Article pubs.acs.org/IECR Solid-State Foaming of Cyclic Olefin Copolymer by Carbon Dioxide Zhenhua Chen,†,‡ Changchun Zeng,*,†,‡ Zhen Yao,*,∥ and Kun Cao§,∥ † High Performance Materials Institute, Florida State University, Tallahassee, Florida 32310, United States Department of Industrial and Manufacturing Engineering, FAMU-FSU College of Engineering, Florida State University, 2525 Pottsdamer Street, Tallahassee, Florida 32310, United States § State Key Laboratory of Chemical Engineering, Zhejiang University, Hangzhou 310017, China ∥ Institute of Polymerization and Polymer Engineering, Department of Chemical and Biological Engineering, Zhejiang University, Hangzhou 310017, China ‡ ABSTRACT: Solid-state foaming of cyclic olefin copolymer (COC, Topas 6017) using carbon dioxide (CO2) was investigated. Before the foaming experiment, the solubility of CO2 in the polymer at various saturation pressure (5−10 MPa) and saturation temperature (40 °C) was measured using the ex-situ gravimetric method. The effects of the saturation pressure, foaming temperature, and foaming time on the foam structure were systematically investigated. Ultramicrocelluar foams with a cell size of 0.5 μm and cell density over 1012 cells/cm3 were successfully prepared. Moreover, it was observed that the system, which was strongly plasticized by CO2, would transition from foaming to crazing under certain conditions. The phenomenon was examined in detail and understood by a proposed mechanism that accounts for both homogeneous nucleation and stress fields induced crazing. an increase in both the effective yield strain and the elongation at the break of the polymers, which are beneficial for bubble nucleation and growth. The solid-state foaming process has been used to prepare microcellular foams from a number of amorphous and semicrystalline polymers, such as poly(vinyl chloride),15 polystyrene,16 polycarbonate,14 poly(acrylonitrile− butadiene−styrene),17 poly(methyl methacrylate),4 poly(ethylene terephthalate),18 polyetherimide,19,20 poly(ether sulfone),10,19,21 and poly(aryl ether ketone),22 etc. One of the key processes to determine the cellular structure is the nucleation and growth of gas bubbles. The resulting structure depends on a balance between several mechanisms, such as the gas solubility and plasticization effect, gas diffusion, and the momentum and thermal transfer in the polymer, which depend inherently on the polymer and process conditions.23 Cell nucleation phenomena in microcellular foaming have been primarily studied by applying classical nucleation theory.24−26 A different mechanism was proposed by Holl et al.27 in solid-state foaming, in which cell nucleation is caused by a triaxial tensile failure mechanism which is similar to the explosive decompression failure observed in elastomers used under high-pressure environments.28 Cycloolefine copolymers (COCs) are a family of amorphous copolymers of ethylene (E) and norbornene (NB) with varying E/NB ratios, which largely determines their mechanical and thermomechanical properties. For example, a higher norbornene fraction in the COC would result in a polymer with a higher glass transition temperature (Tg). COCs are a relatively new thermoplastic with many excellent properties such as good 1. INTRODUCTION Microcellular foams were initially developed by Martini and coworkers at MIT1 using inert gases (carbon dioxide and/or nitrogen) as the physical-blowing agents. Generally, these foam materials are characterized by a cell density higher than 109 cells/cm3 and a cell size of 10 μm or less. They have attracted a great deal of interest in the past few decades due to their unique capability of offering a new range of insulating and mechanical properties with concurrent reduction in materials weight and cost.2,3 Microcellular polymers have been shown to possess high impact strength,4−6 high toughness,7 high stiffness-toweight ratio,4,5 high thermal stability,8 and low thermal conductivity9 when compared to their unfoamed counterparts. As a result, they have been increasingly used in a variety of applications for food packaging, insulation, filtration membrane, sports equipment, automobiles, aircraft parts, etc. Several techniques have been developed to prepare microcellular foams by physical foaming using gases in their supercritical or subcritical states.10−13 These techniques are based on the following common principles: (1) dissolution of gas in a polymer to form a polymer−gas solution; (2) quenching of the polymer−gas solution into a supersaturated state by either reducing pressure (pressure quenching method) or increasing temperature (temperature jumping method); (3) nucleation and growth of gas bubbles till the thermodynamic and/or mass transport driving forces vanish. Both pressure quenching and temperature jump methods take advantage of the plasticization effect of the dissolved gas, which depresses the intrinsic glass transition temperature (Tg) of the polymer. If the plasticizing effect and Tg depression are prominent, foaming may take place at temperatures below the normal glass transition temperature of the polymer, leading to a solid-state foaming process.14 During the process, the plasticization effect also causes a significant reduction in the Young’s modulus and © XXXX American Chemical Society Received: March 14, 2013 Revised: June 8, 2013 Accepted: June 13, 2013 A dx.doi.org/10.1021/ie400833e | Ind. Eng. Chem. Res. XXXX, XXX, XXX−XXX Industrial & Engineering Chemistry Research Article a function of time. Sample transfer was completed as quickly as possible (∼40 s after opening the relief valve) to minimize gas loss before measurement. 2.3. Solid-State Foaming. COC sample was saturated for 28 h with CO2 under designated pressure and temperature using the same apparatus for gas solubility measurements. From the solubility study it was determined that under the chosen saturation time equilibrium was reached in all experiments. Following rapid release of the pressure, the sample was removed from the pressure vessel and quickly transferred (<1 min) to an ethylene glycol oil bath for foaming. Foaming time was kept constant at 30 s unless otherwise indicated. The foamed sample was then quenched in an ice/water mixture to fix the foam morphology. 2.4. Foam Characterization. The foam cell morphology was observed with a JEOL (JSM-7401 F) scanning electron microscope (SEM). Samples were freeze fractured in liquid nitrogen to expose the cross-sectional cellular structure before being mounted on stubs and sputter coated with Au/Pd. Images were acquired at an accelerating voltage of 10 kV. To obtain the average cell size and cell density, image analysis of the SEM micrographs was conducted using Image J (National Institutes of Health). Cell diameter was obtained by calculating the average cell diameter of at least 100 cells in the SEM micrographs. Cell density (Nf), number of cells per cubic centimeter of foam, was determined from eq 140 mechanical properties, high transparency, high glass-transition temperatures, high thermal stability, low dielectric constants and dielectric loss, excellent chemical resistance, low moisture uptake, and good barrier properties.29 Because of this unique combination of properties, COCs are being explored as materials for various applications, e.g., optical data storage, lenses, packaging, medical equipment, etc.30,31 In addition to these applications which take advantage of the excellent transparency of COCs, there are many other actual or potential uses of COCs, such as housings, gears, powder coatings, toner binder resin, and filter media for air and other gases.31,32 Moreover, the combination of excellent mechanical and thermal properties make them viable alternatives to polyolefin. To date, a few studies have shown that carbon dioxide has substantial solubility10,33 and a strong plasticization effect10 in several COCs. The feasibility of preparing COC foams using carbon dioxide has been demonstrated.10,33−37 However, systematic investigation of the processing−morphology relationship of CO2 foaming of COCs is generally lacking. Moreover, the foaming behavior of COCs with very high norbornene content, which is important in thermally demanding applications, has not been studied. Inspired by the excellent foamability of other amorphous high Tg polymers such as polyetherimide,19,20 polysulfone,19 and fluorinated poly (aryl ether)38 from which even nanocellular structure can be obtained by the solid-state foaming process, COC microcellular foaming by the solid-state foaming was systematically investigated in this work using a COC polymer with a high norbornene content and glass transition temperature (180 °C). The effects of process parameters such as saturation pressure, foaming temperature and foaming time on the microstructure and morphology of foams were systematically studied. An “unusual” phenomenon was revealed and analyzed in detail, where the increase of saturation pressure was detrimental to solid-state foaming. The phenomenon, while seemingly inconsistent with the classical nucleation theory, can be readily understood by a new mechanism proposed in the study. ⎛ nM2 ⎞3/2 Nf = ⎜ ⎟ ⎝ A ⎠ (1) where n is the number of cells in the SEM micrograph, M the magnification factor, and A the area of the micrograph (cm2). 3. RESULTS AND DISCUSSION 3.1. Solubility of CO2 in COC. CO2 solubility was measured by the desorption experiments described earlier. In this work, the saturation temperature was fixed at 40 °C. This temperature, which is slightly higher than the critical temperature of the CO2 (Tc = 31 °C), was deliberately selected to avoid potential issues associated with CO2 liquification at high saturation pressure. Moreover, compared to the slow sorption process at room temperature (∼25 °C) that is common for solid-state foaming,14,20 CO2 sorption was greatly expedited because of the vastly increased diffusion rate enabled by the higher temperature and converting CO2 into supercritical state.41 The dissolved CO2 per unit weight of polymer (Mgas) was determined from the weights of the polymer−CO2 sample at different desorption times and the initial polymer. Values were plotted versus the square root of desorption time as shown in Figure 1. A good linear relationship was observed, indicating the desorption process can be adequately described by Fickian diffusion.42 In addition, the solubility of CO2 in the samples can be obtained by linearly extrapolating the desorption curve to zero desorption time according to the one-dimensional Fickian diffusion equation42 2. EXPERIMENTAL SECTION 2.1. Materials. The COC sheets used in the study (Topas 6017) with a thickness of 1 mm were kindly supplied by the manufacturer (TOPAS Advanced Polymers Inc.). The polymer has a density of 1.02 g/cm3 and a glass transition temperature of 180 °C. Sheets were cut into specimens of size 10 mm ×25 mm and dried in a vacuum oven for 24 h before use. Bone dry grade carbon dioxide (i.e., 99.9% purity) was purchase from Airgas Inc. and used as received. 2.2. Solubility Measurements. To determine the equilibrium solubility of CO2 in the COC, desorption experiments were performed using an ex-situ gravimetric method (mass-loss analysis).39 The COC specimen was placed in a high-pressure vessel and evacuated. The vessel was immersed in a thermostat by which the saturation temperature was controlled (ECO-E20G, LAUDA). CO2 was then delivered into the pressure vessel using a high-pressure syringe pump (260D, Teledyne ISCO) to reach the designated saturation pressure. The saturation pressure was controlled (by the pump) with 0.5% full-scale accuracy, and the saturation temperature was controlled (by the thermostat) with an accuracy of ±0.01 °C. After a prespecified sorption time, pressure was released and the saturated COC sample was rapidly removed and transferred to a semimicroanalytical balance (CPA225D Sartorius; precision, 0.01 mg) to record the sample weight as ⎛ Dt ⎞1/2 Mgas = M∞ − 4⎜ 2 ⎟ M∞ ⎝ πL ⎠ (2) where Mgas is the mass of residual gas in the polymer sample, M∞ is the mass uptake at the experimental saturation time, t is B dx.doi.org/10.1021/ie400833e | Ind. Eng. Chem. Res. XXXX, XXX, XXX−XXX Industrial & Engineering Chemistry Research Article model (eq 3), which has been extensively studied and widely used to describe the sorption of gases in glassy polymers43 C = KDPS + C H′ bPS 1 + bPS (3) where C is the equilibrium gas solubility, KD is Herny’s gas dissolution constant, PS is the gas saturation pressure, C′H is the capacity constant for Langmuir mode sites, and b is the affinity constant for the Langmuir mode adsorption. At high gas saturation pressure, eq 3 can be simplified to C = KDPS + C H′ bPS C′ b = KDPS + 1 H ≈ KDPS + C H′ 1 + bPS +b P s (4) Thus, for saturation processes occurring at high pressure, the dual-mode sorption model predicts a linear relationship between the saturation pressure and the equilibrium gas solubility, which was observed in the current study. From the zero pressure intercept and slope of the fitting line in Figure 3, CH′ and KD were estimated to be 0.029 (g of gas/g of polymer) and 0.0074 (g of gas/g of polymer/MPa). 3.2. Foaming Process and Foam Structure. The COCs were foamed under a broad range of saturation and foaming conditions to investigate the effects of typical processing parameters on the foam structure. Specifically, the influence of foaming time, foaming temperature, and saturation pressure was studied in detail. 3.2.1. Effect of Foaming Time. Foaming was conducted at a series of times using a foaming temperature of 170 °C. Prior to foaming all samples were saturated at 7 MPa. Note that even though the foaming temperature is lower than the glass transition temperature of the COC polymer, the polymer was heavily plasticized by the saturated CO2 and foaming readily proceeded. Cellular structures of the foams were examined by SEM and are shown in Figure 3. At short foaming time (10 s), the foam bubbles appeared round in shape and were largely isolated from each other. As foaming time increased and foam cell grew, the cell size increased and the foam cells started to impinge against each other. This was accompanied by the change of the cell shapes, from initial spherical shape to polygonal and elongated polygonal that was oriented along the thickness direction of the foam. At prolonged foaming time, the cell interference became more severe and extensive rupture of cell walls was observed. This may be attributed to the low stiffness of the polymer matrix at the high temperature and high gas concentration and, thus, failure to resist tensile deformation of the cell wall when a relatively long foaming time was applied. This will be discussed in more detail later. Cell size and cell density of the foams were analyzed by image analysis and are summarized in Figure 4. Ultramicrocellular foam with a cell size of 0.5 μm and cell density of 1.4 × 1012 cells/cm3 was prepared using a foaming time of 10 s. Increasing foaming time led to an increase of cell size and concomitant decrease of cell density, resulting from cell growth and coalescence. As foaming proceeds, the dissolved gas was depleted by bubble growth and diffusion out of the matrix. Both led to the reduction of gas concentration and concentration gradient and hence the decrease of bubble growth. Meanwhile, the matrix stiffened as a result of gas depletion and increase of the system Tg and matrix modulus. Eventually the system selfvitrified and foam morphology was frozen. In the present study after 50 s of foaming both cell size and cell density appeared to Figure 1. Desorption CO2 from COC6017 after being saturated for 28 h at 40 °C and 10 MPa. the desorption time, and D and L are the diffusion coefficient and specimen thickness, respectively. To determine the saturation time necessary to achieve equilibrium sorption, CO2 solubility was measured under a series of saturation times for pressures ranging from 5 to 10 MPa. In all cases the solubility increased asymptotically with increasing saturation time, and 28 h was sufficient for the systems to reach equilibrium. This saturation time was hence used for all foaming experiments to ensure the homogeneity of the cellular structures of the foams prepared, as the concentration gradient from nonequilibrium sorption can lead to graded cell size and density.41 Figure 2 shows the equilibrium solubility of CO2 in the COC as a function of saturation pressure at 40 °C. The solubility Figure 2. Equilibrium CO2 solubility in COC6017 as a function of saturation pressure at 40 °C. increased linearly with increasing saturation pressure, consistent with other reports.43 The linear relationship between the solubility and the saturation pressure observed in Figure 2 implies that the sorption behavior may be modeled by the dual-mode sorption C dx.doi.org/10.1021/ie400833e | Ind. Eng. Chem. Res. XXXX, XXX, XXX−XXX Industrial & Engineering Chemistry Research Article Figure 3. SEM micrographs of COC saturated at 7 MPa and foamed at 170 °C for various foaming times: 10 (a), 30 (b), 50 (c), 70 (d), 110 (e), and 170 s (f). have reached the plateau values, ∼1 μm and ∼6 × 1011 cells/ cm3, respectively. Further extending the foaming time did not impact the cell size and cell density. The seemingly minute change in cell size and cell density after 70 s is most likely due to variation associated with image analysis and calculation. 3.2.2. Effect of Foaming Temperature. Foaming temperature has a profound influence on cell morphology because it affects both cell nucleation and growth. Thus, samples were foamed at a series of temperatures (120−180 °C) using samples saturated at 5, 6, and 7 MPa. Cell size and cell density were analyzed and are shown in Figure 5. In all cases, cell size increases with increasing temperature. This is consistent with the fact that higher temperatures accelerate cell growth because of the combined effects of higher gas diffusivity and lower D dx.doi.org/10.1021/ie400833e | Ind. Eng. Chem. Res. XXXX, XXX, XXX−XXX Industrial & Engineering Chemistry Research Article Figure 4. Average foam cell diameter (left) and cell density (right) as a function of foaming time; saturation pressure and foaming temperature were 7 MPa and 170 °C, respectively. Figure 5. Average foam cell diameter (left) and cell density (right) as a function of foaming temperature at different saturation pressures. which decreases as temperature increases. Higher temperature thus has two opposite effects on the cell nucleation rate. On one hand, it facilitates the increase of nucleation by greatly enhancing the exponential term; on the other hand, it may reduce the nucleation rate because of the reduction in available gas. More importantly, as a result of the increased gas diffusivity and lower polymer viscosity at higher temperatures, cell growth becomes more prominent. As both nucleation and growth compete for available gas, nucleation is further suppressed. Therefore, the observed temperature dependency of cell density is a result of the convoluted effects associated with temperature. It can argued that below Tmax at which the maximum of the cell density was achieved the nucleation enhancement effect was more prevalent and hence the increase of cell density. At temperatures higher than Tmax the situation was reversed and the suppression effect prevailed, resulting in the decrease of cell density. Furthermore, the cell coalescence at higher temperature was more prominent, which may also contribute to the decrease of cell density and increase of cell size. stiffness of the polymer−gas mixture. On the other hand, with increasing temperature the cell density first increased to reach a maximum value and then decreased thereafter. Kumar et al.44 and Krause et al.10 observed a similar behavior in their studies of foaming of several polymers. The phenomenon can be readily understood by the competing effects of temperature on cell nucleation and competition (for gas) between nucleation and growth of bubbles. Classical nucleation theory45,46 was widely used in foaming to illustrate experimental phenomena due to its simplicity. According to classical nucleation theory, the homogeneous nucleation rate is expressed as * /kT ) Nhom = f0 C0exp( −ΔG hom (5) where f 0 is the frequency factor representing the frequency that gas molecules join the embryo nucleus, C0 is the concentration of the gas molecules which decreases with increasing temperature, and T is the foaming temperature. The homogeneous nucleation free energy ΔGhom * ∝ γ3bp, where γbp is the interfacial tension at the gas bubble polymer interface E dx.doi.org/10.1021/ie400833e | Ind. Eng. Chem. Res. XXXX, XXX, XXX−XXX Industrial & Engineering Chemistry Research Article Figure 6. SEM micrographs of COC saturated at 5 MPa and foamed at 120 (a), 130 (b), 140 (c), 150 (d), 160 (e), and 170 °C (f). Samples were saturated at higher pressures (6 and 7 MPa) and foamed over the same temperature range, and cell size and cell density were again analyzed and are shown in Figure 5. For both pressures, the temperature dependencies of cell size and cell density followed a similar trend observed previously, but they were significantly less sensitive to the temperature increase. Weller et al.14 observed similar phenomena in their study of solid-state foaming of polycarbonate. In their study the cell density under high saturation pressure was independent of temperature. They suggested that unlike foaming in a polymer melt where the temperature dependency of the cell nucleation rate followed the Arrhenius relationship, cell nucleation in solid-state foaming was fundamentally different. Under high saturation pressure the rate would reach a plateau value solely dependent on pressure. The stronger temperature dependency of cell size at lower pressure was also observed in their study and explained from the standpoint of bubble growth. To maintain the bubble of a certain size, a pressure differential between the growing bubble and the surrounding polymer−gas mixture is necessary, whose value was proportional to 1/r F dx.doi.org/10.1021/ie400833e | Ind. Eng. Chem. Res. XXXX, XXX, XXX−XXX Industrial & Engineering Chemistry Research Article Figure 7. SEM micrographs of COC saturated at 7 MPa and foamed at 120 (a), 130 (b), 140 (c), 150 (d), 160 (e), and 170 °C (f). bubble size would be smaller due to the higher required pressure differential. In summary, our study on the temperature effects on the cell morphology in solid-state foaming of COC agreed well with Weller et al.14 on the study of solid-state foaming of polycarbonate, suggesting that these effects are not system specific and are determined by the foaming mechanisms. It is also noteworthy that when a saturation pressure of 7 MPa was employed, ultramicrocellular foams with a cell size of ∼0.3−1 μm and cell density of ∼1012 cells/cm3 can be following the Laplace equation, where r is the radius of the bubble. Therefore, a lower pressure differential is needed when the bubble size is larger. This is the case for foaming at lower saturation pressures. When the temperature is increased by some increment, a relatively large change in the bubble size can be achieved since the bubble can grow more easily. On the other hand, bubble size is much smaller at higher saturation pressure, due to the simple fact that many more bubbles grow at the same time as a result of enhanced cell nucleation. Therefore, for the same temperature increment, the change of G dx.doi.org/10.1021/ie400833e | Ind. Eng. Chem. Res. XXXX, XXX, XXX−XXX Industrial & Engineering Chemistry Research Article Figure 8. Average foam cell density (left) and cell diameter (right) as a function of saturation pressure at different foaming temperatures. Lines are to assist reading. Lines are not drawn if only two data points are available. former and weak tensile stress field as a result of a lack of concentration gradient. 3.2.4. Effect of Saturation Pressure. Cell size and cell density were replotted as functions of saturation pressure for the series of foaming temperatures (Figure 8). Upon increasing the saturation pressure from 5 to 6 MPa, cell density increased about 2 orders of magnitude (from 109 to 1011 cell/cm3) over the entire foaming temperature range. This can be readily understood by the homogeneous nucleation theory (eq 5). When the saturation pressure was increased, the gas solubility increased, enhancing nucleation. Furthermore, dissolved CO2 is well known to reduce the surface tension.23,49 This would lead to a reduction of the nucleation free energy and significantly enhance the nucleation rate. As significantly more bubbles grow simultaneously, growth of each bubble is limited by the available gas, leading to a substantially smaller size as shown in Figure 8b. However further increasing the saturation pressure from 6 to 7 MPa had little impact on cell size and cell density. This does not follow the prediction from homogeneous nucleation theory and is in contrast with previous results on solid-state foaming of polycarbonate14 and poly(ether imide) (PEI),19,20 where in the latter case transition from micrometer to nanosized bubbles occurred upon increasing saturation pressure. To further investigate the pressure effect, samples were saturated at 10 MPa and foamed. Unexpectedly, instead of cellular structure, many lines that were orthogonal to the thickness direction were observed in the samples for foaming temperatures as low as 60 °C (Figure 9). High-magnification micrographs revealed that the lines were similar to the crazes observed by Weller et al.14 in PC foams. Furthermore, cracks, which presumably evolved from the crazes, were observed. Similar morphology containing crazes or cracks was also observed with foamed samples saturated at 9 and 8 MPa (micrographs not shown here). It appeared that when the saturation pressure was increased beyond 7 MPa, there was a fundamental change in the system behavior in that upon heating the system may transition from foaming to crazing. This transition may be heat-temperature dependent. To further study and verify this phenomenon, morphology evolution was investigated by heating the samples, produced over a very broad temperature range below the Tg of the pure polymer. 3.2.3. Cell Anisotropy. SEM micrographs of the foams prepared with saturation pressures of 5 and 7 MPa are shown in Figures 6 and 7, respectively. Samples saturated at 7 MPa showed anisotropic cell structure when foamed at high foaming temperature, where cells were elongated along the thickness direction of the sample (these micrographs are the crosssection view of the foam sample cryo-fractured parallel to the width direction). Anisotropic cellular structures were reported in several studies. Arora et al.12 prepared foams in a tube, during which foam expansion was constrained radially (by the tube wall) and favored axially (free expansion). They attributed the geometric confinement responsible for the observed cell anisotropy. Zirkel et al.47 discussed the influence of sample geometry on the foaming process and cell anisotropy. They suggested that for samples with disparate dimensions such as a thin sheet or film, the resistance against bubble growth in plane was higher than that for out of plane (thickness). This resulted in faster expansion in the thickness direction and elongated cells. Antunes et al.48 suggested that cell anisotropy resulted from preferential cell growth in the direction of pressure release. Herein we argue that the observed cell anisotropy is the result of a new mechanism as follows. The COC polymer used in the present study was severely plasticized by carbon dioxide,10,43 and the plasticizing effect became more prominent with increasing pressure. When highly plasticized samples were subjected to high temperature during foaming, rapid diffusion of the CO2 from the polymer matrix to the environment (a competing process against bubble growth) led to development of a distinct gas concentration gradient. As we shall discuss in great detail in section 3.3.2, this concentration gradient would lead to a localized tensile field developed along the gas diffusion direction (the thickness direction of the thin sheet sample) and elongation of cells in that direction. The process was facilitated by a weak matrix and rapid establishment of the concentration gradient. The absence of cell anisotropy in foams prepared at lower pressure (5 MPa) and lower temperatures may be attributed to the less prominent matrix plasticization in the H dx.doi.org/10.1021/ie400833e | Ind. Eng. Chem. Res. XXXX, XXX, XXX−XXX Industrial & Engineering Chemistry Research Article Figure 9. SEM of COC samples saturated with 10 MPa and heated at 60 (a), 100 (b), 110 (c), 120 (d), and 130 °C (e) showing extensive parallel lines that are crazes and cracks. An example is shown in f, which is the higher magnification view of the marked region of a. investigated time range, the bubble density and size increased with increasing heating time. As discussed previously, at a saturation pressure of 7 MPa, the eventual prevailing morphology is foam bubbles (for example, see Figure 7). However, further investigation showed that crazing may also take place at this saturation pressure. An example is shown in Figure 11. Crazes appeared soon after the sample was heated, whereas bubble appeared at a later time. Moreover, in the vicinity of the crazes large areas of unfoamed which were saturated at three different CO2 pressures (5, 7, and 10 MPa), for different amounts of time and their morphology observed by SEM (Figures 10−12). Cellular structure was formed in samples saturated at 5 MPa (Figure 10). Bubbles first appeared in the outer region of the sample when the heating time was increased to 2 s, since the temperature there reached the effective glass transition temperature earlier as the direction of thermal transfer was from the boundary to the interior of the polymer. Within the I dx.doi.org/10.1021/ie400833e | Ind. Eng. Chem. Res. XXXX, XXX, XXX−XXX Industrial & Engineering Chemistry Research Article Figure 10. SEM micrographs of COC saturated at 5 MPa and heated at 150 °C for various heating times: 2 (a), 4 (b), 6 (c), and 10 s (d). Bubble nucleation and formation of cellular structure is evident. the other hand, a transition to cellular cavitation was observed when the temperature was close to and above Tg. They attributed this to the different mechanisms for advancement of the craze tip, i.e., a switch from interface convolution at low stresses to cellular cavitation at high stresses. In this section we set out to analyze the transition phenomenon in detail and explore the possible physical processes involved at different pressures to provide more indepth understanding of the mechanisms. 3.3.1. Foaming under Low Saturation Pressure. Cell nucleation in solid-state foaming of amorphous polymer has been satisfactorily explained by the classical homogeneous nucleation. When the polymer is removed from the highpressure environment, the dissolved gas becomes oversaturated and the polymer−gas mixture is in thermodynamically nonequilibrium state. The oversaturation is the driving force for homogeneous bubble nucleation during which fluctuation of gas clusters and aggregation thereof lead to formation of viable nuclei. This is followed by diffusion of gas into the nuclei and bubble growth. However, these processes are kinetically prohibited when the temperature is lower than the effective Tg of the polymer−gas mixture due to the rigidity of the matrix. Alternatively, the gas can also slowly diffuse out of the matrix. When the system is heated to a temperature higher than Tg, the polymer chains gain sufficient mobility to reorient and accommodate bubble nucleation and growth. Such mechanism polymer were present, as shown in the Figure 11c. Apparently the crazing process not only relieved the internal stress and decreased the system free energy but also by phase separation it also depleted the nearby gas and suppressed bubble nucleation and growth in the vicinity. When the saturation pressure was further increased to 10 MPa, cellular morphology completely disappeared and only extensive crazes or cracks were present (Figure 12). Initially nanosized voids were developed (Figure 12a), and as time elapsed they grew in size and were increasingly elongated in the thickness direction, which led to eventual formation of fibrils between the voids (Figure 12b) and the craze structure whereby the two surfaces were bridged by a network of fibrils of drawn polymer (Figure 12c). The process resembles remarkably that of a brittle failure of glassy polymers under uniaxial tensile stress, and the craze structures formed also share strong similarity. 3.3. Analysis of Foaming and Crazing Process. The foaming investigation conducted earlier suggested that when subjected to heating the COC−CO2 system may transition from foaming to crazing, and the transition is largely dictated by the saturation pressure. Investigation on foaming and crazing transition is scarce in the literature. In one report Argon et al.50 studied the mechanism of crazing growth in glassy polymer using the uni- or multiaxial stress experiments and found that crazing took place when the polymer was in the glass state. On J dx.doi.org/10.1021/ie400833e | Ind. Eng. Chem. Res. XXXX, XXX, XXX−XXX Industrial & Engineering Chemistry Research Article Figure 11. SEM micrographs of COC saturated at 7 MPa and heated at 170 °C for various heating times: 3 (a), 5 (b), 7 (c), and 7 s (d). Scant crazes were formed at an earlier time (a and b) than bubble appearance. Terminal morphology is dominantly cellular morphology (d) with scant crazes (c) in the sample. Note that c and d were taken from different regions of the same sample. was commonly adopted to explain solid-state foaming,10,51 and in the current study, the behavior of the COC 6017 at low saturation pressure (5 MPa) also follows well this mechanism. 3.3.2. Crazing under High Saturation Pressure. We now discuss a physical phenomenon and its influence in solid-state foaming that has received less attention: stress field development within the polymer−gas mixture. When the gas-saturated polymer is removed from the high-pressure environment, a pressure differential between the specimen and the ambient is established. Since polymer chain relaxation is prohibited (due to matrix rigidity) to relieve this internal stress, such pressure differential creates a uniform triaxial tension or negative hydrostatic pressure in the specimen. The magnitude of this triaxial tension would increase with increasing gas concentration and temperature. In addition, it is well known that polymers are swollen by the dissolved gas. This generates dilation strain or volumetric strain that is proportional to the gas concentration. After the sample is removed from the high-pressure environment, gas desorption take places, resulting in development of a concentration and fugacity profile that decreases from the center toward the surface of the sample. For the thin sheet sample used in the current study, gradients are steepest in the thickness direction. The characteristics of these gradients are primarily determined by the gas diffusivity that itself is concentration, pressure, and temperature dependent.52 Similarly, the resulting volumetric strain also decreases from the interior to the exterior in the thickness direction. This nonuniform strain distribution would give rise to internal stresses,53 and the interior of the polymer will exert extensive stress to its outer portion. To the first order of approximation this internal stress profile can be considered as a pseudotensile stress field in the thickness direction. A higher saturation pressure results in a higher gas concentration in the polymer and higher gas diffusivity (due to a higher degree of plasticization54), leading to a steeper concentration gradient and larger pseudotensile stress. To summarize the above analysis, the stress profile within the polymer−gas mixture is comprised of a triaxial tension field and a pseudotensile stress field (in the thickness direction). The two types of stress field are schematically shown in Scheme 1. Both types of stress fields are more intensified in samples saturated at higher pressure. It shall be noted that such concept on the stress field evolution in gas swollen polymers was fairly extensively studied for elastomers, and the stresses were attributed to be the major source of gas-induced damage observed in those materials.53,55−58 In particular, a detailed experimental and analytical investigation by Briscoe et al. showed that the combined effects of the triaxial tension and pseudotensile stress field could result in rupture of elastomer.53 K dx.doi.org/10.1021/ie400833e | Ind. Eng. Chem. Res. XXXX, XXX, XXX−XXX Industrial & Engineering Chemistry Research Article Figure 12. SEM micrographs of COC saturated at 10 MPa and heated at 130 °C for various heating times showing crazes formation and growth: 2 (a), 6 (b), and 10 s (c). Circle in a indicates the nanosized voids. (d) Low-magnification view of c, where the square mark indicates the region where image c is taken from. Note the extensive parallel lines consisting of crazes. Scheme 1. Triaxial Tension and Tensile Stress Development Resulting from the Variation of CO2 Concentration (C) and Resulting Variation of Pressure (P) and Dilative Strain (ε) plasticization, the chain mobility of the polymer is significantly enhanced and matrix strength reduced. Furthermore, the small CO2 molecules can wet the surface and reduce the energy required to create new surfaces in the polymer matrix, which is beneficial to formation of new phase, such as cavitation.59 When all these are taken into account, it can be argued that It is plausible that such combined effects of the triaxial tension and pseudotensile stress field are responsible for the craze formation observed in the current study. When the polymer is saturated under higher pressure, the higher CO2 concentration results in significantly higher triaxial tension in the polymer−gas mixture. Additionally, because of intensive L dx.doi.org/10.1021/ie400833e | Ind. Eng. Chem. Res. XXXX, XXX, XXX−XXX Industrial & Engineering Chemistry Research Article of the COC−CO2 solution as a function of CO2 pressure, and the prediction is shown in Figure 13. The system exhibits when materials are saturated under high pressure, the triaxial tension may have surpassed a critical level and cavity or void nucleation readily proceeds even at temperatures below the effective Tg, which is the minimum temperature for foaming to occur. The voids can grow and coalesce when the sample is subject to heating, during which development of both stress and strain fields may be accelerated as a result of rapid diffusion and matrix softening. At the same time, the voids are stretched and increasingly elongated in the thickness direction because of the evolving pseudotensile stress field, forming highly oriented polymer fibrils between the voids. Note that this process is very similar to the crazing of brittle polymer under uniaxial tension, and the final craze structure consisted of the two surfaces bridged by a network of fibrils of drawn polymer. Thus, the craze structures observed in the current study can be satisfactorily explained by the mechanisms discussed above, as those samples were indeed saturated under higher pressure, which would promote nucleation and propagation of crazes during heating. Furthermore, the crazing process relieves the internal stress and depletes the gas dissolved in the bulk. The prominent nucleation and propagation of crazes (in systems saturated under high pressures) would quickly exhaust the available gas. Consequently, bubble nucleation may be totally suppressed and cellular morphology prohibited. This was the case for samples saturated at 10 MPa. Crazing in solid-state foaming has been reported,14 but to our knowledge, the current study is the first to report the foaming−crazing transition phenomenon. The occurrence is due to the unique combination of several factors in the present systems that work in concert and lead to the behavior. First, the COC−CO2 possesses a strong interaction that enables substantial CO2 solubility but also the polymer matrix is strongly plasticized and glass transition temperature severely depressed. Indeed, Kraus et al.10 showed in the three polymers they studied (COC, poly(ether sulfone) (PES), and polysufone (PSU)), with the same amount of dissolved CO2, COC was the most severely plasticized and showed the highest degree of Tg depression. For the COC charged with between 3.9 and 7.8 wt % CO2, the plasticization index was −14.24 °C/wt % CO2. By comparison, for polystyrene, a polymer with moderate interaction with CO2, the index was only −6.8 °C/wt % CO2. In their study a COC with lower norbornene content (Topas COC6013) was used. The higher norbornene content in COC 6017 used in the current study gives rise to even higher CO2 solubility when compared to the COC 601343 and hence even higher Tg depression. Second, the strongly plasticized polymer matrix shall have greatly reduced strength, so that upon heating the stress fields (in the thickness direction) developed quickly and exceed the threshold and the matrix in essence goes through the brittle failure similar to when subjected to uniaxial tension. The lack of strength in extremely plasticized COC appears to be supported by Gendron et al,33 where they showed that at temperatures significantly higher than the glass transition temperature the COC did not show appreciable melt strength and strain hardening. To further examine the COC 6017−CO2 system behavior, we measured CO2 solubility in COC 6017 at temperatures higher than the pure polymer Tg and modeled the behavior by the Sanchez−Lacombe equation of state (SL EoS). Following the thermodynamic framework developed by Condo et al.60 and applying the Gibbs−DiMarzio thermodynamic criterion for glass transition, we calculated the glass transition temperature Figure 13. Prediction of Tg of COC 6017−CO2 solution vs CO2 pressure showing retrograde vitrification phenomenon. Sample was saturated at 330.15 K (57 °C) and 13.8 MPa (indicated by the point in the figure) for 48 h, after which the pressure was quickly released. Resulting sample (photograph shown as the insert) contains a massive amount of lines that consist of crazes and cracks. This provides direct support to the proposed mechanisms on the role of foaming nucleation and crazing discussed in the text. retrograde behavior,60,61 where the system exhibits two glass transition temperatures. With decreasing temperature the system transitions from rubbery to glassy state and once again to rubbery state. The second transition is a direct consequence of the strong polymer−CO2 interaction and the significantly enhanced CO2 solubility in the polymer as temperature is lowed. As a result, the polymer chain is heavily plasticized and gains sufficient mobility that the system transitions from the glassy to the rubber state. Such phenomenon has been observed in several polymer−CO2 systems that show strong interactions, and ultramicro- or nanosized foams were prepared62−64 enabled by the exceptionally high CO2 solubility and the existence of a rubbery state at low temperature. To test whether the high CO2 solubility and substantially reduced matrix strength are responsible for the observed crazing, a sample was saturated at 57 °C and 13.8 MPa for 48 h before releasing the pressure. According to the model prediction (indicated in Figure 13), these conditions reside outside the retrograde envelop and the temperature is higher than the glass transition temperature of the polymer−gas mixture (T > Tg). However, after the pressure was released and sample taken out of the pressure vessel, no foaming was observed. Instead, a massive amount of macroscopic lines were observed in the sample (insert of Figure 13). The lines were cracks that can only result from growth of crazes. Under the saturation condition used, the COC matrix plasticization is extremely severe and hence the severe reduction of matrix strength. In addition, an arguably enormous stress field is developed after the sample was taken out of the pressure vessel because of the extremely high gas solubility and enhanced diffusivity. The experimental conditions used herein thus exemplify the driving forces for crazing to an extreme extent. M dx.doi.org/10.1021/ie400833e | Ind. Eng. Chem. Res. XXXX, XXX, XXX−XXX Industrial & Engineering Chemistry Research Article Scheme 2. (a) Bubble Formation from Homogeneous Nucleation under Low Saturation Pressure; (b) Crazing Formation from the Cooperative Effects of Triaxial Tension and Tensile Stress Field under High Saturation Pressure; (c) Simultaneous Formation of Bubble and Crazing under Medium Saturation Pressure described well by the dual-sorption model. Systematic investigation of the effects of the process parameters, e.g., foaming temperature, foaming time, and saturation pressure, was conducted to tailor microcellular foam structures. The resulting foam structures which had average cell diameter in the range of 0.3−7.5 μm and cell density on the order of 109−1012 cells/cm3 can be controlled by manipulating processing conditions. An ultramicrocelluar foam with a cell size of ∼0.3 μm and cell density of ∼1012 cells/cm3 can be obtained at saturation pressures of 6 and 7 MPa. Solid-state foaming of COC was found to be dictated by two competing processes, bubble nucleation vs crazing. The crazing process was analyzed in detail based on the stress field development. The triaxial tensile stress field in the gas− polymer mixture resulted from the volumetric swelling or dilation strain giving rise to nucleation of voids, and the pseudotensile stress field from diffusion resulted in a gas concentration gradient that facilitates growth of the crazes and cracks. As a result of the strong COC−CO2 interaction and accompanied depression of the glass transition temperature and reduction of the matrix rigidity, crazing becomes more prominent with increasing saturation pressure and eventually completely dominates. The significant plasticization of the polymer sample at high saturation pressure facilitated the crazing process at temperatures below the effective glass transition temperature, which is the minimum temperature required for foaming. That under these conditions the nucleation for crazing completely dominates and foaming has been completely suppressed in the sample provides direct experimental support for the discussed mechanisms. 3.3.3. Coexistence of the Foaming and Crazing Process under Medium Pressure. To summarize sections 3.3.1 and 3.3.2, both foaming and crazing are possible in solid-state foaming of COC. Foaming is favored when the saturation pressure is low and suppressed when the saturation pressure is high. Crazing, on the other hand, has an opposite pressure dependency. It is conceivable that in the intermediate-pressure regimes both the foaming and the crazing processes may be possible. On one hand, voids and crazes may still form because of the stress fields. On the other hand, the less severe stress fields lead to fewer and discrete crazes, whose growth is slow with limited consumption of gas. Thus, bubble nucleation and growth are still possible when the effective Tg is reached, which would compete for the available gas. Furthermore, some of the voids resulting from the triaxial tension may not be able to grow into crazes before the effective Tg is attained. As discussed by Holl et al.,27 they would instead serve as nuclei for bubble growth. These three different processes proceeding under different saturation pressure are schematically illustrated in Scheme 2. 4. CONCLUSIONS In this study, we investigated the solid-state batch foaming of a cycloolefin copolymer (COC) using carbon dioxide. The gas sorption properties that are critical for foaming were characterized first. It was found that gas sorption can be N dx.doi.org/10.1021/ie400833e | Ind. Eng. Chem. Res. XXXX, XXX, XXX−XXX Industrial & Engineering Chemistry Research ■ Article (20) Miller, D.; Chatchaisucha, P.; Kumar, V. Microcellular and nanocellular solid-state polyetherimide (PEI) foams using sub-critical carbon dioxide I. Processing and structure. Polymer 2009, 50, 5576− 5584. (21) Sun, H.; Sur, G.; Mark, J. Microcellular foams from polyethersulfone and polyphenylsulfone - Preparation and mechanical properties. Eur. Polym. J. 2002, 38, 2373−2381. (22) Wang, D.; Jiang, W.; Gao, H.; Jiang, Z. Preparation, characterization, and mechanical properties of microcellular poly(aryl ether ketone) foams. J. Polym. Sci., Part B: Polym. Phys. 2007, 45, 173− 183. (23) Goel, S.; Beckman, E. Nucleation and growth in microcellular materials - supercritical CO2 as foaming agent. AlChE J. 1995, 41, 357−367. (24) Colton, J. S.; Suh, N. P. The nucleation of microcellular thermoplastic foam with additives. Part I. Theoretical considerations. Polym. Eng. Sci. 1987, 27, 485−92. (25) Colton, J. S.; Suh, N. P. The nucleation of microcellular thermoplastic foam with additives. Part II. Experimental results and discussion. Polym. Eng. Sci. 1987, 27, 493−9. (26) Colton, J. S.; Suh, N. P. Nucleation of microcellular foam: theory and practice. Polym. Eng. Sci. 1987, 27, 500−3. (27) Holl, M.; Kumar, V.; Garbini, J.; Murray, W. Cell nucleation in solid-state polymeric foams: Evidence of a triaxial tensile failure mechanism. J. Mater. Sci. 1999, 34, 637−644. (28) Campion, R. Explosive decompression in elastomers - internal blistering and fracturing in rubbers after high-pressure exposure to gases. Cell. Polym. 1990, 9, 206−228. (29) Weller, T.; Hatke, W. Cycloolefin Copolymers. In Encyclopedia of Materials: Science and Technology, 2nd ed.: Buschow, K. H. J., Cahn, R. W., Flemings, M. C., Ilschner, B., Kramer, E. J., Mahajan, S., Veyssière, P., Eds.; Elsevier: Oxford, 2001; pp 1963−1965. (30) Shin, J.; Park, J.; Liu, C.; He, J.; Kim, S. Chemical structure and physical properties of cyclic olefin copolymers - (IUPAC technical report). Pure Appl. Chem. 2005, 77, 801−814. (31) Lamonte, R.; McNally, D. Uses and processing of cyclic olefin copolymers. Plast. Eng. 2000, 56, 51−55. (32) Lamonte, R.; McNally, D. Cyclic olefin copolymers. Adv. Mater. Proc. 2001, 159, 33−36. (33) Gendron, R.; Champagne, M.; Tatibouet, J.; Bureau, M. Foaming Cyclo-Olefin Copolymers With Carbon Dioxide. Cell. Polym. 2009, 28, 1−24. (34) Sun, X.; Liu, H.; Li, G.; Liao, X.; He, J. Investigation on the cell nucleation and cell growth in microcellular foaming by means of temperature quenching. J. Appl. Polym. Sci. 2004, 93, 163−171. (35) Sun, X.; Li, G.; Liao, X.; He, J. An investigation on the cell nucleation and cell growth of microcellular foaming by means of dual depressurization. Acta Polym. Sin. 2004, 93−97. (36) Sun, X.; Liu, C.; Yu, J.; Li, G.; He, J. Preparation and characterization of microcellular cyclic olefin copolymer blends using supercritical CO2. Acta Polym. Sin. 2004, 132−136. (37) Gendron, R.; Chaudhary, M. In Challenging the Paradigm of Microcellular Foams: Mechanical Properties of Low-Density Cyclo-Olefin Copolymer Foams, Chicago, IL, 2009; SPE ANTEC: Chicago, IL, 2009; pp 920−924. (38) Wang, D.; Jiang, W.; Gao, H.; Jiang, Z. Controlling Cellular Morphology by Supercritical Carbon Dioxide. J. Appl. Polym. Sci. 2009, 111, 2116−2126. (39) Pantoula, M.; Panayiotou, C. Sorption and swelling in glassy polymer/carbon dioxide systems - Part I. Sorption. J. Supercrit. Fluids 2006, 37, 254−262. (40) Zeng, C.; Hossieny, N.; Zhang, C.; Wang, B. Synthesis and processing of PMMA carbon nanotube nanocomposite foams. Polymer 2010, 51, 655−664. (41) Zhou, C.; Wang, P.; Li, W. Fabrication of functionally graded porous polymer via supercritical CO2 foaming. Composites , Part B: Eng. 2011, 42, 318−325. AUTHOR INFORMATION Corresponding Author *E-mail: zeng@eng.fsu.edu (C.Z.); yaozhen@zju.edu.cn (Z.Y.). Notes The authors declare no competing financial interest. ACKNOWLEDGMENTS Funding from the Center of Excellence in Advanced Materials by the State of Florida is acknowledged. ■ ■ REFERENCES (1) Martini, J. The Production and Analysis of Microcellular Foam. PhD Dissertation, Massachusetts Institute of Technology, 1981. (2) Martini, J.; Suh, N.; Waldman, F. Microcellular closed cell foams and their method of manufacture. US4473665 A, 1984. (3) Kumar, V. Microcellular polymers - novel materials for the 21stcentury. Cell. Polym. 1993, 12, 207−223. (4) Fu, J.; Jo, C.; Naguib, H. Effect of processing parameters on cellular structures and mechanical properties of PMMA microcellular foams. Cell. Polym. 2005, 24, 177−195. (5) Matuana, L.; Park, C.; Balatinecz, J. Structures and mechanical properties of microcellular foamed polyvinyl chloride. Cell. Polym. 1998, 17, 1−16. (6) Collias, D.; Baird, D.; Borggreve, R. Impact toughening of polycarbonate by microcellular foaming. Polymer 1994, 35, 3978− 3983. (7) Bureau, M.; Kumar, V. Fracture toughness of high density polycarbonate microcellular foams. J. Cell. Plast. 2006, 42, 229−240. (8) Shimbo, M.; Baldwin, D.; Suh, N. The viscoelastic behavior of microcellular plastics with varying cell-size. Polym. Eng. Sci. 1995, 35, 1387−1393. (9) Ruiz, J.; Saiz-Arroyo, C.; Dumon, M.; Rodriguez-Perez, M.; Gonzalez, L. Production, cellular structure and thermal conductivity of microcellular (methyl methacrylate) - (butyl acrylate) - (methyl methacrylate) triblock copolymers. Polym. Int. 2011, 60, 146−152. (10) Krause, B.; Mettinkhof, R.; van der Vegt, N.; Wessling, M. Microcellular foaming of amorphous high-T-g polymers using carbon dioxide. Macromolecules 2001, 34, 874−884. (11) Krause, B.; Kloth, M.; van der Vegt, N.; Wessling, M. Porous monofilaments by continuous solid-state foaming. Ind. Eng. Chem. Res. 2002, 41, 1195−1204. (12) Arora, K.; Lesser, A.; McCarthy, T. Preparation and characterization of microcellular polystyrene foams processed in supercritical carbon dioxide. Macromolecules 1998, 31, 4614−4620. (13) Park, C.; Behravesh, A.; Venter, R. Low density microcellular foam processing in extrusion using CO2. Polym. Eng. Sci. 1998, 38, 1812−1823. (14) Weller, J.; Kumar, V. Solid-State Microcellular Polycarbonate Foams. I. The Steady-State Process Space Using Subcritical Carbon Dioxide. Polym. Eng. Sci. 2010, 50, 2160−2169. (15) Juntunen, R.; Kumar, V.; Weller, J. Impact strength of high density microcellular poly(vinyl chloride) foams. J. Vinyl Addit. Technol. 2000, 6, 93−99. (16) Collias, D.; Baird, D. Tensile toughness of microcellular foams of polystyrene, styrene-acrylonitrile copolymer, and polycarbonate, and the effect of dissolved-gas on the tensile toughness of the same polymer matrices and microcellular foams. Polym. Eng. Sci. 1995, 35, 1167−1177. (17) Murray, R.; Weller, J.; Kumar, V. Solid-state microcellular acrylonitrile-butadiene-styrene foams. Cell. Polym. 2000, 19, 413−425. (18) Kumar, V.; Juntunen, R.; Barlow, C. Impact strength of high relative density solid state carbon dioxide blown crystallizable poly(ethylene terephthalate) microcellular foams. Cell. Polym. 2000, 19, 25−37. (19) Krause, B.; Sijbesma, H.; Munuklu, P.; van der Vegt, N.; Wessling, M. Bicontinuous nanoporous polymers by carbon dioxide foaming. Macromolecules 2001, 34, 8792−8801. O dx.doi.org/10.1021/ie400833e | Ind. Eng. Chem. Res. XXXX, XXX, XXX−XXX Industrial & Engineering Chemistry Research Article (42) Zhao, J.; Zhao, Y.; Yang, B. Investigation of sorption and diffusion of supercritical carbon dioxide in polycarbonate. J. Appl. Polym. Sci. 2008, 109, 1661−1666. (43) Hu, C.; Lee, K.; Ruaan, R.; Jean, Y.; Lai, J. Gas separation properties in cyclic olefin copolymer membrane studied by positron annihilation, sorption, and gas permeation. J. Membr. Sci. 2006, 274, 192−199. (44) Kumar, V.; Weller, J. Production of microcellular polycarbonate using carbon-dioxide for bubble nucleation. J. Eng. Ind.-T ASME 1994, 116, 413−420. (45) Abraham, F. F., Homogeneous Nucleation Theory. The Pretransition Theory of Vapor Condensation. In Advances in Theoretical Chemistry, Suppl. 1; Academic Press: New York, 1974. (46) Laaksonen, A.; Talanquer, V.; Oxtoby, D. W. Nucleation: Measurements, Theory, and Atmospheric Applications. Annu. Rev. Phys. Chem. 1995, 46, 489−524. (47) Zirkel, L.; Jakob, M.; Munstedt, H. Foaming of thin films of a fluorinated ethylene propylene copolymer using supercritical carbon dioxide. J. Supercrit. Fluids 2009, 49, 103−110. (48) Antunes, M.; Velasco, J.; Realinho, V.; Solorzano, E. Study of the Cellular Structure Heterogeneity and Anisotropy of Polypropylene and Polypropylene Nanocomposite Foams. Polym. Eng. Sci. 2009, 49, 2400−2413. (49) Tomasko, D.; Li, H.; Liu, D.; Han, X.; Wingert, M.; Lee, L.; Koelling, K. A review of CO2 applications in the processing of polymers. Ind. Eng. Chem. Res. 2003, 42, 6431−6456. (50) Argon, A.; Salama, M. Growth of crazes in glassy polymers. Philos. Mag. 1977, 36, 1217−1234. (51) Wang, D.; Gao, H.; Jiang, W.; Jiang, Z. Microcellular processing and relaxation of poly(ether ether ketone). J. Polym. Sci., Part B: Polym. Phys. 2007, 45, 2890−2898. (52) Briscoe, B.; Gritsis, D.; Liatsis, D. The Concentration and Pressure Dependent Diffusion of Carbon-Dioxide in Nitrile Rubbers. Philos. Trans. R. Soc. London, A 1992, 339, 497−519. (53) Briscoe, B.; Liatsis, D. Internal crack symmetry phenomena during gas-induced rupture of elastomers. Rubber Chem. Technol. 1992, 65, 350−373. (54) Lin, S.; Yang, J.; Yan, J.; Zhao, Y.; Yang, B. Sorption and Diffusion of Supercritical Carbon Dioxide in a Biodegradable Polymer. J. Macromol. Sci., Part B: Phys. 2010, 49, 286−300. (55) Briscoe, B.; Savvas, T.; Kelly, C. Explosive Decompression Failure of Rubbers - A Review of The Origins of Pneumatic StressInduced Rupture in Elastomers. Rubber Chem. Technol. 1994, 67, 384− 416. (56) Briscoe, B.; Zakaria, S. Gas-Induced Damage in Elastomeric Composites. J. Mater. Sci. 1990, 25, 3017−3023. (57) Denecour, R.; Gent, A. Bubble Formation in Vulcanized Rubbers. J. Polym. Sci., Part B:Polym. Phys. 1968, 6, 1853−&. (58) Gent, A. Cavitation in Rubber - A Cautionary Tale. Rubber Chem. Technol. 1990, 63, G49−G53. (59) O’Connell, P. A.; McKenna, G. B. Yield and Crazing in Polymers. In Encyclopedia of Polymer Science and Technology; John Wiley & Sons, Inc.: 2002. (60) Condo, P.; Sanchez, I.; Panayiotou, C.; Johnston, K. Glasstransition behavior including retrograde vitrification of polymers with compressed fluid diluents. Macromolecules 1992, 25, 6119−6127. (61) Kikic, I.; Vecchione, F.; Alessi, P.; Cortesi, A.; Eva, F.; Elvassore, N. Polymer plasticization using supercritical carbon dioxide: Experiment and modeling. Ind. Eng. Chem. Res. 2003, 42, 3022−3029. (62) Handa, Y.; Zhang, Z. A new technique for measuring retrograde vitrification in polymer-gas systems and for making ultramicrocellular foams from the retrograde phase. J. Polym. Sci., Part B: Polym. Phys. 2000, 38, 716−725. (63) Handa, Y.; Zhang, Z.; Wong, B. Solubility, diffusivity, and retrograde vitrification in PMMA-CO2, and development of submicron cellular structures. Cell. Polym. 2001, 20, 1−16. (64) Nawaby, A.; Handa, Y.; Liao, X.; Yoshitaka, Y.; Tomohiro, M. Polymer-CO2 systems exhibiting retrograde behavior and formation of nanofoams. Polym. Int. 2007, 56, 67−73. P dx.doi.org/10.1021/ie400833e | Ind. Eng. Chem. Res. XXXX, XXX, XXX−XXX