Langmuir 2002, 18, 683-687
683
Buried Interface Modification Using Supercritical Carbon Dioxide Xinqiao Jia and Thomas J. McCarthy* Polymer Science & Engineering Department, University of Massachusetts, Amherst, Massachusetts 01003 Received May 24, 2001. In Final Form: October 30, 2001 Chemical modification of the buried interfaces between silicon wafers and either polystyrene or poly(methyl methacrylate) (PMMA) with the reagent (tridecafluoro-1,1,2,2-tetrahydrooctyl)dimethylchlorosilane (FDCS) was studied using liquid and supercritical carbon dioxide as the solvent and infusing agent. Above the critical point of CO2, FDCS reacts with surface silanols at the SiO2/polystyrene interface to form a monolayer, which was analyzed after removal of the polystyrene film. This is not the case below the critical point, and likely the lower diffusivity and solvent properties of the CO2-swollen film are the causes. The reaction was much less successful at the SiO2/PMMA interface. We suspect that because of strong hydrogen bonding between PMMA and the silicon substrate, modification of the SiO2/PMMA interface is achieved to only a limited extent.
Introduction Supercritical carbon dioxide (scCO2) has tremendous potential for the modification and processing of polymers. Previous studies in this area have used scCO2 for separation, extraction, fractionation,1 swelling,2 polymer impregnation,3 polymeric membrane conditioning,4 and polymer precipitation.5 It also has been used as a reaction medium for polymer modification6 and polymer synthesis,7 and to produce blends,8 semi-interpenetrating networks,9 inorganic/polymer nanocomposites,l0 microcellular foams,11 and fibers.12 scCO2 is a desirable solvent for polymer processing because of its high diffusivity and low viscosity and the absence of surface tension. The solvent strength of scCO2 can be adjusted by simply changing temperature or pressure. Furthermore, it is nontoxic, nonflammable, and inexpensive. After processing, CO2 removal from the polymer is accomplished simply by decreasing the pressure. Chemical modification of inorganic surfaces by covalent attachment of organosilanes has been utilized to improve adhesion of organic coatings to inorganic substrates,13 to protect metal surfaces from corrosive environments,14 to make water repellent surfaces,15 to enhance biocompat(1) McHugh, M. A.; Krukonis, V. J. Supercritical Fluid Extraction: Principles and Practice, 2nd ed.; Butterworth-Heinemann: Boston, MA, 1994. (2) Wissinger, R. G.; Paulaitis, M. G.; Paulaitis, M. E. J. Polym. Sci., Part B: Polym. Phys. 1991, 29, 631-633. (3) Berens, A. R.; Huvard, G. S.; Korsmeyer, R. W.; Kunig, F. W. J. Appl. Polym. Sci. 1992, 46, 231. (4) Hsu, J. H.; Tan, C. S. J. Membr. Sci. 1993, 81, 273. (5) Tom, J. W.; Debenedetti, P. G.; Jerome, R. J. Supercrit. Fluids 1994, 7, 9. (6) Hayes, H. J.; McCarthy, T. J. Macromolecules 1998, 31, 4813. (7) DeSimone, J. M.; Maury, E. E.; Menceloglu, Y. Z.; McClain, J. B.; Romack, T. J.; Combes, J. R. Science 1994, 265, 356. (8) Kung, E.; Lesser, A. J.; McCarthy, T. J. Macromolecules 1998, 31, 4160. (9) Rajagopalan, P.; McCarthy, T. J. Macromolecules 1998, 31, 4791. (10) Watkins, J. J.; McCarthy, T. J. Chem. Mater. 1995, 7, 1991. (11) Arora, K. A.; Lesser, A. J.; McCarthy, T. J. Macromolecules 1998, 31, 4614. (12) Mawson, S.; Johnaton, K. P.; Combes, J. R.; Desimone, J. M. Macromolecules 1995, 28, 3182. (13) Plueddemann, E. P. Silane Coupling Agents, 2nd ed.; Plenum Press: New York, 1991. (14) Kim, H.; Jang, J. Polymer 1998, 39, 4065.
ibility of solid surfaces,16 to tailor the chromatographic characteristics of solid supports,17 and to form monolayers for lithography.18 Further development of this method for the modification of surfaces has led to investigations of the use of monolayers as sensors,19 and for the immobilization of catalysts and other chemically or optically active species.20,21 Monofunctional silanes form covalently attached monolayers on silica surfaces, while trifunctional silanes can form self-assembled monlayers.22,23 Our group has conducted fundamental research on trialkylsilane monolayers that are covalently attached to silicon surfaces under different reaction conditions: in the vapor phase, in toluene solution, and in supercritical carbon dioxide.24,25 The highest contact angles (indicating the densest monolayers) were obtained using vapor phase reactions, and the fastest reactions were found in supercritical CO2. Figure 1 shows the rationale behind the work reported here. Reactions in organic solids (polymers) are invariably mass transport limited because of low diffusivity. We have shown8-l0 that swelling of polymers with scCO2 permits reaction rate-limited chemistry in solids (Figure la). We have also shown25 that monofunctional organosilanes can react reproducibly in liquid and supercritical CO2 with SiO2 to form covalently attached monolayers (Figure lb). Can we “reach inside” composite materials and conduct chemistry at inorganic/organic “buried interfaces” (Figure (15) Chen, W.; Fadeev, A. Y.; Hsieh, M. C.; Oner, D.; Youngblood, J.; McCarthy, T. J. Langmuir 1999, 15, 3395. (16) Rezania, A.; Johnson, R.; Lefkow, A. R.; Healy, K. E. Langmuir 1999, 15, 6931. (17) Unger, K. K. Porous Silica, its Properties and Use as a Support in Column Liquid Chromatography; Journal of Chromatography Library, v.16; Elsevier: Amsterdam, 1979. (18) Dulcey, C. S.; Georger, J. H.; Krauthamer, V.; Stenger, D. A.; Fare, T. L.; Calvert, J. M. Science 1991, 252, 551. (19) Biosensor and chemical sensor technology: process monitoring and control; Mulchandani, A., Zhou, W., Rogers, K. R., Eds.; ACS Symposium Series Vol. 613; American Chemical Society: Washington, DC, 1995. (20) Modified Silanes in Adsorption, Chromatography and Catalysis; Lisichkin, G. V., Ed.; Khimiya: Moscow, 1986. (21) Park, Y. S.; Ito, Y.; Imanishi, Y. Chem. Mater. 1997, 9, 2755. (22) Hoffmann, P. W.; Stelzle, M.; Rabolt, J. F. Langmuir 1997, 13, 1877. (23) Tripp, C. P.; Hair, M. L. Langmuir 1995, 11, 149. (24) Fadeev, A. Y.; McCarthy, T. J. Langmuir 1999, 15, 3759. (25) Cao, C.; McCarthy, T. J. Langmuir 2001, 17, 757.
10.1021/la010768c CCC: $22.00 © 2002 American Chemical Society Published on Web 01/29/2002
684
Langmuir, Vol. 18, No. 3, 2002
Figure 1. Schematic representation of the rationale behind this work: (a) chemistry inside a scCO2-swollen polymer substrate; (b) chemistry at a silica surface using scCO2; (c) chemistry at an inorganic/organic “buried” interface using scCO2.
lc)? This may allow us to rehabilitate polymer coatings and reinforce polymer composites. In this work, we use flat silicon wafers coated with polymer films as model systems to study buried interface modification using (tridecafluoro-1,1,2,2-tetrahydrooctyl)dimethylchlorosilane (FDCS) and scCO2 as a solvent and swellant. The polymer films used were polystyrene and poly(methyl methacrylate) (PMMA), both of which can be effectively plasticized by CO2.26 The choice of FDCS came from two considerations: (1) van der Waals interactions between the fluorinated sites of FDCS and carbon dioxide render high solubility of FDCS in ScCO2.27 (2) The high sensitivity of fluorine in X-ray photoelectron spectroscopy makes it easy to study the extent of surface modification (after dissolution of the film). Reaction conditions and kinetics have been studied, and an explanation for the different results obtained using polystyrene and PMMA as polymer films is proposed. We have demonstrated that, by choosing appropriate conditions, buried interfaces can be modified using supercritical CO2. Experimental Section General. Toluene, ethanol, 2-propanol, tetrahydrofuran, chloroform, cyclohexane, ethyl acetate, acetic acid, methyl ethyl ketone, sulfuric acid, and hydrogen peroxide (30%) were obtained from Aldrich and used as received. (Tridecafluoro-1,1,2,2tetrahydrooctyl)dimethylchlorosilane was obtained from Gelest and used as received. Polystyrene (Mw ) 26.7 K, Mw/MN ) 1.03) and poly(methyl methacrylate) (Mw ) 118 K, Mw/MN ) 2.34) were prepared by standard anionic and free radical procedures, respectively. House-purified (reverse osmosis) water was further purified using a Millipore Milli-Q system that involves reverse osmosis, ion-exchange, and filtration steps (1018 Ω/cm). Substrate Preparation. Silicon wafers were obtained from International Wafer Service (100 orientation, P/B doped, resistivity from 20 to 40 Ω cm). The thickness of the native silicon oxide was determined from ellipsometry to be 2.0-2.5 nm. Disks were cut into 1.5 cm × 1.2 cm samples. The samples were held in a custom designed (slotted glass cylinder) holder and were rinsed with water and submerged in a freshly prepared mixture of 7 parts concentrated sulfuric acid containing dissolved sodium dichromate (∼4 wt %) and 3 parts 30% hydrogen peroxide. The solution turns from red-brown to green, warms to 80-90 °C, and foams extensively because of the formation of oxygen and ozone. The wafers were submerged in the solution overnight, rinsed (26) Condo, P. D.; Paul, D. R.; Johnston, K. P. Macromolecules 1994, 27, 365. (27) Dardin, A.; DeSimone, J. M.; Samulski, E. T. J. Phys. Chem. B 1998, 102, 1775.
Jia and McCarthy with aliquots of water, and placed in a clean oven at 120 °C for 2 h. Polymer films were spun-cast from toluene solution (7 wt %) on the silicon surfaces immediately after treating the plates in this fashion. Polymer film thickness was maintained at ∼3200 Å. Reaction of FDCS at the SiO2/Polymer Interface in CO2. A silicon wafer coated with polymer was placed in a 316 stainless steel high-pressure reaction vessel, which was sealed, purged with CO2, and equilibrated at the desired reaction temperature. The reaction vessel was subsequently filled with CO2 at the desired reaction pressure. FDCS (350 µL) was dissolved in scCO2 in a mixing unit at lower pressure to ensure complete dissolution of FDCS in CO2. The solution was then transferred to the reaction vessel by CO2 at the pressure desired, using an ISCO syringe pump. The vessel was immersed in a circulating controlledtemperature bath for the desired time. After the reaction, the FDCS/CO2 solution was released and the sample was thoroughly rinsed with cyclohexane (40 °C), toluene, tetrahydrofuran, and ethyl acetate (in this order) to remove the polystyrene overlayer, or with toluene, methyl ethyl ketone, acetic acid, and chloroform (in this order) to remove the PMMA film. Samples were subsequently sonicated in toluene for 20 min and rinsed with 2-propanol, ethanol, ethanol/water (1:1), and water (in this order). The samples were dried in an oven (120 °C) for 10 min before the ellipsometry, contact angle, and XPS measurements. Characterization. X-ray photoelectron spectroscopy (XPS) was performed with a Perkin-ElmersPhysical Electronics 5100 with Mg KR excitation (400 W). Spectra were obtained at two different takeoff angles, 15° and 75° (between the plane of the surface and the entrance lens of the detector optics). Contact angle measurements were made with a Ram6-Hart telescopic goniometer and a Gilmont syringe with a 24-gauge flat-tipped needle. The probe fluids used were water, purified as described above, and hexadecane, purified by vacuum distillation. Dynamic advancing (θA) and receding angles (θR) were recorded while the probe fluid was added to and withdrawn from the drop, respectively. Ellipsometry was performed using a Rudolph Research AutoEL-II automatic ellipsometer. The light source was a He-Ne laser (λ ) 6328 Å), the incident angle was 70.0°, and the compensator was set at -45°. The fit of tan φ and cos ∆ allows calculation of thickness using dafIBM software. A doublelayer model was used to calculate the monolayer thickness, with the lower layer native silicon oxide and the upper layer fluorinated silane. We chose 1.462 and 1.345 as refractive indices for the two layers, respectively.
Results and Discussion Reactions at SiO2/Polystyrene Interfaces. Previous research has shown that monofunctional dimethylsilanes react with surface silanols to form covalently attached monolayers.24,25 In the absence of a nitrogen-containing base, room temperature reaction of alkylchlorosilanes generally proceeds by a two-step mechanism. The chlorosilanes are first hydrolyzed by water (either on the surface or in solution) to form silanols, which subsequently condense onto the surface. Fluoroalkylchlorosilanes can react directly with the surface silanol groups at room temperature without the presence of either base or water. The presence of water, however, leads to faster reaction.28 Equation 1 shows this chemistry using FDCS. The extent
of reaction at the buried interface depends on several factors: the diffusivity of the silane reagent in the CO2plasticized polymer film, the accessibility of the silanol groups on the surface, and the reaction kinetics of FDCS with the surface silanols. The pressure and temperature (28) Tripp, C. P.; Veregin, R. P. N.; Hair, M. L. Langmuir 1993, 9, 3518.
Buried Interface Modification Using scCO2
Langmuir, Vol. 18, No. 3, 2002 685
Table 1. Characterization of SiO2/PS Interfaces Modified with FDCS in Liquid CO2 (1100 psi, 23 °C) reaction time (h)
monolayer thickness (Å)
1
1.5
3
3
5
4
atomic concentrationb (%)
θA/θRa (deg)
Si
C
O
F
39/17 18/15 36/18 16/14 37/17 17/14
10.86 33.44 12.56 36.63 8.77 36.01
50.15 15.60 55.73 12.61 56.13 15.91
36.23 50.81 29.40 50.76 32.17 45.69
2.76 0.15 2.31 0.00 2.92 2.39
a
Upper rows are water contact angles, and lower rows are hexadecane contact angles. b Upper rows are 15° takeoff angle data, and lower rows are 75° takeoff angle data.
of CO2 must be chosen to result in effective plasticization of the polymer film so that the silane reagent (dissolved in CO2) is able to reach the interface. The silanols at the interface likely have different reactivity; those beneath a weakly bound layer of polymer film are more likely to react with FDCS, and those that interact strongly with the polymer film are likely less accessible to FDCS. Our cleaning process for the silicon wafer not only removes organic impurities but also results in a homogeneous and fully hydrated surface, with four to five SiOH per square nanometer after the sample is dried at 120 °C.29 Tripp and co-workers30 point out that, unlike conventional solvents, scCO2 extracts adsorbed water from the silica surface. Since FDCS can react with surface silanols with or without water, and the physisorbed water is buried by the spun-cast polymer film, scCO2 extraction should not have any detectable effect on the reaction. Control experiments were carried out to confirm that surface fluorine detected by XPS is solely the result of chemistry at the interface in scCO2. After reaction in CO2 (2500 psi, 40 °C, 8.3 h), XPS was performed on the sample without dissolving the polymer (PS) film. Fluorine contents of 1.45% at a 15° takeoff angle (assesses the composition of the outer ∼10 Å) and 0.24% at a 75° takeoff angle (assesses the composition of the outer ∼40 Å) were measured. This indicates that there is almost no FDCS absorbed in the polymer film after venting (CO2 effectively extracts the FDCS and thus eliminates the possibility of the solution reaction between FDCS and the silicon wafer during the subsequent washing process). Uniform and complete polymer films (a small amount of dewetting was observed by optical microscopy) were recovered if CO2 was released very slowly (over a period of 10-15 h). Faster decompression caused foaming and delamination, and this was used to facilitate the washing process. Table 1 summarizes XPS and contact angle results for reaction at 1100 psi and 23 °C. After the reaction, the polystyrene film was dissolved to analyze the interface. Within 5 h, only a very small amount of fluorine was detected on the surface. The contact angles for both water and hexadecane are low, indicating a SiO2-rich surface. The critical point of CO2 is 31.1 °C and 1070 psi. At 1100 psi and 23 °C, CO2 is liquid and it does not function as an effective enough plasticizer. The lack of chain mobility makes it difficult for FDCS to diffuse to the buried interface and difficult for chain segments adsorbed to the interface to rotate and accommodate reaction with the silane. Buried interface modification was readily achieved by choosing higher temperatures and pressures, and the conditions of 1200 psi and 35 °C proved effective (Table (29) Iler, R. K. The Chemistry of Silica; John Wiley & Sons: New York, 1979. (30) Tripp, C. P.; Combes, J. R. Langmuir 1998, 14, 7348.
Table 2. Characterization of SiO2/PS Interfaces Modified with FDCS in scCO2 (1200 psi, 35 °C) reaction time (h)
monolayer thickness (Å)
1
2
3.2
6.3
5.7
7.8
8
7
atomic concentrationb (%)
θA/θRa (deg)
Si
C
O
F
51/37 26/15 72/59 31/22 73/59 33/24 76/60 37/24
17.37 39.38 17.35 38.61 16.62 36.13 14.41 35.32
39.35 9.06 38.63 11.52 35.77 11.76 39.44 12.13
37.45 49.65 30.76 46.07 32.03 45.63 30.10 44.36
5.83 1.91 13.26 3.80 15.58 6.48 16.15 8.19
a Upper rows are water contact angles, and lower rows are hexadecane contact angles. b Upper rows are 15° takeoff angle data, and lower rows are 75° takeoff angle data.
Figure 2. Contact angle characterization of the chemical modification at the SiO2/polystyrene interface with scCO2 (1200 psi, 35 °C): (b) advancing water contact angle; (O) receding water contact angle; (9) advancing hexadecane contact angle; (0) receding hexadecane contact angle.
2). scCO2 is known to reduce the Tg of many glassy polymers. We emphasize that it is possible to change the degree of swelling and the diffusion rate by slight changes in the system pressure and temperature; the polymer free volume can easily be tuned. CO2 dissolved in the glassy polymer significantly enhances the mass transfer properties of solutes in the polymer matrix. We have reported25 that the formation of monolayers in CO2 is insensitive to pressure and temperature changes and that the reaction is very fast, complete within ∼20 min. The reaction is slower in the case of the buried interface. It takes 3 h for the interface to become moderately hydrophobic, with θA/θR ) 72°/59°. Longer reaction times lead to a further increase in both fluorine content and contact angles, but with a tendency to reach a plateau. This indicates that the buried interface between the silicon wafer and polystyrene is saturated with the reactive silane reagent under these specific conditions. Although there is no strong interaction between polystyrene and the silicon wafer, its presence hinders the formation of a monolayer to some extent. The reaction of chlorosilanes with silica is highly solvent-dependent,24 and CO2-swollen polystyrene is not the best solvent for this reaction. CO2 alone25 and toluene24 are superior. There is a natural relationship between the contact angles and the fluorine concentrations of the modified interfaces that can be seen in Figures 2 and 3. To examine the effect of pressure and temperature on the reaction, the reaction was carried out at higher CO2 pressure and temperature (2500 psi, 40 °C). After reaction
686
Langmuir, Vol. 18, No. 3, 2002
Figure 3. Kinetics of FDCS chemical modification at the SiO2/ PS interface with scCO2 at 1200 psi and 35 °C monitored by XPS fluorine atomic concentration at a 15° takeoff angle (9) and a 75° takeoff angle (0).
Jia and McCarthy
Figure 5. Schematic of the formation of hydrogen bonds between the PMMA film and the silicon substrate. Table 3. Characterization of SiO2/PMMA Interfaces Modified with FDCS in scCO2 under Different Conditionsa Pb (psi)
Tc (°C)
monolayer thickness (Å)
1200
35
3
2500
40
8
4500
50
6
θA/θRd (deg) 47/34 26/15 62/32 24/16 48/30 21/14
atomic concentratione (%) Si
C
O
F
19.94 36.19 12.34 34.69 14.53 29.86
29.65 7.25 48.33 16.20 45.43 18.39
43.47 52.58 32.44 45.92 30.43 46.70
6.94 3.97 6.89 3.19 9.61 5.05
a Reaction time: 6 h. b P: reaction pressure. c T: reaction temperature. d Upper rows are water contact angles, and lower rows are hexadecane contact angles. e Upper rows are 15° takeoff angle data, and lower rows are 75° takeoff angle data.
Figure 4. Comparison of the reaction kinetics at the SiO2/PS interface using scCO2 at 1200 psi, 35 °C (O) and 2500 psi, 40 °C (b) (15° takeoff angle data).
for 8.3 h, water contact angles as high as 95°/76° were obtained. Using the Israelachvili equation (eq 2),31
(l + cos θ)2 ) fl(1 + cos θ1)2 + f2(1 + cos θ2)2 (2) f 1 + f2 ) 1 we calculated the surface coverage by FDCS, assuming that a complete monolayer of FDCS gives θ1 ) 106° 32 and a pure silanol surface gives θ2 ) 0°. Thus, the highest surface coverage obtained in this study is ∼90%, which demonstrates the effectiveness of this modification technique. Covalent attachment of FDCS to the silicon surface in scCO2 under the same reaction conditions in the absence of the polymer film (a clean silicon wafer) gives an XPS fluorine content of 22.26% (15° takeoff angle), water contact angles of 101°/90°, and an ellipsometry thickness of ∼8 Å. It is apparent that a complete monolayer is difficult (31) Israelachvili, J. N.; Gee, M. L. Langmuir 1989, 5, 288. (32) Hozumi, A.; Ushiyama, K.; Sugimura, H.; Takai, O. Langmuir 1999, 15, 7600.
to obtain at the buried interface by this method because of confinement and solubility effects. Still fairly high surface coverage by FDCS was achieved. The XPS atomic concentration results are compared with those obtained at 1200 psi and 35 °C in Figure 4. It is clear that higher pressure and temperature result in higher fluorine content. The difference may be an effect of kinetics or a solvent effect. The polymer chains are more mobile at higher pressure and temperature, permitting faster diffusion and also exposing more silanol groups for reaction. Reactions at SiO2/PMMA Interfaces. It is known that CO2 acts as a Lewis acid when interacting with PMMA and, thus, is an effective plasticizer for PMMA.2 On the other hand, the ester groups of PMMA interact with the surface silanols through strong hydrogen bonding, and it is estimated that one to two silanols per square nanometer interact with the absorbed polymer.33 The carbonyl group can form hydrogen bonds with either one or two silanol groups depending on the distribution of the silanols as well as upon the concentration of carbonyl groups near the surface (Figure 5). Although hydrogen bonding increases the nucleophilicity of silanol groups, it also limits chain mobility at the interface. The data (Table 3) indicate that the interplay of these two factors results in a lower extent of silanization at the interface compared with that for SiO2/polystyrene. The pressure used in these experi(33) Berquier, J. M.; Arribart, H. Langmuir 1998, 14, 3716.
Buried Interface Modification Using scCO2
ments is high enough to lower the glass transition of PMMA to room temperature; therefore, similar results as those obtained with polystyrene would be expected if there were no other contributing factors. We surmise that strong hydrogen bonding between PMMA and surface silanols inhibits the reaction and the buried SiO2/PMMA interface can be chemically modified to only a limited extent. Summary We have studied a model system for the modification of buried interfaces using supercritical CO2. We demonstrate the effectiveness of this method for the SiO2/ polystyrene interface and point out that supercritical conditions are necessary. The modification at the Si/ PMMA interface is less effective because of the strong hydrogen-bonding interaction between Si and PMMA. The
Langmuir, Vol. 18, No. 3, 2002 687
selectivity of this technique (higher degree of modification at the buried interface where the interaction between the substrate and the polymer coating is weaker) makes it potentially desirable because only weak interfaces need to be rehabilitated. Detailed work on in situ modification of the buried interface between glass fiber and polymer matrixes is currently in progress in our group. By choosing appropriate silane reagents (with one reactive site able to bind to the glass surface and the other to the polymer matrix), we expect improvement in mechanical properties of composites. Acknowledgment. We thank the Office of Naval Research and the NSF-sponsored Materials Research Science and Engineering Center for financial support. LA010768C
