Langmuir 2006, 22, 1609-1618
1609
Physically Self-Assembled Monolayers (PSAMs) of Lecithin Lipids at Hydrophilic Silicon Oxide Interfaces Tze-Lee Phang and Elias I. Franses* School of Chemical Engineering, Purdue UniVersity, 480 Stadium Mall DriVe, West Lafayette, Indiana 47907-2100 ReceiVed August 15, 2005. In Final Form: December 14, 2005 A new method of making physically self-assembled monolayers (PSAMs) on hydrophilic solid surfaces is presented. This method uses a mixture of a nonpolar solvent, such as hexane, and a strong polar solvent, such as ethanol, to dissolve the lipids. The deposition of two lecithin lipids, dipalmitoylphosphatidylcholine (DPPC) and dilauroylphosphatidylcholine (DLPC), has been studied. These lipids physically self-assemble, or adsorb, onto hydrophilic silicon oxide/silicon surfaces when such surfaces are in contact with the lipid solution. The adsorbed layers were probed with ex-situ attenuated total reflection infrared (ATR-IR) spectroscopy, ellipsometry, contact angle measurements, and atomic force microscopy (AFM). The thicknesses of the adsorbed monolayers are about 2.8 ( 0.2 nm for DPPC and 2.0 ( 0.2 nm for DLPC, as determined by ellipsometry and AFM. Smooth, uniform monolayers of controlled surface density are formed. The surface density of adsorbed layers is comparable to those of close-packed lipid monolayers, as calculated from the ellipsometry and ATR-IR results. Producing controlled-thickness monolayers has applications in boundary lubrication, biomaterials, sensor technologies, and electronics. The method can be used for depositing many biological surfactants or lipids without the need to modify these surfactants chemically to form chemical bonds with the surfaces, as required by the usual chemical SAMs. Moreover, the new method has several advantages compared to the Langmuir-Blodgett (LB) method.
1. Introduction Monomolecular (monolayer) films are important in boundary lubrication,1,2 biomaterials,3 corrosion inhibition,4 novel sensor technology and bioassaying,5 and electronics.6,7 Despite many advances in monolayer preparation, a robust general method is needed for preparing controlled-surface-density monolayers of lipids or other surfactants on hydrophilic surfaces, such as that of silicon oxide or mica. The well-established Langmuir-Blodgett (LB) technique may not work robustly with a transfer ratio of 1.0 because of problems associated with nonuniform or nonconstant dynamic contact angles, film instability and crystal growth, and finding “windows” of suitable surface pressures and feasible immersion/withdrawing rates.8 Often, one needs to make use of bivalent ions (e.g., Ca2+) or other physicochemical modifications and conduct empirical trials to find such windows.8 Moreover, the LB method works for flat surfaces but may fail for patterned substrates. Methods for making chemically self-assembled monolayers (SAMs) require that a chemically reactive group is present, and they clearly do not work for many biological surfactants that lack such groups, including phospholipids. The production of many popular SAMs of alkylchlorosilanes on hydrophilic surfaces is quite sensitive to reaction conditions such as temperature, concentration, and moisture content.4 If a lipid is insoluble in aqueous solutions and only dispersible in the form of liposomes (multilamellar) or vesicles (unilamellar), then the method of * To whom correspondence should be addressed. Tel: (765)494-4078. Fax: (765)494-0805. (1) Ren, S.; Yang, S.; Zhao, Y.; Zhou, J.; Xu, T.; Liu, W. Tribol. Lett. 2002, 13, 233. (2) Carpick, R. W.; Salmeron, M. Chem. ReV. 1997, 97, 1163. (3) Malmsten, M. Interface Sci. 1997, 5, 159. (4) Ulman, A. Chem. ReV. 1996, 96, 1533. (5) Fan, F. Q.; Maldarelli, C.; Couzis, A. Langmuir 2003, 19, 3254. (6) Gleiche, M.; Chi, L. F.; Fuchs, H. Nature 2000, 403, 173. (7) Aizenberg, J.; Black, A. J.; Whitesides, G. M. Nature 1998, 394, 868. (8) Ulman, A. An Introduction to Ultrathin Organic Films: From LangmuirBlodgett to Self-Assembly; Academic Press: New York, 1991; pp 101-219.
physical adsorption may work poorly for forming monolayers and may lead to supported bilayers or other types of films.9-15 The latter method requires “controlled vesicle fusion”, which is difficult to accomplish or control and may require tailoring the solution conditions. Moreover, special procedures are needed for forming uniform rather than heterogeneous layers.9-15 With vesicular dispersions or with molecular or micellar solutions, the question is whether adsorption will be limited at all to monolayers, bilayers, or multilayers. To avoid the complications associated with developing an effective LB method, a SAM method, or a vesicle- or liposomebased method, we have developed and used single-phase molecular solutions in organic solvents from which molecular (rather than particulate) adsorption is expected to occur. Our goal is to develop methods that are simple, general, and require no chemical modification of the sorbate or the sorbent. In this article, we report a study of monolayer formation of two well-known zwitterionic lipids, dipalmitoylphosphatidylcholine (DPPC) and dilauroylphosphatidylcholine (DLPC). DPPC is a major component of biomembranes and lung surfactant.16-19 It has practically “zero” solubility in water (or ,10-6 wt %) and a chain-melting transition temperature of 41 °C, above which fluid multilamellar liposomes form.20,21 Below this temperature, (9) Reviakine, I.; Brisson, A. Langmuir 2000, 16, 1806. (10) Egawa, H.; Furusawa, K. Langmuir 1999, 15, 1660. (11) Jass, J.; Tja¨rnhage, T.; Puu, G. Biophys. J. 2000, 79, 3153. (12) Tamm, L. K.; Bo¨hm, C.; Yang, J.; Shao, Z. F.; Hwang, J.; Edidin, M.; Betzig, E. Thin Solid Films 1996, 285, 813. (13) Richter, R.; Mukhopadhyay, A.; Brisson, A. Biophys. J. 2003, 85, 3035. (14) Seantier, B.; Breffa, C.; Fe´lix, O.; Decher, G. Nano Lett. 2004, 4, 5. (15) Warr, G. G. Curr. Opin. Colloid Interface Sci. 2000, 5, 88. (16) Alberts, B.; Bray, D.; Lewis, J.; Raff, M.; Roberts, K.; Watson, J. D. Molecular Biology of the Cell; Garland Publishing: New York, 1994. (17) Tanford, C. The Hydrophobic Effect: Formation of Micelles and Biological Membranes; 2nd ed; Wiley: New York, 1980. (18) Notter, R. H.; Finkelstein, J. N. J. Appl. Physiol. 1984, 57, 1613. (19) Notter, R. H. Lung Surfactants; Marcel Dekker: New York, 2000. (20) Wen, X.; Franses, E. I. Langmuir 2001, 17, 3194. (21) Pinazo, A.; Wen, X.; Liao, Y.-C.; Prosser, A. J.; Franses, E. I. Langmuir 2002, 18, 8888.
10.1021/la0522202 CCC: $33.50 © 2006 American Chemical Society Published on Web 01/20/2006
1610 Langmuir, Vol. 22, No. 4, 2006
the liposomes become “frozen”. Unilamellar vesicles can be prepared by sonication above 41 °C. After cooling to 25 °C, these vesicles are generally stable for hours and days. DLPC is a lower-molecular-weight homologue (two C11 vs two C15 aliphatic chains), with a chain-melting temperature of around 5 °C and a finite solubility in water of 4 ppm (0.0004 wt %).21 We used such dispersions to form adsorbed layers but were unable to form quality monolayer or bilayer films. (Data are not shown here but will be summarized in ref 22.) Using a nonpolar alkane organic solvent, such as hexane, may appear to be a possible solution, but it leads to uncontrolled multilayer formation, as shown below (section 3.1). We have discovered that the method works well and also involves using, in addition to hexane, a “strong polar solvent” such as ethanol (or another polar molecule) to ensure the dissolution of the lipid and to control the lipid-solvent and lipid-sorbent interactions. The results have possible implications in making monolayer sensors of such lipids or other surfactants and in applications in normal-phase HPLC chromatography, with no chemical reaction with the substrate. It has been known in the HPLC literature that adding a strong polar solvent to a hydrocarbon may lead to socalled “localized solute adsorption”,23 which probably means controlled solute adsorption. The use of lipid solutions in certain hexane/ethanol solutions is shown below to be successful for making uniform lipid monolayers of DLPC and DPPC of controlled surface density. Further studies may be needed to establish how general the method is for other lipids and surfactants and other binary or multicomponent nonpolar-polar liquid solutions. The adsorbed DLPC or DPPC physically self-assembled monolayers (PSAMs) were produced with various procedures and were characterized quantitatively with ex-situ (at the air/ solid interface) attenuated total reflection infrared (ATR-IR) spectroscopy, ellipsometry, contact angle measurements, and atomic force microscopy (AFM). The DLPC or DPPC films have lower contact angles (they are apparently less hydrophobic) compared to those of chemical SAMs. We present some indirect evidence and argue that the hydrocarbon chains, rather than the polar headgroups, are exposed to air at the air/solid interface. A possible phenomenological surface thermodynamic explanation is given. For adsorption from hexane/ethanol solutions, the effects of ethanol concentration, lipid concentration, and adsorption time are reported. For these two lipids, we also report a limited study on which types of layers, with respect to thickness and uniformity, form with the standard use of the LB method. We then compare the LB films with the PSAMs, with respect to film uniformity and the robustness of the method. 2. Experimental Section 2.1. Materials. Synthetic L-R-dipalmitoylphosphatidylcholine (DPPC, 99+% pure) and L-R-dilauroylphosphatidylcholine (DLPC, 99% pure) were purchased from Sigma Chemical Co. (St. Louis, MO). Octadecyltrichlorosilane (OTS) (90+%) was purchased from Aldrich Chemical Co. (Milwaukee, WI). Carbon tetrachloride (99.9% pure) and n-hexane (99%) were purchased from Sigma-Aldrich Co. (St. Louis, MO). Ethyl alcohol (200 proof) was purchased from Aaper Alcohol and Chemical Co. (Shelbyville, KY). Hexane was dried over 3 Å molecular sieves prior to use to remove traces of water. The other materials were used as received. The lipid solutions were prepared on a weight basis. The pure water used for all samples (22) Phang, T.-L. Ph.D. Thesis, Purdue University, West Lafayette, IN, expected 2006. (23) Snyder, L. R.; Kirkland, J. J.; Glajch, J. L. Practical HPLC Method DeVelopment; 2nd ed., Wiley-Interscience: New York, 1997; pp 268-276.
Phang and Franses was first distilled and then passed through a Millipore four-stage cartridge system, resulting in a water resistivity of 18 MΩ cm at the exit port. A trapezoidal silicon (SiO2/Si) internal reflection element IRE with 45° angles at the ends (50 × 10 × 3 mm3, from Wilmad Glass Co., Buena, NJ) was cleaned with a detergent solution and rinsed with Millipore water. The IRE was immersed in ethanol and sonicated using a sonicator bath (Branson 3510 Ultrasonic cleaner, Branson Ultrasonics Corporation, Danbury, CT) for 10 min. The IRE was then rinsed with water and was finally treated in a plasma cleaner (PDC-3XG, Harrick Scientific Co., Ossining, NY) for 20 min. The IRE had an ultrathin film of native oxide. Similar procedures were used to clean the silicon wafers (with ∼500 Å of a thermal oxide layer, from Silicon Quest International, Santa Clara, CA). The silicon wafers were used as substrates for the AFM experiments. 2.2. Lipid Films Preparation. A lipid monolayer formed on a substrate (a SiO2/Si IRE or a silicon wafer) by a dip-coating technique24-26 or by the Langmuir-Blodgett (LB) method.8,27,28 Dipcoated films were prepared by immersing the substrate vertically into a DPPC or DLPC solution. The solvent was hexane or hexane containing ethanol at various small concentrations, 0.05 to 10 vol %. After a preset immersion time, the substrate was pulled vertically upward at a constant speed of 17 mm/min. Films containing some entrained solution were dried in air for 1 h. The substrate speed and drying time can be quite flexible and robust. To ensure that lipid is deposited on the substrate via adsorption, and not via substantial liquid entrainment and drying, low lipid concentrations were used, from 10 to 100 ppm. The monolayer films in contact with water are quite stable, indicating no desorption.22 LB films were prepared with a KSV 5000 computer-controlled Langmuir trough (purchased from KSV Instruments, Finland), which has a platinum Wilhelmy plate connected to an electrobalance. One hundred microliters of a 1 mg/mL DPPC or DLPC solution in a 9/1 volume ratio of hexane/ethanol was spread on water at ca. 22 °C. After 15 min (to allow the hexane/ethanol to evaporate), the monolayer was compressed at a constant rate of 5 mm/min to the deposition pressure. After surface pressure stabilization (it typically took 30-40 min), the monolayer was deposited onto the substrate at a constant surface pressure at a preset pulling speed. The transfer ratio was found to vary with the substrate pulling speed. Z-type LB films were formed by immersing the substrate under the surface before spreading the monolayer and pulling it upward. Attempts were made (unsuccessfully) to produce X-type LB monolayer films by dipping (downward) the substrate through the monolayer after the monolayer formed and was compressed. To ensure that no further deposition occurred upon withdrawal of the substrate out of the water, the remaining monolayer was first removed from the water surface by aspiration, and the substrate was withdrawn at a rate of 50 mm/min. All films were then dried in air for 1 h. 2.3. Contact Angle Measurements. Advancing contact angles for water, θa, were measured with a Rame´-Hart goniometer (Rame´Hart, Inc., Mountain Lakes, NJ). A small (3-6 µL), mostly spherical droplet was formed at the end of a microsyringe needle, which was slowly lowered until it touched the substrate surface. The reported θa value was the maximum angle measured when the volume of the droplet was increased with no substantial change in the solid/liquid interface area. At least two independent measurements were taken for each surface or each location at the same surface. For “benchmarking”, θa for a self-assembled monolayer of OTS on a silicon wafer was obtained. It was found that θa is 109 ( 2°, which is in good agreement with a literature value of 111 ( 1°.1,29 2.4. Ellipsometry. A Rudolph Research (now Rudolph Technologies, Flanders, NJ) Auto ELII automatic null ellipsometer was (24) Lee, Y.-L.; Chen, C.-Y. Appl. Surf. Sci. 2003, 207, 51. (25) Lee, Y.-L.; Fang, T.-H.; Yang, Y.-M.; Maa, J.-R. Int. Commun. Heat Mass 1998, 25, 1095. (26) Allara, D. L.; Nuzzo, R. G. Langmuir 1985, 1, 45. (27) Blodgett, K. B.; Langmuir, I. Phys. ReV. 1937, 51, 964. (28) KSV 5000 Instruction Manual; KSV Instruments Ltd: Helsinki, Finland, 2000. (29) Tillman, N.; Ulman, A.; Penner, T. L. Langmuir 1989, 5, 101.
PSAMs of Lecithin Lipids at SiO2 Interfaces
Langmuir, Vol. 22, No. 4, 2006 1611
used for measuring ellipsometry angles, ∆ and Ψ,30,31 of adsorbed DPPC or DLPC layers at the air/solid interface. Measurements were taken at a wavelength (λ) of 633 nm with an incident angle (φo) of 60 or 70° measured from the surface normal. The silicon substrate was placed on the standard sample stage. Measurements of ellipsometric angles, ∆o and Ψo, for 5 to 10 spots on the cleaned (no adsorbate) substrate were obtained and then averaged. Similarly, for the adsorbed lipids on the substrate, several measurements of ∆ and Ψ were obtained and then averaged. A typical precision in measuring ellipsometric parameters, δ∆ ≡ ∆ - ∆o and δΨ ≡ Ψ - Ψo, is (0.05 to 0.5°, depending on the angle and the wavelength. For this system, the ellipsometric parameters generally increased as the surface density (mg/m2) or the thickness of the adsorbed surface layer increased.32,33 A three-layer model consisting of air/silicon oxide film/silicon, along with standard ellipsometry theory,30,31 and the measured values of ∆o and Ψo at two angles were used to determine simultaneously the refractive index of silicon (n3 - ik3) and the refractive index and thickness (n2 and d2) of the native silicon oxide layer. Four equations were solved for four unknowns: n3, k3, n2, and d2. An optimization routine for total error minimization described in refs 32 and 33 was used to determine the values of these unknowns. These values, together with the measurements of δ∆ and δΨ (or ∆ and Ψ) and a four-layer model consisting of air/lipid film/silicon oxide film/ silicon, were used to determine the lipid film properties. The ellipsometry ∆-Ψ “trajectories” of films were calculated at various film thicknesses (df) for a series of film refractive indices (nf). The measured ∆ and Ψ values were found to fall into the range of the calculated trajectories with nf ) 1.3 to 1.6. By fixing the film refractive index nf to 1.47 for DPPC and 1.46 for DLPC, for a close-packed monolayer34,35 the apparent value of df was determined. If a smaller value of nf were used, as for more dilute monolayers or for nonuniform multilayers, then the calculated value of df would be higher and vice versa. The estimated accuracy was (0.2 nm for df because of the uncertainty of nf. The surface density Γ was calculated from the following expression, which is well established for isotropic films36 Γ ) df
( )( ) 2 M nf - 1 Af n 2 + 2 f
(1)
where M is the molecular weight (g/mol) and Af is the molar refractivity (cm3/mol), which is calculated by using the molar refractivities of the atoms or atom groups according to Eisenlohr’s system.37 The calculated M/Af value is 3.505 for DPPC and 3.605 for DLPC. The value of Γ was determined with a lower relative uncertainty (more “robustly”) than those of nf or df because for a given ∆ and Ψ as the calculated value of nf increases, df decreases.32,33 2.5. Fourier Transform Infrared Spectroscopy. All infrared spectra were obtained by a Nicolet Prote´ge´ 460 Fourier transform infrared spectrometer equipped with a liquid-nitrogen-cooled mercury-cadmium-telluride (MCT) detector. To determine the molar absorptivities of the lipids, transmission IR (TIR) spectra of each lipid in CCl4 solutions (1 to 2 wt %) were obtained using a Specac Omni-Cell (Specac Inc., Woodstock, GA) with a 12 µm Mylar spacer. The path length of the sample was calculated using a fringe counting technique.38 Attenuated total reflection (ATR) spectroscopy was used to probe the surface densities of the adsorbed (30) Azzam, R. M. A.; Bashara, N. M. Ellipsometry and Polarized Light; North-Holland: Amsterdam, 1979. (31) Tompkins, H. G. A User’s Guide to Ellipsometry; Academic Press: New York, 1993. (32) McClellan, S. J.; Franses, E. I. Colloids Surf., A 2005, 260, 265. (33) McClellan, S. J. Ph.D. Thesis, Purdue University, West Lafayette, IN, May 2005. (34) Polverini, E.; Arisi, S.; Cavatorta, P.; Berzina, T.; Cristofolini, L.; Fasano, A.; Riccio, P.; Fontana, M. P. Langmuir 2003, 19, 872. (35) Phang, T.-L.; McClellan, S. J.; Franses, E. I. Langmuir 2005, 21, 10140. (36) Kop, J. M. M.; Cuypers, P. A.; Lindhout, T.; Hemker, H. C.; Hermens, W. T. J. Biol. Chem. 1984, 259, 3993. (37) Batsonov, S. S. Refractometry and Chemical Structure; Consultants Burea: New York, 1961; pp 17-25.
films. The ATR spectra of the lipid films on a silicon IRE were collected using a custom home-built ATR accessory. Each background spectrum was taken right before the lipid was adsorbed on the silicon IRE. The absorbance spectra of the lipid films were rationed against the background spectra of clean silicon IRE. The IR instrument was continuously purged with dry air from a Balston purge gas generator to reduce the water vapor and carbon dioxide in the sample chamber. The spectra were collected using 512 scans at 4 cm-1 resolution. Happ-Genzel apodization and one level of zero filling were employed, yielding the same data spacing as when the spectra were taken at 2 cm-1 resolution. The spectrum of the water vapor was subtracted from the sample spectra to eliminate the water vapor bands’ interference in the polar group region. The spectra were taken using unpolarized light. The ATR-IR spectra of lecithin lipids in the region between 3000 and 2800 cm-1 displayed peaks typically around 2958 (symmetric stretching of CH3), 2920 (asymmetric stretching CH2), and 2850 cm-1 (symmetric stretching CH2), and in the polar region between 1800 and 1600 cm-1, the spectra displayed a peak around 1738 cm-1 (symmetric stretching CdO). The peak positions of the methylene groups were used to assess the chain conformation, and the peak position of the carbonyl CdO group was used to assess any lipid hydrolysis (none detected). The surface densities of the adsorbed lipid layers were estimated from the peak heights and peak areas of the carbonyl and methylene groups. The latter was estimated without deconvolution of the overlapping peaks. Hence, the latter estimates are expected to be less accurate than the former. The equation for surface density calculations from ATR-IR of thin adsorbed layers has been discussed extensively.39-45 For exsitu experiments in which the IRE that contained the adsorbed film was exposed to air, the absorbance per reflection is given as
( )
2de A Γ ) N dp
(2)
where A is the absorbance, N is the total number of internal reflections (encountering film), de is the effective thickness, dp is the depth of penetration, is the molar absorptivity, and Γ is the adsorbed lipid surface density. Hence, the surface density can be calculated directly if the values of A/N, , de, dp, n21 (refractive index ratio), and θ (angle of incidence) are known. The refractive index of air n2 is 1, and the refractive index of silicon is 3.433 at 2925 cm-1 and 4.080 at 1738 cm-1 (determined from data by extrapolation).46 The dp value is ∼0.23 µm at 2925 cm-1 and ∼0.32 µm at 1738 cm-1. 2.6. Atomic Force Microscopy (AFM). AFM images were acquired in air at room temperature using a Dimension 3100 AFM (a Nanoscope IIIa controller, from Digital Instruments, Santa Barbara, CA). The tapping mode imaging was done using etched silicon cantilevers (Veeco Probes, Santa Barbara, CA) with a resonance frequency of ∼320 kHz, a spring constant of ∼42 N/m, and a tip radius of 0, we infer that γSL is even lower than γSV, possibly because some water penetrates the monolayers and hydrates the polar headgroup. Hence, this explanation seems to be a plausible hypothesis. One would need, of course, further corroborating information or detailed testing from molecular theories, such as those published by Koplic and co-workers.56 Molecular models of DLPC and DPPC show that the zwitterionic group contains a positively
1618 Langmuir, Vol. 22, No. 4, 2006
Phang and Franses
Table 5. Contact Angle, Ellipsometry, and ATR-IR Data for Adsorbed LB Lipid Layers on Silicon IRE ellipsometry data data set
Π mN/m
speed mm/min
type
transfer ratio
θ deg
1 2 3 4
30 30 40 40
5 5 5 5
Z Z Z X
1.13 1.04 1.62 0.27
36 44-46 49-50 30-32
5 6 7 8 9 10 11 12 13 14 15
10 15 15 20 30 30 40 40 40 40 20
5 1 5 8 1 20 1 1 5 10 8
Z Z Z Z Z Z Z Z Z Z X
1.26 2.23 1.33 1.27 1.44 1.06 1.76 1.56 1.10 1.07 0.12
39-37 40-45 37-41 28-37 50-54 35-40 50-58 55-56 36-81 56-57
