pubs.acs.org/Langmuir © 2010 American Chemical Society
Rediscovering Silicones: Molecularly Smooth, Low Surface Energy, Unfilled, UV/Vis-Transparent, Extremely Cross-Linked, Thermally Stable, Hard, Elastic PDMS Peiwen Zheng and Thomas J. McCarthy* Polymer Science and Engineering Department, University of Massachusetts, Amherst, Massachusetts 01003, United States Received July 12, 2010. Revised Manuscript Received November 15, 2010 We demonstrate the preparation of extremely cross-linked poly(dimethylsiloxane) (PDMS)-based materials and report optical, mechanical, and surface properties. Transparent monolithic molded objects are prepared catalytically with no byproducts; parts per million levels of platinum (catalyst) remain in the articles. Essentially the same material was prepared in 1993 and described as a “hard transparent glass.” We confirm the thermal stability and chemical structure described in this report. We show that the catalytic reaction used, which was reported in 1999 always to exhibit a “violent exotherm”, can be controlled conveniently using a low (parts per million) catalyst concentration. The combination of low surface energy, transparency, hardness, elasticity, and thermal stability makes this an unusual and interesting material. That it can be prepared from commercially available low-viscosity monomers adds to its interest. We comment that the class of materials known as siloxanes or silicones and PDMS in particular is not currently generally well understood (or taught) and review aspects of the structure, properties, and cross-linking chemistry of PDMS.
Applications of silicone-based materials span many aspects of daily life in 2010 and many technologies as well.1-5 That these materials can be prepared, in principle, from sand and carbon dioxide (inexhaustible on earth) suggests that they are sustainable and that they will continue to be an important class of materials in the future. Poly(dimethylsiloxane) (PDMS), which is by far the most important silicone, is a material with very special and unusual properties. These properties and their molecular basis were understood in the 1940s but are not generally or widely appreciated today. The abbreviation PDMS is rather loosely used in the literature to describe commercial elastomers that are used in microfluidics devices, soft lithography, adhesion tests, and many other applications. PDMS, however, is a liquid with very interesting properties, and commercial elastomers are cross-linked materials that are compounded with about 50 wt % inorganic filler (silica) and other additives. Cross-linked poly(dimethylsiloxane)s without fillers are useful as coatings but have very poor bulk mechanical properties that are similar to those of tofu. The commercial success of silicone elastomers depends upon compounding with silica. We make several points as introductory comments concerning linear silicones, their properties, and their cross-linking chemistry that will seem didactic to some but are obviously generally warranted today. Although these points were wrestled with, debated, understood, and put to good use 60 years ago,3,4 silicones and silicone *Corresponding author. E-mail:
[email protected]. (1) Reference 2 is to the original report on PDMS synthesis. References 3 and 4 were chosen to emphasize that 60 years ago there was both a deep understanding of silicone chemistry and an appreciation of the wide range of applications of these materials. Reference 5 is a more modern, concise description of the broad field of silicones. (2) Rochow, E. G.; Gilliam, W. F. J. Am. Chem. Soc. 1941, 63, 798. (3) Rochow, E. G. An Introduction to the Chemistry of Silicones; Wiley: New York, 1946. Rochow, E. G. An Introduction to the Chemistry of Silicones, 2nd ed.; Wiley: New York, 1953. (4) McGregor, R. R. Silicones and Their Uses; McGraw-Hill: New York, 1954. (5) Clarson, S. J.; Semlyen, J. A. Siloxane Polymers; Prentice Hall: Englewood Cliffs, NJ, 1993.
Langmuir 2010, 26(24), 18585–18590
chemistry are not currently part of any common curriculum in chemistry or materials science education. The reasons for this are multiple and include the issues that (1) the chemistry and materials science of silicones were developed during World War II and were considered U.S. national security secrets,6-8 (2) the research leading to this significant body of knowledge was carried out primarily by two competitive companies, General Electric7 and Corning Glass (Dow Corning was formed in 1943),8 (3) academic institutions in the U.S. (in particular, chemistry and materials science departments) did not embrace polymers as an important academic subject, and (4) silicones and silicone chemistry do not fit “neatly” into organic, inorganic, or polymer chemistry curricula and are barely mentioned in textbooks on these subjects. Rochow states, in the preface to reference 5, “Methyl silicone was so different in composition, in structure and in physical and chemical properties that it was outside the ordinary day-to-day thinking of chemists and engineers fifty years ago.” This also happens to be true today concerning most chemists, engineers, and materials scientists. The chemical structure of PDMS is shown in Figure 1 using the repeat unit structure D (suggested by Rochow and used by silicone chemists). We note in this Figure that the Si-O and Si-C bond lengths (1.63 and 1.90 A˚) are significantly longer than the C-C (1.53 A˚) bonds and that the Si-O-Si bond angle (143°) is much greater than the C-C-C bond angles (109°). These differences separate methyl groups on individual silicon atoms and pairs of methyl groups from one another significantly more than they are separated in the carbon-based backbone polymer, polyisobutene (PIB). The electronegativity difference between silicon and oxygen also sets PDMS apart from carbon-based polymers. Pauling values for electronegativity (O, 3.5 and Si, 1.8) (6) References 7 and 8 offer historical accounts of the development of silicones at General Electric and Corning/Mellon Institute/Dow Corning and provide leading patent references to the development of silicones. (7) Liebhavsky, H. A. Silicones under the Monogram; Wiley: New York, 1978. (8) Warrick, E. L Forty Years of Firsts; McGraw-Hill: New York, 1990.
Published on Web 11/29/2010
DOI: 10.1021/la104065e
18585
Letter
Figure 1. (a) Repeat unit structure of PDMS indicating Rochow’s definition of D and the structure of the principal thermal decomposition product D3. (b) Structures of PDMS and polyisobutene (PIB) drawn to represent the C-C (1.53 A˚), Si-O (1.63 A˚), and Si-C (1.90 A˚) bond lengths and the Si-O-Si (143°) and C-C-C bond angles (109°) accurately. (c) Depictions of the motions of the dimethylsilyl group and oxygen atom (Si-O-Si bond angle) due to the 51% ionic backbone (calculated from Pauling electronegativity values).
suggest that the Si-O bond is 51% ionic.9,10 The less directional nature (relative to covalent C-C bonds) of the partially ionic bonds and the longer bond lengths permit rotational and vibrational degrees of freedom that are not possible in carbon-based polymers (Figure 1). Rochow’s conclusion11 that “it must be that the silicon atom and its associated pair of methyl groups swings as a unit, as though the silicon-oxygen bond were a ball and socket joint” and Mark’s statement12 regarding the 143° Si-O-Si bond angle, “in addition, this bond angle is so flexible that it can readily pass through the 180° state” suggest polymer chain motions that are counterintuitive to most today. Despite this flexibility, the Si-O backbone is thermodynamically stronger than the C-C backbone, giving PDMS better thermal stability than hydrocarbon polymers. The formation of the cyclic trimer of dimethylsiloxane (conserving Si-O bonds) limits its thermal stability (Figure 1). The cross linking of PDMS to form elastomers was also discovered in the 1940s13-15 and first carried out by bimolecular radical coupling induced by stoichiometric reactions with benzoyl peroxide (Figure 2). We note that because PDMS-derived radicals do not contain β-hydrogens or β-C-C bonds, PDMS is particularly well suited for radical cross linking. (Fragmentation and disproportionation reactions do not compete with coupling as they do in hydrocarbon polymers.) This method was improved by the replacement of some methyl groups with vinyl groups,15 (9) Rochow discusses the 51% ionic character of Si-O bonds on page 114 of reference 3 (2nd edition). (10) Pauling addresses criticisms of the 51% ionic character of Si-O bonds in Pauling, L. Am. Mineral. 1980, 65, 321. (11) Reference 3 (2nd edition), p 115. (12) Mark, J .E. Acc. Chem. Res. 2004, 37, 946. (13) Cross linking of silicones is reviewed by Thomas in Chapter 12 of reference 5. (14) Wright, J. G. E.; Oliver, C. S. U.S. Patent 2,448,565, September 7, 1948. (15) Marsden, J. U.S. Patent 2,445,794, July 27, 1948.
18586 DOI: 10.1021/la104065e
Zheng and McCarthy
Figure 2. Bond-forming events that occur in stoichiometric benzoyl peroxide (BPO) induced the cross-linking of PDMS and a platinum-catalyzed cross-linking (hydrosilylation) of vinyl groups and hydridosilyl groups. Note that the structures of the two cross links (ethylene bridges) are identical.
which contributes to a radical chain reaction and reduces the amount of benzoyl peroxide required, and further improved by the discovery of platinum-catalyzed hydrosilylation16 (Figure 2). We note that the structures (ethylene bridges) formed by the radical coupling of two methyl groups are identical to those formed by the hydrosilylation of a vinyl group with a hydridosilane. We also note that catalyst development (to the level of individually active platinum atoms) made platinum-catalyzed hydrosilylation-cured silicones economically feasible.17 A final point in these introductory comments concerns the low cohesive energy of PDMS. This is discussed in a number of 1940s publications.18-23 The low strength of silicone rubber (tofulike) was initially thought18 to be due to low molecular weight, but molecular weight measurements discounted this premise when values greater than 1 106 g/mol were measured. The low strength as well as the low boiling points of volatile siloxanes (lower than those of alkane analogs) and viscosity properties of silicone oils could be explained entirely by the low intermolecular attractive forces that are due to the free motion of dimethylsilyl groups in the “ionic” chain. These low intermolecular forces impart very low surface energies (γ ≈ 20 dyn/cm) to silicones.24 We describe here a product of the platinum-catalyzed hydrosilylation reaction between 1,3,5,7-tetramethylcyclotetrasiloxane (D4H) and 1,3,5,7-tetravinyl-1,3,5,7-tetramethylcyclotetrasiloxane (D4V) (Figure 3).25 Reactions between these two monomers (16) Speier, J. L.; Webster, J. A.; Barnes, G. H. J. Am. Chem. Soc. 1957, 79, 974. (17) Karstedt, B. D. U.S. Patent 3,775,452, Nov. 27, 1973. (18) Reference 3 (2nd edition), pp 119-120. (19) Wilcock, D. F. J. Am. Chem. Soc. 1946, 68, 691. (20) Sauer, R. O.; Mead, D. J. J. Am. Chem. Soc. 1946, 68, 1794. (21) Scott, D. W. J. Am. Chem. Soc. 1946, 68, 1877. (22) Hunter, M. J.; Warrick, E. L.; Hyde, J. F.; Currie, C. C. J. Am. Chem. Soc. 1946, 68, 2284. (23) Roth, W. L. J. Am. Chem. Soc. 1947, 69, 474. (24) Wu, S. In Polymer Handbook, 4th Ed.; Brandrup, J., Immergut, E. H., Grulke, E. A., Eds.; John Wiley & Sons: New York, 1999; pp VI-524. (25) All samples described in this report were prepared using a 2:1 ratio of D4H/D4V. D4V (1.70 g, 4.93 mmol) and D4H (2.40 g, 9.98 mmol) were mixed well, followed by the addition of 100 μL of 5.0 10-4 g/mL (2.56 10-4 mmol Pt, 12 ppm Pt based on the silicone product mass) Karstedt’s catalyst. The solution was cured at room temperature for 24 h and then cured at 150 °C for 24 h. D4H and D4V were purchased from Gelest Inc. and used as received. Karstedts’ catalyst was purchased from Gelest Inc. and diluted in toluene to 5 10-4 g/mL (based on Pt).
Langmuir 2010, 26(24), 18585–18590
Zheng and McCarthy
Letter
Figure 3. Structures of D4H and D4V that react by platinum-catalyzed hydrosilylation to form D4H/D4V silicone. The sample in the photograph is ∼1 mm thick and has a diameter of 8.5 cm.
and their products have been described in the literature by several groups;26-29 however, the multiple qualities, properties, and potentials of both the reaction and the product are not appreciable from these reports. Michalczyk and co-workers first described26 this material in 1993 as a “hard transparent glass” and studied the thermal properties of materials prepared from 1:1 and 1:2 molar ratios of D4V and D4H. They found that these materials are significantly more thermally stable than PDMS and attributed this to the absence of a pathway to form the cyclic trimer, which is the major thermal decomposition product of PDMS (Figure 1). Stein et al.27 showed that networks prepared from 1:2 and 2:1 molar ratios of D4V and D4H are amorphous by X-ray diffraction and that single platinum atoms from Karstedt’s precatalyst17 are (26) Michalczyk, M. J.; Farneth, W. E.; Vega, A. J. Chem. Mater. 1993, 5, 1687. (27) Stein, J.; Lewis, L. N.; Gao, Y.; Scott, R. A. J. Am. Chem. Soc. 1999, 121, 3693. (28) Redondo, S. U. A.; Radovanovic, E.; Torriani, I. L.; Yoshida, I. V. P. Polymer 2001, 42, 1319. (29) Schiavon, M. A.; Radovanovic, E.; Yoshida, I. V. P Powder Technol. 2002, 123, 232.
Langmuir 2010, 26(24), 18585–18590
the active catalytic species. We point directly to a warning in the Experimental Section of this 1999 paper27 concerning the reaction of a 1:2 molar ratio of D4V and D4H to emphasize the thermodynamics of this reaction: “Warning: A violent exotherm was always noted upon addition of platinum, and the contents burst into flames in one case.” This reaction is indeed strongly exothermic, more so than almost any other polymerization, but the exotherm and the reaction rate can be controlled easily by adjusting the platinum concentration and reaction temperature. To prepare the samples that we report here, we used a very low catalyst concentration25 (12 ppm Pt) and a two-stage cure at different temperatures. For certain applications (e.g., reaction injection molding), an exotherm and rapid reaction may be desirable, and these could be tuned with the catalyst concentration. We emphasize that this reaction involves two low-viscosity liquids that are catalytically converted, using parts per million platinum atoms, to a hard transparent glass with no side products. Figure 3 shows a representative structure of the product of the reaction for a 2:1 ratio of D4H and D4V. The structure is shown through a ∼1-mm-thick, 8.5-cm-diameter sample of cured D4H/D4V DOI: 10.1021/la104065e
18587
Letter
Zheng and McCarthy
Figure 5. Load-displacement data that were used to obtain mech-
anical properties data by (a) nanoindentation using an ∼150-nmradius diamond tip (indentation/displacement was ∼400 nm) and with a contact adhesion analysis instrument (b) using a 5-mmdiameter glass hemispherical probe (displacement was ∼30 μm). Note that the displacement is plotted left to right in panel a and right to left in panel b by convention.
Figure 4. (a) Photograph of ∼1-cm-thick D4H/D4V silicone shown to emphasize the transparency of this material. (b) UV-vis spectrum of an ∼1-mm-thick sample of D4H/D4V silicone. (c) TGA data for a sample of D4H/D4V silicone heated at 10 °C/min to 900 °C. The slight (0.24%) mass increase is presumably due to the oxidation of Si-H bonds.
silicone and indicates that there are residual Si-H bonds (on onethird of the Si atoms) in this material. These hydridosilane groups are reactive to both hydrolysis and oxidation, and these reactions will increase the cross-link density (form additional Si-O-Si bonds) over time upon exposure to air. The 2:1 sample was chosen for detailed study after characterizing samples prepared from 1:1, 1.5:1, 2:1, and 2.5:1 ratios of D4H/D4V using solid-state 13 C NMR30 and nanoindentation.31 The product prepared using the 2:1 ratio showed an absence of vinyl groups by NMR and a (30) Solid-state 13C and 29Si NMR spectra were recorded on a Bruker DSX300 spectrometer. Signals for the vinyl carbons that are apparent at ∼δ137 in samples prepared using a 1:1 rato of D4H/D4Vare absent in samples prepared with a 2:1 ratio. Ratios of intensities from 29Si signals at ∼δ-21 and ∼δ-34 were used to assess the conversion. See Supporting Information and ref 26. (31) Nanoindentation experiments were performed using a Hysitron TI 900 triboindenter with a Berkovitch-type diamond tip (radius of curvature of ∼150 nm). The load is applied to the surface while monitoring the penetration depth of the tip. From the loading and unloading curves, nanomechanical properties can be obtained. The same diamond tip was used to obtain AFM images of the surface at the specified indent positions after indentation.
18588 DOI: 10.1021/la104065e
higher modulus than for the other samples by nanoindentation. Solid-state 29Si NMR30 of the 2:1 ratio product indicates that 64% of the silicon atoms are attached to ethylene bridges (cross links in the network structure). This value is close to the value (66.7%) predicted by the stoichiometry. Solid-state 13C and 29 Si NMR spectra of cross-linked samples prepared with both 1:1 and 2:1 ratios of D4H/D4V are provided as Supporting Information (Figures S1 and S2). Figure 4a shows a photograph of a thicker (∼1 cm) D4H/D4V sample that is clearly very transparent and a transmission UV-vis spectrum (Figure 4b) of a 1-mm-thick sample of D4H/D4V. This sample exhibits >90% transmittance above λ = 355 nm and 50% transmittance at 275 nm. There is no organic chromophore in the material; the absorbance in the near-UV can be attributed to the platinum catalyst. Figure 4c shows thermal gravimetric analysis data for a sample of D4H/D4V (2:1 molar ratio) heated in N2 from room temperature to 900 °C at 10 °C/min. This confirms the thermal stability of this material as reported by Michalczyk.26 The sample began to lose mass at ∼400 °C and lost ∼1% of its mass by the time 500 °C was reached. Differential scanning calorimetry was performed from -150° to 150° (both heating and cooling), and no presence of a Tg was indicated. The mechanical properties of this material were quantitatively assessed using nanoindentation,31 contact adhesion analysis,32 (32) A custom-built contact adhesion analysis instrument was used. A hemispherical glass probe (5 mm radius) was brought into contact with the sample surface at a fixed rate. The force, displacement, and contact area were recorded, and the stiffness and modulus could be calculated accordingly.
Langmuir 2010, 26(24), 18585–18590
Zheng and McCarthy
and standard tensile strain measurements.33 Figure 5a shows load-displacement data that were obtained by increasing the load on a diamond tip from 0 to 500 μN over ∼7 s, holding the force at 500 μN for ∼4 s and then releasing the force over 4 s. When the force reached 500 μN, the displacement was 405 nm. After an additional 4 s at constant force (500 μN), the displacement increased (only) to 412 nm; this qualitatatively indicates that the material is “hard.” The reduced modulus and hardness of the silicone network were calculated from the data to be 1.69 ( 0.04 and 0.34 ( 0.02 GPa, respectively, by fitting the unloading curve to a power law. The unloading curve in Figure 5a shows that the displacement returns to its original value, indicating elasticity. Atomic force microscopy revealed that the diamond tip does not blemish the sample; no indication that the sample had been indented was visible. Using a value of 0.5 for Poisson’s ratio for an elastomer, a Young’s modulus of 1.29 ( 0.4 GPa can be calculated for D4H/D4V silicone from the measured reduced modulus.34 Contact adhesion analysis,32 which is a more macroscopic measurement than nanoindentation, indicated a Young’s modulus of 1.46 ( 0.08 GPa (Figure 5b). ASTM tensile tests33 on “dogbone”-shaped samples yielded a Young’s modulus of 1.59 ( 0.16 (std dev) GPa. To put these numbers into perspective, silicone elastomers exhibit Young’s moduli in the range of 0.1-5 MPa35 and bisphenol A-polycarbonate (polycarbonate is a material that is ∼130 °C below its glass-transition temperature at room temperature) exhibits a Young’s modulus of 2.3 GPa.36 The D4H/D4V silicone that was previously described26 as a hard transparent glass can also be described as an extremely crosslinked poly(dimethylsiloxane). We emphasize that these properties are achieved by chemical cross linking with C-C bonds, not by the incorporation of an inorganic component, and that D4H/ D4V silicone can be regarded as unfilled, which is unusual for a monolithic silicone material. Compression testing37 was also carried out, and samples that were compressed to 20% compression (80% of their initial thickness value) recovered to their original thickness dimensions. (33) Tensile measurements were conducted using ASTM D638 type V samples with a thickness of 0.73 mm. These were punched from a partially cured D4H/D4V sheet, and the samples were then heated to 150 °C to complete the curing. Sample dimensions were measured using calipers. An Instron 5800 R fitted with a 1 kN load cell was used and controlled by using the Merlin software package. Four samples were tested at a constant cross-head speed of 1 mm/min at room temperature. A preload was applied to eliminate compressive forces on specimens and improve the consistency of the measurements. An average Young’s modulus of 1.59 ( 0.16 (standard deviation) GPa was obtained from linear slopes of the stress-strain curves. This result is consistent with nanoindentation and contact adhesion data. Samples broke at a strain of 0.15-0.32%, and no yield behavior was observed. (34) Bhushan, B. Springer Handbook of Nanotechnology; Springer: New York, 2004; Vol. 1, p 677. (35) Possart, W. Adhesion: Current Research and Applications; Wiley-VCH: Berlin, 2005, p 36. (36) McCrum, N. G.; Buckley, C. P.; Bucknall, C. B. Principles of Polymer Engineering; Oxford University Press: New York, 1977; p 373. (37) Compression tests involved three cylindrical specimens with height-todiameter ratios of 1:1. These were compressed using a 50 kN load cell at a 1%/min strain rate using the same Instron that was used for tensile tests. After release from up to 20% compression, all three samples recovered to their original dimensions; this was confirmed by caliper measurements. Minor crack propagation and barreling were observed during the compression; therefore, a modulus was not calculated. This test must be regarded as qualitative, but bulk elasticity was certainly demonstrated. (38) A low-cross-link-density silicone was prepared using vinyl-terminated PDMS (Mw ≈ 28 000, Gelest), trimethylsilyl-terminated poly(hydridomethylsiloxane) (Mw ≈ 1400-1800, Gelest), and ∼2 ppm Karstedt’s catalyst. The ratio of vinyl/hydridosilane groups was 1:5. (39) Contact angle measurements were made with a Rame-Hart telescopic goniometer and a Gilmont syringe with a 24-gauge flat-tipped needle. Dynamic advancing (θA) and receding angles (θR) were recorded while Milli-Q water was added to and withdrawn from the drop, respectively. That the values (θA/θR) of these two silicone samples are within 1° is likely coincidental. We have recently reported41 extensive contact angle data on supported poly(dimethylsiloxane) samples with negligible hysteresis.
Langmuir 2010, 26(24), 18585–18590
Letter
Figure 6. AFM height images and section analyses of the free (air) surface of an as-prepared sample of D4H/D4V silicone (upper data) and a delaminated D4H/D4V silicone sample that was cured in contact with the free surface of an as-prepared sample.
Surface analysis of this D4H/D4V silicone was performed by using water contact angle analysis and atomic force microscopy. To compare contact angle data for the D4H/D4V silicone with that of conventional cross-linked PDMS, a sample was prepared with a much lower cross-link density.38 Water contact angles39 for the two materials are indistinguishable, with the D4H/D4V silicone exhibiting θA/θR = 109°/92° and the lightly cross-linked silicone showing θA/θR = 109°/93°. The extremely cross-linked D4H/D4V silicone thus exhibits the low surface energy that is a characteristic of methylsilicones. Atomic force microscopy40 reveals that the free surfaces of D4H/D4V silicone are molecularly smooth, showing rms roughness values of ∼0.1 nm. Figure 6 shows two AFM height images and section analyses. One image is of the free (air) surface of an as-prepared25 sample, and the other is of a delaminated D4H/D4V silicone sample that was cured (under the same conditions25) in contact with the free surface of an asprepared sample. This suggests that it may be possible to mold and replicate very small features using D4H/D4V silicone. In summary, we report optical, mechanical, and surface properties of a material that can be regarded as an extremely cross-linked poly(dimethylsiloxane). We confirm the thermal stability and chemical structure of the material that was reported in 1993 as a hard transparent glass and show that the violent exotherm warned of in a 1999 paper can be controlled conveniently by using low catalyst concentrations. The ease of preparation of D4H/D4V from commercially available monomers combined with its properties of low surface energy, transparency, hardness, elasticity, and thermal stability (40) Atomic force images were obtained using a Digital Instruments Dimension3000 AFM in tapping mode. (41) Krumpfer, J. W.; McCarthy, T. J. Faraday Discuss. 2010, 146, 103.
DOI: 10.1021/la104065e
18589
Letter
makes this an unusual material that has multiple uses that we are currently exploring. Acknowledgment. We thank Yufeng Zhang for preliminary experiments with D4H/D4V silicones and Prof. Al Crosby’s group for help with mechanical measurements. We thank the Centers for Hierarchical Manufacturing (CMMI-0531171) and the Materials Research Science and Engineering Center
18590 DOI: 10.1021/la104065e
Zheng and McCarthy
(DMR-0213695) at the University of Massachusetts for support and 3M and Henkel and Shocking Technologies for unrestricted funding. Supporting Information Available: 13C NMR (75 MHz) and 29Si NMR (59 MHz) spectra of silicones. DSC data for and IR spectrum of D4H/D4V. This material is available free of charge via the Internet at http://pubs.acs.org.
Langmuir 2010, 26(24), 18585–18590
