Interconnected porous poly(ɛ-caprolactone) tissue engineering scaffolds fabricated by microcellular injection molding

Abstract

In tissue engineering applications, a scaffold containing an interconnected porous structure is often highly desirable since these interconnected pores allow nutrients and signaling molecules to reach all of the cultured cells. In this study, microcellular injection molding, a mass production method for foamed plastic components, was combined with chemical foaming and particulate leaching methods to fabricate an interconnected porous structure using poly(ɛ-caprolactone) (PCL). Sodium bicarbonate (SB) was employed as the chemical foaming agent while carbon dioxide (CO₂) was used as the physical foaming (blowing) agent. The results showed that interconnected porous structures of PCL, which depend on the composition of the materials used, could be successfully produced. Sodium bicarbonate not only generated CO₂ to supplement the supercritical fluid microcellular injection molding, but also served as the nuclei for heterogeneous cell nucleation. Sodium bicarbonate and its byproduct, sodium carbonate, were also the porogens in the particulate leaching process, which further enhanced the porosity and interconnectivity. The morphologies and mechanical properties of the samples with different material compositions and porosities were discussed. The results of cell viability assays of 3T3 fibroblasts suggested that the resulting interconnected porous PCL scaffolds exhibited good biocompatibility. Cell spreading was affected by the porosity of the scaffold because of the physical restriction effect on the cell migration. Highly improved interconnectivity of the scaffold provided more space for the cells to spread.

Keywords

Cell growth chemical foaming/blowing agent compression behaviors decomposition injection foam molding physical foaming/blowing agent

Introduction

A tissue engineering scaffold is an artificial biological structure for supporting cells to form three-dimensional tissues in vitro or in vivo, which is a promising way to repair, restore, or maintain the functions of damaged or diseased tissues. Neo-tissues that grow on the tissue engineering scaffolds that mimic the natural extracellular matrix (ECM) could also be used to conduct drug screening.¹ There are numerous means by which an ECM influences cells and their behaviors such as attachment, proliferation, migration, and differentiation. When developing synthetic ECM substrates, it is important to ensure that the materials are effective at accommodating cells and directing cell growth and proliferation in three dimensions.^2–4 Highly porous and interconnected structures made of biocompatible, biodegradable, and bioresorbable materials are essential for the tissue engineering scaffold for transporting nutrients, oxygen, and waste, as well as for cell migration for tissue vascularization and the formation of new tissues. The porosity and pore size of 3D scaffolds directly influence their functionality during biomedical applications.^4–6

Many methods for fabricating scaffolds using biomaterials, such as natural polymers and synthetic biopolymers, have been developed. These methods include electrospinning, solvent casting and particulate leaching, freeze drying, thermally induced phase separation, 3D printing, gas foaming, and so on.^7–17 However, most of these methods use organic solvents for dissolving the biopolymers, which may be detrimental to cell survival and tissue growth.^18,19 Moreover, the majority of scaffold fabrication techniques are batch processes, which might not be able to fulfill the need for cost-effective manufacturing processes. Therefore, the techniques of mass production of plastics were taken into account. Microcellular injection molding technology combined with particulate leaching was first developed by Kramschuster and Turng.¹⁹ Microcellular injection molding was achieved by a conventional injection molding machine equipped with microcellular injection molding technology (Trexel. Inc.). In this process, atmospheric gas (such as nitrogen (N₂) or carbon dioxide (CO₂)) at the supercritical state (supercritical fluid (SCF)) is introduced into the injection molding barrel and subsequently dissolved in the polymer melt to create a single-phase polymer–gas solution. The pressure inside the barrel is generally kept around 200 bars to facilitate efficient and homogeneous mixing of the gas and polymer, and to maintain the polymer–gas solution in single phase before the injection stage. During injection, the polymer–gas solution is injected into the cavity where the material takes the shape of the cavity and forms the final component(s) through solidification. During the injection stage, the associated pressure drop triggers the thermodynamic instability of the polymer–gas solution. Consequently, the gas emerges from the polymer–gas solution and numerous microscale cells are formed, resulting in a highly porous structure.^19–21 Not only can this process impart porosity into the molded part, but the SCF can also reduce the viscosity of the polymer–gas solution by acting as a plasticizer and filling the interstitial spaces between the polymer molecules, which improves the fluidity and moldability of the materials.^22–26

In Kramschuster and Turng’s study, polylactic acid (PLA), the water-soluble polymer of polyvinyl alcohol (PVOH), and salt particles as particulates were compounded by a twin-screw extruder and then injection molded by microcellular injection molding. The pore size in the scaffolds was controlled by salt particulates and the interconnectivity was achieved by the co-continuous blending morphology of the biodegradable PLA matrix with the water-soluble PVOH. The SCF CO₂ was employed as a plasticizer enabling the blend’s moldability even with an ultra-high filler content of salt. Cui et al. conducted a similar study to produce scaffolds of poly( $ɛ$ -caprolactone) (PCL) and PCL/HA with a different water-soluble polymer of polyethylene oxide (PEO).²⁷ However, even though the porous structures in their studies were interconnected, the size of the channels connecting neighbored pores was not large enough for cells to migrate into.⁷ This was because the co-continuous morphology of the biopolymer and the water-soluble polymer was in the range of several microns, resulting in interconnected channels in the same microscale range after leaching. Therefore, increasing the size of the interconnected channels is necessary to enable cell infiltration, and it can also increase the effective porosity of the scaffold.

In this study, a chemical foaming agent, sodium bicarbonate, was employed to fabricate the porous structure together with the physical foaming agent in the microcellular injection-molding process. The microstructure and mechanical properties of the scaffolds were characterized by scanning electron microscopy (SEM) and an Instron mechanical test unit, respectively. The pore size and porosity of the scaffolds, which depend on material composition and processing, were characterized. Cell proliferation and attachment assays were conducted using the fabricated scaffolds to investigate biocompatibility and cell–scaffold interactions.

Experimental procedure

Materials and material compositions

In this study, particulate leaching techniques were combined with microcellular injection molding to fabricate the porous structure. The schematic of the whole process is shown in Figure 1. Materials selection and samples preparation are described below.

Materials selection

Poly(ɛ-caprolactone) (PCL, CAPA 6500 from Perstorp, UK) was employed as the matrix material in the experiment, as it is a suitable material for cartilage scaffolds owing to its high toughness and matching properties.^28,29 The molecular weight of PCL used in this study was 50,000 g/mol and the melt flow index was 7 g/10 min (2.16 kg, 160℃).

Polyethylene oxide (PEO, POLYOX), a water-soluble polymer, was purchased from DOW (Midland, MI, USA). It had a melting temperature of around 68℃, which was close to the melting temperature of the PCL at 65℃. Having similar matching melting temperatures facilitated the mixing and blending of these two polymers. A preliminary test showed that in a fairly broad volume ratio range (PCL/PEO volume ratio from 60/40 to 40/60), PEO and PCL were able to form a co-continuous morphology, which was desirable for creating interconnected channels in the leached matrix.

Sodium bicarbonate (SB) was applied as both the porogen and the chemical foaming agent, which could be gradually decomposed to sodium carbonate, CO₂, and water (in vapor form) at temperatures above 50℃.³⁰ When the gas and vapor were released through the decomposition reaction process, it would generate more pores in the matrix and improve the overall porosity of the resultant structure. SB powder was purchased from Sigma and the density was 2.159 g/cm³.

Volume ratio of the components

Based on preliminary work, three different kinds of PCL/PEO/SB volume blend ratios—i.e. A, B, and C—were prepared in this study, as shown in Table 1, together with their corresponding weight ratios.

Table 1.

Volume ratio and weight ratio of the material components.

Materials	A (%)		B (%)		C (%)
Materials	Volume ratio	Weight ratio	Volume ratio	Weight ratio	Volume ratio	Weight ratio
PCL	33.3	23.6	37.5	24.2	28.6	21.2
PEO	33.3	23.0	25	21.2	42.8	30.9
SB	33.3	53.4	37.5	54.6	28.6	47.9

Sample preparation

Blending PEO, SB, and PCL

PEO, SB, and PCL in different volume ratios (cf. Table 1) were blended using a twin-screw extruder (screw diameter is 27 mm and L/D ratio is 42) and then pelletized for later use. The processing temperatures were set at 50℃, 60℃, 75℃, 85℃, and 70℃ at the various heating zones of the extruder to minimize the decomposition of the SB. The rotation speed of the screw was kept at 80 r/min. Before compounding, PCL was dried in an oven at 30℃ overnight to reduce moisture.

Microcellular injection molding

Three different blend materials (i.e., A, B, and C) were then molded using an injection molding machine (Arburg Allrounder 320 S) equipped with microcellular injection molding technology. The mold was shaped as a standard tensile bar according to ASTM-638 and the overall dimensions were 166 mm × 12.9 mm × 3.3 mm. The SCF CO₂ was employed in the microcellular injection molding process as the physical foaming agent. The processing temperatures were set as 100℃, 130℃, 140℃, 150℃, and 155℃ at the various heating zones of the injection barrel to allow the SB to decompose at the maximum rate; the mold temperature was 24℃. The injection parameters were set as 20 ccm/s for the injection shot volume speed, 193 bar (2800 psi) for the SCF injection pressure, 2 s for the SCF injection time, 60 bars for the back pressure, and 0.22 g/s for the flow rate of CO₂. The residence time of the material in the barrel was about 180 s.

Leaching process

After injection molding, the molded samples were sliced into small pieces of roughly 10 mm × 10 mm. They were then leached in a deionized water bath for up to 48 h using the setup depicted in Kramschuster and Turng.¹⁹ A continuous supply of water was provided to flush the PEO as it absorbed water and swelled before dissolving. A water pump (Via Aqua1300A) continuously pumped water into the top of the reservoir. The water was discharged into the bottom of the reservoir and cycled back through the pump. The water was changed every 12 h. The residual SB, sodium carbonate (which is the byproduct of SB’s decomposition), and the PEO were then leached until only the PCL with a porous structure remained in the samples. All the samples were vacuum dried for further use.

Characterization

SEM

The injection-molded parts and leached porous parts were fractured in two directions (cf. Figure 2) after being frozen in liquid nitrogen. This allowed the sample cross sections to be observed both parallel to and perpendicular to the flow direction of the injection molded parts. The samples were observed using a LEO1530 field-emission scanning electron microscope (FE-SEM) with an accelerating voltage of 5 kV. All samples were gold-coated before image analysis.

Porosity calculation

In the process of the porosity calculation, the samples were removed from the water bath, and the water on the surface of the sample was carefully cleaned with disposable task wipers. Then the weight of the sample (W₀), which included the weight of the scaffold and the water remaining in the pores, was recorded. Subsequently, the samples were dried in a vacuum oven, and the weight loss was monitored continuously until the weight of the sample remained constant, which was then recorded as (W₁). The porosity of the sample was calculated by equation (1).²⁷

P = \frac{\frac{W_{0} - W_{1}}{ρ_{W}}}{\frac{W_{0} - W_{1}}{ρ_{W}} + \frac{W_{1}}{ρ_{PCL}}} \times 100 %

(1)

Mechanical properties

Mechanical properties for the samples were measured using an Instron 9567 mechanical properties test unit. The samples were placed onto the sample stage along the direction of thickness and a compression strain rate was set as 0.005 in./in./min. The compression modulus and strength were then obtained.

Biological assays

Cell culture

Swiss mouse NIH 3T3 ECACC fibroblasts were employed for the biological assays. These cells were cultured in high-glucose DMEM (Invitrogen) which was supplemented with 20% fetal bovine serum (WiCell), 1 unit/mL penicillin (Invitrogen), 1 µg/mL streptomycin (Invitrogen), and 2 mM L-glutamine (Invitrogen). Tissue culture-treated polystyrene 6-well plates (BD Falcon) were used for routine maintenance. Media was replaced every other day and cells were passaged with EDTA at a 1:40 ratio every six days during regular maintenance.

Cell seeding

The scaffold samples were sterilized by UV exposure for 30 min before seeding. 3T3 cultures were dissociated for 5 min in EDTA and washed with 3T3 media. Cells were then spun down for 3 min at 1000g and the supernatant was removed. Cells were resuspended, counted, and seeded at a density of 10,000 cells/cm² onto the scaffolds, resting in an ultra-low attachment to the culture plates (Corning). The utilization of these plates ensured that only the cells being assayed were those that bound to the scaffold material under examination. Cells were carefully fed with 3T3 media every two days (as described above). Live/dead assays were taken on days 3 and 10, and all tests were done with at least three replicates.

Live/dead assay

Viability was assessed via LIVE/DEAD Viability/Cytotoxicity Kits (Invitrogen) after 3 and 10 days of culture. The stain utilized red fluorescent ethidium homodimer-1 (EthD-1) to indicate cell death by penetrating damaged cellular membranes, and green fluorescent Calcein-AM to target esterase activity within the cytoplasm of living cells. Cells were imaged with an A1RSi confocal microscope with an attached Photometrics CoolSNAP HQ2 camera. NIS Elements D Advanced Research (v.3.22) software was used for the image analysis.

Results and discussion

Morphological analysis

Physical foaming and chemical foaming

Physical foaming was achieved through the use of SCF CO₂, which occurred during the microcellular injection molding process. Chemical foaming, on the other hand, was done via the decomposition of SB, the byproduct of which (i.e. CO₂ and water vapor) could also form cells or pores in the matrix. Figure 3 shows the cross section images of as-molded, pre-leached samples made of the blend materials of A, B, and C by microcellular injection molding. Several pores ranging from tens of microns to more than 200 microns were apparent in the samples, and there were some areas showing many interconnected pores (as seen in the squares in Figure 3), which might have been formed by the chemical foaming agent SB. Recall that the pores in the parts produced by microcellular injection molding were usually closed pores that were isolated from each other. However, the chemical foaming agent SB decomposed,^30,31 and the released CO₂ and H₂O (vapor) further increased the porosity by providing more foaming agent.³² Therefore, some large and connected pores were produced. However, because the pressure in the barrel was very high, the injected physical foaming agent, CO₂, from the SCF injection unit was maintained in the SCF state, and accordingly the decomposition reaction of SB could be hampered by the existing CO₂ in the injection barrel,³³ until the blend was injected into the mold cavity. During the filling stage, as the pressure in the cavity dropped suddenly, the CO₂ emerged rapidly from the single-phase polymer–gas solution and a great number of cells started to grow. On the other hand, it still took a period of time for temperature of the sample to decrease to the value of mold temperature, hence the SB decomposed at a very high rate, leading to the generation of additional pores. It can be seen that the cell numbers and size of sample A, B, and C varied with the content of SB in blends (cf. Figure 3). It is known that the melt vicosity is a key factor for both cell nucleation and growth.³⁴ Lower viscosity is favorable for nucleation while higher one is benefical for pore growth. In the PCL/PEO/SB/CO₂ system, the viscosity of the blends was affected by the addition of solid particles, SB, and the more SB in the blends the more viscos it was in the blend. Therefore, smaller amount of cells was shown in sample B with the cell size being larger than samples A and C.

A mixture of chemical and physical foaming agent obtained the advantages of both agents.³² Physical foaming agents provide additional gas during foaming and improve the filling behavior by forming polymer–gas solution acting as a plasticizer. The nucleation and growth of cells from both sources could easily lead to connected pores, the reason is that CO₂ release from SB decomposition was localized, and there might be CO₂ releasing among the pores generated by physical foaming which might be able to breakthrough the neighbored pores, which was also reported in our previous work.³⁵ As a result, a number of interconnected pores were created in the matrix, which leveraged both of the chemical foaming agent and microcellular injection molding. Based on the thermal decomposition of SB and the physical foaming process, a rough estimation was done to further investigate the two kind of foaming process.³⁰ For physical foaming, the CO₂ uptake for each shot was 0.44 g (gas flow rate was 0.22 g/s, and 2 s per shot) because each volume injected was about 7.1 cm³ (166 mm × 12.9 mm × 3.3 mm), and this indicated that the amount of CO₂ was between 10% and 20% with respect to the PCL phase. On the other hand, for the SB decomposition, according to the reaction kinetics curve in the literature,³³ it can be estimated that the conversion fraction of SB was about 10% (assuming that in the whole cycle (60 s) the melt temperature was kept at about 140℃), and the chemical foaming agent generated by the decomposition can be shown as about 0.28 g CO₂ which was about 10% with respect to the PCL phase. Even though there is no direct data to describe the solubility of CO₂ in PCL at the temperature above 100℃,^36,37 it can be estimated that the solubility might be smaller than 0.03, and this number is much less than the CO₂ in the melt which indicates that most of the CO₂ did not contribute to the gas foaming process.

The pores were prone to grow around the center of the part, where the temperature decreased much slower than in the skin layer, which contained far fewer pores. The skin layer was about 200–400 µm as shown between the short lines in Figure 3. In the image with a larger magnification (Figure 4), the sodium carbonate or undecomposed sodium bicarbonate can be clearly observed in the skin area. These particulates have a size of tens of microns and would be leached in the subsequent particulate leaching process.

Porous structure

After leaching the particulates in the injection-molded parts, more pores were visible in the cross sections. It can be clearly observed in Figure 5 that the particulates of sodium carbonate or sodium bicarbonate observed in the skin layers of the samples (cf. Figure 4) were leached completely, and highly porous structures were produced. After a very careful comparison of the leached parts with the un-leached parts, it was found that the pores with smooth cell geometries were likely created by the foaming process, whereas the pores with sharp edges were presumably generated by the particulate leaching process (cf. Figure 6). The pores generated by gas foaming and those generated by particulate leaching connected with each other after removing the PEO phase, undecomposed SB and one of the byproducts of SB decomposition, sodium carbonate, which increased the interconnectivity of the pores in the scaffold samples.

The flow behavior of the melt also affected the morphology of the pores. In Figure 7, it can be seen that the pores were elongated along the flow direction. This was because the pores were forming and growing in the melt during the filling process. Evidently, the shear stress in the melt drew the pores from a spherical shape to be an elliptical shape. In this process, there was also the possibility of neighboring pores connecting with each other. It can be seen from Figure 7 that the pores were interconnected along the longitudinal direction. The channel consisted of several connected pores and had a length of more than 1 mm, which was one-third of the sample thickness (Figure 7(c)). Additionally, the diameter of the interconnected channels was up to 100 µm, which was large enough for cells to migrate through. These results indicated that the interconnectivities of the scaffolds made of A, B, and C were improved by combining the dual physical and chemical foaming mechanisms.

Pore size, porosity, and interconnectivity

The quantitative analysis of the pore size on cross section was performed using the ImageJ software by measuring the short axis of a pore because this dimension might limit the penetration of living cells. As shown in Figure 8(a), the pore sizes were all around 35 µm for samples of A, B, and C, which were uniform even though the component ratios varied for different samples. The results of porosity are shown in Figure 8(b). The porosities of samples of A, B, and C were 68.2%, 73.8%, and 70.9%, respectively. However, it was interesting to see that the calculated porosities were very close to the theoretical value of the targeted porosity (sum of PEO and SB volume ratios), namely 66.7%, 62.5%, and 71.4%, respectively. This suggested the effective removal of all of the water soluble materials from the matrix. For sample A, the porosity was slightly higher than the theoretical value of 66.7%. For sample B, the porosity was about 10.3% higher than the corresponding theoretical value. For sample C, the porosity was lower than, but very close to, the theoretical value which might because a small amount of SB had decomposed in the twin-screw extrusion blending process. This discrepancy comes from the additional voids resulting from the microcellular injection-molding process combined with SB decomposition, as well as the possible loss of some isolated PCL domains during leaching. By comparing the porosity with their theoretical values and recognizing that sample B had the highest SB content, followed by samples A and C, it can be concluded that CO₂-assisted foaming contribute less than the phase leaching process in pore generation, which was also reported in Kramschuster and Turng.¹⁹ It is interesting to note that the measured pore sizes among the three samples were fairly uniform. Therefore, it can be hypothesized that most of the pores were generated by particulate leaching.

Mechanical properties

The mechanical properties of a scaffold are another important characteristic attribute for tissue engineering. Figure 9 shows the compression test results of samples A, B, and C. Note that in the mechanical testing, the wet condition (before being vacuum-oven dried) was utilized. The compressive moduli of samples A, B, and C were 7.9 MPa, 2.5 MPa, and 7.2 MPa, respectively. In addition, the compressive strengths of samples A, B, and C were 0.26 MPa, 0.16 MPa, and 0.21 MPa, respectively. The compressive properties of sample B were much lower than the other samples because the porosity of sample B was higher than the other samples. This suggests that the porosity significantly affected the mechanical properties and that it was a challenge to balance the porosity and mechanical properties for scaffold fabrication.

Cell proliferation and attachment

Preliminary biological assays were done to gain an understanding of how the cells responded to these scaffolds based on their various compositions and porosities. The florescence images of 3T3 cells cultured on different scaffolds are shown in Figure 10. As described in the experimental section, living cells fluoresced green and dead cells fluoresced red due to the calcein-AM and EthD-1 stains, respectively. It can be seen that most of the cells were alive after three days of culture, suggesting that the cells that attached to the substrate remained viable. This revealed that the scaffolds had good biocompatibility. The result of 10 days of culture showed an increase in cell spreading. Based on three-dimensional confocal microscope imaging, it was observed that the cells migrated throughout the interconnected channels and arrived in the deeper portions of the scaffold as labeled in the images. This suggested that the improved interconnectivity provided more space for the cells to spread and penetrate. It should be pointed out that cells did not spread as well into the deeper portions for sample C because the porosity of C was the lowest among the three samples. This was likely due to the physical restrictions in the migration of the cells, and reduced nutrition diffusion to the cells into the scaffolds. Sample B showed the best cell proliferation and spreading, as the cells nearly occupied the entire area of the cross section, both on the surface and inside of the scaffold. This was likely attributed to sample B having the highest porosity, which enabled the highest mass transportation rate. The cell response was also good on sample A, even though it had a lower porosity than that of sample B. The cell proliferation and attachment results showed that the porosity played a very important role in cell–scaffold interactions and cell–cell interactions. More specific cell tests are needed, however, to further evaluate the clinical implications of these scaffolds.

Conclusions

By combining microcellular injection molding with chemical foaming, biopolymer-based tissue engineering scaffolds were successfully produced. The results revealed that interconnectivity was improved by the introduction of sodium bicarbonate as both a chemical foaming agent and a porogen. The porosity of the scaffold was highly affected by the material composition, represented by the volume ratios of water-soluble polymers and sodium bicarbonate. As expected, the scaffolds with a higher porosity had lower mechanical properties. The cell attachment and proliferation tests showed that porosity played an important role in cell proliferation and penetration, and that improved interconnectivity enhanced cell migration. The results suggested that the combination of microcellular injection molding and chemical foaming could be a very promising method for fabricating highly porous and interconnected tissue engineering scaffolds.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The authors are grateful for the financial support from the Chinese Scholarship Council (CSC) and the Wisconsin Institute for Discovery (WID) at the University of Wisconsin–Madison. The authors also acknowledge financial support from the International Science & Technology Cooperation Program of China (grant no. 2015DFA30550), the National Science Foundation of China (grant no. 11372286), the Key Project of Science and Technology of the Education Department of Henan Province (grant no. 15A430047), and the Key Project of International Cooperation of Henan Province (grant no. 152102410013).

References

Langer

Vacanti

. Tissue engineering. Science 1993; 260: 920–926.

Salerno

Di Maio

Iannace

et al.

Tuning the microstructure and biodegradation of three-phase scaffolds for bone regeneration made of PCL, Zein, and HA. J Cell Plast 2011; 47: 245–60.

Yang

Leong

et al.

The design of scaffolds for use in tissue engineering. Part 1. Traditional factors. Tissue Eng 2001; 7: 679–689.

Salerno

Netti

Di Maio

et al.

Engineering of foamed structures for biomedical application. J Cell Plast 2009; 45: 103–117.

Hollister

. Porous scaffold design for tissue engineering. Nat Mater 2005; 4: 518–524.

Causa

Netti

Ambrosio

. A multi-functional scaffold for tissue regeneration: the need to engineer a tissue analogue. Biomaterials 2007; 28: 5093–5099.

Loh

Choong

. Three-dimensional scaffolds for tissue engineering: role of porosity and pore size. Rev 2013; 19: 484–502.

Wiesmann HP and Lammers L. Scaffold structure and fabrication. In: Meyer U, Handschel J, Wiesmann HP and Meyer T (eds) Fundamentals of Tissue Engineering and Regenerative Medicine. Berlin, Heidelberg: Springer Berlin Heidelberg, 2009, pp. 539–49.

Yang

Leong

et al.

The design of scaffolds for use in tissue engineering. Part II. Rapid prototyping techniques. Tissue Eng 2002; 8: 1–11.

10.

Hollister

. Porous scaffold design for tissue engineering. Nat Mater 2005; 4: 518–524.

11.

Nam

Yoon

Park

. A novel fabrication method of macroporous biodegradable polymer scaffolds using gas foaming salt as a porogen additive. J Biomed Mater Res 2000; 53: 1–7.

12.

Cuénoud

Bourban

P-E

Plummer

CJG

et al.

Plasticization of polylactide foams for tissue engineering. J Cell Plast 2012; 48: 409–432.

13.

Sannino

Madaghiele

. Tuning the porosity of collagen-based scaffolds for use as nerve regenerative templates. J Cell Plast 2009; 45: 137–155.

14.

Jing

H-Y

Salick

et al.

Morphology, mechanical properties, and shape memory effects of poly(lactic acid)/ thermoplastic polyurethane blend scaffolds prepared by thermally induced phase separation. J Cell Plast 2014; 50: 361–379.

15.

Bhamidipati

Scurto

Detamore

. The future of carbon dioxide for polymer processing in tissue engineering. Tissue Eng B Rev 2013; 19: 221–232.

16.

Pintado-Sierra

Delgado

Aranaz

et al.

Surface hierarchical porosity in poly (ɛ-caprolactone) membranes with potential applications in tissue engineering prepared by foaming in supercritical carbon dioxide. J Supercrit Fluid 2014; 95: 273–284.

17.

Ogunsona

D'Souza

. Characterization and mechanical properties of foamed poly(-caprolactone) and Mater-Bi blends using CO₂ as blowing agent. J Cell Plast 2014; 51: 245–268.

18.

Agrawal

Ray

. Biodegradable polymeric scaffolds for musculoskeletal tissue engineering. J Biomed Mater Res 2001; 55: 141–150.

19.

Kramschuster

Turng

L-S

. An injection molding process for manufacturing highly porous and interconnected biodegradable polymer matrices for use as tissue engineering scaffolds. J Biomed Mater Res B: Appl Biomater 2009; 92: 366–376.

20.

Kramschuster

Cavitt

Ermer

et al.

Quantitative study of shrinkage and warpage behavior for microcellular and conventional injection molding. Polym Eng Sci 2005; 45: 1408–1418.

21.

Zhao

Yan

Zhao

et al.

Microcellular injection molded polylactic acid/poly (ɛ-caprolactone) blends with supercritical CO₂: correlation between rheological properties and their foaming behavior. Polym Eng Sci 2016; 56: 39–946.

22.

H-Y

Jing

Turng

L-S

. Fabrication of porous synthetic polymer scaffolds for tissue engineering. J Cell Plast 2015; 51: 165–196.

23.

Kwag

Manke

Gulari

. Rheology of molten polystyrene with dissolved supercritical and near-critical gases. J Polym Sci B: Polym Phys 1999; 37: 2771–2781.

24.

Royer

Gay

Desimone

et al.

High-pressure rheology of polystyrene melts plasticized with CO₂: experimental measurement and predictive scaling relationships. J Polym Sci B Polym Phys 2000; 38: 3168–3180.

25.

Kwag

Manke

Gulari

. Effects of dissolved gas on viscoelastic scaling and glass transition temperature of polystyrene melts. Indus Eng Chem Res 2001; 40: 3048–3052.

26.

Markočič

Škerget

Knez

. Effect of temperature and pressure on the behavior of poly(ɛ-caprolactone) in the presence of supercritical carbon dioxide. Indus Eng Chem Res 2013; 52: 15594–15601.

27.

Cui

Nelson

Peng

et al.

Fabrication and characterization of injection molded poly (ɛ-caprolactone) and poly (ɛ-caprolactone)/hydroxyapatite scaffolds for tissue engineering. Mater Sci Eng C 2012; 32: 1674–1681.

28.

Martinez-Diaz

Garcia-Giralt

Lebourg

et al.

In vivo evaluation of 3-dimensional polycaprolactone scaffolds for cartilage repair in rabbits. Am J Sport Med 2010; 38: 509–519.

29.

Schagemann

Chung

Mrosek

et al.

Poly-ɛ-caprolactone/gel hybrid scaffolds for cartilage tissue engineering. J Biomed Mater Res Part A 2010; 93A: 454–463.

30.

Keener

Frazier

Davis

. Thermal decomposition of sodium bicarbonate. Chem Eng Commun 1985; 33: 93–105.

31.

Ball

Snelling

Strachan

et al.

Thermal decomposition of solid sodium bicarbonate. J Chem Soc, Faraday Transact 1: Phys Chem Condens Phase 1986; 82: 3709–3715.

32.

Al-Moameri

Zhao

Ghoreishi

et al.

Simulation blowing agent performance, cell morphology, and cell pressure in rigid polyurethane foams. Indus Eng Chem Res 2016; 55: 2336–2344.

33.

Hartman

Svoboda

Pohořelý

et al.

Thermal decomposition of sodium hydrogen carbonate and textural features of its calcines. Indus Eng Chem Res 2013; 52: 10619–10626.

34.

Nofar

Majithiya

Kuboki

et al.

The foamability of low-melt-strength linear polypropylene with nanoclay and coupling agent. J Cell Plast 2012; 48: 271–287.

35.

Wang

Jiang

Hou

et al.

Fabrication of porous structure of poly(ɛ-caprolactone) via microcellular injection molding combined with chemical foaming. CIESC J 2014; 65: 2386–2392.

36.

Leeke

Cai

Jenkins

. Solubility of supercritical carbon dioxide in polycaprolactone (CAPA 6800) at 313 and 333 K. J Chem Eng Data 2006; 51: 1877–1879.

37.

Pastore Carbone

Di Maio

Scherillo

et al.

Solubility, mutual diffusivity, specific volume and interfacial tension of molten PCL/CO₂ solutions by a fully experimental procedure: effect of pressure and temperature. J Supercrit Fluids 2012; 67: 131–138.