13496
Langmuir 2008, 24, 13496-13501
Why does Silane Enhance the Protective Properties of Epoxy Films? Peng Wang and Dale W. Schaefer* Department of Chemical and Materials Engineering, UniVersity of Cincinnati, Cincinnati, Ohio 45221-0012 ReceiVed August 27, 2008. ReVised Manuscript ReceiVed September 26, 2008 Using neutron reflectivity, the protection mechanisms of a novel one-step epoxy-silane coating system were investigated in terms of coating structure and water response behavior. By comparing pure epoxy and epoxy-silane mixtures in various aqueous environments, the effects of the addition of silane were determined. Specifically, a bridged bis-silane coupling agent with six alkoxy moieties and a polysulfur bridge was investigated. The key mechanisms of silane-enhanced protection are (1) the silane is enriched at the substrate-coating interface, forming a hydrophobic dense interfacial layer and good adhesion to the substrate, and (2) the silane serves as a cross-linker, resulting in a denser and less hydrophilic bulk film compared to the neat epoxy. The hydrophobic nature of bis-sulfur silane also increases the overall hydrophobicity of the mixed film.
1. Introduction Although there is an enormous body of applied coating research, few papers address what the components of coating formulation actually do. The addition of certain silanes to epoxy-based coatings, for example, dramatically improves corrosion protection properties. 1 This paper addresses the mechanisms that underlie the observed performance enhancement for silane-laced novolac epoxy resins. Bis-type silanes improve the corrosion performance of novolac epoxy resins compared to neat epoxy resin or neat silane, especially on aluminum alloys. Moreover, this epoxy-silane coating system is also a direct-to-metal, one-step coating system, which eliminates the need for a priming step. Novolac epoxy resins are produced by acid-catalyzed condensation of a phenolic precursor with formaldehyde, followed by glycidation with epichlorohydrin. The relatively high epoxide functionality leads to an increased cross-link density, which results in better chemical resistance and thermal stability compared to that achieved with standard bisphenol A or bisphenol F resins in parallel systems. However, protection failures caused by the poor adhesion to the metal substrates have also been reported. The epoxy coating tends to be vulnerable to delamination. In addition, epoxy resins are not good water barriers. The water uptake can be as high as 8 wt % at elevated temperatures.2,3 Silanes as corrosion inhibitors are not as well understood as silanes as adhesion promoters. Bis-silanes with the general formula of (RO)3Si(CH2)3-R′-(CH2)3Si(OR)3, where OR represents an alkoxy group and R′ is an organic functionality, show better performance than monosilanes because bis-silanes have a higher number of hydrolyzable groups per molecule, which ultimately leads to denser films.4-8 Also, the bridging functionality * Corresponding author. E-mail:
[email protected]. (1) van Ooij, W. J.; Seth, A.; Maguda, T.; Pan, G.; Schaefer, D. W. A novel self-priming coating for corrosion protection. Presented at the International Surface Engineering Congress and Exposition, Orlando, FL, August 2-4, 2004. (2) Ji, W. G.; Hu, J. M.; Zhang, J. Q.; Cao, C. N. Corros. Sci. 2006, 48(11), 3731–3739. (3) Xiao, G. Z.; Shanahan, M. E. R. J. Polym. Sci., Part B: Polym. Phys. 1997, 35(16), 2659–2670. (4) Plueddemann, E. P. Silane Coupling Agents, 2nd ed.; Plenum Press: New York, 1991. (5) Puomi, P.; Fagerholm, H. M. J. Adhesion Sci. Technol. 2001, 15(5), 509– 533. (6) Susac, D.; Sun, X.; Mitchell, K. A. R. Appl. Surf. Sci. 2003, 207(1-4), 40–50. (7) van Ooij, W. J.; Child, T. Chem. InnoVation 1998, 28(2), 26–35.
group (R′) can be altered to optimize the substrate wettability, adhesion strength, as well as bulk film hydrophobicity.9 According to performance tests by van Ooij et al., bis-silane films are comparable to chromate conversion coating in terms of protection of and adhesion to metal substrates.7,8,10 On the other hand, previous studies showed that neat silane films are porous and are not strict water barriers. More than 30 vol % water can be trapped in neat silane films in the molecular level free space.11-16 On the basis of the properties of neat epoxy and silane, it is clear that the performance improvement does not come from the simple mixing of two components, but from the alteration of the chemistry and physics of the film. Numerous questions arise. What is the structure of this epoxy-silane coating film? What happens at the coating-substrate interface? How is the silane minority phase distributed in the epoxy matrix? How and to what extent are the barrier properties affected by the addition of silane? By answering these questions, the underlying protection mechanisms can be illuminated. In order to reveal the mechanism of corrosion protection, the epoxy-silane coating system was simplified to the main ingredients: novolac epoxy resin, silane, and curing agent in the ratio of the original recipe formulated by van Ooij et al.1 Pure epoxy and epoxy-silane films were made and investigated by neutron reflectivity (NR). Pure epoxy refers to the simplified recipe without silane. NR is an ideal probe because NR can easily examine interfaces that are buried well within a sample. NR can elucidate film (8) van Ooij, W. J.; Zhu, D. Corrosion 2001, 57(5), 413–427. (9) Pan, G.; Yim, H.; Kent, M. S.; Majewski, J.; Schaefer, D. W. Effect of bridging group on the structure of bis-silane water-barrier films. In Silanes and Other Coupling Agents; Mittal, K. L., Ed.; VSP/Brill: Utrecht, The Netherlands, 2004; Vol. 3, pp 39-50. (10) van Ooij, W. J.; Zhu, D.; Prasad, G.; Jayaseelan, S.; Fu, Y.; Teredesai, N. Surf. Eng. 2000, 16(5), 386–396. (11) Pan, G.; Schaefer, D. W. Are silane films water barriers? In Silanes and Other Coupling Agents; Mittal, K. L., Ed.; VSP/Brill: Leiden, The Netherlands, 2007; Vol. 4, pp 3-16. (12) Pan, G.; Schaefer, D. W.; Ilavsky, J. J. Colloid Interface Sci. 2006, 302(1), 287–293. (13) Pan, G.; Schaefer, D. W.; van Ooij, W. J.; Kent, M. S.; Majewski, J.; Yim, H. Thin Solid Films 2006, 515, 2771–2780. (14) Pan, G.; Watkins, E.; Majewski, J.; Schaefer, D. W. J. Phys. Chem. C 2007, 111, 15325–15330. (15) Wang, Y.; Wang, P.; Kohls, D.; Hamilton, W. A.; Schaefer, D. W. Water absorption and transport in bis-amino silane films. In Silanes and Other Coupling Agents; Mittal, K. L., Ed.; VSP/Brill: Leiden, The Netherlands, 2008; Vol. 5. (16) Wang, Y.; Watkins, E.; Ilavsky, J.; Metroke, T. L.; Wang, P.; Lee, B.; Schaefer, D. W. J. Phys. Chem., B 2007, 111, 7041–7051.
10.1021/la8028066 CCC: $40.75 2008 American Chemical Society Published on Web 11/08/2008
Enhanced Protection of Epoxy-Silane Films
Figure 1. Monomer of EPON SU-8 resin.
Figure 2. Molecular structure of bis[3-(triethoxysilyl)propyl]tetrasulfide (bis-sulfur silane).
composition normal to the substrate surface on scales from 10 Å to 2000 Å. Contrast over the thickness of the film arises from differences in the neutron scattering length density (SLD), which is determined by the material’s density and chemical composition. The SLD represents the scattering power of a substance and can be calculated if its chemical composition and density are known. The structure of the film normal to the substrate surface is captured by the SLD profile (SLD vs distance from the substrate surface) as determined by fitting the measured specular NR data using standard inversion techniques. Neutron SLD is sensitive to different isotopes. By replacing normal H2O with D2O, water transport kinetics, equilibrium water content and water distribution in the film can be determined.
2. Experimental 2.1. Materials. EPON resin SU-8 is a polymeric solid epoxy novolac resin with average epoxide group functionality around eight (Resolution Performance Products, Houston, TX). The molecular structure is shown in Figure 1. SU-8 contains reactive epoxide functionalities and is intended for high performance applications that require maximum chemical and solvent resistance and/or elevated temperature service. Utilizing a commercial modified polyamine adduct curing agent, EPIKURE 6870-W-53, acquired from Resolution Performance Products, the epoxy-cure system forms a clear, highly cross-linked, tough, chemically resistant film. Advantages associated with SU-8, compared to a standard bisphenol A epoxy resin, include improved elevated temperature performance, increased water and solvent resistance, and improved chemical resistance. Bis[3-(triethoxysilyl)propyl]tetrasulfide (bis-sulfur silane) was provided by OSi Specialties (Tarrytown, NY, and now named Momentive Performance Materials). The molecular structure is shown in Figure 2. The silane was used without further purification. The silicon slabs used as substrates were one-side-polished 3-in. diameter single crystal (111) wafers with a thickness of 5 mm obtained from Wafer World, Inc. (West Palm Beach, FL.). Sulfuric acid, hydrogen peroxide, toluene and tetrahydrofuran (THF) were obtained from Sigma-Aldrich and used as received. 2.2. Procedures. Films were deposited on silicon slabs by spin coating. The silicon slabs as substrate were cleaned by immersion in a freshly prepared “Piranha” solution at room temperature for 30 min. Piranha solution is a mixture of H2SO4 and 30% H2O2 at a volume ratio of 7:3. After immersion, the substrates were rinsed repeatedly with deionized (DI) water. The silicon slabs were ready for spin coating after desiccation. The epoxy-silane mixture precursor solution is made by dissolving the epoxy resin, curing agent, and bis-sulfur silane at a weight ratio of 7:2:1 in a mixture of THF and toluene (volume ratio of THF to toluene is 7:3, which produce the smoothest films). For the precursor solution of pure epoxy samples, the only difference is the absence
Langmuir, Vol. 24, No. 23, 2008 13497 of bis-sulfur silane in the recipe (the epoxy resin, curing agent, and bis-sulfur silane at a weight ratio of 7:2:0). The bis-sulfur silane was prehydrolyzed following the procedures of Pan et al.13 before mixing with epoxy and curing agent. The thickness of the coated film can be controlled by either the concentration of precursor solution or the spin-coating speed. Here, 1% (by weight) precursor solution yielded a film of thickness around 700 Å, with spinning speed of 2000 rpm. The coating procedure was carried out by using a Laurell singlewafer spin processor (WS-400A-6NPP-Lite, North Wales, PA). The precursor solution was pipetted onto the slabs to cover and wet the whole surface of the silicon slab for one minute before the spin coating at 2000 rpm for 30 s. The samples were then dried and cured in an oven at 150 °C for 1 h. Samples were kept in a desiccated environment before measurement. NR was performed on the Surface Profile Analysis Reflectometer (SPEAR) at the Los Alamos National Laboratory. The method of NR is thoroughly discussed in the literature;17-20 therefore, only a brief description is given here. The neutron reflectivity, R(q), defined as the intensity ratio between reflected and incident neutron beam, is measured as a function of the scattering vector, q ) (4π/λ) sin θ, where θ is the angle of incidence and λ is the neutron wavelength. For our measurements, q was varied by collecting intensity for a range of different wavelengths at a fixed angle of incidence. The incident wavelength distribution ranged from 1.4 to 16 Å. The reflectivity curves were obtained by merging data from two angles of incidence. The reflectivity curve is determined by the neutron SLD profile normal to the surface of the sample. The neutron SLD is a function of density and atomic composition, described as eq 1:
SLDneutron ) F
NA atoms b M i)1 i
∑
(1)
where bi is the coherent scattering length of the ith atom, F is density, M is molecular weight, and NA is Avogadro’s number. The curve fitting was done by using the free software “Parratt 32,” which is based on recursive Parratt formalism.17 To obtain agreement between the simulated and measured reflectivity, Parratt optimizes the parameters of a candidate real space SLD profile model by means of nonlinear regression. The simplest reasonable model is chosen if more than one real-space model fits the data. A sealed Al can was used as a sample holder for the in situ water conditioning study. The thin Al wall causes almost no flux loss. The reflectivities of the as-prepared and redried films were measured with desiccant inside the Al can. For the D2O vapor conditioning measurements, the desiccant was removed, and several drops of D2O were introduced in the notch at the bottom of the Al can. The coating films reached equilibrium saturation when no further change in the reflectivity with time was observed. The D2O liquid contact conditioning experiments were performed by mounting the sample with the coating side against the D2O and shooting from the silicon substrate side through the silicon slab. Because the absorbed water usually resides in molecular-level free space inside the film,15 the D2O-equivalent volume fraction, φD2O, can be calculated with the following equation assuming no swelling:
φD2O )
SLDsaturated - SLDas-prepared SLDD2O
(2)
where the subscripts “saturated” and “as-prepared” mean the state of equilibrium with D2O vapor or liquid environment and the virgin dry state, respectively. (17) Parratt, L. G. Phys. ReV. Lett. 1954, 95(2), 359–369. (18) Roe, R. J. Methods of X-ray and Neutron Scattering in Polymer Science; Oxford University Press: New York, 2000. (19) Russell, T. P. Annu. ReV. Mater. Sci. 1991, 21, 249–268. (20) Russell, T. P. Physica B 1996, 221, 267–283.
13498 Langmuir, Vol. 24, No. 23, 2008
Wang and Schaefer
Figure 3. (a) NR from a silicon wafer coated with the pure epoxy in the as-prepared state. The curve through the data points is the best fit corresponding to the SLD profile in (b). (b) Best-fit SLD.
Figure 4. (a) NR from a silicon wafer coated with the epoxy-silane mixture in the as-prepared state. The curve through the data points is the best fit corresponding to the SLD profile in (b). (b) Best-fit SLD profile.
In the presence of swelling or shrinking, eq 2 is modified to eq
For the epoxy-silane film, a layered model is required to get good fit to the reflectivity curve. As shown in Figure 4b, there is a bottom layer at the silicon dioxide interface with a thickness of (40 ( 5) Å plus a top layer accounting for the rest of thickness of the film (620 ( 5 Å). The mean neutron SLD of the bottom layer is (1.8 ( 0.05) × 10-6 Å-2, while the top layer is (1.60 ( 0.01) × 10-6 Å-2. The bottom layer has an interfacial width comparable to its thickness, which indicates the bottom layer is actually a transitional region gradually changing its composition to the top bulk layer. The mean SLD of the interfacial layer is greater than the SLD of condensed pure bis-sulfur film (0.67 × 10-6 Å-2) reported by Pan et al.13 and the measured value for the cured pure epoxy film (1.40 × 10-6 Å-2). If the compositions and densities of the epoxy and silane ingredients in the mixture retain their neat values in the cured state, the mixture’s SLD should be smaller than 1.4 × 10-6 Å-2. If all silane molecules occupied free space in the epoxy network, which is not expected considering the size of the silane molecule, the maximum possible SLD of the mixture is around 2.07 × 10-6 Å-2, which is still smaller than the film SLD at the film-substrate interface (2.2 ( 0.05) × 10-6 Å-2. Therefore, the elevated SLD must be caused by the formation of a denser film and/or by a composition change. The key to understanding the unique interfacial layer of epoxysilane coating lies in the nature of the silane. After hydrolysis, each silane molecule forms six silanol functional groups
3:
φD2O )
SLDsaturated × tsaturated - SLDas-prepared × tas-prepared (3) SLDD2O × tsaturated
where t is the thickness of the film.
3. Results and Discussion 3.1. Film Structure. The virgin structure of each sample is revealed by the SLD profile in the as-prepared state. The reflectivity curves from as-prepared pure epoxy and epoxy-silanecoated samples are shown in Figure 3a and Figure 4a, respectively. For the pure epoxy coated sample, a simple one-layer uniform model fits the experimental data well (Figure 3a). In the SLD profile (Figure 3b), the top surface of silicon slab was picked as the origin. On the left side of the origin is the substrate, including the single crystal silicon slab and a 15-20 Å thick surface silicon dioxide layer with an SLD of 3.475 × 10-6 Å-2. The single crystal silicon layer was treated as infinitely thick with an SLD of 2.073 × 10-6 Å-2. Roughness of substrate is set to less than 5 Å, on the assumption of a sharp interface between substrate and bulk film. To the right-hand of the origin, the epoxy forms a uniform film with a thickness of 845 ( 5 Å. The SLD of the epoxy layer is (1.40 ( 0.01) × 10-6 Å-2. The roughness of epoxy layer top surface is 25 Å.
Enhanced Protection of Epoxy-Silane Films
Langmuir, Vol. 24, No. 23, 2008 13499
Figure 5. Molecular structure of fully hydrolyzed bis-sulfur.
(-Si-OH, Figure 5).4 These groups are hydrophilic and ready to react with the -OH4,21-23 on the substrate surface:
-Si-OH + Sisubstrate-OH f Sisubstrate-O-Si- + H2O
(4)
When the precursor solution is applied on the substrate surface, the above reaction consumes the silane near the substrate-solution interface. The resulting concentration gradient drives silane diffusion to the substrate surface, leading to a silane rich interface layer. Assuming the process is Fickian,24 the silane diffusion distance, S, is proportional to the square root of the product of diffusion coefficient D and time t:
S ) √Dt
(5)
The diffusion coefficient can be calculated by eq 6:24
D ) kT ⁄ (6πηRH)
(6)
where k is the Boltzmann constant, T is temperature, and η is viscosity. The viscosities of THF and toluene at 20 °C are 5.5 ( 10-4 kg/m/s and 5.9 ( 10-4 kg/m/s, respectively. It is safe to take 6 ( 10-4 kg/m/s as the viscosity of the THF-toluene mixture. The fully stretched hydrolyzed silane backbone contains 15 covalent bonds. Assuming the length of each bond is 1.5 Å, the end-to-end length (L) of the silane molecule is around 20 Å. By simplifying the silane molecule to a rod, the hydrodynamic radius (RH) of the silane molecule can be calculated according to eq 7:25
RH3 )
3LR2 4
(7)
where R is the radius of the rod, which should be around 2 Å. On the basis of the above analysis, the silane molecules can diffuse a distance of 0.23 mm in 1 min, which is about 50% of the solution thickness above the substrate surface, ensuring that the silane molecules can arrive at the substrate surface before spinning. According to eq 4, silane molecules act as coupling agents improving the adhesion by providing a high density of functional groups capable of coupling the bulk film to the substrate via covalent bonds. In addition, the high functionality of the silane leads to full condensation of the silanols groups, leading to a dense SiO2-like structure at the film-substrate interface. At the same time, the functional groups on the other end of the silane molecules can either react with other silane molecules (eq 8)4,21-23 or with the hydroxy groups produced when epoxide cross-links (eq 9),4,21-23,26 which enhances the film-substrate adhesion, further increases the cross-link density, and densifies the interfacial region. According to eq 9, the reaction between silane and epoxy also decreases the concentration of H atoms, which will lead to an 11% SLD increase in the case of complete reaction (if density is not affected).
-Si-OH + -Si-OH f -Si-O-Si + H2O
(8)
(9)
Because the silane molecules are enriched near the interface, the effect of the SLD increase is more significant than the bulk top layer. Ashirgadey and Yin et al. demonstrated the existence of this silane-rich interfacial layer by scanning electron microscopy (SEM) with energy dispersive X-ray analysis (EDX) when scanning the artificial scratch in the epoxy-silane coating.27,28 Compared with the interfacial layer, the bulk film is more uniform. The bis-sulfur silane molecules serve as cross-linkers, condensing with each other and with hydroxy groups on the epoxy network. The SLD of the bulk film is 14% greater than, but still close to the pure epoxy film. The fact that the SLD of top layer is close to pure epoxy film indicates the main ingredient is epoxy. Since the elimination of H in the form of H2O during the cure reaction can increase the SLD by 11% at most, the 14% increase in SLD also implies a slightly denser film. The fact that epoxy-silane results in a thinner film from the same concentration precursor solution also implies a denser film than pure epoxy. The air-side surface roughness is 25 Å, which is the same as the pure epoxy-coated sample. 3.2. WaterBarrierProperties.3.2.1. D2OVaporConditioning. In situ D2O vapor conditioning experiments were carried out for both pure epoxy and epoxy-silane films. The reflectivity curves from the as-prepared film and the same film exposed to saturated D2O vapor at room temperature for various time intervals (0.25, 0.5, 1, 3 and 6 h) are shown in Figure 6a and Figure 7a for pure epoxy and epoxy-silane samples, respectively. For both pure epoxy and epoxy-silane coated samples, the reflectivity increases relative to the as-prepared film due to the absorption of D2O. The equilibrium with saturated D2O vapor is reached within 15 min. Water enters the film readily and penetrates the entire film in a short period. This result indicates that neither film is a strict water barrier. After 15 min, no further change in film SLD or thickness is observed up to 6 h of vapor conditioning. Thus, for clarity, only the profiles of the as-prepared state and the equilibrium state (after 6 h conditioning) are shown in Figures 6b and 7b. For pure epoxy (Figure 6), there is no significant film swelling. The SLD of the saturated bulk film increased to (1.9 ( 0.05) × 10-6 Å-2, implying the D2O volume fraction is 7.9%. A distinct water-rich top layer was observed at the air-side surface (55 Å), in which the D2O concentration is 14 vol %. Interestingly, at the substrate-epoxy interface, there is also a thin (10 Å) D2O-rich layer. Interfacial enrichment is essential to fit the experimental data. The SLD of this layer is (3.6 ( 0.1) × 10-6 Å-2, implying the D2O volume fraction is 35%. This observation indicates that both the air-side surface and the substrate-epoxy interface are hydrophilic. For the air side, the hydrophilic hydroxy groups attract D2O from aqueous environment. The defects of the epoxy network due to the presence of the surface provide more molecular-level free space to absorb more D2O than bulk film. For the substrate side, besides the hydroxy groups in epoxy network, hydroxy groups on the silicon substrate make the interface more hydrophilic. Also, the loosely bonded substrate-epoxy interface (21) Abel, M. L.; Digby, R. P.; Fletcher, I. W.; Watts, J. F. Surf. Interface Anal. 2000, 29(2), 115–125. (22) Abel, M. L.; Rattana, A.; Watts, J. F. J. Adhes. 2000, 73(2-3), 313–340. (23) Abel, M. L.; Watts, J. F.; Digby, R. P. J. Adhes. 2004, 80, 291–312. (24) Smith, W. F. Foundations of Materials Science and Engineering, 3rd ed.; McGraw-Hill: Boston, MA , 2004. (25) Strobl, G. R. The Physics of Polymers: Concepts for Understanding Their Structures and BehaVior; Springer: New York, 1996. (26) Pocius, A. V. Adhesion and AdhesiVes Technology, 2nd ed.; Hanser Gardner Publications: Cincinnati, OH, 2002. (27) Ashirgadey, A. Masters Thesis, University of Cincinnati, Cincinnati, OH, 2006. (28) Yin, Z. Ph.D Thesis, University of Cincinnati, Cincinnati, OH, 2009.
13500 Langmuir, Vol. 24, No. 23, 2008
Figure 6. (a) NR data from pure epoxy-coated sample at different time intervals when treated with saturated D2O vapor. The curves through the data points correspond to the best fits using model SLD profiles in (b). (b) Best-fit SLD profiles.
may result in extra free volume near the interface capable of harboring more D2O compared to the bulk film. For the epoxy-silane coated sample, at equilibrium, the SLD of bulk layer increases to (2.2 ( 0.05) × 10-6 Å-2, indicating the D2O volume fraction is 3%, which is less than half of the pure epoxy sample (7.9% in volume). The air-side surface D2Orich region is almost negligible compared to neat epoxy sample. No D2O-rich layer is observed at the substrate-coating interface. The dense interfacial layer absorbs only 1.6 vol % D2O. The hydrophobic nature of the epoxy-silane mixture coating could be due to several causes. At the substrate-coating interface, the enriched silane distribution induces the SiO2-like dense molecular layer. In the bulk film, the silane densifies the film, reducing the free space capable of harboring D2O molecules. The silane also decreases the hydrophilic sites (hydroxy groups) of the coating. Finally, bis-sulfur silane itself contains a hydrophobic bridging group formed by four sulfur atoms, which also contributes to the film hydrophobicity. 3.2.2. Liquid D2O Contact Conditioning and after Redrying State. 3.2.2.1. Liquid D2O Contact Conditioning. Because the SLD of liquid D2O is 6.36 × 10-6 Å-2, which is much higher than the coating, flip-mounting against D2O and shooting from the silicon side are required to get information associated with liquid-exposed films. From the liquid contact (∆) curves in Figure 8 and Figure 9, the critical edge, qc (where neutron beam begins to penetrate into the film) significantly shifts to higher q compared to the as-prepared state. Since this sample is shot from the backside through the silicon substrate, the critical edge is determined by the SLD of the liquid D2O layer instead of silicon. The SLD
Wang and Schaefer
Figure 7. (a) NR data from epoxy-silane mixture coated sample at different time intervals when treated with saturated D2O vapor. The curves through the data points correspond to the best fits using model SLD profiles in (b). (b) Best-fit SLD profiles.
profiles at equilibrium with liquid D2O imply similar behavior as found for vapor conditioning. D2O penetrates the film, and the SLD of the film increases. However, slight differences were also observed when treated by liquid D2O. For pure epoxy sample, saturated with D2O liquid, the SLD of bulk film increases to (2.1 ( 0.05) × 10-6 Å-2, resulting in a D2O volume fraction of 11%. The thickness of the total coating increases from 845 Å to 890 Å, a 5% swelling effect. Swelling generates more space capable of harboring D2O molecules, which induces a further SLD increase in the bulk epoxy film. More D2O was absorbed at the substrate-epoxy interface. The SLD of this layer is (4.0 ( 0.1) ( 10-6 Å-2, implying the D2O volume fraction is as high as 41%. The large amount of D2O suggests a porous and hydrophilic interface region, therefore, implying poor bonding of the coating. For the epoxy-silane-coated sample, at equilibrium, the SLD of bulk film increased to (1.89 ( 0.05) × 10-6 Å-2, corresponding to a D2O volume fraction of 4.2%, which is less than half of the pure epoxy sample (11% in volume). A 15 Å increase in total thickness was observed compare to the as-prepared state. This 1.5% swelling leads to greater D2O absorption compared to the vapor conditioned films. No D2O-rich layer was observed at the substrate-coating interface. The interface layer behaves the same as when treated with saturated D2O vapor. The SLD is (2.2 ( 0.05) × 10-6 Å-2, which implies a D2O volume fraction is only1.6 vol %. Thus, the interface remains as good water barrier even when directly contacted with liquid D2O.
Enhanced Protection of Epoxy-Silane Films
Figure 8. (a) NR data from pure epoxy-coated samples in the as-prepared state, at equilibrium with liquid D2O and the redried state. The curves through the data points correspond to the best fits using the model SLD profiles in (b). (b) Best-fit SLD profiles.
3.2.2.2. Redried State. Figure 8 and Figure 9 also give the NR data from redried samples. According to these SLD profiles, both pure epoxy and epoxy-silane samples fully recover physically and chemically. The SLDs returns to the same as the as-prepared state. No significant thickness change is observed.
4. Conclusions The pure epoxy film is hydrophilic and has a uniform structure. When treated with water (vapor or liquid D2O), water penetrates the entire neat epoxy film within 15 min. Considerable water is absorbed in bulk film. Excess water is observed at both air and substrate interfaces. The water excess at the substrate-epoxy interface indicates an open interfacial structure and poor bonding between the epoxy coating and the substrate. The film fully recovers to the as-prepared state after redrying, which indicates that the water was physically absorbed in the film. The epoxy-silane film is hydrophobic compared to pure epoxy film. A layered structure is required to fit the experimental data. Although water still penetrates the entire film within 15 min, the water fraction in the bulk film is less than half-that of the pure epoxy film. The water-rich layer on the air-side interface is negligible compared to the pure epoxy sample. Bis-sulfur silane is rich at the substrate-coating interface, which makes the interface dense and more hydrophobic compared to pure epoxy. Only 1.5 vol % water is absorbed in this region. The small amount
Langmuir, Vol. 24, No. 23, 2008 13501
Figure 9. (a) NR data from epoxy-silane-coated sample in the asprepared state, at equilibrium with liquid D2O and in the redried state. The curves through the data points correspond to the best fits using the model SLD profiles in (b). (b) Best-fit SLD.
of water absorption also implies good adhesion between the coating and the substrate. The water is physically absorbed since the film fully recovers the as-prepared state after redrying. Bis-sulfur silane has several effects on the epoxy-silanecoated film. Silane serves as a curing agent, increasing the crosslink density, densifying the film, and reducing the total free space capable of harboring D2O molecules. All these factors imply that the film is more hydrophobic. The reactions with the hydroxy groups both inside the epoxy network and on substrate surface further enhance the hydrophobicity of the coating. Good adhesion between substrate and coating is ensured by enriched bis-sulfur silane near the substrate. Finally, the bridging group of bissulfur silane is itself hydrophobic. All these factors underlie the improved anticorrosion performance of silane-epoxy films. Acknowledgment. We benefited from numerous useful discussions with Professor Wim van Ooij. We thank Jaraslaw Majewski, Erik Watkins and Hillary Smith for their effort in collecting the reflectivity data. Work at University of Cincinnati was sponsored by the Strategic Environmental Research and Development Program (www.serdp.org). Work performed at Surface Profile Analysis Reflectometer (SPEAR) at Lujan Neutron Scattering Center at Los Alamos National Laboratory was supported by Los Alamos National Laboratory under DOE contract W7405-ENG-36, and by the DOE Office of Basic Energy Sciences. LA8028066
