2184
Langmuir 2004, 20, 2184-2190
Synthesis of Vinyl Polymer-Silica Colloidal Nanocomposites Prepared Using Commercial Alcoholic Silica Sols M. J. Percy, J. I. Amalvy,† D. P. Randall, and S. P. Armes* Department of Chemistry, School of Life Sciences, University of Sussex, Falmer, Brighton, BN1 9QJ, United Kingdom
S. J. Greaves and J. F. Watts The Surface Analysis Laboratory, School of Engineering, University of Surrey, Guildford, Surrey, GU2 7XH, United Kingdom Received October 7, 2003. In Final Form: January 2, 2004 The surfactant-free synthesis of vinyl polymer-silica nanocomposite particles has been achieved in aqueous alcoholic media at ambient temperature in the absence of auxiliary comonomers. Styrene, methyl methacrylate, methyl acrylate, n-butyl acrylate, and 2-hydroxypropyl methacrylate were homopolymerized in turn in the presence of three commercially available ultrafine alcoholic silica sols. Stable colloidal dispersions with reasonably narrow size distributions were obtained, with silica contents of up to 58% by mass indicated by thermogravimetric analysis. Particle size distributions were assessed using both dynamic light scattering and disk centrifuge photosedimentometry. The former technique indicated that the particle size increased for the first 1-2 h at 25 °C and thereafter remained constant. Particle morphologies were studied using electron microscopy. Most of the colloidal nanocomposites comprised approximately spherical particles with relatively narrow size distributions, but in some cases more polydisperse or nonspherical particles were obtained. Selected acrylate-based nanocomposites were examined in terms of their film formation behavior. Scanning electron microscopy studies indicated relatively smooth films were obtained on drying at 20 °C, with complete loss of the original particle morphology. The optical clarity of solutioncast 10 µm nanocomposite films was assessed using visible absorption spectrophotometry, with 93-98% transmission being obtained from 400 to 800 nm; the effect of long-term immersion of such films in aqueous solutions was also examined. X-ray photoelectron spectroscopy studies indicated that the surface compositions of these nanocomposite particles are invariably silica-rich, which is consistent with their long-term colloidal stability and also with aqueous electrophoresis measurements. FT-IR studies suggested that in the case of the poly(methyl methacrylate)-silica nanocomposite particles, the carbonyl ester groups in the polymer are hydrogen-bonded to the surface silanol groups. According to differential scanning calorimetry studies, the glass transition temperatures of several poly(methyl methacrylate)-silica and polystyrene-silica nanocomposites can be either higher or lower than those of the corresponding homopolymers, depending on the nature of the silica sol.
Introduction The field of polymer nanocomposites is of rapidly growing interest. Intimate mixing of polymers with inorganic clays or silica on a nanoscale can lead to superior mechanical properties and improved fire retardancy.1-3 Other applications include nanolithography,4,5 polymer light-emitting diodes (LEDs),6 solid polymer electrolytes,7 * To whom correspondence should be addressed. † Member of research Career of CIC, Buenos Aires, Argentina, on leave from CIDEPINT (Centro de Investigacio´n y Desarrollo en Tecnologı´a de Pinturas), Av. 52, entre 121 y 122 s/n, (B1906AYB) La Plata, Buenos Aires, Argentina. (1) Gilman, J. W.; Jackson, C. L.; Morgan, A. B.; Harris, R.; Manias, E.; Giannelis, M.; Wuthenow, E. P.; Hilton, D.; Phillips, S. H. Chem. Mater. 2000, 12, 1866. (2) Zhu, J.; Morgan, A. B.; Lamelas, F. J.; Wilkie, C. A. Chem. Mater. 2001, 13, 3774. (3) (a) Manias, E.; Touney, A.; Wu, L.; Strawhecker, K.; Lu, B.; Chung, T. C. Chem. Mater. 2001, 13, 3516. (b) Kashiwagi, T.; Morgan, A. B.; Antonucci, J. M.; VanLandingham, M. R.; Harris, R. H.; Awad, W. H.; Shields, J. R. J. Appl. Polym. Sci. 2003, 89, 2072. (4) Hu, Y. Q.; Wu, H. P.; Gonsalves, K. E.; Merhari, L. Microelectron. Eng. 2001, 56, 289. (5) Gonsalves, K. E.; Merhari, L.; Wu, H. P.; Hu, Y. Q. Adv. Mater. 2001, 13, 703. (6) Lee, T.-W.; Park, O. O.; Yoon, J.; Kim, J.-J. Adv. Mater. 2001, 13, 211.
molecular imprinting,8 and immunodiagnostic assays.9 In recent years, there has been increasing interest in the synthesis of particulate or colloidal organic/inorganic hybrids.10 For example, Huang and Brittain have reported the synthesis of poly(methyl methacrylate)-clay nanocomposites using both suspension and emulsion polymerization techniques.11 Using wide-angle X-ray diffraction, these authors reported that exfoliated nanostructures were initially produced from both synthetic routes, with a mixture of intercalated and exfoliated nanostructures being obtained after melt processing. Bourgeat-Lami and Lang reported the synthesis of polystyrene/silica composite particles by the nonaqueous dispersion polymerization of styrene in alcoholic media in the presence of surfacefunctionalized silica particles using polymeric stabilizers such as poly(N-vinyl pyrrolidone).12,13 The silica particles ranged from 13 nm to more than 600 nm in diameter and (7) Bronstein, L. M.; Joo, C.; Karlinsey, R.; Ryder, A.; Zwanziger, J. W. Chem. Mater. 2001, 13, 3678. (8) Marx, S.; Liron, Z. Chem. Mater. 2001, 13, 3624. (9) Pope, M. R.; Armes, S. P.; Tarcha, P. J. Bioconjugate Chem. 1996, 7, 436. (10) Bourgeat-Lami, E. J. Nanosci. Nanotechnol. 2002, 2, 1. (11) Huang, X.; Brittain, W. J. Macromolecules 2001, 34, 3255. (12) Bourgeat-Lami, E.; Lang, J. J. Colloid Interface Sci. 1998, 197, 293.
10.1021/la035868s CCC: $27.50 © 2004 American Chemical Society Published on Web 02/17/2004
Vinyl Polymer-Silica Colloidal Nanocomposites
contained polymerizable methacrylate groups which had been surface-modified using siloxane chemistry. Smaller silica sols led to the formation of nanocomposite particles, whereas larger silica sols produced core-shell morphologies (silica cores, polystyrene shells). In related work, Furusawa et al. utilized hydropropylcellulose as a polymeric binder/stabilizer in the aqueous emulsion polymerization of styrene in order to promote the formation of polystyrene/silica composite particles.14 Very recently, Bourgeat-Lami and co-workers15 have described the preparation of poly(methyl methacrylate)/silica composite particles and Bourgeat-Lami10 has published a comprehensive review of colloidal nanocomposites. Recently, we reported that vinyl polymer-silica nanocomposite particles of colloidal dimensions are readily prepared by the (co)polymerization of an auxiliary monomer, 4-vinylpyridine (4VP), in the presence of an ultrafine silica sol.16-18 Particle diameters were typically 100-200 nm, with silica contents varying from 10 to 50% silica by mass. Transparent, abrasion-resistant nanocomposite coatings with low water permeability were obtained with film-forming comonomers such as n-butyl acrylate.18 These syntheses were carried out in the absence of any added surfactants in aqueous media, and as little as 10 mol % of the 4VP auxiliary was required to ensure nanocomposite particle formation. Antonietti and co-workers recently modified this 4VP auxiliary protocol in order to synthesize colloidal nanocomposite particles using miniemulsion polymerization. This approach required the addition of both surfactant and cosurfactant but had the great advantage of dramatically increasing the silica aggregation efficiency.19 Nevertheless, it would be preferable if such nanocomposite particles could be prepared without recourse to auxiliary comonomers. In a recent communication,20 we showed that replacing the aqueous silica sol with a commercially available alcoholic silica sol (13 nm diameter, donated by Clariant) enabled poly(methyl methacrylate)/ silica nanocomposite particles to be prepared in the absence of any auxiliary comonomers. In the present study, we have extended this initial work to include styrene, acrylates, and other methacrylate monomers and also two further commercial silica sols, thus providing access to a wide range of vinyl polymer-silica nanocomposite particles. These inorganic/organic hybrid particles can be readily prepared with reasonably uniform size distributions in aqueous alcoholic media at ambient temperature, see Figure 1. Experimental Section Materials. Highlink 502-31 (12 nm diameter silica sol dispersed in 2-propanol, supplied by Clariant, France), MA-ST-M (22 nm diameter silica sol dispersed in methanol, supplied by Nissan Chemicals, U.S.), and IPA-ST (a 13 nm silica sol dispersed in 2-propanol, supplied by Nissan Chemicals) were all used as received. 2-Hydroxypropyl methacrylate (HPMA) monomer (13) Bourgeat-Lami, E.; Lang, J. J. Colloid Interface Sci. 1999, 210, 281. (14) Furusawa, K.; Kimura, Y.; Tagawa, T. J. Colloid Interface Sci. 1986, 109, 69. (15) Luna-Xavier, J. L.; Bourgeat-Lami, E.; Guyot, A. Colloid Polym. Sci. 2001, 279, 947. (16) Barthet, C.; Hickey, A. J.; Cairns, D. B.; Armes, S. P. Adv. Mater. 1999, 11, 408. (17) Percy, M. J.; Barthet, C.; Lobb, J. C.; Khan, M. A.; Lascelles, S. F.; Vamvakaki, M.; Armes, S. P. Langmuir 2000, 16, 6913. (18) Amalvy, J. I.; Percy, M. J.; Armes, S. P.; Wiese, H. Langmuir 2001, 17, 4770. (19) Tiarks, F.; Landfester, K.; Antonietti, M. Langmuir 2001, 17, 5775. (20) Percy, M. J.; Armes, S. P. Langmuir 2002, 18, 4562.
Langmuir, Vol. 20, No. 6, 2004 2185
Figure 1. Schematic representation of the formation of vinyl polymer-silica nanocomposite particles by the free radical homopolymerization of various vinyl monomers at 25 °C in aqueous alcohol mixtures in the presence of an ultrafine silica sol. It is emphasized that neither surface pretreatment nor any auxiliary comonomer is required for successful nanocomposite formation. (donated by Cognis Performance Chemicals, Hythe, U.K.) and all other vinyl monomers (Aldrich) were treated in turn with basic alumina to remove inhibitor and stored at -20 °C until required. Nanocomposite Synthesis. The nanocomposite particles were prepared by free radical homopolymerization of various vinyl monomers (see Table 1 for details) in the presence of one of the three ultrafine alcoholic silica sols in various water/alcohol mixtures. Thus, in a typical experiment, the alcoholic solution of the silica sol (equivalent to 4.0 g of dry weight silica) was mixed with deionized water (30 mL) and monomer (5.0 mL) in a Schott bottle. The total volume was made up to 47.0 mL with distilled water, and the bottle was sealed with a screw cap with two drill-holes that was fitted with an inner rubber gasket. This setup allowed both degassing and addition of initiator using syringe needles. The mixture was degassed via nitrogen purge and placed in a circulating water bath maintained at 25 °C and equipped with a magnetic stirrer. Ammonium persulfate initiator (APS; 25.0 mg; 1.0 wt % based on monomer) dissolved in water (1.5 mL) and N,N,N′,N′-tetramethylethylenediamine (TEMED; 25.5 mg; the APS/TEMED molar ratio was unity) dissolved in water (1.5 mL) were separately degassed and added to the reaction vessel. The reaction mixture was stirred at 25 °C for 24 h. The milky-white dispersions were purified by at least three centrifugation-redispersion cycles with each successive supernatant being decanted and replaced with aqueous HCl (pH 4) until no excess silica sol was observed by electron microscopy. Care was taken to avoid excessive centrifugation rates and times since these would result in the unwanted sedimentation of the excess silica sol and also make redispersion of the nanocomposite particles more difficult. Particle Size Analysis. Disk centrifuge photosedimentometry (DCP; Brookhaven Instruments) was used to obtain the weight-average particle size distribution of the nanocomposite particles. Solid-state particle densities were measured by helium pycnometry (Micromeritics AccuPyc 1330 instrument). The centrifugation rate was in the 3000-12000 rpm range, and the dispersions were assumed to have the same scattering characteristics as aqueous poly(methyl methacrylate) or polystyrene latex. Typical DCP runs lasted 30 min. Dynamic light scattering (DLS) measurements were performed using a Malvern 4700 instrument equipped with a 75 mW argon ion laser; all measurements were carried out at 25 °C at a fixed angle of 90° on highly dilute aqueous solutions. Surface Area Analysis. Brunauer-Emmett-Teller (BET) surface area measurements were performed using a Quantachrome Nova 1000 instrument using dinitrogen gas as an adsorbate at 77 K. The area per molecule for dinitrogen was taken to be 16.2 Å2. Helium pycnometry measurements indicated average particle densities of 2.11, 2.16, and 2.13 g cm-3 for the Highlink, MA-ST-M, and IPA-ST silica sols, respectively. Mean particle diameters for the silica sols were calculated using As ) 6/FD, where As is the BET specific surface area, F is the particle density,
2186
Langmuir, Vol. 20, No. 6, 2004
Percy et al.
Table 1. Summary of the Synthesis Conditions, Silica Contents, Particle Densities, and Particle Size Data for the Vinyl Polymer-Silica Nanocomposites Prepared Using Three Commercial Ultrafine Silica Solsa entry number 1 2 3 4 5 6 7 8 9 10 11 12
silica sol type
nominal silica sol diameterb (nm)
measured silica sol diameterc (nm)
Highlink Highlink Highlink Highlink MA-ST-M MA-ST-M MA-ST-M IPA-ST IPA-ST IPA-ST IPA-ST IPA-ST
13 13 13 13 22 22 22 13 13 13 13 13
12 12 12 12 21 21 21 13 13 13 13 13
monomer type MMA St MA n-BuA MMA St HPMA MMA St MA n-BuA n-BuMA
silica contentd (wt %)
particle densitye (g cm-3)
32 40 43 38 39 44 39 54 48 58 62 43
1.39 1.35 1.47 1.27 1.47 1.44 1.45 1.58 1.40 1.64 1.54 1.45
particle diameter DLS Dif DCP Dwg TEM Dn (nm) (nm) (nm) 340 310 230 280 770 360 450 180 180 120 130 160
220 ( 40 150 ( 40 210 ( 70 230 ( 50 160 ( 30 180 ( 15 300 ( 60 140 ( 30 140 ( 30 110 ( 40 110 ( 40 130 ( 30
230 170 260 280 110 230 150h 100 140 110 130 140
a All syntheses were conducted at 25 °C using 5.0 mL of monomer and 4.0 g of silica sol in 50 mL of an aqueous alcohol solution, together with 1.0 wt % of a 1:1 ammonium persulfate/N,N,N′,N′-tetramethylethylenediamine complex based on monomer. b Nominal silica sol diameter based on manufacturers’ specifications. c Silica sol diameter as determined by BET surface area measurements. d Determined from thermogravimetric analysis assuming that the incombustible residues were SiO2. All values were corrected to account for the surface moisture content of the silica sols. e As measured by helium pycnometry (Micromeritics AccuPyc 1330 instrument). f Intensity-average diameter of the diluted reaction solution as determined by dynamic light scattering studies. g Weight-average diameter of the redispersed, purified nanocomposite particles as determined by disk centrifuge photosedimentometry. h Most particles were 150 nm. However, a minor fraction of particles were of 560 nm diameter.
Table 2. Summary of the Bulk Compositions (Calculated from Thermogravimetric Analyses and Carbon Microanalyses) and Surface Compositions (Determined by X-ray Photoelectron Spectroscopy) Obtained for Selected Vinyl Polymer-Silica Nanocompositesa XPS surface composition entry number (see Table 1) 2 6 9 10 FDd 10 RTd
bulk composition Si % C% 17.5 20.2 21.6 26.1 26.1
55.6 51.2 48.1 24.4 24.4
bulk atomic ratio Si/C
Si (atom %)
0.13 0.17 0.19 0.46 0.46
12.4 14.5 18.4 11.7 16.1
C (atom %)
Si/C atom ratio
surface/bulk atomic ratio (Si/C)surface/(Si/C)bulk
inference
52.8 43.2 27.8 49.0 30.8
0.23 0.34 0.66 0.24 0.52
1.8 2.0 3.5 0.5 1.1
silica-rich silica-rich silica-rich polymer-rich uniform composition
a These compositions are expressed in terms of Si/C atomic ratios: a normalized Si/C atomic ratio is used to indicate the relative surface composition of the nanocomposite particles (see main text). b From elemental microanalyses (in wt %). c As determined by thermogravimetric analysis, assuming that the incombustible residues are SiO2. d FD, freeze-dried; RT, dried at room temperature (20 °C).
and D is the particle diameter. These were then compared with the manufacturers’ nominal diameters, see Table 1. Differential Scanning Calorimetry. Differential scanning calorimetry (DSC) data were obtained using a Perkin-Elmer DSC-7 instrument, between 20 and 150 °C for polystyrene/silica nanocomposites and between 20 and 200 °C for poly(methyl methacrylate)/silica nanocomposites, at a heating rate of 20 °C min-1. Samples were heated to 200 °C at 40 °C min-1 and cooled at 50 °C min-1 before undertaking any measurements. A temperature correction was made using three fixed melting point standards over the range 90-160 °C. Scanning Electron Microscopy (SEM). Samples were sputter-coated with gold to prevent sample charging and examined using a Stereoscan 420 instrument at a operating voltage of 20 kV and a beam current of 15 pA. Chemical Composition. Thermogravimetric analyses were performed using a Perkin-Elmer TGA-7 instrument. Nanocomposite dispersions were dried at 50 °C overnight to yield dried powders. These powders were heated in air to 800 °C at a scan rate of 20 °C min-1, and the observed mass loss was attributed to the quantitative pyrolysis of the (co)polymer component. The incombustible residues remaining after pyrolysis were assumed to be pure silica (SiO2), and the inorganic contents of the nanocomposites were calculated after correcting for loss of surface moisture of the silica sol at elevated temperature. CHN microanalyses were carried out at an independent external analytical laboratory (Medac Ltd., Egham, Surrey, U.K.). Aqueous Electrophoresis Measurements. Aqueous electrophoresis data were obtained using a Malvern Instruments ZetamasterS instrument. The zeta potential, ζ, was calculated21 from the electrophoretic mobility (u) using the Smoluchowsky relationship, ζ ) ηu/, where it is assumed that κa . 1 (where η is the solution viscosity, is the dielectric constant of the medium, and κ and a are the Debye-Hu¨ckel parameter and the
particle radius, respectively). The solution pH was adjusted by the addition of either NaOH or HCl. X-ray Photoelectron Spectroscopy (XPS) Measurements. X-ray photoelectron spectra were recorded using a VG Scientific Sigma Probe spectrometer fitted with a standard twin anode X-ray source providing either Al KR or Mg KR photons. The spectra reported in this paper were all acquired using Al KR radiation, and the analysis area was approximately 500 µm in diameter. The spectrometer was operated in the fixed analyzer transmission mode, and pass energies of 20 and 100 eV were used to acquire the high-resolution spectra and the survey spectra, respectively. For each sample, a survey spectrum was recorded together with core level spectra of the elements of interest (C1s, O1s, Si2p). Quantitative surface elemental compositions were determined from the peak areas of these spectra together with the appropriate sensitivity factors, using the software provided on the VG Scientific Eclipse data system. Identification of the Si2s peak was verified by XPS analysis of the original silica sols used to prepare the vinyl polymer-silica nanocomposites. The elemental ratios were estimated to within (10% by adjusting the integrated peak areas using known sensitivity factors.22 The chemical composition data presented in Table 1 were used to determine the corresponding bulk Si/C atomic ratios for the nanocomposites summarized in Table 2.
Results and Discussion The commercial ultrafine alcoholic silica sols examined in this work were characterized by BET surface area (21) Butterworth, M. D.; Maeda, S.; Johal, J.; Corradi, R.; Lascelles, S. F.; Armes, S. P. J. Colloid Interface Sci. 1995, 174, 510. (22) Percy, M. J.; Amalvy, J. I.; Barthet, C.; Armes, S. P.; Greaves, S. J.; Watts, J. F.; Wiese, H. J. Mater. Chem. 2002, 12, 697.
Vinyl Polymer-Silica Colloidal Nanocomposites
measurements and helium pycnometry. The mean BET particle diameter calculated for each sol was in very good agreement with the manufacturer’s nominal diameter, see Table 1. These sols are significantly more hydrophobic (i.e., have lower surface charge densities) than the ultrafine aqueous silica sols previously investigated.20 This subtle difference allows the facile preparation of colloidal nanocomposites without recourse to either surface pretreatments or auxiliary comonomers. Homopolymerization of either methyl methacrylate, styrene, methyl acrylate, n-butyl acrylate, or 2-hydroxypropyl methacrylate in the presence of the commercial ultrafine alcoholic silica sols led to the formation of colloidally stable nanocomposite particles. Details of the various syntheses are summarized in Table 1. Mean silica contents were relatively high, ranging from 32 to 62% by mass as judged by thermogravimetric analyses. Hence the average nanocomposite particle densities were as high as 1.64 g cm-3, which is significantly higher than conventional latex particle densities. The particle density is an important input parameter for the DCP studies, which report a weight-average particle diameter. In contrast, DLS provides an intensity-average diameter, and thus this technique “oversizes” relative to DCP since it is much more sensitive to the presence of larger particles, even if these constitute a minor fraction of the size distribution. In most cases, the difference in mean particle diameters obtained from the DLS and DCP studies was relatively small, but in one or two cases larger discrepancies were observed. In particular, the apparent DLS diameter of 770 nm obtained for entry 5 probably indicates that this dispersion is weakly flocculated. The mean electron microscopy diameters reported in the final column of Table 1 should be treated with some caution, since only a relatively small number of particles (95%), but in some cases (particularly with syntheses involving the IPA-ST silica sol) lower conversions were obtained. Allowing for the incorporation of the silica sol into the nanocomposite particles, the final solids contents ranged from 7 to 17% solids. By inspection of Table 1, it is clear that using the IPAST silica sol led to nanocomposites with relatively high silica contents compared to the other two sols. For example, in the case of the three poly(methyl methacrylate)/silica nanocomposite particles (see entries 1, 5, and 8) the silica content was 54% for the IPA-ST sol but only 32% and 39% for the methanolic Highlink and the MA-ST-M sols, respectively. A similar trend is evident for the three polystyrene/silica nanocomposites (compare entries 2, 6, and 9). This may reflect subtle differences in the surface chemistries of the three sols or may be simply due to the change in the alcoholic cosolvent from IPA to methanol. Further studies are required to examine this aspect. Considering just the two methanolic sols, it seems that using the larger silica sol leads to slightly higher silica contents in the final nanocomposites (compare entries 1 and 5 and also 2 and 6). In almost all syntheses, the monomer is only partially miscible with the continuous phase, despite the presence of the alcohol cosolvent; hence the polymerizations were conducted under emulsion polymerization conditions. The exception is 2-hydroxypropyl methacrylate (see entry 7), which is completely miscible prior to polymerization. Interestingly, this synthesis produced a rather polydisperse, bimodal size distribution. Our various attempts to prepare vinyl polymer-silica nanocomposite particles using basic monomers such as 4-vinylpyridine or 2-(diethylamino)ethyl methacrylate were unsuccessful, with
Langmuir, Vol. 20, No. 6, 2004 2187
Figure 2. Particle growth vs time curves for nanocomposite particles prepared at 25 °C using the IPA-ST silica sol in conjunction with the following three monomers: (a) methyl methacrylate (entry 8 in Table 1), (b) methyl acrylate (entry 10 in Table 1), and (c) styrene (entry 9 in Table 1).
macroscopic precipitates being obtained in each case. Again, this suggests that there is a significant difference between the aqueous silica sols employed in our earlier work and the hydrophobic silica sols used in the present study. Given its sensitivity to particle aggregation, DLS was used to monitor the nanocomposite particle growth with time, see Figure 2. No further increase in particle diameter was taken to indicate that the polymerization had ceased. As expected, acrylic monomers were polymerized more rapidly than methacrylic monomers, which in turn reacted faster than styrene. Thus, in the case of methyl acrylate polymerization ceased after around 30 min at 25 °C, whereas methyl methacrylate required approximately 80 min under the same conditions. The formation of polystyrene/silica nanocomposite particles was considerably slower, but eventually a constant particle diameter of 180 nm was achieved after 16 h (not shown in Figure 2). Treatment of one of the poly(methyl methacrylate)/silica nanocomposites (entry 5 in Table 1) with excess HF led to complete digestion of the silica component (confirmed by thermogravimetric analyses) and allowed isolation of the poly(methyl methacrylate) component. Analysis of the extracted polymer by gel permeation chromatography (GPC) using a THF eluent and poly(methyl methacrylate) calibration standards indicated an Mn of 265 000 and a Mw/Mn of 3.1. Similar GPC data were recently reported by Kashiwagi et al. for poly(methyl methacrylate)-silica monolithic nanocomposites3b and are typical for free radical polymerizations that proceed to high conversion. This extracted poly(methyl methacrylate) sample also proved to be a useful reference material for both FT-IR spectroscopy and DSC studies (see below). SEM was used to examine the size distributions and morphologies of the nanocomposite particles, see Figure 3. At low magnification, the polystyrene/silica nanocomposite particles prepared using the 13 nm IPA-ST silica sol (entry 9 in Table 1) appeared to be approximately spherical and had a relatively uniform size distribution (Figure 3a). In contrast, the polystyrene/silica nanocomposite particles prepared using the 22 nm MA-ST-M silica sol (entry 6 in Table 1) were significantly larger and had a distinctive framboidal or “raspberry” particle morphology, see Figure 3b. Poly(methyl methacrylate)/silica nanocomposites prepared using the IPA-ST silica sol were relatively small (around 100 nm diameter as judged by SEM) and rather misshapen, since in this case the sol diameter is relatively large compared to the nanocomposite
2188
Langmuir, Vol. 20, No. 6, 2004
Percy et al.
Figure 3. Representative scanning electron micrographs for (a) polystyrene/silica (entry 9 in Table 1), (b) polystyrene/silica (entry 6 in Table 1), (c) poly(methyl methacrylate)/silica (entry 8 in Table 1), and (d) poly(methyl acrylate)/silica (entry 10 in Table 1). The scale bars are (a) 1 µm, (b) 300 nm, (c) 200 nm, and (d) 1 µm.
Figure 4. Zeta potential vs pH curves obtained for two polystyrene/silica nanocomposites: (O) prepared with the Highlink silica sol (entry 2 in Table 1, silica content ) 40 wt %) and (b) prepared with the IPA-ST silica sol (entry 9 in Table 1, silica content ) 48 wt %). The corresponding Highlink (4) and IPA-ST (2) silica sols are included as reference materials.
particle diameter. Poly(methyl acrylate)/silica nanocomposite particles prepared using the same sol were filmforming, even at a silica content of around 58% (see entry 10 in Table 1). SEM studies confirmed that coalescence was complete; there was no evidence for the original nanocomposite particles in the solution-cast film, see Figure 3d. Aqueous electrophoresis versus pH curves for two polystyrene/silica nanocomposites and their corresponding silica sols are shown in Figure 4. In both cases, the zeta potentials of the nanocomposites remained negative across the whole pH range examined. For a given pH, each nanocomposite has a more negative zeta potential than its corresponding silica sol. This is most likely due to the presence of anionic sulfate groups derived from the
persulfate initiator. These data are consistent with the observed long-term colloid stabilities of the nanocomposite dispersions, since they suggest that the surface of the nanocomposite particles is silica-rich. Similar observations were made by both Butterworth et al. for various conducting polymer-silica nanocomposite particles21 and also Percy and Armes for a poly(methyl methacrylate)/ silica nanocomposite prepared using the Highlink silica sol.20 XPS can also be used to assess the surface compositions of these nanocomposite particles (see Table 2). In our XPS studies, the Si2p and C1s signals were used as unique elemental markers for the silica and polymer components, respectively. Since XPS is very surface-specific (its typical sampling depth is 2-10 nm), the Si/C atomic ratio determined by this technique indicates the near-surface composition of the nanocomposite particles. Four nanocomposites were examined in turn (see entries 2, 6, 9, and 10 in Table 1), and their surface Si/C atomic ratios were compared with the corresponding bulk Si/C atomic ratios, which were calculated from a combination of thermogravimetric analyses and carbon microanalyses, respectively. The “normalized” Si/C atomic ratio indicates whether the surface composition differs significantly from the bulk composition of the nanocomposite particles. If this ratio is greater than unity, the surface composition is silica-rich, whereas if the ratio is less than unity, the surface composition is polymer-rich (or silica-depleted). On the other hand, if the ratio is approximately unity, this indicates that there is little or no difference between the interior and exterior of the nanocomposite particles.22 Inspecting Table 2, it is clear that the surface compositions of each of the three polystyrene/silica nanocomposites are silica-rich, which is consistent with the aqueous electrophoresis data (see above). XPS studies were also conducted on a poly(methyl acrylate)/silica nanocomposite dispersion (entry 10 in Table 1) that had been either freezedried from water or dried at ambient temperature. Under
Vinyl Polymer-Silica Colloidal Nanocomposites
Figure 5. Visible absorption spectra obtained for two solutioncast, acrylate-based film-forming nanocomposites: (a) poly(methyl acrylate)/silica nanocomposites prepared with IPA-ST silica sol (entry 10 in Table 1, 58% silica); (b) poly(n-butyl acrylate)/silica nanocomposites prepared with Highlink silica sol (entry 4 in Table 1, 38% silica).
the former conditions, the original particle morphology should be largely retained, whereas film formation should occur at room temperature. XPS studies indicate that the thermal history of this nanocomposite makes a significant difference to its near-surface composition (see the last two entries in Table 2). However, the surface of the nanocomposite film comprising coalesced particles was less polymer-rich (i.e., it had a higher surface silica concentration) than the original nanocomposite particles. This unexpected observation is in contrast to that reported previously for poly(n-butyl acrylate-stat-4-vinylpyridine)silica nanocomposite particles,18 where surface depletion of the silica sol was observed. In this earlier study, displacement of the hydrophilic silica sol from the solid/air interface by the mobile hydrophobic copolymer was rationalized on the basis of a simple surface thermodynamics argument. Thus the most likely explanation for the XPS results observed in the present work is that the alcoholic silica sols used to prepare the nanocomposite particles are simply more hydrophobic than the poly(methyl acrylate) component. Regardless of the true explanation for silica enrichment at the surface, we note that such nanocomposites may offer improved abrasion resistance and flame retardancy performance. The optical clarity of selected film-forming nanocomposite formulations was assessed using visible absorption spectroscopy, see Figure 5. Two acrylate-based nanocomposite films were prepared by solvent casting at ambient temperature in a modified quartz cell. The film prepared using a poly(methyl acrylate)-silica nanocomposite (entry 4 in Table 1; silica volume fraction ) 0.26; film thickness ) 10 µm) had a transmittance of at least 95% over the whole visible spectrum range (400-800 nm). Similarly, a poly(n-butyl acrylate)-silica nanocomposite film (entry 10 in Table 1; silica volume fraction ) 0.42) had a mean transmittance of more than 96% from around 550 to 800 nm. These observations indicate that the ultrafine silica sol is well dispersed within both nanocomposite films, since appreciable flocculation would inevitably lead to significant light scattering and hence reduced transmittance. Similar results were reported previously for film-forming acrylate-based nanocomposites prepared using a 4-vinylpyridine auxiliary comonomer in combination with an ultrafine aqueous silica sol.18 The poly(methyl acrylate)silica nanocomposite film was then immersed in doubly
Langmuir, Vol. 20, No. 6, 2004 2189
Figure 6. FT-IR spectra (KBr disk) recorded for (a) HFextracted poly(methyl methacrylate), (b) poly(methyl methacrylate)/silica nanocomposite (entry 1 in Table 1, 32% silica content), (c) poly(methyl methacrylate)/silica nanocomposite (entry 5 in Table 1, 39% silica), and (d) poly(methyl methacrylate)/silica nanocomposite (entry 8 in Table 1, 54% silica).
distilled water at 20 °C for 24 h. This treatment led to a 19% reduction in transmittance at 500 nm. This is attributed to water uptake by the film, since the initial transparent film becomes white after immersion. Transmittance versus time studies on this water-immersed film (not shown) indicated that a constant reduced transmittance was obtained within 40 min at 20 °C. We have recently collaborated with Titman’s group,23 who utilized a combination of two-dimensional protonsilicon-29 and proton-carbon-13 solid-state NMR correlation spectroscopy to probe the molecular interaction between the silica and the polymer components of one of the polystyrene/silica nanocomposites described in the present study. The aliphatic and aromatic proton signals of the polystyrene chains were resolved, and it was shown that the most likely interaction was between the aromatic rings on the styrene residues and the silanol groups on the surface of the silica particles. However, insufficient spectral resolution precludes such NMR experiments for nanocomposites that contain wholly aliphatic polymers such as poly(methyl methacrylate). FT-IR spectra are shown in Figure 6. The HF-extracted poly(methyl methacrylate) has a strong carbonyl band centered at around 1732 cm-1, as expected. The IR spectra obtained for several poly(methyl methacrylate)/silica nanocomposite particles also have this feature, as expected. In addition, a weak shoulder at a lower wavenumber (1720 cm-1) is also discernible. According to Berquier and Arribart, this shoulder indicates a hydrogenbonding interaction between the carbonyl ester group and the silanol groups on the surface of the silica sol.24 Thus these studies suggest that, at least in the case of the poly(methyl methacrylate)/silica nanocomposite particles, a hydrogen-bonding interaction between the organic and inorganic components is most likely responsible for nanocomposite particle formation. Differential scanning calorimetry was used to study the physical properties of the nanocomposites (see Figure 7), with the poly(methyl methacrylate) obtained from the HF extraction of entry 5 in Table 1 being used as a reference (23) Agarwal, G. K.; Titman, J. J.; Percy, M. J.; Armes, S. P. J. Phys. Chem. B 2003, 107, 12497. (24) Berquier, J.-M.; Arribart, H. Langmuir 1998, 14, 3716.
2190
Langmuir, Vol. 20, No. 6, 2004
Percy et al.
in the course of preparing this paper Arrighi and coworkers30 have reported that adding ultrafine silica nanoparticles to a styrene-butadiene rubber either can have little effect on the Tg or can lower the Tg, depending on the surface properties (i.e., hydrophobicity) of the silica sol. Thus it is possible that the apparently anomalous behavior of the Highlink silica-based nanocomposites simply reflects the greater hydrophobicity of this particular sol.
Figure 7. Differential scanning calorimetry curves for (a) HFextracted poly(methyl methacrylate) obtained from poly(methyl methacrylate)/silica nanocomposite (entry 5 in Table 1), (b) poly(methyl methacrylate)/silica nanocomposite (entry 1 in Table 1), (c) poly(methyl methacrylate)/silica nanocomposite (entry 5 in Table 1), and (d) poly(methyl methacrylate)/silica nanocomposite (entry 8 in Table 1).
material. The glass transition temperature, Tg, of a polymer corresponds to the onset of significant segmental chain motion; hence this property is sensitive to the local environment and flexibility of the polymer chains. According to a number of reports concerning polymer-based nanocomposites, the change in specific heat observed at the Tg of the organic phase is usually reduced relative to that found for the pure polymer.26-28 This is simply because the inorganic phase acts as a diluent, reducing the sensitivity of the DSC measurement. In the present study, the HF-extracted poly(methyl methacrylate) had an onset Tg of around 114 °C, which is in reasonable agreement with the literature.25 As expected, this phase transition was shifted to 121 °C in the case of two poly(methyl methacrylate)/silica nanocomposites (entries 5 and 8 in Table 1). This shift is usually interpreted in terms of reduced mobility of the polymer chains due to their interaction with the inorganic phase.29 However, for the poly(methyl methacrylate)/silica prepared with the Highlink silica (entry 1 in Table 1) the onset Tg was reduced to 111 °C. Very similar results were obtained with the corresponding polystyrene/silica nanocomposites, see Figure 1 in the Supporting Information. A charge-stabilized polystyrene latex prepared in the absence of any silica sol had an onset Tg of 92 °C and was used as a reference material. The two nanocomposites prepared using the IPAST and MA-ST-M silica sols had higher Tg’s of 93 and 95 °C, respectively, whereas the nanocomposite prepared with the Highlink silica sol had a lower onset Tg of 86 °C. We are unable to explain these unexpected results at the present time, although we note that Hajji et al. have reported similarly perplexing observations.26 However, (25) Colom, X.; Garcia, T.; Sunol, J. J.; Saurina, J.; Carrasco, F. J. Non-Cryst. Solids 2001, 287, 308. (26) Hajji, P.; David, L.; Gerard, J. F.; Pascault, J. P.; Vigier, G. J. Polym. Sci., Polym. Phys. Ed. 1999, 37, 3172. (27) Motomatsu, M.; Takahashi, T.; Nie, H.-Y.; Mizutani, W.; Tokumoto, H. Polymer 1997, 38, 177. (28) Vaia, R. A.; Ishii, H.; Giannelis, E. P. Chem. Mater. 1993, 5, 1694. (29) Porter, C. E.; Blum, F. D. Macromolecules 2000, 33, 7016.
Conclusions In summary, the convenient synthesis of a range of new colloidal polymer-silica nanocomposites using three different ultrafine alcoholic silica sols is described. No surface pretreatments or auxiliary comonomers were required: styrene, methyl methacrylate, methyl acrylate, n-butyl acrylate, and 2-hydroxypropyl methacrylate were simply homopolymerized in turn in the presence of the silica sol to produce colloidally stable inorganic/organic hybrid particles with reasonably narrow size distributions and silica contents of up to 58% by mass, as indicated by thermogravimetric analysis. Most of the colloidal nanocomposites comprised approximately spherical particles with relatively narrow size distributions, but in some cases more polydisperse or nonspherical particles were obtained. Acrylate-based nanocomposites proved to be film-forming, and scanning electron microscopy studies indicated complete loss of the original particle morphology in these films. Selected nanocomposite films exhibited transmittances of more than 93% in the visible spectrum, but their optical clarity was compromised by immersion in water. X-ray photoelectron spectroscopy studies indicated that the surface compositions of selected nanocomposite particles are silica-rich, which is consistent with their long-term colloidal stability and also with aqueous electrophoresis measurements. HF etching of the nanocomposites allowed the organic component to be isolated for both spectroscopic and GPC analysis. FT-IR studies suggested that, in the case of the poly(methyl methacrylate)-silica nanocomposite particles, the carbonyl ester groups in the polymer are hydrogen-bonded to the surface silanol groups. According to differential scanning calorimetry studies, the Tg’s of several poly(methyl methacrylate)-silica and polystyrene-silica nanocomposites can be either higher or lower than the Tg of the corresponding homopolymer, depending on the nature of the silica sol. Acknowledgment. S.P.A. acknowledges EPSRC for a ROPA postdoctoral research grant (GR/R79814) that supported M.J.P. and J.I.A. The Comisio´n de Investigaciones Cientı´ficas de la Provincia de Buenos Aires (CICPBA, Argentina) is thanked for allowing J.I.A. a oneyear secondment at the University of Sussex. Clariant (France) and Nissan Chemicals (U.S.) are thanked for the donation of the ultrafine silica sols. Dr. P. McKenna of Cognis Performance Chemicals (Hythe, U.K.) is acknowledged for the kind donation of the HPMA monomer. Supporting Information Available: Differential scanning calorimetry curves for polystyrene latex obtained by emulsifier-free emulsion polymerization and polystyrene/silica nanocomposites (entries 2, 6, and 9 in Table 1). This material is available free of charge via the Internet at http://pubs.acs.org. LA035868S (30) Arrighi, V.; McEwen, I. J.; Qian, H.; Serrano Prieto, M. B. Polymer 2003, 44, 6259.
