THERMOPLASTIC NANOCELLULAR FOAMS WITH LOW RELATIVE DENSITY USING CO 2 AS THE BLOWING AGENT

Size: px

Start display at page:

Download "THERMOPLASTIC NANOCELLULAR FOAMS WITH LOW RELATIVE DENSITY USING CO 2 AS THE BLOWING AGENT"

Mildred Simpson
6 years ago
Views:

1 SPE FOAMS 2011 Conference September 2011 (Iselin, NJ) 1 THERMOPLASTIC NANOCELLULAR FOAMS WITH LOW RELATIVE DENSITY USING CO 2 AS THE BLOWING AGENT Stéphane Costeux, Dow Building Solutions, The Dow Chemical Company, Midland, MI Lingbo Zhu, Core R&D Materials Science, The Dow Chemical Company, Midland, MI Abstract It is expected than nano-cellular foams (defined as having cell size of about 200 nm or less) will possess optical, mechanical or thermal properties not observed with microcellular foams. For instance, the insulation performance of nanofoams should benefit from the Knudsen effect, the reduction of air thermal conductivity when cell size approaches the mean free path between air molecules (75 nm). Despite several attempts using thin films or engineered polymer templates, producing thick nanofoams by a gas-blowing process has remained elusive, in particular at low densities, due to the difficulty in nucleating and expanding a very large number of cells without coalescence. Here we report the successful production of thick, homogeneous nano-cellular foams with high pore volume by carbon dioxide foaming of thermoplastic polymers. Nanofoams with 100 nm average cell size, relative density of 0.15 and cell densities on the order of cells/cm 3 were made using an acrylic polymer and nanoscale particles to enhance cell nucleation. The broader validity of the approach is also exemplified for a styrenic polymer, and with different nanoparticles. Introduction Nanoporous structures have found numerous applications, from catalysis to filtration and electronic materials. The methods used to produce these structures with polymeric materials have typically involved an organic solvent, and relied on either a phase separation (membranes) or a gelation (aerogels and xerogels), followed by a removal of the solvent, a difficult step given the capillary forces generated during circulation in narrow channels, especially for thick samples. Yet, effective thermal insulation requires thicker cellular structures, at least a few millimeters thick. The finest nano-structured insulation materials used to-date are organic and inorganic aerogels, which have illustrated the benefit of having a nanoporous structure to exploit the Knudsen effect. Thermal conductivities as low as 10 mw/m-k were obtained for aerogels [1] with pore size in the range of 10 to 50 nm, i.e. smaller than the mean free path of air molecules at ambient temperature and pressure (ca. 75nm), combined with a low density (high void fraction or porosity close to 85-90% [1] ) which reduces the contributions on heat conduction via phonon transport through the solid. Gas foaming remains the most attractive method to produce porous materials at relatively low cost. Microcellular foams can be produced relatively easily in batch process, and a significant effort was devoted to the development of continuous processes. [2] Such foams have cell densities of about /cm 3. [Note: Cell density is commonly defined, as proposed by Kumar and Suh [3], as the final number of cells in the foam, in a volume corresponding to 1 cm 3 of the polymer prior to foaming.] Yet the benefit for insulation is rather insignificant when cells are in the micron range, as shown in Figure 1. Indeed, the theoretical gas contribution to thermal conductivity, k gas, of a foam is reduced about by half when the cell size approaches 200 nm, and cells need to be about 100 nm to see a three-fold reduction. k gas (mw/m.k) Gas thermal conductivity in air-filled foams 1/ nm Pore size (micron) Figure 1. Predicted reduction of gas thermal conductivity with cell size due to the Knudsen effect Generating and stabilizing a large number of cells at the nanoscale (ca. 100 nm) by a gas foaming process is a major challenge that is getting increased consideration in academia and industry. A major step was a very thorough study by Krause et al., [4] who showed how CO 2 foaming could be used with high Tg polymers to produce 100 m thick membranes with very fine cells ( nm). However, cells were only observed in the core of the samples, while a significant portion of the material close 1/2

2 SPE FOAMS 2011 Conference September 2011 (Iselin, NJ) 2 to the free surfaces was essentially unfoamed. As a result, the overall relative density was close to 1. Handa and Zhang [5] proposed to improve upon processes to make microcellular poly(methyl methacrylate) (PMMA) using high pressure CO 2 by making use of the retrograde vitrification behavior of PMMA, an approach also followed by Nawaby et al. [6] using ABS copolymers. These materials appear to develop a second, low Tg, under which the material becomes rubbery. The CO 2 soaking conditions are set in this narrow region, right above the liquid-gas line of CO 2, at about 0 C and 3.4 MPa. Nuclei are formed upon depressurization without visible expansion. Foaming is then produced by immersing the material in a heated water bath to produce ultra-microcellular foams with cell size around 0.5 microns and cell densities close to cells/cm 3. Ruckdäschel et al. [7] compared the cell densities and cell size achieved in work reported in the published literature before 2010, which reveals that few submicron foams have been produced with relative densities below 0.5, or with cell densities above /cm 3. Miller et al. [8] showed at the FOAMS 2009 conference polyesterimide foams with cell density of the order of /cm 3 and relative densities reaching 0.47 with rather broad cell size distribution around ca. 300 nm. At the FOAMS 2010 conference, Dumon et al. [9] showed how addition of a triblock acrylic copolymer could decrease cell size in PMMA, and they reported foams with 300nm cells and relative density of 0.7. [Note that the high cell densities reported in this reference are in error, as they exceed the theoretical limit for a given cell size by a large factor]. Thus, despite recent progress in the field, producing thick foams with smaller cells (ca. 100 nm) and higher cell densities (ca /cm 3 ) expected to be required for super insulation materials is a problem that has remained unsolved. Additionally, the presence of micron size pores in these foams could be detrimental to the performance, as was explained by Ohshima. [10] Thus homogeneous cell size distributions are preferred. By a proper choice of polymer, nucleating additive and foaming conditions, we solved these issues [11] to produce unique CO 2 -blown nanoporous materials with homogeneous cell size around nm. The approach proved successful not only in drastically improving upon the cell densities reported to-date by two orders of magnitude, but also in allowing controlled cell growth resulting in high porosities corresponding to relative density as low as Such nanofoams are good candidates for high performance insulation. Materials Experimental PMMA is a homopolymer powder of methyl methacrylate (Sigma-Aldrich, 120kg/mol, Tg=118 C). PMMA-co-EA is a low melt flow polymer with 9% ethyl acrylate randomly distributed (VM100 from Arkema, Tg=98 C). PMMA-co-EMA is a random copolymer of MMA with 50wt% ethyl methacrylate (360 kg/mol, Tg=96 C) obtained from SP2 (Ontario, NY). SAN is a styrene copolymer with 31wt% acrylonitrile (Dow Chemical Tyril 125, Tg=107 C). POSS is a methacryl-substituted polyhedral oligomeric silsesquioxane (Sigma-Aldrich). SiO 2 describes colloidal silica particles (5-7 nm) prepared by reaction of TEOS with ethanol and water (37/37/26 vol%) at 23 C and ph=2 (adjusted with 0.5M HCl). Gelation of the primary particles is obtained by adding NaOH to ph=6.5. Specimen Preparation Composites of PMMA and PMMA-co-EA with SiO 2 were obtained by adding the silica gel directly to the polymer powder, and melt-blended in a Haake Rheomix at 180 C for 10 min. POSS was first diluted to a 3wt% solution in ethanol, added to either SAN or PMMA-co- EMA, and melt-blended at 180 C. A nitrogen blanket was used to minimize ethanol vapors during compounding. All samples were pressed into 3 mm thick plaques by compression molding at 180 C. Foaming Foaming was carried out in an autoclave enclosed in an oven. A syringe pump (Teledyne ISCO 100D) provided CO 2 at constant pressure. Once the CO 2 sorption reached equilibrium, an actuated valve depressurized the autoclave in less than 1 second. After depressurization, the foamed sample was promptly removed from the autoclave and annealed (post-foaming, PF ) in a water bath for 3 min to complete degassing of CO 2 and release internal stresses, which resulted in a secondary expansion. Foam Characterization Density ρ f was measured on foams after complete CO 2 desorption (over 24 hrs after foaming) by the Archimedes method. Porosity was calculated as the void volume fraction p%=100 (1-ρ f /ρ p ), where ρ p is the polymer density (1.08 for SAN, and for the acrylic polymers). Average cell size,, was determined from high resolution SEM images, by measuring at least 200 representative cells on a foam cross section prepared by cryo-fracture.

SPE FOAMS 2011 Conference 12-15 September 2011 (Iselin, NJ) 3 Results For each polymer, a systematic exploration was conducted to identify the properties of the polymer-co 2 system (e.g.

3 SPE FOAMS 2011 Conference September 2011 (Iselin, NJ) 3 Results For each polymer, a systematic exploration was conducted to identify the properties of the polymer-co 2 system (e.g. CO 2 solubility, viscosity) and choose foaming pressure and temperature conditions that ensure sufficient CO 2 concentration to maximize the driving force towards nucleation. Nucleating additives with sizes less than 30 nm were then introduced, and selected for each polymer based on the ease of dispersion, and their effectiveness in increasing cell density. Post-foaming conditions were then adjusted to maximize porosity. To further increase cell density while achieving high porosity, a high molecular weight PMMA-co-EMA was foamed under conditions suitable to achieve CO 2 solubility in excess of 25 wt%. Foams with porosities as high as 85% and cell size of 120 nm or less were achieved after annealing at 70 C for 3 min. Smaller cells were produced when further reducing the foaming temperature to increase CO 2 solubility, and when lowering the annealing temperature to limit post-expansion. As for all nucleated examples in Table 1, the cell size and structure was homogeneous throughout the bulk of the 7-10 mm thick foams, except for a 5-10 μm thick integral skin. Table 1 shows the pressure/temperature conditions and results of a representative number of foaming experiments with PMMA, PMMA-co-EA, PMMA-co- EMA and SAN with various types and levels of nanoparticles, which illustrate the benefit of the approach. Foamed at 60 C and under a CO 2 pressure of 30 MPa, conditions at which CO 2 solubility is close to 20 wt%, PMMA without SiO 2 nanoparticles produced a bimodal foam with a large fraction of cells of average size 130μm and sub-micron cells in the material connecting these large cells. The addition of 3 wt%, 5 nm SiO 2 particles prior to foaming eliminated all large cells and produced foams with significantly reduced cell size, while maintaining a product porosity of 77%. The foams are homogeneous and isotropic, i.e. the cell distribution is narrow, does not depend on the orientation of the fracture surface or on the location within the foamed sample (apart from the 2-10 micron skin layer at the surface of the foams). Since the presence of the silica did not induce meaningful changes to the CO 2 solubility or viscosity, it is hypothesized that SiO 2 particles behave as very effective nucleating centers. This contrasts with findings by Ramesh et al. [12] that particles much smaller than about 300 nm are not very effective nucleators. The same strategy was applied to a random acrylic copolymer, PMMA-co-EA. Due to the absence of chemical inhomogeneity in these random copolymers, the polymers do not exhibit any phase separation, based on Small Angle X-Ray Scattering experiments prior to the addition of SiO2 nanoparticles. We found the level of SiO 2 could be reduced due to improved nucleation in the pure polymer at higher pressure and lower temperature. Optimization led to a foam with 82% porosity (0.18 relative density) and over cells/cm 3. The approach was extended to another random copolymer, SAN, for which a similar observation was made. Upon addition of 0.25 wt% POSS, a sub-nanometer silica cage-like molecule, the average cell size in SAN foams was reduced 10 times (to 200 nm) while maintaining 70 vol% voids in the foam. Fraction (%) Direct cell density count: - 2 μm x 2 μm section: cell count = 424; N a = 106 /μm 2 - Expansion ratio: R=100/(100-p%) = # cells/cm 3 of foam: N v=(n a) 3/2 =1.1x10 15 /cm 3 - # cells/cm 3 of unfoamed polymer (cell density): N=R * N v 7 x10 15 /cm 3 Number fraction (%) Volume fraction (%) Cells averaged: 1020 Cell size = 99 nm Porosity p = Cell density: p 1 N 3 / 6 1 p N=1x10 16 /cm Cell size (nm) Figure 2. Cell size distribution and cell density calculation for a nanofoam made from PMMA-co-EMA wt% POSS (average cell size: 99 nm; void volume: 83.8 %; density: g/cm 3 )

SPE FOAMS 2011 Conference 12-15 September 2011 (Iselin, NJ) 4 Figure 2 shows an example of a representative SEM of the PMMA-co-EMA nanofoam with 99 nm cells and 83.8% porosity.

4 SPE FOAMS 2011 Conference September 2011 (Iselin, NJ) 4 Figure 2 shows an example of a representative SEM of the PMMA-co-EMA nanofoam with 99 nm cells and 83.8% porosity. Well defined cells are visible throughout the foam. The cell size distribution was obtained by measuring the length between two opposite cell walls of 1020 randomly chosen cells. The cell density was calculated, assuming spherical cells, by dividing the pore volume by the average cell volume to obtain the total number of cells in 1 cm 3 of foam. This was then divided by the foam-to-polymer density ratio (1-p) to yield the cell density per cm 3 of unfoamed polymer usually reported. [3] For this foam, a cell density of /cm 3 is obtained, as listed in Table 1. Also shown is the cell density obtained by directly counting all visible cells in a section of the SEM. This method yields a similar number (7x10 15 /cm 3 ), presumably lower due to the difficulty in accounting for every single small cell in the image. Figure 3 shows an example of SEM for the foam in the last row of Table 1. The average cell size of 65 nm is remarkable in that it is below the mean free path of air, which is unique for homogeneous nanofoams made by a gas-blowing process. Interestingly, although the foam is white and opaque, cross-sections of the materials have a bluish hue reminiscent of that observed in aerogels. This is attributed to the same Rayleigh scattering effect. Efforts to further increase porosity are underway. Figure 3. SEM of foam (7mm thick) made from PMMAco-EMA with 0.25wt% POSS, with 65 nm average cell size and 74 vol% voids (see Table 1) Conclusions Thick nanofoams with homogeneous cells and low relative densities were produced by a batch foaming process with supercritical CO 2 using non-phase segregated acrylic and styrenic polymers. Nanoparticles were shown to be effective in increasing the cell density. Using POSS and PMMA-co-EMA, 100 nm nanocellular foams with cell densities of the order of cells/cm 3 were achieved, with porosities up to 85% (0.15 relative foam density). This unique combination of properties is a significant step beyond previous efforts to produce homogeneous gas-blown nanofoams, as recently summarized by Ruckdäschel et al. [7] (see Fig. 36 in this reference), and towards the production of true nanofoams with improved thermal insulation performance. Further improvement is expected by proper modification of the polymer, [13] optimization of the interaction between the polymer and the nanoparticles to maximize dispersion, and refinement in the foaming conditions to fine-tune thermodynamic and rheological properties of the polymer/co 2 system. References 1. L. W. Hrubesh and R. W. Pekala, J. Mater. Res. 9, (1994). 2. C. B. Park, et al., Polym. Eng. Sci. 38 (11), (1998). 3. V. Kumar and N. P. Suh, Polym. Eng. Sci. 30 (20), (1990). 4. B. Krause, et al., Macromolecules 34 (4), (2001). 5. Y. P. Handa and Z. Zhang, J. Polym. Sci., Part B: Polym. Phys. 38 (5), (2000). 6. A. V. Nawaby, et al., Polym. Int. 56 (1), (2007). 7. H. Ruckdäschel, et al., Adv. Polym. Sci. 227, (2010). 8. D. Miller, et al., in Proceedings of the SPE FOAMS 2009 Conference, Iselin, NJ (September 16-17, 2009). 9. M. Dumon and J. A. Reglero Ruiz, in Proceedings of the SPE FOAMS 2010 Conference, Seattle, WA (September 30 October 1, 2010). 10. N. Achaba and M. Ohshima, presented at the SPE FOAMS 2009 Conference, Iselin, NJ (September 16-17, 2009). 11. S. Costeux et al., Nanoporous Polymeric Foam Having High Porosity, PCT Patent appl. WO , Dow Global Technologies, filed on Nov 25, N. S. Ramesh, et al., Polym. Eng. Sci. 34 (22), (1994). 13. S. Costeux et al., Nanoporous Polymeric Foam Having High Cell Density Without Nanofiller, Patent appl. PCT/US11/025782, Dow Global Technologies, filed on March 10, Key Words: Nanofoam, Nanocellular, CO 2, Insulation, Nucleation, Knudsen effect, PMMA, SAN, acrylic

5 SPE FOAMS 2011 Conference September 2011 (Iselin, NJ) 5 Table 1. Representative Nanofoams Polymer Additive Foaming P(MPa) / T( C) Soaking time Post-foaming T( C) Cell size / Porosity p Cell density N PMMA - 30 MPa / 60 C 4.5 hrs No PF 900 nm+130 μm / 81% < /cm 3 PMMA 3 wt% SiO2 30 MPa / 60 C 4.5 hrs No PF 280 nm / 77% 3 x /cm 3 PMMA-co-EA 0.5wt% SiO2 33MPa / 40 C 6 hrs 80 C (25 s) 180 nm / 82% 1.5 x /cm 3 SAN - 33 MPa / 30 C 20 hrs 60 C 2 μm/70% 4.5 x /cm 3 SAN 0.25 wt% POSS 33 MPa / 30 C 20 hrs 60 C 200 nm/70% 5.5 x /cm 3 PMMA-co-EMA 0.25 wt% POSS 33 MPa / 50 C 6 hrs 70 C 120 nm/85.1% 6 x /cm 3 PMMA-co-EMA 0.25 wt% POSS 33 MPa / 40 C 6 hrs 70 C 99 nm / 83.8% 1 x /cm 3 PMMA-co-EMA 0.25 wt% POSS 33 MPa / 35 C 6 hrs 55 C 65 nm / 74% 2 x /cm 3

Abstract. Concept. Introduction

Abstract. Concept. Introduction FOAM INJECTION-MOLDING PROCESS DESIGNED TO PRODUCE SUB-MICRON CELLS Stéphane Costeux, Hyunwoo Kim, Devin Foether, The Dow Chemical Company, Midland, MI, U.S.A. Abstract Significant progress has been made