pubs.acs.org/Langmuir © 2009 American Chemical Society
Segmental Dynamics in Poly(methyl acrylate) on Silica: Effect of Surface Treatment Burak Metin† and Frank D. Blum*,‡ Departments of Chemistry and Materials Science and Engineering, Missouri University of Science and Technology, Rolla, Missouri 65409-0010. †Current Address: Solopower Inc., 5981 Optical Court, San Jose, CA 95138. ‡Address as of Jan. 2010, Department of Chemistry, Oklahoma State University, Stillwater, OK 74078 Received September 30, 2009. Revised Manuscript Received November 3, 2009 The effect of surface treatment on the dynamics of adsorbed poly(methyl acrylate) (PMA) was studied using deuterium NMR and temperature-modulated differential scanning calorimetry (TMDSC). The solid-state deuterium NMR experiments were performed using PMA-d3, deuterated on the methyl group. The line shape changes for PMA-d3 were followed as a function of temperature and compared for the polymer on untreated silica, organically modified (treated) silica (reacted with hexamethyltrisilazane), and in bulk. The dynamics of PMA-d3 on treated silica was found to be intermediate between that of the polymer adsorbed on untreated silica and that of the bulk polymer, i.e., the treated silica caused a restriction on the dynamics of the polymer as compared to bulk, but not as dramatically as that on untreated silica. Similar to the dynamics on untreated silica, the dynamics on treated silica showed a broad heterogeneity with a superposition of more-mobile and less-mobile components. Two molecular mass samples were also studied (38 and 77 kDa) with the molecular mass dependence on the treated or untreated silica being weaker than that in bulk. The TMDSC thermograms of the samples were consistent with the NMR results, with the glass transition region for the PMA-d3 on the treated silica being in between that of the bulk and that on the untreated silica.
Introduction The behavior of polymers at interfaces is of great importance to a variety of applications, and several techniques have been used to study them.1-5 Of particular focus has been the determination of the glass transition temperature (Tg) of interfacial polymers. It has been found that the Tg depends on the nature of the interface (gas, liquid, or solid substrate) and the interaction of the polymer with the other species (e.g., nature of the substrate). Much of the research to date has been performed on the behavior of thin films, either supported or unsupported. Techniques such as ellipsometry,3 calorimetry,6,5 and optical techniques7,8 have been effective for films, including those containing particles.9 There is also another regime of interest, namely, that of polymers adsorbed on particles and in porous media. These particle systems can be probed with some of the same techniques, such as calorimetry,10,11 or different ones, such as transmission *Corresponding author. E-mail: fblum@mst.edu. (1) Fleer, G. J.; Cohen-Stuart, M. A.; Scheutjens, J. M. H. M.; Cosgrove, T.; Vincent, B. Polymers at Interfaces; Chapman & Hall: London, 1993. (2) Wool, R. P. Polymer Interfaces: Structure and Strength; Hanser Publishers: New York, 1995. (3) Forrest, J. A.; Dalnoki-Veress, K. Adv. Colloid Interface Sci. 2001, 94, 167– 196. (4) Alcoutlabi, M.; McKenna, G. B. J. Phys.-Condens. Matter 2005, 17, R461– R524. (5) Fryer, D. S.; Peters, R. D.; Kim, E. J.; Tomaszewski, J. E.; de Pablo, J. J.; Nealey, P. F.; White, C. C.; Wu, W. L. Macromolecules 2001, 34, 5627–5634. (6) Fryer, D. S.; Nealey, P. F.; de Pablo, J. J. Macromolecules 2000, 33, 6439– 6447. (7) Frank, B.; Gast, A. P.; Russell, T. P.; Brown, H. R.; Hawker, C. Macromolecules 1996, 29, 6531–6534. (8) Ellison, C. J.; Torkelson, J. M. Nat. Mater. 2003, 2, 695–700. (9) Rittigstein, P.; Torkelson, J. M. J. Polym. Sci., B: Polym. Phys. 2006, 44, 2935–2943. (10) Sargsyan, A.; Tonoyan, A.; Davtyan, S.; Schick, C. Eur. Polym. J. 2007, 43, 3113–3127. (11) Blum, F. D.; Krisanangkura, P. Thermochim. Acta 2009, 492, 55–60. (12) Fontana, B. J.; Thomas, J. R. J. Phys. Chem. 1961, 65, 480–487. (13) Kulkeratiyut, S.; Kulkeratiyut, S.; Blum, F. D. J. Polym. Sci., B: Polym. Phys. 2006, 44, 2071–2078.
5226 DOI: 10.1021/la903705p
infrared spectroscopy,12,13 magnetic resonance,14-18 or electron spin resonance.19 These latter techniques can, in favorable circumstances, probe the behavior of very small amounts of adsorbed species at interfaces and in materials, especially when the materials are not optically clear. For a polymer such as poly(methyl acrylate) (PMA) on silica, it has previously been shown (with solid-state deuterium NMR) that some segments of adsorbed PMA-d3 had enhanced while others had reduced segmental mobility.20,21 Consequently, the adsorbed polymer had Tg’s that were higher or lower than those for the bulk polymer. The regions of mobility that were different from those of bulk were assigned to those segments near the silica surface and at the interface with air. The location of the (few) more-mobile segments, at the air interface, was verified by overlaying the adsorbed polymer at the interface with unlabeled polymers.22 The less-mobile segments, at the substrate interface, had mobilities that were consistent with the attractive interaction between the polymer carbonyls and surface silanols. The behavior of adsorbed PMA was consistent with that of poly(methyl methacrylate) (PMMA)23 and poly(vinyl acetate) (PVAc)24 on silica, as verified by NMR and calorimetry. The primary reason for the decreased mobility in all of these systems was the H-bonding through the carbonyl groups on the polymers.13 (14) Cosgrove, T.; Griffiths, P. C. Adv. Colloid Interface Sci. 1992, 42, 175–204. (15) Blum, F. D. Annu. Rep. NMR Spectrosc. 1994, 28, 277–321. (16) Mirau, P. A.; Heffner, S. A. Macromolecules 1999, 32, 4912–4916. (17) Ayalur-Karunakaran, S.; Bl€umich, B.; Stapf, S. Eur. Phys. J. E 2008, 26, 43–53. (18) Fortier-McGill, B.; Reven, L. Macromolecules 2009, 42, 247–254. (19) Hommel, H. Adv. Colloid Interface Sci. 2008, 141, 1–23. (20) Lin, W.-Y.; Blum, F. D. Macromolecules 1997, 30, 5331–5338. (21) Lin, W.-Y.; Blum, F. D. Macromolecules 1998, 31, 4135–4142. (22) Lin, W. Y.; Blum, F. D. J. Am. Chem. Soc. 2001, 123, 2032–2037. (23) Blum, F., D.; Young, E., N.; Smith, G.; Sitton, O., C. Langmuir 2006, 22, 4741–4. (24) Blum, F. D.; Xu, G.; Liang, M.; Wade, C. G. Macromolecules 1996, 29, 8740–8745.
Published on Web 12/01/2009
Langmuir 2010, 26(7), 5226–5231
Metin and Blum
Article
Thermal analysis studies of these systems were consistent with the NMR results. The effects of molecular mass were also shown to have unusual trends with respect to the Tg of the adsorbed polymer, i.e., a polymer of intermediate molecular mass showed the most restricted mobility.25 In the present report, we extend our knowledge of the behavior of adsorbed PMA on silica with a focus on how the surface treatment of silica with hexamethyldisilazane affects the dynamics of adsorbed PMA. Treatment with hexamethyldisilazane methylates the surface of the silica and changes it from hydrophilic to hydrophobic.26 We have previously reported Fourier transform infrared (FTIR) and calorimetric studies of PMMA on treated silica27 and found that there was likely some residual H-bonding between the polymer and the treated silica. Fryer et al.5 also used local thermal analysis for PMMA on surfaces with different treatments and found that the Tg of the adsorbed polymer scaled with the surface energy of the substrate. With PMA-d3, we probed the effect of the methylation of the silica on the polymer dynamics and also correlated the results with temperature-modulated differential scanning calorimetry (TMDSC).
occur from one site to one of the three nearest neighbors with equal probability. To a reasonable extent, the soccer ball model mimics the small angle motions experienced by amorphous polymer segments without substantially increasing the amount of computation required for truly small angle jumps. A set of simulated spectra, composed of 94 individual spectra at different jump rates, was generated using jump rates from 1.0 102 Hz to 1.0 1011 Hz. A mathematical routine was applied to fit the experimental line shapes to a superposition of simulated spectra with MATLAB (The Mathworks, Inc., Natick, MA). Details of the line-shape simulation are given elsewhere.28 The TMDSC experiments were preformed on a TA Instruments 2920 modulated differential scanning calorimeter. A heating rate of 3 K/min and a modulation of ( 0.5 K every 50 s were used. The Universal Analysis software package was used to process the data that is displayed as the derivative of the reversing heat flow rate. The derivative mode makes broad transitions easier to discern, and the nonreversing component has significant noise and is not as useful in observing the glass transition region for low cases.
Experimental Section
The deuterium NMR spectra of the bulk PMA-38K sample are shown in Figure 1a at different temperatures. The dots in Figure 1 (and Figure 2) represent the fits to the experimental line shapes. A Pake powder pattern was present at 25 C for the bulk PMA-38K sample. At 35 C, evidence of motion appeared as a broadening of the “horns” of the powder pattern with the middle partially filled in. The powder pattern collapsed to a fairly broad triangle-shaped resonance at 45 C. This resonance became narrower with further increases in temperature. No residual powder pattern was apparent in the spectrum at higher temperatures, up to 80 C. The 2H NMR spectra of the PMA-38K sample adsorbed (0.84 mg/m2) on an untreated (strongly adsorbing) silica surface are shown in Figure 1b. A rigid powder pattern appeared from 25 to 45 C with no significant distortion in the line shape. At 55 C, a motionally narrowed component appeared in the middle of the spectrum. The intensity of this mobile component increased with temperature. However, a significant amount of residual solid powder pattern was present even at temperatures of 60, 70, and 80 C. Thus, the sample showed characteristics of motional heterogeneity in at least two separate regimes. The 2H NMR line shapes of the PMA-38K sample (0.82 mg/ 2 m ) adsorbed on a treated (weakly adsorbing) silica surface are shown in Figure 1c. No significant changes were observed in the powder patterns with an increase in temperature from 25 to 35 C. However, at 45 C, the powder pattern started to collapse, and a motionally narrowed component appeared in the middle of the spectrum. A further increase in temperature resulted in an increase in intensity of the more-mobile, middle component. The residual solid powder pattern lasted until the temperature reached 80 C, when its intensity became very weak. A higher molecular-mass sample, PMA-77K, was also studied using 2H NMR. The spectra for the bulk PMA-77K samples at different temperatures are shown in Figure 2a. The temperature region at which the powder pattern collapsed to a single resonance shifted to a higher temperature as a result of a molecular-mass effect. Around 45 C, effects of the increasing mobility of the segments were seen in the powder pattern with the narrowing in quadrupole splitting and the presence of a middle component in the spectrum. At 55 C, a single broad resonance resulted, and a further increase in temperature resulted in narrower resonances due to increased mobility of the segments. The 2H NMR spectra for the PMA-77K sample (0.90 mg/m2) adsorbed on the untreated (strongly adsorbing) silica surface are
A detailed description of the synthesis of the methyl acrylate-d3 is given elsewhere.28 PMA-38K (Mw=38000 Da) and PMA-77K (Mw=77000 Da) samples, deuterated on the methyl groups, were synthesized by atom transfer radical polymerization (ATRP).29 The polydispersities were 1.15 and 1.26, respectively. Amorphous untreated fumed silica (M-5 grade, Cabot Corp., Tuscola, IL) was dried in an oven at 450 C for 24 h before use. Amorphous treated silica (TS-530 grade, Cabot Corp., Tuscola, IL) was dried at 110 C to prevent the degradation of the treated groups on the silica surface. For TS-530, most of the surface-hydroxyl groups on M-5 grade silica (strongly adsorbing surface) were replaced with trimethylsilyl groups on the TS-530 grade silica (weakly adsorbing surface).30 Solutions of PMA in toluene were prepared and mixed with silica samples in separate tubes to prepare the adsorbed samples. These mixtures were placed on a mechanical shaker for 48 h and centrifuged for an hour after that. The supernatant liquids were decanted, and the solid-containing portions were kept in a vacuum oven at 70 C for 36 h. Thermogravimetric analysis was used to determine the adsorbed amounts of polymer samples. NMR spectra were obtained using a VARIAN VXR-400/S spectrometer. The quadrupole-echo pulse sequence (delay-90y-τ90x-τ-acquisition) was used with the 2H frequency at 61.39 MHz. The 90 pulse width was 2.85 μs with an echo time (τ) of 30 μs. Two hundred fifty-six scans were used for the bulk polymers and 4096-8192 scans were used for the adsorbed samples. The experimental NMR line shapes were fit using a series of simulated line shapes that were produced using the MXQET31,32 program. Experimental NMR settings (given above) and a reduced quadrupole-coupling constant (QCC) of 55 kHz were used in the simulation program. The reduced QCC accounts for the reductions due to the fast rotation of the methyl groups. The simulated line shapes were generated using a jump model based on truncated icosahedron (soccer ball) geometry.28 The jump sites were defined as the vertices of a soccer ball, and the jumps would (25) Metin, B.; Blum, F. D. J. Chem. Phys. 2006, 125, 054707/1–054707/9. (26) Slavov, S. V.; Sanger, A. R.; Chuang, K. T. J. Phys. Chem. B 2000, 104, 983–989. (27) Kabomo, M. T.; Blum, F. D.; Kulkeratiyut, S.; Kulkeratiyut, S.; Krisanangkura, P. J. Polym. Sci., B: Polym. Phys. 2008, 46, 649–658. (28) Metin, B.; Blum, F. D. J. Chem. Phys. 2006, 124, 054908/1–054908/10. (29) Xia, J.; Matyjaszewski, K. Macromolecules 1997, 30, 7692–7696. (30) Anonymous CAB-O-SIL Untreated Fumed Silica, Properties and Function; Cabot Corp.: Tuscola, IL, 2000. (31) Greenfield, M. S.; Ronemus, A. D.; Vold, R. L.; Vold, R. R.; Ellis, P. D.; Raidy, T. E. J. Magn. Reson. 1969, 72, 89–107. (32) Vold, R. R.; Vold, R. L. Adv. Magn. Opt. Reson. 1991, 16, 85–171.
Langmuir 2010, 26(7), 5226–5231
Results
DOI: 10.1021/la903705p
5227
Article
Metin and Blum
Figure 1. 2H NMR spectra of the PMA-38K sample at different temperatures for (a) bulk, and adsorbed on (b) untreated (0.84 mg/m2) and (c) treated (0.82 mg/m2) silica surfaces. Fits of the experimental line shapes are shown as dots.
Figure 2. 2H NMR spectra of the PMA-77K sample at different temperatures for (a) bulk, and adsorbed on (b) untreated (0.90 mg/m2) and (c) treated (0.84 mg/m2) silica surfaces. Fits of the experimental line shapes are shown as dots.
shown in Figure 2b. A powder pattern for a rigid sample was present at 25 C. Further increases in temperature did not result in any significant change in the NMR line shapes until the temperature reached 60 C, when a mobile component appeared in the powder pattern and its intensity increased with temperature. Above 60 C, the line shapes consisted of a superposition of a more-mobile (middle) component and a less-mobile (residual) powder pattern. The residual powder pattern was present in the spectra due to the highly restricted conformations of the segments on the strongly adsorbing (untreated) silica surface. The NMR line shapes of the PMA-77K sample (0.84 mg/m2) adsorbed on treated (weakly adsorbing) silica are shown in 5228 DOI: 10.1021/la903705p
Figure 2c. The rigid powder pattern observed at 25 C became slightly distorted at 35 C. The horns became less sharp, and the distance between them (quadrupole splitting) decreased due to an increase in mobility. A more-mobile (middle) component appeared at 45 C in the powder pattern. Intensity in the middle, as a result of more-mobile components, increased with temperature. The spectrum was again a superposition of more-mobile and less-mobile components as in Figure 2b. Although the powder pattern collapsed to a single resonance at 80 C, a very small amount of residual powder pattern was still present. Plots of the averages of the logarithms of the jump rates calculated based on the simulation data,28 are shown in Langmuir 2010, 26(7), 5226–5231
Metin and Blum
Article Table 1. Tg Values and Widths from TMDSC for Bulk and Adsorbed PMA-d3 TMDSC adsorbed on bulk
Figure 3. Plots of the average of the log jump rates (Ælog kæ) for PMA-38K (open symbols) and PMA-77K (open symbols) on treated (squares) and untreated (triangles) silica and in bulk (circles).
Figure 4. TMDSC thermograms for the bulk and adsorbed PMA samples on treated and untreated silica surfaces.
Figure 3 for the molecular mass PMA-d3 polymers on both surfaces. For comparison, the values for the bulk polymer are also shown.28 The Ælog kæ values were calculated as Ælog kæ ¼
X
ai logðkÞ
ð1Þ
where the sum is over the simulated spectra with different jump rates (I=1, 94) used for the fitted spectra, ai is the contribution (intensity) from each jump rate, and k is the jump rate. At the higher temperatures studied, the Ælog kæ values for all samples scaled as expected, with the PMA-d3 on the treated silica showing faster motions than those on the untreated silica. The values for the 38k samples were faster than those for the 77k samples at the same temperatures. In the lower temperature range, the Ælog kæ values were more or less indistinguishable, as significant fractions of the polymers had motions in the slow limit, for which the powder patterns were insensitive. In the region where the line shapes were sensitive to the motions of PMA-d3, the state of the sample, for both molecular masses, was ordered such that Ælog kæ(bulk) was greater than Ælog kæ(treated), which was greater than Ælog kæ(untreated). The results of the TMDSC experiments are shown in Figure 4 for both PMA-38K and PMA-77K samples. The results are plotted as the derivative of the reversing heat flow rate so that Langmuir 2010, 26(7), 5226–5231
treated silica
polymer
Tg (C)
width (C)
Tg (C)
width (c)
PMA-38K PMA-77K
3.5 12.1
8.9 7.0
17.8 19.9
10.1 9.5
untreated silica Tg (C) 24.3, 42.3 23,2, 40.4, 60.7
width (c) 35.1 39.6
the width, especially the broader widths, can be observed. The maxima in the derivative curves were taken as the reported Tg’s of the samples, and the widths were taken as the full widths at half height. It is noted that the glass transitions occurred over a range of temperatures; in fact, a particularly wide range was observed for the strongly adsorbed samples. The details of the TMDSC transitions are listed in Table 1. The derivative of the reversing heat flow curve for the bulk PMA-38K sample peaked at a Tg of 3.5 C. The curve was very broad for the PMA-38K sample adsorbed on the strongly adsorbing (untreated) surface with significant peaks of Tg at 23.3 and 42.3 C. The Tg occurred at 17.8 C for the same sample on the treated (weakly adsorbing) silica surface. The width of the transition on the treated silica was broader than bulk, but not as broad as that on the untreated silica. The TMDSC curve for the PMA-77K samples were similar to those for PMA-38K. For the bulk PMA-77K sample, the Tg was 12.1 C. In bulk, the role of the chain ends in reducing the Tg was responsible for the PMA-38K sample having a broader transition. Peaks occurred at 23.2, 40.4, and 60.7 C for the sample on an untreated (strongly adsorbing) surface with a very broad transition width. The broader features were also present in the TMDSC thermogram for the PMA-77K adsorbed on the treated (weakly adsorbing) surface with a Tg of 19.9 C.
Discussion A variety of reviews exist on the use of NMR to understand the dynamics of polymers.33-36 In a single crystal and in the absence of motion, where only one C-D bond angle would exist with respect to the external magnetic field, the 2H NMR spectra would be composed of a doublet because of two transitions for an I=1 deuteron. However, for an amorphous polymer, where an isotropic distribution of C-D bond orientations exists, a Pake powder pattern results from the superposition of resonances from each possible orientation. For a methyl deuteron with an axial symmetry, the splitting between the horns (ΔνQ) of the powder pattern will be given by37,38 3 1 ΔνQ ¼ ðe2 qQ=hÞ Æ3cos2 βðtÞ - 1æð3cos2 φ - 1Þ 4 2
ð2Þ
The first term, (e2qQ/h) contains the parameters that define the QCC. The second term, within the brackets (Ææ), represents the average of all orientations (β(t)) that the symmetry axis of the methyl group can make with the magnetic field, and the third term (33) Schmidt-Rohr, K.; Spiess, H. W. Annu. Rep. NMR Spectrosc. 2002, 48, 1– 29. (34) DeAzevedo, E. R.; Bonagamba, T. J.; Reichert, D. Prog. Nucl. Magn. Reson. Spectrosc. 2005, 47, 137–164. (35) Reichert, D. Annu. Rep. NMR Spectrosc. 2005, 55, 159–203. (36) Geppi, M.; Borsacchi, S.; Mollica, G.; Veracini, C. A. Appl. Spectrosc. Rev. 2009, 44, 1–89. (37) Seelig, J. Q. Rev. Biophys. 1977, 10, 353–418. (38) Schmidt-Rohr, K.; Spiess, H. W. Multidimensional Solid-State NMR and Polymers; Academic Press: London/San Diego, 1994.
DOI: 10.1021/la903705p
5229
Article
Metin and Blum
Table 2. Tg Values and Widths from 2H NMR for Bulk and Adsorbed PMA-d3 2
H NMR adsorbed on
bulk polymer PMA-38K PMA-77K
treated silica
Tg (C)
width (C)
Tonset (C)
40 55
35 35
45 55
Tmid (C) 55 70
untreated silica Tonset (C) 45 60
Tmid (C) 60 80
defines the contribution due to the orientation (φ) between C-D bonds and the methyl-group symmetry axis. For φ, being ∼70.5 for a methyl group, the quadrupole splitting is reduced to 1/3 of its original value for a rapidly rotating methyl group. With the asymmetry parameter being approximately 0 for the axially symmetric C-2H bond,38 several other contributions may be dropped, yielding eq 1. The powder pattern for an amorphous polymer collapses to a single liquid-like resonance in a 2H NMR experiment with an increase in temperature. The temperature where the powder pattern collapses can be considered the Tg(NMR). In general, the collapse of the 2H powder pattern in the glass transition region provides information not available from, say, calorimetry, as previously shown for PMA-d320,28 and PVAc-d3.39 For example, the sensitivity of 2H NMR in the vicinity of the glass transition is greater in terms of the temperature range where relevant events occur. The Tg’s for the two bulk PMA-d3 are shown in Table 2 and are considerably higher than the Tg(DSC) because of the differences in the frequencies probed by the two techniques. For the PMA-77K sample, the Tg(DSC) is about 12 C, and the Tg(NMR) is about 55 C. These values are consistent with the differences in the two time scales of the measurements. An estimate of the difference between the temperatures associated with the two time scales may be made from the Williams, Landel, and Ferry (WLF) treatment as40 logðτ=τ0 Þ ¼ - C1 ðT - T0 Þ=ðC2 þ T - T0 Þ
ð3Þ
where the τ’s are the correlation times, and the reference T0 is taken as Tg(DSC). The τ0 is taken as 100 s, and τ is taken as the reciprocal of the quadrupole splitting (about 1/(40 kHz)). Estimates for Tg(NMR) - Tg(DSC) from viscosity (C1 =16.7, C2 = 60),41 NMR (C1 =9, C2 =17),42 and simulation (C1 =19, C2 = 72)43 predict differences of between 38 and 46 K. A 43 K difference is found experimentally, which is in good agreement with the predictions. The 2H NMR spectra of all of the PMA samples adsorbed on untreated (strongly adsorbing) and treated (weakly adsorbing) silica surfaces indicated the heterogeneity of the segmental dynamics in a way that the bulk samples did not. Specifically, in the glass transition region, only part of the powder patterns for the adsorbed samples collapsed, to yield a superposition of narrow (more-mobile) and wide (less-mobile) components at higher temperatures. This difference is due to a motional heterogeneity of the adsorbed polymers at the polymer-air, polymer-silica, and polymer-polymer interfaces.20,24,25 For PMA-d3 on the treated silica, only a small fraction from highly motionally (39) Nambiar, R. R.; Blum, F. D. Macromolecules 2008, 41, 9837–9845. (40) Ferry, J. D. Viscoelastic Properties of Polymers, 3d ed.; Wiley: New York, 1980. (41) Berry, G. C.; Fox, T. G. Adv. Polym. Sci. 1968, 5, 261–357. (42) Gaborieau, M.; Graf, R.; Kahle, S.; Pakula, T.; Spiess, H. W. Macromolecules 2007, 40, 6249–6256. (43) Soldera, A.; Metatla, N. Phys. Rev. E 2006, 74, 061803/1–061803/6.
5230 DOI: 10.1021/la903705p
restricted segments was seen at the highest temperature studied, with a larger fraction easily observed on the treated silica. The configurations of randomly adsorbed polymers can be thought of as composed of the segments in trains, loops, and tails.1 For a polymer with an attractive interaction with the surface, loops and tails are more mobile than trains due to their lack of direct interactions with the surface.44 The mobile components in the spectra of adsorbed polymers are from the moremobile loops and tails, particularly those near the polymer-air interface.22 The residual powder patterns for the adsorbed samples at higher temperatures are due to the presence of trains that are in more or less close contact with the surface at the polymer-silica interface. It is noted that the definition of trains (e.g., the extent that “train” segments extend from the surface) depends on the experimental probe.1,11 The segmental dynamics of the PMA-38K and PMA-77K samples were different on both the untreated and treated surfaces. In bulk, the PMA-38K sample showed the effects of its low molecular mass through a lower Tg and lower temperature collapse of the 2H powder pattern than the higher molecular mass sample did. On either surface, the molecular mass effect was observed, but not as strongly as it appeared in the bulk polymer.28 Even on the treated silica, the molecular mass effect was not as dramatic as that in bulk. It is possible that, on untreated silica, the presence of trains reduces the effect of the chain ends on polymer segmental dynamics. Infrared studies on PMMA suggest the possibility that PMMA may H-bond to some residual surface silanols.27 One study has shown that, after 4 h of reaction of HS-5 Cab-O-Sil45 with HMDS, 10% of the non-H-bonded surface silanols remain. We believe that there may be a few residual H-bonds between PMA-d3 and the treated silica. The presence of a very small amount of rigid material at a higher temperature is consistent with a limited number of hydrogen-bonded trains. For both molecular masses of PMA-d3 on silica, the majority of adsorbed PMA-d3 segments became more rubbery on the treated silica surface at lower temperatures than they did on the untreated surface. In this case, the mobility of PMA-d3 on the treated surface was intermediate between that in bulk and that on the untreated silica. The rubbery nature of many segments is undoubtedly due to the weaker interaction of the polymer with the trimethylsilyl groups on the surface, although there is likely some anchoring to unreacted surface silanols. As an example, the spectrum for the PMA-77K sample on the treated surface at 70 C is similar to the one at 90 C on the untreated surface. The ordering, based on the mobility evident in the spectra, was also evident in the Ælog kæ calculated from the simulated spectra (Figure 3). The PMS-38K bulk jump rates were faster due to the relatively low molecular mass of that polymer, even when it was adsorbed. The Ælog kæ values at low temperatures were insensitive to the slower motions that occurred in the polymers at these temperatures, i.e., the slow motions did not affect the onedimensional (1-D) line shapes. At higher temperatures, differences in Ælog kæ values were obvious. The bulk samples, with no filler to restrict motions, had the fastest jump rates. The PMA-d3 samples on the treated surfaces were intermediate in terms of jump rates. The PMA-d3 samples on the untreated surfaces had much slower jump rates because of the much stronger interaction with the surface silanols. The thermal analysis experiments were quite consistent with the NMR experiments with respect to the molecular mass effect (44) Barnett, K. G.; Cosgrove, T.; Vincent, B.; Cohen-Stuart, M.; Sissons, D. S. Macromolecules 1981, 14, 1018–20. (45) Gun’ko, V. M.; Vedamuthu, M. S.; Henderson, G. L.; Blitz, J. P. J. Colloid Interface Sci. 2000, 228, 157–170.
Langmuir 2010, 26(7), 5226–5231
Metin and Blum
Article
and the effect of surface treatment. Bulk polymers from both experiments had the narrowest ranges and the lowest transition temperatures. The polymer adsorbed on the untreated surface had the highest Tg and was also the broadest. The behavior of the PMA-d3 on the treated surface was intermediate in temperature, but the width of the glass transition region was similar to that found in bulk for DSC. It is difficult to quantify a single Tg(NMR) for the surface polymer, as different segments vary in mobility at different temperatures. In this sense, the NMR experiment was extremely sensitive to the motional heterogeneity in surface samples. To at least partially unravel the Tg behavior, we define the temperature where the first hint of a narrow central resonance is evident, as Tonset. We characterize the point where the intensity of the middle resonance is about half of the total intensity as Tmid. The end of the transition is difficult to determine, as a broad powder pattern is difficult to observe in the presence of an intense narrow resonance. In any case, these working definitions can give an indication of at least half of the breadth of the transition as seen by 2H NMR. These values are tabulated in Table 2 along with information on the bulk polymer. It is interesting to note that Fryer et al.5 observed glass transitions for thin films of PMMA that scaled with the surface energy of the octadecylsilane-treated silicon. For their thinnest films, an increase in the Tg was found for high-energy surfaces. As the surface energy decreased, a lower Tg was found. The ordering of the effect of surface treatment for our samples is similar to their results. However, since their film thicknesses (20 nm and larger) represented much more polymer than we had in our samples (0.8 mg polymer/m2), our results showed greater detail through the glass transition region.
(treated) silica and strongly absorbing (untreated) silica. Both techniques identified glass transition regions consistent with their own frequency range and both identified the bulk polymer as having the lowest and the narrowest Tg. The bulk polymer was also homogeneous with respect to segment location and the onset of local segmental mobility. On the surface, the dynamics of PMA-d3 was heterogeneous with respect to segment position. The segments near the polymer-air interface had the lowest Tg, and those near the polymer-silica interface had the highest Tg (or onset of segmental motions). The breadths of the glass transition regions from either TMDSC or NMR provided evidence as to heterogeneity. The effect of surface treatment was also apparent from both the NMR or TMDSC experiments. Both experiments found that the glass transition for PMA-d3 adsorbed on treated silica (where it was more weakly absorbed) had shifted to lower temperatures and was reduced in breadth as compared to PMA-d3 adsorbed on untreated silica (where it was more strongly adsorbed). The glass transition regions for the PMA-d3 on untreated silica were 4050 K wide, while those for the polymer on treated silica were about 20-30 K wide. Clearly, the polymer adsorbed on treated silica had motional rates in between those of bulk and those on untreated silica. The effect of molecular mass was more prominent in bulk than when it was adsorbed on either treated or untreated silica. The end groups still gave the PMA-38K sample more mobility on either surface than the PMA-77K sample did. However, the molecular mass effect was smaller on the surface, possibly due to either interaction with the trimethylsilyl groups or residual silanol groups.
Conclusions Both 2H NMR and TMDSC have provided important insights for a valid comparison of the glass transition behavior of bulk PMA-d3 and the absorption on weakly absorbing
Langmuir 2010, 26(7), 5226–5231
Acknowledgment. The financial support of the National Science Foundation under Grant DMR-0706197 is acknowledged.
DOI: 10.1021/la903705p
5231