonderheiden and Eldridge, 1963; Ohe, 1990; Chang et al., 1997; Gonzalez et al., 2002). Moreover, the possibility that the solute can modify binary VLE has to be taken into account, but the modification caused by the presence of polymers on the binary VLEs is not available. Therefore, a screening of SAA operating conditions has been first performed and, in preliminary experiments on PMMA and PLLA, saturator operating pressure and temperature have been varied in the range between 70 and 90 bar and between 70 and 90°C, respectively. For both the polymers, the best results in terms of stability of the process and polymer precipitated particles have been observed operating at 80 bar and 80°C. Therefore, these pressure and temperature conditions have been used in all the subsequent SAA experiments proposed in this work.

VI.1.1.2 Effect of the mass flow ratio CO2 / liquid solution (R)

Fixed pressure and temperature, the mass flow ratio (R) determines the operating point of the process in the ternary VLE diagram of the polymer/CO2/liquid solvent system.

The formation of an homogenous liquid phase in the saturator is required, but in presence of the solute the vapor-liquid equilibrium could be significantly modified, especially in the case of high-molecular weight compounds like polymers. Therefore, at least an empirical study of stability of the process and the effect of R on particle morphology is necessary.

At fixed value of solute concentration (Csol=50 mg/mL) R values ranging between 0.9 and 1.6 were explored in the experiments on PMMA.

The influence of R on SAA processed PMMA can be qualitatively evaluated from Figure VI.1, in which SEM images of processed material at various R are reported. For R values ranging between 0.9 and 1.2, PMMA microparticles have been successfully produced with a morphology formed by well-defined spheres and collapsed spheres (doughnuts); whereas, at R=1.6 a marked particles coalescence has been observed. Moreover, during the experiments performed at R=1.6, pressure fluctuations have occurred and precipitated material has been found in the saturator. These observations lead to the hypothesis that an antisolvent effect of SC-CO2 on the system PMMA+acetone takes place at this R value; i.e., pressure fluctuations are caused by PMMA precipitation in the saturator.

Chapter VI

48

Figure VI.1: SEM images of PMMA precipitated by SAA from acetone at different mass flow ratios (R), at Csol=50 mg/mL and 60°C in the precipitator.

SEM images proposed in Figure VI.1 are taken at the same enlargement (20K); therefore, they also allow a qualitative evaluation of the increase in particle size and polydispersity when R varies between 0.9 and 1.2. PSDs at different R values have been calculated from SEM images analysis on the basis of particles number; then, they have been converted in terms of volumetric distributions and plotted in a cumulative form. Volume based distributions are reported in Figure VI.2.

R=1.2 R=1.6

R=0.9 R=1

RESULTS PART I: SAA CARRIERS PROCESSING

49

0.0 0.3 0.6 0.9 1.2 1.5 1.8

20

40

60

80

100R=1.2

R=1.0

R=0.9

Vo

lum

e,

%

Particle size, µm

Figure VI.2: PSDs of micronized PMMA at different R. Calculations in terms of particle volume percentages.

All particles are smaller than about 1.6 µm and a decrease of the mean particle size can be observed when R increases; for instance, at R=0.9, 90% of the total volume of PMMA powder is related to particles smaller than 1.27 µm, whereas at R=1.2 the same percentage is related to particles smaller than 0.95 µm.

A possible explanation of the effect of R on particle size and distribution can be proposed in terms of the molar fraction of CO2 dissolved in the solution formed in the saturator. In the hypothesis that all feed CO2 dissolves in the liquid, R=0.9 and 1.2 correspond to mole fractions of CO2 in acetone of 0.54 and 0.61, respectively. The higher is R, the greater is the quantity of CO2 solubilized in the liquid phase; therefore, a stronger decompressive atomization results and smaller microparticles are precipitated.

In the experiments performed on PLLA, R values ranging between 1.2 and 2 have been explored at a fixed solute concentration of 20 mg/mL. SEM images of particles recovered at different R are reported in Figure VI.3.

Chapter VI

50

Figure VI.3: SEM images of PLLA precipitated by SAA from DCM at different mass flow ratios (R), at Csol=20 mg/mL and 57°C in the precipitator.

Well-defined microparticles have been obtained at R ranging between 1.5 and 2, whereas coalescing particles resulted at R=1.2. Therefore, this R value has been no more used in PLLA experiments.

Also in this case, SEM images (Figure VI.3) are proposed at the same magnification (20K) and allow a qualitative evaluation of the increase in particle size and polydispersity as R varies from 1.5 to 2. Volumetric PSD are reported in Figure VI.4; they show that all particles are smaller than 3.5 µm and confirm the former qualitative considerations. Indeed, 90% of the total volume occupied by PLLA micronized powder is related to particles smaller than 1.44 µm at R=1.5 and to particles smaller than 0.49 µm at R=2.

R=1.2 R=1.5

R=2 R=1.8

RESULTS PART I: SAA CARRIERS PROCESSING

51

0 1 2 3 4

20

40

60

80

100

R=1.8

R=2

R=1.5

Vo

lum

e,

%

Particle size, µm

Figure VI.4: PSDs of micronized PLLA at different R. Calculations in terms of particle volume percentages.

The experiments performed at different R for both polymers have showed that morphology and PSD of PMMA and PLLA microparticles are influenced by this process parameter.

The overall effect of the mass flow ratio could be explained by a qualitative thermodynamic interpretation of the results. In the case of PMMA, a qualitative vapour-liquid diagram at 80°C of the system acetone-CO2 is reported in Figure VI.5, on which the different SAA operating points have been placed. The miscibility hole at 80°C has been obtained using the Soave-Redlich-Kwong Equation of State (SRK EoS). The same equation has been also used to fit experimental data at different temperatures when available; the results are reported in Figure VI.5. Points A, B, C and D represent the operating conditions for R=0.9, 1.0, 1.2 and 1.6, respectively. If the diagrams reflects the VLEs at the SAA operating conditions, points A, B, C are related to experiments performed in the liquid-rich phase of the diagram, that have produced well-defined PMMA microparticles with the PSD depending on R. Moving from point C to D, the operating point probably falls in the miscibility hole, resulting in the formation of two phases (one rich in CO2 and one rich in liquid). In the gas phase, CO2 probably plays as an antisolvent on the system acetone+PMMA and causes the precipitation of the polymer in the saturator; in the liquid phase, the higher concentration can cause an increase of the boiling point of the solution, resulting in the non-complete drying of the droplets and the coalescence of particles collected in the precipitator.

Chapter VI

52

0.0 0.2 0.4 0.6 0.8 1.0

2

4

6

8DCBA

30°C 35°C 40°C 80°C

P,

MP

a

xCO

2

Figure VI.5. Vapour-liquid equilibria for the system CO2-acetone: (○) at 30°C; (▲) at 35°C; (■) at 40°C [40]. (▬) Vapour-liquid equilibrium curve at 80°C using SRK EoS. Points A, B, C and D are referred to R-values of 0.9, 1.0, 1.2 and 1.6, respectively.

Similar considerations could be applied to the system PLLA-DCM-CO2, since PSD of PLLA microparticles shows a similar dependence on R to PMMA. The coalescence of PLLA microparticles observed at R=1.2 could be caused by various factors, such as interactions of PLLA with DCM and CO2 or a non-efficient pneumatic atomization.

VI.1.1.3 Effect of the operating temperature in the precipitator

Temperature optimization in the precipitator is required to assist droplets evaporation minimizing the stress on the treated compound.

Temperatures lower than PMMA and PLLA glass transition temperature have been used in the precipitator, to prevent the plasticization of micronized polymer particles; indeed, plasticization can induce particles coalescence, that results in the failure of the micronization process.

DSC analysis performed on untreated polymers has shown glass transition temperatures (Tg) of 125°C and ranging from 60 to 80°C for PMMA and PLLA, respectively. DSC thermograms have shown only a slight decrease of the Tg of SAA processed PMMA and PLLA with respect to the untreated polymers, probably related to SAA processing.

SAA experiments on PMMA have been performed at 60 and 70°C in the precipitator; i.e., very far from the Tg of this polymer. PMMA microparticles

RESULTS PART I: SAA CARRIERS PROCESSING

53

have shown no significative differences in morphology and well-defined spherical or “doughnut” particles have been obtained.

Experiments on PLLA have been performed at temperatures of 57°C and 67°C; i.e., very near the Tg of this polymer. Therefore, although an increment of only 10°C has been tested, very different results have been obtained in terms of particles morphology: well-defined PLLA microparticles have been recovered in the precipitator at 57°C, as shown in the SEM image in Figure VI.3; whereas, a tow has been recovered in the precipitator when the temperature of 67°C has been used. It was formed by very thin PLLA filaments networked each other to form a web. SEM image reported in Figure VI.6, shows micrometric particles trapped in these nanometric filaments. Probably, droplets had been successfully produced, but at 67°C the plasticization of the polymer has been induced and particles have been “extruded” in form of filaments.

Figure VI.6: SEM image of PLLA precipitated by SAA from DCM at a precipitator temperature of 67°C.

In conclusion, as expected, a successful micronization of PMMA and PLLA by SAA is possible when the temperature in the precipitator is sufficiently far from the glass transition temperature of the polymer.

In the case of PMMA, operating at temperatures lower than 80°C (down to 60°C) an efficient evaporation of acetone has been obtained. In the case of PLLA, the range of operating temperatures in the precipitator is narrower, since Tg is near DCM boiling temperature. At 57°C PLLA microparticles have been obtained with an efficient evaporation of DCM; therefore, this temperature value has been selected as the operating temperature for all the experiments on PLLA.

VI.1.1.4 Effect of solute concentration

Systematic experiments have been performed operating at different PMMA concentrations (Csol) in acetone between 1.5 and 50 mg/mL, at

Chapter VI

54

R=1.2. PMMA powder recovered in these experiments has been always formed by well-defined non-coalescing microparticles. Examples of the particles collected at different Csol are shown in SEM images reported in Figure VI.7. SEM images have been produced at the same enlargement (20K), thus allowing a qualitative evaluation of the increase in particle size with solute concentration.

Figure VI.7: SEM images of micronized PMMA obtained at different Csol

values, at R=1.2 and 60°C in the precipitator.

Volumetric PSDs at Csol = 1.5, 10 and 50 mg/mL are reported in Figure VI.8; they confirm the qualitative analysis performed in Figure VI.7; 90% of the total volume occupied by PMMA micronized powder is due to particles smaller than 0.95 µm at Csol=50 mg/mL and smaller than 0.58 µm at Csol=1.5 mg/mL.

10mg/mL 1.5mg/mL

50mg/mL 30mg/mL

RESULTS PART I: SAA CARRIERS PROCESSING

55

0.0 0.2 0.4 0.6 0.8 1.0 1.2

20

40

60

80

100

C=50 mg/mL

C=30 mg/mL

C=1.5 mg/mL

Vo

lum

e,

%

Particle size, µm

Figure VI.8: PSDs of micronized PMMA at different Csol values. Calculations in terms of particle volume percentages.

The effect of the solute concentration on particles size can be explained considering some physical characteristics of the solution, such as viscosity and surface tension. An increase of Csol causes an increase of viscosity and of surface tension of the liquid solution, resulting in the formation of larger primary droplets and influences also the formation and the dimension of the secondary droplets.

VI.1.1.5 Conclusions on PMMA and PLLA

SAA technique can successfully produce PMMA and PLLA microparticles.

Morphology, particle size and distributions of PMMA and PLLA microparticles have shown a dependence on the mass flow ratio R. The study of the effect of the temperature in the precipitator on polymer morphology have revealed that this is a key parameter in the polymer processability: a successfully micronization of PMMA and PLLA is possible at temperature values in the precipitator far from their glass transition temperature, but high enough to allow a fast solvent evaporation.

Obtained microparticles can be used as the starting material for the production of drug loaded microparticles for drug controlled release.

Chapter VI

56

VI.2 NATURAL POLYMER MICROPARTICLES PRODUCTION:

CHITOSAN

Chitosan (CH) is a non-toxic, biodegradable and biocompatible polymer. It is a natural linear biopolyaminosaccharide derived by alkaline deacetylation of chitin, the second most abundant natural biopolymer found in nature after cellulose. Chitin is the principal component of protective cuticles of crustaceans such as crab, shrimp, prawns, lobsters and cell walls of some fungi such as aspergillus and mucor (No and Meyers, 1989). CH is insoluble at neutral and alkaline pH values, but forms salts with inorganic and organic acids such as glutamic acid, hydrochloric acid, lactic acid and acetic acid. Chitosan salts are soluble in water, the solubility being dependent on the degree of acetylation and pH. Usually, 1–3% aqueous acetic acid solutions are used to solubilize CH.

Over the last several years, chitinous polymers, especially chitosan, have received increased attention as promising renewable polymeric materials. CH presents special properties, due to its chemical structure, that make it useful in many fields. Particularly, properties such as biodegradability, low toxicity and good biocompatibility make CH particularly suitable for use in biomedical and pharmaceutical formulations. It can be used for hypobilirubinaemic and hypocholesterolemic effects, antacid and antiulcer activities, immobilization of enzymes and living cell and in ophthalmology (Felt et al., 1998). CH has shown excellent properties as excipients and has been used as a vehicle for directly compressed tablets, as a disintegrant, as a binder, as a granulating agent, in ground mixtures, as well as a co-grinding diluent for the enhancement of dissolution rate and bioavailability of water insoluble drugs (Kristmundsdottir et al., 1995; Illum, 1998; Miyazaki et al., 1981; Shiraishi et al., 1990).

CH has also been widely investigated for its potential in the development of controlled release drug delivery systems. Particularly, it has mucoadhesive properties due to molecular attractive forces formed by electrostatic interaction between positively charged CH and negatively charged mucosal surfaces, and thus promotes drugs transmucosal absorption. It has been exploited for nasal and oral delivery of polar drugs, to include peptides and proteins, for vaccine delivery (Lehr et al., 1992; He et al., 1998), as well as a drug carrier for sustained release preparations (Miyazaki et al., 1990; Akbuga, 1993; Kristl et al., 1993).

CH microspheres are used to provide controlled release of many drugs and to improve the bioavailability of degradable substances such as protein or enhance the uptake of hydrophilic substances across the epithelial layers. CH microspheres are being investigated for parenteral and oral drug delivery and are the most widely studied drug delivery systems for the controlled

RESULTS PART I: SAA CARRIERS PROCESSING

57

release of drugs such as antibiotics, antihypertensive agents, anticancer agents, proteins, peptide drugs and vaccines.

Different methods have been used to prepare CH particulate systems. Some conventional processes for CH microparticles production are emulsion cross-linking, coacervation/precipitation, ionic gelation, reverse micelles formation, solvent evaporation and spray-drying. Many reports and reviews are available on the preparation of CH microspheres (Brannon-Peppas, 1995; Sinha et al., 2004; Agnihotri et al., 2004). These conventional processes show several drawbacks, such as the use of organic solvents and hard-to-separate surfactants (with many consequent following treatments to reduce solvent residue below the safety limits), high temperatures and limited control of particle size and particle size distribution (Luisi et al., 1988; Akbuga and Durmaz, 1994; Genta et al., 1994; Lorenzo-Lamosa et al., 1998). But, factors such as particle size requirement, thermal and chemical stability of the active agent, reproducibility of the release kinetic profiles, stability of the final product and residual toxicity associated with the final product are particularly relevant for pharmaceutical formulations.

CH porous structures obtained using supercritical carbon dioxide (SC-CO2) (Le et al., 2005) and CH-indomethacine composites via SCF assisted impregnation have been proposed (Gong et al., 2005); but, to date, CH microparticles have not been produced using SC-CO2 based processes. Indeed, solubility of CH in SC-CO2 is very low; therefore, processes based on the solvent or solute effect of CO2 (such as in RESS and PGSS) cannot be successfully applied to CH micronization. Moreover, since usually 1–3% aqueous acetic acid solutions are used to solubilize CH, processes based on the antisolvent effect of CO2 cannot be used, because SC-CO2 is not an effective solvent for diluted acetic acid-water solutions (Briones et al., 1987).

The SAA characteristic of being able to operate not only with organic solvents, but also with aqueous solutions, suggests the possibility of CH microparticles production using this technique. Therefore, the processability of CH by SAA and the performance of this technique have been investigated using as solvent a 1% v/v acetic acid aqueous solution. The effect of process parameters such as temperature in the precipitator and solute concentration in the liquid solution on CH particles morphology, size and distribution have been studied to evaluate the possibility of particle size tailoring.

Low molecular weight chitosan with degree of deacetylation, 84.0%; viscosity, 70 MPa⋅s (measured in 1% w/v of chitosan in 1% v/v acetic acid aqueous solution) was used as raw material.

Chapter VI

58

VI.2.1 Results and Discussion

SAA experimentation on CH has been characterized by the following original aspects:

� the high value product treated, because of its special properties and extensive applications in the pharmaceutical and biomedical industries;

� the challenge of processing this particular compound by a SCF based-technique;

� the innovative use of a multi-component mixture as solvent in the SAA process.

Since 1% v/v acetic acid aqueous solution had never been employed in SAA, a preliminary feasibility experiment has been necessary. Process conditions have been selected according to previous experimentations performed using pure water (Reverchon and Della Porta, 2003a, 2003b, 2003c; Reverchon et al., 2003a). In particular the following conditions have been used: mass feed ratio of 1.8 between CO2 and solution; pressure and temperature in the saturator at 80 bar and 85°C; temperature in the precipitator at 110°C. This experiment has presented no process problems and the powder recovered was formed by well-defined spherical microparticles, as can be observed in Figure VI.9.

Figure VI.9: SEM image of chitosan microparticles precipitated by SAA from 1% acetic acid aqueous solution at R=1.8, Csol = 5 mg/mL and Tp=110°C.

Once verified the processability of CH by SAA, a systematic study on the effect of process parameters on particles morphology and size has been performed.

RESULTS PART I: SAA CARRIERS PROCESSING

59

VI.2.1.1 Selection of the saturator operating parameters

The maximum quantity of CO2 that can be solubilized in a 1% acetic acid aqueous solution depends on the temperature and pressure in the saturator. To set pressure and temperature conditions in the saturator, precise values of CO2 solubility in the liquid solvent should be known.

Data on high pressure solubilities for acetic acid-water mixtures containing SC-CO2 are available in literature only in a limited range of temperature (Briones et al., 1987). Moreover, the possibility that the solute can modify ternary VLEs has to be taken into account, but this information is not available. In every case, SC-CO2 solubility in aqueous solutions is low at all operating conditions. Therefore, CO2 in excess with respect to the expected solubilization concentration has been used. This condition also assures a readily obtainment of the process pressure in the saturator and an enhanced pneumatic atomization. A mass feed ratio (R) between CO2 and the liquid solution of 1.8 has been used in all the experiments, since it showed to be the most suitable for aqueous solutions processing in previous SAA works (Reverchon and Della Porta, 2003a, 2003b, 2003c; Reverchon et al., 2003a).

According to these considerations, some tests have been performed by setting the saturator operating conditions in a pressure range from 80 to 150 bar and in a temperature range between 70 and 90 °C. The best results in terms of stability of the process and morphology of CH precipitated particles have been observed operating at 105 bar and 85°C. Therefore, these pressure and temperature conditions have been used in all the subsequent SAA experiments proposed in this work.

VI.2.1.2 Effect of temperature in the precipitator

The effect of temperature on CH particle morphology has been observed performing experiments at temperatures ranging between 87 and 135°C. All these experiments have been performed at a CH concentration in the liquid solution of 5 mg/mL; the results are summarized in Table VI.1.

Chapter VI

60

Table VI.1: Effect of the chitosan concentration in the liquid solution and of the temperature in the precipitator on the morphology of CH particles precipitated by SAA from 1% acetic acid aqueous solution.

CH concentration Temperature in the precipitator Morphology

1 mg/mL 95°C Spherical

microparticles

87 °C Solvent condensation

95°C Spherical

microparticles

105 °C Spherical

microparticles

5 mg/mL

135 °C Spherical

microparticles

95°C Solvent condensation

10 mg/mL 106°C

Spherical microparticles

106°C Solvent condensation

15 mg/mL 126°C

Spherical microparticles

At precipitation temperatures between 90 and 135°C, well-defined and

non-coalescing spherical microparticles have been obtained, as shown by SEM images reported in Figure VI.10. At temperatures lower than 90°C substantially no powder has been recovered on the precipitator walls and part of the liquid solvent has been found at the bottom of the precipitator. SEM image of the very small amount of powder collected in the precipitator have revealed that collapsed and coalescing particles were obtained, as can be observed in Figure VI.10a. This result can be explained considering that these precipitation temperatures can induce a partial recondensation of the solvent on the precipitated particles, or even a non-efficient evaporation of the droplets; the liquid is, then, forced to pass through the stainless filter by the N2 stream and no solute is collected in the precipitator.

Therefore, the subsequent experiments on CH have been performed at 95°C in the precipitator to avoid solvent condensation and to minimize the thermal stress on the precipitated microparticles.

RESULTS PART I: SAA CARRIERS PROCESSING

61

Figure VI.10: SEM images of chitosan microparticles precipitated by SAA from 1% acetic acid aqueous solution at precipitation temperatures of 87 °C (a), 95°C (b) and 135°C (c) (R=1.8, Csol = 5 mg/mL).

VI.2.1.3 Effect of solute concentration

Previous works on SAA showed that solute concentration in the liquid solution is a relevant parameter in the control of particle size and particle size distribution (Reverchon, 2002, Reverchon and Della Porta, 2003a, 2003b, 2003c; Reverchon et al., 2004). Indeed, an increase of solute concentration enhances viscosity and surface tension of the liquid solution, resulting in the formation of larger primary droplets and also influences the formation and the dimensions of the secondary droplets. Therefore, systematic experiments were performed operating at different CH concentrations (C) in 1% acid acetic aqueous solution to explore the influence of this process parameter on the materials object of this work.

CH concentrations between 1 and 15 mg/mL have been explored. Performing experiment at larger C, a marked increase of the boiling temperature of the solution has been observed: condensation phenomena have occurred in the precipitator. Therefore, higher precipitation temperatures have been needed as C was increased from 5 to 15 mg/mL, as summarized in Table VI.1.

c

b a

Chapter VI

62

Figure VI.11: SEM images of chitosan microparticles precipitated by SAA from 1% acetic acid aqueous solution at concentrations and precipitation temperatures of 1 mg/mL and 95°C (a), 5 mg/mL and 95°C (b) and 10 mg/mL and 106°C (c), respectively.

The powder recovered in these experiments has always been formed by well-defined non-coalescing microparticles. Examples of the particles collected at different C are shown in SEM images reported in Figure VI.11. These images have been produced at the same enlargement (10K); therefore, a qualitative evaluation of the increase in particle size with solute concentration is possible.

Quantitative measurement of PSDs has been performed by SEM image analysis and by laser scattering analysis (LS). PSDs in terms of number of particles and particle volume have been calculated using image analysis method and from LS data. PSDs obtained by image analysis have shown that SAA precipitated CH powders are formed by particles ranging between 0.1 and 11 µm, with 90% of the total volume of the powder occupied by particles smaller than 4.5 µm when precipitated at C=1 mg/mL, smaller than 6.6 µm at C= 5 mg/mL and smaller than 8.5 µm at C= 10 mg/mL, respectively. However, SEM image technique does not allow one to verify if particles are really well separated; therefore, LS analysis has been

a b

c

RESULTS PART I: SAA CARRIERS PROCESSING

63

performed. PSDs obtained by LS are reported in Figure VI.12 and 13 in terms of number of particles and particles volume, respectively, and they substantially confirm the results of image analysis. These figures show the effect of solute concentration on particles size: particle size increases and distribution broadens when more concentrated solutions are processed. When the particle number distributions are considered (Figure VI.12), C-effect is visible on particles smaller than 0.3 µm; whereas, fairly good overlap of distributions occurs for larger particle: i.e., when less concentrated solutions are processed the number of small particles increases, whilst few large particles are produced.

0.1 1 10

20

40

60

80

100 C = 1 mg/mL C = 5 mg/mL C= 10 mg/mL

Particle size, µm

Par

ticle

s, %

Figure VI.12: PSDs in terms of number of particles percentages of micronized chitosan at C = 1 mg/mL and Tp = 95°C (□), C = 5 mg/mL and Tp = 95°C (●) and C = 10 mg/mL and Tp = 106°C (∆).

In the case of volumetric curves (Figure VI.13), even a very limited number of large particles has a relevant impact on the overall distribution, because the particle volume increases as the cube of its diameter. For this reason, C-effect is more evident in volumetric distributions obtained by LS. 90% of the volume is occupied by particles with diameters smaller than 4.1 µm operating at C = 1 mg/mL and by particles smaller than 8.1 µm at C = 10 mg/mL, in fair good agreement with data obtained by image analysis and confirming that no agglomeration of CH microparticles is present.

Chapter VI

64

0.1 1 100

20

40

60

80

100 C = 1 mg/mL C = 5 mg/mL C= 10 mg/mL

Vol

ume,

%

Particle size, µm

Figure VI.13: PSDs in terms of particle volume percentages of micronized chitosan at C = 1 mg/mL and Tp = 95°C (□), C = 5 mg/mL and Tp = 95°C (●) and C = 10 mg/mL and Tp = 106°C (∆).

In conclusion, the concentration of the solution is confirmed to be a parameter influencing CH particle size and distribution, although its effect is less marked than in previous SAA works (Reverchon, 2002, Reverchon and Della Porta, 2003a, 2003b, 2003c; Reverchon et al., 2004). This is probably due to the relatively narrow range of concentrations examined; i.e., processing more concentrated solutions a more marked effect would be visible. However, as previously explained, an increase in concentration over 15 mg/mL causes a strong increase in viscosity and boiling temperature, that interfere with process efficiency, due to difficulties in solution pumping and to the necessity to use higher temperatures in the precipitator. Therefore, a concentration of 5 mg/mL will be selected for further experimentations on CH.

VI.2.1.4 Solid state characterization

Dissolution and drying procedures can influence CH solid state. In particular, a decrease in crystallinity can occur depending on the drying process; e.g., freeze-drying produces less crystalline CH than oven drying (Yui et al., 1994; Salmon and Hudson, 1995). Therefore, X-ray and DSC analysis were performed on untreated and SAA-processed CH to evaluate the effect of SAA process on the solid state of this polymer.

Diffraction patterns of untreated and SAA-processed CH precipitated at temperatures ranging between 103 and 134 °C are reported in Figure VI.14. X-ray analysis reveals that untreated CH has a semi-crystalline structure.

RESULTS PART I: SAA CARRIERS PROCESSING

65

After SAA processing, CH powders show a lower crystallinity and, particularly, the higher is the precipitation temperature, the larger is the amorphous content in CH microparticles.

10 20 30 40 50 60 70

CH processed at 103°C

CH processed at 105°C

CH processed at 134°C

Untreated CH

Inte

nsity

2θ

Figure VI.14: X-Ray diffraction patterns of untreated and SAA processed CH at different temperatures in the precipitator.

The solid state of SAA processed CH depends on the precipitation temperature. A possible explanation can take into account a physico-chemical modification of the polymer. Several authors observed that the type of crystalline structure, the degree of crystallinity and the average crystallite size depends on the deacetylation degree of the material: the higher the deacetylation of the sample, the lower its crystallinity and the smaller crystallites size (Urbanczyk and Lipp-Symonowicz, 1994; Jaworska et al. 2003; Zhang et al. 2005). Indeed, dissolving CH in acid aqueous solutions and vaporizing the solvent at high temperatures can cause a change in both the chemical structure and the supramolecular structure of the polymer; processes such as further deacetylation and cross linking of the polymer can take place in the thermal treatment of CH salts (Kittur et al. 2002; Harish Prashanth et al., 2002; Zotkin t al., 2004).

Chapter VI

66

Figure VI.15: DSC curves of untreated and SAA processed CH at different temperatures in the precipitator (Tp).

DSC thermograms on untreated and SAA precipitated CH at different temperatures are reported in Figure VI.15. The endothermic peak related to the evaporation of water is expected to reflect physical and molecular changes of the polymer. Polysaccharides usually have a strong affinity for water and in the solid state these macromolecules may have disordered structures, that can be easily hydrated. The hydration properties of polysaccharides depend on their primary and supramolecular structures (Kacurakova et al., 1998; Phillips et al., 1996). As can be observed in Figure VI.15, the first thermal event registered in all the samples is a wide endothermic peak centred between 90 and 105 °C, due to water evaporation. Figure VI.15 shows some differences between SAA precipitated samples and the untreated material in this peak area and position, indicating that these samples differ in water holding capacity and in the strength of water-polymer interaction. Micronized CH powders show larger peak areas and lower peak temperatures with respect to the raw material, i.e. higher water content and weaker water-polymer interactions, respectively. Moreover, the powder obtained at 135°C shows a lower endothermic peak temperature and a larger area with respect to the powders precipitated at 103 and 105 °C, thus pointing out that SAA precipitation temperature affects hydration properties of CH microparticles. These results agree with the X-ray analysis, since the less ordered structure due to chemical modification may contribute significantly to the increase in the content of sorbed water. The bound water content of the samples can also depend on the structure of the polysaccharide chain; e.g. by the number of amine and carboxymethyl groups that are hydrophilic centres.

50 100 150 200 250 300

Tp = 103°C

Tp = 105°C

Tp = 135°C

Untreated CS

W g

-1

Temperature, °C

Untreated CH

RESULTS PART I: SAA CARRIERS PROCESSING

67

The second thermal event is an exothermic peak occurring at temperatures higher than 250°C: SAA precipitated powders show the exothermic peak shifted to lower temperatures with respect to the untreated material. Also this behaviour can be attributed to a decrease in thermal stability as a consequence of the decreased acetyl content and degree of polymerization.

In conclusion, SAA treated CH microparticles show a crystalline degree depending on the temperature set in the precipitator. This process parameter can be easily modulated and allows the production of CH microparticles characterized by different degrees of crystallinity, according to the selected application. This aspect can be particularly relevant, for example, in modifying drug release mechanisms in CH/drugs formulations.

VI.2.1.5 Conclusions on Chitosan

SAA process is able to produce CH well-defined spherical microparticles with narrow PSDs. Particle size tailoring is possible by modulation of some process parameters, such as the solute concentration. The precipitation temperature influences the crystallinity of the precipitated microparticles.

SAA micronized CH microparticles are expected to be promising vehicles for drug delivery. Moreover, the positive results obtained by SAA micronization of pure chitosan introduce the possibility of direct CH-drug composite microparticles production using the SAA technique.

Chapter VI

68

VI.3 CYCLODEXTRIN MICROPARTICLES PRODUCTION

Cyclodextrins (CDs) are biocompatible compounds widely employed for a variety of purposes, because of their multi-functional characteristics and bioadaptability. They are used in many industrial products and analytical methods: in food and flavors, cosmetics, bioconversion, packing, textiles, separation processes, environment protection, fermentation and catalysis (Szejtli, 1998; Hedges, 1998; Szente and Szejtli, 2004 ). Indeed, as a result of their molecular structure and shape, they possess the ability to act as molecular containers entrapping guest molecules in their internal cavity (as described in Chapter II.2).

Cyclodextrins have also been exploited in the pharmaceutical field for a variety of purposes: e.g., drug formulations, control of polymorphic transition and crystallization of drugs by amorphous CDs, uses in peptide and protein formulations, application in gene therapy. They can be used in drug formulations either for complexation or as auxiliary additives such as carriers, diluents, solubilizers and tablet excipients. Inclusion complex formation usually results in advantageous modification of the physico-chemical properties of the complexed drug: improvement of physical and chemical stability, enhancement of the bioavailability of poorly soluble drugs, lowering the incidence of side effects and promoting a faster therapeutic action (Uekama et al., 1998; Fernandes and Veiga, 2002 ).

The production of CD microparticles is a relevant aspect for many applications, especially for pharmaceutical ones. For example, the formulation of drug/CD micronized complexes could improve current administration routes and exploit novel delivery systems. In particular, CD particle size reduction may enhance the interaction between drug and CD molecules, thus accelerating the complex formation even in physical mixtures. CD microparticles can be used as excipients in inhalation powders to increase drug stability, dissolution rate and bioavailability or to decrease local irritation of an inhaled drug (Kinnarinen et al., 2003).

Conventional methods to obtain CD microparticles are grinding, spray-drying, freeze-drying and co-evaporation. However, they do not assure an efficient control of the particle size, can cause thermal or chemical degradation and low reproducibility among different batches has been reported.

SCFs have been applied to CD micronization and CD/drug complexes formation to overcome the drawbacks of conventional techniques. For example, impregnation of CDs microparticles using supercritical CO2 (SC-CO2), or SC-CO2 modified with organic solvents, has been proposed to produce naproxen/β-CD complexes (Junco et al., 2002a, 2002b ). Ibuprofen/β-CD microcomposite particles have been obtained using the Rapid Expansion of Supercritical Solutions (RESS) as an impregnation

RESULTS PART I: SAA CARRIERS PROCESSING

69

technique (Cristini et al., 2003). Supercritical Antisolvent precipitation (SAS)-based processes have been investigated to precipitate complexes of celecoxib (Fabing et al. 2002) and naproxen (Foster et al., 2004) with hydroxypropyl- and methyl-β-CDs. However, the impregnation-based techniques showed a low efficiency and, although CDs are soluble in water, this solvent is difficult to be employed in SAS-based processes.

The potential of SAA in the production of cyclodextrin microparticles has been investigated to overcome the limitations of the previously described techniques. In particular, SAA processability of α-cyclodextrin (α-CD) and hydroxypropyl-β-cyclodextrin (HP-β-CD), two of the most employed CDs, has been evaluated using water as solvent. The effect of the solute concentration in the liquid solution on CDs particles morphology, size and distribution has been studied to obtain particle size tailoring. Solid state characterization of untreated and SAA processed cyclodextrins has been performed by X-ray and calorimetric analyses.

VI.3.1 Results And Discussion

Water has been chosen as solvent for CDs; the use of water is particularly advantageous when pharmaceutical compounds are processed, because it gives no problems of toxicity (unlike the organic solvents) and, thus, there is no problem of solvent residues in the precipitate.

Untreated α-CD and HP-β-CD consist of irregularly shaped particles ranging between 20 and 500 µm. α-CD and HP-β-CD show a solubility in water (at 25°C) of 145 mg/mL and larger than 2000 mg/mL, respectively.

VI.3.1.1 Selection of the saturator operating parameters

As previously discussed, the solubilization of SC-CO2 in the liquid solution inside the saturator is one of the key steps controlling the efficiency of the SAA process and the selection of adequate gas and liquid flow rates is also relevant for the achievement of long residence times in the saturator (from 5 to 6 min) to assure gas saturation in the liquid solution. Moreover, the mass flow ratio between CO2 and liquid solution influences the equilibrium in the saturator of the ternary system CO2/solvent/solute. Data reported in the literature give information about the binary system CO2/water (Takenouchi and Kennedy, 1964); however, no data is available for the ternary systems CO2/water/α-CD and CO2/water/HP-β-CD. Therefore, in the selection of the mass feed ratio, we relied on our previous experience in this process. Thus, a mass feed ratio (R) of 1.8 between CO2 and the liquid solution has been used in all the experiments performed in this work, since it has been showed to be the most appropriate for SAA of aqueous solutions (Reverchon and Della Porta, 2003a, 2003b, 2003c; Reverchon et al., 2003a).

Chapter VI

70

According to these considerations, some preliminary tests have been performed setting the saturator conditions in the pressure range from 75 to 150 bar and in the temperature range between 70 and 90 °C. The best results in terms of stability of the process and of CDs precipitated particles have been observed operating at 85 bar and 85°C. Therefore, these pressure and temperature conditions have been used in all the subsequent SAA experiments on CDs.

VI.3.1.2 Effect of temperature in the precipitator

The effect of temperature on particle morphology has been observed performing experiments on α-CD at temperatures ranging between 100 and 186°C, at a α-CD concentration in the liquid solution of 50 mg/mL. SEM images of SAA processed α-CD precipitated at 108°C, at 118°C and at 186°C are proposed in Figure VI.16.

Figure VI.16: SEM images of α-CD microparticles precipitated by SAA at temperatures of 108 °C (a), 118 °C (b) and 186 °C (c) in the precipitator (Csol =50 mg/mL).

In the experiments performed at 108°C, or lower temperatures, coalescing particles have been recovered in the precipitator due to a low

c

b a

RESULTS PART I: SAA CARRIERS PROCESSING

71

efficiency of solvent evaporation (Figure VI.16a); whereas, in all the experiments performed at temperatures higher than 118°C well-defined spherical microparticles have been produced. No substantial differences in particle morphology and size distribution have been observed when the temperature was further increased, as can be observed in Figure VI.16b and 16c.

Experiments performed on HP-β-CD at precipitation temperatures ranging between 118 and 160°C confirmed that well-defined spherical microparticles can be obtained also for this CD. Therefore, 118°C has been set as the operating temperature in the precipitator in the following experiments on α-CD and HP-β-CD.

VI.3.1.3 Effect of solute concentration

As previously discussed, an increase of solute concentration may enhance viscosity and surface tension of the liquid solution, resulting in the formation of larger primary droplets and also influencing the formation and the dimensions of the secondary droplets. Therefore, systematic experiments have been performed operating at different α-CD and HP-β-CD concentrations (C) in water to explore the influence of this process parameter on the materials object of this work.

α-CD concentrations between 20 and 130 mg/mL have been explored. The powder recovered in these experiments was formed by well-defined non-coalescing microparticles. Examples of the particles collected at different C are shown in SEM images reported in Figure VI.17. These images have been produced at the same enlargement (20K); therefore, a qualitative evaluation of the increase of particle size with solute concentration is possible.

Chapter VI

72

Figure VI.17: SEM images of α-CD microparticles precipitated by SAA at solute concentrations of 20 mg/mL (a), 50 mg/mL (b) and 130 mg/mL (c) in the liquid solution (T=118°C in the precipitator).

Quantitative measurement of PSDs has been performed by laser scattering analysis and by image analysis on the powder produced at different concentrations. PSDs in terms of particle volume and number of particles have been calculated from LS data and using SEM image analysis method. The distributions obtained using the two methods were compared as, for example, in Figure VI.18, where histograms in terms of number of α-CD microparticles precipitated at 20 mg/mL are reported.

c

b a

RESULTS PART I: SAA CARRIERS PROCESSING

73

0 1 2 3

20

40

60

80

Particle size, µm

Part

icle

s, %

SEM

LS

Figure VI.18: Histograms in terms of number of particles percentages of micronized α-CD at C = 20mg/mL. Comparison between PSDs obtained using LS (■) and SEM (□) analysis. Expanded X-axis below 3 µm.

Both techniques show that more than 90% of particles have diameters lower than 2 µm; but, SEM images analysis overestimates small submicrometric particles and underestimates particles larger than 3 µm with respect to LS. Indeed, LS indicates that particles with diameters ranging between 0.2 and 0.7 µm are the major component of the distribution and also shows a long tail formed by very small percentages of particles up to 20 µm. The long tail of LS distribution can be observed in Figure VI.19 as well, where cumulative distributions based on the number and on the volume of α-CD microparticles precipitated at 20 mg/mL are reported (in Figure VI.18 X-axis was expanded below 3 µm to allow a comparison of the performance of the two analytical techniques).

Volumetric cumulative distributions obtained using the two methods clearly show the influence of a small number of large particles on the overall distribution with respect to the number of particles based distributions.

Chapter VI

74

20

40

60

80

100

0 3 6 9 12 15 18 21

Pa

rtic

les,

%

SEM

LS

SEM

LS

0 3 6 9 12 15 18 210

20

40

60

80

100

Vol

ume

, %

Particle size, µm

Figure VI.19: PSDs in terms of number of particles and volume percentages of micronized α-CD at C = 20mg/mL. Comparison between PSDs obtained using LS (—) and SEM (– –) analysis.

It is possible to give an explanation of these discrepancies. If the number of particles based distributions are considered, 95% of the particles produced has a diameter smaller than 0.9 µm in the LS curve and smaller than 1.45 µm in SEM analysis. Moreover, 99.8% and 99.9% of the particles have diameters lower than 5 µm for LS and SEM analysis, respectively. Therefore, only the remaining about 0.2% of particles contributes to the formation of the long tail shown by volumetric LS curve. Indeed, in the case of volumetric PSDs a very limited number of large particles has a relevant impact on the overall distribution. For example, particles with diameters differing of 10 units have volumes that differ 1000 times. These particles, in a number of particles distribution give the same contribution; but, in a volumetric distribution, the larger one gives a contribution that is one thousand times larger.

RESULTS PART I: SAA CARRIERS PROCESSING

75

To verify if these larger particles are really present in the powders or if they are artifacts formed by coalescence, a further accurate morphological analysis has been performed by SEM. A landscape image of micronized α-CD is reported in Figure VI.20. Qualitative evaluation of this image shows α-CD particles with a very uniform morphology, a fairly narrow distribution and a maximum diameter of 5-6 µm, confirming the previous analysis. However, scanning several samples by SEM it was possible to observe a very limited number of larger particles, confirming the long tail in the number of particles based distributions and the consequent enlargement of the volumetric distributions obtained using LS. These few large particles are probably produced at non-steady state conditions during start up and the end of the process.

Figure VI.20: SEM image at magnification of 3K reporting a landscape of α-CD microparticles precipitated at C=130 mg/mL.

Number of particles and volume based distribution data obtained by LS and SEM analysis on α-CD micronized at different concentrations are reported in Table VI.2 and Table VI.3. PSDs obtained at 50 and 130 mg/mL confirm the previous discussion based on distributions at 20 mg/mL. According to LS, 99.7% and 95.1% of particles have diameters lower than 5µm at 50 and 130 mg/mL, respectively. Both LS and SEM analysis at different C show particle size increase and distribution broadening when more concentrated solutions are processed.

Chapter VI

76

Table VI.2: Number of particles PSDs data on of micronized α-CD at different concentration values. Comparison between LS and SEM analysis.

Number of Particles Distribution Percentages [µµµµm]

50% 90%

Particles with Diameter

< 5µµµµm

αααα-CD

C values

[mg/mL] LS SEM LS SEM LS SEM

20 0.35 0.26 0.5 0.98 99.8% 99.9%

50 0.41 0.37 0.48 1.54 99.7% 99.7%

130 1.92 0.62 3.99 2.72 95.1% 95.7%

Table VI.3: Volumetric PSDs data of micronized α-CD at different concentration values. Comparison between LS and SEM analysis.

Volumetric Distribution Percentages [µµµµm]

50% 90% αααα-CD

C values [mg/mL]

LS SEM LS SEM

20 5.4 3.7 13.1 6.8

50 6.9 5.0 17.6 8.9

130 7.0 8.7 18.3 12.7

Volumetric distribution percentages obtained at C = 20, 50 and 130

mg/mL are also reported in Table VI.3. 90% of the total volume of α-CD powder is due to particles with diameters up to 13.1 µm at C=20 mg/mL and up to 18.3 µm at C=130 mg/mL, respectively. This data have also been reported in terms of cumulative distributions in Figure VI.21.

RESULTS PART I: SAA CARRIERS PROCESSING

77

0 10 20 30 400

20

40

60

80

100

C=130 mg/mL

Vol

ume

, %

Particle size, µm

C=20 mg/mL

Figure VI.21: PSDs in terms of particle volume percentages of micronized α-CD at different C values, obtained by LS.

The effect of solute concentration has been investigated also in the case of HP-β-CD micronization. Concentration values ranging between 20 and 100 mg/mL have been explored and in all the experiments well-defined spherical microparticles have been precipitated.

In Figure VI.22, SEM images of particles produced at different C are reported. Also in this case, images have been produced at the same magnification (10K), allowing the qualitative evaluation of the increase in particle size with solute concentration. Data from LS and SEM analysis used to produce number of particles and volume based distributions at C = 20, 50 and 100 mg/mL are reported in Table VI.4 and Table VI.5, respectively. A comparison between LS volumetric curves at 20 and 100 mg/mL is also shown in Figure VI.23: 90% of the total volume occupied by micronized HP-β-CD is related to particles with diameters up to 8.45 µm at C=20 mg/mL and up to 12.04 µm at C=100 mg/mL, respectively. But, 96.6% and 96.3% of particles have diameters lower than 5 µm at 20 and 100 mg/mL, respectively (Table VI.4).

Chapter VI

78

Figure VI.22: SEM images of HP-β-CD microparticles precipitated by SAA at solute concentrations of 20 mg/mL (a), 50 mg/mL (b) and 100 mg/mL (c) in the liquid solution (T=118°C in the precipitator).

Table VI.4: Number of particles PSDs data on of micronized HP-β-CD at different concentration values. Comparison between LS and SEM analysis.

Number of Particles Distribution Percentages [µµµµm]

50% 90%

Particles with Diameter

< 5µµµµm

HP-ββββ-CD

C values

[mg/mL] LS SEM LS SEM LS SEM

20 0.45 0.32 0.71 1.18 99.6% 99.7%

50 0.76 0.64 0.93 1.96 98.5% 98.8%

100 1.48 1.22 2.84 3.29 96.3% 97.1%

c

b a

RESULTS PART I: SAA CARRIERS PROCESSING

79

Table VI.5: Volumetric PSDs data of micronized HP-β-CD at different concentration values. Comparison between LS and SEM analysis.

Volumetric Distribution Percentages [µµµµm]

50% 90% HP-ββββ-CD

C values [mg/mL]

LS SEM LS SEM

20 3.8 4.0 8.4 6.3

50 4.8 4.1 9.1 6.7

100 6.1 5.5 12.0 8.7

0 5 10 15 20 250

20

40

60

80

100

C=100 mg/mL

Vo

lum

e, %

Particle size, µm

C=20 mg/mL

Figure VI.23: PSDs in terms of particle volume percentages of micronized HP-β-CD at different C values, obtained by LS.

In conclusion, the comparison between LS and SEM analysis data can compensate the weaknesses of the two techniques. SEM analysis can underestimate larger particles: when a limited number of large micrometric particles is contained in the overall sample, they could not be all captured in SEM image analysis. LS analyses a large number of particles; but, tends to underestimate the smaller particles.

Chapter VI

80

VI.3.1.4 Solid state characterization

X-ray and DSC analyses have been performed on untreated and SAA-processed α-CD and HP-β-CD to evaluate the effect of SAA process on the solid state of these compounds.

Diffraction patterns of untreated and SAA-processed α-CD and HP-β-CD are reported in Figure VI.24a. X-ray analysis reveals that SAA micronized α-CD and HP-β-CD are amorphous; whereas, the raw material is crystalline in the case of α-CD and amorphous in the case of HP-β-CD. This result is confirmed by DSC thermograms (Figure VI.24b). Only untreated α-CD shows definite fusion peaks that are characteristic of crystalline structures. SAA processed α-CD, untreated and SAA micronized HP-β-CD show an endothermic peak in the range of 90-160°C that can be attributed to a dehydration process.

10 20 30 40 50 60

SAA-processed α-CD

SAA-processed HP-β-CD

Untreated α-CD

Untreated HP-β-CD

2θ

50 100 150 200 250 300

Untreated

HP-β-CD

Untreated αCD

SAA processed

HP-β-CD

W g

-1

Temperature, °C

SAA processed αCD

Figure VI.24: X-Ray diffraction patterns (a) and DSC thermograms (b) of untreated and SAA processed α-CD and HP-β-CD.

a

b

(a)

(b)

RESULTS PART I: SAA CARRIERS PROCESSING

81

The amorphous solid state of CD microparticles can be a relevant aspect for their pharmaceutical applications. Indeed, crystal modifications significantly affect various pharmaceutical properties such as solubility, dissolution rate, stability and bioavailability of drugs. As a consequence, the rational control of crystal growth, habit and polymorphic transition, using pharmaceutical additives, such as CDs, becomes an attractive and interesting area of drug research and development. Many reports have shown that crystalline drugs can be converted to an amorphous form by complexation with amorphous CDs (Hirayama et al., 1997; Kimura et al., 1999). For example, amorphous CD such as HP-β-CD is useful for inhibition of polymorphic transition and crystallization rates of poorly water-soluble drugs during storage, which can consequently maintain the higher dissolution characteristics and oral bioavailability of the drugs (Hirayama et al., 2001). VI.3.1.5 Conclusions on Cyclodextrins

SAA can be used for cyclodextrins micronization using water as solvent. It has been successfully tested on α-CD and HP-β-CD and well-defined spherical and amorphous microparticles have been produced. The effect of the solute concentration on particle size and distribution has been evaluated; all the micronized powders were formed by at least 95% of particles with diameters ranging between 0.1 and 5 µm. This size range is particularly relevant for pharmaceutical applications.

In conclusion, cyclodextrin microparticles can be easily produced by SAA with morphological, granulometric and solid-state characteristics useful for pharmaceutical applications. Moreover, new prospects for the direct production of micronized drug/CD complexes using SAA technique are opened by this experimentation.

Chapter VI

82

Chapter VII RESULTS PART II:

COPRECIPITATION Once selected the optimal process conditions for the successful

micronization of the carriers, some coprecipitation tests with model drugs have been performed. In particular, among synthetic polymers we have selected PMMA, because of its high Tg, and among natural carriers we have used chitosan.

Coprecipitation tests on PMMA have been performed using as model drug Medroxyprogesterone acetate (MEPA); chitosan has been coprecipitated with Ampicillin trihydrate.

VII.1 PMMA-MEPA

VII.1.1 Medroxyprogesterone acetate (MEPA)

MEPA is a synthetic hormone obtained from progesterone. It is largely used in gynaecological therapies and as contraceptive. For this purpose, formulations allowing a sustained release of this compound are very interesting, since a reduced number of administrations and a control of drug blood level fluctuations would be possible.

Previous SAA experimentations (Williams, 2002) were performed on pure MEPA using methanol and acetone as process fluids. Successful micronization of this compound was obtained and the best R-value for the experiments in acetone was found to be R=1.

Chapter VIII

84

Figure VII.1: SEM image of MEPA microparticles produced by SAA (adapted from: Williams, 2002)

The morphology of micronized MEPA produced at R=1 and Tpr=60°C is reported in the SEM image proposed in Figure VII.1. At these process conditions, micrometric and well-defined particles of MEPA were obtained.

VII.1.2 PMMA and MEPA coprecipitation

The individual experimentations performed on PMMA and MEPA using acetone allowed the selection of SAA process conditions at which the polymer and the drug can be successfully micronized. Particularly, a flow mass ratio of 1 and a precipitation temperature of about 60°C resulted the most suitable conditions to perform a coprecipitation. Therefore, coprecipitation has been carried out using acetone as solvent at the following conditions:

• PMMA:MEPA ratio of 4:1;

• R=1;

• Pmix=76 bar;

• Tmix=80 °C;

• Tpr=60 °C.

RESULTS PART II: COPRECIPITATION

85

Figure VII.2: SEM image of PMMA and MEPA coprecipitates

The morphology of the powder collected in the precipitator is reported in Figure VII.2. Well-defined and spherical microparticles have been produced. Moreover, during the experiments, no pressure fluctuations have been observed and no material precipitation in the saturator has occurred; it means that no antisolvent effect has occurred.

SEM images of pure PMMA, pure MEPA and coprecipitates obtained at the same process conditions have been studied using the image analysis software to measure PS and PSDs, that have been compared in Figures VIII.3 and VIII.4, reporting the PDS in terms of number of particles and volume percentages, respectively.

0 1 2 3 4 50

10

20

30

40

MEPA

PMMA+MEPA

PMMA

Par

ticle

s, %

Particle size, µm

Figure VII.3: PSD of PMMA, MEPA and PMMA/MEPA SAA coprecipitated particles. Calculations in terms of number of particles percentages.

Chapter VIII

86

An increase in particle size for coprecipitates is evidenced with respect to the single compounds. The mode moves from 0.5 µm for PMMA, to 1.2 µm for MEPA and to 1.5 µm for the coprecipitated powder. An enlargement of the size distribution is also obtained when compared to those of pure compounds.

0 1 2 3 4 50

20

40

60

80

100

PMMA+MEPA

MEPA

PMMA

Vol

ume,

%

Particle size, µm

Figure VII.4: PSD of PMMA, MEPA and PMMA/MEPA coprecipitated particles. Calculations in terms of particle volume percentages.

The diagram in Figure VII.4 shows the same information of Figure VII.3, but in terms of volume percentage: 95% of PMMA, of MEPA and of coprecipitated powder is formed by particles sizing less than 0.94 µm, 2.75 µm and 3.96 µm, respectively. These distributions, however, enhance the differences in PS observed in Figure VII.3.

The results obtained and the comparison between the morphologies of the pure compounds and the coprecipitated powder allow the formulation of some preliminary considerations about the quaternary system. Although, PMMA and MEPA are contemporary solubilized in the liquid solution, no CO2 antisolvent effect has occurred, proved by no material precipitation and no pressure fluctuations in the saturator. But, the physical characteristics of the liquid solution PMMA/MEPA/acetone are different from those of PMMA/acetone and MEPA/acetone ones, resulting in increased viscosity of the solution and, therefore, increased particles size and broadened PSD.

Both these results indicate that interactions occur between PMMA, MEPA, acetone and CO2 into the saturator, resulting in a complex quaternary system with characteristics significantly different from the ternary PMMA/acetone/CO2 and MEPA/acetone/CO2 systems, but the operating conditions used are still adequate to successful coprecipitation. To

RESULTS PART II: COPRECIPITATION

87

confirm these results, a solid state analysis has been carried out on coprecipitated PMMA/MEPA microparticles.

VII.1.2.1 Solid state characterization

DSC analysis give information about the crystalline structure of materials. In the case of coprecipitates, a different solid state is expected for coprecipitates with respect to physical mixture of the drug and the polymer. Therefore, DSC has been performed on MEPA as raw material, PMMA microparticles produced by SAA and coprecipitated PMMA/MEPA microparticles. A physical mixture in the same ratio of coprecipitates (polymer:drug = 4:1) of PMMA microparticles and MEPA has also been analysed as control.

DSC thermograms are reported in Figure VII.5. PMMA has a completely amorphous structure (trace d), whereas, MEPA shows the characteristic endothermic peak of crystal melting at 210°C (trace a). The physical mixture formed by polymeric microparticles and the drug, shows that each of the two components does not change its thermal behaviour with respect to the raw materials. Indeed, the endothermic peak of MEPA is still present at 210°C; it is only smaller due to the smaller amount of the drug with respect to the polymer (trace b).

30 60 90 120 150 180 210 240

bcd

a

[Wg^

-1]

Temperature [°C]

Figure VII.5: DSC thermograms of: a) raw MEPA; b) physical mixture of PMMA microparticles/MEPA (ratio 4/1); c) SAA coprecipitated PMMA/MEPA microparticles; d) PMMA microparticles.

Chapter VIII

88

A deep change in drug thermal behaviour is shown by SAA coprecipitates with respect to the pure compounds and to the physical mixture: MEPA endothermic peak is less pronounced and shifted at 198°C. Lowering of the melting temperature and enthalpy of MEPA suggest an interaction due to the physical mixture of the compounds (Dubernet, 1995; Singhal et al., 1999; Wang et al. 2004; El-Gibaly, 2002; Ceschel et al., 2003). This behavior can be explained if SAA coprecipitated particles are formed by a PMMA matrix in which MEPA molecules are entrapped, thus obstructing the drug molecules organization into a crystalline structure.

VII.1.2.2 Release rate studies

The previous results support the hypothesis of a solid solution formed by the drug and the polymer during SAA precipitation. However, we still do not know how close is the interaction between these two components are and if this formulation really allows a slow MEPA release. Therefore, drug release rate tests have been performed to verify the efficiency of the SC-CO2 encapsulation process.

The analysis has been performed on tablets using UV-vis spectrometer, according to the procedure described in Chapter V.5, and as the dissolution medium a physiological saline solution (pH 7.2).

All tablets analyzed have been prepared by directly compressing of 200 mg of the powder, at 150 bar for 10 min. Tablets have been placed into 1000 mL of physiological solution, kept at 37°C and stirred at 200 rpm. At predetermined time intervals, the concentration of drug was assayed using UV spectrophotometer.

Drug release rates from tablets formed by raw MEPA and SAA coprecipitated PMMA/MEPA microparticles have been compared in Figure VII.6. It clearly shows a slower release from SAA coprecipitate with respect to untreated MEPA, that completely dissolves in about 1h. A MEPA prolonged release up to nearly 4 days is exhibited by SAA coprecipitated powder, that, furthermore, exhibits no burst effect: i.e., no initial fast release of the drug is observed for PMMA/MEPA coprecipitates. This characteristic suggests that no drug is concentrated on the particles surface.

These results agree with the previous considerations derived by morphological and solid state analysis, confirming that SAA produced microparticles are formed by a uniform dispersion of MEPA in PMMA matrix.

RESULTS PART II: COPRECIPITATION

89

0 2 4 6 8 10 120

20

40

60

80

100

ME

PA

Rel

ease

d, %

Time, h

0 12 24 36 48 60 72 84 960

20

40

60

80

100

ME

PA

Rel

ease

d, %

Time, h

Figure VII.6: Drug release rate: (∆) raw MEPA; (ο) SAA coprecipitated PMMA/MEPA microparticles. Drug release rates from coprecipitates are reported up to 12 h (left) and to 96 h (right).

VII.1.2.3 Conclusions on PMMA/MEPA coprecipitation

The tests of coprecipitation performed using as model carrier PMMA and as model drug MEPA have given satisfying results: spherical and well defined microparticles with a uniform morphology have been produced. DSC and drug release rate analyses have confirmed that MEPA is entrapped in a PMMA matrix, forming a solid solution of the two components that allows a prolonged release of the drug. Using a PMMA:MEPA ratio of 4:1, a coprecipitated formulation able to release the drug up to 4 days has been obtained. The possibility of modulating the polymer/drug ratio can be taken into account in further experimentation to modify drug release, according to the pharmaceutical final purpose. For example, if the coprecipitate will be administered orally, a MEPA release time of 24 h is needed, since then the tablet is expelled by the human body; if instead a subcutaneous implant of PMMA/MEPA is required, prolonged times of the order of months are desired.

Chapter VIII

90

VII.2 CHITOSAN-AMPICILLIN

Ampicillin trihydrate is a broad spectrum antibiotic, commonly used for systemic therapy as well as locally for gastric or intestinal infections. It is acid resistant; therefore, it can be administered orally, but it has a short biological half-life of 0.75–1.5 h. In order to make the application of ampicillin more effective, research has been directed to design formulations for its sustained and controlled release. It is possible to design particulate drug carrier systems with mucoadhesive properties that may protect the drug from degradation during the passage through the gastrointestinal tract, enhance the uptake by the epithelium and act as a controlled release system resulting in prolonged blood concentrations.

The mucoadhesive properties of Chitosan suggest the employment of this polymer for the production of CH/Ampicillin microparticles. Indeed, CH is characterized by molecular attractive forces due to electrostatic interactions between positively charged CH and negatively charged mucosal surfaces and, thus, promotes drugs transmucosal absorption.

For these reasons, the possibility of producing CH/Ampicillin microparticles for drug controlled release using SAA has been investigated.

VII.2.1 Ampicillin

A previous SAA work on ampicillin trihydrate, using water as solvent, has showed that well defined spherical microparticles can be produced, with uniform morphology and sizes ranging between 1 and 5 µm (Reverchon et al., 2003a). However, no previous experimentation was conducted on this drug using acid water as solvent, that is the only solvent that can be used to solubilize both drug and chitosan. Therefore, before coprecipitation experiments, some micronization tests on pure ampicillin trihydrate have been performed at the optimum precipitation conditions exploited for CH (Chapter VII.2). In particular, the experiments on ampicillin have been conducted using:

• 1% acetic acid aqueous solution as solvent; • mass flow ratio between CO2 and solution R = 1.8; • temperature and pressure in the saturator set at 85°C and 105 bar,

respectively; • temperature in the precipitator Tpr=95°C; • solute concentration C=5 mg/mL.

Successful micronization has been obtained: well-defined microparticles with uniform morphology have been produced, as shown in Figure VII.7.

RESULTS PART II: COPRECIPITATION

91

Figure VII.7: SEM image of ampicillin microparticles precipitated at Tpr=95°C from 1% acid acetic aqueous solution.

VII.2.2 Chitosan-Ampicillin coprecipitation

Verified the SAA processability of the two pure compounds, coprecipitation tests have been performed to prepare Ampicillin (Amp) loaded microparticles from the ternary system formed by CH and Amp dissolved in 1% acid acetic aqueous solution. The same process conditions described in the previous paragraph have been used.

Different amounts of drug have been solubilized into the polymeric solution, to investigate the effect of the polymer/drug ratio on the release rate. In particular, CH and ampicillin have been coprecipitated in the ratios:

• CH : Amp = 1:1 • CH : Amp = 2:1 • CH : Amp = 5:1 • CH : Amp = 8:1

Drug concentrations refer to anhydrous CH/ampicillin (w/w) ratios. Successful micronization has been obtained for theses systems, though a

complex quaternary system has to be considered during the solubilization of SC-CO2 in the saturator, as in the case of PMMA/MEPA. We have concluded that simultaneous presence of CH and Amp in the aqueous solution does not develop particular interactions leading to process instability and, therefore, the system behaves like the two ternary ones CH/acid water/CO2 and Amp/acid water/CO2. Well-defined spherical microparticles with very uniform morphology have been produced at all polymer/drug ratios tested. For example, SEM images of coprecipitated microparticles are reported in Figure VII.8.

Chapter VIII

92

Figure VII.8: SEM images of CH/ampicillin microparticles coprecipitated by SAA at Tpr=95°C from 1% acid acetic aqueous solution, using polymer/drug ratios of 1:1 (a), 2:1 (b), 5:1 (c) and 8:1 (d).

A qualitative evaluation performed on SEM images suggests that particles produced at different polymer/drug ratios range between about 0.2 and 3 µm, with a narrow size distribution; moreover, PSDs do not seem to depend on the polymer/drug ratio used in the liquid solutions.

Quantitative measurements of PSDs of the coprecipitated microparticles have been performed using laser scattering analysis (LS) confirming the qualitative considerations. PSDs in terms of number of particles and volume calculated from LS data are reported in Figure VII.9, where the curves obtained from all polymer/drug ratios are compared.

a b

c d

RESULTS PART II: COPRECIPITATION

93

2

4

6

8

10

12

0 1 2 3 4 5 6

Par

ticle

s, %

0 1 2 3 4 5 60

20

40

60

80

100

CH : Amp = 1:1 CH : Amp = 2:1 CH : Amp = 5:1 CH : Amp = 8:1

Vol

ume,

%

Particle size, µm

Figure VII.9: PSDs in terms of number of particles and volume percentages of CH/Amp microparticles coprecipitated by SAA at polymer/drug ratios of 1:1 (∆), 2:1 (ο), 5:1 (□) and 8:1 (�).

In conclusion, no systematic effect of CH/ampicillin ratio on the PSDs of coprecipitates can be noted, since a substantial overlap of distributions occurs. Very narrow PSDs are obtained: microparticles range between 0.1 and 6.3 µm, with the mode of all the number of particles based distributions centred at about 0.3 µm. Volumetric PSDs are less overlapping when compared to the particle of number based curves. These differences are due to some fluctuations of process conditions, that can cause the formation of few larger particles: a very limited number of large particles has a relevant impact on the overall volumetric distributions. This hypothesis is supported by volumetric PSDs data, reported in Table VII.1, in which D50 and D90, respectively, represent the maximum diameter exhibited by the 50% and the 90% of the total volume occupied by the powder: D50 values are nearly the

Chapter VIII

94

same for all distributions; whereas D90, is more influenced by the presence of few large particles and varies between 2.18 and 2.95 µm.

Table VII.1: Volumetric PSDs percentages of CH/Amp microparticles coprecipitated by SAA at polymer/drug ratios of 1:1, 2:1, 5:1 and 8:1.

CH : Amp D50, µµµµm D90, µµµµm 1:1 1.15 2.93 2:1 1.14 2.63 5:1 1.33 2.95 8:1 1.01 2.18

Morphological and particle size analyses have shown that the processing

of this multicomponent liquid solution by SAA is possible, with the formation of well-defined spherical microparticles with narrow PSDs. However, these analyses give no information about the distribution of the drug with respect to the polymer. It is still not known if the drug is entrapped inside the polymer matrix and how it is distributed. Therefore, coprecipitated microparticles have been characterized by X-ray, DSC, EDX and UV-vis analysis, to investigate their solid state, the spatial distribution of the compounds and drug release rates, respectively.

VII.2.2.1 Solid state characterization

X-ray analysis X-ray analysis give information about the solid state of materials. In the

case of coprecipitates an amorphous state is expected, in particular of the drug, since Amp molecules organization could be hindered by the polymer chains.

The diffraction pattern of coprecipitated microparticles in the ratios 1:1 is reported in Figure VII.10 (g). Moreover, diffractograms of raw Amp (a), raw CH (b), SAA produced Amp microparticles (c), SAA produced CH microparticles (d), physical mixture of SAA produced CH microparticles and raw Amp in the ratio 1:1 (e), physical mixture of SAA produced CH and Amp microparticles in the ratio 1:1 (f) have also been produced and are reported in Figure VII.10 for comparison purposes.

RESULTS PART II: COPRECIPITATION

95

0 10 20 30 40 50 60 70

g

f

e

d

c

b

a

Inte

nsity

2θ, degrees

Figure VII.10: X-Ray diffraction patterns of: a) raw Amp; b) raw CH; c) SAA produced Amp microparticles; d) SAA produced CH microparticles; e) physical mixture of SAA produced CH microparticles and raw Amp in the ratio 1:1; f) physical mixture of SAA produced CH and Amp microparticles in the ratio 1:1; g) SAA coprecipitated microparticles in the ratio 1:1.

The diffraction patterns evidence that: • Sharp peaks are produced by raw Amp (a), indicating a crystalline

solid state. • Raw CH is semi-crystalline with two characteristic peaks at 11° and

20° (b). • SAA processed Amp is nearly amorphous, but still exhibits three

peaks at 6°, 16° and 18°; the peak at 6° can be taken as reference (c). • The two peaks at 11° and 20° disappear in the SAA processed CH

(d) indicating a complete amorphous state. • The physical mixture of SAA produced CH and raw Amp shows all

the peaks characteristic of the drug (e).

Chapter VIII

96

• The physical mixture of SAA produced CH and Amp microparticles shows the three peaks of SAA processed Amp, especially the one at 6° (f).

• The pattern of coprecipitated powder shows a complete amorphous solid state, with no Amp peaks; in particular, the peak at 6° is not detected.

The comparison between the coprecipitated powder and the physical mixture of SAA produced CH and Amp microparticles supports the hypothesis that, during the particle formation in the SAA process, the drug and the polymer have no time to separate and precipitate in different respective microparticles, but both compounds remain entrapped in the same particle. The formation of this solid solution is probably due to the rapid evaporation of the solvent from droplets that causes the entrapment of drug molecules into the polymeric matrix. The organization of Amp molecules into a crystalline structure is hindered by the presence of the polymer and the powder results completely amorphous.

To confirm this behaviour of coprecipitated powders, the comparison between physical mixtures and coprecipitates has been performed for the powder produced using a polymer/drug ratio of 5:1 as well. Figure VII.11 shows the diffraction patterns of raw Amp (a), raw CH (b), SAA produced Amp microparticles (c), SAA produced CH microparticles (d), physical mixture of SAA produced CH microparticles and raw Amp in the ratio 5:1 (e), physical mixture of SAA produced CH and Amp microparticles in the ratio 5:1 (f) and SAA coprecipitated microparticles in the ratio 5:1 (g).

Also in this case, the Amp peak at 6° is visible in the physical mixtures; obviously, it is less pronounced than the mixture at ratio 1:1, because of the greater amount of polymer with respect to the drug in the mixture. SAA coprecipitated microparticles are completely amorphous, as the powder coprecipitated at the 1:1 ratio.

RESULTS PART II: COPRECIPITATION

97

0 10 20 30 40 50 60 70

g

f

e

d

c

b

a

Inte

nsity

2θ, degrees

Figure VII.11: X-Ray diffraction patterns of: a) raw Amp; b) raw CH; c) SAA produced Amp microparticles; d) SAA produced CH microparticles; e) physical mixture of SAA produced CH microparticles and raw Amp in the ratio 5:1; f) physical mixture of SAA produced CH and Amp microparticles in the ratio 5:1; g) SAA coprecipitated microparticles in the ratio 5:1.

SAA coprecipitates produced at ratios 2:1 and 8:1 have also been analyzed and their diffractograms have been compared together with those relative to coprecipitates at 1:1: and 5:1 in Figure VII.12. All coprecipitated powders show no Amp peaks and a complete amorphous solid state, thus confirming the hypothesis that they are formed by a solid solution of the drug and the polymer.

Chapter VIII

98

0 10 20 30 40 50 60 70

Inte

nsity

2θ, degrees

8:1

5:1

2:1

1:1

Figure VII.12: X-Ray diffraction patterns of SAA coprecipitated microparticles in the ratio 1:1, 2:1, 5:1 and 8:1.

DSC analysis Differential scanning calorimetry also allows to investigate the solid state

of materials; therefore, physical mixtures and coprecipitate powders have been analyzed by DSC to confute or confirm X-ray results.

Figure VII.13 reports thermograms of raw Amp (a), SAA produced CH microparticles (b), physical mixture of SAA produced CH microparticles and raw Amp in the ratio 1:1(c) and SAA coprecipitated microparticles in the ratio 1:1 (d).

RESULTS PART II: COPRECIPITATION

99

0 50 100 150 200 250 300

d

c

b

dH

/dt,

W/g

T, °C

a

End

o

Figure VII.13: DSC thermograms of: a) raw Amp; b) SAA produced CH microparticles; c) physical mixture of SAA produced CH microparticles and raw Amp in the ratio 1:1; d) SAA coprecipitated microparticles in the ratio 1:1.

Raw ampicillin trihydrate shows a broad endothermic peak at about 130°C, due to the evaporation of the bond water, and a second endothermic peak at about 205°C, due to the fusion of the crystalline habit (Han et al., 1998; Liu et al., 2004; Miller et al. 2004; Shefter et al. 1973), thus confirming X-ray results that showed the complete crystalline state of untreated Amp. At temperatures higher than 250°C the drug undergoes thermal degradation.

SAA processed CH shows a broad endothermic peak between 60 and 130°C, related to evaporation of water absorbed from air humidity and an exothermic peak at a temperature at about 290°C, due to a decrease in acetyl content and degree of polymerization (Kittur et al., 2002), as already discussed in Chapter VII.2.

Physical mixture of SAA produced CH microparticles and raw Amp in the ratio 1:1 shows four thermal events: the evaporation of water from CH (≃70°C) and from Amp (≃120°C), respectively; the fusion of crystalline Amp (≃205°C) and the deacetylation of CH (≃290°C).

SAA coprecipitated powder (d), instead, presents only three thermal events:

• A wide endothermic peak at about 90°C, related to the water evaporation, and an exothermic peak at about 290°C; both peaks are characteristic of CH.

Chapter VIII

100

• A broad endothermic peak at 225°C that can be exclusively ascribed to the evaporation of Amp bond water. Indeed, this peak cannot represent a fusion event, since X-ray analysis shown that SAA coprecipitated powder is completely amorphous.

These results not only confirm X-ray analysis, but also show that ampicillin entrapped into coprecipitated microparticles is more stable than the raw drug; indeed, the loss of bond water takes place at temperatures higher with respect to the untreated Amp and to the physical mixture ones. Therefore, the interaction between CH and Amp results in a stabilizing effect of CH on Amp.

To confirm these considerations, DSC analyses have been performed also on the SAA coprecipitated powders at other polymer/drug ratios. Figure VII.14 reports thermograms of raw Amp (a), SAA produced CH microparticles (b), physical mixture of SAA produced CH microparticles and raw Amp in the ratio 5:1 (c) and SAA coprecipitated microparticles in the ratio 5:1 (d).

0 50 100 150 200 250 300

d

b

c

dH/d

t, W

/g

T, °C

a

End

o

Figure VII.14: DSC thermograms of: a) raw Amp; b) SAA produced CH microparticles; c) physical mixture of SAA produced CH microparticles and raw Amp in the ratio 5:1; d) SAA coprecipitated microparticles in the ratio 5:1.

The physical mixture again presents four thermal events characteristic of both CH and raw Amp; obviously, the areas of the peaks related to Amp are smaller than peaks at ratio 1:1, because of the smaller amount of drug with respect to the polymer.

RESULTS PART II: COPRECIPITATION

101

SAA coprecipitated powder shows no fusion peak and a very smooth peak at about 225°C, related to the bond water evaporation of Amp. In this case, this latter event not only occurs at a higher temperature than the untreated drug and than the physical mixture, but it is also less pronounced than the peak exhibited by the SAA coprecipitates at ratio 1:1, thus, indicating an enhanced stabilizing effect of CH. This effect is more evident when the thermograms of coprecipitated microparticles at different polymer/drug ratios are compared in the same graph, as shown in Figure VII.15: the richer in polymer is the powder, the lower is the area of the peak at 225°C. Moreover, the evaporation temperature of bond water slightly moves from 223°C to 230°C as the ratio varies from 1:1 to 8:1. An explanation of this effect can be related to the degree of drug dispersion into the polymer matrix; i.e., the richer in CH is the liquid solution, the higher is the dispersion of Amp into the polymer matrix of final microparticles. Consequently, when the powder is exposed to heat, the polymer shields the drug and hinders the evaporation of Amp bond water.

0 50 100 150 200 250 300

dH/d

t, W

/g

T, °C

8:1

5:1

2:1

1:1

Raw Amp

End

o

Figure VII.15: DSC thermograms of SAA coprecipitated microparticles in the ratio 1:1, 2:1, 5:1 and 8:1. DSC thermogram of raw Amp is reported as reference.

VII.2.2.2 EDX analysis

SAA coprecipitated powders have also been characterized by EDX microanalysis to verify if each microparticle is formed by both the drug and the polymer, or if the two compounds precipitate separately. Indeed,

Chapter VIII

102

microanalysis not only reveals what elements are present in the sample, but can also produce an elemental maps showing the spatial variation of elements in the sample.

In Figure VII.16 ampicillin and chitosan chemical formulas are reported, showing that both compounds are formed by oxygen (O), carbon (C), nitrogen (N) and hydrogen (H); the only element that differentiates ampicillin from CH is a sulphur (S) atom. Therefore, S is the element that can indicate the location of the drug in the coprecipitated powders.

Ampicillin

Chitosan

Figure VII.16: Ampicillin and chitosan chemical formulae.

Physical mixtures of SAA produced CH microparticles with raw Amp and with SAA processed Amp have been EDX analysed to compare it with SAA coprecipitated microparticles.

EDX analysis on the physical mixture of SAA produced CH microparticles and raw Amp in the ratio 1:1 is reported in Figure VII.17, showing the SEM image of the analyzed area, the Oxygen map reported as reference and the Sulphur map. In the SEM image of the analyzed area, the spherical microparticles represent CH, whereas Amp is present as crystals. This figure clearly shows that Oxygen (O) is uniformly distributed in both CH microparticles and Amp crystals, whereas Sulphur (S) is contained only in the crystals. This analytical technique, therefore, allows a good localization of Amp with respect to CH in a physical mixture.

RESULTS PART II: COPRECIPITATION

103

Figure VII.17: EDX microanalysis of physical mixture of SAA produced CH microparticles and raw Amp: SEM image of the analysed area (SEM); Oxygen map (O); Sulphur map (S).

A physical mixture in the ratio 1:1 of CH and Amp microparticles, both produced by SAA, has also been analyzed to verify if the localization of the drug is possible even when both compounds show the same morphology. The SEM image of the analyzed area and EDX maps are reported in Figure VII.18. In this case, the difference between CH and Amp is less marked than the previous mixture, because Amp particles have small size and very fine nanometric particles (of both Amp and CH) are dispersed onto larger ones. Nevertheless, elemental maps show that Oxygen is uniformly distributed in all the area, as expected, whereas Sulphur is not present in all microparticles, but concentrated in few spots, i.e. in the Amp microparticles. For example, the yellow circles in the Figure VII.18 point out three microparticles in the SEM image. These three particles are clearly visible also in the Oxygen map, whereas only two of them are present in the Sulphur map, demonstrating that CH only constitutes the third particle.

SEM

O S

10 µm

Chapter VIII

104

Figure VII.18: EDX microanalysis of physical mixture of SAA produced CH microparticles and SAA produced ampicillin microparticles: SEM image of the analysed area (SEM); Oxygen map (O); Sulphur map (S).

Finally, coprecipitated microparticles in the ratio 1:1 have been analyzed. SEM images and elemental maps are reported in Figure VII.19 and Figure VII.20, respectively showing a landscape at high magnification and two particulars at low magnifications of the analyzed area. Unlike the physical mixture, in Figure VII.19 not only Oxygen, but also Sulphur is uniformly spread over all the area and the maps of the two elements practically overlap. This is also confirmed by Figure VII.20, reporting three particles selected at random in the analyzed area: Sulphur is revealed in each particle, i.e. each particle is formed by both CH and Amp.

These results definitively confirm that during SAA of CH/Amp/acid water, the drug and the polymer do not precipitate separately, but coprecipitate into microparticles formed by a solid solution of the two components.

SEM

O S

10 µm

RESULTS PART II: COPRECIPITATION

105

Figure VII.19: EDX microanalysis of SAA coprecipitated CH/Amp microparticles. Landscape of: SEM image of the analysed area (SEM); Oxygen map (O); Sulphur map(S).

Figure VII.20: EDX microanalysis of SAA coprecipitated CH/Amp microparticles. Particulars of: SEM image of the analysed areas (SEM); Oxygen maps (O); Sulphur maps (S).

S O

SEM

20 µm

S

S O SEM

O SEM

5 µm

5 µm

Chapter VIII

106

VII.2.2.3 Drug release studies

X-ray, DSC and SEM-EDX analysis have confirmed that SAA coprecipitated microparticles are formed by a solid solution of CH and Amp, and that the drug is entrapped in an amorphous state into the polymeric matrix. However, none of the previous analysis is able to determine if the drug is really uniformly dispersed into the particle or if, for instance, it is concentrated only onto its surface and how it will be release from the polymer.

The real efficiency of CH/Amp as controlled release delivery systems has been, therefore, analyzed performing drug release tests.

The release rate is influenced by the drug spatial distribution into the particles: • if the drug is concentrated on the particle surface, a burst effect occurs,

i.e. an initial fast release on the drug is obtained, followed by a secondary slower release rate;

• if the drug is concentrated in the core of the particle, a nearly zero initial release occurs (delayed release), followed by a secondary faster release once the dissolution medium has reached the core;

• if the drug is uniformly distributed in the polymeric matrix, neither burst effect nor delayed release occurs.

Drug content determination and drug release rate analysis have been performed on SAA coprecipitated powders is the ratios 1:1, 2:1, 5:1 and 8:1.

Drug content in SAA coprecipitates has been determined suspending a known amount of Amp-loaded microparticles into physiological saline solution at pH 7.2. Suspension was kept at 37°C and stirred at 200 rpm for 15 days. Then, the amount of drug incorporated was assayed by spectrophotometric analysis, revealing that the ratio between CH and Amp in the liquid solution is maintained, also after SAA processing.

Drug release profiles have been obtained using a physiological saline solution (pH 7.2) as dissolution medium. A known amount of powder to be analyzed has been prepared into tablets obtained by direct compressing of powder at 150 bar for 10 min using a standard tablet preparation device. Tablets have been placed into 1000 mL of physiological solution, kept at 37°C and stirred at 200 rpm. At predetermined time intervals, the concentration of drug has been assayed using an UV spectrophotometer.

The weight of the coprecipitate to be analyzed at the different polymer/drug ratios has been calculated in order to load 50 mg of Amp in each tablet. After the analysis, tablets have been dried and then weighed to verify the amount of dissolved Amp.

To verify the improvement in drug release rate of SAA composite microparticles with respect to physical mixtures, Figure VII.21 reports Amp release profiles from: the physical mixture of raw CH and raw Amp in the ratio 5:1 (a), from the physical mixture of SAA produced CH microparticles

RESULTS PART II: COPRECIPITATION

107

and raw Amp in the ratio 5:1 (b), from the SAA coprecipitated CH/Amp microparticles in the ratio 5:1 (c).

0 20 40 60 80 100 120

20

40

60

80

100

c

b

a

Phys. Mix. Raw CH : Raw Amp=5:1 (a) Phys. Mix. SAA CH : Raw Amp=5:1 (b) CH : Amp = 5:1 (c)

Am

pici

llin

Rel

ease

d, %

Time, min

0 12 24 36 48 60 72 84 96

20

40

60

80

100c

ba

Phys. Mix. Raw CH : Raw Amp=5:1 (a) Phys. Mix. SAA CH : Raw Amp=5:1 (b) CH : Amp = 5:1 (c)

Am

pici

llin

Rel

ease

d, %

Time, hours

Figure VII.21: In vitro Amp release profiles from: physical mixture of raw CH and raw Amp in the ratio 5:1 (a); physical mixture of SAA produced CH microparticles and raw Amp in the ratio 5:1 (b); SAA coprecipitated CH/Amp microparticles in the ratio 5:1 (c). Drug release rates from coprecipitates are reported up to 120 min (left) and to 96 h (right).

Raw Amp dissolution profile is not reported in Figure VII.21, since it dissolves completely in only 15 min.

This figure clearly shows that SAA coprecipitated powder exhibits the slowest drug release rate: physical mixture of raw materials releases 100% of the drug after only 38 min; whereas, during the same time, the physical mixture of SAA CH microparticles with raw Amp and SAA coprecipitated microparticles released about 36% and 10% of the drug, respectively. After 13.5 hours, the physical mixture of SAA CH microparticles and raw Amp has completely released the drug; whereas, SAA coprecipitates have released only 64% of the encapsulated drug. After 4 days SAA coprecipitate has released the 96% of the drug. Moreover, no burst effect or delayed release is observed in drug release profiles from SAA coprecipitated powders, indicating no drug concentration on the particles surface.

Table VII.2 reports the time values related to the 50% (t50) and 90% (t90) of Amp released from physical mixtures and SAA coprecipitates at ratio 5:1, pointing out that the drug release times have the order of magnitude of minutes and hours for the physical mixtures, and of days for the SAA coprecipitates.

Chapter VIII

108

Table VII.2: Time values related to the 50% (t50) and 90% (t90) of Amp released from physical mixtures and SAA coprecipitates at ratio 5:1.

Sample t50 t90

Physical mixture of Raw CH and Raw Amp (5:1) 9.6 min 22.8 min

Physical mixture of SAA CH microparticles and Raw Amp (5:1)

1 h 6 h

SAA composite CH/Amp microparticles (5:1) 6 h 3.6 days

It is interesting to observe that a physical mixture prepared using

micronized CH causes the slowing of drug release with respect to the physical mixture of raw materials; this is due to a larger dispersion of Amp and CH particles. The degree of drug dispersion is the key parameter affecting release rate and, therefore, explains the slowest release rate of SAA coprecipitate with respect to physical mixtures, since CH and Amp are molecularly dispersed into microparticles. According to this hypothesis, the richer in polymer are microparticles, the slower should be the drug release rate. Therefore, drug release has also been measured from microparticles precipitated at different CH/Amp ratios and results have been reported in Figure VII.22.

0 3 6 9 12 15 18 21 240

20

40

60

80

100

c

b

a

CH : Amp = 1:1 (a)CH : Amp = 5:1 (b)CH : Amp = 8:1 (c)

Am

pic

illin

Re

leas

ed

, %

Time, hours

0 12 24 36 48 60 72 84 960

20

40

60

80

100

cb

a

CH : Amp = 1:1 (a)CH : Amp = 5:1 (b)CH : Amp = 8:1 (c)

Am

pic

illin

Rel

ease

d, %

Time, hours Figure VII.22: In vitro Amp release profiles from SAA coprecipitated CH/Amp microparticles in the ratio 1:1 (a), 5:1 (b) and 8:1 (c). Drug release rates from coprecipitates are reported up to 24 h (left) and to 96 h (right).

These analyses reveal that, in this case, the polymer/drug ratio is an important parameter to modulate drug release rate from microparticles. As expected, the higher is the ratio, the slower is drug release. Table VII.3

RESULTS PART II: COPRECIPITATION

109

reports the time values related to the 50% (t50) and 90% (t90) of Amp released from SAA coprecipitates at different ratios.

Table VII.3: Time values related to the 50% (t50) and 90% (t90) of Amp released from SAA coprecipitates at different CH:Amp ratios.

Sample t50 t90

CH : Amp = 1:1 52.8 min 2 days

CH : Amp = 5:1 6 h 3.6 days

CH : Amp = 8:1 17.2 h 3.7 days

Both Figure VII.22 and Table VII.3 clearly show that release rates are

mostly influenced by the loading ratio in the first part of the dissolution profiles: the initial slope of the curves decrease as the ratio increases. This effect can be explained considering the mass transfer mechanism between microparticles and dissolution medium. Indeed, the higher is drug content in microparticles (i.e. at low ratios), the higher is the concentration gradient between microparticles and dissolution medium, and the faster is drug release. Obviously, the maximum gradient is at the start of release, where the differences between the release rates at different polymer/drug ratios are more marked.

The possibility of modulating drug release rate by changing the loading ratio is very important for the pharmaceutical tailoring of this kind of formulation. If, for example, the optimal releasing time for an oral administration of Amp is 24 h, the CH/Amp coprecipitation ratio showing the most suitable release rate for this purpose is 1:1; indeed, after 24 it has released more than 80% of the loaded drug.

Finally, no burst effect is observed in all drug release rates, thus confirming the good dispersion of the Amp into the particles and the formation of a drug/polymer solid solution between CH and Amp.

VII.2.2.4 Interpretation of drug release results and

mechanisms

The aim of this thesis was to investigate the possibility of applying the SAA technique to the microparticles production for drug controlled release. Results obtained for CH and Amp coprecipitation have proved that this goal can be successfully reached using SAA.

For completeness, in the next paragraphs, we try to give an interpretation of drug release kinetics from SAA coprecipitates.

Before focusing on CH/Amp coprecipitates, a general description of swelling controlled systems is provided; indeed, it is known that CH

Chapter VIII

110

produces controlled drug release matrices based on its gelling capacity. A possible model to describe Amp release kinetics from SAA coprecipitates is suggested. Indeed, the mathematical modelling of drug release is of great importance in pharmaceutical science and engineering, because it allows to predict the effects of the composition and geometry on drug release profiles; this is very helpful to the design of new drug delivery systems. Obviously, the modelling of release kinetics is not the aim of this thesis; therefore, relatively simple models described in the literature have been used.

General characteristics of swelling controlled release systems Over the last 25 years, hydrogels have become popular carriers for drug

delivery applications due to their biocompatibility and resemblance to biological tissues (Langer and Peppas, 2003; Lowman et al., 2004). From a structural point of view, hydrogels are three-dimensional hydrophilic polymer networks that swell in water or biological fluids without dissolving as a result of chemical or physical crosslinks. Hydrogels can be used to target the release of a drug or protein to a specific area of the body and simultaneously control the release kinetics due to their three-dimensional structure (Brazel and Peppas, 1999).

Hydrogel-based devices belong to the group of the swelling-controlled drug delivery systems (Colombo et al., 2000b, 2004). When the polymer network comes in contact with aqueous solutions, the thermodynamic compatibility of the polymer chains and water causes the polymer to swell. As water penetrates inside the glassy network, the glass transition temperature of the polymer decreases and the hydrogel becomes rubbery. The drug trapped inside the hydrogel dissolves with the imbibed water and starts diffusing out of the network. Three driving forces contribute to this phenomenon: a penetrant concentration gradient, a polymer stress gradient and osmotic forces.

In the case of non-swelling controlled delivery systems, the relaxation rate of the polymer is very slow in comparison to the water transport inside the hydrogel. Then, the transport mechanism in this type of systems follows Fickian diffusion. When the macromolecular chain relaxation is the dominating driving force, non-Fickian transport is observed. However, in many swelling-controlled delivery systems, non-Fickian transport mechanism has been observed, characterized by an intermediate Fickian diffusion and non-Fickian transport.

In general, drug release from swellable matrix tablets is based on glassy-rubbery transition of polymer as a result of water penetration into the matrix. Whereas interactions between water, polymer and drug are the primary factors for release control, various formulations variables, such as polymer grade, drug/polymer ratio, drug solubility, drug and polymer particle size and compaction pressure, can influence drug release rate to greater or lesser degree. However, the central element of the mechanism of drug release is the

RESULTS PART II: COPRECIPITATION

111

gel layer (rubbery polymer) that forms around the matrix and is capable of preventing matrix disintegration and further rapid water penetration. Water penetration, polymer swelling, drug dissolution and diffusion and matrix erosion are the phenomena determining gel layer thickness. Therefore, drug release is controlled by drug diffusion through the gel layer and/or by erosion of the gel layer.

The gel layer strength is important in the matrix performance and is controlled by the concentration, viscosity and chemical structure of the rubbery polymer. During drug delivery, the gel layer is exposed to continuous changes in its structure and thickness. It begins when the polymer becomes hydrated and swells. The polymer chains are strongly entangled and the gel layer is highly resistant. However, moving away from this swelling position, the gel layer becomes progressively hydrated and, when sufficient water has accumulated, the chains disentangle and the polymer dissolves. Thus, the gel layer thickness dynamics in swellable matrix tablet show three distinct regimes: it increases when the penetration of water is the fastest phenomenon (Figure VII.23 a, b), is constant when the disentanglement rate is similar to the penetration (Figure VII.23 b, c), and decreases when all of the polymer is in the rubbery phase (Figure VII.23 c,d) (Harland et al., 1988).

Figure VII.23: Time dependence of normalized gel layer thickness in presence of drug (adapted from Harland et al., 1988).

The swelling behaviour of swellable matrices is mechanistically described by front positions, where ‘front’ indicates the position in the matrix where the physical conditions sharply change. Thus in swellable matrix tablets three fronts could be expected, as shown in Figure VII.24:

1. the swelling front, the boundary between the still glassy polymer and its rubbery state,

Time

Nor

mal

ized

gel

laye

r th

ickn

ess,

δ

a

b

c

d

Chapter VIII

112

2. the diffusion front, the boundary in the gel layer between the solid, as yet undissolved, drug and the dissolved drug,

3. the erosion front, the boundary between the matrix and the dissolution medium.

Figure VII.24: Schematic illustration (cross-section view) o a swellable matrix tablet during radial drug release

Drug release kinetics is strictly associated with the dynamics of the gel layer. Initially it ranges from the Fickian to non-Fickian, and subsequently from quasi-constant to constant, becoming first order at the end. However, swellable matrices rarely show all three described regimens during the release time of the drug.

An example of the front movement of swellable cylindrical matrices is reported in Figure VII.25, showing a photograph of the upper base of a HPMC cylindrical matrix containing buflomedil pyridoxalphosphate after one hour of swelling–release (Colombo et al., 2000a). The drug has a light yellow colour; whereas, its aqueous solutions range from yellow to intense orange (depending on the solution concentration); during drug release the fronts are visible on the matrix as concentric circles, corresponding to a sharp change of colour. This figure clearly shows the three fronts during the swelling–dissolution process: from centre to matrix periphery, the swelling front (polymer glassy–rubbery transition boundary), the diffusion front (solid drug–drug solution boundary) and the erosion front (swollen matrix–solvent boundary) can be identified. The erosion front moves outwards, due to the swelling of the matrix, or inwards when the matrix dissolves; whereas, the swelling front moves inwards after water penetration. The gel layer thickness is measured as the distance between erosion and swelling front.

RESULTS PART II: COPRECIPITATION

113

Figure VII.25: Photograph of the upper base of a HPMC cylindrical matrix containing 60% w/w of buflomedil pyridoxalphosphate, placed in between two transparent discs after one hour of swelling–release (adapted from Colombo et al., 2000a).

The relevance of the front movement to gel-layer thickness behaviour is related to drug release kinetics. It was demonstrated that, when the polymer is soluble enough, the gel-layer thickness can remain constant and the drug can be released at a constant rate. This behaviour is explained by the front synchronization movement, that is, the swelling and erosion fronts move in parallel and are strongly dependent on the polymer characteristics (because when the gel layer synchronized, polymer dissolution dominates the delivery kinetics). This means that drug solubility becomes not relevant in rate determination.

A qualitative interpretation of Amp release kinetics exhibited by SAA coprecipitates can be given on the basis of the front movements. Rate and kinetics of drug release from CH matrix tablets are controlled by the dynamics of gel layer thickness. At the beginning of the release process, there is a steep water concentration gradient at the CH/water interface, resulting in water imbibition into the matrix. Due to the imbibition of water CH swells, resulting in dramatic changes of polymer and drug concentrations, increasing dimensions of the system and increasing macromolecular mobility. Upon contact with water, the Amp rapidly dissolves (since it is highly soluble in water) and, due to concentration gradients, diffuses out of the device.

A schematic representation of these hypothesized mechanisms is shown in Figure VII.26, in which water penetration and drug diffusion are supposed axial; indeed, due to the small height of the tablet with respect to its diameter, radial diffusion can be neglected as a first approximation.

Chapter VIII

114

Water

Water

Glassy polymer

(a) (b)

Glassy polymer

Rubbery polymer

δδδδ

Drug

Drug

Figure VII.26: Schematic representation of swellable tablet containing CH/Amp coprecipitate during axial drug release.

Initially, the water penetration is faster than chain disentanglement and a quick build-up of the gel layer thickness takes place; the Amp transport mechanism could follow Fickian diffusion. But, when water penetrates slowly, due to the increase of the diffusional distance, a small change in gel thickness is obtained because water penetration and polymer disentanglement rate are similar (Lee, 1981) and drug release kinetic ranges from the initial Fickian to the non-Fickian transport. When all CH is in the rubbery phase, the gel layer thickness is constant, since erosion mechanism in CH is negligible. Indeed, all CH/Amp tablets dried after Amp release showed a weight equal or higher than the tablet itself before the release test, proving no significant CH loss. When the gel layer is constant, a first-order drug release can occur.

Dimensionless parameters and models to describe water and drug transport

Water uptake and drug delivery from swelling-controlled release systems can be described by two dimensionless parameters: the diffusional Deborah number (De), which relates the net water motion to the rate of polymer relaxation, and the Swelling Interface number (Sw), relating water penetration into a network to a dispersed solute diffusion from the polymer. These parameters are sensitive to both polymer and drug properties.

The diffusional De is expressed as a ratio between the characteristic polymer relaxation time and a characteristic diffusion time:

[ ]2

2,1

)(t

DDe

δλ

θλ == (Eq.1)

where λ is the characteristic relaxation time for the polymer when subjected to swelling stresses, and θ is the characteristic penetrant diffusion time into swelling sample. θ is defined as the square of the half-thickness of a thin-disc sample divided by the diffusion coefficient of water in the

RESULTS PART II: COPRECIPITATION

115

polymer (δ2/D1,2). If either the relaxation time (De>>1) or solvent diffusion (De<<1) dominates the swelling process, the time dependence is Fickian; however, if De is on the order of 1, the two processes will occur on the same time scale, leading to non-Fickian transport behaviour.

The swelling interface number (Sw) is important in describing the balance between solvent penetration and solute diffusion:

21,3DSw rυδ= (Eq.2)

where υ is the velocity of the moving glassy/rubbery front, δ is the thickness of the swollen gel layer, and D3,21 is the diffusion coefficient of solute in the polymer. Sw is sensitive to both the polymer structure (as it affects penetrant uptake) and drug properties through the solute diffusion coefficient. When Sw is significantly greater or smaller than 1, either water penetration or drug diffusion will control the release pattern and the time dependence will be Fickian; however, when Sw is of the order of 1, non-Fickian behaviour prevails.

Both dimensionless numbers, De and Sw, have been shown to vary with time, as the diffusion coefficients and swelling front velocity are not constant during the swelling process. Therefore, either the initial or equilibrium values of Sw and De need to be experimentally determined.

A successful mechanistic analysis of drug release in relation to a matrix swelling can be done using the swelling interface number, Sw. This number is analogous to the Peclet number in that it compares a pseudo-convective process to a diffusional process. However, whereas the Peclet number defines these processes for the same diffusant, the swelling interface number relates transport phenomena of a penetrant and a solute. As the term ‘ν δ(t)’ of Eq. 2 has units of area over time, a new dimensionless number, similar to Sw, describes the significant three-dimensional expansion of the matrix due to swelling and its associated influence on drug release. This is important because Sw is inherently related to one-dimensional transport (disc or films), whereas the improved dimensionless number can describe the behaviour of truly three-dimensional matrices. As the swellable matrix tablet is characterized by a major change of surface area, the water front mobility is replaced in the new dimensionless number by a matrix swelling expansion characteristic. The increase of the releasing area produced by the matrix during swelling is used as a measure of the matrix expansion. Therefore, the new number, the Swelling Area number (Sa) is defined as:

Ddt

dASa

1= (Eq.3)

Chapter VIII

116

where dA/dt is the expansion rate if the surface area of the swellable matrix (Colombo at al., 1992). This Sa number suggests a dependence of drug release on matrix releasing area.

An analysis of release behaviour using dimensionless numbers need their experimental determination and the analysis of tablet size modifications during release, that are not available in the case of Amp release from coprecipitates studied in this thesis. Therefore, an analysis based on models is more suited to our case.

The quantity of drug released from matrix tablets is often analyzed as a function of the square root of time (Higuchi’s law); this is typical for systems where drug release is governed by pure diffusion. However, the use of this relationship in swellable systems is not completely justified, as such systems can be erodible and the contribution of the relaxation of polymeric chains to drug transport has to be taken into account. Therefore, analysis of drug release from swellable matrices must be performed with a flexible model that can identify the different contribution to overall kinetics.

An empirical equation able to predict the drug release and the water concentration profiles with simultaneous swelling was developed by Peppas et al. (1980). The equation is based on a power law dependence of the fraction released on time. The exponent n has values that can range between 0.43 and 1, according to the geometry and the prevalence of the Fickian or the case II (relaxation) transport. The equation has the following form:

nt ktM

M =∞

(Eq.4)

where Mt is the drug released at time t, M∞ is the amount of drug released at infinite time, k is a kinetics constant, and n is the diffusional exponent.

These authors (Davidson and Peppas, 1986) studied various sample geometries and determined appropriate n values for spherical, cylindrical and slab geometries. In addition, they defined an aspect ratio that could be used to determine the appropriate exponent for a system.

A binomial equation (Eq. 5), similar in meaning to Eq. 4, was also proposed by Hopfenberg (1978) and adapted to pharmaceutical problems by Peppas and Sahlin (1989). In this equation the contribution of the relaxation or erosion mechanism and of the diffusive mechanism can be quantified:

mmt kkM

M 221 ττ +=

∞

(Eq.5)

where k1 is the diffusional constant, k2 is the relaxational constant and m is the diffusional exponent.

Eq. 5 allows the quantitative estimation of the mechanisms involved, by means of calculation of the ratio between the relaxation contribution (R) and the diffusional contribution (F), according to the following formula:

RESULTS PART II: COPRECIPITATION

117

mtk

k

F

R

1

2= (Eq.6)

In this model, the phenomenon of polymer dissolution is not considered. Successive models, accounting for concomitant swelling and dissolution

of the polymer, were proposed, for example, by Lee and Peppas (1987), Harland et al. (1988), Siepmann and Peppas (2000). These models take into consideration the position of the relevant fronts and the pertinent parameters of drug and polymer. The considered fronts, i.e., the swelling front (R) and the erosion front (S), are illustrated in Figure VII.25 together with the gradient involved in the analysis.

Thickness

Con

cent

ratio

n

Glassy state Rubbery state Dissolution medium

Ccp

Ccd

C*

Cs

Cb

Cd

1-Cd-Cb

1-Cs-C*

R S a

Figure VII.27: Definition of main parameters in swellable-soluble matrix tablet.

The parameters are defined as follows:

─ Ds is the water diffusion coefficient in the drug-polymer matrix. ─ Dd is the drug diffusion coefficient in the swollen polymer. ─ Cs is the drug solubility (expressed as volume fraction) at the drug

core interface (S). ─ C* is the polymer volume fraction at the gel-solution interface (S). ─ Cd is the polymer volume fraction at S. ─ Ccd is the drug volume fraction in the glassy core. ─ Ccp is the polymer volume fraction in the glassy core.

For a cylindrical tablet having thickness a and surface area of basis A,

from which the release is one-dimensional, the normalized gel layer

Chapter VIII

118

thickness (δ) is given by the relative position of swelling and dissolution fronts, according to the expression:

a

RS )( −=δ (Eq.7)

The gel layer thickness varies with time, according to the sketch proposed in Figure VII.27. The behavior of the gel layer thickness reveals two different kinds of drug release in time. The first one is related to the early swelling region where the thickness of gel layer increases. In this case, water and drug diffusion can be described by Fick’s second law for cylindrical devices, taking into account axial and radial mass transfer with concentration-dependent diffusion coefficients. Mass balance, therefore, can be expressed as:

∂∂

∂∂+

∂∂

∂∂+

∂∂

∂∂=

∂∂

z

CrD

z

C

r

D

r

CrD

rrt

C kk

kkkk

k

θθ1

(Eq.8)

where Ck and Dk are the concentration and diffusion coefficient of the diffusing species (water and drug), r is the radial coordinate, z is the axial coordinate, θ is the angular coordinate, and t represents time. Figure VII.28 shows a schematic illustration of the matrix for mathematical analysis (a), the symmetry planes in axial and radial direction for water and drug concentration profiles (b) and the “sequential layer” structure for numerical analysis (c). As there is no concentration gradient of any component with respect to θ, Eq. 8 can be transformed into:

∂∂

∂∂+

∂∂+

∂∂

∂∂=

∂∂

z

CD

zr

C

r

D

r

CD

rt

C kk

kkkk

k (Eq.9)

RESULTS PART II: COPRECIPITATION

119

Figure VII.28: (a) Scheme of the matrix for mathematical analysis, with (b) symmetry planes in axial and radial direction for the water and drug concentration profiles, (c) “sequential layer” structure for numerical analysis.

A special approximate solution of mass balance equation gives rise to a prediction of the amount of drug released, that can be expressed as:

ttM

M t γα +=∞

2/1 (Eq. 10)

Chapter VIII

120

where α and γ are constants (Colombo et al., 2000b). It is interesting to observe that Eq. 10 is similar to the empirical Eq. 5

proposed by Peppas and Sahlin (1989) for the analysis of swellable systems, where the two mechanisms controlling the release, both relaxation and diffusion, occur. The equation predicts that the drug released is described by the addition of a relaxation term (with t-dependence) to a diffusional term (with t1/2-dependence).

The second kind of drug quantity released in time is relative to the synchronization region, when the value of δ remains constant. The amount released in this situation is expressed as:

tM

M t εζ +=∞

(Eq.11)

where ζ and ε are constants. In this part of drug release, the delivery equation can be expressed as a zero-order equation. This indicates that in the synchronization region (gel layer thickness constant), drug release is a function of drug loading (Ccd) receiving solution concentration (Cb) and true dissolution parameter of polymer (Cd).

The study about mechanisms of drug release via a swellable and dissoluble hydrophilic polymer matrix is not as extensive as for purely diffusion, swelling or polymer dissolution controlled drug release systems since all these processes are coupled, thus making the models more intricate and difficult to solve. Both empirical and theoretical more complex mathematical models have been proposed for drug release via either cross-linked or uncross-linked polymer matrix in literature (Wu et al., 2005; Mustafin et al., 2005; Serra et al., 2006).

Fitting of Ampicillin release kinetics from SAA coprecipitates

In the context of this Ph.D. thesis, we believe that it is convenient to use a simple empirical model as the first modelling attempt to describe Amp release from SAA coprecipitates. For this purpose, the equation proposed by Peppas and Sahlin (1989) is suitable to describe CH-based swelling systems; indeed, in this model the phenomenon of polymer dissolution is not considered, i.e. the case of CH that is completely insoluble in water.

Therefore, the equation used to fit drug release data is Eq. 5:

mmt tktkM

M 221 +=

∞

where

t is the release time;

Mt is the Ampicillin released at time t;

RESULTS PART II: COPRECIPITATION

121

M∞ is the total amount Ampicillin loaded in the tablet, i.e. 50 mg in each tablet;

k1 is the diffusional constant;

k2 is the relaxational constant;

m is the diffusional exponent. Data fitting has been carried out using Micromath Scientist 3.0 software.

This software employs a least squares minimization procedure based on a modification of Powell's algorithm. The goodness of fit has been evaluated by the coefficient of determination, defined as:

∑

∑

=

=

−

−−= n

idatai

n

ifitidatai

yy

yyR

1

2,

1

2,,

2

)(

)(1 (Eq.12)

where

dataiy , is Mt/M∞ value of experimental data

fitiy , is Mt/M∞ value calculated from Eq. 5

y is the arithmetical mean of experimental Mt/M∞ values The model should be able to predict the release kinetics at different

polymer/drug ratios (r). Therefore, the effect of this factor must be taken into account.

To determine which parameter is depending on r, as first step the initialization of the parameters has been performed on the arithmetical mean of values calculated from independent curves, i.e. k1, k2 and m pure constants for each kinetic; these values have been reported in Table VII.4.

Table VII.4: k1, k2 and m values calculated as pure constants using Eq. 5 for Amp release kinetics from SAA coprecipitates at different CH/Amp ratios (r).

r = 1 r = 5 r = 8

k1 18.897580 3.718750 0.860900

k2 0.000978 0.000908 0.000912

m 0.130208 0.254740 0.358530

Chapter VIII

122

Then, k1, k2 and m have been plotted vs. r and a very good exponential fit has resulted for k1 and k2, whereas a linear fit has resulted for m, as shown in Figure VII.29.

RESULTS PART II: COPRECIPITATION

123

0 1 2 3 4 5 6 7 8 9 100

2

4

6

8

10

12

14

16

18

20

k 1

r

0 1 2 3 4 5 6 7 8 9 100.00088

0.00090

0.00092

0.00094

0.00096

0.00098

0.00100

k 2

r

0 1 2 3 4 5 6 7 8 9 100.0

0.1

0.2

0.3

0.4

0.5

m

r

Figure VII.29: Exponential fit (─) of k1 and k2 values vs. r and linear fit (─) of m values (●) vs. r.

y = 28.5733 e-0.4130 x

R2 = 0.9973

y = 0.00091 + 0.00037e-1.70704 x

R2 = 0.9974

y = 0.0325 x + 0.096

R2 = 0.9991

Chapter VIII

124

These results point out that both diffusional and relaxational contributions depends on the polymer/drug ratio. Therefore, Eq. 5 has been modified, and k1, k2 and m have been considered no more as pure constants, but as functions of r, thus leading to the following system of equations:

)(22

)(1 )()( rmrmt rkrk

M

M ττ +=∞

r]a[ak 21101+= (Eq. 13)

r]a[ak 43102+=

raam 65 +=

Eq. 13 have been used to calculate the fitting curves of Amp release data at different r shown in Figure VII.30; Table VII.5 reports the related k1, k2 and m values calculated as functions of r. The model shows a particular good prediction for Amp release at r = 8 (R2 = 0.99) and a satisfying prediction of drug release at r = 1 (R2 = 0.95) and r = 5 (R2 = 0.95).

0 12 24 36 48 60 72 84 960

20

40

60

80

100 experimental data at r = 1 experimental data at r = 5 experimental data at r = 8

model curve at r = 1 (R2=0.9495)

model curve at r = 5 (R2=0.9496)

model curve at r = 8 (R2=0.9916)

Am

pici

llin

Rel

ease

d, %

Time, h

Figure VII.30: Fitting curves of experimental Amp release data calculated using k1, k2 and m as functions of r (Eq. 13).

RESULTS PART II: COPRECIPITATION

125

Table VII.5: k1, k2 and m values calculated as functions of r using Eq 13 for Amp release kinetics from SAA coprecipitates.

r = 1 r = 5 r = 8

k1(r) 19.62959 3.18513 0.81431

k2(r) 3.67 E-07 4.54 E-07 5.32 E-07

m(r) 0.12726 0.26646 0.37086

In order to ascertain the importance of the two mechanisms for drug

release, the ratio of relaxational (R) and Fickian (F) contributions can be calculated using Eq. 6. The values of the R/F vs. Amp released fraction are presented in Figure VII.31.

0 20 40 60 80 100

1

2

3

4

5

6

7

8

r = 1 r = 5 r = 8

Rel

axat

ion/

Fic

kian

, R/F

* 1

0-5

Ampicillin Released, %

Figure VII.31: Relaxation/Fickian ratio (R/F) as a function of the fractional Amp release obtained by coprecipitates at different CH/Amp ratios (r).

This graph shows that as the polymer/drug ratio increases, the contribution of the relaxational release mechanism increases with respect to the Fickian one. This effect is probably due to the increase of gel thickness when r ranges from 1 to 8: the thicker is the layer, the longer is the time during which water penetration and polymer chain relaxation rate are similar, causing drug non-Fickian transport. The relaxational contribution is considerable in the case of curve at r = 8, leading to an almost non-Fickian transport. These results confirm that a Fickian transport cannot be used in the case of drug release from CH-based systems modelling.

Chapter VIII

126

VII.2.2.5 Conclusions on Chitosan-Ampicillin coprecipitation

Coprecipitation of CH, as carrier, and Amp, as model drug, has been successfully obtained using the SAA technique. Well-defined spherical microparticles, with a uniform morphology and a narrow distribution have been precipitated.

X-ray, DSC and EDX and drug release analysis confirmed that the drug is entrapped in an amorphous solid state into the polymeric matrix, thus forming a solid solution of the two components. Moreover, CH avoids the degradation of Amp.

A prolonged release is obtained from SAA coprecipitates with respect to raw drug and physical mixtures of CH and Amp, and the polymer/drug ratio has revealed to be a controlling parameter for drug release; in particular, it allows the modulation of drug release rate from coprecipitates.

CH/Amp systems has shown a swelling-controlled behavior, with drug release kinetics depending on front movements inside the tablet. Results show that drug release depends on both relaxation and diffusive mechanisms. Indeed, the empirical binomial equation proposed by Peppas and Sahlin (1989), well describe Amp release rate if both the relaxational and the diffusional parameters are function of the polymer/drug ratio.

In conclusion, SAA is successful in formulating CH/Amp microparticles for oral administration: Amp can be protected from degradation during the passage through the gastrointestinal tract, with an enhanced uptake by the epithelium due to the mucoadhesive properties of CH, and with a controlled release rate allowing a prolonged Amp blood concentrations.

Chapter VIII CONCLUSIONS AND FUTURE

DEVELOPMENTS

The aim of this thesis was to broaden the field of SAA applications, demonstrating that this supercritical based technique can be applied not only to particle size reduction of pure compounds, but also to the production of composite microparticles for drug controlled release.

This work was developed in two main parts: an initial investigation about the SAA processability of different kinds of carriers for drug release, followed by coprecipitation tests with model drugs.

Synthetic (PMMA and PLLA) and natural (chitosan) polymers and cyclodextrins were successfully processed, obtaining well-defined spherical microparticles at optimal process conditions, for all individual compounds. Narrow particle size distributions resulted, showing a dependence on process parameters, such as the solute concentration or the mass flow ratio. Results showed that the precipitation temperature is a key parameter for polymer processability: successful micronization is possible if the temperature in the precipitator is far lower than the polymer glass transition temperature, but high enough to allow a fast solvent evaporation.

Medroxyprogesterone acetate and Ampicillin trihydrate were selected as model drugs for the first attempts of SAA coprecipitation, and they were processed with PMMA and chitosan, respectively.

In the case of PMMA/MEPA coprecipitation, satisfying results were obtained: spherical and well defined microparticles with a uniform morphology were produced. Analyses confirmed MEPA entrapment into the PMMA matrix, i.e. a solid solution of the two components, leading to a prolonged release of the drug.

SAA CH/Amp coprecipitation was successful as well, in this case also producing spherical microparticles ranging between 0.1 and 6 µm. Furthermore, analysis confirmed that the Amp is entrapped and homogeneously dispersed, in an amorphous solid state, into the CH matrix and thermal stabilization of Amp shielded by the polymer results.

Chapter VIII

128

Wider experimentation was performed on the couple CH-Amp, investigating not only their SAA processability, but also the effect of the ratio between the polymer and the drug; moreover, an interpretation of CH/Amp system behaviour and of drug release kinetics was provided.

Drug release analyses showed a prolonged release from SAA coprecipitates with respect to the raw drug and physical mixtures of CH and Amp. In addition, the polymer/drug ratio is clearly a controlling parameter for drug release, allowing the modulation of drug release rates from coprecipitates.

A drug release mechanism characteristic of swelling-controlled systems has been observed for CH/Amp, in which both relaxation and diffusive mechanisms have a significant contribution. As a first attempt of modelling Amp release rates from a CH matrix, an empirical binomial equation in which two parameters are functions of the polymer/drug ratio was used, showing fairly good agreement with experimental data. In future developments of this work, mathematical models could be developed in order to design target-oriented drug delivery systems.

The results obtained with PMMA/MEPA and CH/Amp systems have also produced interesting information concerning the hypothesized mechanism of SAA. Indeed, interactions between phases in the saturator could eventually cause the failure of the experiment, since complex quaternary systems are formed in the case of coprecipitates and no previous results about them are available. However, no process problems arose in any coprecipitation test, indicating that even complex systems can be successfully treated by SAA. Moreover, the interactions in the quaternary systems can also result in behaviours favourable to SAA processing. For example, in the case of PMMA/MEPA, no antisolvent effect of acetone occurred, unlike the PMMA/acetone/CO2 ternary system. In the future, interesting information about these complex systems could be obtained by using a view cell, in which saturator conditions could be reproduced, thus allowing the visualization of the phases produced at operating pressure and temperature conditions.

The future direction of bioadhesive drug delivery systems, lies in the formulation of vaccines that adhere to the mucosal surface and result in mucosal immunity. This is especially relevant in nasal and oral vaccination programmes. Therefore, an interesting investigation in the future could regard drug release rates directly obtained from microparticles, immobilizing them onto a polymeric inert support reproducing nasal mucosa. The results of this experimentation would undoubtedly demonstrate the efficiency and the performance of the SAA technique in the production of composite polymer/drug microparticles.

REFERENCES

Agnihotri S.A., Mallikarjuna N.N., Aminabhavi T.M. (2004) Recent advances on chitosan-based micro- and nanoparticles in drug delivery. J. Contr. Rel., 100, 5-28.

Akbuga J., Durmaz G. (1994) Preparation and evaluation of crosslinked chitosan microspheres containing furosemide. Int. J. Pharm., 111, 217–222.

Akbuga, J. (1993) Effect of the physicochemical properties of a drug on its release from chitosan malate matrix tablets. Int. J. Pharm., 100, 257–261.

Alfrey, J. T., Gurnee, E. F., Lloyd, W. G. (1966) Diffusion in glassy polymers. Journal of Polymer Science, Part C: Polymer Symposia, 12, 249-61.

Arshady, R. (1991) Preparation of biodegradable microspheres and microcapsules: 2. Polylactides and related polyesters, J. Control. Rel., 17, 1-22.

Benoit, J.P., Rolland, H., Thies, C., Vande Velde V. (2000) Method of coating particles and coated spherical particles, US Patent 6087003.

Bertucco, A. (1999) Precipitation and crystallization techniques. In: Chemical synthesis using supercritical fluids (Jessop P.G., Leitner W., editors). Weinheim: Wiley-VCH, p. 108–26.

Bettini, R., Colombo, P., Massimo, G., Catellani, P.L., Vitali, T. (1994) Swelling and drug release in hydrogel matrixes: polymer viscosity and matrix porosity effects. Eur. J. Pharm. Sci., 2, 213-219.

Bettini, R., Peppas, N.A., Colombo, P. (1998) Polymer relaxation in swellable matrices contributes to drug release. Proc. of the Int. Symp. Control Rel. Bioact. Mater., 25, p. 26-37.

130

Bistrat, M., Moshashaee, S. (2002) Budesonide particles and pharmaceutical compositions containing them, US Patent 6346523.

Blanco, J., Vila-Jato, J. L., Otero, F., Anguiano, S. (1991) Influence of method of preparation on inclusion complexes of naproxen with different cyclodextrins, Drug Dev. Ind. Pharm., 17, 943-957.

Bleich, J., Kleinebudde, P., Muller, B. W. (1994) Influence of Gas Density and Pressure on Micropaticles Produced with the ASES Process, Int. J. Pharmaceutics, 106, 77-84.

Bodmeier, R., McGinity, J.W. (1988) Poly(lactic acid) microcapsules containing quinidine base and quinidine sulfate prepared by solvent evaporation technique. III. Morphology of the microspheres during dissolution studies, J. Microencapsul., 5, 325-330.

Bortner, M. J., Baird, D. G. (2004) Absorption of CO2 and subsequent viscosity reduction of an acrylonitrile copolymer, Polymer, 45, 3399-3412.

Brannon-Peppas, L., (1995) Recent advances on the use of biodegradable microparticles and nanoparticles in the controlled drug delivery, Int. J. Pharm., 116, 1– 9.

Brazel, CS, Peppas, NA. (1999) Mechanisms of solute and drug transport in relaxing, swellable, hydrophilic glassy polymers. Polymer,40, 3383–3398.

Briones J.A., Mullins J.C., Thies M.C, Kim B.-U. (1987) Ternary phase equilibria for acetic acid-water mixtures with supercritical carbon dioxide, Fluid Phase Equil., 36, 235-246.

Bulmus V., Ayhan H., Piskin E., (1997) Modified PMMA monosize microbeads for glucose oxidase immobilization. Chem. Eng. J., 65, 71-76.

Cavalier, M., Benoit, J.P., Thies, C. (1986) The formation and characterization of hydrocortisone-loaded poly((±)-lactide) microspheres, J. Pharm. Pharmacol., 38, 249-253.

Ceschel G.C., Badiello R., Ronchi C., Maffei P. (2003) Degradation of components in drug formulations: a comparison between HPLC and DSC methods. J. of Pharm. & Biomed. Analysis, 32, 1067-1072.

Chang C.J., Day C.-Y., Ko C.-M., Chiu K.-L. (1997) Densities and P-x-y diagrams for carbon dioxide dissolution in methanol, ethanol, and acetone mixtures. Fluid Phase Equil., 131, 243-258.

131

Chiou, J. S., Barlow, J. W., Paul, D. R. (1985) Polymer crystallization induced by sorption of CO2 gas. J. Appl. Polym. Sci., 30, 3911-3924.

Chiou, J.S., Barlow, J.W. and Paul, D.R. (1985). Plasticization of Glassy Polymers by CO2 , J. Appl. Polym. Sci., 30, 2633–2642.

Colombo, P., Bettini, R., Santi, P., De Ascentiis, A., Peppas, N. A. (1996) Analysis of the swelling and release mechanisms from drug delivery systems with emphasis on drug solubility and water transport. J. Control. Rel., 39, 231-237.

Colombo, P., Bettini, R., Massimo, G., Castellani, P.L., Santi, P., Peppas, N. A (1995) Drug diffusion front movements is important in drug release control from swellable matrix tablets. J. Pharm. Sci., 84, 991-997.

Colombo, P., Bettini, R., Santi, P., Peppas, N.A. (2000a) Swellable matrices for controlled drug delivery: gel layer behaviour, mechanisms and optimal performance. Pharm. Sci. Technol. Today, 3, 198-204.

Colombo, P., Catellani, P. L., Peppas, N. A., Maggi, L., Conte, U. (1992) Swelling characteristics of hydrophilic matrixes for controlled release. New dimensionless number to describe the swelling and release behavior. Int. J. Pharm., 88, 99-109.

Colombo P, Santi P, Bettini R, Brazel CS, Peppas NA. (2000b) Drug release from swelling-controlled systems. In Handbook of pharmaceutical controlled release technology (D.L. Wise, L. Brannon-Peppas, A.M. Klibanov, R.S. Langer, A.G. Mikos, N.A. Peppas, D.J. Trantolo, G.R. Wnek, M.J. Yaszemski Eds). New York: Dekker, p. 183–209.

Colombo P., Santi P., Bettini R., Brazel C., Peppas N. (2004) Drug release from swelling controlled system. In Handbook of Controlled Release, Wiley Science Eds.

Connors, K. A., Mollica, J. A. (1966) Theoretical analysis of comparative studies of complex formation. J. Pharm. Sci., 55, 772-780.

Couvreur P., Grislain L., Lenaerts V., Brasseur F., Guiot P. (1986) In Polymeric Nanoparticles and Microspheres (P. Guiot, P. Couvreur Eds), CRC Press, Boca Raton, FL.

Cristini, F., Delalonde, M., Joussot-Dubien, C. Bataille, B. (2003) Elaboration of ibuprofen microcomposites in supercritical CO2. Proc. of the 6th International Symposium on Supercritical Fluids

132

(Brunner G., Kikic I., Perrut M., Eds.), Versailles (France), p. 1917-1922.

Davidson, G.W.R., Peppas, N.A. Solute and penetrant diffusion in swellable polymers. V. (1986) Relaxation-controlled transport in P(HEMA-co-MMA) copolymers. J. Control. Rel., 3, 243-258.

Dulieu, C., Richard, J., Benoit, J.P. (2001) Prolonged release microspheres for injection delivery and preparation method, WO Patent 8948.

Dubernet C. (1995) Thermoanalysis of microspheres. Thermochimica Acta, 248, 259-269.

Edelman, R., Russell, R.G., Losonsky, G., Tall, B.D., Tacket, C.O., Levine, M.M., Lewis, D.H. (1993) Immunization of rabbits with enterotoxigenic E. coli colonization factor antigen (CFA/I) encapsulated in biodegradable microspheres of poly(lactide-coglycolide), Vaccine, 11, 155-118.

El-Gibaly I. (2002) Development and in vitro evaluation of novel floating chitosan microcapsules for oral use: comparison with non-floating chitosan microspheres. Int. J. Pharm., 249, 7-21.

Elvassore, N., Bertucco, A., Caliceti, P. (2001) Production of insulin loaded poly(ethylene glycol)/poly(L-lactide) (PEG/PLA) nanoparticles by gas antisolvent techniques. J. Pharm. Sci., 90, 1628-1636.

Fabing, I., Leboeuf, F. Jung, J., Perrut, M. (2002) Method and apparatus for making very fine particles containing a complex of an active principle and a cyclodextrin, WO Patent 02/32462, EP 1 330 266.

Felt, O., Buri, P., Gurny, R. (1998) Chitosan: a unique polysaccharide for drug delivery. Drug Dev. Ind. Pharm., 24, 979– 993.

Fernandes, C. M., Veiga, F. J. B. (2002) The Cyclodextrins as Modelling Agents of Drug Controlled Release, J. Inclusion Phenom. Macro. Chem., 44, 79-85.

Fisher, W., Muller, W. (1988) Method and appararatus for the manufacture of a product having a substance embedded in a carrier, EP Patent 0322687.

Fixe, F., Dufva, M., Telleman, P., Christensen, C. B. V. (2004) Functionalization of poly(methyl methacrylate) (PMMA) as a substrate for DNA microarrays, Nucleic Acids Res., 32, 1.

133

Flick-Smith, H.C., Eyles, J.E., Hebdon, R., Waters, E.L., Beedham, R.J., Stagg T.J., Miller J., Alpar H.O., Baillie L.W.J., Williamson E.D. (2002) Mucosal or Parenteral Administration of Microsphere-Associated Bacillus anthracis Protective Antigen Protects against Anthrax Infection in Mice. Infection and Immunity, 70, 2022-2028.

Fong, J. W (1988) Microencapsulation by solvent evaporation and organic phase separation process. In Controlled release systems: fabrication technology, vol. 1 (Hsieh D. Eds), Boca Raton, FL: CRC Press, p. 81-108.

Foster, N., Mammucari, R., Dehghani, F. (2004) Dense gas Antisolvent Processing of Cyclodextrins from Aqueous Solutions, Proc. of the 9th Meeting on Supercritical Fluids, Trieste, Italy.

Foster, N.R., Dehghani, F., Bustami, R.T., Chan, H.K. (2003) Generation of lysozyme–lactose powders using the ASES process. In Proc. of the 6th International Symposium on supercritical fluids (Brunner G., Kikic I., Perrut M. Eds), Versailles, France.

Genta, I., Pavanetto, F., Conti, B., Giunchedi, P., Conte, U. (1994) Spray-drying for the preparation of chitosan microspheres. Proc. of the Int. Symp. Control. Rel. Bioact. Mater., 21, 616– 617.

Ghaderi, R., Artursson, P., Carlfors, J. (2000) A new method for preparing biodegradable microparticles and entrapment of hydrocortisone in DL-PLG microparticles using supercritical fluids. Eur. J. Pharm. Sci., 10, 1–9.

Gong, K., Darr, J. A, Rehman, I. U., Anwar, J. (2005) Chitosan-Indomethacine Composites via SCF Assisted Impregnation. In International Symposium on Supercritical Fluids, Orlando, Florida #390.pdf .

Gonzalez, A., Tufeu, R., Subra, P. (2002) High-Pressure Vapor-Liquid Equilibrium for the Binary Systems Carbon Dioxide + Dimethyl Sulfoxide and Carbon Dioxide + Dichloromethane. J. Chem. Eng. Data, 47, 492-495.

Han J., Sangeeta G., Surynarayanan R. (1998) Applications of pressure differential scanning calorimetry in the study of pharmaceutical hydrates. II. Ampicillin trihydrate. Int. J. Pharm., 170, 63-72.

Harish Prashanth, K.V., Kittur, F.S., Tharanathan, R.N. (2002) Solid state structure of chitosan prepared under different N-deacetylating conditions. Carbohydr. Polym, 50, 27-33.

134

Harland, R. S., Gazzaniga, A., Sangalli, M. E., Colombo P., Peppas N. A. (1988) Drug/polymer matrix swelling and dissolution. Pharm. Res., 5, 488-494.

He, P., Davis, S.S., Illum, L. (1998) In vitro evaluation of the mucoadhesive properties of chitosan microspheres. Int. J. Pharm., 166, 75–88.

Hedges, A. R. (1998) Industrial Applications of Cyclodextrins. Chem. Rev., 98, 2035-2044.

Hénon, F.E., Carbonell, R.G., DeSimone, J.M. (2002) Effect of polymer coatings from CO2 on water-vapor transport in porous media. AIChE J., 48, 941–952.

Henry, A.C., McCarley R.L. (2001) Selective deposition of metals on plastics used in the construction of microanalytical devices: photo-directed formation of metal features on PMMA. J. Phys. Chem. B, 105, 8755-8761.

Henry, A.C., Tutt, T.J., Galloway, M., Davidson, Y.Y., McWhorter, C.S., Soper S.A., McCarley R.L. (2000) Surface modification of poly(methyl methacrylate) used in the fabrication of microanalytical devices. Anal. Chem., 72, 5331-5337.

Higuchi, T., Connors, K. A. (1965) Phase solubility techniques. Adv. Anal. Chem. Instrum., 4, 117-212.

Hirayama, F., Usami, M., Rimura, K., Uekama, K. (1997) Crystallization and Polymorphic Transition Behavior of Chloramphenicol Palmitate in 2-Hydroxypropyl-β-cyclodextrin Matrix. Eur. J. Pharm. Sci., 5: 23-30.

Hirayama, F., Zaoh, K., Harata, K., Saenger, W., Uekama, K. (2001) Transparent, Adhesive Film Formation of Per-O-valeryl-β-cyclodextrin. Chem Lett., 636-637.

Hopfenberg, H.B. and Hsu, K.C. (1978) Swelling-controlled, constant rate delivery systems. Polym. Eng. Sci., 18, 1186-1191.

Huimin, L. (1999) Science and engineering of the droplets. In Fundamentals and application (William Andrew Publishing) pp. 47-50.

Illum, L., (1998) Chitosan and its use as a pharmaceutical excipient. Pharm. Res. 15, 1326–1331.

Iwata, M., McGinity, J.W. (1992) Preparation of multi-phase micro-spheres of poly(d,l-lactic acid) and poly(d,l-lactic-co-glycolic acid)

135

containing a W/O emulsion by a multiple emulsion solvent evaporation technique. J. Microencapsulation, 9, 201–214.

Jain, R. A. (2000) The manufacturing techniques of various drug loaded biodegradable poly(lactide-co-glicolide) (PLGA) devices. Biomaterials, 21, 2475-2490.

Jalil, R., Nixon, J. R. (1990) Biodegradable poly(lactic acid) and poly(lactide-co-glycolide) microcapsules: problems associated with preparative techniques and release properties, J. Microencapsulation, 7, 297-325.

Jaworska, M., Sakurai, K., Gaudon, P., Guibal, E. (2003) Influence of chitosan characteristics on polymer properties. I: Crystallographic properties. Polym. Int., 52, 198–205.

Junco, S., Casimiro, T., Ribeiro, N., Da Ponte, M.N., Marques, H.C. (2002a) Optimization of supercritical carbon dioxide systems for complexation of naproxen: Beta-cyclodextrin. J. Incl. Phenom. Macro. Chem., 44, 69-73.

Junco, S., Casimiro, T., Ribeiro, N., Da Ponte, M.N., Marques, H.C. (2002b) A comparative study of naproxen-beta cyclodextrin complexes prepared by conventional methods and using supercritical carbon dioxide. J. Incl. Phenom. Macro. Chem. 44, 117-121.

Jung, J., Perrut, M. (2001) Particle design using supercritical fluids: literature and patent survey. J. Supercrit. Fluids, 20, 179–219.

Kacurakova, M., Belton, P.S., Hirsch, J., Ebringerova, A. (1998) Hydration properties of xylan-type structures: an FTIR study of xylooligosaccharides. J. Sci. Food Agric., 77, 38-44.

Kerc, J., Srcic, S., Knez, Z., Sencar-Bozic, P. (1999) Micronization of drugs using supercritical carbon dioxide. Int. J. Pharmaceutics, 182, 33-39.

Khan, K.A. (1975) The concept of dissolution efficiency. J. Pharm. Pharmacol., 27, 48-49.

Kimura, K., Hirayama, F., Arima, H., Uekama, K. (1999) Solid-state 13C Nuclear Magnetic Resonance Spectroscopic Study on Amorphous Solid Complexes of Tolbutamide with 2-Hydroxypropyl-α and β-Cyclodextrins. Pharm. Res., 16, 1729-1734.

136

Kinnarinen, T., Jarho, P., Jarvinen, K., Jarvinen, T. (2003) Pulmonary deposition of a budesonide/γ-cyclodextrin complex in vitro. J. Control. Rel., 90, 197–205.

Kittur, F.S., Harish Prashanth, K.V., Udaya Sankar, K., Tharanathan, R.N. (2002) Characterization of chitin, chitosan and their carboxymethyl derivatives by differential scanning calorimetry. Carbohydr. Polym., 49, 185-193.

Köse, G.T., Tezcaner, A., Hasirci, V. (2002) Fundamentals of tissue engineering: Carrier materials and an application. Techn. Health Care, 10, 187-201.

Kristl, J., Smidx-Korbar, J., Struc, E., Schara, M., Rupprecht, H. (1993) Hydrocolloids and gels of chitosan as drug carriers. Int. J. Pharm., 99, 13–19.

Kristmundsdottir, T., Ingvarsdottir, K., Saemundsdottir, G. (1995) Chitosan matrix tablets: the influence of excipients on drug release. Drug Dev. Ind. Pharm., 21, 1591–1598.

Langer, R., Peppas, N.A. (2003) Advances in biomaterials, drug delivery, and bionanotechnology. AIChE, 49, 2990–3006.

Lavernia, E. J., Wu, Y. (1996) In Spray Atomization and Deposition, John Wiley and Sons Eds.

Le, M., Dehghani, F., Poole-Warren, L., Foster, N. (2005) A novel Method of fabrication of porous chitosan for biomedical or pharmaceutical applications. In International Symposium on Supercritical Fluids, Orlando, Florida, #188.pdf .

Lee, H.K., Park, J.H., Kwon, K.C. (1996) Double walled microparticles for HBV single shot vaccine, Proc. of the Int. Symp. Control. Rel. Bioact. Mater., 23, p. 333-334.

Lee, P.I. (1981) Controlled drug release from polymeric matrixes involving moving boundaries., Proc. of the 7th Controlled Release Pestic. Pharm., Plenum Publishing, New York, pp. 39-48.

Lee, P.I. and Kim C. (1991) Probing the mechanisms of drug release from hydrogels. J. Control. Rel., 16, 229-236.

Lee, P.I., Peppas, N.A. (1987) Prediction of polymer dissolution in swellable controlled-release systems. J. Control. Rel., 6, 207-15.

137

Lehr, C.M., Bouwstra, J.A., Schacht, E.H., Junginger, H.E. (1992) In vitro evaluation of mucoadhesive properties of chitosan and some other natural polymers. Int. J. Pharm., 78, 43–48.

Lewis, D.H. (1990) Controlled release of bioactive agents from lactide/glycolide polymers. In Biodegradable polymers as drug delivery systems (M. Chasin and R. Langer Eds), New York: Marcel Dekker, p. 1-41.

Liggins, R.T., Burt, H.M. (2004) Paclitaxel loaded poly(L-lactic acid) (PLLA) microspheres. II. The effect of processing parameters on microsphere morphology and drug release kinetics. Int. J. Pharm., 281, 103.

Liu, C-L., Wu S-M., Chang T-C., Liu M-L., Chiang H-J. (2004) Analysis of a mixture containing ampicillin anhydrate and ampicillin trihydrate by thermogravimetry, oven heating, differential scanning calorimetry, and X-ray diffraction techniques. Anal. Chim. Acta, 517, 237–243.

Lockemann, C., Luddecke, E., Horn, D. (2002) Process and apparatus for producing stably fine-particle powders, US Patent 6375873.

Lorenzo-Lamosa, M.L., Remunan-Lopez, C., Vila-Jato, J.L., Alonso, M.J. (1998) Design of microencapsulated chitosan microspheres for colonic drug delivery. J. Control. Rel., 52, 109-118.

Lowman, A.M., Dziubla, T.D., Bures, P., Peppas, N.A. (2004) Structural and dynamic response of neutral and intelligent networks in biomedical environments. In Molecular and cellular foundations of biomaterials (N.A. Peppas, M.V. Sefton Eds) New York: Academic Press., p. 75–130.

Luisi, P.L., Giomini, M., Pileni, M.P, Robinson, B.H. (1988) Riverse micelles as hosts for proteins and small molecules. Biochim. Biophys. Acta, 947, 209- 246.

Mammucari, R., Dehghani, F., Foster, N. (2002) Coprecipitation of pharmaceuticals using gas antisolvent technique, Proc. of the 8th meeting on Supercritical Fluids (Besnard M., Cansell F., Edts), Bordeaux, France, Tome 1, p. 321-326.

Mandal, T.K., Shakleton, M., Trinh, K.B., Le, T.N. (1995) Application of a novel phase separation technique for the encapsulation of diltiazem and metoprolol, Proc. of the Int. Symp. Control. Rel. Bioact. Mater., 22, p. 772-773.

138

Mandel, F., Wang, J.D., Howdle, S.M., Shakesheff, K.M. (2002) Controlled release pharmaceuticals prepared by supercritical fluid processing techniques, WO Patent 02/20624.

Marongiu, B., Piras, A., Porcedda, S., Delogu, G., Fabbri, D., Dettori, M.A. (2002) Cyclodextrin inclusion compounds by supercritical carbon dioxide, Proc. of the 8th meeting on Supercritical Fluids (Besnard M., Cansell F., Eds), Bordeaux, France, Tome 1, p. 327-332.

Matson, D.W., Fulton, J.L., Petersen, R.C., Smith, R.D. (1987) Rapid expansion of supercritical fluid solution: solute formation of powders. Thin films and fibers. Ind. Eng. Chem. Res., 26, 2298-2306.

Matsuoka, S., Kwei, T.K. (1979) In Physical Behaviour of Macromolecules (F.A. Bowey and F.H. Winslow Eds.). Academic Press, New York. p. 355.

Miller, J.M., Kale, U.J., Kelvin Lau, S.M., Greene, L., Wang, H.Y. (2004) Rapid estimation of kinetic parameters for thermal decomposition of pennicillins by modulated thermogravimetric analysis. J. Pharm. Biomed. Anal., 35, 65-73.

Miyazaki, S., Ishii, K., Nadai, T. (1981) Use of chitin and chitosan as drug carriers. Chem. Pharm. Bull., 29, 3067-3069.

Miyazaki, T., Komuro, T., Yomota, C., Okada, S. (1990) Use of chitosan as a pharmaceutical material effectiveness as an additional additive of sodium alginate. Eisei Shikenjo Hokoku, 108, 95–97.

Mosharraf, M., Nystrom, C. (1995) The effect of particle size and shape on the surface specific dissolution rate of micronized practically insoluble drugs, Int. J. Pharmaceutics, 122, 35-47.

Moyano, J.R., Ginés, M.J., Rabasco, A.M. (1995) Study of the dissolution characteristics of oxazepam via complexation with b-cyclodextrin. Int. J. Pharm., 114, 95-102.

Mustafin, R.I., Protasova, A.A., Van den Mooter, G., Kemenova, V. A. (2005) Diffusion transport in interpolyelectrolyte matrix systems based on chitosan and Eudragit l100. Pharm. Chem. Journal, 39, 663-666.

No, H.K., Meyers, S.P. (1989) Crawfish Chitosan as a Coagulant in Recovery of Organic Compounds from Seafood Processing Streams. J. Agric. Food Chem., 37, 580-583.

139

O’Donnell, P. B., McGinity, J.W. (1997) Preparation of microspheres by the solvent evaporation technique. Adv. Drug Deliv. Rev., 28, 25-42.

O’Donnell, P.B., Iwata, M., McGinity, J.W. (1995) Properties of Multiphase Microspheres of Poly(D,L-lactic-co-glycolic acid) Prepared by a Potentiometric Dispersion Technique. J. Microencapsulation, 12, 155–163.

Ohe, S. (1990) In Vapour-liquid equilibrium data at high pressure. Elsevier, Amsterdam, , p. 436.

Pallado, P., Benedetti, L., Callegaro, L. (2001) Nanospheres comprising a biocompatible polysaccharide, US Patent 6214384.

Peppas, N.A., Gurny, R., Doelker, E., Buri, P. (1980) Modelling of drug diffusion through swellable polymeric systems. J. Membr. Sci, 7, 241-53.

Peppas, N.A., Sahlin, J.J. (1989) A simple equation for the description of solute release. III. Coupling of diffusion and relaxation. Int. J. Pharm., 57, 169-172.

Perrut, M. (2001) Method for collecting and encapsulating fine particles, WO Patent 49407.

Perrut, M., Jung, J., Leboeuf, F. (2004) Enhancement of dissolution rate of poorly-soluble active ingredients by Supercritical Fluid processes. Part II: Preparation of composite particles. Proc. of the 9th Meeting on Supercritical Fluids. Trieste, Italy

Phillips, G.O., Takigami, S., Takigami, M. (1996) Hydration characteristics of the gum exudate from Acacia Senegal. Food Hydrocoll., 10, 11-19.

Reverchon, E., Della Porta, G.,Taddeo. R. (1995) Solubility and micronization of griseofulvin in supercritical CHF3. Ind. Eng. Chem. Res., 34, 4087-4091.

Reverchon, E. (1999) Supercritical antisolvent precipitation of microand nano-particles. J. Supercrit. Fluids, 15, 1-21.

Reverchon, E., Della Porta, G. (1999) Production of antibiotic micro- and nano-particles by supercritical antisolvent precipitation. Powder Technol., 106, 23-29.

140

Reverchon, E. (2002a) Supercritical assisted atomization to produce micro and/or nano particles of controlled size and distribution. Ind. Eng. Chem. Res., 41, 2405-2411.

Reverchon, E. (2002b) Micro and Nano particles Produced by Supercritical Fluids Assisted Techniques: Present Status and Future Developments. Proc. of the 4th International Symposium on High Pressure Process Technology and Chemical Engineering, Venice, Italy.

Reverchon, E., Della Porta, G., Spada, A. (2002) Micronization of Rifampicin by Supercritical Assisted Atomization. Proc. of the 4th Internat. Symp. on High Pressure Process Technology and Chemical Engineering, Venice, Italy.

Reverchon, E. (2003) Process for production of micro and/or nano particles, WO Patent 03004142.

Reverchon, E, Della Porta, G. (2003a) Micronization of some antibiotics by supercritical assisted atomization. J. Supercrit. Fluids, 26, 243-252.

Reverchon, E., Della Porta, G. (2003b) Terbutaline microparticles suitable for aerosol delivery produced by supercritical assisted atomization. Int. J. Pharm., 258, 1-9.

Reverchon, E., Della Porta, G. (2003c) Particle Design Using Supercritical Fluids. Chem. Eng. Technol., 26, 840-845.

Reverchon, E., Della Porta, G., Spada, A. (2003a) Ampicillin micronization by supercritical assisted atomization. J. Pharm. Pharmacol., 55, 1465-1471.

Reverchon, E., Volpe, M. C., Caputo, G. (2003b) Supercritical fluid processing of polymers: composite particles and porous materials elaboration. Curr. Opin. Solid State Mater. Sci., 7, 391-397.

Reverchon, E., Della Porta, G., Spada, A., Antonacci, A. (2004) Griseofulvin micronization and dissolution rate improvement by supercritical assisted atomization, J. Pharm. Pharmacol., 56, 1379-1387.

Reverchon, E., Spada, A. (2004) Erytromycin micro-particles produced by supercritical fluid atomization, Powder Tech., 141, 100-108.

Rhodes, M. (1999) In Introduction to particle technology (Wiley&Sons Eds), New York.

141

Ribero Dos Santos, I., Richard, J., Pech, B., Thies, C., Benoit, J.P. (2002) Microencapsulation of protein particles within lipids using a novel supercritical fluid process. Int. J. Pharm., 242, 69–78.

Richard, J., Dulieu, C., Le Meurlay, D., Benoit, J.P. (2001) Microparticles for pulmonary administration, WO Patent 12160.

Ritger, P.L., Peppas, N.A. (1987) A simple equation for description of solute release. II. Fickian and non-Fickian release from swellable devices. J. Control. Rel., 5, 37-42.

Salmon S. and Hudson S.M. (1995) Shear-precipitated chitosan powders, fibrids, and fibrid papers: Observations on their formation and characterization. J. Polym. Sci. B, Polym. Phys., 33, 1007–1014.

Sellers, S.P., Clark, G.S., Sievers, R.E., Carpenter, J.F. (2001) Dry powders of stable protein formulations from aqueous solutions prepared using supercritical CO2-assisted aerosolization. J. Pharm. Sci., 90, 785-797.

Sencar-Bozic, P., Srcic, S., Knez, Z., Kerc, J. (1997) Improvement of nifedipine dissolution characteristics using supercritical CO2. Int. J. Pharm., 148, 123-130.

Serra, L., Domenech, J., Peppas, N.A. (2006) Drug transport mechanisms and release kinetics from molecularly designed poly(acrylic acid-g-ethylene glycol) hydrogels. Biomaterials, 27, 5440–5451.

Shefter, E., Fung, H.L., Mok, O. (1973) Dehydratation of crystalline theophylline monohydrate and ampicillin trihydrate. J. Pharm. Sci., 62, 791-794.

Shekunov, B., Chattopadhyay, P., Seitzinger, J. (2003) Engineering of composite particles for drug delivery using supercritical fluid technology, PMSE, 89, 681–682.

Shimoda, J., Onishi, H., Machida, Y. (2001) Bioadhesive characteristics of chitosan microspheres to the mucosa of rat small intestine. Drug Dev. Ind. Pharm., 27, 567–576.

Shiraishi, S., Arahira, M., Imai, T., Otagiri, M. (1990) Enhancement of dissolution rates of several drugs by low-molecular weight chitosan and alginate. Chem. Pharm. Bull., 38, 185–187.

Siepmann, J., Peppas, N. A. (2000) Hydrophilic Matrices for Controlled Drug Delivery: An Improved Mathematical Model to Predict the

142

Resulting Drug Release Kinetics (the “Sequential Layer” Model). Pharm. Res., 17, 1290-1298.

Sievers, R.E. and Karst., U. (1997) Methods for fine particle formation. US Patent No. 5639441.

Sievers, R.E., Milewski, P.D., Sellers, S.P., Miles, B.A., Korte, B.J., Kusek, K.D., Clark, G.S., Mioskowski, B., Villa, J.A. (2000) Supercritical and Near-critical Carbon Dioxide Assisted Low-Temperature Bubble Drying. Ind. Eng. Chem. Res., 39, 4831-4836.

Singhal, R., Nagpal, A.K., Mathur, G. N. (1999) Study of Ethyl Cellulose-Benzoic Acid interactions in matrices for controlled drug release by DSC. J. Thermal Anal. Calorim., 58, 28-38.

Sinha, V.R., Singla, A.K., Wadhawan, S., Kaushik, R., Kumria, R., Bansal, K., Dhawan, S. (2004) Chitosan microspheres as a potential carrier for drugs. Int. J. Pharm., 274, 1 –33.

Subramaniam, M.R., Rajewski, R.A., Snavely, K. (1997) Pharmaceutical processing with supercritical carbon dioxide. J. Pharm. Sci., 86, 885-890.

Szejtli, J. (1998) Introduction and general overview of cyclodextrin chemistry. Chem. Rev., 98, 1743-1753.

Szejtli, J. (1994) Medicinal applications of cyclodextrins, Med. Res. Rev., 14, 353-386.

Szente, L., Szejtli, J. (2004) Cyclodextrins as food ingredients. Trends Food Sci. Tech., 15, 137-142.

Szente, L., Strattan, E. (1988) Hydroxypropyl β-cyclodextrins, preparation and physicochemical properties. In New Trends in cyclodextrins and derivatives, (D. Duchene Ed), Editions de Sante’, Paris.

Takada, S., Uda. Y., Toguchi, H., Ogawa, Y. (1995) Application of a spray drying technique in the production of TRH-containing injectable sustained-release microparticles of biodegradable polymers. PDA J. Pharm. Sci. Technol., 49, 180-184.

Takenouchi, S., Kennedy, G.C. (1964) The binary system H2O-CO2 at high temperatures and pressures. Am. J. Sci., 262, 1055-1074.

Tamber, H., Merkle, H.P., Steward, M.W., Partidos, C.D., Gander, B. (1996) In vitro release of a synthetic measles virus antigen from PLGA

143

microspheres. Proc. of the International Conference on Advances in Controlled Delivery, Baltimore, MD, pp. 129-130.

Tice, T.R., Tabibi, E.S. (1991) Parenteral drug delivery: injectables. In: Treatise on controlled drug delivery: fundamentals optimization, applications (A. Kydonieus Ed) New York, Marcel Dekker, pp. 315-39.

Tsai, D.C, Howard, S.A, Hogan, T.F, Malanga, C.J, Kandzari, S.J, Ma, J.K.H. (1986) Preparation and in vitro evaluation of polylactic acidmitomycin C microcapsules, J. Microencapsul., 3, 181-193.

Uekama, K., Hirayama, F., Irie, T. (1998) Cyclodextrin Drug Carrier Systems, Chem. Rev., 98, 2045-2076.

Urbanczyk, G.W., Lipp-Symonowicz, B. (1994) The influence of processing terms of chitosan membranes made of differently deacetylated chitin on the crystalline structure of membranes. J. Appl. Polym. Sci, 51, 2191-2194.

Vonderheiden, F.H., Eldridge, J.W. (1963) The System Carbon Dioxide. Methylene Chloride. Solubility, Vapor Pressure, Liquid Density, and Activity Coefficients. J. Chem. Eng. Data, 8, 20-21.

Wada, R., Hyon, S-H, Ike, O., Watanabe, S., Shimizu, Y., Ikada, Y. (1988) Preparation of lactic acid oligomer microspheres containing anti-cancer drug by o/o type solvent evaporation process. Polym. Mater. Sci. Eng., 59, 803-806.

Waddell, E., Wang, Y., Stryjewski, W., McWhorter, S., Henry, A.C., Evans, D., McCarley, R.L., Soper, S.A. (2000) High-resolution near-infrared imaging of DNA microarrays with time resolved acquisition of fluorescence lifetimes. Anal. Chem., 72, 5907-5918.

Wagenaar, B.W., Muller, B.W. (1994) Piroxicam release from spray-dried biodegradable microspheres. Biomaterials, 15, 49-54.

Wake, M.C., Gerecht, P.D., Lu, L., Mikos, A.G. (1998) Effects of biodegradable polymer particles on rat marrow-derived stromal osteoblasts in vitro. Biomaterials, 19, 1255-1268.

Wang, X., Michoel, A., Van den Mooter, G. (2004) Study of the phase behavior of polyethylene glycol 6000–itraconazole solid dispersions using DSC. Int. J. of Pharm., 272, 181-187.

Weidner, E., Knez, Z., Novak, Z. (1995) Process for preparing particles or powders. WO Patent 21688.

144

Williams, A. (2002) In Application of Supercritical Assisted Atomization to acetonic and ethanolic solutions. MSc thesis. University of Salerno, Italy.

Wissinger, R. G., Paulaitis, M.E. (1991) Glass transitions in polymer/CO2 mixtures at elevated pressures. J. Polym. Sci., 29, 631-633.

Wu, X.S. (1995) Preparation, characterization, and drug delivery applications of microspheres based on biodegradable lactic/glycolic acid polymers. In: Encyclopedic handbook of biomaterials and bioengineering (Wise et al. Eds). New York, Marcel Dekker, pp. 1151-200.

Wu, N., Wang, L-S., Tan, D.C-W., Moochhala, S.M., Yang, Y-Y. (2005) Mathematical modeling and in vitro study of controlled drug release via a highly swellable and dissoluble polymer matrix: polyethylene oxide with high molecular weights. J. Control. Rel., 102, 569–581.

Zhang, Y., Xue, C., Xue, Y., Gaoa, R., Zhang, X. (2005) Determination of the degree of deacetylation of chitin and chitosan by X-ray powder diffraction. Carbohydr. Res., 340, 1914–1917.

Zotkin, M.A., Vikhoreva, G.A., Smotrina, T.V., Derbenev, M.A. (2004) Thermal modification and study of the structure of chitosan films. Fiber Chem., 36, 16-20.

Yui, T., Imada, K., Okuyama, K., Obata, Y., Suzuki, K., Ogawa, K. (1994) Molecular and Crystal Structure of the Anhydrous Form of Chitosan. Macromolecules, 27, 7601–7605