Relationship between chemical degradation and thickness loss of an amine-cured epoxy coating exposed to different UV environments.

The relationship between chemical degradation and thickness loss of an unpigmented, nonUV-stabilized, crosslinked amine-cured epoxy coating exposed to three UV conditions was investigated. Spin-coated samples having a thickness of approximately 7 [micro]m on an Si substrate were prepared from a stochiometric mixture of a bisphenol A epoxy resin and a tetra-functional amine curing agent. Samples were exposed outdoors and to two accelerated laboratory UV environments. Chemical degradation and thickness loss were measured by transmission Fourier transform infrared spectroscopy (FTIRS) and laser scanning confocal microscopy (LSCM), respectively. In addition, surface roughness and morphological changes were measured by atomic force microscopy (AFM) and LSCM. Substantial chemical degradation, thickness loss, and morpholocal changes occurred in the exposed films, and the rate of chemical degradation was greater than that due to the thickness loss. This additional chemical loss was attributed to an inhomogeneous degradation process in which nanoscale localized depressions initiate at certain sites on the surface, which then enlarge and deepen with exposure time. The results of this study provide a better understanding of the degradation mechanism and should lead to the development of scientific-based models for predicting the service life of crosslinked amine-cured epoxy coatings.

Keywords: Atomic force microscopy, FTIR, ATR, photodegradation, durability, service life prediction, architectural, weatherability, epoxy resins, epoxy, thermoset 

**********

Amine-cured epoxies are used extensively in exterior coatings and polymer-fiber composites. However, epoxy-based materials are known to undergo degradation when exposed to ultraviolet (UV) radiation. (1) In the presence of oxygen, amine-cured epoxies undergo photooxidation, which generates various oxidation products and results in chain scissions. (1-3) Photochemical degradation of the cured epoxy leads to physical changes in the coating films, such as thickness loss, gloss loss, and cracking. Fourier transform infrared spectroscopy (FTIRS) is the most common technique to measure photochemical degradation of polymeric coatings, and the information obtained is commonly used to predict long-term performance. From FTIR data of unpigmented coatings exposed to commercial weathering devices, previous studies have indicated that thickness change in weathered coatings is approximately linear with FTIR intensity loss of the polymer CH groups. (4,5) Further, due to this linear relationship, these studies have asserted that a measurement of the remaining thickness of the weathered film provides an estimate of the degradation (erosion) rate of polymeric coatings. This apriority is predicated on the assumption that chemical degradation (represented by the CH loss) can be entirely attributed to a homogeneous and uniform ablation process, i.e., a uniform thickness loss from the film surface. The correctness of this assumption has great practical implications. If true, then the warranty service life of a clear coating on an automobile can be guaranteed by ensuring that a prespecified minimum film thickness is applied over the entire car. If this assumption is not correct, ensuring a minimum coating thickness may not be the optimum strategy.

There is evidence to suggest that uniform ablation is not the principal degradation mode of polymeric coatings exposed to UV environments. For example, typical unpigmented coating surfaces before exposure are very smooth (mean root square roughness <10 nm) and highly glossy. If the photodegradation is a uniform ablation process, then the surfaces should remain smooth and retain their high glossiness during exposures. However, UV-stabilized and non-UV-stabilized clear polymeric films are known to undergo gradual loss of gloss during exposures to UV. (6,7) This means that the coating surface roughness increases with exposure and that thickness loss is not a uniform process from the surface. Another strong piece of evidence is the extensive data of recent atomic force microscopic (AFM) studies for a variety of crosslinked coatings exposed to UV environments, including using narrow band pass filter and full spectral UV radiation and in the presence or absence of humidity. (8-11) AFM images of these studies have clearly shown that the UV-exposed surfaces degrade nonhomogeneously through the formation of pits, which deepen and enlarge with time. Further, for model acrylic-melamine and acrylic urethane coatings, we have also shown experimentally that photochemical degradation is, for the most part, greater than one would expect relative to an independent measure of thickness loss, indicating that chemical degradation is attributable to additional processes other than thickness loss. (11)

[FIGURE 1 OMITTED]

The main objective of this study is to assess the relationship between the chemical degradation and the thickness loss for a crosslinked amine-cured epoxy coating exposed outdoors and to two accelerated laboratory environments. Results obtained from this study will provide a better understanding of the photodegradation mode and assist in the development of improved models for predicting the service life of amine-cured epoxy coatings. This work is part of a greater effort at the National Institute of Standards and Technology (NIST) on application of the reliability theory approach for predicting the service life of polymeric coatings exposed outdoors from the accelerated laboratory tests.

EXPERIMENTAL PROCEDURES (b)

Materials and Specimen Preparation

An unpigmented, non UV-stabilized, model crosslinked amine-cured epoxy coating was used. The coating was a stoichiometric mixture of a highly pure diglycidyl ether of bisphenol A (DGEBA) with an epoxy equivalent of 172 (grams of resin containing one gram equivalent of epoxide) (DER 332 from Dow Chemicals) and 1,3-bis (aminomethyl)-cyclo-hexane (1,3 BAC) (Aldrich) as the aliphatic amine curing agent. Hereafter, 1,3 BAC-DGEBA system is designated as the amine-cured epoxy coating.

The coating was applied to 100-mm diameter Si wafers by spin casting. The Si substrate was chosen because it is suitable for the quantitative study of coating chemical degradation by transmission FTIR spectroscopy and for the nondestructive measurement of film thickness changes by laser scanning confocal microscopy (LSCM) in the reflection mode. Appropriate amounts of both epoxy resin and amine curing agent in toluene solvent at 90% mass fraction were mixed thoroughly using a mechanical stirrer. The mixture was degassed, deposited onto the substrates, and spun at 210 rad/sec (2000 rpm) for 30 sec. The spin coating operation was carried out in an essentially C[O.sub.2]-free, dry air glove box to minimize the reactions of C[O.sub.2] and moisture with the amine curing agent. Coated samples were cured at room temperature for 24 hr, followed by heating in an air-circulated oven at 130[degrees]C for two hours. The glass transition temperature, [T.sub.g], of the cured film was 123[degrees]C [+ or -] 2[degrees]C (by dynamic mechanical analysis). (The number after the [+ or -] sign indicates one standard deviation.) The average thickness of the coated films was 7.2 [micro]m [+ or -] 0.8 [micro]m, as measured by LSCM.

After curing, the 100 mm diameter coated Si wafers were cut into 25 x 25 mm samples using a diamond saw. These cut samples were exposed to UV environments. For measuring the coating bulk microstructure, cryo-fractured 3 mm thick free films were prepared. In addition to the curing schedule used for chemical degradation analysis samples, the 3 mm thick films were subjected to an additional two hours of post-curing at 130[degrees]C to assure total cure, as verified by attenuated total reflection (ATR) FTIR. The cryo-fractured surfaces were prepared by notching a 12 mm x 12 mm x 3 mm section with a band saw, immersing the notched section in liquid nitrogen for one hour, and cleaving the notch with a screwdriver.

The main chemical structures of the components and the cured film are illustrated in Figure 1. It should be noted that, due to steric hindrance and restricted transport at the late curing stages, unreacted epoxide and amino groups and monosubstituted nitrogen chains are expected to be present in the cured films.

UV Exposure Environments

Samples were exposed outdoors and in two accelerated laboratory UV environments that differed only in relative humidity (RH) levels. The experimental setup for the accelerating laboratory exposures is schematically shown in Figure 2. The UV radiation source consisted of a 1000 W Xenon arc lamp emitting radiation having wavelengths from approximately 275 nm to 800 nm. The light source was equipped with dichroic mirrors that removed most of the infrared and some visible radiation and a photofeedback controller to regulate light intensity. Therefore, if no external heat was supplied, the temperature inside the irradiated chamber was only a few degrees above the ambient room temperature (24[degrees]C). In this study, photodegradation of specimens exposed to accelerating laboratory environments at two levels of RH, 9% (dry) and 75% (humid) and at 50[degrees]C, was conducted. Specimens were placed in two exposure cells. Each exposure cell was equipped with a thermocouple and an RH sensor. They also contained an inlet and an outlet that allowed fresh air with the desired temperature and RH to continuously enter and vent from the cell chamber. A humidity generation and supply system based on a mixture of dry air and moisture-saturated air delivered the RH to the cell chambers. All specimens within each exposure cell were exposed to the same relative humidity. The RH in each exposure cell could be independently controlled and maintained to within approximately [+ or -]3% of their preselected values from 0% to 90%, and the temperature within each cell was controlled to within [+ or -] 1[degrees]C. Complete details on the UV light source, instrumentation, exposure cells, and control systems have been described previously, (12) as well as information on the RH generation system. (13) The laboratory UV/9%RH/50[degrees]C and UV/75%RH/50[degrees]C environments are designated as dry UV and humid UV environments, respectively.

The outdoor exposure was carried out in Gaithersburg, MD. For this experiment, specimens in a lidless polycarbonate container were placed in a custom built outdoor environmental chamber that was tilted at 5[degrees] from the horizontal plane and faced south (Figure 3). The upper surface of the chamber bottom was laid with a 3-mm thick polytetrafluoroethylene sheet to provide some thermal control in the chamber through reflection of the solar radiation. The top of the chamber was covered with a borosilicate glass plate, and all sides were enclosed with a breathable cloth that allowed water vapor, but prevented dust from entering the chamber. The borosilicate glass used in this study contained a low amount of iron, which provides improved transmission of UV, visible, and infrared light over soda-lime glass. UV-visible spectral results showed that this borosilicate glass had a cut off at 270 nm wavelength and did not alter the solar spectrum after more than one year of outdoor exposure in Gaithersburg. All the outdoor exposures were started in April 2003. The temperature and RH in the outdoor chamber were recorded continuously. The temperature inside the chamber was found to be as much as 20[degrees]C higher than the surrounding ambient temperature, and the RH in the chamber often reached 100%, indicating that water often condensed in the outdoor chamber during exposure.

It should be emphasized that the UV light source used in the laboratory experiments contained radiation with wavelengths below 300 nm, which are absent in the terrestrial sunlight. Therefore, the data generated under accelerated and outdoor exposures are strictly used for relating the thickness loss and chemical degradation under their respective exposure environments and should not be used for relating in-service performance with accelerating tests.

[FIGURE 2 OMITTED]

Measurements of Degradation

CHEMICAL DEGRADATION: Coating chemical degradation as a function of exposure time was measured by Fourier transform infrared spectroscopy in the transmission mode (FTIR-T). For accelerating laboratory exposures, the specimen-containing exposure cells were removed from the chambers at preselected time intervals and were fitted into a demountable 150-mm diameter ring of an autosampler described previously, (14) using spring-loaded clips to ensure that the specimens were precisely located. The ring was rotated and translated to cover the entire sampling area. This automated sampling device allowed for rapid FTIR-T spectra recording of all specimens in each exposure cell before or after each exposure time. For outdoor exposures, after being removed from the chamber at a preselected time interval, each specimen was precisely positioned in a transmission sample holder. Since all specimens were mounted precisely on the autosampler or sample holder, errors due to variation of sampling at different exposure times were essentially eliminated. FTIR spectra were recorded at a resolution of 4 [cm.sup.-1] using dry air as the purge gas and an FTIR spectrometer equipped with a liquid nitrogen-cooled mercury cadmium telluride (MCT) detector. All spectra were the average of 128 scans. In all cases, the specimen residence time in the FTIR spectrometer compartment was <15 min. The peak height was used to represent IR intensity, which is expressed in absorbance, A. All FTIR results for each environment were the average of four specimens.

[FIGURE 3 OMITTED]

[FIGURE 4 OMITTED]

[FIGURE 5 OMITTED]

THICKNESS CHANGES: Coating thicknesses before and after exposures for different times were measured nondestructively using a Zeiss model LSM510 laser scanning confocal microscope (LSCM) in the reflection mode. LSCM was also used to characterize surface morphology and roughness of exposed samples. LSCM utilizes coherent light and collects light exclusively from a single plane (a pinhole sits conjugated to the focal plane) and rejects light out of the focal plane. The wavelength, numerical aperture (N.A.) of the objective, and the size of the pinhole determine the resolution in the thickness or axial direction. (15) By moving the focal plane, single images (optical slices) can be combined to build up a three-dimensional stack of images that can be digitally processed. The LSCM images reported were in 2D intensity projection and representative of a series of overlapping optical slices (a stack of z-scan images) with each z-step measuring 100 nm. These images provide surface and subsurface features of a material. The laser wavelength used was 543 nm.

The procedure for measuring the thickness of a polymer film on an Si substrate using LSCM is illustrated in Figure 4. As light is shone onto a polymer-coated Si substrate, ray 1 is reflected from the air/polymer interface and ray 2 is reflected from the polymer/Si interface. The film thickness [d.sub.p] is equal to [n.sub.p] x [d.sub.o], where [d.sub.o] is the distance between the two interfaces, which is measured by LSCM, and [n.sub.p] is the refractive index of the polymer coating. An [n.sub.p] value of 1.56 (16) was used for both initial and degraded epoxy samples. Thickness measurements were performed on five different predetermined locations on each of the four specimens used for the FTIR analysis. Each position was approximately 2 mm apart. Further, thickness measurements were performed only at locations that were free of pits and holes. If a pit had formed at a predetermined location, the measurement was carried out at a nearby pit-free area. A specially designed device attached to the stage of the confocal microscope was used to hold the specimen. All thickness results were the average of 20 measurements. Thicknesses of the unexposed samples were also measured by the scanning electron microscopy (SEM) technique using cross-sections of the same specimens. For these thickness measurements, three 25 mm x 25 mm specimens cut from the same 100-mm diameter coated Si plates were used. SEM thickness results were compared with those obtained by the LSCM technique.

[FIGURE 6 OMITTED]

SURFACE MORPHOLOGICAL CHANGES: Both LSCM and atomic force microscopy (AFM) can give data on roughness and surface morphological changes due to exposure. These techniques provide complementary topographic features ranging in size from a few nanometers to hundreds of micrometers. Surface morphology measurement with an LSCM was described briefly above.

For AFM imaging, specimens (different from those used for FTIR and LSCM measurements) were periodically removed from the laboratory and outdoor chambers and analyzed using a Dimension 3100 scanning probe microscope (Digital Instruments) operated in tapping mode and commercial Si microcantilever probes. Phase changes during scanning in the tapping mode can often provide significantly more contrast than the topographic images, and are useful for studies of surface microstructures. (17) Because a low force is applied and the duration of tip-sample contact is short, tapping mode operation causes minimum damage to the specimen while maintaining a high spatial resolution. (18) AFM images were obtained using a resonance frequency of approximately 300 kHz for the probe oscillation and a free oscillation amplitude of 62 nm [+ or -] 2 nm. The set-point ratio (the ratio of set point amplitude to the free amplitude) ranged from 0.60 to 0.80. Root mean square (RMS) roughness values obtained from LSCM and AFM techniques were based on the 450 [micro]m x 450 [micro]m scans and 20 [micro]m x 20 [micro]m scans, respectively, and were the average of four measurements (two measurements on each of two specimens).

[FIGURE 7 OMITTED]

[FIGURE 8 OMITTED]

RESULTS AND DISCUSSION

Microstructure of Amine-Cured Epoxy Coatings

Analyses of the surface and bulk microstructures of a polymeric coating provide crucial information for understanding how a polymeric material degrades during environmental exposure. Topographic and phase AFM images taken at the surface (air) and from within the bulk (fractured) of the amine-cured epoxy are displayed in Figures 5a and 5b, respectively. The surface images were taken on the same specimens used for chemical degradation analyses, and the bulk images were from a cryo-fractured surface of the 3-mm thick section. For each pair in Figure 5, topographic images are on the left side and phase images are on the right side. The bright and dark areas in the topographic images correspond to the peaks and valleys, respectively.

[FIGURE 9 OMITTED]

The topographic and phase AFM images of the surface both show that it is nearly featureless but has a detectable pattern consisting of bright and dark domains. Examination of the phase images taken in this and previous studies (19,20) reveals that the surface of the amine-cured epoxies is generally featureless or appears as an unclear pattern. The lack of a clear pattern of the surface sample is probably due to

the presence of a very thin layer of a rather homogeneous, lower (than the bulk) surface-free energy material. Direct evidence to support this postulation for an amine-cured epoxy is not available. However, extensive experimental data obtained by AFM and other surface analytical techniques reported in the literature show that the outer surface of a multicomponent polymer system is generally enriched with a lower surface-free energy component to minimize the polymer-air interfacial energy. (21,22) A thin layer of such material would likely mask the microstructure underneath, and the surface should appear homogeneous and featureless. Since AFM operates in the near field (i.e., very close to the surface), it images only the top layer of rigidly bound atoms and does not image the layer beneath the outermost surface.

The phase image of the bulk (Figure 5b, right) clearly shows that this amine-cured epoxy has a two-phase microstructure consisting of a matrix that appears bright, and interstitial regions dispersed throughout the matrix that appear dark. This varied feature of the bulk is similar to that observed previously (23) for the surface of DGEBA cured with bis(para-aminocyclohexyl) methane and is analogous to that of the microtomed and interface samples of other amine-cured epoxies, (19,20) polyester film surface, (24) and acrylic-melamine coating interfaces. (25) Although the exact contrast mechanism in phase imaging is not fully understood, the bright domains in the phase image have been interpreted as being due to a mechanically harder area, and the dark surrounding region as being due to more compliant material. (26,27) From these assignments, it is reasonable to conclude that the bulk microstructure of an amine-cured epoxy is heterogeneous, consisting of softer regions (dark) dispersed in a harder matrix (bright). The harder, nodular domain in the matrix has been attributed to the high crosslinked material and the soft interstitial regions to the less crosslinked, low molecular mass material. (23,24,28,29) Bascom (30) surmised that such heterogeneous structures are formed in crosslinked materials because, as the high molecular mass segments polymerize and terminate to form a network, some unreacted and partially polymerized molecules are unable to merge into the network structure and are left at the periphery of the network units.

Degradation During UV Exposure

CHEMICAL CHANGES: Figures 6a, 6b, and 6c display FTIR-T spectra for three exposure times of amine-cured epoxy specimens before and after exposures in the three environments. One can see that these spectra contain the interference fringes. These fringes arise from the constructive and destructive interference of the multiply-reflected components with the first transmitted components, or equivalently, with the front surface reflected component. Further, both the spacing and intensity of the interference fringes vary with exposure time (i.e., degree of degradation in samples). Similar to that observed for other coatings, (5,11) this type of fringe, which is prominent when an Si substrate is used, would contribute to errors in the intensity measurements of the weak bands. In this study, chemical degradation analysis was based on the strong absorption bands, and thus the errors due to the interference fringe should be small.

The bands of interest in an unexposed amine-cured epoxy material (Figure 6, initial) are those at 1037 [cm.sup.-1], due to aliphatic-ether stretching; at 1245 [cm.sup.-1], due to aryl-ether stretching; at 1510 [cm.sup.-1], due to C=C stretching of the benzene ring; and near 3400 [cm.sup.-1], due to hydrogen-bonded OH stretching. As the coating degraded, the intensities of these existing bands decreased and new bands at 1730 [cm.sup.-1], due to C=O stretching of a ketone, and 1660 [cm.sup.-1], assigned to amide C=O stretching, appeared. The formation of these oxidation products is in good agreement with a photooxidative mechanism proposed by Bellinger and Verdu (2,3) for epoxy cured with aliphatic amines. Ketone formation is derived from the secondary hydroxyl groups, and amide is generated from the abstraction of a hydrogen of the methylene groups adjacent to the crosslink. Radicals and oxygen are required for both reactions. Although Figure 6 provides useful information about coating degradation, the decrease in intensities of the existing bands and the formation of new bands resulting from exposures are more clearly seen in the difference spectra illustrated in Figure 7. These spectra were obtained by subtracting the spectrum of the unexposed specimen from that of the exposed specimen after adjusting for any baseline shift. In a difference spectrum, the bands below and above the zero baseline indicate loss and formation, respectively, of certain functional groups. As seen in Figure 7, in addition to obvious changes of the prominent bands, difference spectra can also clearly reveal the formation and depletion of minor components in a degraded amine-cured epoxy film.

[FIGURE 10 OMITTED]

[FIGURE 11 OMITTED]

In this study, the bands at 2925 [cm.sup.-1] (due to CH stretching of the C[H.sub.2] group), 1245 [cm.sup.-1], 1510 [cm.sup.-1], 1660 [cm.sup.-1], and 1730 [cm.sup.-1] were used to follow various degradation processes of the UV exposed amine-cured epoxy; these results are depicted in Figure 8. Each data point in this figure is the average of four replicate specimens, and the standard deviations are shown by the error bars, which indicate a good reproducibility of the chemical degradation data. A decrease of both the 1510 [cm.sup.-1] and 1245 [cm.sup.-1] bands signify a chain scission process because these bands represent the chemical groups present in the main chains of the epoxy resin. Figure 8 clearly shows that unstabilized amine-cured epoxies undergo rapid chemical degradation under the three UV environments. For example, after being exposed for 55 days in humid UV, dry UV, and outdoor, both the 1510 [cm.sup.-1] (aromatic ring) and the 1245 [cm.sup.-1] (aryl-ether groups) bands lost more than 50%, 40%, and 45% of their intensities, respectively. Similar relative changes were observed for the C[H.sub.2] groups (2925 [cm.sup.-1] band) under the same environments and for the same duration. On the other hand, the formation of amide (1660 [cm.sup.-1] band) and ketone (1730 [cm.sup.-1] band) appeared to reach a maximum after approximately 20 days in the laboratory environments and 40 days of exposure to the outdoors despite the fact that the degradation of other functional groups still continued thereafter. Such behavior was probably due to a combination of two factors: depletion of the oxidizable groups and loss of the oxidized products that were formed in the films.

Thickness Changes

As indicated in the Experimental Section, SEM technique was employed to measure coating thickness, and the results were used to verify the thickness values obtained by the LSCM technique. Figure 9 shows representative SEM images at two magnifications of a coated cross section before exposure. They reveal the uneven surface topography of the epoxy film and the fractures along the epoxy/Si substrate interface. These features would likely contribute to the experimental errors of thickness values obtained by the SEM technique. From the SEM images, the average thickness of these specimens taken from 18 different measurements (six different locations for each of the three specimens) was 5.8 [micro]m [+ or -] 0.5 [micro]m. The corresponding average LSCM thickness of the same samples was 6.4 [+ or -] 0.7 [micro]m. The difference between the two techniques was within the experimental measurement errors. A previous study showed that thickness values of polymer films measured by the LSCM technique agree well with those obtained by ellipsometry. (11) These results validate the use of LSCM to measure thickness of a polymer film on an Si substrate.

One aspect of film thickness measurement by LSCM that deserves comment is the values obtained for degraded specimens. As indicated in the Experimental Section, thickness measurements were performed on locations that were devoid of pits. For the early exposure stages, where pit formation is still negligible, the average value for a set of four specimens is considered to represent the true average of the film. However, at later degradation stages where more substantial pitting has occurred, the measured thickness values, which account only for the nonpitted areas, are probably somewhat higher than the average remaining thickness of the entire film, (including both nonpitted and pitted areas). However, the difference between the actual and the measured values in all cases should be small because the total areas occupied by the pits and surface deformation are still small and the pits are mostly shallow. In addition, the thickness of any film, porous films such as membranes, or nonporous films, is generally given based on the nonpore areas of the film. Therefore, thickness values of the degraded coating reported here were consistent with the universal definition of film thickness for both porous and nonporous polymer films.

Figure 10 shows film thickness as a function of time exposed to the three UV environments. Each data point given in this figure is the average of 20 measurements, and the error bars represent one standard deviation. In both laboratory environments, humid UV and dry UV, little change in thickness was detected in the first 10 days of exposure. Between 10 days and 130 days, the thickness decreased almost linearly with exposure time. After 130 days of exposure in humid UV and dry UV environments, the films lost approximately 37% and 28%, respectively, of their initial thickness. For outdoor exposure, little change in thickness was observed for the first 25 days. Thereafter, the thickness decrease was almost linear with exposure time, reaching a thickness loss of nearly 18% after 75 days. Note that the standard deviations of the thickness values for samples exposed to the dry UV environment were larger than those exposed to the humid UV or outdoor, indicating a greater surface topographical variability in the specimens exposed to the dry UV environment. This was consistent with the changes in roughness data, to be discussed later.

Morphological and Roughness Changes

Figures 11a, 11b, and 11c are 2D (upper row) and 3D (bottom row) AFM topographic images taken after exposing the specimens to humid UV for 38 days, dry UV for 45 days, and the outdoors for 40 days, respectively. AFM images of the film surface before exposure is shown in Figure 5. For film exposed to the laboratory humid UV condition, a surface degradation pattern was clearly observed, as evidenced by the regular topographical feature shown in the 3D image (Figure 11a). The RMS roughness (from AFM 20 [micro]m scan) of these specimens increased from approximately 7 nm [+ or -] 1 nm for the unexposed specimens to 30 nm [+ or -] 2 nm after 38 days of exposure. For specimens exposed to dry UV (Figure 11b), except for the dark colored dots and strings, which are the dewetting pattern and present on the film surface before exposure, morphological changes were less visible than those of the other two exposure environments. The RMS roughness (from AFM data) of these samples increased from approximately 7 nm [+ or -] 1 nm for the unexposed specimen to 18 nm [+ or -] 3 nm after 45 days of exposure. Outdoor-exposure specimens showed a definitive degradation pattern (Figure 11c) after 40 days. A close examination of the circular features in this image at higher magnifications indicated that they are actually shallow pits, having raised rims (bright color) at the pit mouths surrounding the shallow, depressed areas (dark color) inside. The origin of these degradation features, which have always been observed in outdoor exposure samples for this amine-cured epoxy regardless of the exposure time, is unknown at this time. Experiments are being conducted to identify the effects of various factors, including condensed water and unreacted amine, on degradation.

[FIGURE 12 OMITTED]

[FIGURE 13 OMITTED]

Typical 450 [micro]m x 450 [micro]m scan LSCM images and their corresponding surface line profiles of specimens exposed to the three environments for more than two months are shown in Figure 12. Initially, the surface appears smooth and featureless. However, evidence from the LSCM images and their corresponding line profiles displayed in Figure 12 clearly indicates that pitting and roughening have occurred on the surfaces of these specimens after exposure.

[FIGURE 14 OMITTED]

Changes in RMS roughness with time of exposure to three UV environments are given in Figure 13. These roughness values were measured by LSCM and were the average of four 450 [micro]m x 450 [micro]m scans from two different specimens. As exposure time increased, the surface roughness of specimens exposed to humid UV exhibited several different stages: a relatively rapid rise for the first 30 days of exposure, essentially unchanged between 30 days and 90 days, then a sharp increase again between 80 days and 110 days, and nearly constant thereafter. Although there are a few data points that do not follow in line with the general trend, specimens exposed to dry UV showed a sharp rise in roughness during the first 20 days, followed by a low rate of roughness increase. For outdoor exposure, the roughness appeared to vary little in the first 30 days, but increased rapidly thereafter reaching a value of 0.45 [micro]m after 75 days.

Surface morphological and roughness changes observed in Figures 11 to 13 for specimens exposed to three different UV environments provide strong evidence to suggest that the degradation of this non-UV-stabilized, crosslinked, amine-cured epoxy coating was not a uniform thickness reduction process. This observation is similar to those reported previously for photodegradation (8-11) and hydrolytic degradation (immersion in both acid and alkaline solutions) (25,26,31,32) for polymers and coatings. In both types of degradation, the film surface degraded nonhomogeneously through the formation of localized nanoscale pits, which deepened and enlarged with time. For hydrolysis, the degradation has been postulated as an autocatalytic process occurring at the hydrolysis-susceptible, hydrophilic regions in the films. (25,31) However, the nature of the areas where degradation initiates under UV radiation is not certain. Work is in progress in our laboratory to address this question. It should be mentioned that the accelerating UV exposure environments used in references 4 and 5 included liquid water sprayed onto the samples. Therefore, the surface morphological changes during exposure of samples used in those studies may be different from those observed above for epoxy and previously in our laboratory for acrylic-melamine and acrylic-urethane coatings. (8-11)

Relation Between Thickness Loss and Chemical Changes

The relationship between the thickness loss and the FTIR intensity decreases (chemical degradation) for three different functional groups in an amine-cured epoxy coating exposed to the three environments are illustrated in Figure 14. Both the thickness loss and intensity decrease are expressed as relative changes with respect to the initial values. Normalization against the bands' initial values eliminates any errors due to variations between individual specimens. Further, because the intensity of an IR band is directly proportional to its absorption coefficient (extinction coefficient), such normalization would allow for a direct comparison of the rates of different degradation processes. In Figure 14, if the rates of relative change in the chemical degradation were totally explained by thickness loss, the relationship would be a straight line with a slope of 1 (as shown in the figure).

For both outdoor and laboratory UV exposures, the relative changes of the three chemical degradation processes, namely chain scission (aryl-O bond at 1245 [cm.sup.-1]), C[H.sub.2] group loss (2925 [cm.sup.-1] band), and aromatic ring loss (1510 [cm.sup.-1] band), followed each other closely. Therefore, any of these processes can be used to express the photochemical degradation of an amine-cured epoxy. It is clear from Figure 14 that, for all three exposure environments, a substantial chemical degradation occurred before the loss in thickness was detected, and that this disparity is greater for the outdoor than for the laboratory exposures. Further, although the thickness loss and chemical degradation appear to generally follow a nearly linear relation after an initial period, the rate of thickness change is much lower than that of the chemical degradation. If the early stage degradation data is included for the entire degradation period, the rate of thickness change is only approximately 60% of the rate of chemical degradation. These results clearly indicate that the chemical change is due to other mechanisms besides film ablation.

One question that needs to be addressed is how coating surface topography changes during exposure would affect the measured FTIR intensity. As the film degrades, pits of various sizes develop and the surface roughness increases. These surface topographic changes will scatter IR radiation. If the topographic changes are sufficiently severe, such scattering could reduce the amount of IR radiation reaching the FTIR detector. This means that the transmission ratio between the detected intensity I and the incident intensity [I.sub.o] (I/[I.sub.o], the transmittance T) should decrease due to IR radiation scattering. Since absorption is an inverse function of transmission (A=log100/T), a decrease of T will result in an increase of FTIR absorption of the films. Therefore, if surface topography were a factor in the FTIR intensity changes during exposure, the intensity loss/time results shown in Figure 8 would represent a lower-bound estimate of the IR absorption. Thus the actual difference between the thickness loss and chemical degradation should be greater than those shown in Figure 14.

SUMMARY AND CONCLUSIONS

Previous studies have asserted that the thickness loss of an unpigmented, crosslinked coating film exposed to UV is directly proportional to the loss of the CH groups in the film and that the film thickness measured as a function of exposure time can be used to estimate the erosion rate of polymeric coatings. This apriority is based on the assumption that chemical degradation can be entirely attributed to a homogeneous and uniform ablation process. However, the gloss loss phenomenon commonly observed on weathered coatings and recent AFM results of coatings exposed to UV suggest that uniform ablation may not be the principal degradation process of polymeric coatings and polymers. This study investigates the relationship between chemical degradation and thickness loss of an amine-cured epoxy coating exposed to outdoors and accelerating laboratory UV environments. Samples of an unpigmented, crosslinked, non-UV-stabilized amine-cured epoxy applied to an Si substrate were exposed to the outdoors and to a Xenon arc light source with a ~275 nm to 800 nm wavelength, at 9%, and 75% RH, at 50[degrees]C. Chemical degradation, thickness loss, and surface physical degradation as a function of exposure time were measured by FTIR spectroscopy, laser scanning confocal microscopy, and atomic force microscopy. Based on microscopic and spectroscopic results obtained in this study, the following main conclusions are drawn:

(1) The rate of chemical degradation for an amine-cured epoxy coating is always greater than that of the thickness loss.

(2) Photodegradation of an amine-cured coating is not a uniform thickness reduction (ablation) but is an inhomogeneous erosion process, with formation of localized nanometer-micrometer depressions and pits.

(3) Using the thickness change with exposure to estimate the chemical degradation rate will result in an underestimate of the chemical degradation rate.

Data on the rates of chemical degradation and thickness loss and on the surface topographic changes provided by this study can contribute to the development of better conceptual and mathematical models for predicting the service life of amine-cured epoxy coatings.

ACKNOWLEDGMENTS

This research is part of a government/industry consortium on Service Life Prediction of Coatings at NIST. Companies involved in this consortium include Akzo Nobel, Arkema Inc., Atofina, Atlas Material Testing Solutions LLC, Dow Chemical Company, and The Sherwin-Williams Company. Federal Highway Administration and the Forest Products Laboratory also provided additional funds for this research. We also thank Susan Falcone of Dow Chemical Company for providing the epoxy resin.

References

(1) Rabek, J.F., Polymer Photodegradation--Mechanisms and Experimental Methods, Chapman & Hall, New York, pp. 269-278, 1995.

(2) Bellinger, V., Bouchard, C., Claveirolle, P., and Verdu, J., "Photooxidation of Epoxy Resins Cured by Non-Aromatic Amines," Photochemistry, 1, 69 (1981).

(3) Bellinger, V. and Verdu, J., "Oxidative Skeleton Breaking in Epoxy-Amine Networks," J. Appl. Polym. Sci., 30, 363 (1985).

(4) Gerlock, J.L., Smith, C.A., Nunez, E.M., Cooper, V.A., Liscombe, P., Cummings, D.R., and Dusibiber, T.G., "Measurements of Chemical Change Rates to Select Superior Automotive Clear-coats," in Polymer Durability, in Advances in Chemistry Series 249, Clough, R., Billinham, N., and Gillen, K. (Eds.), American Chemical Society, Washington, DC, pp. 335-347, 1996.

(5) Croll, S.G. and Skaja, A.D., "Quantitative Spectroscopy to Determine Effects of Photodegradation on a Model Polyester-Urethane Coating," J. COAT. TECHNOL., 75, No. 945, 85 (2003).

(6) Wernsath, K.L., "Service Life Prediction of Automotive Coatings, Correlating Infrared Measurements and Gloss Retention," Polym. Degrad. Stab., 54, 57 (1996).

(7) Decker, C. and Zahouily, K., "Photodegradation and Photooxidation of Thermoset and UV-Cured Acrylate Polymers," Polym. Degrad. Stab., 64, 293 (1999).

(8) Nguyen, T., Gu, X., VanLandingham, M.R., Ryntz, R., Nguyen, D., and Martin, J.W., "Nanoscale Characterization of Coatings Surface Degradation with Tapping Mode AFM," Proc. Adhesion Society Meeting, Anderson, G.L. (Ed.), p. 505, 2003.

(9) VanLandingham, M.R., Nguyen, T., Byrd, W.E., and Martin, J.W., "On the Use of the Atomic Force Microscope to Monitor Physical Degradation of Polymeric Coating Surfaces," J. COAT. TECHNOL., 73, No. 923, 43 (2001).

(10) Nguyen, T., Gu, VanLandingham, M.R., Byrd, E., Martin, D., Ryntz, R., and Martin, J.W., "Degradation Modes of Polymeric Coatings Exposed to UV," Proc. Coatings for Plastics Symposium, Troy, MI, July 2004.

(11) Sung, L., Jasmin, J., Gu, X., Nguyen, T., and Martin, J.W., "Use of Laser Scanning Confocal Microscopy for Characterizing Changes in Film Thickness and Local Surface Morphology of UV Exposed Polymer Coatings," J. COAT. TECHNOL. RES., 1, No. 4, 267 (2004).

(12) Martin, J.W., Nguyen, T., Byrd, W.E., Dickens, B., and Embree, N., "Relating Laboratory and Outdoor Exposures of Coatings: I. Cumulative Damage Model and Laboratory Experiment Apparatus," Polym. Degrad. Stab., 75, 1, 193 (2002).

(13) Nguyen, T., Martin, J.W., Byrd, W.E., and Embree, N., "Relating Laboratory and Outdoor Exposure of Coatings: III. Effects of Relative Humidity on Moisture-Enhanced Photolysis of Acrylic-Melamine Coatings," Polym. Degrad. Stab., 77, 1,1 (2002).

(14) Nguyen, T., Martin, J.W., Byrd, W.E., and Embree, N., "Relating Laboratory and Outdoor Exposures of Coatings: II. Effects of Relative Humidity on Photodegradation of Acrylic Melamine," J. COAT. TECHNOL., 74, No. 932, 31 (2002).

(15) Corle, T.R. and Kino, G.S., Confocal Scanning Optical Microscopy and Related Imaging Systems, Academic Press, 1996.

(16) Brandrup, J., Immergut, E.H., and Grulke, E.A. (Eds.), Polymer Handbook, John Wiley, Inc., New York, NY, p. VI-576, 1999.

(17) Magonov, S.N. and Heaton, M.G. "Atomic Force Microscopy, Part 6: Recent Developments in AFM of Polymers," Am. Lab., 30, 9 (May 1998).

(18) Sauer, B.B., McLean, R.S., and Thomas, R.R., "Tapping Mode AFM Studies of Nano-Phases on Fluorine-Containing Polyester Coatings and Octadecyltrichlorosilane Monolayers," Langmuir, 14, 3045 (1998).

(19) Gu, X., Raghavan, D., Ho., D.L., Sung, L., VanLandingham, M.R., and Nguyen, T., "Nanocharacterization of Surface and Interface of Different Epoxy Networks," Mater. Res. Soc. Symp., 710, DD10.9 (2002).

(20) Gu, X., Martin, D., Martin, W.J., and Nguyen, T., "Surface, Interface, and Bulk Microstructure of Amine-Cured Epoxies--An Intensive AFM Study," Proc. Adhesion Society Meeting, Chaudhury, M. (Ed.), p. 507, 2004.

(21) Chen, X., McGurk, S.L., Davies, M.C., Roberts, C.J., Shakesheff, K.M., Tendler, S.J.B., and Williams, P.M., "Chemical and Morphological Analysis of Surface Enrichment in a Biodegradable Polymer Blend by Phase-Detection Imaging Atomic Force Microscopy," Macromole-cules, 31, 2278 (1998) and references therein.

(22) Pienka, Z., Oike, H., and Tezuka, Y., "Atomic Force Microscopy Study of Environmental Responses on Poly(vinyl alcohol)-graft-Polystyrene Surfaces," Langmuir, 15, 3197 (1999) and references therein.

(23) VanLandingham, M.R., Eduljee, R.F., and Gillespie, J.W., "Relationship Between Stoichiometry, Microstructure, and Properties of Amine-Cured Epoxies," J. Appl. Polym. Sci., 71, 699 (1999).

(24) Gu, X., VanLandingham, M.R., Raghavan, D., and Nguyen, T., "Characterization of Polyester Degradation in Alkaline Solution with AFM," Polym. Degrad. Stab., 74, 139 (2001).

(25) Nguyen, T., Martin, J.W., Byrd, W.E., and Embree, N., "Relating Laboratory and Outdoor Exposures of Coatings: IV. Mode and Mechanism of Hydrolysis of Acrylic-Melamine Coatings," J. COAT. TECHNOL., 75, No. 941, 37 (2003).

(26) Bar, G., Thomann, Y., Brandsch, R., and Whangbo, M.H., "Factors Affecting the Height and Phase Images in Tapping Mode Atomic Force Microscopy. Study of Phase-Separated Polymer Blends of Poly(ethene-co-styrene) and Poly(2,6-dimethyl-1, 4-phenylene oxide)," Langmuir, 13, 3807 (1997).

(27) Magonov, S.N., Elings, V., and Whangbo, M.H., "Phase Imaging and Stiffness in Tapping Mode Atomic Force Microscopy," Surf. Sci., 375, L385 (1997).

(28) Cuthrell, R.E., "Epoxy Polymers. II. Macrostructure," J. Appl. Polym. Sci., 12, 1263 (1968).

(29) Giraud, M., Nguyen, T., Gu, X., and VanLandingham, M.R., "Effects of Stoichiometry and Epoxy Molecular Mass on Wettability and Interfacial Microstructures of Amine-Cured Epoxies," Proc. Adhesion Society Meeting, Emerson, J.A. (Ed.), p. 260, 2001.

(30) Bascom, W.D., "Water at the Interface," J. Adhes., 2, 168 (1970).

(31) Leadley, S.R., Shakesheff, K.M., Davies, M.C., Heller, J., Franson, N.M., Paul, A.J., Brown, A.M., and Watts, J.M., "The Use of SIMS, XPS, and In Situ AFM to Probe the Acid Catalysed Hydrolysis of Poly(orthoesters)," Biomaterials, 19, 1353 (1998).

(32) Gopferich, A. and Langer, R., "The Influence of Microstructure and Monomer Properties on the Erosion Mechanism of a Class of Polyanhydrides," J. Polym. Sci., Part A: Polym. Chem., 31, 2445 (1993).

Aziz Rezig, Tinh Nguyen, David Martin, Lipiin Sung, Xiaohong Gu, Joan Jasmin, and Jonathan W. Martin--National Institute of Standards and Technology (a)

Presented at the 82nd Annual Meeting of the Federation of Societies for Coatings Technology, October 27-29, 2004, in Chicago, IL.

(a) 100 Bureau Dr., Gaithersburg, MD 20899.

(b) Certain commercial products or equipment are described in this article in order to specify adequately the experimental procedure. In no case does such identification imply recommendation or endorsement by the National Institute of Standards and Technology, nor does it imply that it is necessarily the best available for the purpose.