Understanding Solvation in Hydrofluoroalkanes: Ab Initio Calculations

Jun 20, 2007 - Solvation in hydrofluoroalkanes: how can ethanol help? Denise S. Conti , Jordan Grashik , Lin Yang , Libo Wu , Sandro R. P. da Rocha. J...
2 downloads 11 Views 212KB Size
Download PDF
8096

J. Phys. Chem. B 2007, 111, 8096-8104

Understanding Solvation in Hydrofluoroalkanes: Ab Initio Calculations and Chemical Force Microscopy Libo Wu, Robson P. S. Peguin, and Sandro R. P. da Rocha* Department of Chemical Engineering and Materials Science, Wayne State UniVersity, 5050 Anthony Wayne DriVe, Detroit, Michigan 48202 ReceiVed: February 12, 2007; In Final Form: May 15, 2007

Understanding solvation in hydrofluoroalkane (HFA) propellants is of great importance for the development of novel pressurized metered-dose inhaler (pMDI) formulations. HFA-based pMDIs are not only the most widely used inhalation therapy devices for treating lung diseases, but they also hold promise as vehicles for the systemic delivery of biomolecules to and through the lungs. In this work we propose a combined microscopic experimental and computational approach to quantitatively relate the chemistry of moieties to their HFAphilicity. Binding energy calculations are used to determine the degree of interaction between a propellant HFA and candidate fragments. We define a new quantity, the enhancement factor E, which also takes into account fragment-fragment interactions. This quantity is expected to correlate well with the solubility and the ability of the moieties of interest to impart stability to colloidal dispersions in HFAs. We use a methylbased (CH) segment and its fluorinated analog (CF) to test our approach. CH is an important baseline case since it represents the tails of surfactants in FDA-approved pMDIs. CF was chosen due to the improved solubility and ability of this chemistry to stabilize aqueous dispersions in HFAs. Adhesion force from Chemical Force Microscopy (CFM) is used as an experimental analog to the binding energy calculations. The force of interaction between a chemically modified AFM tip and substrate is measured in a model HFA, which is a liquid at ambient conditions. Silanes with the same chemistry as the fragments used in the ab initio calculations allow for direct comparison between the two techniques. The CFM results provide an absolute scale for HFA-philicity. Single molecule (pair) forces calculated from the CFM experiments are shown to be in very good agreement to the E determined from the ab initio calculations. The ab initio calculations and CFM are corroborated by previous experimental studies where propellants HFAs are seen to better solvate the CF functionality.

1. Introduction Aerosol inhalation therapy has been traditionally used in the treatment of asthma and chronic pulmonary disorders (chronic bronchitis and emphysema).1 However, there is a significant thrust in expanding the role of aerosols as vehicles for the systemic delivery of biomolecules that can be used in the treatment of various diseases including diabetes, cancer, osteoporosis, and multiple sclerosis,2 and as vaccine delivery systems.3 Such interest stems largely from the fact that the bioavailability of biomolecules in the lungs is significantly improved compared to that seen in alternative delivery routes, due to the very large surface area and slow clearance from the deep lungs, and the lower activity of enzymes compared to those present in the liver and kidneys.4 Another important factor that makes aerosol therapy so attractive is its noninvasive nature, which not only increases patient compliance, but may also decrease risks for transmitting blood-born pathogens.5 There are several devices that can be used for pulmonary drug delivery. Those include nebulizers, dry powder inhalers (DPIs), and pressurized metered-dose inhalers (pMDIs).6 While all approaches have advantages and disadvantages, pMDIs are of great significance because they are inexpensive, portable, relatively easy to use, and provide a sealed environment for * To whom correspondence should be addressed. Tel: +1-313-577-4669. Fax: +1-313-577-3810. E-mail: [email protected].

the drug formulation.7 Despite significant inroads made by DPIs in recent years,8 pMDIs still occupy the largest market share worldwide, with sales exceeding 2 billion dollars a year.9 The increased importance of DPIs can be directly linked to the challenges faced by the pharmaceutical industry in reformulating pMDIs with the non-ozone depleting hydrofluoroalkane (HFA) propellants.10 The very limited number of propellant-based drug formulations currently available in the market is a clear testimony of the difficulties in developing new HFA-based pMDIs.7,10 Today’s pMDIs contain small solutes either in solution (drugs soluble in the propellant) or in suspension (with the help of a stabilizer) in HFA propellants.11 Polar compounds, including water and biomolecules, have extremely limited solubility in the low dielectric HFAs.12,13 The ability to disperse biomolecules of interest in pMDI formulations is also affected by the incompatibility of the methyl-based surfactants used in FDAapproved formulations with the somewhat polar semi-fluorinated propellants.14 While significant progress has been made in terms of developing new particle technologies with enhanced stability in pMDIs,15 very little is known regarding the solvation properties of HFAs.16 Understanding solvation forces in HFAs is of great relevance for the development of novel suspension-based pMDIs. The prevalent mechanism for imparting stability to colloidal particles in low dielectric media, such as the HFA propellants, is steric

10.1021/jp071205y CCC: $37.00 © 2007 American Chemical Society Published on Web 06/20/2007

Solvation in Hydrofluoroalkanes

J. Phys. Chem. B, Vol. 111, No. 28, 2007 8097

stabilization. Steric stabilization is dominated by solvation effects,17 which may be achieved by selecting amphiphiles that have an anchor segment that strongly interacts with the dispersed phase, and a well-solvated tail segment that extends into the bulk/dispersing medium. When two colloidal particles approach each other, the stabilizing segments may interpenetrate. If the bulk phase is a good solvent for the stabilizing tail, the interpenetration is thermodynamically disfavored.18 Strong enthalpic interactions (good solvation) between the solvent and the stabilizing moiety, relatively weak interaction between the tail groups on the surface of opposing colloidal particles, and the entropic loss upon entanglement of the stabilizing moieties will all favor colloidal stability.17 It is clear from the discussion above that the ability to quantitatively relate the chemistry of a moiety to its HFAphilicity is paramount in order to design molecules capable of stabilizing dispersions in HFAs. In this work we show how a combined computational and experimental microscopic approach can be used to characterize the HFA-philicity and tail-tail interaction of moieties of interest. Ab initio calculations are used to calculate the binding energy between 1,1,1,2-tetrafluoroethane (HFA134a) and tail fragments (i.e., to probe solvent-tail interaction), and also to determine the tail-tail enthalpic interactions. The proposed methodology is tested with a methylbased tail, and its fluorinated counterpart. The methyl-based tail serves as a baseline in our studies as it represents the FDAapproved surfactants that have very limited solubility and are not able to stabilize suspensions in HFAs (even in the presence of cosolvents).14 The fluorinated-based tail represents the fluorinated surfactants that have been somewhat successful in forming and stabilizing water-in-HFA (W/HFA) microemulsions.19,20 The results from the ab initio calculations are corroborated and complemented by Chemical Force Microscopy (CFM) studies. Adhesion forces between chemically modified AFM tip and substrate are measured in the presence of 2H,3H-perfluoropentane (HPFP), a mimicking solvent to HFA propellants. The chemistries that modify the tip and substrate are the same as those in the ab initio studies, thus allowing for a direct comparison of both studies. 2. Methods and Materials 2.1. Ab Initio Calculations. Binding energies (Eb) were computed using the supermolecule approach:21

Eb ) Est - Es - Et

(1)

where Est is the total energy of the complex HFA134a + fragment of interest, and Es and Et are the energies of the isolated HFA134a and fragment, respectively. Using this approach, more negative Eb’s represent more energetically favorable pairs. Representative fragments are investigated instead of the whole moiety in order to make such computational intensive calculations feasible. In this case we have selected propane as methylbased fragment (represented by the nomenclature CH), and perfluoropropane as fluorinated-based fragment (represented by the nomenclature CF). Both raw Est, and those corrected for basis set superposition error (BSSE) are reported here. The BSSE was determined using the counterpoise (CP) method of Boys and Bernadi.22 Calculations were carried out using Gaussian 03.23 Complete geometry optimizations calculations were done for each fragment and the complexes at the secondorder Møller-Plesset perturbation method (MP2), and 6-31+g(d,p) basis set. Single point MP2 energy calculations have been carried out with the Dunning basis set, aug-cc-pVDZ.24 The

results of these optimizations correspond to energy minima since no imaginary frequencies were observed.25 The average of the raw and counterpoise corrected interaction energies has been shown to be a good approximation to the complete basis set limit, thus avoiding computationally expensive calculations at higher levels of theory and larger basis sets.26-28 Also, the overestimation of the energy at the MP2 level of theory is minimized when using the average.27,29 Therefore, whenever both raw and CP-corrected energies are available, we discuss the results in terms of average energies. We assumed an error of (0.10 kcal‚mol-1 in the binding energy values. This value is taken from the difference in Eb’s for the CF4 dimer calculated with the aug-cc-pVDZ and the basis set limit, both at the MP2 level of theory.30 This value is larger than the difference in Eb’s for the same complex that arises on going from MP2 to CCSD(T), with the agu-cc-pVDZ basis set.31 2.2. Experimental. 2.2.1. Materials. Octyltrichlorosilane (C8TS, 97%) and trichloro (1H,1H,2H,2H-perfluorooctyl)silane (FC8TS, 97%) were purchased from Aldrich Chemicals Co. Isooctane (ISO, 99.4%), chloroform (99.9%), acetone (99.8%), sulfuric acid (95.8%), hydrogen peroxide (30%), and hexane (99.9%) were purchased from Fisher Scientific. 2H,3H-Perfluoropentane (HPFP, 98%) was obtained from SynQuest Labs., Inc. All chemicals were used as received. Deionized water (NANOpure DIamond UV ultrapure water system: Barnstead International), with a resistivity of 18.2 MΩ·cm and surface tension of 73.8 mN‚m-1 at 296 K, was used in all experiments. Glass slides (22 mm2, No. 2) were obtained from Corning Labware and Equipment. Si3N4 V-shaped contact mode cantilevers with integrated pyramidal tips were purchased from Veeco Instruments Inc. (Model: NP-20). The reported (manufacturer) spring constant is 0.06 N‚m-1, and the tip radius is 20-60 nm. For tips within the same batch, the spring constant was found to be in the range between 0.058-0.061 N‚m-1, and the tip radius between 28-38 nm. The spring constant was determined using our AFM, and the MI Thermal K 1.02 software (Molecular Imaging Co.). A power spectrum of the AC signal was used to determine the mean-squared amplitude of the cantilevers, which is used to solve for spring constant.32,33 The tip radii were measured using a TGT01 calibration grating template (MikroMasch), and the Scanning Probe Image Processor (SPIP software from Image Metrology A/S). 2.2.2. Pretreatment of Substrates before Chemical Modification. The glass microscope slides were first degreased under ultrasonic stirring in chloroform for 10 min, and then placed in a freshly prepared piranha solution (a 7:3 v/v mixture of 95.8% H2SO4 and 30% H2O2) at 373 K for 20 min. (Warning: Piranha solution should be used with great care since it is extremely caustic and reacts violently with organic solvents.) This step serves to remove organic contaminants and produce a glass surface with high concentration of hydroxyl groups, which act as attachment points for the trichlorosilanes. Subsequently, the substrates were rinsed with deionized water and blown-dried with a light stream of dry nitrogen. The AFM Si3N4 probes were initially rinsed with chloroform and immersed in piranha solution for 20 min at 373 K. They were then rinsed with deionized water and dried in a vacuum oven for 15 min. 2.2.3. Surface Chemical Modification. Solution Deposition. C8TS and FC8TS were deposited from solution according to established literature procedures.34-37 Immediately after the pretreatment described above, the dried glass slides and AFM probes were immersed in a 2.5 mM solution of C8TS or F8TS in hexane, for various deposition times, at room temperature. The glass slides and AFM probes were subsequently removed

8098 J. Phys. Chem. B, Vol. 111, No. 28, 2007

Wu et al.

Figure 1. Schematic diagram of the chemical force microscopy (CFM) measurements: (a) force-distance curve (force curve); (b) approach/ retract cycle of the force curve.

from the silane solution, and immediately rinsed with copious amounts of chloroform. Finally, the glass slides were blowndried with nitrogen. The AFM probes were dried in air. Vapor-Phase Deposition. An alternative approach to the surfacemodificationdescribedaboveisthegas-phasetechnique.35-37 Silane self-assembled monolayers were also deposited from the vapor phase onto the pretreated substrates. The deposition was achieved by placing the substrates and AFM probes inside a desiccator containing a Petri dish with a small amount of either pure C8TS or F8TS. The deposition was performed under light vacuum at room temperature. The samples were exposed to the silane vapor for various deposition times. After deposition, the glass slides and AFM probes were washed thoroughly with chloroform. The glass slides were dried with nitrogen. The AFM probes were allowed to dry in air. 2.2.4. Contact-Angle Measurements. The static contact angle of the modified substrates was determined in order to follow monolayer deposition and quality. Contact angle was determined on a goniometer (KSV CAM 200)38 equipped with a CCD camera. All data were collected at room temperature. 2.2.5. Chemical Force Measurements. Figure 1 is a schematic illustration of a typical measurement of the force of interaction between the silane-modified AFM tip and substrate during an approach/retract cycle. The adhesion force (Fad) is the product of the cantilever deflection (∆x) and the spring constant of the tip (k). Fad measurements were carried out on a PicoSPM LE atomic force microscope (Molecular Imaging Co.), at room temperature. A fluid cell was used to conduct the experiments in liquid environment in order to evaluate the solvation effects. The force-distance curves (force curves) were measured in ISO and HPFP. HPFP is a liquid HFA at ambient conditions. It has been extensively used as a mimicking solvent to HFA propellants. All probes and glass slide substrates were rinsed with chloroform and dried just prior to mounting them into the AFM fluid cell. Several contact points randomly distributed on the sample surfaces were picked for the measurements. At each contact point, 25 force curves were recorded. The average Fad was determined by fitting a Gaussian profile to the histogram of forces.39-41 The tip-substrate combination was approached and retracted in a range of 2000 nm, with a speed of 4 s‚cycle-1. Repeated Fad measurements for several tip-substrate combinations indicated a small variation on the average Fad. 3. Results and Discussion In order to design moieties with high affinity for HFAs, it is necessary to quantitatively relate the chemistry of such func-

tionalities to their HFA-philicity. The solubility and ability of HFA-philes in stabilizing aggregates in hydrofluoroalkane propellants is directly related to how well they are solvated in HFAs. The combined computational and experimental approach shown below provides a rational framework for designing HFAphiles, and can be extrapolated to other solvent systems as well. Such studies are relevant not only in the context of the traditional (commercially available) solution and dispersion pMDI formulations, which generally require surfactants as excipients,16 but also for the development of novel HFA-based formulations that can be potentially used for the systemic delivery of biomolecules to and through the lungs, as for example reverse aqueous microemulsions in HFAs,19,20,42-44 and other particle-based approaches.45,46 3.1. The “Enhancement Factor”: Solvation, Solubility, and Dispersion Stability in HFAs. 3.1.1. HFA134a-Fragment Interaction Energy. In an effort to quantify the HFA-philicity of different chemistries, we determined the interaction energy between HFA134a and candidate moieties. Due to the computational costs associated with calculations of many atoms at high levels of theory and large basis sets, we have performed calculations with fragments that are representative of the moieties of interest. A methyl-based fragment (propane, CH) and its perfluorinated counterpart (perfluoropropane, CF) have been selected to demonstrate the applicability of the proposed approach. CH serves as a baseline moiety. It represents the tailgroups of surfactants that make up the FDA-approved pMDI formulations. CH-based surfactants have extremely limited solubility in HFAs, and are not capable of stabilizing dispersions in the semi-fluorinated propellants.16 The CF fragment represents the tails of fluorinated-based surfactants that have been capable of forming and stabilizing aqueous dispersions in HFAs.19,20,47 CH and CF are the prototypes of the C8TS and FC8TS tails, respectively, which are used in the CFM studies discussed later in this work. The single point Eb’s between HFA134a and the CH and CF fragments were calculated at the MP2/aug-cc-pVDZ, from the optimized geometries at the MP2/6-31g+(d,p) level of theory. These results are summarized in Table 1, along with the binding energy for the CH-CH and CF-CF fragments that have been previously reported in the literature.30 The binding energies reveal that HFA134a interacts more strongly with the CF fragment (Eb ) -2.43 kcal‚mol-1) than with its hydrogenated counterpart (Eb ) -1.83 kcal‚mol-1). As previously reported,48 differences in binding energies of about 0.1 kcal‚mol-1 or smaller may be deemed not significant. Even though the

Solvation in Hydrofluoroalkanes

J. Phys. Chem. B, Vol. 111, No. 28, 2007 8099

TABLE 1: Interaction (Binding) Energy of the HFA134a-Fragment (CH or CF) Pairs (Estb ), and the Fragment-Fragment (Ettb ) Pairs, and the Calculated Enhancement Factor (E ) Estb /Ettb ) binding energy (kcal‚mol-1) HFA134a-fragm ent

fragment-fragment

enhancement factor

fragment

average Estb a

CP-corrected Estb a

CP-corrected Ettb b

Ec

CF CH

-2.43 -1.83

-1.20 -1.10

-1.45 -1.85

0.83 0.59

a This work. Optimization and single point energy calculation performed at the MP2/6-31g+(d,p) and MP2/aug-cc-pVDZ level of theory, respectively. b Tsuzuki et al.30 Optimization and single point energy calculation performed at MP2/6-311g(d) and MP2/aug(df,pd)-6-311G** level of theory, respectively. c E is calculated from the CP-corrected Estb and Ettb .

magnitude of the difference in energy between HFA134a-CH and -CF is relatively small, this value is significantly larger than the error expected here. Moreover, the magnitude in energy is expected to be significantly augmented in bulk, thus justifying the differences experimentally observed between methyl and fluorinated moieties in HFAs. These results are in agreement with experimental observations. Alkyl-based moieties show limited solubility in HFAs,14 and are not capable of stabilizing dispersions in such propellant systems.16 Such surfactants perform very poorly as dispersing agents even in the presence of cosolvents.49 On the other hand, fluorinated moieties have been, to some extent, successful in stabilizing reverse aqueous aggregates in HFAs.19,20,47 The Eb’s shown here can be contrasted with those reported for another compressible solvent, namely compressed CO2. Complexes of hydrocarbons and fluorocarbons with CO2 have lower (less negative) Eb’s, of ∼ -1.0 kcal‚mol-1.50-52 The enhanced binding energies observed in HFA systems are attributed to its more polar nature and higher polarizability, and the larger polarizability of the tail fragments. We have previously investigated the solvation of longer methyl-based fragments in HFA134a.43 The average Eb of the HFA134a-isohexane complex was found to be -2.72 kcal‚mol-1. Although HFA134aisohexane and HFA134a-propane complexes show the same number of interaction sites, the longer alkane chain makes less bifurcated C-F‚‚‚H contacts with HFA134a. The bifurcation present in the system with the smaller alkane (CH) causes the bonds to be further apart from each other, thus decreasing the overall energy (less negative). The larger polarizability of isohexane will also enhance the interaction energy of the first tail with HFA134a. Our previous results also show that HFA134a interacts significantly better with a more polar fragment (propyleneoxide, PO), with an average Eb of -6.36 kcal‚mol-1. The results suggest the following order of HFAphilicity PO > CF > CH, which is corroborated by interfacial activity, aggregation behavior, and solubility studies.42,43,53 Further insight into the nature of the interactions between HFA134a and the tail fragments can be gained by analyzing the optimized geometries of these complexes. The interaction points in HFA134a-CF and HFA134a-CH are indicated by dashed lines in panels a and b, respectively, of Figure 2. The distances are given in Å. The Cartesian coordinates of the complexes are summarized in Tables S1 and S2 in Supporting Information. In total, we can identify five interaction points between HFA134a and the CH moiety, and one between HFA134a and CF. Similar interactions were observed previously in the HFA134a-isohexane complex.43 All short C-F‚‚‚H distances present in both complexes are inside of the range observed in crystalline fluorobenzenes (2.36-2.86 Å).54 The longest C-F‚‚‚H interaction in the HFA134a-CH complex is inside of the range reported in complexes formed between fluorobenzene and N-methylformamide or benzene (2.18-2.98

Figure 2. Optimized geometry of the (a) HFA134a-CF and (b) HFA134a-CH complexes at the MP2/6-31g+(d,p) level of theory. Interatomic distances (in bold) are in Å.

Å).55 All C-F‚‚‚H interactions are inside of the range of short contacts reported in CHF3 dimers (2.63 and 3.02 Å).56 Based on a distance criterion, the interactions are within the typical range for weak H-bonds.57 The single interaction in the HFA134a-CF complex demonstrates the loss of attractive electrostatic interaction between hydrogen and fluorine atoms due the repulsive electrostatic interaction between fluorine atoms (complex).56 Even though the HFA134a-CH complex has more interaction points, it is still less energetically favorable than the HFA134a-CF complex. As observed earlier for some CHF3 dimer configurations, the orientation of the C-F‚‚‚H bond in the HFA134a-CH is not sufficient to enhance its binding energy.56 Even though there are several C-F‚‚‚H bonds in the HFA134a-CH complex (weaker), those are not enough to enhance the binding, highlighting the orientation dependence of such interactions. 3.1.2. The Enhancement Factor. Combining the SolVentFragment and Fragment-Fragment Interaction Energies. While the Estb ’s provide quantitative information on the degree of interaction between the propellant and the tail fragments, the solubility of a moiety in HFA134a will depend not only on solvent-solute (fragment) interactions, but also on the interactions arising due to the solute-solute favorable contacts. In this case, a better indicator of the solubility of a fragment (chemistry of interest), can be defined as the ratio of the binding energy of the solvent-fragment pair (Estb ) over the binding energy for the fragment-fragment pair (Ettb ). Here we define this ratio as the solubility enhancement factor, E. Based on this definition, a

8100 J. Phys. Chem. B, Vol. 111, No. 28, 2007

Wu et al.

Figure 3. Variation of the static water contact angle (θ) with deposition time for the C8TS- and FC8TS-modified glass slides: (a) solution deposition; (b) vapor deposition.

larger E represents enhanced solubility, and stronger solventfragment interaction. E can, therefore, be related to the ability of a moiety to impart stability to dispersions in HFAs. If the fragments are well solvated, and do not interact strongly with each other (weak tail-tail interaction), dispersion stability is enhanced.17 Fragment-Fragment Interaction. The calculated value for the single point (interaction) energy of the CH-CH dimer (Ettb ) was -2.50 kcal‚mol-1, with a CP-corrected value of -1.69 kcal‚mol-1, determined at the MP2/aug-cc-pVDZ, with geometry optimized at the MP2/6-31g+(d,p) level of theory. The optimized geometry for the propane dimer obtained in this work has similar orientation to the configuration “G” studied by Tsuzuki.30 The Cartesian coordinates for the CH-CH pair are given in Table S3 in Supporting Information. The reported CPcorrected Eb of configuration ‘G’ (MP2/aug(df,pd)-6-311G**) at the potential minimum is -1.85 kcal‚mol-1.30 The Ettb for the CH-CH pair calculated at MP2/aug(df,pd)-6-311G** and MP2/aug-cc-pVDZ level of theory are seen to differ by less than 0.2 kcal‚mol-1. Given the good agreement in the calculated energies for the CH-CH pair, and the small difference in binding energies of the CF4 (-0.69 and -0.44 kcal‚mol-1) and C2F6 (-1.02 and -0.82 kcal‚mol-1) dimers calculated at aug(df)-6-311G* and aug-cc-pVDZ basis set, respectively,30 we decided to use the results for the C3F8 dimer from the same work. While one should be careful in extrapolating the energies from the aug(df)-6-311G* to the aug-cc-pVDZ basis set as done here, we feel comfortable with the approach since (i) the difference in energy between the two basis sets of the smaller perfluoroalkanes is observed small, and (ii) even if the binding energy value for the C3F8 dimer at the aug-cc-pVDZ basis set turns out to be 0.25 kcal‚mol-1 less negative then the value employed here (as is the worse case scenario for the smaller perfluoroalkanes), the conclusions of this work will remain unchanged. The CP-corrected value reported for the CF dimer is -1.45 kcal‚mol-1.30 These results are summarized in Table 1, along with the values for the enhancement factor. E is calculated with the CP-corrected Ett and Est values, instead of the average of the raw and CP-corrected values, since the raw value for the dimers has not been reported.30 The Ett results show that the CH fragments interact more strongly to one another than their perfluorinated counterparts. This has been attributed to the greater steric repulsion between CF fragments, which increases the intermolecular separation.30 Dispersion forces are considered to be the major contribution to the attractive energy for both dimers.31,58-60 The larger E for the HFA134a-CF system (0.83) compared to that of HFA134a-CH (0.59) suggests that the fluorinated moiety is a

better HFA-phile. Such fact is corroborated by the experimental evidence discussed earlier. 3.2. Single Molecule (Pair) Force: Solvation from Chemical Force Microscopy (CFM). Perhaps the closest microscopic experimental analog to the calculations described above (the E factor) is the adhesion force (Fad) obtained by CFM. The CFM technique is based on the modification of an AFM tip by covalently attaching an organic monolayer with predefined functional groups.40 The force of interaction between the modified tip and substrate of interest is subsequently measured in air or in liquid. Fad can be used to characterize the solvation capacity of both polar and nonpolar solvents.39,41 It is a direct measure of the enthalpic penalty for creating the interfaces between the modified tip and solvent, and modified substrate and solvent, when the tip-substrate contact is broken upon retraction of the cantilever. Large Fad will be characteristic of a poor solvent environment for the (tail) groups on the substrate and AFM tip. On the other hand, small Fad will be detected when the tail groups are well solvated by the solvent environment.39 CFM can, therefore, provide quantitative information on the relationship between tail chemistry and its HFA-philicity. The CFM results are discussed below. 3.2.1. Substrate Modification. Contact angle (θ) measurements were performed to confirm the formation and quality of the C8TS and FC8TS monolayers on the piranha-treated glass substrates. The relationship between deposition time and static water θ is shown in Figure 3, for monolayers deposited from both solution and the vapor phase. Although, saturation is reached at earlier times when using a solution-based deposition, slightly more compact monolayers are seen with the vapor-phase deposition. This is in agreement with previous reports.37 The θ at saturation for the C8TS-modified substrate was 104.3° and 106.9° for the solution and vapor-phase deposition, respectively. A similar trend was observed for the FC8TS-modified substrates, with saturation θ‘s of 108.1° and 113.0° for the solution- and gas-phase deposition, respectively. These numbers are in agreement with previous literature values,34,35,61,62 and indicate formation of a compact layer modifying the substrates. 3.2.2. Tip Modification. The Si3N4 AFM tips were modified with C8TS and FC8TS using both the vapor phase- and solutionbased methods. Based on the fact that both techniques successfully produced compact monolayers on the cover glass substrates, and on the reported success in modifying AFM tips with solution-based deposition methods,34-37,61 one can also expect that the tips would have been effectively modified with the procedures described abovesat least when using the solutionbased method. To confirm the surface modification of the tips with the vapor-phase technique, we performed Fad measurements

Solvation in Hydrofluoroalkanes

J. Phys. Chem. B, Vol. 111, No. 28, 2007 8101 TABLE 3: Adhesion Force (Fad) between C8TS- and FC8TS-Modified Tip and Substrate in HPFP and Isooctane (ISO), at 298 Ka Fad/R (mN‚m-1) CFM deposition technique vapor solution

modified tip and substrate C8TS (CH-CH) FC8TS (CF-CF) C8TS (CH-CH) FC8TS (CF-CF)

solvent medium HPFP

ISO

68.2 ( 17.6 6.4 ( 1.8 82.7 ( 12.1 10.9 ( 3.9

0 39.1 ( 26.4 0 38.8 ( 17.3

a The Fad has been normalized by the radius of curvature (R) of the AFM tip. The results for both the vapor phase and solution deposition procedure are shown.

Figure 5. Average adhesion force for C8TS- (CH-CH) and FC8TS- (CF-CF) modified tip and substrate in HPFP and isooctane (ISO). Fad has been normalized to the radius of curvature (R) of the AFM tip. Tips and substrates were prepared using vapor-phase deposition, and data were collected at 298 K.

Figure 4. Adhesion force (Fad) histograms of: (a) unmodified tip; (b) piranha-treated tip; and (c) vapor-phase C8TS-modified tip, against the vapor-phase C8TS-modified glass slide. All experiments were performed in HPFP at 298 K.

TABLE 2: Average and Deviation of the Adhesion Force (Fad) between Si3N4 AFM Tip at Different Stages of Surface Treatment, and C8TS-Modified Substrate (Vapor Deposition) in HPFP and Isooctane (ISO) Solvent, at 298 K Fad (nN) CFM

solvent medium

AFM tip

modified substrate

unmodified piranha-treated vapor C8TS-modified

C8TS (CH-) C8TS (CH-) C8TS (CH-)

HPFP

ISO

0.25 ( 0.28 0.18 ( 0.05 0.64 ( 0.09 0.55 ( 0.12 2.25 ( 0.58 0

between the AFM tips, at different stages of the surface treatment, and a C8TS-modified substrate in both HPFP and ISO. The histograms of Fad in HPFP are shown in Figure 4. The histograms in ISO were also determined, but only the average forces are reportedsTable 2. The trends and observed changes in Fad that accompanied each surface treatment strongly indicate that each step was successfully accomplished. It is interesting to note that relatively narrow Fad distributions were observed for the unmodified and piranha-treaded tipsspanels a and b, respectively, of Figure 4. This shows that the substrate modification was fairly uniform. A significantly broader distribution was seen for Fad measured between the C8TS-modified

tip and substratesFigure 4c. Because Fad is affected by the number of contact pairs at the rupture point, larger variations on the number of contacts pairs at each approach/retract cycle will tend to increase the width of the Fad distribution. This may be attributed to either the surface roughness of the glass substrate and/or defects in the deposited monolayer on both the substrate and the tip. A similar trend was observed for the Fad in ISO. The treatment of the tip with piranha solution results in an increase in Fad. This is an expected trend because the more polar tip will interact less favorably with the alkane solvent. After the modification of the tip and substrate with C8TS, the Fad between them decreases to zero since no (or minimal) work is required to separate them in ISO. Similar results are reported in the literature.36 Combined, these results indicate that the surface modification of both tip and substrate was successfully accomplished. 3.2.3. Adhesion Force and SolVation. A summary of all the Fad results for the baseline alkyl moiety (C8TS, which represents the tails of the surfactants present in FDA-approved pMDI formulations) and its fluorinated counterpart (FC8TS) is given in Table 3. The Fad measurements for substrate/tip modified by the solution and the vapor deposition methods, in both isooctane (ISO) and HPFP, were performed at 298 K. The average force curves, along with the Fad normalized by the radius of curvature (R) of the tip, are shown in Figure 5. The results represent an average of approximate 200 force curves for each system. The Fad result for C8TS-modified tip and substrate (here we represent C8TS as CH since it contains the methyl group) in ISO is of great relevance. It represents a reference state, one which is closest to “ideal solvation” (Fad ≈ 0). The force of interaction is not actually zerosthe tethered alkyl-based silanes

8102 J. Phys. Chem. B, Vol. 111, No. 28, 2007

Wu et al.

interact with each other through dispersion interactions, and work must be performed to break those “bonds”, just outside of the detection limit of the experiment. The reported Fad for CF-CF in perfluorooctane is also zero,63,64 again suggesting ideal solvation for that system. These results serve as a guide to understand the solvation of different tail chemistries by hydrofluoroalkanes. The Fad was determined in HPFP, an HFA that is a liquid at ambient conditions. The dielectric constant and dipole moment of HPFP is close to that of HFA propellants.65 In order to put the HPFP results in perspective, force-distance curves between CF-CF modified tip and substrate were also collected in ISO. The low Fad between CF-CF (tip-substrate) (Fad-CF-CF/R ) 6.4 mN‚m-1) indicates that the fluorinated moiety is significantly better solvated by HPFP than its hydrogenated counterpart (Fad-CH-CH/R ) 68.2 mN‚m-1). However, the solvation is still very far from ideal (Fad-CH-CH/R ) 0 mN‚m-1 in ISO). The smaller Fad-CF-CF is in direct agreement with the more negative Estb and the larger E values obtained for the HFA134a-CF system from the ab initio studies discussed above. CFM results provide, however, an important additional quantitative piece of information, which is the degree of solvation of the candidate moiety relative to ideal solvation. Very similar results were obtained for the systems modified by the solution deposition method (see Table 3), indicating the reliability of the Fad results discussed in this work. 3.2.4. Single Molecule (Pair) Force (Fs). The Fad results discussed above can be further normalized by calculating the single molecule (pair) force Fs. As described by the JohnsonKendall-Roberts (JKR) model,39,40 the adhesion force can be calculated by

Fad ) 1.5πRWtip-medium-substrate

(2)

where R is the radius of curvature of the AFM tip. Wtip-medium-substrate is the work per unit area required to separate the modified tip and substrate in the solvent medium. By applying the JKR model to the adhesion force data described above, an effective contact area between tip and substrate (as the contact ruptures) can be calculated. The contact radius (a) defined by the JKR theory is

a)

(

)

1.5πR2Wtip-medium-substrate K

1/3

(3)

where K, the reduced elastic modulus for the tip-substrate combination, is given by

(

)

2 1 - Vsubstrate2 1 3 1 - Vtip ) + K 4 Etip Esubstrate

(4)

where ν is the Poisson ratio and E is the Young’s modulus.66 The corresponding number of contact groups n at rupture can be determined from the knowledge of the contact radius and the area occupied per molecule A.67,68 In this way, the single molecule (pair) force Fs can be calculated from Fs ) Fad/n. The calculated values for Fs and n are summarized in Table 4. It can be seen that this normalization actually brings the results from the vapor and solution deposition methods in even closer agreement. It is interesting to note the smaller number of contacts n observed in the experiments where both tip and substrate have been modified with the perfluorinated silanes. This trend is in agreement with the larger area per molecule occupied by the bulkier fluorinated species (ACH ) 21 and ACF ) 25.4 Å2‚molecules-1).67,68 The reasons for the observed

TABLE 4: Singe Molecule (pair) Force (Fs) between C8TS(CH-CH) and FC8TS- (CF-CF) Modified Tip and Substrate in HPFP and Isooctane (ISO) at 298 Ka Fs (pN) (n) CFM deposition technique vapor solution

solvent medium

modified tip and substrate

HPFP

ISO

C8TS (CH-CH) FC8TS (CF-CF) C8TS (CH-CH) FC8TS (CF-CF)

156 (14) 86 (2) 166 (16) 102 (3)

0 156 (8) 0 156 (8)

a Fs (Fs ) Fad/n) obtained by normalizing the determined adhesion force (Fad) over the number of contact points (n), calculated using the JKR theory.

differences in n for the CF-CF in the different solvents are less obvious. One possible explanation is that CF-based silanes attached on the sides of the AFM (pyramidal) tip might be weakly interacting with their counterparts on the substrate. Those bonds could be broken before the effective Fad is detected; i.e., the smaller adhesive contacts would not contribute to the total Fad. Due the smaller penalty for breaking such contacts in HPFP, fewer effective interacting pairs or a smaller n is seen in HPFP. Fs for the CF-CF system (86 pN) is only about 55% of that of the CH-CH pair (156 pN), indicating a much better solvation of the fluorinated tail in the hydrofluoroalkane solvent. It is interesting to compare/contrast these results with the E values obtained from the ab initio calculations discussed above, where the E for the HFA134a-CH system is shown to be only ∼70% of that for HFA134a-CF. This remarkable agreement shows the complementary nature of these approaches. The trends in both CFM and E are also corroborated by macroscopic experimental results.14,20,42 4. Conclusions In this work we used a combined microscopic computational and experimental approach to understand solvation in hydrofluoroalkanes (HFAs). Ab initio calculations and chemical force microscopy (CFM) were used to quantitatively relate the chemistry of moieties to their HFA-philicity. A methyl-based (propane, represented by CH) fragment and its perfluorinated counterpart (represented by CF) were chosen to test the approach. Binding energy calculations performed at the MP2/ aug-cc-pVDZ level of theory reveal more favorable intermolecular interactions for the 1,1,1,2-tetrafluoroethane (HFA134a)CF complex (average Estb ) -2.43 kcal‚mol-1), compared to that between HFA134a and CH (average Estb ) -1.83 kcal‚mol-1). We also defined a new quantity, the enhancement factor E ) Estb /Ettb , where Ettb is the fragment-fragment binding energy. This parameter takes into account not only solventsolute interactions, but also solute-solute interactions. E, therefore, is expected to be a better predictor of the solubility of HFA-philes, and their ability to impart stability to colloidal dispersions in HFAs. The results suggest that fluorinated tails (E ) 0.83) are better HFA-philes than methyl-based moieties (E ) 0.59). These results are in direct agreement to previously reported solubility and dispersion stability results.14,20,42 Adhesion force (Fad) measurements from CFM are not only complementary to the ab initio calculations, but also provide a solvation scale which cannot be obtained by binding energy results. The Fad for a methyl-modified (C8TS) tip and substrate in a model HFA propellant is 68.2 mN‚m-1, which is significantly larger than that for the equivalent fluorinated silane (FC8TS), of 6.4 mN‚m-1. This indicates a greater enthalpic penalty

Solvation in Hydrofluoroalkanes for creating an interface between methyl-based groups (on the tip and substrate) in the liquid HFA; i.e., the fluorinated group is a better HFA-phile. The Fad for C8TS-modified tip and substrate in an alkane solvent is zero; i.e., there is no penalty for creating those interfaces since the C8TS is ideally solvated by the alkane solvent. Therefore, Fad results provide an absolute solvation scale, and are of great relevance for the rational design of HFA-philes. Using the Johnson-Kendall-Roberts theory,39,40 information on the force between single pairs of molecules (Fs) can be obtained. Fs for FC8TS in the mimicking HFA is 86 pN, which is only about 55% of that observed between a pair of C8TS molecules (156 pN). This result is in very good agreement with the E factor from the ab initio calculations. The proposed methodology is broad, and it is expected to be applicable to any solvent-solvophile system. Acknowledgment. WSU for start-up funds, a Ph.D. assistantship for LW, and an RA-ship (IMR at WSU) for RP. We also acknowledge Dr. Schlegel’s group at WSU for useful discussions related to the ab initio calculations, GRID/WSU for computer time, and NSF-CBET 0553537 for financial support. Supporting Information Available: Cartesian coordinates for HFA134a-CF and HFA134a-CH complexes, and CHCH dimer at MP2/6-31g+(d,p) level of theory. This material is available free of charge via the Internet at http://pubs.acs.org. References and Notes (1) Hofford, J. M. J. Fam. Practice 1992, 34, 485-492. (2) Laube, B. L. Respir. Care 2005, 50, 1161-1176. (3) Bivas-Benita, M.; Ottenhoff, T. H. M.; Junginger, H. E.; Borchard, G. J. Controlled Release 2005, 107, 1-29. (4) Patton, J. S. AdV. Drug DeliVery ReV. 1996, 19, 3-36. (5) Sullivan, V. J.; Mikszta, J. A.; Laurent, P.; Huang, J.; Ford, B. Expert Opin. Drug DeliVery 2006, 3, 87-95. (6) Knecht, A. J. Aerosol Med. 1991, 4, 189-192. (7) Labiris, N. R.; Dolovich, M. B. Br. J. Clin. Pharmacol. 2003, 56, 600-612. (8) Frijlink, H. W.; De Boer, A. H. Expert Opin. Drug DeliVery 2004, 1, 67-86. (9) Terzano, C. Pulm. Pharmacol. Ther. 2001, 14, 351-366. (10) Smyth, H. D. C. Expert Opin. Drug DeliVery 2005, 2, 53-74. (11) Johnson, K. A. In Inhalation Aerosols: Physical and Biological Basis for Therapy, 2nd ed.; Hickey, A. J., Ed.; Informa Healthcare: New York, 2007; pp 347-371. (12) Patton, J. S. Ther. Proteins 1993, 329-347. (13) Byron, P. R. AdV. Drug DeliVery ReV. 1990, 5, 107-132. (14) Vervaet, C.; Byron, P. R. Int. J. Pharm. 1999, 186, 13-30. (15) Edwards, D. A.; Hanes, J.; Caponetti, G.; Hrkach, J.; Ben-Jebria, A.; Eskew, M. L.; Mintzes, J.; Deaver, D.; Lotan, N.; Langer, R. Science 1997, 276, 1868-1871. (16) Rogueda, P. Expert Opin. Drug DeliVery 2005, 2, 625-638. (17) Napper, D. H. Polymeric Stabilization of Colloidal Dispersions; Academic Press: Orlando, 1983. (18) Tadros, T. F. In Colloid Stability: The Role of Surface Forces, Part I; Tadros, T. F., Ed.; Wiley-VCH Verlag GmbH & Co.: Weinheim, Germany, 2007; Vol. 1, pp 1-22. (19) Patel, N.; Marlow, M.; Lawrence, M. J. J. Colloid Interface Sci. 2003, 258, 354-362. (20) Patel, N.; Marlow, M.; Lawrence, M. J. J. Colloid Interface Sci. 2003, 258, 345-353. (21) Morokuma, K.; Kitaura, K. In Molecular Interactions; Ratajczak, H., Orville-Thomas, W. J., Eds.; Wiley: New York, 1980; Vol. 1, pp 2187. (22) Gutowski, M.; Van Duijneveldt-Van, de Rijdt, J. G. C. M.; Van Lenthe, J. H.; Van Duijneveldt, F. B. J. Chem. Phys. 1993, 98, 47284737. (23) Frisch, M. J. T.; G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Montgomery, J. A., Jr.; Vreven, T.; Kudin, K. N.; Burant, J. C.; Millam, J. M.; Iyengar, S. S.; Tomasi, J.; Barone, V.; Mennucci, B.; Cossi, M.; Scalmani, G.; Rega, N.; Petersson, G. A.; Nakatsuji, H.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Klene, M.; Li, X.; Knox, J. E.; Hratchian, H. P.; Cross, J. B.; Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.; Cammi, R.;

J. Phys. Chem. B, Vol. 111, No. 28, 2007 8103 Pomelli, C.; Ochterski, J. W.; Ayala, P. Y.; Morokuma, K.; Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Zakrzewski, V. G.; Dapprich, S.; Daniels, A. D.; Strain, M. C.; Farkas, O.; Malick, D. K.; Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.; Ortiz, J. V.; Cui, Q.; Baboul, A. G.; Clifford, S.; Cioslowski, J.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.; Peng, C. Y.; Nanayakkara, A.; Challacombe, M.; Gill, P. M. W.; Johnson, B.; Chen, W.; Wong, M. W.; Gonzalez, C.; and Pople, J. A. Gaussian 03, Revision A.1 ed.; Gaussian, Inc.: Pittsburgh, 2003. (24) Dunning, T. H., Jr. J. Chem. Phys. 1989, 90, 1007-1023. (25) Foresman, J. B.; Frisch, A. Exploring Chemistry with Electronic Structure Methods: A Guide to Using Gaussian, 2nd ed.; Gaussian: Pittsburgh, 1998. (26) Baradie, B.; Shoichet, M. S.; Shen, Z.; McHugh, M. A.; Hong, L.; Wang, Y.; Johnson, J. K.; Beckman, E. J.; Enick, R. M. Macromolecules 2004, 37, 7799-7807. (27) Kilic, S.; Michalik, S.; Wang, Y.; Johnson, J. K.; Enick, R. M.; Beckman, E. J. Ind. Eng. Chem. Res. 2003, 42, 6415-6424. (28) Feller, D.; Jordan, K. D. J. Phys. Chem. A 2000, 104, 9971-9975. (29) Hyla-Kryspin, I.; Haufe, G.; Grimme, S. Chem. Eur. J. 2004, 10, 3411-3422. (30) Tsuzuki, S.; Uchimaru, T.; Mikami, M.; Urata, S. J. Chem. Phys. 2004, 121, 9917-9924. (31) Tsuzuki, S.; Uchimaru, T.; Mikami, M.; Urata, S. J. Chem. Phys. 2002, 116, 3309-3315. (32) Hutter, J. L.; Bechhoefer, J. ReV. Sci. Instrum. 1993, 64, 18681873. (33) Burnham, N. A.; Chen, X.; Hodges, C. S.; Matei, G. A.; Thoreson, E. J.; Roberts, C. J.; Davies, M. C.; Tendler, S. J. B. Nanotechnology 2003, 14, 1-6. (34) Fadeev, A. Y.; McCarthy, T. J. Langmuir 1999, 15, 37593766. (35) Fadeev, A. Y.; M., T. J. Langmuir 2000, 16, 7268-7274. (36) Ito, T.; Namba, M.; Buhlmann, P. Langmuir 1997, 13, 4323-4332. (37) Jung, G.-Y.; Li, Z.; Wu, W.; Chen, Y.; Olynick, D. L.; Wang, S.Y.; Tong, W. M.; Williams, R. S. Langmuir 2005, 21, 1158-1161. (38) KSV “CAM 200-Optical Contact Angle Meter”; 2001. (39) Noy, A.; Frisbie, C. D.; Rozsnyai, L. F.; Wrighton, M. S.; Lieber, C. M. J. Am. Chem. Soc. 1995, 117, 7943-7951. (40) Noy, A.; Vezenov, D. V.; Lieber, C. M. Annu. ReV. Mater. Sci. 1997, 27, 381-421. (41) Takano, H.; Kenseth, J. R.; Wong, S.-S.; O’Brien, J. C.; Porter, M. D. Chem. ReV. 1999, 99, 2845-2890. (42) Steytler, D. C.; Thorpe, M.; Eastoe, J.; Dupont, A.; Heenan, R. K. Langmuir 2003, 19, 8715-8720. (43) Selvam, P.; Peguin, R. P. S.; Chokshi, U.; da Rocha, S. R. P. Langmuir 2006, 22, 8675-8683. (44) Peguin, R. P. S.; Selvam, P.; da Rocha, S. R. P. Langmuir 2006, 22, 8826-8830. (45) Williams, R. O., III; Liu, J. Eur. J. Pharm. Sci. 1999, 7, 137-144. (46) Liao, Y.-H.; Brown, M. B.; Jones, S. A.; Nazir, T.; Martin, G. P. Int. J. Pharm. 2005, 304, 29-39. (47) Eastoe, J.; Steytler, D. C.; Robinson, B. H.; Heenan, R. K. J. Chem. Soc., Faraday Trans. 1994, 90, 3121-3127. (48) Kilic, S.; Michalik, S.; Wang, Y.; Johnson, J. K.; Enick, R. M.; Beckman, E. J. Macromolecules 2007, 40, 1332-1341. (49) Blondino, F. E. NoVel Solution Aerosols for Inhalation, Virginia Commonwealth University: 1995. (50) Cece, A.; Jureller, S. H.; Kerschner, J. L.; Moschner, K. F. J. Phys. Chem. 1996, 100, 7435-7439. (51) Diep, P.; Jordan, K. D.; Johnson, J. K.; Beckman, E. J. J. Phys. Chem. A 1998, 102, 2231-2236. (52) Raveendran, P.; Wallen, S. L. J. Phys. Chem. B 2003, 107, 14731477. (53) Ridder, K. B.; Davies-Cutting, C. J.; Kellaway, I. W. Int. J. Pharm. 2005, 295, 57-65. (54) Thalladi, V. R.; Weiss, H.-C.; Blaeser, D.; Boese, R.; Nangia, A.; Desiraju, G. R. J. Am. Chem. Soc. 1998, 120, 8702-8710. (55) DerHovanessian, A.; Doyon, J. B.; Jain, A.; Rablen, P. R.; Sapse, A.-M. Org. Lett. 1999, 1, 1359-1362. (56) Tsuzuki, S.; Uchimaru, T.; Mikami, M.; Urata, S. J. Phys. Chem. A 2003, 107, 7962-7968. (57) Alonso, J. L.; Antolinez, S.; Blanco, S.; Lesarri, A.; Lopez, J. C.; Caminati, W. J. Am. Chem. Soc. 2004, 126, 3244-3249. (58) Tsuzuki, S.; Uchimaru, T.; Mikami, M.; Tanabe, K. J. Phys. Chem. A 1998, 102, 2091-2094. (59) Jalkanen, J.-P.; Mahlanen, R.; Pakkanen, T. A.; Rowley, R. L. J. Chem. Phys. 2002, 116, 1303-1312. (60) Tsuzuki, S.; Honda, K.; Uchimaru, T.; Mikami, M. J. Phys. Chem. A 2004, 108, 10311-10316. (61) Headrick, J. E.; Berrie, C. L. Langmuir 2004, 20, 4124-4131.

8104 J. Phys. Chem. B, Vol. 111, No. 28, 2007 (62) Ulman, A. An Introduction to Ultrathin Organic Films: From Langmuir-Blodgett to Self-Assembly; Academic Press: San Diego, 1997. (63) Nakagawa, T.; Ogawa, K.; Kurumizawa, T.; Ozaki, S. Jpn. J. Appl. Phys., Part 2 1993, 32, L294-296. (64) Nakagawa, T.; Ogawa, K.; Kurumizawa, T. J. Vac. Sci. Technol., B 1994, 12, 2215-2218. (65) Rogueda, P. G. A. Drug DeV. Ind. Pharm. 2003, 29, 39.

Wu et al. (66) Barsoum, M. W. Fundamentals of Ceramics; McGraw-Hill: New York, 1997. (67) Kim, H. I.; Koini, T.; Lee, T. R.; Perry, S. S. Langmuir 1997, 26, 7192-7196. (68) Maoz, R.; Sagiv, J.; Degenhardt, D.; Moehwald, H.; Quint, P. Supramol. Sci. 1995, 2, 9-24.