Long-term weathering behavior of UV-curable clearcoats: depth profiling of photooxidation, UVA, and HALS distributions.

The long-term weathering performance of two UV-curable clearcoat systems was studied using in-plane microtomy in combination with infrared spectroscopy, UV spectroscopy, and ESR spectroscopy. Oxygen transport characteristics were also studied using the half-time method. The photooxidation versus depth profile was highly dependant on the presence of hindered amine light stabilizers (HALS) for both coating systems. Ultraviolet light absorbers (UVA) had little effect on the photooxidation profile. A photooxidation gradient was formed in both clearcoats due to a reduction in oxygen solubility when compared to standard thermoset clearcoats. This gradient was only seen in formulations not containing HALS.

Keywords: UVA/HALS, FTIR, ATR, UV spectroscopy, photodegradation, UV, EB, radiation cure, weatherability

**********

Automotive clearcoats continue to evolve due to a variety of driving forces that include environmental regulations, customer performance demands, and cost pressures. High-solids solventborne 1K and 2K clearcoats are currently the industry standards in North America. However, super high-solids, powder, waterborne, and slurry clearcoats either have been recently introduced or are in advanced development. Each of these clearcoats offers reduced VOCs in comparison to conventional thermoset clearcoats, but all come with performance or cost compromises as well.

Another class of clearcoats being developed, but not yet commercial for automobile bodies, is UV-curable clearcoats. UV-curable clearcoats typically crosslink via free radical polymerization upon exposure to UV radiation. These coating systems are typically mixtures of oligomers, monomers, solvents, one or more photoinitiator (PI), and other additives for weatherability enhancement, including ultraviolet light absorbers (UVA) and hindered amine light stabilizers (HALS). (1) UV-curable coatings are well established in such industries as telecommunications, wood finishing, and flooring manufacturing, but have only recently been seriously considered by the automotive industry for body paint applications. These coatings are of major interest for a variety of reasons, the most widely discussed of which is the potential for outstanding scratch resistance. This improved scratch resistance is mainly due to the extremely high crosslink density that can be achieved by UV-curable coatings. (2,3) In addition, UV-curable clearcoats have the potential to offer lower VOCs, improved process efficiency, shorter curing times, and lower energy consumption. These potential process advantages must be weighed against issues associated with reparability, adhesion, yellowing, and curing shadowed areas on highly contoured objects. One solution that addresses the shadowing issue is to utilize both thermal and UV curing to process the clearcoat. These systems, known as dual cure systems, allow the partial thermal cure of shadowed regions while giving the properties of full UV-cure systems in unshadowed regions.

Because UV-curable coatings have not been used extensively in outdoor environments, their long-term weathering behavior is largely unknown. The research to date, mainly by Decker, (4-7) Valet, (8) and the results presented in previously authored work, (9) has shown that the addition of UVAs, and particularly HALS, significantly improve their weathering performance.

In previous work, Fourier Transform Infrared Spectroscopy (FTIR) experiments showed that the addition of HALS greatly reduced the amount of photooxidation during the accelerated weathering testing of two urethane acrylate clearcoats. In the absence of HALS, the clearcoats photooxidized very rapidly. UV spectroscopy results showed that the clearcoat's UVA loss rate was measurably higher in the absence of HALS. Fracture energy results showed that both clearcoats embrittled during weathering, with the HALS-containing formulations embrittling more slowly than the HALS-free formulations. However, Raman spectroscopy showed that the embrittlement was more closely tied to the rate at which any residual acrylate double bonds were consumed during weathering than to the rate of photooxidation.

Techniques now exist that allow the creation of photooxidation, UV absorbance, and active HALS depth profiles utilizing in-plane microtomy techniques. These measurements have a resolution of 5 [micro]m and allow for a more comprehensive look at the clearcoats that were studied in previous work. (9) In this work, we report on the photooxidation, UVA, and HALS longevity as a function of depth from the coating surface.

EXPERIMENTAL

Materials

Two urethane acrylate clearcoat systems were used in this study. Clearcoat A was a monocure clearcoat formulated to cure using only UV light. Clearcoat B was a dual cure clearcoat formulated to cure both thermally and by exposure to UV light. The effects of stabilizer additives were studied by investigating four different formulations of both clearcoats containing: HALS and UVA, HALS only, UVA only, and no HALS or UVA. The formulations containing both HALS and UVA are referred to as fully stabilized, while the formulations with no HALS or UVA are referred to as no additive. Details of the formulations are summarized in Table 1.

Clearcoat A contained a mixture of short and long wavelength PIs. The long wavelength PI possessed an absorption peak at 300 nm and an absorption tail that extended out to ~450 nm. The short wavelength PI possessed a peak absorbance at ~240 nm with a long wavelength tail that extended no further than 380 nm. The UVA used in Clearcoat A was a triphenyltriazine type, whose absorbance tail ended at ~390 nm, giving a window for absorption of long wavelength UV light by the PI that absorbs at wavelengths >390 nm. The HALS used was typical of those found in automotive clearcoats. Clearcoat B contained only one PI possessing two absorption peaks, one at ~260 nm and the other at ~210 nm, as well as an absorption tail that ended by 325 nm. The UVA used was a triphenyltriazine type that absorbed out to ~390 nm. The UV spectra of the UVA and PIs used in both clearcoats are shown in Figure 1. The HALS used was again typical of those used in automotive clearcoats. Unlike Clearcoat A, Clearcoat B was a two-component mixture, which was mixed prior to application and curing.

[FIGURE 1 OMITTED]

Curing Conditions

The clearcoat was applied to panels coated with a standard automotive primer. These coated panels were used for in-plane microtomy experiments, as well as PAS-FTIR measurements. Prior to clearcoat application, the panels were hand sanded and solvent wiped to improve adhesion of the clearcoat to the primer. Clearcoat was then applied with a Bird applicator, which produced dry film builds of ~40-50 [micro]m. Panels were then cut to a size of 8 cm X 14 cm with a band saw.

Clearcoat A was cured using a two-step process. After clearcoat application, the sample was heated to 70[degrees]C for 5 min to flash off excess solvent. The sample was then cured using an F300S Ultraviolet Lamp System, containing a microwave-powered H-class bulb (Fusion UV Systems, Inc.). An LC-6 conveyor system (Fusion UV Systems, Inc.) was used to transport the samples under the UV bulb for curing. Clearcoat A was cured by passing the sample through the UV system one time at a belt speed of 3 m/min for a total of ~3 J/[cm.sup.2] of UV energy.

Clearcoat B was cured using a three-step process. After clearcoat application, the sample was heated to 60[degrees]C for 10 min to flash off any solvent in the clearcoat. The sample was then passed through the same F300S Ultraviolet Lamp System three times at a belt speed of 11 m/min for a total of ~2.5 J/[cm.sup.2] of UV energy. The final step involved thermally curing the sample at 95[degrees]C for 12 min.

Free films of Clearcoat A (no additives) were made by applying the clearcoat to a glass panel using a Bird applicator that produced dry film builds of ~70-80 [micro]m. After curing, the system was heated to 120[degrees]C to relax any residual stress that can cause curling after the film is removed from the glass panel. Film removal was accomplished by using a flat, one-sided razor blade to lift the film off the glass. Free films were then placed in 35 mm slide holders for accelerated weathering.

Weathering/Heat Aging

Accelerated weathering of the two clearcoat systems was carried out in an Atlas Xenon arc Weather-Ometer[R] running test method SAE J1960 Jun89 with borosilicate inner and outer filters at 0.55 W/[m.sup.2] @ 340 nm irradiance. Panels were exposed for 970 hr and 1868 hr. Free films were exposed in a light only weatherometer (no dark or wet cycles) with the same borosilicate inner and outer filters for 1000 hr.

Samples for PAS-FTIR analysis were heat-aged in a forced air oven at 60[degrees]C for 1656 hr.

In-Plane Microtomy

Sample slices measuring 2.5 cm X 3 cm X ~5 [micro]m were cut using a Polycut E microtome (Leica SM2500 E, Cambridge Instruments GmbH). This was done by first placing a 20 cm X 40 cm X 4 cm block of polypropylene in the microtome's sample holder and then microtoming the surface of the block flat using a tungsten carbide blade with a 50[degrees] cutting angle. Next, a ~3 cm X 7.5 cm section of flat test panel was cut from a 7.5 cm X 12.5 cm weatherometer panel using a band saw. The panel section was glued to the surface of the polypropylene block with double-sided adhesive paper. The tungsten carbide cutting blade was advanced toward the sample surface in 1-[micro]m increments at a cutting speed of 3-4 mm/sec until the blade made contact with the surface. Once contact was made, the cutting increment was increased to 5 [micro]m. Slices were taken for the clearcoat and top few layers of primer.

[FIGURE 2 OMITTED]

[FIGURE 3 OMITTED]

[FIGURE 4 OMITTED]

[FIGURE 5 OMITTED]

Free film samples were sliced using a slightly modified method. Two different free film samples were sliced; one from the exposed side down, where the surface that faced away from the xenon lamp during weathering was taped down using the double-sided adhesive paper; the other was sliced from the rear side forward, where the surface that faced directly toward the xenon lamp was taped down. This method allowed us to create a photooxidation profile from each surface down into the bulk of the film.

Analysis

Panel sample slices were analyzed in the following order: transmission UV spectroscopy (UVA profile), followed by transmission FTIR (photooxidation profile), followed by ESR spectroscopy (HALS profile). PAS-FTIR, oxygen transport, and free film samples were analyzed independently.

Transmission UV Spectroscopy

For UVA analysis, a ~5 [micro]m thick coating slice measuring ~1.0 cm X 1.2 cm was wetted with perfluorinated oil, fixed between two 25 mm diameter NaCl plates, and placed in a CO50303 sample holder (Perkin Elmer). The sample was wetted with the perfluorinated oil to reduce the light scattering associated with a dry sample. The UV spectrum of the slices was recorded from 280-440 nm using a Perkin Elmer Lambda 18 UV/VIS spectrometer run in transmission mode. A background spectrum of the NaCl plates wetted with perfluorinated oil was used. Due to scattering from primer pigments, UV spectra were only taken of the clearcoat.

Transmission FTIR

For photooxidation analysis, the same ~5 [micro]m coating slice wetted with perfluorinated oil and fixed between the two NaCl plates were used. While scattering was not as severe for IR as compared to UV spectroscopy, the oil did reduce the baseline in the 4000-2000 [cm.sup.-1] region. The sample was placed in an FTIR sample holder and a transmission FTIR spectrum was recorded over the range of 4000-500 [cm.sup.-1] using a Nicolet Magna-IR 560 spectrometer. A background spectrum of the NaCl plated wetted with perfluorinated oil was again used. Photooxidation was tracked using the same method described in previous work, and originally developed by Gerlock et al. This method uses FTIR measurements and monitors the growth of the -OH, -NH region over time. The greater the growth in -OH, -NH area, the more photooxidized the coating has become. This area is normalized to the area of the -CH region in all measurements. (10) Since both clearcoats were of similar chemical composition, it was assumed that the -CH area of both clearcoats would show similar changes during weathering. This removed any need for a -CH area versus thickness calibration. (11)

HALS Analysis

A technique using electron spin resonance spectroscopy (ESR) to quantify active HALS concentrations has been described in previous works. (12,13) Briefly, active HALS, or species that have the ability to quench radicals, can be converted into nitroxyl radicals with p-nitroperbenzoic acid. The quantity of nitroxyl (NO*) radicals present in the sample can then be determined by performing a double integration on the derivative of the ESR signal. From the microtomed slices, 5-10 mg of coating was removed and placed into the ESR sample tube, weighted, and then swelled with 0.2-0.3 ml of C[H.sub.2] [Cl.sub.2]. The mixture was stirred with a thin glass rod and then placed in an ultrasonic bath for ~1 min to create a homogeneous slurry. After sonication, the sample was reweighed to determine the amount of C[H.sub.2][Cl.sub.2] added. Next, 0.5-1 mg of p-nitroperbenzoic acid was added to the sample tube. The slurry was then stirred with a thin glass rod and sonicated for a 5-10 sec. Finally, the ESR spectrum of the sample was recorded intermittently over a ~30 min period. The maximum nitroxyl radical signal intensity was recorded for use in the calculation of active HALS. Active HALS concentration is reported in [mu]-moles of nitroxyl radical per gram of clearcoat. The active HALS concentration was determined for each slice of clearcoat and the top few layers of primer.

PAS-FTIR

PAS-FTIR spectra were recorded using a Mattson Cygnus 100 rapid scan FTIR system. The scan velocity was 3.6 KHz and the sampling depth was between 12 and 16 [micro]m into the clearcoat. That data was collected and measured using Mattson WinFirst software. Analysis of the PAS-FTIR spectra were performed to quantify photooxidation in the clearcoats by using the same method utilized for transmission FTIR data.

Oxygen Solubility/Permeability

Oxygen transport (solubility, permeability, and diffusion) measurements were taken on free films of both Clearcoat A and a standard OEM thermoset clearcoat, each of which were made by drawing down a film on a glass panel resulting in a dry film build of ~75 [micro]m. Oxygen transport was measured by using a Multi-Tran 4/40 that utilizes a TCD (thermal conductivity detector) and two filaments, one with carrier only, the other with carrier plus flux. Permeability (P) was found using an isostatic test procedure while the diffusion constant (D) was found using the half-time method. (14,15) Solubility (S) could then be determined from the following equation:

S = P / D (1)

[FIGURE 6 OMITTED]

[FIGURE 7 OMITTED]

[FIGURE 8 OMITTED]

RESULTS

Transmission FTIR-Photooxidation

Plots of (-OH, -NH)/-CH ratio values as a function of depth, obtained using transmission FTIR spectroscopy on silicon discs, are shown in Figures 2, 3, and 4. The (-OH, -NH)/-CH ratio is a quantitative measure of the concentration of photooxidation products that remain in the clearcoat after weathering; the greater the ratio, the greater the level of photooxidation. It is most important to note how the ratio changes from the unexposed sample to the weathered sample, as even unweathered samples have a non-zero (-OH, -NH)/-CH ratio. Each data point represents the (-OH, -NH)/-CH value for a 5 [micro]m piece of the clearcoat, allowing the construction of a profile of photooxidation versus depth into the coating. The photooxidation profile of the no additive and HALS only formulations of Clearcoat A after 1000 hr of accelerated weathering is shown in Figure 2. The photooxidation profile for each of the four formulations (plus unexposed) of Clearcoat A after 1868 hr of accelerated weathering is show in Figure 3. The photooxidation profile for each of the four formulations (plus unexposed) of Clearcoat B after 1868 hr of accelerated weathering is shown in Figure 4. As was found in previous work, HALS dramatically reduced the rate of photooxidation in each of the two clearcoats. (9) UVA again appeared to do little to reduce the amount of photooxidation. In both clearcoats, the surface of the no additive and UVA only samples photooxidized more rapidly than the surface of the HALS only and fully stabilized samples. Each of the clearcoats also showed a photooxidation gradient in both the no additive and UVA only formulations after 1868 hr of weathering. Neither of the clearcoats' HALS only nor fully stabilized formulations showed any such gradient.

Ultraviolet Light Absorber Longevity

UVA retention results for Clearcoats A and B are shown in Figures 5-8. The UVA depth profiles of the fully stabilized and UVA only formulations of Clearcoats A and B, respectively, after 1868 hr of accelerated weathering, are shown in Figures 5 and 6. The absorbance summed over each of the slices for each of the clearcoat formulations normalized to 45 [micro]m (standard automotive clearcoat thickness), as a function of weathering time, is shown in Figures 7 and 8. Previous works have shown that UVA loss behavior follows a linear trend, hence the large extrapolation shown in Figure 7, despite only having two data points available. (16) Each depth profile shows a typical gradient in UVA after accelerated weathering. The surface of the clearcoat has lost some of the UVA, while the bulk has retained much of its UVA. The addition of HALS slowed the loss rate of UVA during accelerated weathering in both Clearcoats A and B. This loss rate is comparable to the UVA loss rate of standard OEM thermoset clearcoats.

Hindered Amine Light Stabilizer Longevity

The concentration of active HALS in both Clearcoats A and B are shown in Figures 9 and 10 for paint systems weathered for 1868 hr. Active HALS is a measure of the amount of HALS that has the potential to scavenge radicals. Typically, this means the species either is in the Denisov cycle or in the parent HALS form. (12) Measurement of active HALS is accomplished by oxidizing any active HALS into nitroxyl radicals using p-nitroperbenzoic acid. However, not all active HALS species can be oxidized by this acid, so some background information on the parent HALS used must be known. (12) Results for both the fully stabilized and HALS only samples, as well as an unexposed sample, are shown in each plot. In both systems, the active HALS slowly disappeared from the surface, producing a slight gradient in active HALS. The UVA appeared to have no effect on the HALS behavior. In both clearcoats, the concentration of active HALS stayed at measurable levels through the maximum weathering time studied.

Free Film Photooxidation

A plot of the photooxidation profile produced from two free films of Clearcoat A (no additive) is shown in Figure 11. One set of data was taken from the front (side facing the light during accelerated weathering), while the other was taken from the back (side facing away from the light during accelerated weathering). Each was taken after 1000 hr of light only accelerated weathering. Each sample showed a gradient in photooxidation that disappeared after a depth of ~40 [micro]m. Little difference in the gradient was observed for each side.

Oxygen Solubility

Oxygen solubility, permeability, and diffusion constants for Clearcoat A and a standard OEM thermoset clearcoat are shown in Table 2. Clearcoat A had an oxygen permeability constant approximately one order of magnitude less than the thermoset clearcoat. This was a result of an order of magnitude reduction in the oxygen solubility constant of Clearcoat A versus the thermoset clearcoat. The oxygen diffusion constants for both clearcoats were approximately equal.

[FIGURE 9 OMITTED]

DISCUSSION

The accelerated weathering behavior observed in the two UV-curable clearcoats studied here is, in many respects, similar to the behavior observed in traditional thermoset automotive clearcoats. Particular commonalities include the observation that accelerated weathering leads to a gradient in UVA concentration through the thickness of the clearcoat, a gradient in photooxidation products (in UVA containing formulations), and a small gradient in HALS concentration. The most striking differences observed between the behavior of UV-curable clearcoats and traditional thermoset clearcoats were: (1) the presence of a gradient in photooxidation products in the clearcoat formulations that contain no UVA or HALS and (2) the dramatic effectiveness of HALS in reducing the rate of photooxidation in the UV-curable clearcoats.

[FIGURE 10 OMITTED]

Ultraviolet Light Absorber Retention

The gradient in UVA concentration through the thickness of the clearcoat has been observed in other clearcoat systems and is presumed to form due to the self-protection of the UVA. (17) UVA molecules deeper within the coating experience a much lower intensity of light than those at the top due to absorption of UV light by the UVA molecules near the surface of the coating. Previous work has shown that the loss of UVA from a coating is due primarily to its chemical destruction and not due to exudation from the surface. (17, 18) This destruction rate is proportional to the rate of photooxidation, and thus, the UVA is more readily consumed at the surface of the coating where the intensity of light is highest and the rate of photooxidation is the most rapid. A reduction in the rate of photooxidation of the clearcoat by the addition of HALS leads to a shallower gradient in UVA concentration, as seen in Figures 5 and 6.

The presence of the UVA gradient, however, is not fully explained by the chemical consumption of the UVA. Other research has shown that the mobility of the UVA molecules is sufficient enough to allow for appreciable diffusion through the clearcoat on the time scale of accelerated weathering, which should erase any significant concentration gradient. (19) Yet, the gradient in UVA concentration persists and is observed in almost all weathered automotive paint systems. A satisfactory explanation for this observation has yet to be determined.

[FIGURE 11 OMITTED]

Hindered Amine Light Stabilizer Behavior

Clearcoats A and B had similar active HALS depth profiles. Active HALS were slowly consumed at the surface of the clearcoat in both systems. In much the same way as the gradient in UVA formed, photooxidation also slowly destroyed the active HALS in the clearcoat, producing a similar yet less severe gradient. Despite the presence of the gradient, both Clearcoats A and B (fully stabilized and HALS only) had measurable active HALS at the surface and significantly high levels in the bulk of the clearcoat after 1868 hr of weathering. The concentration of active HALS required to successfully inhibit photooxidation in different coatings is not known. However, these coatings where the concentration of active HALS drops to zero at the surface are known to rapidly photooxidize and embrittle. (20)

A comparison between the active HALS concentration of the weathered and unexposed samples is difficult, as the parent HALS used in both clearcoats proved to be unoxidizable using p-nitroperbenzoic acid. However, once the parent HALS was "activated," species present in the Densiov cycle with the ability to scavenge radicals were oxidizable by the acid. (12) The amount of active HALS in the unexposed samples appeared lower than that of the weathered samples due to this inability to oxidize the parent HALS. In general, it takes ~1200 hr (equivalent to one year of Florida exposure) of SAE J1960 Borosilicate/Borosilicate accelerated weathering to activate all of the parent HALS in the clearcoat. (12)

Photooxidation

Formulations of Clearcoats A and B that did not contain HALS developed a photooxidation gradient during accelerated weathering. This gradient was not surprising, and has been seen previously in systems containing UVA (due to the protecting nature of the UVA). (17) However, in no additive samples, the gradient was still unexpectedly present. A number of possible explanations for this photooxidation gradient were hypothesized and tested:

(1) Hydroperoxides could form during cure due to the reaction of radicals with oxygen (oxygen inhibition). Oxygen inhibition is very common in free radically cured coatings. These hydroperoxides could thermally decompose during weathering, oxidizing the coating. The concentration of hydroperoxides would likely be depth dependant, with a greater concentration at the surface due to oxygen diffusion limitations during the curing process. Thus, the observed oxidation gradient would mirror the hydroperoxide gradient that may be present before weathering.

(2) Chromophores could form during the curing process, mostly likely as byproducts of photoinitiator/UV radiation interactions. These chromophores could act like a UVA and filter out UV light during weathering, reducing the amount of photooxidation in the depths of the clearcoat. This would produce a photooxidation gradient similar to that seen in the no additive formulations of Clearcoats A and B.

(3) Oxygen transport could have been significantly lower than is typically seen in conventional clearcoats, which could reduce the amount of photooxidation in the depth of the clearcoats. While oxygen could still penetrate the top 5-20 [micro]m to allow photooxidation, the deeper sections of the clearcoat would have insufficient amounts of oxygen to produce significant photooxidation products. This would give rise to a photooxidation gradient similar to that seen in the no additive formulations of Clearcoats A and B.

In order to determine if hydroperoxides existed in sufficient quantities to cause the oxidation gradient seen in the no additive formulations, panels were made and heat-aged for 1656 hr at 60[degrees]C to thermally decompose any hydroperoxides present. This temperature was chosen because it corresponds to the temperatures seen in accelerated weathering tests. This decomposition would oxidize the clearcoat preferentially where the hydroperoxides existed, which can be monitored using PAS-FTIR. After heat aging for 1656 hr, no change in the PAS-FTIR (-OH, -NH/-CH) ratio was observed. This implied that hydroperoxides were not present in high enough concentrations to cause the oxidation gradient that was seen in the accelerated weathering samples, and that UV radiation is necessary to produce the observed gradient in photooxidation products.

The other two possibilities for the photooxidation gradient are the existence of chromophores or unusually low oxygen transport within the clearcoat. Both of these hypotheses were addressed by exposing a free film of clearcoat in an accelerated weathering chamber.

If the gradient was caused by the presence of a chromophore in the UVA-free coating, the shape of the depth profile should have looked something like Figure 12a. The chromophore would effectively filter out much of the UV radiation, thereby greatly reducing photooxidation deep within the bulk of the free film. Since oxygen is still available at the back surface of the film, this shape could only be caused by a light filtering species being present within the coating and not low oxygen transport constants. This scenario was not a strong possibility as no UV absorbing species were seen in the UV spectra of the UVA-free clearcoat. If the gradient was caused only by a reduction in oxygen transport within the coating, the resulting free film exposure data should look similar to Figure 12b. In this case, the same light intensity was reaching every part of the coating, but oxygen could not continually penetrate the surface of the coating into the bulk. This would have caused a reduction in photooxidation deep within the bulk, but not at either surface. This would have produced a "U" shaped photooxidation profile. Examination of the actual photooxidation data (Figure 11) shows a shape similar to that in Figure 12b, the "U" shaped profile. This implied that the photooxidation profile was likely caused by an unexpected decrease in oxygen transport within the clearcoat and not by a chromophore.

This final hypothesis was confirmed by the data in Table 2. Results indicated that the permeability and solubility of Clearcoat A were an order of magnitude lower than that of a standard thermoset clearcoat that does not show this type of photooxidation gradient in no additive formulations.

Diffusion-limited oxidation has been commonly reported on in literature. The studies have only looked at thick sections of organic materials such as elastomers or plastics. (21,22) In coatings, the accepted belief has been that the availability of oxygen was never a rate limiting process in the photooxidation of the binder. However, the results presented here strongly suggest that for these UV-curable coatings, the rate of photooxidation is high enough and the permeability of oxygen low enough to exhaust the available supply of oxygen in the regions of the coating below a depth of 25 [micro]m. The phenomenon has not been observed in other coatings where the inherent photooxidation rate is lower and the oxygen permeability is higher. The lack of a gradient in the HALS only samples implies that even though the permeability of the oxygen was likely not changed by the addition of HALS, the reduced photooxidation rate alleviated the oxygen availability constraint and no gradient was produced. This finding has important implications for the durability of multilayer paint systems where the upper layer may act as a barrier against oxygen, preventing the lower layers from photooxidizing, even without protection from UVA stabilizers.

[FIGURE 12 OMITTED]

CONCLUSIONS

The accelerated weathering behavior of two UV-curable clearcoats was studied and compared to the behavior of thermally cured automotive clearcoats. Both of the UV-curable clearcoats photooxidized slowly and uniformly during accelerated weathering testing when stabilized with HALS. In formulations without HALS, the clearcoats photooxidized rapidly during accelerated weathering. The UV-curable clearcoat formulated without both HALS and UVA developed a gradient in photooxidation into the depth of the clearcoat. The cause of this gradient was determined to be an unexpected reduction in the solubility of oxygen in both UV-curable clearcoats as compared to that in thermally cured clearcoats. It is possible that this gradient would be reduced or eliminated if the paint system was exposed outdoors, due to the lower rate of chemical composition change and the increased time for oxygen to diffuse into the coating. The photooxidation depth profile of both clearcoats was relatively unaffected by the addition of UVA, implying that UVAs do little to reduce the photooxidation of the clearcoats. However, UVAs are still often necessary to protect underlying layers in a coating system from photooxidation. The use of low oxygen solubility coatings to reduce the photooxidation of deeper layers in multilayer paint systems is an area that warrants further exploration.

Table 1 -- Stabilizers Present in Each Formulation Classification

                       Stabilizer
Formulation        UVA            HALS

No additive        Not present    Not present
UVA only           Present        Not present
HALS only          Not present    Present
Fully stabilized   Present        Present

Table 2 -- Oxygen Solubility, Permeability, and Diffusion Constants for
Clearcoat A and a Standard Thermoset Clearcoat. Two Samples of Each
Clearcoat Were Tested

                         Permeability         Diffusion      Solubility
Sample               cc*mil/[m.sup.2]*day   [cm.sup.2]/sec     cc/cc

Clearcoat A (1)             175.7              3.29E-06       1.57E-04
Clearcoat A (2)             205.1              3.29E-06       1.83E-04
Std. Thermoset (1)         3050.8              4.11E-06       2.18E-03
Std. Thermoset (2)         3248.6              4.12E-06       2.32E-03

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

References

(1) Fouassier, J.P., Photoinitiation, Photopolymerization, and Photocuring: Fundamentals and Applications, Hanser, Munich, 1995.

(2) Lange, J., Luisier, A., and Hult, A., "Influence of Crosslink Density, Glass Transition Temperature and Addition of Pigment and Wax on the Scratch Resistance of an Epoxy Coating," JOURNAL OF COATINGS TECHNOLOGY, 69, No. 72, 77 (1997).

(3) Evans, D.C. and Lancaster, J.K., Treat. Mater. Sci. Tech., 13, 85 (1979).

(4) Decker, C. and Zahouily, K., Polym. Degrad. Stab., 64, 293 (1999).

(5) Bendaikha, T. and Decker, C., J. Rad. Cure., April, 6 (1984).

(6) Decker, C., Moussa, K., and Bendaikha, T., J. Polym. Sci., Part A: Polym. Chem., 29, 739 (1991).

(7) Decker, C. and Biry, S., Prog. Org. Coat., 29, 81 (1996).

(8) Valet, A., Prog. Org. Coat., 35, 223 (1999).

(9) Seubert, C.M., Nichols, M.E., Gerlock, J.L., and Cooper, V.A., Polym. Degrad. Stab., 81, 103 (2003).

(10) Gerlock, J.L., Smith, C.A., Cooper, V.A., Dusbiber, T.G., and Weber, W.H., Polym. Degrad. Stab., 62, 225 (1998).

(11) Gerlock, J.L., et al., Service Life Prediction Methodology and Metrologies, ACS, Oxford University Press, 212, 2002.

(12) Kucherov, A.V., Gerlock, J.L., and Matheson, R.R., Jr., Polym. Degrad. Stab., 69 (1), 1 (2000).

(13) Gerlock, J.L., Kucherov, A.V., and Smith, C.A., Polym. Degrad. Stab., 73 (2), 201 (2001).

(14) "MOCON Technical Bulletin," #124-C, Minneapolis, MN 55428, February 1996.

(15) Alger, M.M. and Stanley, T.J., J. Appl. Polym. Sci., 36, 1501 (1988).

(16) Pickett, J.E., Service Life Prediction Methodology and Metrologies, ACS, Oxford University Press, 250, 2002.

(17) Decker, C., Polym. Prepr., 34, 139 (1993).

(18) Gerlock, J., Tang, W., Dearth, M., and Korniski, T., Polym. Degrad. Stab., 48, 121 (1995).

(19) Cliff, N., Kanouni, M., and Peters, C., Proc. International Waterborne, High-Solids, and Powder Coatings Symposium, New Orleans, LA, 2004.

(20) 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," JOURNAL OF COATINGS TECHNOLOGY, 73, No. 918, 45 (2001).

(21) O'Donnell, B. and White, J.R., Polym. Degrad. Stab., 44, 211 (1994).

(22) Clough, R.L. and Gillen, K.T., Polym. Degrad. Stab., 31 (1), 47 (1992).

C.M. Seubert, ([dagger]) M.E. Nichols, and A.V. Kucherov** -- Ford Research and Advanced Engineering*

* Materials Research and Advanced Engineering, Dearborn, MI 48121-2053.

([dagger]) Author to whom correspondence should be addressed: P.O. Box 2053, MD 3182, SRL-Room 2321, Dearborn, MI 48124; 313.322.3070; [e-mail].

** Visiting scientist: Zelinsky Institute of Organic Chemistry, RAS, Moscow, Russia.