A new supercritical fluid-based process to produce scaffolds for tissue replacement

Available online at www.sciencedirect.com

J. of Supercritical Fluids 45 (2008) 365–373

A new supercritical fluid-based process to produce scaffolds for tissue replacement E. Reverchon, S. Cardea ∗ , C. Rapuano Department of Chemical and Food Engineering, University of Salerno, Via Ponte Don Melillo 1, 84084 Fisciano, Italy Received 1 October 2007; received in revised form 19 December 2007; accepted 4 January 2008

Abstract Temporary scaffolds can be used as surrogates of the extracellular matrix to favour the regeneration of tissue and organs. Several techniques have been proposed for scaffolds fabrication; but they have various limitations. Above all, it is very difficult to obtain the coexistence of the macro and microstructural characteristics necessary for a successful application. To overcome these limitations, we tested a new supercritical fluid assisted technique for the formation of 3D scaffolds, that consists of three subprocesses: the formation of a polymeric gel loaded with a solid porogen, the drying of the gel using supercritical CO2 (SC-CO2 ), the washing with water to eliminate the porogen. We obtained poly(l-lactic acid) (PLLA) scaffolds with elevated porosity (>90%) and interconnectivity, with good mechanical properties (compressive modulus up to 81 kPa). The most innovative characteristic of these scaffolds is the fibrous nanostructure coupled to micronic cells of controllable size. We also produced scaffolds with predetermined shape and size in a short time (<30 h) and without an appreciable solvent residue (<5 ppm). © 2008 Elsevier B.V. All rights reserved. Keywords: Scaffolds; Supercritical fluids; Poly(l-lactic acid); Tissue engineering

1. Introduction Temporary scaffolds, built using synthetic or natural materials, can be used as surrogates of the extracellular matrix to favour the regeneration of tissue and organs, working as support for cells proliferation [1]. These materials are aimed at eliminating the problems due to donor scarcity, immune rejection and pathogens transfer. The temporary substitution of different biological materials requires different scaffold characteristics; but, they share a series of common peculiarities that have to be simultaneously obtained: • A high regular and reproducible 3D structure (macrostructure) similar to the one of the tissue to be substituted. • A porosity exceeding 90% and an open pore geometry that allows cells growth and reorganization; indeed, porosity and pore interconnectivity directly affect the diffusion of nutrients and the removal of metabolic wastes. • A suitable cell size depending on the specific tissue to be replaced. For example, in the regeneration of bone tissues, ∗

Corresponding author. Fax: +39 089964057. E-mail address: [email protected] (S. Cardea).

0896-8446/$ – see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.supflu.2008.01.005

cell sizes ranging between 200 and 400 ␮m and up to 500 ␮m are commonly indicated [1,2]. • High internal surface areas; the scaffold should be not only highly porous and interconnected but should present nanostructural surface characteristics that allow cell adhesion, proliferation and differentiation. The cells of different tissues such as bones, tendon, cartilage, cardiac valves and blood vessels, are submerged in a collagen matrix with a fibrous structure with diameters ranging between 50 and 500 nm. As a consequence a fibrous structure should be formed on the walls of the scaffold to allow cells growth [2]. • Mechanical properties to maintain the predesigned tissue structure; the correct values depend on the organs to be repaired; for example, a bone temporary scaffold should present a compressive modulus of the order of 100 kPa. • Biodegradability, biocompatibility and a proper degradation rate to match the rate of the neo-tissue formation; for this reason, it is very important the choice of the polymer because the biodegradability of the various polymers deeply varies from one polymer to the other. Many natural and synthetic polymers have been proposed for scaffolding applications. Poly(glycolic acid) (PGA), poly(lactic acid) (PLA) and their copolymers poly(lactic acid-co-glycolic

366

E. Reverchon et al. / J. of Supercritical Fluids 45 (2008) 365–373

acid) (PLGA) are a family of linear aliphatic polyesters that are the most frequently used in tissue engineering. These polymers degrade through hydrolysis of the ester bonds. PGA is the most widely used since it is relatively hydrophilic and degrades rapidly in aqueous solutions or in vivo. It loses mechanical integrity in 2–4 weeks. Other linear aliphatic polyesters, such as poly(␧-caprolactone) (PCL) and poly(hydroxy butyrate) (PHB) are used in tissue engineering research. Natural polymers such as proteins and polysaccharides are also used. For example Collagen is a fibrous protein and the major natural extracellular matrix component; it has been proposed for various tissue regeneration applications [2]. Polysaccharides such as alginate, chitosan and hyaluronate have also been suggested as porous solid-state tissue engineering scaffolds [2]. Several techniques have been proposed for scaffolds fabrication that include: fiber bonding, solvent casting, particulate leaching, melt moulding, solid free form fabrication, gas foaming and freeze drying combined with particulate leaching. Details about these process can be found in several reviews [2,3]. Conventional scaffolds fabrication techniques suffer various limitations; particularly, it is very difficult to obtain the coexistence of the macro and microstructural characteristics that have been previously described. For example, it is difficult to obtain: • Large porosity together with a control of the pore diameter, of the pore connectivity and of the mechanical resistance. • Large porosity together with the rough pore surface that allows cell adhesion. • Complex—3D structures; several techniques are limited to the production of thin disks or films. • The removal of toxic solvents that are located deep inside the structure. • The removal of solid porogens in structures with a large thickness.

• Formation of a polymeric gel loaded with a solid porogen. • Drying of the gel using SC-CO2 to form a supercritical mixture with the solvents. • Washing with water to eliminate the porogen. The scaffolds produced using this process should have the following advantages: • Accurate reproduction of the shape and 3D structure of the tissue to be substituted. • Controlled and large porosity >90%. • Large internal surface areas. This properties is conferred to the scaffold by the original fibrous nanometric substructure that is characteristic of several polymer gels [18–22]. The high porosity of the structure obtained should also aid the efficient removal of the porogen. • Very large connectivity at micronic and nanometric levels. • Short processing time. • Absence of solvent residues; indeed, SC-CO2 shows a very large affinity with almost all the organic solvents and is completely miscible with them even at very mild operating conditions (for example 100 bar, 40 ◦ C) [23]. Moreover, characteristic mass transfer properties of the supercritical fluids allow the almost complete removal of any solvent residue, as it has been previously demonstrated in other SC-CO2 -based processes [22]. In this way the potential disadvantages of the porogen-based techniques (the elimination of the organic solvents and the difficulties in the removal of the porogen in deep structures) can be easily overcome by the use of SC-CO2 . Therefore, the aim of this work is to study the feasibility and the influence of the process parameters on the formation of poly(l-lactic acid) (PLLA) scaffolds designed for bone replacement using an ethanol–dioxane PLLA gel loaded with d-fructose particles of selected diameter. Microscopic and mechanical properties of the produced scaffolds will be also characterized. 2. Materials and methods

Until now the use of supercritical CO2 (SC-CO2 ) in this field, has been limited to gas-foaming techniques in which it is used as a porogen to produce polymeric foams [4–6]. The process is solventless and very efficient in producing a porous structure; but, generally closed-cells structures are generated during this process. Indeed, though Mooney et al. [5] claimed that open porous structures could be obtained using this process; the same author [6] in a subsequent work adopted a CO2 -foaming plus solid porogen technique to overcome the limited cell connectivity of this technique. SC-CO2 has been also used as an alternative non-solvent in phase inversion processes to generate polymeric and biopolymeric porous structures [7–17]. In this case, the structures obtained present good interconnectivity and high porosity but, they do not posses the fibrous substructure on cell walls. In this work we propose the feasibility of a different supercritical fluid assisted technique for the formation of scaffolds, that consists of the following subprocesses:

Poly(l-lactic acid) with an inherent viscosity ranging between 2.6 and 3.2 dl/g (0.1% in chloroform, 25 ◦ C) was purchased from Boehringer Ingelheim (Ingelheim, Germany), d-fructose (mp 119–122 ◦ C), Dioxane and Ethanol (99.8% purity) were bought from Sigma–Aldrich (S. Louis, MO, USA); CO2 (99% purity) was purchased form SON (Societ`a Ossigeno Napoli—Italy). All materials were used as received. 2.1. Preparation of PLLA scaffolds In the first set of experiments, we produced some PLLA aerogels. They were prepared following one of the procedures adopted in literature [18]: we added 10% of polymer in dioxane using ethanol as the non-solvent with a dioxane/ethanol ratio of 1.7. The solution was heated at 60 ◦ C until it became homogeneous; than, it was poured in steel containers, obtaining samples with a diameter of 2 cm and height of 4 mm. Subsequently, it

E. Reverchon et al. / J. of Supercritical Fluids 45 (2008) 365–373

was placed into a freezer at −18 ◦ C for 1 h to obtain the gel. Then, the gel was dried using SC-CO2 [22]. The drying vessel was filled from the bottom with SC-CO2 up to the desired pressure using a high pressure pump (Milton Roy-Milroyal B, Pont-Saint-Pierre, France). The optimized supercritical CO2 extraction of the solvent from the polymeric gel lasted 4 h. A depressurization time of 10 min was used to bring back the system at atmospheric pressure. Ascertained the possibility of producing a dried gel, we added fructose into the starting solution. The sugar particles were sieved to select the desired size range (250–500, 500–700 ␮m); then, we prepared solutions with PLLA concentrations of 10%, 12%, 15% (w/w) in dioxane and added ethanol as the non-solvent with a dioxane/ethanol ratio of 1.7 to induce the gelation. After the supercritical drying, the obtained structures were poured in distilled water for 24 h to eliminate the fructose. Subsequently, the scaffolds were put in a oven for 4 h at 40 ◦ C, to evaporate water.

367

30 m length, 0.53 mm i.d., 1 ␮m film thickness and the second (model Cp Sil 5CB CHROMPACK, Stepbios, Italy) connected to the injector; 25 m length, 0.53 mm i.d., 5 ␮m film thickness. GC conditions were the one described in the USP 467 Pharmacopoeia with some minor modifications: oven temperature from 45 to 210 ◦ C for 15 min. The injector was maintained at 135 ◦ C (split mode, ratio 4:1), and Helium was used as the carrier gas (5 ml/min). Head space conditions were: equilibration time, 30 min at 95 ◦ C; pressurization time, 0.15 min; and loop fill time, 0.15 min. Head space samples were prepared in 20 ml vials filled with internal standard DMI (3 ml) and 500 mg of NaCl and water (0.75 ml) in which samples of PLLA scaffold were suspended. 3. Results and discussion

2.2. Characterizations

An aerogel is a polymeric structure characterized by a very large porosity and can be formed by a network of fibres with nanometric diameters. Two major problems in aerogel production are:

2.2.1. Scanning electron microscopy (SEM) Poly(l-lactic acid) scaffolds were cryofractured using liquid Nitrogen; then, the sample was sputter coated with gold (Agar Auto Sputter Coater mod. 108 A, Stansted, UK) at 30 mA for 180 s and was analyzed by a scanning electron microscope (SEM mod. LEO 420, Assing, Italy) to analyze cell and pore size and the overall scaffolds structure.

• to find the right formulation of the solvents and of the process to induce the “gelation”; • to eliminate the solvents from the structure avoiding its collapse: evaporation of the solvent produces the destruction of the nanostructure due to the force exerted by the surface tension of the evaporating liquid.

2.2.2. Scaffold porosity The porosity (ε) represents the “void space” of the scaffold and was calculated from the density of the scaffold ρS = (scaffold volume/scaffold weight) and the density of untreated PLLA ρP = 1.24 (g/cm3 ): ρS ε=1− . ρP The scaffold density was determined by measuring its volume and weight. 2.2.3. Mechanical tests We measured the compressive mechanical properties of the scaffolds using an INSTRON 4301 (Instron Int. Ltd., High Wycombe, UK). Cylindrical samples with a diameter of 2 cm and a thickness of 4 mm were compressed at a cross-head speed of 1 mm/min. The compressive modulus is defined as the initial linear modulus on the stress–strain curves. Seven specimens were tested for each sample. 2.2.4. Solvent residue analysis Dioxane residue was measured by a headspace (HS) sampler (model 7694E, Hewlett Packard, USA) coupled to a gas chromatograph (GC) interfaced with a flame ionization detector (GC–FID, model 6890 GC-SYSTEM, Hewlett Packard, USA). Dioxane was separated using two fused-silica capillary columns connected in series by press-fit: the first column (model Carbowax EASYSEP, Stepbios, Italy) connected to the detector,

Supercritical CO2 gel extraction is a new process that allows the drying of gels through the formation of a supercritical solution between CO2 and the liquid solvent. The supercritical mixture shows no surface tension and can be easily eliminated in a single step by the continuous flushing of SC-CO2 in the drying vessel [22]. Aerogel structures can possess some of the nanometric characteristics useful to produce scaffolds; but, large interconnected cavities are generally missing. In the first part of the work, we verified and optimized PLLA gel formation and drying. We decided to dry PLLA gels by SCCO2 at a pressure of 200 bar and a temperature of 33 ◦ C for 4 h. The operative temperature used is the lowest possible to maintain CO2 at supercritical conditions and was selected to avoid the crossing of PLLA glass transition (Tg ) temperature during the drying process. It has also to be take into account that the presence of SC-CO2 , as a rule, lowers the Tg of the polymers [24]. The operative pressure was selected to minimize the drying process time; indeed, increasing the pressure, the SCCO2 solvent power increases and, as a consequence, the solvent elimination is faster. We consistently obtained aerogels of the same size and shape of the starting gels: 2 cm of diameter and 4 mm of thickness, as shown in Fig. 1. We also successfully processed, by SC-CO2 drying, larger and more complex 3D gel structures (i.e., bone-shaped gels 8 cm long and 2 cm large) as shown in Fig. 2, that confirm the possibility of producing aerogels with a specific geometry without size limitations.

368

E. Reverchon et al. / J. of Supercritical Fluids 45 (2008) 365–373

Fig. 2. Picture of PLLA bone-shaped aerogel dried at 200 bar and 33 ◦ C.

Fig. 1. Pictures of PLLA aerogel dried at 200 bar and 33 ◦ C.

We analyzed by SEM the aerogels produced; examples of the electronic images of the section and of the surface of the PLLA aerogels obtained in this study are reported in Fig. 3. It is possible to observe, that, as expected, the internal structure is formed by nanometric fibers with a mean diameter lower than 200 nm and the surface is characterized by an highly porous structure. Moreover, the porosity of the aerogels obtained is about 95%, measured using the techniques illustrated in Section 2. These results confirm the feasibility of PLLA aerogel formation by SC-CO2 processing and the velocity in obtaining a dried solventless structure. The overall process time is about 4 h; whereas, the traditional process (i.e., freeze drying) requires approximately 7 days [25,26]. Then, in a second set of experiments, we tried to obtain also large micronic cells by superimposing to the drying process, a supercritical foaming. We tried to take advantage of the presence of SC-CO2 in the vessel and after the drying process we

Fig. 3. SEM images of PLLA aerogel obtained at 10% (w/w), 200 bar and 33 ◦ C; (a) section and (b) surface.

E. Reverchon et al. / J. of Supercritical Fluids 45 (2008) 365–373

369

Fig. 5. SEM image of fructose particles used as porogen.

Fig. 4. SEM images of PLLA aerogel obtained at 10% (w/w), 200 bar and 60 ◦ C; (a) magnification at 250× and (b) magnification at 2500×.

increased the operating temperature up to 60 ◦ C; i.e., at temperature higher than the Tg of PLLA. Operating in this manner, the aerogel structure was subjected to nucleation and growth of CO2 bubbles during the depressurization process. Large cellular structures were obtained (Fig. 4) confirming the possibility of superimpose the gas foaming to the drying process; but, the structure was also partially collapsed. Indeed, the thickness of the aerogel dramatically decreased from 4 to 0.5 mm. Moreover, the diameter of the cells was not homogeneous and the porosity considerably decreased (approximately to 56%). We increased the depressurization time and the relative timing between drying and foaming but the results were unsatisfactory. Therefore, we further modified the process and proposed an hybridization of the supercritical drying process with the particulate leaching, previously proposed in the literature [23,25,26]. An alternative to produce large cells inside a polymeric matrix is represented by the use, as porogen, of an insoluble compound that in our process has to be introduced into the polymeric solution before the gelation. We selected d-fructose particles of controlled diameter as the porogen (Fig. 5), since it has a large solubility in water, it is insoluble in SC-CO2 and has been already successfully used as porogen [20]. The selection of poro-

gen particles diameter was obtained by mechanical sieving of the starting material. The introduction of a large quantity of solid in the liquid solution before gelation can strongly interfere with the gelation process itself. Therefore, we performed several tests on PLLA gels with different concentrations of polymer: 10%, 12% and 15% (w/w) and adding fructose particles with diameters ranging between 250–500 and 500–700 ␮m. Gelation and supercritical drying, performed using the same procedure adopted for the drying of PLLA gels without the use of porogens, were successful. Even when large quantities of porogen were added to PLLA, the gel was formed and the produced 3D structure did not present any macroscopic or microscopic indication of instability. In Table 1 we report the operative conditions used, the size and weight of fructose introduced into each scaffold and the porosity of the scaffolds obtained: porosity was always higher than 95%. Moreover, we also performed an analysis of the grade of interconnection of the material and on the surface area by N2 adsorption–desorption. We verified that both the sample at 10% and at 15% presented a completely interconnected structure (i.e., a grade of interconnection of about 100%) and large surface areas (i.e., about 45 m2 /g). In Fig. 6, a SEM image of a PLLA scaffold containing dfructose particles is shown. A homogeneous distribution of the sugar particles inside the polymeric matrix has been obtained. In Fig. 7 we report two SEM images related to the scaffolds formed by PLLA + porogen obtained using different loadings and particles diameters, after the washing step. The complete elimination Table 1 Operative conditions of supercritical drying and porosity of the PLLA + fructose structures obtained with a dioxane/ethanol ratio of 1.7, with 1.2 g of fructose, at 200 bar and 33 ◦ C PLLA dioxane (%)

Solution (g)

Fructose (␮m)

Porosity (%)

10 12 15 10 12 15

0.44 0.45 0.43 0.4 0.41 0.45

250–500 250–500 250–500 500–700 500–700 500–700

97.2 96.7 96.0 97.6 96.9 96.3

370

E. Reverchon et al. / J. of Supercritical Fluids 45 (2008) 365–373

Fig. 6. SEM image of a PLLA scaffold containing the d-fructose particles (particle size 500–700 ␮m).

Fig. 8. SEM image of a PLLA scaffold obtained by supercritical gel drying and the compression of the suspension liquid polymer + porogen.

of porogen is evidenced in all the images. A very regular structure is obtained, characterized by a large cellular structure due to the porogen, embedded into a network of nanometric fibers that represent the original gel structure. Then, we also optimized the interconnection among the cells by pressurizing the suspension liquid polymer + porogen before gelation. We used pressures of 10, 20 and 30 bar. The best results were obtained operating at 10 bar, since this pressure is sufficient to induce the contact among the fructose particles that produces interconnection among the cells in the scaffold (Fig. 8). Larger pressure produced the breakage of the porogen particles and the resulting scaffold was partly collapsed. SEM images produced at large magnifications are reported in Fig. 9. In these images, it is possible to observe that the fibrous substructure characterizes not only the polymeric network, but also the cells wall. In particular, fibers size ranges between 50 and 500 nm, adds further porosity and interconnectivity and produces the walls “roughness” that should be the key factor for cellular growth. Fig. 9b and c are particularly significative in representing the surface structure of the cells; it is also relevant their similarity with collagen nanostructures characteristic of bone cell walls. 3.1. Mechanical tests

Fig. 7. SEM images of PLLA scaffolds obtained by supercritical gel drying combined with particulate leaching. (a) Scaffold obtained using 500–700 ␮m dfructose particles; (b) scaffold obtained using 250–500 ␮m d-fructose particles.

We measured the compressive modulus of some of the scaffolds produced to verify if mechanical resistance is compatible with the one required for bones scaffolding application. We first analyzed PLLA scaffolds produced starting from 12% to 15% (w/w) polymeric solution and using sugar particles ranging between 500 and 700 ␮m. In Fig. 10, we reported two curves representing the stress vs. deformation behavior. To obtain a statistically significant value, the compressive modulus has been measured on seven identical samples (2 cm × 4 mm) for each kind of scaffold produced. An example of these samples is shown in Fig. 11. The scaffolds obtained starting from 15% (w/w) solution present a higher compressive modulus: indeed, 12% (w/w) scaffolds show a mean compressive modulus of 50 kPa (ranging

E. Reverchon et al. / J. of Supercritical Fluids 45 (2008) 365–373

371

Fig. 10. Stress deformation curve of two PLLA scaffold samples: 12% and 15% (w/w), obtained using fructose particles ranging between 500 and 700 ␮m.

can give an higher resistance to the structure. The mechanical resistance required depend on the organs to be repaired; as usual, 100 kPa are necessary for bone scaffolds application [27]. For this reason, we focused our attention on the improvement of the compressive modulus of the PLLA scaffolds. In particular, we tested scaffolds obtained using fructose particles

Fig. 9. SEM images of the same part of a PLLA scaffold obtained at different magnifications: (a) image of a single cavity; (b) enlargement of part of the cavity; (c) further enlargement of the cavity.

between 42 and 58 kPa); whereas a mean compressive modulus of 65 kPa (ranging between 60 and 74 kPa) was obtained for 15% (w/w) scaffolds. This result can be easily explained considering that, when the scaffolds are obtained starting from a more concentrated polymeric solution, the higher amount of polymer

Fig. 11. Series of scaffolds reproduced with similar characteristics (diameter 2 cm and height 4 mm).

372

E. Reverchon et al. / J. of Supercritical Fluids 45 (2008) 365–373

Table 2 Effect of porogen size and polymer concentration on compressive modulus of PLLA scaffolds Polymer concentration (%)

d-Fructose particle size (␮m)

Mean compressive modulus (kPa)a

12 12 15 15

500–700 250–500 500–700 250–500

50 68 65 81

a

Mean calculated on seven samples.

ranging between 250 and 500 ␮m. Since the compressive modulus is inversely proportional to the cell size of the scaffold (i.e., to the porogen size) [20]. This trend is confirmed by the stress–strain analysis performed. The 12% and 15% (w/w) scaffolds obtained using fructose particles of 250–500 ␮m show a higher compressive modulus with respect to the same scaffolds obtained using the porogen of 500–700 ␮m. This result is evident in Table 2: varying the fructose particle size from 500–700 to 250–500 ␮m, the mean compressive modulus of 12% (w/w) scaffolds increases from 50 to 68 kPa; whereas, the mean compressive modulus of 15% (w/w) scaffolds increases from 65 up to 81 kPa. 3.2. Solvent residue analysis We also measured solvent residue present in the scaffolds. In Fig. 12 we reported a typical curve resulting from the analysis of the scaffold samples. The quantity of dioxane was evaluated as the area of the peak corresponding at the time of the eluition of dioxane (20.6 min). The quantity of dioxane residue in the PLLA scaffold considered in Fig. 12 is 263 ppm; it is lower than the limit of USP 467 Pharmacopoeia that is 380 ppm. However, we performed longer drying experiments to verify the possibility of further

Table 3 Process duration vs. solvent residue Process duration (h)

Solvent residue (ppm)

4 6 8

263 <5 <5

decreasing the dioxane residue, since it can strongly interfere with the scaffold application proposed. We modified the drying process time (using times ranging from 4 to 8 h) and analyzed its effect on the solvent residue by headspace-GC. The results are reported in Table 3. As expected, increasing the drying time, dioxane residue sensibly decreases down to a minimum of less than 5 ppm. Therefore, the possibility of eliminating by SC-CO2 , the organic solvents used to produce the gel is confirmed. 4. Conclusions and perspectives We showed that it is possible to generate PLLA scaffolds using a SC-drying process combined with particulate leaching. We obtained scaffolds presenting the nano, micro and macrostructural characteristics that have to be simultaneously verified for tissue engineering applications; indeed, they are characterized by elevated porosity (up to 97%) and surface area (45 m2 /g), by a fibrous structure (that mimes the extracellular matrix), cells of variable size (depending on size of porogen introduced), adequate mechanical properties and no solvent residues. Moreover, it is possible to obtain scaffolds with a predetermined shape and size in a relatively short time. In the next future, we will focus our attention on the inclusion of hydroxyapatite particles inside the polymeric structure with the aim of modifying the scaffold resistance and improving the formation of an environment suitable for bone cells growth; indeed, hydroxyapatite can behave as a support for the structure, increasing the compressive modulus [20] and it allows to create an environment that better mimes the extracellular matrix. Moreover, it is our intention to induce osteocytes growth on the scaffolds produced, to verify the compatibility between the structure and the human cells. References

Fig. 12. Curve resulting from the headspace-GC analysis of PLLA scaffold samples.

[1] R. Langer, J.P. Vacanti, Tissue engineering, Science 260 (5110) (1993) 920–926. [2] P.X. Ma, Scaffold for tissue fabrication, Mater Today 7 (5) (2004) 30– 40. [3] X. Liu, P.X. Ma, Polymeric scaffolds for bone tissue engineering, Ann. Biomed. Eng. 32 (3) (2004) 477–486. [4] R.A. Quirk, R.M. France, K.M. Shakesheff, S.M. Howdle, Supercritical fluid technologies and tissue engineering scaffolds, Curr. Opin. Solid State Mater. Sci. 8 (2004) 313–321. [5] D.J. Mooney, D.F. Baldwin, N.P. Suh, J.P. Vacanti, R. Langer, Novel approach to fabricate porous sponges of (d,l-lactic-co-glycolic acid) without the use of organic solvents, Biomaterials 17 (1996) 1417–1422. [6] D.J. Mooney, B.S. Kim, L.D. Harris, Open pore biodegradable matrices formed with gas foaming, J. Biomed. Mater. Res. 42 (1998) 396–402. [7] Y.W. Kho, D.S. Kalika, B.L. Knutson, Precipitation of Nylon 6 membranes using compressed carbon dioxide, Polymer 42 (2001) 6119–6127.

E. Reverchon et al. / J. of Supercritical Fluids 45 (2008) 365–373 [8] H. Matsuyama, A. Yamamoto, H. Yano, T. Maki, M. Teramoto, K. Mishima, K. Matsuyama, Formation of porous flat membrane by phase separation with supercritical CO2 , J. Membr. Sci. 194 (2001) 157–163. [9] H. Matsuyama, A. Yamamoto, H. Yano, T. Maki, M. Teramoto, K. Mishima, K. Matsuyama, Effect of organic solvents on membrane formation by phase separation with supercritical CO2 , J. Membr. Sci. 204 (2002) 81–87. [10] E. Reverchon, S. Cardea, Formation of cellulose acetate membranes using a supercritical fluid assisted process, J. Membr. Sci. 240 (2004) 187–195. [11] E. Reverchon, S. Cardea, Formation of polysulfone membranes by supercritical CO2 , J. Supercrit. Fluids 35 (2005) 140–146. [12] Q. Xu, M. Pang, Q. Peng, J. Li, Y. Jiang, Application of supercritical carbon dioxide in the preparation of biodegradable polylactide membranes, J. Appl. Polym. Sci. 94 (2004) 2158–2163. [13] Q. Xu, M. Pang, Q. Peng, J. Li, Y. Jiang, H. Wang, M. Zhu, Effect of different experimental conditions on biodegradable polylactide membranes prepared with supercritical CO2 as nonsolvent, J. Appl. Polym. Sci. 98 (2005) 831–837. [14] E. Reverchon, S. Cardea, Generation of biopolymeric membranes by supercritical fluids, Int. J. Artif. Organs 28 (2005) 538. [15] I. Tsivintzelis, E. Pavlidou, C. Panayiotou, Porous scaffolds prepared by phase inversion using supercritical CO2 as antisolvent. I. Poly(l-lactic acid), J. Supercrit. Fluids 40 (2006) 317–322. [16] E. Reverchon, E. Schiavo Rappo, S. Cardea, A flexible supercritical CO2 assisted process for poly (methyl methacrylate) membranes formation, Polym. Eng. Sci. 46 (2006) 188–197. [17] S. Cardea, A. Gugliuzza, E. Schiavo Rappo, M. Aceto, E. Drioli, E. Reverchon, Generation of PEEK-WC membranes by supercritical fluids, Desalination 200 (2006) 58–60.

373

[18] C. Schugens, V. Maquet, C. Grandfils, R. Jerome, P. Teyssie, Polylactide macroporous biodegradable implants for cell transplantation. II. Preparation of polylactide foams by liquid–liquid phase separation, J. Biomed. Mater. Res. 30 (1996) 449–461. [19] Y.S. Nam, T.G. Park, Porous biodegradable polymeric scaffolds prepared by thermally induced phase separation, J. Biomed. Mater. Res. 47 (1999) 8–17. [20] R. Zhang, P.X. Ma, Synthetic nano-fibrillar extracellular matrices with predesigned macroporous architectures, J. Biomed. Mater. Res. 52 (2000) 430–438. [21] M.K. Woo, V.J. Chen, P.X. Ma, Nano-fibrous scaffolding architecture selectively enhances protein adsorption contributing to cell attachment, J. Biomed. Mater. Res. Part A 67 (2003) 531–537. [22] C. Daniel, D. Alfano, V. Venditto, S. Cardea, E. Reverchon, D. Larobina, G. Mensitieri, G. Guerra, Aerogels with a microporous crystalline host phase, Adv. Mater. 17 (2005) 1515–1518. [23] G. Brunner, Gas extraction, in: Steinkopff Darkmstadt, Springer, New York, 1994. [24] E. Reverchon, S. Cardea, Production of controlled polymeric foams by supercritical CO2 , J. Supercrit. Fluids 40 (2007) 144–152. [25] G. Wei, P.X. Ma, Macroporous and nanofibrous polymer scaffolds and polymer/bone-like apatite composite scaffolds generated by sugar spheres, J. Biomed. Mater. Res. Part A 78 (2006) 306–315. [26] P.X. Ma, R. Zhang, Synthetic nano-scale fibrous extracellular matrix, J. Biomed. Mater. Res. 46 (1999) 60–72. [27] J. Zhang, H. Zhang, L. Wu, J. Ding, Fabrication of three dimensional polymeric scaffolds with spherical pores, J. Mater. Sci. 41 (2006) 1725–1731.