9606
Langmuir 2007, 23, 9606-9610
Fluorocarbon Crowning: Langmuir-Blodgett Deposition versus Self-Assembly at Molecularly Rough Surfaces Donald H. McCullough III, Ruslan Grygorash, and Steven L. Regen* Department of Chemistry, Lehigh UniVersity, Bethlehem, PennsylVania 18015 ReceiVed May 8, 2007. In Final Form: June 12, 2007 Langmuir-Blodgett deposition of a single monolayer of 1,2,4,5-tetrakis[(N-(perfluoroundecanoamidoethyl)-N,Ndimethylammonium)methyl]benzene tetrabromide (1) onto a thin film made from alternating layers of poly(diallydimethylammonium chloride) (PDADMA) and poly(4-styrenesulfonate) (PSS) ions affords a uniform fluorinated surface of low energy. An analogous surface that has been constructed by self-assembly shows the same critical surface tension of 16.5 dyn/cm. A comparison of Zisman plots for these two modified films, in combination with analysis by X-ray photoelectron spectroscopy, indicates that Langmuir-Blodgett deposition produces a higher quality and more densely packed fluorocarbon surface that is Very hydrophobic. In sharp contrast, the use of a single-chain analog (i.e., N-(perfluoroundecanoamidoethyl)-N,N,N-trimethylammonium bromide) (2)) affords relatively high energy surfaces by Langmuir-Blodgett deposition and by self-assembly.
Introduction Organic thin films are of considerable current interest for a wide range of applications, including the fabrication of optical devices, sensors, transducers, protective coatings, permeationselective barriers, patternable surfaces, and biomaterials.1,2 Although the structure and composition of the surface of organic thin films play an important role in many applications, “bruteforce” methods of surface modification remain widely used.3 One important example is the fluorination of organic polymers by plasma treatment, which affords etched surfaces having wettability properties that are similar to those of poly(tetrafluoroethylene) (PTFE).4-7 Recent efforts that have been made to prepare PTFE-like surfaces under much milder conditions have focused on the use of fluorocarbon-based polymers as surfacemodifying agents. For example, fluorinated polymeric surfactants have been deposited onto solid substrates by use of conventional Langmuir-Blodgett transfer methods.8 In a second approach, alternating layers of cationic and anionic fluorocarbon-based polymers have been deposited onto solid surfaces using layerby-layer self-assembly protocols.9 In both cases, multilayers of fluorocarbon have been used to form low-energy surfaces. The primary aim of the work that is reported herein was to compare the Langmuir-Blodgett (LB) transfer of single surfactant monolayers with self-assembly (SA) as synthetic tools for modifying molecularly rough polymeric surfaces.10,11 To our * Corresponding author. E-mail:
[email protected]. (1) Swalen, J. D.; Allara, D. L.; Andrade, J. D.; Chandross, E. A.; Garoff, S.; Israelachvili, J.; McCarthy, T. J.; Murray, R.; Pease, R. F; Rabolt, J. F.; Wynne, K. J.; and Yu, H. Langmuir 1987, 3, 932-950. (2) Forrest, S. R. Nature 2004, 428, 911-918. (3) Koberstein, J. T. J. Polym. Sci., Part B: Polym. Phys. 2004, 42, 29422956. (4) Schwarzenbach, W.; Derouard, J.; Sadeghi, N. J. Appl. Phys. 2001, 90, 5491-5406. (5) Niino, H.; Yabe, A. Appl. Phys. Lett. 1993, 63, 3527-3529. (6) Occhiello, E.; Morra, M.; Garbassi, F.; Bargon, J. Appl. Surf. Sci. 1989, 36, 285-295. (7) Bariya, A. J.; Shan, H.; Frank, C. W.; Self, S. A.; McVittie, J. P. J. Vac. Sci. Technol., B 1991, 9, 1-7. (8) Aminuzzaman, M.; Kado, Y, Mitsuishi, M.; Miyashita, T. J. Mater. Chem. 2004, 14, 3014-3018. (9) Jisr, R. M.; Rmaile, H. H.; Schlenoff, J. B. Angew. Chem., Int. Ed. 2005, 44, 782-785. (10) Gaines, G. L. Jr. Insoluble Monolayers at Liquid-Gas Interfaces; Interscience: New York, 1966.
knowledge, this is the first report where such a comparison has been made. Although the transfer of a preassembled surfactant monolayer from the air/water interface to a molecularly rough polymeric surface (i.e., LB deposition) is expected to lead, initially, to a tightly packed array, the diffusion of portions of the assembly into the bulk polymer could result in gaps along the surface. Alternatively, whereas the self-assembly technique has the potential to fill in gaps, it also requires a sufficient surface density of polymeric counterions to produce tightly packed surfactant monolayers. Thus, it is not clear whether deposition by LB or self-assembly methods would result in more efficient surface coverage. A secondary aim of this study was to compare the use of oligomeric versus monomeric surfactants as surfacemodifying agents. Because of the importance of fluorocarbon surfaces in general, we sought appropriate fluorocarbon-based surfactants to address both of these specific aims. Experimental Section Materials. Poly(sodium 4-styrenesulfonate) (PSS, Mw ) 70 000) was purchased from Polysciences Inc. Poly(diallydimethylammonium chloride) (PDADMAC, Mw ) 100 000-200 000) and poly(ethyleneimine) (PEI, Mw ) 70 000) were purchased from Aldrich and used as obtained. Methyl perfluoroundecanoate was purchased from Oakwood Prod., Inc. (West Columbia, SC) and used without further purification. House-deionized water was purified using a Millipore Milli-Q filtering system containing one carbon and two ion-exchange stages. A Nima 612D film balance (Nima Technologies, Coventry, England) was used for all monolayer experiments. A Lab-Line Instruments environ shaker (model 3527) was used for mixing in all self-assembly experiments. 1H NMR spectra were recorded on a Bruker DRX 500 spectrometer (500 MHz). N-[2-(Dimethylamino)ethyl]-perfluoroundecanoic acid amide and 1,2,4,5-tetrakis[(N(perfluoroundecanoamidoethyl)-N,N-dimethylammonium)methyl]benzene tetrabromide (1) were synthesized using experimental procedures described elsewhere.12 N-(Perfluoroundecanoamidoethyl)-N,N,N-trimethylammonium Bromide (2). A solution of N-[2-(dimethylamino)ethyl]perfluoroundecanoic acid amide (844 mg, 1.32 mmol) and MeI (0.29 g, 2 mmol) in 5 mL of CH3CN was stirred for 12 h at 45 °C, followed (11) Ulman, A. An Introduction to Ultrathin Films: From Langmuir-Blodgett to Self-Assembly; Academic Press: New York, 1991. (12) McCullough D. M., III; Grygorash, R.; Hsu, J. T.; Regen, S. L. J. Am. Chem. Soc. 2007, 129, 8663-8667.
10.1021/la7013336 CCC: $37.00 © 2007 American Chemical Society Published on Web 08/14/2007
Fluorocarbon Crowning
Langmuir, Vol. 23, No. 19, 2007 9607 Chart 1
Scheme 1
by additional stirring at 75 °C for 12 h. The solvent was then removed under reduced pressure, and the residue was recrystallized from CH3CN to give 500 mg (48%) of the iodide form of 2. Conversion to the bromide form was then accomplished by dissolving the entire 500 mg (0.64 mmol) in 50 mL of DMF/H2O (3/1 v/v) and stirring the solution with 10 g of the bromide form of an anion-exchange resin (Dowex 1 X2-200) at room temperature for 24 h. The resin was separated by filtration, and the solution was concentrated under reduced pressure. Recrystallization from CH3CN gave 421 mg (44%) of 2. 1H NMR (CF3CO2D) δ: 4.03 (t, 2H); 3.72 (t, 2H); 3.24 (s, 9H). Anal. Calcd for C16H14BrF21N2O: C, 26.36; H, 1.94; F, 54.72; N, 3.84. Found: C, 26.37; H, 2.01; F, 54.54; N, 3.80. Preparation of [Silicon-PEI/PSS/(PDADMA/PSS)2] Substrates. The silicon wafers that were used in this work was obtained from WaferNet, Inc. (San Jose, CA). These wafers were cut into 12 × 20 mm2 pieces and were treated with concentrated H2SO4 and 30% H2O2 (70/30 v/v) at 70 °C for 4 h. (Caution! Piranha solution reacts Violently with many organic materials and should be handled with great care.) The wafers were then rinsed with distilled water and dried in a stream of nitrogen. A silicon wafer was directly immersed into 125 mL of an aqueous solution of PEI (1.0 mg/mL) for 2 min at room temperature, and then it was removed and washed by immersing in 125 mL of pure water for 5 min, followed by air drying. This washing and drying cycle was then repeated once. Similar depositions were carried out with the resulting substrate using 125 mL of an aqueous solution of polymer (1 mg/mL) in the sequence PSS, PDADMA, PSS, PDADMA, and PSS to give the desired substrate: [silicon-PEI/PSS/(PDADMA/PSS)2]. The ellipsometric thickness of this layer-by-layer deposited thin film was typically 3.0 ( 0.5 nm. Monolayer Formation at the Air/Water Interface and Langmuir-Blodgett Transfers. Typically, 500 µL of a surfactant solution (0.2 mg/mL using CF3CH2OH/CH3OH (1/1 v/v) as the solvent) was spread onto a pure water subphase that was maintained at 25 °C. After allowing the solvent to evaporate for 30 min, the film was compressed at a speed of 25 cm2/min to a given surface pressure. For LB transfers, the substrate was first immersed in the subphase prior to deposition of the monolayer at the air/water interface. Monolayers were transferred to the substrate by a single vertical up-trip using constant surface pressure. The transfer speed was 2
mm/min. The substrate was allowed to dry in air for 24 h before further characterization. Surface Viscosity Measurements. For surface viscosity experiments, a home-built canal viscometer (192 mm × 40 mm solid Teflon block having a centrally located 6.0 mm slit) was placed in front of the compressing barrier, and the monolayer was compressed to a surface pressure of 20 dyn/cm. After the monolayer was allowed to equilibrate at this pressure for 60 min, the film was decompressed using a maximum speed of 120 cm2/min, leaving the canal viscometer in its original position. The surface pressure was then recorded as a function of time. The rate of surface pressure decrease was used as a measure of the relative surface viscosity. Self-Assembly of Fluorocarbon Surfactants onto [Silicon-PEI/ PSS/(PDADMA/PSS)2]. Typically, 14 mL of a solution containing a surfactant [1 mg/mL in CF3CH2OH/CH3OH (1/1 v/v)] was added to a 20 mL glass vial that was equipped with a Teflon cap. The substrate was then immersed in the solution in the vial, with the PDADMA/PSS layer facing “up” and the lid screwed on tightly. The vial was then placed in a Lab-Line Instruments environ shaker and was shaken for 24 h at 24 °C and 150 rpm. The substrate was then removed and submerged in surfactant-free CF3CH2OH/CH3OH (1/1 v/v) for 1 min and washed briefly (30 s) with methanol and water. The substrate was then allowed to dry in air for 24 h. Water Contact Angle and Ellipsometric Measurements. Water contact angles were measured using a Rame-Hart NRL model 100 goniometer. A minimum of six measurements were made for three independent drops that were placed at different locations along the substrate. Ellipsometric measurements were made using an automatic null ellipsometer (Rudolph Auto-EL III) equipped with a heliumneon laser (λ ) 632.8 nm) that was set at an incident angle of 70°. Measurements were taken at four spots along the surface of each sample, and the mean and the standard deviations were calculated. Film thickness was determined using the manufacturer’s program, assuming refractive indices of 1.465 for the silicon oxide layer and 1.500 for the surfactant layer. X-ray Photoelectron Spectroscopy (XPS) Measurements. Atomic composition was measured with a Scienta ESCA-300 spectrometer (Scienta Instruments AB, Uppsala, Sweden) with a takeoff angle of 15°. Total fluorine and sulfur atomic compositions were normalized to 100%.
9608 Langmuir, Vol. 23, No. 19, 2007
McCullough et al.
Results Design and Synthesis of Fluorocarbon-Based Surfactants. Specific surfactants that were chosen for this work were 1,2,4,5tetrakis[(N-(perfluoroundecanoamidoethyl)-N,N-dimethylammonium)methyl] benzene tetrabromide (1) and N-(perfluoroundecanoamidoethyl)-N,N,N-trimethylammonium bromide) (2) (Chart 1).12 In principle, the oligomeric nature of 1 should keep this surfactant firmly anchored at surfaces that contain polymeric counterions (e.g., thin films made from alternating layers of poly(diallydimethylammonium) (PDADMA) and poly(4-styrenesulfonate) (PSS) ions (Chart 1). Specifically, 1 was expected to undergo ionic cross linking (i.e., “gluing”) with the polymer.13-15 In addition, the presence of multiple fluorocarbon chains increases the probability of tight packing of the surfactant because only two cationic sites are required for cross linking; that is, even if the remaining pendant chains were not paired with polymeric counterions, they would still serve as nearest neighbors. In contrast, smaller monomeric analog 2 cannot undergo ionic cross linking, and its potential for leaving fluorocarbon gaps along the surface as a result of “sinking” into the bulk film is greater than with 1. Scheme 1 outlines the method of synthesis that was used to prepare 1. Thus, direct condensation of methyl perfluoroundecanoate with N,N-dimethylethylenediamine followed by quaternization with 1,2,4,5-tetrakis(bromomethyl)benzene afforded 1. The synthesis of 2 from a common intermediate, N-[2(dimethylamino)ethyl]-perfluoroundecanoic acid amide, is also shown in Scheme 1. Evidence that 1 Can Undergo Ionic Cross Linking from Monolayer Measurements at the Air/Water Interface. To obtain evidence that 1 is capable of undergoing ionic cross linking by polymeric counterions, we compared its monolayer properties with that of 2 in the absence and in the presence of PSS. Figure 1 shows the surface pressure-area isotherms that were recorded for both surfactants over pure water and over an aqueous subphase that was 5 mM in PSS (repeat unit concentration). For both surfactants, the presence of PSS led to a shift of the isotherm to larger occupied areas per molecule. The fact that PSS exhibited no detectable surface pressure, by itself, indicates that this polyanion associates with these assemblies and prevents tight packing as the monolayer undergoes compression. Relative surface viscosities further revealed a substantial increase in the surface viscosity for 1 when PSS was present in the subphase (Figure 1).16 In sharp contrast, monolayers of 2 exhibited negligible surface viscosity in the absence and in the presence of PSS (not shown); that is, their viscosity profiles were identical with that found for 1 in the absence of PSS. Taken together, these results provide strong inferential evidence for ionic cross linking of 1 by PSS at the air/water interface. Langmuir-Blodgett Deposition of Single Monolayers of 1 and 2. To examine the surface-modifying properties of 1 and 2 using conventional LB methods, we first introduced two alternating layers of PDADMA and PSS ions onto a silicon wafer via layer-by-layer deposition.17-20 Using a strategy similar to (13) Yan, X.; Janout, V.; Hsu, J. T., Regen, S. L. J. Am. Chem. Soc. 2003, 125, 8094-8095. (14) Li, J.; Janout, V.; McCullough, D. H. III; Hsu, J. T.; Troung, Q.; Wilusz, E.; Regen, S. L. Langmuir 2004, 20, 8214-8219. (15) McCullough, D. H. III; Regen, S. L. Chem. Commun. 2004, 2787-2791. (16) Hendel, R. A.; Nomura, E.; Janout, V.; Regen, S. L. J. Am. Chem. Soc. 1997, 119, 6909-6918. (17) Decher, G. Science 1997, 277, 1232-1237. (18) Rouse, J. H.; Ferguson, G. S. Langmuir 2002, 18, 7635-7640. (19) Koestler, S.; Delgado, A. V.; Ribitsch, V. J. Colloid Interface Sci. 2005, 286, 339-348. (20) Chen, W.; McCarthy, T. J. Macromolecules 1997, 30, 78-86.
Figure 1. (A) Surface pressure-area isotherms for (a) 2 over water, (b) 2 over 5 mM PSS, (c) 1 over water, and (d) 1 over 5 mM PSS. (B) Surface pressure as a function of time of exposure toward a 6.0 mm slit opening of a canal viscometer for (a) 1 over water and (b) 1 over 5 mM PSS. All measurements were made at 25 °C. Prior to exposure, each monolayer was maintained for 60 min at 20 dyn/cm. During the equilibration period, the decrease in surface area was less than 5%.
that described elsewhere, poly(ethyleneimine) (PEI) was employed as an adhesive layer.9 The resulting thin film [siliconPEI/PSS/(PDADMA/PSS)2] exhibited an advancing contact angle for water of 30 ( 1° and an ellipsometric film thickness that was typically 3.0 ( 0.5 nm. This substrate was then passed through a monolayer of 1 at the air/water interface by a single vertical up-trip (45 dyn/cm, 2 mm/min, 25 °C, transfer ratio ) 1.0 ( 0.1) to give a surface having an advancing contact angle for water of 122 ( 2° and a receding angle of 99 ( 4°. Analysis by ellipsometry further indicated an overall increase in the thickness of the film of 1.65 ( 0.17 nm. All attempts to remove 1 from this surface by extensive washing with CF3CH2OH/CH3OH (1/1 v/v), water, hexane, 2-propanol, acetone, n-butanol, and toluene failed, as judged by the complete retention of its wetting properties and film thickness. A Zisman plot of the cosine of the contact angle versus surface tension for water, toluene, n-butanol, acetone, 2-propanol, and hexane is shown in Figure 2. On the basis of these data, extrapolation to cos Θ0 ) 1 yields a critical surface tension of 16.5 dyn/cm. To put this value into perspective, the critical surface tension of PTFE has been reported to be ca. 19.2 dyn/cm.21 The (21) Chan, R. K. S. J. Colloid Interface Sci. 1970, 32, 492-499.
Fluorocarbon Crowning
Langmuir, Vol. 23, No. 19, 2007 9609
Figure 2. Zisman plot for a monolayer of 1 deposited on siliconPEI/PSS/(PDADMA/PSS)2 by LB transfer (0) and self-assembly (O). The liquids used for these measurements were (a) hexane, (b) 2-propanol, (c) acetone, (d) n-butanol, (e) toluene, and (f) water. The error in these measurements is less than the size of each symbol. Table 1. Fluorine and Sulfur Atomic Compositions as Determined by XPS surfactant
depositiona
fluorine (%)
1 2 1 2
LB LB SA SA
97.3 95.7 95.1 41.7
a
109 ( 2.4°, respectively; the receding angle was 76 ( 4°. For 2, the thickness and advancing contact angle values were 0.96 ( 0.14 nm and 42 ( 2.1°, respectively. Table 1 lists the relative compositions of fluorine and sulfur for each of these films as determined by XPS analysis. Whereas the F/S ratio for selfassembled monolayers of 1 was similar to that found for 2 under LB transfer conditions, the quantity of 2 that was deposited by self-assembly was very low. In addition, extensive rinsing of the PDADMA/PSS film that was modified with 2 with CF3CH2OH/CH3OH (1/1 v/v) resulted in a substantial loss of surfactant from the surface, as evidenced by the fact that the surface had wettability properties and an apparent film thickness that were similar to that of the bare substrate. An examination of the surface of PDADMA/PSS that was modified by 1 by self-assembly via contact angle measurements gave a Zisman plot that is shown in Figure 2. Although both surfaces showed the same critical surface tension (16.5 dyn/cm), the LB-deposited film gave a steeper slope.
Discussion
sulfur (%)
F/S
2.7 4.3 4.9 58.3
36 22 19 0.72
LB, Langmuir-Blodgett; SA, self-assembly.
lower critical surface tension of this modified surface relative to PTFE is, we believe, a likely consequence of the presence of terminal CF3 groups in 1, which are absent in PTFE. In sharp contrast, similar LB deposition of 2 onto a siliconPEI/PSS/(PDADMA/PSS)2 substrate (15 dyn/cm, 2 mm/min, 25 °C, transfer ratio ) 1.0 ( 0.1) afforded a surface that had an advancing contact angle for water of 51 ( 2.8° and an overall increase in film thickness of 1.2 ( 0.21 nm. To test for the possibility of partial loss of 2 from the outer surface of the film due to penetration into the bulk PDADMA/PSS, we measured the F/S atomic ratio of this modified surface by X-ray photoelectron spectroscopy (XPS) using a 15° takeoff angle. This ratio was then compared with the F/S atomic ratio for a PDADMA/ PSS-based film that had been modified with 1. Because the limiting area per fluorocarbon chain in 1 (31 Å2) is similar to that of 2 (25 Å2/molecule), if both surfactants were localized at the outer surface of the support then the F/S atomic ratios should be similar. However, if 2 has a greater tendency to penetrate into the bulk PDADMA/PSS, then its F/S ratio should be lower than that of 1. Table 1 shows the fluorine and sulfur atomic compositions that were observed for 1 and 2 on PDADMA/PSS, where the total fluorine and sulfur content has been normalized to 100%. The fact that the F/S ratio for 2 is lower than that found for 1 provides direct evidence for a lower concentration of 2 at the outer surface of the film, which supports the hypothesis of greater penetration of 2 into the bulk polymer. It is also noteworthy that extensive rinsing with CF3CH2OH/CH3OH (1/1 v/v) resulted in a substantial loss of 2 from the surface, as indicated by film thickness and contact angle measurements. Deposition of 1 and 2 by Self-Assembly. In parallel experiments, 1 and 2 were deposited onto silicon-PEI/PSS/ (PDADMA/PSS)2 by self-assembly using CF3CH2OH/CH3OH (1/1 v/v) as the solvent. Specific procedures that were used are described in the Experimental Section. On the basis of contact angle and film thickness measurements, equilibrium was reached within 24 h. The apparent thickness of 1 and the advancing contact angle for water on the surface were 1.75 ( 0.28 nm and
Very substantial differences have been found for PDADMA/ PSS surfaces that were modified with 1 versus 2 via LangmuirBlodgett deposition and by self-assembly. In the case of 1, a strongly hydrophobic low-energy surface was produced. With 2, the surface was found to be relatively hydrophilic. This large difference is fully consistent with the hypothesis that the surfactant monomers should have a greater tendency to penetrate organic thin films having molecularly rough surfaces as compared with oligomers and that monomers should be less firmly attached to such surfaces. That 2 does, indeed, penetrate the PDADMA/PSS thin film to a greater extent than 1 is supported by F/S atomic ratios determined from XPS analysis. In addition, that 2 is more weakly associated with this film (i.e., incomplete ion exchange) is apparent from the relatively low fluorine content on the surface that was modified under self-assembly conditions. The fact that extensive rinsing with CF3CH2OH/CH3OH (1/1 v/v) removes 2 but not 1 from the surface provides further support for weaker surface binding of 2. The relatively hydrophilic character of the surfaces that have been modified with 2 is a likely consequence of this weaker association. Zisman plots that were made for 1, deposited onto PDADMA/ PSS by LB transfer and by self-assembly, show a similarity and a difference. Specifically, both plots yield the same critical surface tension, but the slopes are different. This difference in slope can be accounted for in terms of a more perfectly packed and denser fluorocarbon array for the LB-modfied surface. In this case, a well-packed monolayer is “mechanically guided” to the outer surface of the thin film, where it undergoes ionic cross linking. Under self-assembly conditions, 1 has to “find its way” to all available polymeric counterions on the film. If lateral diffusion is restricted, as would be expected for an ionic cross-linked surfactant, then this should preclude the possibility of forming a tightly packed monolayer. Thus, the self-assembled monolayer is expected to be more loosely packed and less well organized. The lower F/S atomic ratio for self-assembled monolayers of 1, relative to the LB-deposited monolayer, shows that self-assembly does indeed produce a lower surface density of 1. Finally, the greater contact angle hysteresis found for self-assembled monolayers of 1 (109°/76°) as compared with that of LB transferred monolayers (122°/99°) also points toward a surface that is less structurally homogeneous.22 (22) Gao, L.; McCarthy, T. J. Langmuir 2006, 22, 6234-6237.
9610 Langmuir, Vol. 23, No. 19, 2007
McCullough et al.
Conclusions
feasibility of producing high-quality fluorocarbon surfaces via a single vertical up-trip by LB transfer, together with the fact that methods for continuous monolayer formation at the air/water interface have already been developed, suggests that this approach could be applicable to large-scale synthesis.23
Both Langmuir-Blodgett and self-assembly methods are capable of producing uniform, low-energy, fluorocarbon-based coatings at molecularly rough surfaces. Although the critical surface tension that characterizes the surfaces prepared by both techniques is the same, Zisman plots, in combination with X-ray photoelectron spectroscopic analysis, have shown that LangmuirBlodgett deposition leads to a higher quality surface that is Very hydrophobic. Results that have been obtained with 1 and 2 also reveal the critical importance of using multiply charged crosslinkable surfactants to modify molecularly rough surfaces by Langmuir-Blodgett deposition and by self-assembly. The
Acknowledgment. This work was supported by the Department of Energy (grant DE-FG02-05ER15720). LA7013336 (23) Albrecht, O.; Eguchi, K.; Matsuda, H.; Nakagiri, T. Thin Solid Films 1996, 284-285, 152-156.
