Assessment of various analytical techniques on naturally weathered paints as a predictor of long-term gloss retention for waterborne acrylic architectural paints.

Gloss retention of waterborne acrylic architectural paints is one of many properties defining exterior durability. Assessing gloss retention is a time consuming process, and accelerated weathering protocols (Weather-Ometer[R]) typically have poor correlation with natural exposures for this class of paints. The recent development of new analytical techniques for assessing automotive paint durability raises the possibility of detecting subtle changes in short-term natural exposures that would be predictive of long-term gloss retention; however, extension to architectural paints has yet to be demonstrated. To test the concept, seven paints based on different aqueous emulsion polymers were exposed to natural weathering at the same time for 12 weeks at the Paint Quality Institute (PQI) test site, and were then examined using scanning probe microscopy, optical profilometry, chemiluminescence, XPS, and FTIR. These paints were also exposed for 1000 hr in a Weather-Ometer (WOM). Scanning probe microscopy, optical profilometry, chemiluminescence, FTIR, and the WOM results all showed poor correlation with 96-week natural exposure at the PQI test site. The lack of correlation for the WOM results emphasizes the crucial importance of using natural weathering to assess gloss retention. In contrast, the XPS shows some promise for qualitatively predicting subsequent gloss loss and appears worthy of further exploration. As employed here, the XPS was used to monitor the ratio of pigment (Al, Si, and Ti) to carbon coming from the Ti[O.sub.2] in the top few nanometers of the paint surface. A correlation ([R.sup.2] = 0.731) was found between gloss retention after 96 weeks and the change (2 versus 12 weeks of exposure) in pigment to carbon ratio. This correlation is probably sufficient to rank a series of paints and allow for an assessment against well-established controls. While promising, XPS has some issues. Surface contamination, equipment expense, the considerable expertise required, and precision make utility for screening less than ideal.

Keywords: Atomic force microscopy, ESCA, XPS, auger, SIMS, latexes, colloids, emulsions, accelerated testing, durability, service life prediction, weatherability, architectural, water-based, acrylics

**********

Over the last several decades there have been significant efforts to establish faster and more predictive methods for assessing the exterior durability of paints and coatings. This is driven by a need to reduce product development cycles, thus shortening time-to-market for new products, and a need to better predict the service life of paints and coatings. (1) Better service life prediction has a profound impact on product warranties and liabilities and, thus, has its greatest impact in markets such as automotive coatings and industrial maintenance where product liabilities are high. In fact, shorter product development cycles impact the whole coatings market by reducing both the cost and risk associated with new products.

Given these drivers, it is not surprising that there has been a significant effort to establish both a fundamental understanding of exterior coating failures and a better way to predict exposure-induced failure. The overwhelming bulk of the work has been on automotive coating systems (2) and on corrosion protection (3) of steel structures. The issues of durability in architectural paints are, in general, much more complex due to the variety of substrates, variable conditions of the substrates, and highly variable applications. Despite the recent efforts on automotive coating systems, there continues to be a need for more predictive accelerated tests for paints and coatings, and in particular, waterborne architectural paints. In addition, the issue of durability assessment of waterborne architectural paints has increased in importance as restraints on volatile organic compounds (VOCs) are forcing many paint companies to select softer polymers, which generally provide a less optimized balance of chalk resistance, gloss retention, fade resistance, and dirt resistance, as compared with comparable formulations made with conventional latex emulsion polymers in use today.

Exterior durability is a complicated catch-all phrase which tends to imply a large number of paint performance features and a myriad of different and complicated failure modes, many of which are not related to one another or are actually antagonistic properties, such as dirt pick-up resistance and resistance to cracking. In the context of waterborne architectural paints, some of the more important exterior performance features are:

(1) Resistance to flaking, cracking, and peeling over dimensionally unstable substrates such as wood;

(2) Adhesion to a variety of possible substrates including weathered paint or wood, cementitious materials, metals, and solventborne alkyd paints;

(3) Dirt pick-up resistance;

(4) Resistance to mildew growth;

(5) Efflorescence resistance over cementitious substrates;

(6) Hydrolysis resistance over alkaline surfaces such as fresh or damp masonry;

(7) Gloss retention;

(8) Resistance to color fading; and

(9) Resistance to chalking and erosion.

To date, there are no acceptable means of assessing these properties without doing natural exterior exposure studies. Correlation between accelerated lab tests and actual long-term exposures is very weak to nonexistent, and depends on a complex relationship between paint property, substrate, and environmental conditions. Thus, the current state-of-the-art for assessing exterior durability still involves doing a combination of exterior exposures over multiple substrates in different directions and in different geographical locations. Despite the current lack of success in using accelerated lab tests to predict exterior durability, there continues to be a strong interest in developing better and quicker test methods.

GLOSS RETENTION

The current state-of-the-art technology for gloss durability studies of waterborne architectural paints involves a combination of exterior exposures over several substrates in different directions and at different geographical locations, and accelerated lab tests such as QUV, and Weather-Ometer exposure. Correlation between the accelerated lab tests and exterior exposures is frequently questionable and, thus, not relied upon in the architectural paint industry. QUV and Weather-Ometer tests do not accelerate all processes equally, and, in fact, cannot be practically utilized for some critical properties such as mildew and blister resistance over various surfaces. Efforts to design improved instrumentation to accelerate film degradation have met with limited success.

The only reliable and trusted method for assessing gloss retention of waterborne architectural coatings continues to be natural exterior exposures. The use of accelerated outdoor exposure sites such as in South Florida and Arizona, as well as Q-Trac or Emmaqua[R], can help to accelerate the gloss degradation. However, these more extreme exposure sites and conditions offer only moderate acceleration of gloss degradation, and published long-term studies correlating these sites to actual exposure are lacking for architectural paints. Given this, there continues to be a pressing need for better means of assessing the gloss retention of this class of coatings.

ALTERNATIVE APPROACH TO PREDICTING GLOSS RETENTION

An alternative approach to accelerating the degradation process is the use of highly sensitive analytical techniques to detect and evaluate subtle changes occurring early in the life of a naturally weathering paint, long before these changes are noted by the eye or by conventional instruments. These changes may be early indicators of potential failures, which will occur at longer times; however, it remains to be seen whether this is a viable gloss retention assessment strategy. This approach offers a new and intriguing possibility for the assessment of the gloss retention of waterborne acrylic architectural paints. The success of such a strategy is potentially enabled by developments in powerful analytical techniques, which can probe extremely subtle changes occurring early in the life of a paint film.

Gerlock, Bauer, and co-workers (2) at Ford Motor Company have had a great deal of success in utilizing many analytical techniques to identify the telltale signs of photodegradation that occur in automotive coatings; but to the best of our knowledge these techniques have not been applied to architectural waterborne paints. Since many of these techniques were developed specifically to study multilayer automotive coatings, it is not obvious that they would be applicable to studying acrylic architectural waterborne paints. In particular, many of these analytical methods were designed to determine the fate of photostabilizers, which are typically not present in acrylic architectural waterborne paints and, thus, might not be applicable. In addition, the photodegradation pathway for acrylic polymers is markedly different from the pathway for highly crosslinked polyurethane or aminoplast-based coatings, and may not produce species that are spectroscopically identifiable.

APPROPRIATE ANALYTICAL METHODS FOR ASSESSING COATING DEGRADATION.

There have been many analytical techniques applied to the study of coating degradation. After extensively reviewing the literature it was decided that the following techniques showed the most promise for predictive testing.

Scanning Probe Microscopy

Scanning probe microscopy (SPM) has been utilized widely to reveal the topography of coating surfaces. (4) In the simplest form of the technique, a stylus closely follows the surface contour as it is scanned across the sample. An analysis of the surface contour gives statistics of the contour variation. A wide range of statistics associated with surface topography may be obtained. Two typical measurements are the peak-to-valley, which determines the highest and lowest point on the image, and average roughness values. These values, along with more detailed image analysis, can provide insight on film properties and weathering of the paints, as described above.

Optical Profilometry

Optical profilometry is often utilized to complement the capabilities of SPM. The profilometer is operated in vertical scanning interferometry mode. In this mode, interference fringes are created when reflected light from a reference mirror is combined with light reflected from the sample. The system measures the degree of fringe modulation as the objective lens is scanned through focus. Due to the short coherence length of the white light utilized, the interference fringes are present only over a very limited depth of focus. The vertical position of the sample corresponds to the peak of the low-pass filter signal (where the low-frequency components are removed and the signal is rectified by square-law detection). A curve-fitting interpolation is employed to increase the vertical resolution of the measurement beyond the sampling interval. Thus, vertical resolution of 3 nm can be obtained. An advantage of incorporating optical profilometry into our study is that it offers an alternative approach to SPM (i.e., optical verses mechanical) for measuring surface topography. Also, it extends the sampling length (area analyzed) to the millimeter range.

Chemiluminescence

Chemiluminescence is a recently developed method for monitoring oxidative reactions in polymers. Chemiluminescence occurs when energy from chemical reactions is released as light (e.g., as with the firefly). With the advent of sensitive optical detectors, it was realized that organic materials undergoing oxidation emit low levels of light, and these light levels could be used to monitor the extent of oxidative degradation. While the precise chemical mechanisms giving rise to the optical signal are subjects of debate and no doubt vary from material to material, it is generally accepted (5-7) that chemiluminescence in polymers stems from the recombination of peroxide radicals formed from alkyl bond scission and subsequent oxidation.

These peroxide radicals react to generate excited state carbonyls that return to their ground state via the emission of a photon. Chemiluminescence has been used to monitor the effect of stabilizers on polypropylene (3,8-10) to track the thermal oxidative stability of ABS (acrylonitrile-butadiene-styrene) copolymers, (11-13) to study the photostability of automotive coatings, (14) and to predict the relative UV durabilities of a series of polyester coatings. (15)

In general, the experimental work has fallen into two classes. For polymers such as polypropylene or ABS, the chemiluminescence signal is tracked as the sample is heated in an oxidative environment, and the magnitude and time-dependence of the signal are used to draw conclusions about relative oxidative stability and, in some cases, the kinetics of the degradation processes. The second class of experiments generally involves heating the samples and monitoring the chemiluminescence signal in an inert environment. Here, the signal is believed to arise primarily from polymeric hydroperoxides already present in the sample. Thus, aging the samples prior to the chemiluminescence experiment by exposure to either heat, light, or stress is believed to result in the buildup of polymeric hydroperoxides that can then be quantified during the chemiluminescence experiment. This approach is similar to iodometric titration for peroxide level determination; however, chemiluminescence offers the potential for substantially greater sensitivity. (11,16) In particular, the need for cryo-grinding and the low sensitivity of the iodometric titration often makes the technique impractical. In the work reported here, we used the chemiluminescence signal produced in a nitrogen environment to track the build-up of polymeric hydroperoxides in naturally weathered samples.

Photoacoustic Spectroscopy 

IR spectroscopy has, in the past, been a very useful technique for the assessment of photodegradation in automotive coatings. These automotive paints are typically isocyanate or melamine crosslinked polymers and tend to degrade through photooxidation and hydrolysis in many cases. As the process of photodegradation proceeds, there is an oxidation of the polymer backbone, which can result in increases in the OH stretching region of the spectrum. This spectral change can be used to quantitatively or qualitatively assess the degree of photooxidation. The technique has also been employed to assess the fate of hindered amine light stabilizers as well as UV absorbers commonly used in automotive paint systems to improve durability. (17) In contrast, there has been very little reported work using IR to follow the degradation of waterborne acrylic architectural paints.

The useful technique to monitor the IR absorption bands in the paint film is photoacoustic spectroscopy (PAS). As compared to conventional IR, PAS lends itself to the analysis of paint films, as it does not require the sample to be removed from the substrate. In a PAS experiment, modulated light is absorbed by the sample and converted into heat by a nonradiative de-excitation process, causing periodic temperature fluctuations in the sample, mostly within the optical penetration depth. As this heat-induced thermal wave diffuses through the sample and then to the adjacent medium (usually helium or nitrogen) it induces pressure fluctuations, which are detected as an acoustic wave, and are subsequently converted to an electric signal. (18-20) PAS can also be used to get depth profile information about species present in the film and has been used to study surfactant migration through the film. (21,22) As employed in this study, PAS was used to follow growth in photooxidation peaks to assess whether this change would correlate with long-term gloss changes.

X-ray Photoelectron Spectroscopy 

X-ray photoelectron spectroscopy (XPS) is a widely used surface-specific analytical technique that provides both elemental compositions and oxidation state information from the first 100[Angstrom] of the surface of the paint. Quantitative surface compositional information is obtained through measuring photoelectron currents generated by irradiating a surface with X-rays. XPS has been used to characterize the first few nanometers of polymer and paint surfaces for the past 30 years. (23-27) The thinking here is that as photodegradation proceeds there will be an increase in the prevalence of pigment (alumina and silica coated Ti[O.sub.2]) relative to polymer at the surface. An increase in the titanium, silicon, and aluminum signal relative to the polymeric carbon signal would indicate that polymer was being eroded from the surface. The drawback of this technique is that it is so surface sensitive that migration of surface-active components, such as surfactants or the deposition of extraneous materials such as airborne oil and dirt, can complicate the analysis. The migration of surface-active ingredients, such as surfactants and defoamers in the paints, are particularly problematic for this technique.

ASSESSMENT PROTOCOL

To test the concept, a concerted attempt was made to assess the five analytical techniques as predictors of gloss retention in all-acrylic architectural paints. A series of paints using seven all-acrylic latex emulsion polymers were prepared in a standard gloss paint formulation. The latex emulsion polymers were chosen from a set of known commercial products, which spanned a wide range of gloss retention performance. The [T.sub.g]s of these latex emulsion polymers ranged from 0[degrees]C to about 35[degrees]C. The paints were placed on exterior exposure in Spring House, PA, and at the Q-Panel testing site in southern Florida. In this study, we found good correlation between Spring House and South Florida; so, to simplify the analysis, we focused on the Spring House results. Analytical measurements were made initially and at 2, 12, and 32 weeks on the samples exposed at Spring House, and the gloss was followed for 96 weeks. Correlations were examined between 12-week analytical test results and 96-week gloss results for the seven paints. Testing in an Atlas Ci5000 Xenon Weather-Ometer[R] (ASTM test protocol G26) was also included to assess its utility for predicting gloss durability of this series of paints.

EXPERIMENTAL DETAILS

Paint Formulations

Seven commercial waterborne emulsion polymers were selected which, based on previous history, are known to exemplify a broad range of gloss retention performance. Paints were prepared from these dispersion polymers following the formulation listed in Table 1. The paints all had a total nonvolatile content by volume of 33.6% and a pigment concentration by volume of 20%, which required an appropriate adjustment of the level of emulsion polymer and water in the formulation. To account for the different polymer [T.sub.g]s, the amount of coalescent (Texanol[TM]) was adjusted appropriately. All of the formulated paints were designed to have a minimum film formation temperature at or below 0[degrees]C. The amount of thickeners (Acrylsol[R] RM-2020NPR and Acrylsol[R] RM-8W) were adjusted for each paint to provide a KU viscosity of 85 to 95, and a high shear (ICI) viscosity of 1.2 to 1.5 Poise. After allowing the formulated paints to equilibrate for at least 24 hours, they were applied to aluminum panels so as to give a dry film thickness of ~25 [micro]m. After allowing seven days for the coatings to dry at 70[degrees]F and 50% relative humidity, the panels were placed on exposure at the Rohm and Haas Company Paint Quality Institute's test fence facility in Spring House on April, 4, 2001. A matching set was sent to the Q-Panel exposure testing facility in South Florida (near Miami), and a third set was placed in an Atlas Ci5000 Xenon Weather-Ometer (WOM). The WOM paints were exposed using the G26 protocol. For the Spring House and Q-Panel exposures, the panels were exposed in a south vertical direction. This orientation was chosen to minimize complications from dirt pick-up. Gloss readings were taken after 2, 4, 12, 24, 28, 32, 48, 56, 64, 72, 80, 88, and 96 weeks of exposure. Periodically, samples of the weathered paints were collected and stored in the dark for future analysis by the selected analytical techniques. Discs about 9.5 mm in diameter were punched out of the coated aluminum panels that had been exposed to natural weathering in Spring House. Nonexposed (control) samples were compared to the exposed samples. In this study we have focused on the 12-week Spring House samples for the analytical test methods and the 96-week Spring House samples for gloss measurements.

Gloss Measurements

Gloss readings were taken on the paints using a GL4520 micro-Tri Gloss meter (BYK-Gardner USA). Gloss readings of 20[degrees], 60[degrees], and 85[degrees] were obtained. For the sake of simplicity we have focused on 60[degrees] gloss in this study. The initial gloss values for all of the paints are given in Table 2. The 85[degrees] gloss readings were found to be too insensitive to changes that were occurring in the paints, whereas the 20[degrees] gloss tended to be too sensitive. There is some indication that 20[degrees] gloss changes at 12 weeks are a good indicator of 60[degrees] gloss changes at 96 weeks; however, it is felt that a larger data set is required to confirm this. The gloss data in this study is for samples exposed in the south vertical direction at the Paint Quality Institute's paint exposure station in Spring House. Corresponding data was also collected at the Q-Panel testing facility in South Florida, and in the Ci5000.

Scanning Probe Microscopy

Surface topography of the samples was monitored by scanning probe microscopy (SPM). Average surface roughness, [R.sub.a] ([R.sub.a] is defined as the arithmetic average of the absolute values of the surface height deviations measured from the mean plane, and thus takes into account any tilt that may be present in the sample), and peak-to-valley ([Z.sub.t]) measurements were collected using a Dimension 5000 Scanning Probe Microscope (Digital Instruments, Inc.) in tapping mode. The tapping or intermittent contact mode allows images of the surface to be obtained with minimal damage. In this mode, the tip is oscillated at one of its harmonic frequencies. Reflecting a laser beam onto a position sensitive detector monitors the amplitude. As the tip interacts with the surface, the amplitude changes slightly. However, near constant amplitude can be obtained by vertical movement of the head assembly, which contains the tip. Plotting these height changes as a function of tip position provides a profile of the surface.

Three 20 [micro]m x 20 [micro]m scans at 512 x 512 pixels were collected for each sample, and the average is reported. Compressed air was used to remove any free particles from the surface. This was required because, in initial experiments, image distortion occurred due to particles attaching to the tip. Average roughness (i.e., average deviation from the mean height) values are mainly used for comparison because the percent deviation was less for [R.sub.a] than for [Z.sub.t]. The deviation in peak-to-valley is more likely to be larger because a single data point has a greater effect on [Z.sub.t] than the roughness measurement, which is an average value.

[FIGURE 1 OMITTED]

[FIGURE 2 OMITTED]

Optical Profilometry

The optical profilometry technique was incorporated to increase the length scale in roughness measurements. The Wyko NT 1000 Optical Profiler (Veeco Instrument, Inc.) may be operated at a wide range of magnifications, varying the field of view from 45 [micro]m x 60 [micro]m to 3.8 mm x 5.0 mm. It was operated at lower magnifications to provide complimentary data to the SPM results. Data was recorded at 0.9 mm x 1.2 mm, 1.9 mm x 2.5 mm, and 3.8 mm x 5.0 mm fields of view. A single point was collected at the center of the sample for the lower two magnifications; whereas the sample center and an additional two points were collected for the 0.9 mm x 1.2 mm area. The average [R.sub.a] for this area was used to compare to the SPM results and gloss values. The reported values have incorporated tilt and curvature correction (required due to the nature of the "punched-out" samples). Free particles on the surface were removed with compressed air prior to analysis.

[FIGURE 3 OMITTED]

Chemiluminescence

Chemiluminescence measurements were conducted using an Atlas CL400. Chemiluminescence data was collected while samples were subjected to a temperature ramp under a nitrogen environment. The temperature was increased from room temperature to 150[degrees]C and then held constant at 150[degrees]C until the signal returned to baseline levels, usually within two to four hours. All the samples produced a well-defined chemiluminescence peak, with the signal increasing rapidly as the temperature ramped up and then decreasing gradually to near-baseline levels. For some samples, a full return to baseline did not occur, and the signal leveled off at 10-20% of the peak height and then showed no further decrease with time. Since the area under the chemiluminescence curve was taken as a measure of the peroxide concentration, samples for which the peak never decayed fully to baseline levels posed some difficulties in determining when to end the peak integration. However, we did observe that the shape of the chemiluminescence peak remained consistent for a given coating, regardless of exposure time. Thus, we could monitor percent increase in peroxide levels for a given coating by holding the integration limits constant for that coating. While we could not directly compare absolute peroxide levels between coatings, the rate of peroxide level increase could be compared from sample to sample.

Photoacoustic Spectra

Photoacoustic spectra were acquired on a Bio-Rad (Cambridge, MA) FTS6000 step-scan spectrometer equipped with a MTEC (Ames, IA) 300 Photoacoustic detector. Helium was used as the purge gas and as the coupling gas as this enhances the photoacoustic signal. The spectrometer was equipped with a water cooled ceramic mid-infrared source and a KBr crystal beam-splitter. The experimental parameters are summarized in Table 3.

X-ray Photoelectron Spectroscopy

XPS data were acquired using a Kratos Axis-Ultra instrument. Data was collected using monochromatic Al K[alpha] (1486 eV) X-rays operated at 350 W. Analysis areas were approximately 1 mm in diameter and at least three areas were interrogated per sample. Pass energies of 160 eV and 40 eV were used to collect survey and high-resolution spectra, respectively. All spectra were charge corrected by setting the C-C peak maximum to 284.8 eV.

RESULTS AND DISCUSSION

Gloss

The gloss results are given in Figure 1, and are shown as a percent of 60[degrees] gloss retained. As can be seen in Figure 1, there is a wide range of performance in the gloss retention of the paints, and several exhibit excellent gloss retention and actually show an increase in gloss as the film ages outside. This phenomenon is commonly seen in the early stages of weathering for many of the higher performing coatings. It has been attributed (28) to a densification of the paint film and a concomitant increase in refractive index.

Weather-Ometer

Figure 2 shows a plot of percent gloss loss after 1000 hr in the WOM versus percent gloss loss after 96 weeks of exposure at Spring House. As can be seen, the correlation between the WOM and the exterior exposure results is virtually nonexistent. What is most troubling is that the WOM predicts that some paints will do well and yet they fail badly, and that some will do poorly and yet they hold up fairly well. This data brings into doubt the use of the WOM for predicting gloss retention with this class of paints. It clearly cannot predict pronounced failures, and would potentially cause good paints to be rejected as well. It should be remembered that these results indicate that the G-26 testing is not predictive for these paints, and that it is entirely possible that a modification to the testing protocol might produce better results.

Surface Topography

SPM and optical profilometry are techniques which probe the surface topography and thus provide useful information on the smoothness or roughness of a paint film. As such, these techniques are related to determining the gloss of a coating. In past work, the surface roughness has been found to be related to gloss. (29) At this point there is no clear understanding of the relationship of length scale to effects on gloss. Given that increases in surface roughness are a major factor in reducing the gloss of a coating, it is anticipated that as paint weathers, the gloss would decrease and the surface roughness would increase. In the study, the percent change in [R.sub.a] [%[DELTA][R.sub.a] = ((Ra(12week)-Ra(initial))/Ra(initial))x100] rather than simply [DELTA][R.sub.a] is used to compare to percent gloss change after 96 weeks. This is to compensate for the range in surface roughness for the different paint systems. For example, the [R.sub.a] value for the SPM technique varied on the control samples from 19 nm for paint AA6 to 43 nm for paint AA1. The range for the optical technique was 135 nm for paint AA4 to 181 nm for paint AA5 on the same set of samples.

It might be expected that a correlation would exist between initial gloss readings and [R.sub.a]. It has been found that a moderate correlation exists between initial gloss readings and [R.sub.a] as measured using SPM; however, the correlation is quite weak between initial gloss readings and [R.sub.a] as measured by optical profilometry. In fact, [R.sub.a] as measured using SPM does not correlate well with [R.sub.a] as measured by optical profilometry. As displayed in Figure 3, this lack of correlation is observed in the 12-week %[DELTA][R.sub.a] values. Figure 3 compares the percent change in [R.sub.a] for the scanning probe and optical techniques to the percent change in gloss after 96 weeks. It should be noted that one would not anticipate the magnitude of the percent [R.sub.a] change to be equal between the two techniques or equal to the %[DELTA]G (i.e., percent change in gloss). It is expected that the changes in roughness values, whether determined by OP or SPM, would track changes in gloss.

[FIGURE 4 OMITTED]

In the SPM data, the 12-week %[DELTA][R.sub.a] is greatest for sample AA4 (at 21%), with little change in samples AA1, AA3, and AA7, and samples AA2 and AA6 actually decreased in roughness. An initial reduction in roughness with exposure is not necessarily unexpected considering the paint may continue to cure, and components not fully incorporated into the paint or surfactant blooming to the surface may be washed away. Simple ranking of the samples by %[DELTA][R.sub.a] values does not correspond to the ranking for percent gloss change after 96 weeks (with smaller values ranked highest). This is also the case for ranking the 12-week %[DELTA][R.sub.a] optical data. These values range from 487% for paint AA6 (due to blistering) to -12% for paint AA5. For certain samples (e.g., paints AA3 and AA4), there is excellent correlation between the %[DELTA][R.sub.a] values for the two techniques, and long-term performance. This trend is not observed across the sample set and, in fact, there is negative correlation between SPM, optical profilometry, and gloss data for some paints. For example, paint AA5 demonstrates good correlation between SPM values and long-term performance, but poor correlation to the optical data. The reverse is true for paint AA7. Some of this is due to length scale effects that will be discussed later.

[FIGURE 5 OMITTED]

Figure 4 presents the 12-week images from paints AA5 and AA7 for both the scanning probe and optical techniques. The black box on the optical profilometry images compares the 20 [micro]m x 20 [micro]m area imaged using the SPM technique to the area captured by optical profilometry. Twelve week %[DELTA][R.sub.a] values are displayed along with the [R.sub.a] and peak-to-valley measurements (i.e., [Z.sub.t]) for each image. Comparison of the SPM images demonstrates the relation of roughness values to surface texture. Paint AA5, which appears rougher, does have a higher [R.sub.a] value of 48 nm, while paint AA7 is smoother and has a surface roughness of 43 nm. These [R.sub.a] values are consistent with the observed surface texture when measured at the 20 [micro]m x 20 [micro]m scale. Optical profilometry was run at a lower magnification to reveal the [R.sub.a] at a length scale greater than could be obtained by the SPM. The area of the optical images presented in Figure 4 is approximately 940 [micro]m x 1240 [micro]m. Visual comparison of these images is consistent with the appearance and [R.sub.a] values from the corresponding SPM images. The roughness values for the optical images, however, are in conflict with their general appearance. That is, paint AA5 appears rougher than AA7, but its [R.sub.a] value is less. One factor driving up the [R.sub.a] value for paint AA7 is the density of pits.

The conflict between the [R.sub.a] values for the SPM and the optical profilometry data appears to arise from the different length scales of measurement. The SPM collects data from a relatively small area when compared to the optical profilometer. The small areas will not capture all the surface excursions that are greater than a few microns. The Z scale is limited to six microns in SPM, due to instrument limitations. The optical profilometer, on the other hand, is capable of capturing data over a much greater surface area and Z scale. These larger areas allow more details on surface changes due to pitting, pigment fall-out, and general changes in surface topography to be captured. The smaller length scale monitored by the SPM provides microstructure details such as degree of film formation and onset of micro cracks.

The results obtained from paints AA5 and AA7, as found in Figure 4, are particularly useful for assessing the utility of the two techniques. Paint AA5, by optical profilometry, showed a relatively smooth surface with a few 10 [micro]m peaks. There did not appear to be any major flaws in the surface (i.e., pits, voids, etc.). A small decrease in surface roughness was detected in the measurements over 12 weeks. The 11.7% decrease in %[DELTA][R.sub.a] is near the statistical difference between the samples. A few larger surface excursions in the sampling area may be responsible for these statistical differences. The %[DELTA][R.sub.a] recorded by SPM, 20.5%, indicates an increase in the roughness of the surface that follows the trend of decreasing gloss (i.e., %[DELTA]G = 31.2) in the sample. The SPM will follow changes in the surface roughness that is occurring at the smaller length scale. Paint AA7 revealed the issue of changes in the surface topography the SPM may have missed. Images collected by the optical profilometer showed pitting on the surface and exhibited a %[DELTA][R.sub.a] of 32.8, which was consistent with the paint's poor long-term performance (as illustrated by the 96-week %[DELTA]G of 29.5). These pits are very difficult to measure by the SPM, because the area of the pits is greater than the analysis area. Thus, the resulting %[DELTA][R.sub.a] of 1.9 from SPM does not predict the poor performance of this sample. The smooth areas imaged will provide details on how the smaller features are changing with exposure, yet miss the larger features captured by the optical profilometry. The SPM analysis may suggest the potential for the paint if these larger defects are controlled.

To summarize, it is important to review the images collected by both the optical profilometer and the SPM. The different length scales sampled in the images by the two techniques are sensitive to changes in different surface features. Depending on the mechanism of failure, one technique may correlate to changes in gloss better than another. Using both techniques develops a greater understanding as to the mechanism of failure and how the surface is responding to weathering. At this point it is fair to say that neither of these techniques looks very promising as a general broad screening tool for accelerated gloss retention.

[FIGURE 6 OMITTED]

[FIGURE 7 OMITTED]

Chemiluminescence

Figure 5 plots the chemiluminescence peak areas after 0, 12, and 32 weeks of weathering. As can be seen for coatings AA2 and AA5, the largest peak intensity occurs in the control (preweathered) samples. Since the chemiluminescence experiments were conducted under an inert environment, the peak intensity should reflect the level of peroxides present in the sample. These high initial signal levels may arise from residual peroxides left over from the synthesis process. For samples AA2 and AA5, the chemiluminescence signal levels dropped sharply by the next time the samples were scanned--after 12 weeks of exterior exposure--during which time the peroxides were presumably degraded. Comparison of the 12- and 32-week data reveals that for all coatings, the chemiluminescence peak area increases with increasing exposure time, presumably due to build up of polymeric hydroperoxides formed from bond scission during the weathering process.

Since the initial signal levels may be artificially high due to residual peroxides from synthesis, the chemiluminescence increase from 12 to 32 weeks might best reflect oxidative degradation due to the exterior exposure. The percent chemiluminescence increase over the 12-32 week time span versus the ultimate extent of gloss retention at 96 weeks is plotted in Figure 6. A strong correlation between growth in chemiluminescence during the initial stages of the weathering process and ultimate gloss retention should result in a negative-trending slope. A negative-trending slope is observed, but the weak correlation ([R.sup.2] = 0.5) indicates that the technique's ability to predict gloss performance across the range of emulsion polymers is poor.

It should be remembered that the advantage of chemiluminescence as an analytical screening technique is its extreme sensitivity. With modern photon-counting detectors, peroxide levels in the femtomolar range can be readily detected. However, the technique offers extremely limited specificity and generally does not distinguish between the various types of alkyl peroxides. Some groups have attempted to gain additional information by spectrally resolving the chemiluminescence signal. (4) When we used broadband filters in an attempt to identify the wavelength region of chemiluminescence, we found that all samples emitted primarily within the broad range of 450-520 nm. For complex formulations of paint samples, there are a variety of kinetic processes occurring during weathering, and different types of polymeric peroxides form and decay at different rates. Since all these peroxides can contribute to the signal level, the potential of chemiluminescence seems most likely to be realized when making comparisons among similar chemistries in highly controlled formulations. At this point the chemiluminescence does not appear to be useful as a rapid screening tool.

PAS

Figure 7 shows the PAS-FTIR spectra for a paint based on AA3. It compares the original unweathered paint to a paint exposed for 250 hr in the WOM to the same paint weathered for 12 weeks outside. Examination of the figure shows that WOM weathering leads to chemical compositional changes, as indicated by the appearance of new bands and changes in intensities of certain bands in the spectra; however the changes are very slight. These changes include an increase in the NH/OH bands in the 3600-3050 [cm.sup.-1] region, which is most likely due to polymer hydrolysis, and a slight change at around ~1800 [cm.sup.-1], which could be explained by the formation of anhydrides. These changes are consistent with known photooxidation reactions of acrylic polymers. Other studies (30) have seen similar results in the IR spectra. Unfortunately, these spectral changes were not detectable in the naturally weathered samples. Figure 8 compares the spectra for paint AA3. It compares the original unweathered paint to a paint exposed for 96 weeks outside. Even after 96 weeks the PAS technique is not capable of detecting significant changes.

[FIGURE 8 OMITTED]

[FIGURE 9 OMITTED]

[FIGURE 10 OMITTED]

Following treatments by Gerlock et al., an IR Weathering Factor (IRWF) was defined based on the ratio of the OH and NH to the CH adsorption bands. IRWF (at time t) = [OH & NH]/[CH].

The weathering factor (photooxidation index) for each coating was determined initially and at 12 weeks of natural exposure. The ratio of IRWF (12 weeks) to IRWF (initial) is defined as %IRWF and is a measure of the percent increase in oxidation products. It might be expected that high %IRWF values would indicate higher photooxidation rates and, thus, higher long-term gloss loss. This is of course complicated by the low degree of change seen in IR spectra for the naturally weathered paints.

Figure 9 shows a plot of the percentage 60[degrees] gloss loss at 96 weeks against the %IRFW for paints AA1 through AA7. Examination of Figure 9 indicates poor correlation of the gloss changes with %IRFW. The lack of correlation is most likely due to the fact that the change in the IR signal is too small and cannot be accurately measured.

Although the IR technique is sensitive to chemical compositional changes in a paint film, the results of this study strongly suggest that the chemical changes observed in these waterborne acrylic paints are too small to allow for general correlation with percent gloss loss at early stages, thus limiting its utility for longer-term predictive purpose. Ultimately, the degradation mechanisms are very complicated for outdoor exposures due to a myriad of previously identified factors, and, although the use of multivariate tools can be invoked as a way of improving the IR approach, it is difficult to implement because the approach requires similar samples be used for calibration. (31) It is important to point out that the focus of this work has been on the utility of different techniques, such as IR for rapid assessment; and, while PAS does not show promise in this specific case, this does not mean it is not useful for assessing long-term chemical changes in naturally weathered paints and identification of degradation chemistries.

XPS

A typical XPS survey spectrum obtained from a 12-week weathered paint is displayed in Figure 10. The carbon and oxygen are associated with the polymer. There is a small contribution to the oxygen intensity from the different inorganic oxides, such as Ti[O.sub.2], found at the air interface. The detection of the titanium and aluminum reveal the presence of pigment particles at the air interface. Although detectable, the relative pigment concentrations are quite low. The detection of sodium and sulfur suggest the presence of an anionic surfactant at the surface of the paint. The binding energy of the detected silicon is consistent with silica or silicates coming from the passivation coating on the Ti[O.sub.2] or possibly the silica particles in the defoamer.

[FIGURE 11 OMITTED]

The pigment-to-carbon ratios versus exposure time were used to assess the degree of polymer erosion and subsequent exposure of the pigment from the paint's surface. As a polymer erodes from the surface, the surface should roughen and, hence, reduce the gloss of the paint. The titanium and aluminum concentrations reflect the pigment surface concentrations while the total carbon concentration was used to monitor the polymer surface concentrations. Figure 11 contains the pigment-to-carbon ratios at 0, 2, and 12 weeks of natural exposure. In general, there is an increase in the pigment-to-carbon ratio after only two weeks of exposure. This increase is probably due to the washing away of low molecular weight species and surface-active agents such as surfactants and defoamers from the paint/air interface. For most of the paints, there is another increase in the ratio after 12 weeks of exposure. This increase is most likely due to degradation and erosion of polymeric carbon; however, it is difficult to definitively assign these changes to specific paint ingredients. No significant changes in the carbon or oxygen spectra were detected, suggesting that oxidized polymer species are not prevalent. A very plausible explanation is that photodegradation is leading to degradation of the acrylic polymer into volatile, water-soluble, or easily eroded species that are eliminated from the paint surface, thus leaving a paint surface that would look very similar to the nondegraded acrylic polymer.

[FIGURE 12 OMITTED]

In order to account for the large amount of easily removed material from the paint surface, the XPS analysis focused on changes in the pigment to carbon ratios occurring between 2 and 12 weeks. The correlation between long-term gloss retention and percent change in the pigment to carbon concentration ratio between 2 weeks and 12 weeks of exposure is plotted in Figure 12. One would anticipate the largest percent change in the pigment-to-carbon ratio would correspond to the paints with the worst gloss retention, thus resulting in a negative slope in the data. There is a correlation ([R.sup.2] = 0.731) between the pigment-to-carbon ratio and the long-term gloss retention of the acrylic paints studied here.

These results would indicate that, at least for all acrylic polymers, the use of XPS to follow the onset of polymer erosion appears to warrant further work to assess the breadth to which it might be a useful screening tool for predicting long-term gloss retention. Of particular interest would be a broader study which looks at crosslinked and uncrosslinked paints, as this should alter the kinetics of the weathering processes. The preliminary success of XPS is important because few, if any, analytical techniques have proven to be useful for all these waterborne all-acrylic paints. The lack of any useful analytical technique is probably due to the fact that photodegradation of acrylic polymers results in the removal of the degraded fragments of the polymer from the coating. In particular, studies by Gupta (32) have shown that one of the major byproducts of butyl acrylate photodegradation is butene, which would most certainly volatilize from the film.

CONCLUSIONS

The work presented in this article has focused on the concept of using natural weathering coupled with sensitive analytical techniques as a possible early indicator of gloss retention for waterborne acrylic architectural decorative paints. We surveyed a number of possible techniques, some of which were not useful, while others were and are worthy of a more in-depth study. Correlations between the analytical results and gloss changes after 96 weeks suggest that XPS can be effectively used to observe the onset of polymer degradation, predicting gloss loss far earlier (12 weeks versus 96 weeks) than found by conventional measurement of gloss loss alone. An extension of this work for following other paint durability properties such as loss or fading of color is certainly in order at this time.

In comparison, the experiments with surface topology, chemiluminescence, and PAS-FTIR all showed very poor to nonexistent correlation with long-term gloss loss. The WOM results were equally disappointing. The advantage of XPS over these other techniques most likely stems from its high surface sensitivity and, thus, its ability to detect small compositional changes at the surface. It is important to note that the XPS is so surface sensitive that the issue of surface-active materials blooming to the surface in the first few days of film drying is a complicating factor. These interfering materials seemed to disappear within two weeks of natural weathering so that the two-week weathered sample can be used as an initial baseline. It was found that the correlations are actually poor if this effect is not taken into account.

The poor performance of each of the other techniques is most likely tied to unique issues with each technique. Size domain issues and a general poor understanding of how this relates to gloss complicate the surface topology measurements. (33) It is quite possible that a better treatment of the data in conjunction with a better understanding of how domain size affects gloss could yield a better correlation. The chemiluminescence data is probably affected by the fact that it is a measure of peroxide formation and not a measure of polymer degradation, and thus does not effectively probe the actual degradation of the polymer. Since there are many sources for peroxides, including polymerization initiators, it is difficult to establish a good baseline reading. The PAS-FTIR data are most certainly complicated by the fact that for acrylic polymers the amount of change in the FTIR signal is not very large after 12 weeks of natural exposure, thus making it very difficult to assess any changes that have occurred. An attempt was made to see if using these analytical techniques on the 32-week weathered samples would lead to better correlations and improved predictive power, but at this point the results indicate that this did not offer much improvement and was not pursued further.

The fact that short-term surface measurements do not correlate well with the long-term gloss measurements is probably not surprising, as a good correlation would require that the kinetics of the degradation process do not change dramatically for each of the paints, and as can be seen from Figure 1 where the paints do not display a consistent rate of gloss loss. This makes the results for the XPS even more striking, as the XPS detects the early effects of a degradation mechanism that appears to hold out for 96 weeks.

To summarize, the data presented shows that XPS, with appropriate accounting for surfactant blooming, holds significant promise as a technique for assessing polymer erosion in naturally weathered paints. Further work needs to be done to test this hypothesis, as well as to extend the XPS technique to other paint properties, most notably color retention and fading.

ACKNOWLEDGMENTS

The authors would like to thank the Rohm and Haas Company for supporting this work. The following people have contributed to the design and implementation of this study: Dave Westmoreland, George Daisey, Tod Zwolak, Michaeleen Pacholski, Sanjay Chaturvedi, Michael Bender, Al Prisco, Roy Robertson, Rebecca Smith, John Hook, Susan Fitzwater, Brough Richey, and Patrick Clark.

References

(1) Martin, J.W., Saunders, S.C., Floyd, F.L., and Wineburg, J.P., "Methodologies for Predicting the Service Lives of Coating Systems," Federation Series on Coatings Technology, Brezinski, D. and Miranda, T.J., (Eds.), Federation of Societies for Coatings Technology, Blue Bell, PA, 1996.

(2) (a) Bauer, D.R., "Perspectives on the Weatherability Testing of Automotive Coatings," J. COAT. TECHNOL., 74, No. 924, 33 (2002); (b) Gerlock, J.L., Smith, C.A., Cooper, V.A., Dusbiber, T.G., and Weber, W.H., "On the Use of Fourier Transform Infrared Spectroscopy and Ultraviolet Spectroscopy to Assess the Weathering Performance of Isolated Clearcoats from Different Chemical Families," Polym. Degrad. Stab., 62, p. 225 (1998); (c) Bauer, D.R., Paputa-Peck, M.C., and Carter, R.O. III, "Evaluation of Accelerated Weathering Tests for a Polyester-Urethane Coating Using Photoacoustic Infrared Spectroscopy," J. COAT. TECHNOL., 59, No. 755, 103 (1987); (d) Mielewski, D.F., Bauer, D.R., and Gerlock, J.L., "The Role of Hydroperoxides in the Photo-Oxidation of Crosslinked Polymer Coatings," Polym. Degrad. Stab., 33, p. 93 (1991); (e) Bauer, D.R., "Predicting In-Service Weatherability of Automotive Coatings: A New Approach," J. COAT. TECHNOL., 69, No. 864, 85 (1997); (f) Nichols, M.E., Gerlock, J.L., and Smith, C.A., "Rates of Photooxidation Induced Crosslinking and Chain Scission in Thermoset Polymer Coatings," Polym. Degrad. Stab., 56, p. 81 (1997); (g) Gerlock, J.L., Bauer, D.R., Briggs, L.M., and Dickie, R.A., "A Rapid Method of Predicting Coating Durability Using Electron Spin Resonance," J. Polym. Chem., 28, p. 447 (1984); (h) Bauer, D.R., "Predicting Coating Durability Using Chemical Methods," Prog. Org. Coat., 23, p. 105 (1993); (i) Wernstahl, K.M. and Carlsson, B., "Durability Assessment of Automotive Coatings--Design and Evaluation of Accelerated Tests," J. COAT. TECHNOL., 69, No. 865, p. 69 (1997).

(3) (a) Bierwagen, G.P., He, L., Li, J., Ellingson, L., and Tallman, D.E., "Studies of a New Accelerated Evaluation Method for Coating Corrosion Resistance--Thermal Cycling Testing," Prog. Org. Coat., 39, 67 (2000); (b) Mills, D.J., Bierwagen, G.P., Skerry, B., and Tallman, D., "Investigation of Anticorrosive Coatings by the Electrochemical Noise Method," Mater. Performance, 34, p. 33 (1995).

(4) (a) Bierwagen, G.P., Twite, R., Chen, G., and Tallman, D.E., "Atomic Force Microscopy, Scanning Electron Microscopy and Electrochemical Characterization of Al Alloys, Conversion Coatings, and Primers Used for Aircraft," Prog. Org. Coat., 32, p. 25 (1997); (b) Gerharz, B., Kuropka, R., Petri H., and Butt, H.-J., "Investigation of Latex Particle Morphology and Surface Structure of the Corresponding Coatings by Atomic Force Microscopy," Prog. Org. Coat., 32, p. 75 (1997); (c) Gilicinski, A.G. and Hegedus, C.R., "New Applications in Studies of Waterborne Coatings by Atomic Force Microscopy," Prog. Org. Coat., 32, p. 81 (1997); (d) Joanicot, M., Granier, V., and Wong, K., "Structure of Polymer Within the Coating: An Atomic Force Microscopy and Small Angle Neutrons Scattering Study," Prog. Org. Coat., 32, p. 109 (1997); (e) Keddie, J.L., "Film Formation of Latex," Mater. Sci. Eng. R-Reports, 21, p. 101 (1997); (f) Ho, C.C., and Khew, M.C., "Low Glass Transition Temperature ([T.sub.g]) Rubber Latex Film Formation Studied by Atomic Force Microscopy," Langmuir, 16, p. 2436 (2000); (g) Steward, P.A., Hearn, J., and Wilkinson, M.C., "An Overview of Polymer Latex Film Formation and Properties," Adv. Colloid Interface Sci., 86, p. 195, (2000); (h) Tang, J.S., Daniels, E.S., Dimonie, V.L., Vratsanos, M.S., Klein, A., and El-Aasser, M.S., "Mechanical Properties of Films Prepared from Model High-Glass-Transition-Temperature/Low-Glass-Transition-Temperature Latex Blends," J. Appl. Polym. Sci., 86, p. 2788 (2002); (i) Mallegol, J., Dupont, O., and Keddie, J.L., "Morphology and Elasticity of Waterborne Acrylic Pressure-Sensitive Adhesives Investigated with Atomic Force Microscopy," J. Adhes. Sci. Technol., 17, p. 243 (2003); (j) Winnik, M.A., "Latex Film Formation," Curr. Opin. Colloid Interface Sci., 2, p. 192 (1997).

(5) Mendenhall, G.D., "Analytical Applications of Chemiluminescence," Angew. Chem. Int. Ed. Engl., 89(4), p. 220 (1977).

(6) George, G.A., Egglestone, G.T., and Riddell, S.Z., "Chemiluminescence Studies of the Degradation and Stabilization of Polymers," Polym. Engin. Sci., 23, p. 412 (1983).

(7) Zlatkevich, L., "Chemiluminescence Apparatus and Method for Studying Thermal Oxidative Stability of Polymers," Polym. Eng. Sci., 24, p. 1421 (1984).

(8) Lacey, D.J. and Dudler, V., "Chemiluminescence from Polypropylene. Part 2. The Emission Wavelengths During Prolonged Oxidation," Polym. Degrad. Stab., 51, p. 109 (1996).

(9) Dudler, V., Lacey, D.J., and Krohnke, C., "Chemiluminescence from Polypropylene. Part 3. Application to the Study of Antioxidant Effectiveness," Polym. Degrad. Stab., 51, p. 115 (1996).

(10) Billingham, N.C., Then, E.T.H., and Gijsman, P.J., "Chemiluminescence from Peroxides in Polypropylene. Part I. Relation of Luminescence to Peroxide Content," Polym. Degrad. Stab., 34, p. 263 (1991).

(11) Memetea, T., Vuluga, Z., and Hagiopol, C., "The Thermooxidative Degradation of Acrylonitrile-Butadiene-Styrene Copolymers During Processing as Studied by Chemiluminescence," J. Appl. Polym. Sci., 45, p. 1229 (1992).

(12) Parikh, S.S. and Zlatkevich, L., "Chemiluminescence in Evaluating the Thermal Oxidative Stability of Acrylonitrile-Butadiene-Styrene (ABS) Copolymers. II. ABS/Polycarbonate (PC) Blends," J. Environ. Polym. Degrad., 1, p. 275 (1993).

(13) Parikh, S.S. and Zlatkevich, L., "Chemiluminescence in Evaluating Thermal Oxidative Stability of Acrylonitrile-Butadiene-Styrene (ABS) Copolymers. III. Emulsion-Grade Products," J. Environ. Polym. Degrad., 3, p. 199 (1995).

(14) Dudler, V., Th. Bolle, T., and Rytz, G., "Use of Chemiluminescence to the Study of Photostability of Automotive Coatings," Polym. Degrad. Stab., 60, p. 351 (1998).

(15) Johnson, B.W. and McIntyre, R., "Analysis of Test Methods for UV Durability Predictions of Polymer Coatings," Prog. Org. Coat., 27, p. 95 (1996).

(16) Kron, A., Stenberg, B., Reitberger, T., and Billingham, N.C., "Chemiluminescence from Oxidation of Polypropylene: Correlation with Peroxide Concentration," Polym. Degrad. Stab., 53, p. 119 (1996).

(17) Gerlock, J.L., Kucherov, A.V., and Nichols, M.E., "On the Combined Use of UVA, HALS, Photooxidation, and Fracture Energy Measurements to Anticipate the Long-Term Weathering Performance of Clearcoat/Basecoat Automotive Paint Systems," J. COAT. TECHNOL., 73, No. 918, 45 (2001).

(18) Rosencwaig, A., Photoacoustics and Photoacoustic Spectroscopy, Wiley, New York, p. 93, 1980.

(19) Jones, R.W. and McClelland, J.F., "Quantitative Depth Profiling of Layered Samples Using Phase-Modulation FT-IR Photoacoustic Spectroscopy," Appl. Spectrosc., 50, 1258 (1996).

(20) Jiang, E.Y., Palmer, R.A., Barr, N.E., and Morosoff, N., "Phase-Resolved Depth Profiling of Thin-Layered Plasma Polymer Films by Step-Scan Fourier Transform Infrared Photoacoustic Spectroscopy," Appl. Spectrosc., 51, p. 1238 (1997).

(21) Evanson, K.W. and Urban, M.W., "Surface and Interfacial FTIR Spectroscopic Studies of Latexes. I. Surfactant-Copolymer Interactions," J. Appl. Polym. Sci., 42, p. 2287 (1991).

(22) Urban, M.W., "Surfactants in Latex Films; Interactions with Latex Components and Quantitative Analysis Using ATR and Step-Scan PAS FT-IR Spectroscopy," Prog. Org. Coat., 32, p. 215 (1997).

(23) Clark, D.T. and Thomas, H.R. "Applications of XPS to Polymer Chemistry. XI. Core and Valence Energy Levels of a Series of Polymethacrylates," J. Polym. Sci., 14, p. 1701 (1976).

(24) Briggs, D. and Seah, M.P. (Eds.), Practical Surface Analysis: Auger and X-ray Photoelectron Spectroscopy, 2nd Ed., Vol. 1, John Wiley & Sons, 1983.

(25) Bell, R.T. "Failure Analysis of Coating," Surf. Coat. Australia, 21(8), p. 8 (1984).

(26) Turner, N. and Schreifels, J.A., "Surface Analysis: X-ray Photoelectron Spectroscopy and Auger Electron Spectroscopy," Anal. Chem., 72, p. 99R (2000).

(27) Johansson, L.S., "Analyzing Coated Powders with XPS," Surf. Inter. Anal., 17, p. 663 (1991).

(28) Colling, J.H. and Dunderdale, J., "The Durability of Paint Films Containing Titanium Dioxide--Contraction, Erosion, and Clear Layer Theories," Prog. Org. Coat., 9, p. 47 (1981).

(29) Willmouth, F.M., in Optical Properties of Polymers, Meeten, G.H. (Ed.), Elsevier Applied Science Publishers, New York, 1986.

(30) Nguyen, T., Martin, J., Bird, E., and Embree, N., "Relating Laboratory and Outdoor Exposure of Coatings III. Effect of Relative Humidity on Moisture-Enhanced Photolysis of Acrylic-Melamine Coatings," Polym. Degrad. Stab., 77 p. 1 (2002).

(31) Gerlock, J.L., Smith, C.A., Cooper, V.A., Dusbiber, T.G., and Weber, W.H., "On the Use of Fourier Transform Infrared Spectroscopy and Ultraviolet Spectroscopy to Assess the Weathering Performance of Isolated Clearcoats from Different Chemical Families," Polym. Degrad. Stab., 62(2), 225 (1998).

(32) Liang, R.H., Tsay, F-D., and Gupta, A., "Photodegradation of Poly(n-butyl acrylate). Photochemical Processes in Polymeric Systems. 8," Macromolecules, 15, 974 (1982).

(33) Alexander-Katz, R. and Barrera, R.G., "Surface Correlation Effects on Gloss," J. Polym. Sci., Part B: Polym. Phys., 36(8), 1321 (1998).

Matthew S. Gebhard, Michael B. Clark, Jr., Kenneth F. Willey, Robert F. Antrim, Francis V. Acholla, and John B. Newman -- Rohm and Haas Company*

Kristin Weidemaier -- BD Technologies ([dagger])

* Research Laboratories, 727 Norristown Rd., Spring House, PA, 19477-0904.

([dagger]) 21 Davis Dr., Research Triangle Park, NC 27709.

Table 1 -- Paint Formulation Used to Prepare Paints Using the Seven
Emulsion Polymers

Material                                         Weight in Grams

Cowles Grind
Propylene Glycol                                  35
Tamol[R] 731A (Rohm and Haas Company)              7.2
Byk 022 (BYK Chemie)                               4
Ti-Pure[R] R-706 (Dupont)                        225
Acrysol[R] RM-2020NPR (Rohm and Haas Company)     15
Kathon[R] LX (Rohm and Haas Company)               1.5

Let Down
AA-1                                             565
Texanol[R] (Eastman Chemical Company)             36
Byk 022 (BYK Chemie)                               2
Triton[R] GR-7M (Dow Chemical Company)             2
Acrysol[R] RM-2020NPR (Rohm and Haas Company)      5
Acrysol[R] RM-8W (Rohm and Haas Company)          --
Troysan[R] P-20T (Troy Chemical Company)          12
Ammonia (28% in water)                             1
Water                                            112

Table 2 -- Initial 60[degrees] Gloss Values for the Paints Based on the
Seven Latex Emulsion Polymers

Latex Emulsion Polymer  Initial 60[degrees] Gloss

AA1                     78
AA2                     80
AA3                     76
AA4                     78
AA5                     72
AA6                     83
AA7                     73

Table 3 -- PAS-FTIR Parameters

Mode                                Rapid Scan

Modulation frequency (scan speed),  2.5 kHz (depths of 5-20 [micro]m)
wavelength dependent probing depth  and 40 kHz (depths of 1-5 [micro]m)
depended (4000-400 [cm.sup.-1])
Scan filter                         1.2 (for 2.5 kHz) and 30 (for 40
                                    kHz)
UDR                                 2.0
Resolution                          8 [cm.sup.-1]
Scans                               128 and 500 for 40 kHz
Source                              SRC1
Scan gain range radius              40 points
Effective laser wavenumber (ELWN)   15800.8235
Signal                              Auto subtract average interferogram
                                    value
Scan co-add                         4
Background                          Carbon black-single beam
Scan aperture                       Open
Scan nominal range                  6000-400 [cm.sup.-1]
Scan range                          4000-400 [cm.sup.-1]
Scan delay                          0
MTEC 300 gain set                   8/9