8966
Langmuir 2006, 22, 8966-8974
Effect of Supercritical CO2 in Modified Polystyrene 3D Latex Arrays Albertina Caban˜as,*,† Eduardo Enciso,† M. Carmen Carbajo,‡ M. Jose´ Torralvo,‡ Concepcio´n Pando,† and Juan Antonio R. Renuncio† Departamento de Quı´mica-Fı´sica I and Departamento de Quı´mica Inorga´ nica I, UniVersidad Complutense de Madrid, 28040 Madrid, Spain ReceiVed May 30, 2006. In Final Form: July 21, 2006 The effect of supercritical CO2 (scCO2) in 3D latex arrays formed by monodispersed particles of polystyrene (PS), PS cross-linked with divinylbenzene (PS-DVB), and PS block copolymers with 2-hydroxyethyl methacrylate (PSHEMA), methacrylic acid (PS-MA), acrylic acid (PS-AA), itaconic acid (PS-IA), and a mixture of methacrylic and itaconic acid (PS-IA-MA) has been studied. Sorption of CO2 into the polymer particles leads to a decrease in the glass transition temperature of the polymer and the swelling of the particles and induces their coalescence. 3D-latex arrays of the former compositions were treated in scCO2 at temperatures and pressures ranging from 40 to 80 °C and from 85 to 197 bar, respectively. The effect of CO2 on the polymeric template was assessed by scanning electron microscopy and N2 adsorption analysis. Bare PS and PS-HEMA particles sintered readily in scCO2 at 40 °C and 85 bar. On the other hand, particles containing carboxylic acid groups on their surface (PS-MA, PS-AA, PS-IA, and PS-IA-MA) were, at the same temperature and pressure, more resistant to the CO2 treatment. For a given polymer composition, the sorption of CO2 inside the polymer particles, the swelling, and the degree of coalescence depend on the pressure, temperature, and time of the CO2 treatment. Analysis of the pore size distributions from the N2 adsorption data has allowed us to quantify the degree of coalescence of the particles in the matrix. By careful control of the experimental variables, the coalescence of the 3D latex array could be finely tuned using CO2.
Introduction Monodisperse latex particles, microspheres of controlled size and variable composition, can be arranged in 3D ordered arrays. These materials can be used as templates for structured porous materials including inorganic oxides, carbons, metals, and polymers.1-4 Recently, a new method to produce large surface area SiO2 macroporous materials from the reaction of silicon alkoxides in a supercritical fluid using 3D latex arrays as templates has been developed in our laboratory.5,6 Supercritical carbon dioxide (scCO2) has been used since it is cheap, nontoxic, and nonflammable and has a relatively low critical temperature and pressure (Tc ) 31 °C, Pc ) 73.8 bar).7 The metal alkoxide precursor is dissolved in scCO2 at moderate pressure and temperature conditions.8 The gaslike transport properties of scCO2 (low viscosity, high diffusivity relative to that of liquids, and very low surface tension) allow infiltration of the precursor into the 3D latex array. A catalyst adsorbed at the particle surface promotes hydrolysis and condensation of the silica source. CO2 is a gas at ambient pressure and is eliminated completely upon depressurization. After reaction, the template is calcined, yielding the inverse replica of the original template. The proposed method * To whom correspondence should be addressed. Phone: 34 + 91 3944200. Fax: 34 + 91 3944135. E-mail:
[email protected]. † Departamento de Quı´mica-Fı´sica I. ‡ Departamento de Quı´mica Inorga ´ nica I. (1) Velev, O. D.; Jede, T. A.; Lobo, R. F.; Lenhoff, A. M. Nature 1997, 289, 447-448. (2) Xia, Y.; Gates, B.; Yin, Y.; Lu, Y. AdV. Mater. 2000, 12, 693-713. (3) Stein, A.; Schroden, R. C. Curr. Opin. Solid State Mater. Sci. 2001, 5, 553-564. (4) Ozin, G. A.; Arsenault, A. C., Nanochemistry. A Chemical Approach to Nanomaterials, 1st ed.; RSC Publishing: Cambridge, U.K., 2005. (5) Caban˜as, A.; Enciso, E.; Carbajo, M. C.; Torralvo, M. J.; Pando, C.; Renuncio, J. A. R. Chem. Commun. 2005, 2618-2620. (6) Caban˜as, A.; Enciso, E.; Carbajo, M. C.; Torralvo, M. J.; Pando, C.; Renuncio, J. A. R. Chem. Mater. 2005, 17, 6137-6145. (7) McHugh, M. A.; Krukonis, V. J., Supercritical Fluid Extraction: Principles and Practice; Butterworth: Boston, 1986. (8) Sphul, O.; Herzog, S.; Gross, J.; Smirnova, I.; Arlt, W. Ind. Eng. Chem. Res. 2004, 43, 4457-4464.
overcomes some of the limitations of other techniques, being a faster and more sustainable method of synthesis and, at the same time, rendering materials of unique properties. A limitation to the synthesis of inverse opals in scCO2 using 3D latex array templates is the large solubility of CO2 in many polymers and the subsequent swelling and reduction of the glass transition temperature of the polymer (Tg), which can produce coalescence or sintering of the polymer particles in the template at temperatures much lower than those found at atmospheric pressure. If the polymeric template is not stable in scCO2, the condensation of the inorganic precursor on the surface of the template is not possible. Furthermore, dissolved CO2 causes a considerable reduction in the viscosity of molten polymers.9,10 However, these phenomena are crucial in many applications such as polymer separation and fraction, extraction of additives and impurities, drug impregnation, polymer modification, gas separation membranes, production of microparticles, foams, gels, and fibers, and polymerization reactions in scCO2.9-12 The solubility of scCO2 in different polymers has been widely studied.9 For example, the sorption of CO2 in polystyrene (PS) at 40 °C and 85 bar is 12-13 wt %.13 Although the Tg of PS at atmospheric pressure is ca. 105 °C,14 this value decreases as a function of CO2 pressure.15,16 The Tg for PS decreases ca. 1 °C/bar up to 60 bar and reaches 55 °C at this pressure. A further 10 bar increase of pressure to 70 bar causes Tg to decrease to (9) Tomasko, D. L.; Li, H.; Liu, D.; Han, X.; Wingert, M. J.; Lee, J. L.; Koelling, K. W. Ind. Eng. Chem. Res. 2003, 42, 6431-6456. (10) Nalawade, S. P.; Picchioni, F.; Janssen, L. P. B. M. Prog. Polym. Sci. 2006, 31, 19-43. (11) Woods, H. M.; Silva, M.; Nouvel, C.; Shakesheff, K. M.; Howdle, S. M. J. Mater. Chem. 2004, 14, 1663-1678. (12) Mu¨ller, M.; MacDowell, L. G.; Virnau, P.; Binder K. J. Chem. Phys. 2002, 117, 5480-5496. (13) Aubert, J. J. Supercrit. Fluids 1998, 11, 163-172. (14) Brandrup, J.; Immergut, E. H.; Grulke, E. A. Polymer Handbook, 4th ed.; John Wiley & Sons: New York, 2003. (15) Condo, P. D.; Johnston, K. P. Macromolecules 1992, 25, 6119-6127. (16) Wissinger, R. G.; Paulatis, M. E. J. Polym. Sci., Part B: Polym. Phys. 1991, 29, 631-633.
10.1021/la061539z CCC: $33.50 © 2006 American Chemical Society Published on Web 09/07/2006
Effect of scCO2 in Modified PS 3D Latex Arrays
32 °C. At this temperature, Tg becomes pressure independent up to 110 bar.15 Kazarian et al.17 have shown using IR spectroscopy that polymers possessing electron-donating functional groups (e.g., carbonyl groups) such as poly(methyl methacrylate) (PMMA) exhibit Lewis acid-base specific interactions with CO2, a larger CO2 sorption, and a pronounced effect on Tg. Thus, the Tg of PMMA presents retrograde behavior in CO2 with a maximum pressure at 60 bar.16,18 Above this pressure, the polymer undergoes plastic deformation. In contrast, there is evidence suggesting that polymers with carboxylic acid groups are more resistant to CO2. For example, at 80 bar, the Tg of poly(acrylic acid) (PAA) in CO2 determined using a chromatographic technique is 48 °C,19 while the Tg of PS in CO2 is ca. 32 °C15 (the Tg values of PAA and PS at atmospheric pressure are 105 and 106 °C, respectively14). For films at ambient conditions, Tg depends on the thickness, and for freely standing films, the measured Tg values are generally lower than those found in the bulk.20 In the case of supported films, Tg can be larger or smaller than that of the bulk polymer, depending on the interactions with the surface.20 The Tg of thin polymer films exposed to CO2 also depends on the thickness. In particular, PS and PMMA thin polymer films (e100 nm) supported on SiO2 substrates show lower Tg values than the bulk when they are exposed to CO2.21,22 Similarly, 200 nm latex particles may be expected to exhibit thin film behavior rather than that of the bulk.23 Matsuyama et al.24 have shown that methacrylic acid (MA) groups and 2-hydroxyethyl methacrylate (HEMA) act as stabilizers in the copolymerization of glycidyl methacrylate in scCO2 at 65 °C and 100 bar and were able to produce at these conditions spherical 0.5-5 µm particles. Otake et al.25,26 have also studied the role of carboxylic acid vinyl monomers such as MA or acrylic acid (AA) in the polymerization of methyl methacrylate and observed formation of fine powders. They reported that MA and AA acted as both monomers and surfactants in the polymerization and found that other aliphatic carboxylic acids such as myristic acid could also act as surfactants. The surfactant-like nature was explained as the effect of specific interactions between the carbonyl oxygen atom and CO217 combined with the static repulsion of the hydroxyl groups.26 On the other hand, crosslinked PS microspheres can be produced by unstabilized suspension polymerization in scCO2 without surfactants. In this case, particles are resistant to coagulation due to their rigid, cross-linked surfaces.27 The CO2 effect on the polymers can be exploited to convenience, and CO2 can be use to induce coalescence of the polymer particles forming the colloidal crystal, thus stabilizing (17) Kazarian, S. G.; Vincent, M. F.; Bright, F. V.; Liotta, C. L.; Eckert, C. A. J. Am. Chem. Soc. 1996, 118, 1729-1736. (18) Condo, P. D.; Johnston, K. P. Macromolecules 1992, 25, 6730-3732. (19) KiKic, I.; Vecchione, F.; Alessi, P.; Cortesi, A.; Eva, F. Ind. Eng. Chem. Res. 2003, 42, 3022-3029. (20) Forrest, J. A.; Dalnoki-Veress, K. AdV. Colloid Interface Sci. 2001, 94, 167-196. (21) Pham, J. Q.; Sirard, S. M.; Johnston, K. P.; Green, P. F. Phys. ReV. Lett. 2003, 91, 175503-1/175503-4. (22) Pham, J. Q.; Johnston, K. P.; Green, P. F. J. Phys. Chem. B 2004, 108, 3457-3461. (23) Abramowitz, A.; Parag, S. S.; Green, P. F.; Johnston, K. P. Macromolecules 2004, 37, 7316-7324. (24) Matsuyama, K.; Mishima, K.; Takahashi, K.; Ohdate, R.; Tomokage, H. J. Chem. Eng. Jpn. 2003, 36, 516-521. (25) Kobayashi, M.; Otake, H.; Yoda, S.; Takebayashi, Y.; Sugeta, T.; Nakazawa, N.; Sakai, H.; Abe, M. Proceedings of the 5th International Symposium on Supercritical Fluids, Versailles, France; ISASF: Nancy, France, 2003. (26) Otake, H.; Kobayashi, M.; Ozaki, Y.; Yoda, S.; Takebayashi, Y.; Sugeta, T.; Nakazawa, N.; Sakai, H.; Abe, M. Langmuir 2004, 20, 6182-6186. (27) Cooper, A. I.; Hems, W. P.; Holmes, A. B. S. Macromol. Rapid Commun. 1998, 19, 353-357.
Langmuir, Vol. 22, No. 21, 2006 8967
the structure of the template. This technique represents an alternative to current physical and/or chemical stabilization methods. Physical stabilization is based on the interdiffussion and entanglement of polymer chains between particles and is usually carried out by partially annealing the latex particles at a temperature above the Tg of the polymer.28 Melt pressing can also be used to improve the stability of the template.29 Chemical stabilization involves interdiffusion of polymer chains as well as chemical reaction at the contact points forming a covalent bond and can be achieved by using reactive species on the particle surface and by swelling the polymer with a monomer/photoinitiator and photopolymerizing the array after it is formed.28 Abranowitz et al.23 have recently shown the ability of CO2 to “weld” PS colloidal crystals and have followed the kinetics of the process using in situ measurements of Bragg diffraction. PS photonic crystals were treated in scCO2 at temperatures ranging from 32 to 40 °C up to 256 bar. Upon exposure to CO2, the welds grew tangentially from the points of contact in all directions. After 6 h of exposure to CO2 at 32 °C and 256 bar, film formation was achieved. Near the critical point, an anomalous sorption and swelling of CO2 into PS latex was observed. The irregularities were explained as a consequence of the Gibbs excess adsorption that results from the high compressibility of CO2 in this region. An anomalous adsorption of CO2 on inorganic supports such as activated carbon has also been reported at conditions very close to the critical point of CO2.30 Otake et al.31 have measured the swelling of 50 nm PS latex spheres in water saturated with CO2 at 25 °C up to 350 bar by dynamic light scattering. The swelling of PS latex particles in water induced by CO2 was much larger than the swelling of bulk PS, due to the preferential CO2 sorption at the latex surface to lower the overall interfacial tension. An anomalous swelling of PMMA thin films (e325 nm) has also been observed in conditions close to the critical point of CO2.32 In this paper, we describe our study of the effect of scCO2 on PS latex particles decorated with different acrylic groups on their surface and cross-linked PS latex particles. Selecting the composition of the block copolymer and the pressure and temperature of the CO2 treatment, the coalescence of 3D latex arrays can be finely tuned. The effect of CO2 on the polymeric template is assessed by scanning electron microscopy (SEM) and N2 adsorption analysis. We have previously shown that N2 adsorption data can be employed to study the ordering of particles in 3D latex arrays.33 In the next sections we will show how this tool can be very valuable to study the degree of coalescence induced by the CO2 treatment. Experimental Section Materials. Bare PS latex particles and PS particles decorated with different hydrophilic groups on their surface were prepared by emulsion copolymerization free of surfactant in water.34 Figure 1 shows the different monomers employed in the copolymerization. HEMA, MA, AA, and itaconic acid (IA) were copolymerized with styrene (S) without surfactant using potassium persulfate as initiator following a procedure previously described,33-36 yielding PS(28) Jons, S.; Ries, P.; McDonald, C. J. J. Membr. Sci. 1999, 155, 79-99. (29) Sosnowski, S.; Li, L.; Winnik, M. A.; Clubb, B.; Shivers, R. J. Polym. Sci., Part B: Polym. Phys. 1994, 32, 2499-2508. (30) Humayun, R.; Tomasko, D. L. AIChE J. 2000, 46, 2065-2075. (31) Otake, H.; Webber, S. E.; Munk, P.; Johnston, K. P. Langmuir 1997, 13, 3047-3051. (32) Sirard, S. M.; Ziegler, K. J.; Sanchez, I. C.; Green, P. F.; Johnston, K. P. Macromolecules 2002, 35, 1928-1935. (33) Carbajo, M. C.; Climent, E.; Enciso, E.; Torralvo, M. J. J. Colloid Interface Sci. 2005, 284, 639-645. (34) Carbajo M. C. Ph.D. Thesis, Universidad Complutense, Madrid, 2004. (35) Carbajo, M. C.; Lo´pez, C.; Gomez, A.; Enciso, E.; Torralvo, M. J. J. Mater. Chem. 2003, 13, 2311-2316.
8968 Langmuir, Vol. 22, No. 21, 2006
Caban˜ as et al. Table 1. Experimental Conditions for the Treatment of PS-MA 3D Latex Arrays in CO2, Temperature (T), Pressure (P), CO2 Density (G), CO2 Activity (a), and Treatment Time (t), BET Surface Area (SBET), and Maximum Value in the Pore Size Distribution Calculated from the Desorption Branch of the Isotherm (max)
Figure 1. Monomers employed in the copolymerization. HEMA, PS-MA, PS-AA, and PS-IA, respectively. The ratio of S to the different monomers ranged from 9:1 to 14:1. Copolymerization of PS with a mixture of MA and IA in similar ratios was also carried out (PS-IA-MA). Cross-linked PS particles produced by copolymerization of S and divinylbenzene (DVB) were also studied (PS-DVB). The suspensions were filtered through glass wool and dialyzed for four weeks against water. The morphology and mean diameter of the particles were characterized by transmission electron microscopy (TEM). The diameter of the particles was measured manually from the images of about 70 particles in each sample. The polydispersity indexes in each case were lower than 1.01. The particle size ranged from 160 to 540 nm for the different compositions. 3D latex arrays were prepared by centrifugation, evaporation of the solvent, or membrane filtration. The materials were dried at room temperature under air and at 45 °C in an oven. Particles crystallized in an fcc packing of spheres, although there were less ordered regions. The packing method influences notably the quality of the order in the latex monoliths as has been previously discussed.33 The Tg values of the homopolymers are Tg(PS) ) 100106 °C, Tg(PIA) ) 160 °C, Tg(PHEMA) ) 85 °C, Tg(PMA) ) 228 °C, and Tg(PAA) ) 106 °C. 14 The Tg of the block copolymers was measured using differential scanning calorimetry (DSC) (DSC 22, Seiko Instruments, Inc.) and ranges from 105 to 120 °C.34 These values agree with the estimate obtained using the Fox equation,37 which relates the inverse Tg of the block copolymer to the summation of the inverse Tg of the homopolymers weighed by their weight fractions. Method. Pieces of 3D latex arrays of different compositions (1-5 mm thick) were treated in scCO2 in a ca. 70 mL custom-made cylindrical stainless steel high-pressure reactor. The reactor containing a small amount of solid sample was immersed in a thermostated bath (PolyScience) at the selected temperature and filled with CO2 using a high-pressure syringe pump (Isco, Inc. model 260D). The pressure was measured using a pressure transducer (Druck Ltd.). A safety valve (Swagelok) was fitted to the reactor. Vessel contents were kept at these conditions for a period between 15 min and 4.5 h depending on the material. After the CO2 treatment, the vessel was depressurized through a needle valve in 15-20 min, and the material was kept for analysis. The experimental setup has been described in detail in ref 38. Preliminary experiments were conducted at 50 °C and 85 bar for 15-60 min and have been reported elsewhere.39 Although these conditions are above the solvent-depressed Tg of PS in CO2,15 particle coalescence showed a clear dependence on the composition of the copolymers. To establish the degree of coalescence of the particles as a function of the polymer composition, 3D latex arrays of every composition were treated at 40 °C and 85 bar in CO2 for 1 h. The effects of pressure, temperature, and treatment time were further investigated using PS-IA-MA and PS-MA. In the PS-MA experiments, (36) Carbajo, M. C.; Gomez, A.; Torralvo, M. J.; Enciso, E. J. Mater. Chem. 2002, 12, 2740-2746. (37) Fox, T. G. Bull. Am. Phys. Soc. 1956, 1, 123. (38) Caban˜as, A.; Enciso, E.; Carbajo, M. C.; Torralvo, M. J.; Pando, C.; Renuncio, J. A. R. Microporous Mesoporous Mater., in press. (39) Caban˜as, A.; Enciso, E.; Carbajo, M. C.; Torralvo, M. J.; Pando, C.; Renuncio, J. A. R. Proceedings of the 6th International Symposium on Supercritical Fluids; Trieste, Italy; ISASF: Nancy, France, 2004; CD-rom.
T/ °C
P/ bar
Fa/ (g‚cm-3)
t/h
untreated 40 40 40 60 60 60 60 80 80
85 120 197 120 197 197 197 133 197
0.3539 0.7178 0.8370 0.4344 0.7188 0.7188 0.7188 0.3524 0.5865
1 1 1 1 0.25 1 4.5 1 1
ab
0.983 1.096 1.300 0.962 1.194 1.194 1.194 0.838 1.037
SBET/ (m2/g)
max/ nm
29 29 28 26 26 25 18 3 aggregated aggregated
23 19 20 19 18 15 12 10
a From ref 43. b Calculated using the Peng-Robinson equation of state.45
temperature was varied from 40 to 80 °C and pressure from 85 to 197 bar. The experimental conditions of the treatment of PS-MA monoliths are summarized in Table 1. Characterization. Materials were characterized using TEM, SEM, and N2 adsorption experiments. TEM micrographs were obtained using a JEOL 2000FX electron microscope operating at 200 kV by placing drops of the latex suspensions over copper grids covered with holey carbon support films. SEM images were taken on a JEOL6400 electron microscope working at 20 kV. Samples were gold coated prior to analysis. N2 adsorption-desorption isotherms at 77 K were obtained using ASAP-2020 equipment from Micromeritics. Prior to the adsorption measurements, the samples were outgassed at 80 °C for 10 h. The isotherms were analyzed using standard procedures. The BET equation was used for specific surface calculations, and the external surface area and the micropore volume were determined using the t plot method.40 The pore size distributions were calculated using the Barrett, Joyner, and Halenda (BJH) method for a cylindrical pore model.41
Results and Discussion Effect of the Chemical Composition. Monodisperse spherical particles self-organize in 3D ordered arrays. PS latex particles (400 nm in diameter) show an fcc packing together with some less ordered regions (Figure 2a,b). In the modified PS latex, the presence of hydrophilic segments at the surface of the particles seems to assist the ordering of the spheres.33 Figures 2c and 3 show images of different latex particles after the treatment in CO2 at 40 °C and 85 bar for 1 h. Bare PS particles coalesced readily in scCO2 (Figure 2c). PS-AA and PS-MA 3D latex arrays (410 and 300 nm particles, respectively) did not seem to aggregate at these conditions as shown in part a and b and part c, respectively, of Figure 3. The same was observed for PS-IA (540 nm); the images are shown in Figure 3d,e. On the contrary, PS-HEMA particles (440 nm) aggregated in scCO2 (Figure 3f,g). Treatment of PS-IA-MA particles (350 nm) in scCO2 at these conditions apparently did not induce particle coalescence. In those cases when coalescence took place (PS and PS-HEMA, Figures 2c and 3f,g, respectively), the samples experienced some hardening and a slight volume contraction. No effect was found when pieces of the same samples were treated in N2 at the same pressure and temperature.39 These conditions are above the solvent-induced Tg of pure PS for 85 bar of CO2 pressure.15 However, the presence of carboxylic acid groups in the block (40) Jaroniec, M.; Kruk, M.; Olivier, J. P. Langmuir 1999, 15, 5410-5413. (41) Barret, E.; Joyner, L. G.; Halenda, P. P. J. Am. Chem. Soc. 1951, 73, 373-380.
Effect of scCO2 in Modified PS 3D Latex Arrays
Figure 2. PS 3D latex arrays (a, b) before and (c) after treatment in CO2 at 40 °C and 85 bar for 1 h. Scale bars: (a) 10 µm, (b, c) 6 µm.
copolymer (PS-MA, PS-AA, PS-IA, and PS-IA-MA) seems to stabilize the latex particles against coalescence as previously suggested in the literature.25,26 On the contrary, coalescence was found when bare PS particles were treated in scCO2 at the same conditions. In the case of PS-HEMA, the effect of the CO2 treatment is more pronounced than in bare PS. This is due partly to the specific interaction between CO2 and the carbonyl group,17 as well as to the lower Tg of the homopolymer PHEMA at atmospheric pressure (85 °C).14 Although a specific interaction between CO2 and the carbonyl group can be expected in PSMA, PS-AA, PS-IA, and PS-IA-MA,17 the presence of a
Langmuir, Vol. 22, No. 21, 2006 8969
hydroxyl group directly bound to the carbonyl group may generate a static repulsion which prevents particles from aggregation.26 A common way to enhance the physical resistance into the PS film is cross-linking. Cross-linked PS-DVB particles (440 nm) were also treated in CO2. Particle coalescence depended on the amount of DVB employed in the synthesis. Latex particles with a high degree of cross-linking, ca. 15 wt % DVB, were more resistant to CO2 because of their larger rigidity and the lower CO2 sorption. The effect of CO2 at 50 °C and 85 bar on a PSDVB sample (molar ratio PS:DVB ) 15:1) is illustrated in Figure 4. The ordering of the latex particles in the 3D array does not seem to be disrupted by the CO2 treatment. This may be observed in the SEM images at lower magnification. For the sake of brevity, the lower magnification images are not shown for all the samples studied. As it will be shown later in the paper, structural damage is only observed in some of the samples treated at higher temperatures and pressures due to the rapid depressurization. Effect of Temperature, Pressure, and Treatment Time. A more detailed study of the effect of CO2 was carried out using PS-IA-MA samples (350 nm) by varying the temperature and pressure conditions. PS-MA 3D latex array monoliths composed of 160 nm particles were also treated in scCO2 at a temperature between 40 and 80 °C and a pressure between 85 and 197 bar. The CO2 exposure time ranged from 15 min to 4.5 h. Table 1 summarizes the experimental conditions. Coalescence was studied using SEM and N2 adsorption. The effect of CO2 on these small particles may be more important than that on larger ones, as they may exhibit thin film behavior rather than bulk behavior.23 SEM Analysis. Treatment of PS-IA-MA particles in scCO2 at 40 °C and 85 bar for 5 h did not induce coalescence according to SEM (Figure 5a,b). On the contrary, severe aggregation was found when the sample was treated at 60 °C and 120 bar for 1 h (Figure 5c,d). Figure 5d shows an area with extensive damage caused by CO2 at these conditions probably during depressurization. PS-MA 3D latex arrays were treated in CO2 for 1 h at 40 °C between 85 and 197 bar. SEM pictures (not shown here) showed that there was not much aggregation during the CO2 treatment.
Figure 3. Modified PS 3D latex arrays after treatment in CO2 at 40 °C and 85 bar for 1 h: (a, b) PS-AA, (c) PS-MA, (d, e) PS-IA, (f, g) PS-HEMA. Scale bars: (a, b, d) 6 µm, (c) 3 µm, (e, f) 10 µm, (g) 4 µm.
8970 Langmuir, Vol. 22, No. 21, 2006
Figure 4. PS-DVB 3D latex arrays (a) before and (b) after treatment with CO2 at 50 °C and 85 bar for 1.5 h. Scale bars: 8 µm.
Figure 5. PS-IA-MA 3D latex arrays after treatment in CO2 at (a, b) 40 °C and 85 bar for 5 h and (c, d) 60 °C and 120 bar for 1 h. Scale bars: (a, d) 10 µm, (b) 2 µm, (c) 6 µm.
However, N2 adsorption experiments revealed a slight coalescence in the polymer particles treated at 40 °C and every pressure. PS-MA 3D latex arrays were also treated in CO2 at 60 °C and 120 bar for 1 h and at 60 °C and 197 bar for different periods of time. Coalescence was more important at 60 °C than at 40 °C. SEM images of PS-MA samples treated at 60 °C at the different pressures are shown in Figure 6. The sample treated at 120 bar shows some aggregation (Figure 6b,c), while coalescence is significant in the samples treated at 197 bar for 1 and 4.5 h (parts d and e, respectively, of Figure 6). In Figure 6c, small spots are observed on the surface of the particles, probably created during
Caban˜ as et al.
the depressurization. In the sample treated at 60 °C and 197 bar for 4.5 h, coalescence has evolved to film formation (Figure 6e). Experiments conducted at 80 °C and 133 and 197 bar revealed complete coalescence of particles in both cases (images not shown here). The different degree of coalescence in the samples was studied using N2 adsorption. N2 Adsorption Analysis. The N2 adsorption isotherm of the PS-MA latex array before any CO2 treatment is shown in Figure 7. The isotherm is type IV with an H1 hysteresis loop.42 The BET surface area (SBET) is 29 m2/g. The t plot obtained from the adsorption data using the Harkins-Jura equation for the film thickness shows deviations from the linear behavior in the highpressure region due to capillary condensation. Downward deviations in the low-pressure region are associated with the existence of micropores. The external surface calculated from the slope of the t plot is equal to 24 cm3/g. This value is slightly smaller than the BET surface area, thus indicating that the microporosity in the sample is very small. The geometrical area was also calculated from the mean particle size, and a value of 34 m2/g was obtained. The geometrical area is higher than the external one, because the area lost due to the contacts among particles was not considered in the geometrical analysis. The external pore system in a 3D latex array of spheres is defined by the interstitial cavities created among the latex particles at high volume packing fraction (74% volume occupation). These cavities are created by tetrahedral and octahedral arrangements of particles in the crystal, which are interconnected by windows formed by three contacting particles. The description of this interconnected bimodal pore system is a challenge for current theoretical tools mostly developed for simple morphologies (cylindrical and slit shape). Nevertheless, the pore system in a close-packed system of spheres can be described as interconnected cylindrical channels of variable diameter parallel to the 〈110〉 directions of the fcc packing. For the 160 nm PS-MA particles, the estimated diameters corresponding to the imaginary spheres tangent to the particles inside the octahedral and tetrahedral cavities in the fcc packing are 66 and 36 nm, respectively. The diameter of the circle that can be inscribed in the triangular window in the close-packed layer of spheres is 25 nm. Considering that each particle contributes to six octahedral and eight tetrahedral cavities in the fcc crystal, the volume fractions attributed to tetrahedral and octahedral cavities are 1/3 and 2/3 of the total void volume, respectively. In an ideal fcc packing, both capillary condensation into the octahedral and tetrahedral cavities during adsorption and evaporation of the condensate during desorption take place at particular values of the relative pressure related by the Kelvin equation to the cavities and window diameters, respectively, which should therefore appear in the pore size distributions. The inset in Figure 7 shows the pore size distributions calculated using the BJH method for a cylindrical pore model obtained from the adsorption and desorption branches of the isotherm. Despite the difficulties in modeling the morphology of the pore system, the analysis of the pore size distributions using the BJH method for a cylindrical pore system gives quantitative information about the system. The distribution curve derived from the adsorption branch of the isotherm is much broader than that obtained from the desorption branch and exhibits two maxima at 60 and 46 nm, which are related to the octahedral and tetrahedral cavities, respectively. These values are, however, slightly different from those estimated from the diameter of the PS-MA particles (66 and 36 nm for tetrahedral and octahedral geometries, respectively). The deviations among (42) Gregg, S. J.; Sing, K. S. W. Adsorption, Surface Area and Porosity, 2nd ed.; Academic Press Inc.: London, 1982.
Effect of scCO2 in Modified PS 3D Latex Arrays
Langmuir, Vol. 22, No. 21, 2006 8971
Figure 6. PS-MA 3D latex arrays before (a) and after treatment with CO2 at 60 °C and different pressures: (b, c) 120 bar for 1 h, (d) 197 bar for 1 h, (e) 197 bar for 4.5 h. Scale bars: (a, b, d, e) 3 µm, (c) 1 µm.
Figure 7. N2 adsorption isotherm for the PS-MA 3D latex array. The inset shows pore distribution curves obtained from the adsorption (0, left scale) and desorption (O, right scale) branches.
these values may be related to the connectivity existing between tetrahedral and octahedral cavities. Smaller relative maxima at ca. 25 and 110 nm and a shoulder at 78 nm are also found. The maximum at 25 nm is related to the size of the triangular window, while the other relative maxima could be due to defects or disorder in the fcc array. On the contrary, the distribution curve obtained from the desorption branch only shows a maximum at 23 nm, which matches closely the size of the window formed by three neighboring spheres estimated using geometrical arguments. A similar analysis was carried out on the samples treated in CO2 at different conditions. Table 1 summarizes the results obtained, indicating the BET surface area and the value of the maximum in the pore size distribution curve obtained from the desorption branch at each condition. Although the SEM pictures (not shown here) did not show aggregation for PS-MA particles treated in CO2 for 1 h at 40 °C between 85 and 197 bar, N2 adsorption experiments showed small differences in the amount of gas adsorbed, BET area, and pore size distribution of the samples, which correlate with the CO2 pressure. The SBET values of PS-MA treated in CO2 at 40 °C and 85, 120, and 197 bar
are 28, 28, and 26 m2/g, respectively (slightly lower than the value obtained for the starting material, 29 m2/g). Figures 8 and 9 show the pore size distributions of PS-MA latex arrays before and after the treatment in CO2 at 40 °C at each pressure, obtained from the adsorption and desorption branches of the isotherm, respectively. Pore size distributions from the adsorption branch in the samples treated in CO2 show maxima better defined and differentiated than in the untreated sample. The intensity of the maximum at ca. 60 nm, related to the octahedral cavities, is lower in the samples treated at 40 °C, and the maximum shifts to higher pore sizes as the CO2 pressure increases (Figure 8). Analysis of the pore size distribution obtained from the desorption branch (Figure 9) is simpler, as it only presents a maximum related to the triangular window formed by three neighboring spheres (Table 1). The maximum in the samples treated in CO2 shifts to lower pore sizes, suggesting partial coalescence of the latex particles. The shift is more pronounced at the highest pressure, and the sample treated at 197 bar presents a narrow maximum at 19 nm. The samples treated at 85 and 120 bar show broader pore size distributions, thus indicating a less homogeneous effect of CO2. The maximum in the pore size distribution obtained from the desorption branch is slightly lower in the sample treated at 85 bar than in the sample treated at 120 bar. Furthermore, the width of the pore size distribution is also larger in the sample treated at 85 bar. The pore size distribution of this sample exhibits a significant contribution of pore sizes smaller than those found in the sample treated at 197 bar, which suggests that the effect of CO2 is larger at the lowest pressure. At constant temperature, CO2 sorption into polymers increases with pressure,9 but the increase is anomalous in the proximity of the critical point of CO2, due to its high compressibility.23 Using data from ref 43, the maximum compressibility at 40 °C is shown to correspond to 86 bar; thus, CO2 sorption at these conditions into the polymer is larger, and as a consequence, coalescence of polymer particles at 85 bar may be more important than at 120 bar. The data presented indicate slight coalescence of the PS-MA particles in CO2 at 40 °C (not evident from SEM). As the polymer particles sinter slightly, tetrahedral and octahedral cavities become better defined, and in comparison to the untreated sample, the
8972 Langmuir, Vol. 22, No. 21, 2006
Caban˜ as et al.
Figure 9. Pore size distribution curves calculated from the desorption branch of the isotherm for PS-MA 3D latex arrays untreated (0) and after treatment in scCO2 for 1 h at 40 °C and different pressures: (O) 85 bar, (4) 120 bar, (]) 197 bar.
Figure 8. Pore size distribution curves calculated from the adsorption branch of the isotherm for PS-MA 3D latex arrays untreated (a) and after treatment in scCO2 for 1 h at 40 °C and different pressures: (b) 85 bar, (c) 120 bar, (d) 197 bar. Dashed lines show the position of the maxima in the pore size distribution of the untreated PS-MA 3D latex array.
two maxima related to these cavities become clearer and shift to lower and higher values, approaching the theoretical values estimated from geometrical arguments. In this slightly sintered geometry, the gas adsorption in the tetrahedral and octahedral cavities is less coupled. The coalescence however is weak and nonuniform, the distribution obtained from the desorption branch broadens, and the maximum shifts to lower values. At 197 bar, the maximum in the pore size distribution associated with the octahedral cavities becomes much larger than the estimated value. This pore size enlargement may be explained considering heterogeneous coalescence, as the probability of coalescence for the four particles forming a tetrahedral cavity is higher than that for the six particles forming an octahedral one. Particles forming tetrahedral cavities may aggregate before those forming octahedral
cavities, and as a consequence, in the first step, the octahedral cavities may become slightly larger. Quantitative assessment of the effect of CO2 at 60 °C was made using N2 adsorption. Adsorption data of samples treated in scCO2 at 60 °C and the different pressures revealed changes in the amount of gas adsorbed, BET area, and pore size distributions. The loss of BET surface with the CO2 treatment was more important at 60 °C and 197 bar than at 120 bar (SBET ) 18 and 26 m2/g, respectively). Further experiments were conducted at 60 °C and 197 bar, varying the treatment time (from 15 min to 4.5 h). After 4.5 h, the coalescence was severe as indicated by the SEM images (Figure 6e) and the loss of area almost complete (SBET ) 3 m2/g). Reducing the treatment time to 15 min, the effect of CO2 was minor and a value of 25 m2/g was obtained for SBET. Figure 10 compares the pore size distribution curves obtained from the adsorption branch of the isotherms for samples treated at 60 °C at the different pressures and treatment times used. Inspection of the distribution curves shows that when aggregation is not very important and the surface area reduction is small (60°C and 120 bar, Figure 10a), the pore size distribution is better defined than that of the sample before any treatment (Figure 8a). Partial coalescence among particles delimits better the tetrahedral and octahedral cavities in the fcc structure. At 60 °C, 197 bar, and longer treatment time (Figure 10c,d), however, coalescence is more important and both tetrahedral and octahedral cavities become smaller and their maxima shift to lower values. The maximum of the pore size distribution curve from the desorption branch (Figure 11) shifts to lower values as the CO2 pressure increases. At 60 °C the distribution curves do not show an anomalous variation with pressure, because this temperature is too far away from the critical temperature of pure CO2 and the compressibility of the fluid at these conditions does not diverge.43 The temporal evolution of the coalescent process can also be followed from the distribution curves obtained from the desorption branch, which show a progressive shift of the maxima to the left with time, accompanied by a decrease in the surface area and the adsorbed volume. After 15 min of treatment, the maximum has shifted from 23 to 15 nm and has broadened. Treatment for 1 and 4.5 h shifts the maximum further, which becomes much broader. After 4.5 h of treatment, surface reduction is very large and the volume adsorbed very small (notice that the scale in Figure 10d is different). It is surprising, however, that the main features of the distribution curve are still preserved. (43) NIST, http://webbook.nist.gov/.
Effect of scCO2 in Modified PS 3D Latex Arrays
Langmuir, Vol. 22, No. 21, 2006 8973
Figure 11. Pore size distribution curves calculated from the desorption branch of the isotherm for PS-MA 3D latex arrays untreated (0) and after the treatment in scCO2 at 60 ° C and different pressures and treatment times: (O) 120 bar for 1 h, (]) 197 bar for 15 min, (4) 197 bar for 1 h, and (*) 197 bar for 4.5 h.
Peng-Robinson equation of state45 as the ratio of the fugacity to the fugacity at the reference state. The reference state was set at the temperature of interest and saturation pressure (extrapolated from the saturation curve above the critical temperature). At all the CO2 pressures and temperatures studied, the polymer is above the solvent-depressed glass transition of bulk PS. Apart from the sorption of CO2 into the polymer directly related to the solvent activity, the mobility of the polymer chains and their transport need to be considered. The diffusivity of the polymer chains in the particle depends on both the temperature and time, because the polymer chain entanglements need to be disrupted before the chains diffuse and the particles coalesce in the contact points. Gupta et al.46 have studied the self-difussion of polystyrene (Ds) in a CO2-swollen PS matrix using neutron reflectivity and shown the dependence of Ds on temperature. According to their data, Ds at 60 °C is several orders of magnitude larger than at 40 °C. Thus, the different Ds would explain the larger coalescence found at 60 °C for the same treatment time.
Conclusions
Figure 10. Pore size distribution curves calculated from the adsorption branch of the isotherm for PS-MA 3D latex arrays after the treatment in scCO2 at 60 ° C and different pressures and treatment times: (a) 120 bar for 1 h, (b) 197 bar for 15 min, (c) 197 bar for 1 h, and (d) 197 bar for 4.5 h. Dashed lines show the position of the maxima in the pore size distribution of the untreated PS-MA 3D latex array.
The effect of CO2 in the 3D latex array is more important at 60 °C than at 40 °C at most pressures studied. SBET and pore size distribution curves in the sample treated at 60 °C and 120 bar were comparable to those observed in the sample treated at 40 °C and 197 bar. Even though the CO2 density at 60 °C and 120 bar is much lower than at 40 °C and 197 bar (see Table 1), the effect of CO2 in the template was very similar. If instead of the CO2 density the results are interpreted in terms of the solvent activity,31,44 the coalescence induced by CO2 at 60 °C is also much larger than that observed at 40 °C for a similar solvent activity (Table 1). The activity of CO2 was calculated using the (44) Shim, J.-J.; Johnston, K. P. AIChE J. 1989, 35, 1097-1106.
The effect of scCO2 in a PS 3D latex modified with DVB and acrylic derivative comonomers has been studied. The sorption of CO2 into the polymer reduces the Tg of the polymers and enhances the dynamics of the polymer chains, and at relatively low temperatures, the particles coalesce. The extent of the coalescence depends on the surface composition, since the swelling degree of the polymer particles and the reduction of Tg depend on the solubility of CO2 inside the polymer. Bare PS and PS-HEMA particles sintered readily in scCO2 at 40 °C and 85 bar. These conditions are above the solvent-depressed Tg of bulk PS. The DVB comonomer increases the mechanical strength of the amorphous particles by the interchain links, decreasing the swelling of the particles. The effect of CO2 in PS-HEMA was more important due to the specific interaction of CO2 with the carbonyl groups and the expected larger sorption of CO2, as well as the lower Tg of PHEMA at atmospheric pressure. On the other hand, particles containing carboxylic acid groups on their surface (PS-MA, PS-AA, PS-IA, and PS-IA-MA) were, at the same conditions, more resistant to the CO2 treatment. Although it has been suggested in the literature that the stabilization of the polymer (45) Peng, D. Y.; Robinson, D. B. Ind. Eng. Chem. Fundam. 1976, 15, 59-64. (46) Gupta, R. R.; Lavery K. A.; Francis, T. J.; Webster, J. R. P.; Smith, G. S.; Russell, T. P.; Watkins, J. J. Macromolecules 2003, 36, 346-352.
8974 Langmuir, Vol. 22, No. 21, 2006
particles is due to the static repulsion between the OH groups in the carboxylic acid,26 the stabilization mechanism of the carboxylic acid groups in CO2 is a challenging topic which will require further analysis. By varying the temperature, pressure and exposure time, the coalescence degree of the particles in the latex crystal could be finely tuned. At 40 °C and 85 bar, the effect of CO2 in PS-MA 3D latex arrays is similar to the effect at 120 bar, due to the larger sorption of CO2 into the polymer at the maximum in the compressibility. At comparable CO2 activities, coalescence is more important at 60 °C than at 40 °C due to the improved mobility of polymer chains at the highest temperature. These results highlight the importance of the surface composition to control particle coalescence and open up the possibility of using scCO2 to induce coalesce of latex particles in a controlled way. Reducing the effect of scCO2 in the 3D latex arrays has allowed us to use the latex crystals as templates to react inorganic precursor dissolved in scCO2.5,6 SEM images have been useful for a coarse definition of the effect of scCO2 in the different latexes. However, N2 adsorption
Caban˜ as et al.
data were essential to quantify the degree of coalescence of the particles in the matrix. Despite the crude model considered in the analysis of the adsorption and desorption branches of the isotherm, the distribution curve could be satisfactorily explained using geometrical arguments. These data will surely encourage new theoretical approaches to describe the adsorption and capillary condensation phenomena in the well-defined porous networks formed by monodisperse spherical particles. Acknowledgment. We gratefully acknowledge the financial support of MEC (Spain) through Research Project CTQ200607172, the Universidad Complutense de Madrid (UCM; Spain), Project PR1/06-14425-A, and CAM (Spain), Project MATERYENER. A.C. thanks MEC (Spain) for its support through a “Ramo´n y Cajal” contract. We thank Dr. J. A. R. Cheda and Mr. F. J. Martı´nez for DSC measurements. We also thank the Centro de Microscopı´a Electro´nica at UCM for technical assistance. LA061539Z
