Among the large number of methodologies for combining organic and inorganic building blocks in one material, the sol-gel process represents one of the preferred ways thanks to its mild operative conditions.3 The classical sol-gel process consists of a two-step hydrolysis-condensation reaction starting with metal alkoxides M(OR)4, typically tetraethoxysilane, according to the scheme shown in Table 1.

In the present work, the organic phase consisted of commercial perfluoropolyethers (PFPE),4,5 which show the typical unusual properties of fluoro-products.6 In particular, PFPEs are characterized by very low glass transition temperatures (-120 °C), chemical inertia, solvent and high temperature resistance, barrier properties, low coefficient of friction, hydrophobicity (water repellency), lipophobicity and, in particular, low surface energy. In fact, some of the most interesting features of PFPEs derive from their surface properties, which distinguish them from all other polymers, with the exception of polysiloxane.7

Depending on the molecular weight, the surface tension of linear PFPE varies in the range of 14-25 mN·m-1, which is at least 10-15 mN·m-1 lower than other conventional polymers such as poly(vinyl chloride), polystyrene, poly(methyl methacrylate), polyurethane and poly(ethylene terephthalate). Furthermore, PFPEs are extremely non-polar substances, and their very low solubility parameters lead to a marked thermodynamic incompatibility with most of the common organic polymers and solvents. For this reason, pure PFPE cannot be directly used in the sol-gel process because of the impossibility of finding a common solvent for the fluorinated oligomer and the other reactants (tetraethoxysilane and water).

In order to overcome this drawback we modified the PFPE oligomer by polycaprolactone chain-extension leading to an organic oligomer soluble in solvents commonly used in sol-gel processes. In previous papers8,9 we demonstrated that the fluorinated segment of poly(caprolactone-b-perfluoropolyether-b-caprolactone) (PCL-PFPE-PCL) triblock copolymers tends to migrate to the outer surface due to the strong thermodynamic driving force to minimize the surface energy. The same behavior was also noted in the case of polymeric blends,10,11 and thus a similar mechanism of surface segregation can be expected in the case of hybrid systems containing PFPE segments.

These organic-inorganic hybrid materials can be potentially used as functional coatings to prepare water-repellent glasses or, in general terms, to modify the surface characteristics of several substrates. The present study is concerned with the preparation and the characterization of PCL-PFPE-PCL block copolymers having a,w-triethoxysilane end-groups, which are more reactive than hydroxyl-terminated block copolymers for the subsequent sol-gel reaction.

Hybrid coatings were prepared by varying the organic-inorganic and the PFPE/PCL ratios in the copolymer. The evaluation of surface wetting behavior was carried out by contact angle measurements. Independent information on surface chemical composition was obtained by XPS spectroscopy. Non-fluorinated and fully inorganic reference materials were prepared and characterized for comparison.

Experimental

H-(OC2H4)n-OCH2CF2O-(C2F4O)p(CF2O)q-CF2CH2O-(C2H4O)n-H

where the constituent units -C2F4O- and -CF2O- are randomly distributed along the macromolecular chains with p/q ratio of 0.9 and the average value of n is 1.5. The molar mass of this fluorinated macromer is about 2150 g·mol-1. For the purpose of identification this will be referred to as TX.

2. Polycaprolactone diol (Aldrich, average molecular weight 2000 g/mol); tetraethoxysilane (TEOS), Aldrich; 3-isocyanatopropyltriethoxysilane (ICTES), Fluka; hydrochloric acid at 37% concentration, Carlo Erba; ethanol (EtOH), Carlo Erba; tetrahydrofuran (THF), Carlo Erba; and tin(II) octoate (SnOct2), Aldrich were used, as received, without further purification.

3. e-Caprolactone (CL), Aldrich, was dried over calcium hydride and distilled before use.

Preparation of a,w-hydroxy-terminated
PCL-PFPE-PCL block copolymers
a,w-Hydroxy-terminated poly(caprolactone-b-perfluoropolyether-b-caprolactone) block copolymers with the following structure:
H[O(CH2)5-CO]x-(OC2H4)nOCH2CF2O(C2F4O)p(CF2O)qCF2CH2O(C2H4O)n[CO-(CH2)5O]xH
were prepared by ring-opening polymerization of CL using TX as transfer agent and in the presence of tin octoate as catalyst according to a procedure previously reported.8,12 The final products, coded as TXCL(x) (in which x = 2, 5, 10 represents the number-average degree of polymerization of the polycaprolactone segments), were dried overnight at room temperature under reduced pressure before use. The expected structures were confirmed by 1H-NMR analysis.

Preparation of a,w-triethoxysilane-terminated PCL-PFPE-PCL block copolymers
Triethoxysilane end-capped fluorinated copolymers were prepared by bulk reaction of TXCL(x) with ICTES (molar ratio of 1:2). The reaction was carried out in a 50-ml glass flask equipped with a CaCl2 trap and under magnetic stirring, at 120 °C for 3 h.

The final products, coded as TXCL(x)Si, were dried overnight at room temperature under reduced pressure before use. Polycaprolactone diol (Mn = 2000 g/mol) was directly reacted with ICTES under the same experimental conditions. The correspondent product was coded as PCLSi.

Preparation of PCL-PFPE-PCL/silica hybrids

A typical preparation was as follows: 8 ml of THF were added to 2.0 g of a TXCL(x)Si/TEOS mixture in a screw-thread glass vial and mixed until a homogeneous solution was obtained. Then EtOH, water and HCl (37%wt solution) were added under vigorous stirring at room temperature for about 10 min. The closed vial was placed in an air-circulating oven at 70 °C for 3 h in order to promote the sol-gel reaction. Then the sol-gel solution was deposited onto microscope glass slides (previously washed with EtOH) by manual dip- or spin-coating to obtain a typical coating thickness of 1-3 mm and 0.3-0.4 mm, respectively. Samples, after a period of 30 minutes at room temperature, were subjected to a thermal post-treatment at 100 °C for 2 h. Spin-coating was carried out by using a Laurell WS-400B-6NPP-LITE manual dispenser spin-coater operating at 3000 rpm for 40 s.

The final hybrids were coded as TXCL(x)Si/SiO2 y:z wt/wt, in which y:z represents the nominal final weight ratio of organic and inorganic components assuming the completion of the following sol-gel reactions:

ºSi-OEt + H2O k ºSi-OH + EtOH
ºSi-OH + EtO-Siº k ºSi-O-Siº + EtOH
ºSi-OH + HO-Siº k ºSi-O-Siº + H2O

The final organic-inorganic weight ratio was ranged from 5:95 to 80:20, as reported in Table 2. The same experimental conditions were used to prepare reference materials such as non-fluorinated organic inorganic hybrid (coded as PCLSi/SiO2) and pure silica (coded as SiO2).

Characterization

1H-NMR analysis was performed with a Bruker DPX200 spectrometer by using CDCl3 as solvent and tetramethylsilane as internal reference. 1H-NMR signal assignment for both reactants TXCL(x) and ICTES are reported in Table 3.

Differential Scanning Calorimetry (DSC) was performed with a TA DSC2010 instrument in the range -150 to +150 °C with a heating rate of 20 °C·min-1.

Contact angle measurements were carried out by using a DataPhysics OCA20 apparatus. In order to avoid any surface contamination, all specimens were washed with tetrahydrofuran and accurately dried just before measurement.

Static contact angle determination was carried out on two or three different specimens of the same sample, and an average value of contact angle was determined on the basis of at least 10 measurements.

Surface tension values were determined by applying the Owens and Wendt method13 on static contact angles respectively, evaluated with water, methylene iodide, dimethyl formamide and n-hexadecane as probe liquids.

XPS measurements were carried out on the vacuum-side surface using a V.G. ADES 400 hemispherical analyzer and a single-channel detector system. The Mg Ka1.2 line from a V.G. XR3 dual anode x-ray tube was used as the source, operated at 210 W (14 kV and 15 mA). High-resolution spectra of fluorine 1s (F1s), silicon 2p (Si2p), carbon 1s (C1s) and oxygen 1s (O1s) core level for each sample were acquired. Data were collected at 20° takeoff angle between the sample and the analyzer corresponding to a sampling depth of approximately 10-20 Å. Binding energies were referenced to the C-H level at 285.0 eV.

Results and Discussion

Typical FT-IR spectra of reactants (TXCL(x) and ICTES) and the obtained crude product (TXCL(x)Si) are shown in Figure 1. Spectrum of the triethoxysilane end-capped product (curve c) showed the almost complete disappearance of the peak at 2270 cm-1, indicating the complete reaction of the isocyanate groups of ICTES (curve a).

The expected structure was also confirmed by comparing the 1H-NMR spectra of the final product TXCL(x)Si and of the reactants a,w-hydroxy-terminated TXCL(x) and ICTES. The peak assignment is reported in Table 3, and a detailed discussion on the molecular characterization of these products is reported in a previous paper.14 1H-NMR spectra of the final crude products (not reported here) typically showed the disappearance of the signal at 3.65 ppm, related to the methylene groups adjacent to hydroxyl end-groups of TXCL(x) (1-f in Table 3), indicating that the reaction between terminal groups of TXCL(x) and isocyanate groups has gone to completion. The signal of the above-mentioned methylene groups after reaction is probably covered by other signals.

Finally, the ratio between signals related to internal PFPE block (4.25 ppm, 1-b in Table 3) and to propylsilane groups (0.70 or 3.30 ppm, 2-c and 2-e in Table 3, respectively) is approximately 1:1, within the experimental error. This points to the presence of two triethoxysilane terminal groups for each TXCL(x) chain, according to the following reaction scheme:

H[O(CH2)5-CO]x-(OC2H4)1.5OCH2CF2O(C2F4O)p(CF2O)qCF2CH2O(C2H4O)1.5[CO-(CH2)5O]xH +
+ 2 (EtO)3Si(CH2)3NCO Æ (EtO)3Si(CH2)3-NH-CO-[O(CH2)5-CO]x-

-(OC2H4)1.5OCH2CF2O(C2F4O)p(CF2O)qCF2CH2O(C2H4O)1.5[CO-(CH2)5O]xCO-NH-(CH2)3-Si(OEt)3

DSC analysis carried out on the organic oligomers TXCL(2)Si, TXCL(5)Si and TXCL(10)Si showed evident melting transitions attributable to the PCL segments (Tm values ranging 34.3, 35.5 and 41.0 °C, respectively). Contrary to TXCL(x)Si samples, all the prepared TXCL(x)Si/SiO2 hybrids were fully transparent and DSC analysis showed no melting transitions. Both results represent an indirect indication for the formation of a nanocomposites structure with a morphology in which organic and silica phases have dimensions less than the wavelength of visible light and that PCL chains bonded to silica domains are not able to organize themselves into crystalline domains.

Surface Characterization

Aimed at studying the effect of these parameters on the surface properties (in particular wettability) of the coatings, contact angle measurements were carried out. In order to adequately describe the hydrophobicity and oleophobicity of the prepared coatings, contact angles with water and n-hexadecane were measured and reported in Table 4 for TXCL(x)Si/SiO2 and PCLSi/SiO2 hybrids.

In the case of dip-coated samples, it is worth noting that strong hydrophobic and oleophobic character was shown by all the coatings containing PFPE segments. Water and n-hexadecane contact angles always ranged between 94° and 109° and between 60° and 68°, respectively, indicating a very low wettability with respect to both probe liquids; furthermore, the low values of standard deviation can be considered as evidence of high surface homogeneity and smoothness of the coatings.15

In the case of spin-coated samples, much lower values were obtained, especially for water contact angle. It can be surmised that the very fast solvent elimination during the spin-coating process (with respect to the dip-coating) can hinder the surface segregation of the fluorinated segments onto the air-coating interface leading to lower values of contact angles.

More surprisingly, the organic-inorganic ratio was a parameter with only a little effect on the wettability of the coatings based on TXCL(2)Si, TXCL(5)Si oligomers prepared by dip-coating. Both water and n-hexadecane contact angles are almost constants for coatings based on the same type of fluorinated oligomer for all the compositions investigated. In other words, even coatings with a relatively low content of organic fluorinated phase are able to lead to surfaces with similar hydrophobic and oleophobic character of organic-rich hybrids.

Data indicated that the driving force for the surface segregation of the PFPE segment, expected on the basis of the low surface tension of PFPE, was almost independent on the bulk concentration of organic and inorganic components of the coating. This evidence suggests that the driving force was attributable to segment mobility rather than translational mobility. A more marked effect of the organic-inorganic ratio was noted in the case of TXCL(10)Si-based hybrids, probably due to the lower PFPE/PCL ratio of the organic oligomer, which in turn may reduce its tendency to surface migration.

The only parameter that significantly influences the surface properties of the coating is the type of fluorinated oligomer used for the coating preparation. In fact, water and n-hexadecane contact angles showed a progressive decrease when measured on hybrids belonging to the series based on TXCL(2)Si, TXCL(5)Si or TXCL(10)Si, respectively.

Data seem to indicate that the driving force for PFPE surface segregation can be finely tuned by accurate balancing of the ratio between the fluorinated PFPE and the hydrogenated polycaprolactone segments constituting the organic oligomer TXCL(x)Si. In particular, the lowest wettability was expected for the highest PFPE/PCL and thus, at least in principle, in the case of TXCL(0)Si, i.e., simply a,w-triethoxysilane-terminated PFPE homopolymer. Unfortunately, this type of oligomer was immiscible in THF and it was not possible to find a common solvent or mixture of solvents for both TXCL(0)Si and TEOS.

Very low values of surface tension were measured in all cases, with values ranging from about 14 to 21 mN·m-1. It is interesting to note that the reported surface tension value of Fluorolink® E is 23 mN·m-1, while for unfunctionalized PFPE (i.e. fully fluorinated even in the terminal groups) with comparable molecular weight (about 2000 g/mol)it is about 15 mN·m-1.4

As expected, the fundamental contribution of PFPE segments on these very low surface tension values was clearly evidenced by comparison with the results obtained in the case of the non-fluorinated and completely inorganic reference materials (PCLSi/SiO2 50:50 and SiO2) which showed much higher values of surface tension g (41.9 and 42.5 mN·m-1, respectively) together with an important contribution of the polar component g p (18.4 and 19.9 mN·m-1, respectively).

Data indicated that the very surface of the investigated TXCL(x)Si/SiO2 hybrids was almost completely perfluorinated (i.e., fully covered by PFPE segments) and that the other more polar groups such as ethylene oxide and caprolactone units are buried under the very surface. Furthermore, the covalent bonding of these terminal segments with rigid silica domains inhibits their mobility thus allowing a surface reorganization, which could be responsible for the kinetic hysteresis phenomenon and could lead to a more hydrophilic surface similar to the case of poly(2-hydroxyethylmethacrylate).16

A further analysis of the data reported in Table 5 for TXCL(2)Si and TXCL(5)Si-based coatings indicated that g values were almost independent from the organic-inorganic ratio, consistent with the previously discussed results on wettability (water and n-hexadecane contact angles). The trend of g p clearly indicates the effect of a longer PCL segment in the organic phase, even if the total surface free energy results are only lightly influenced by that.

An independent evaluation of surface composition was obtained by XPS analysis. Data related to the fluorine, silicon, carbon and oxygen atomic surface composition of some TXCL(x)Si/SiO2 coatings, prepared by dip-coating, are reported in Table 6.

Conclusions

Better results in terms of non-wettability were obtained by decreasing the length of the hydrogenated polycaprolactone segment of the a,w-triethoxysilane-terminated PCL-PFPE-PCL copolymers used for the preparation of the hybrid material. The measured surface tension values were always of the same order of magnitude with respect to those typical of unfunctionalized fully perfluorinated PFPE indicating that the very surface was composed almost completely of perfluoropolyether segments. Similar conclusions were also supported by XPS analysis.

References

This paper was presented at the Nano and Hybrid Coatings Conference sponsored by The Paint Research Association, January 2005, Manchester, UK. Conference proceedings can be obtained by contacting Janet Saraty, conference administrator, at j.saraty@pra.org.uk.