Use of reactable light stabilizers to prevent migration and to improve durability of coatings on plastic substrates.

One of the challenges of coating plastic substrates for exterior automotive applications involves the tendency for light stabilizers, which inhibit UV degradation of the coating, to migrate out of the topcoat and into the underlying plastic substrate. The consequent depletion of the stabilizer from the topcoat could account for significantly poorer durability for coatings applied on plastic than on a nonpermeable substrate such as steel.

These studies investigate the migration of both nonreactable and reactable UV absorbers (UVAs) and hindered amine light stabilizers (HALS) in fully formulated 2K urethane coatings applied on thermoplastic olefin (TPO) and steel substrates. Extensive migration of the nonreactable stabilizers occurs during accelerated exposure of the cured coating, and this migration correlates with poorer performance on TPO. Functionalization of the stabilizer with a group which covalently reacts into the coating will prevent much of this migration, and the improved degree of retention of the reactable light stabilizers in the topcoat correlates with improved performance on TPO substrates. In addition, stabilizer variables, such as the type and extent of functionalization as well as their chemical structure, have an effect on both the migration and performance of the stabilizers.

Keywords: UV absorbers, hindered amine light stabilizers, electron paramagnetic spectroscopy (EPR), crosslinking, cure, photodegration, stabilization, photostabilizers, durability, thermoplastic olefins, UV spectroscopy

**********

The use of plastics for automotive structural components has increased dramatically in recent years. Most of these plastic parts require coatings that are not only matched with the steel body in color and appearance, but that also meet the same durability criteria. Historically, coatings on plastic parts have not been able to match steel coatings in weatherability. Problems in durability may result from the somewhat different composition of the flexible plastic coating versus rigid coatings. They may also be caused by the migration of low molecular weight components from the plastic into the coating (1) as well as the migration of light stabilizers from the coating into the plastic. (2-4) In this article, we demonstrate that the latter mechanism--migration of conventional stabilizers out of the topcoat and into the plastic substrates--can account for a significant portion of the poorer performance observed in coatings applied on plastic versus on steel.

Figure 1 illustrates a typical coating system construction for coatings on steel and on TPO. Although some limited migration of light stabilizers from the topcoat (basecoat + clearcoat) into the primer layers on the steel can occur, this is largely prevented by prebaking the primers before applying the basecoat and clearcoat layers, as well as by their highly crosslinked nature. In any case, any migration would be limited to the coating layers, since no migration into the steel substrate could take place. In contrast, all the paint layers on the TPO were applied wet-on-wet-on-wet. In addition, the TPO, which is much thicker than the two primer layers on steel, and is also noncrosslinked, can act as a virtually unlimited "sink" for the migration of stabilizers, resulting in depletion of the stabilizers from the topcoat.

A number of studies have demonstrated significant migration in both model clear-over-plastic systems and real basecoat/clearcoat systems. Haacke (3) and Cliff, et al. (5) have shown that nonreactable stabilizers migrate into plastic substrates in model studies where clearcoat is applied directly on the plastic and cured under normal conditions. Recent studies have also extended this characterization to studies of total systems applied on primed TPO substrates. In addition, weathering studies on these coating systems have shown that the most significant migration occurs during weathering exposure rather than during application and cure. Significant (almost complete) depletion of nonreactable UV absorbers and hindered amine light stabilizers (HALS) from the topcoat into the TPO substrate was seen after only a few thousand hours of xenon exposure. (6)

[FIGURE 1 OMITTED]

[FIGURE 2 OMITTED]

One solution to the migration problem is modification of the light stabilizer structure to include a functional group that will react into the typical topcoat applied on these substrates. (5) There are many examples of various types of reactable stabilizers cited in the literature (7-12); there are also a number of applications in which reactable light stabilizers have demonstrated utility. These include the incorporation of co-condensible or copolymerizable UV absorbers into very thin coatings which require high levels of UVAs, (13) incorporation of copolymerizable light stabilizers into waterborne coatings via emulsion polymerization, (14,15) and condensation of silanol-functional UVAs into silicone-based hardcoats. (16-18) However, until recently, the need for these stabilizers has been in relatively small, specialized markets. In conventional automotive basecoat/clearcoat systems applied on nonpermeable substrates such as steel, reactability shows little observable advantage. (6) With the increased use of painted plastic for automotive exterior body parts, the need for reactable light stabilizers is increasing. The most widely used substrate for exterior automotive parts is currently thermoplastic olefin. Some of TPO's most desirable characteristics, such as excellent flexibility and thermoplasticity, make it also relatively easy for low molecular weight stabilizers to be absorbed into it from the topcoat.

This article presents the results of two studies. In the first study (study #1), data is presented which shows a weaker performance in xenon exposure of the same coating system when applied on TPO rather than on steel, and shows the effectiveness of reactable or partially reactable commercially available light stabilizers in improving the performance of coatings on TPO to the same level as achieved on steel. It does not differentiate between the effects of reactable UVA and reactable HALS, nor does it include "model" compounds to allow for further definition of the effects of partial versus full reactability, and compatibility and solubility characteristics of the light stabilizers.

In addition to confirming a direct relationship between the migration of light stabilizers and the performance of stabilized coatings in a xenon weatherometer, study #2 attempts to answer some of the questions about the requirements, usefulness, behavior, and testing of reactable light stabilizers in coatings on plastics by further defining which characteristics of the stabilizers have the greatest effect on either migration or actual performance in coatings. These characteristics include degree of reactability, primary versus secondary hydroxyl functionality, reactable UV absorber versus reactable HALS, effect of structural variations in UVAs and HALS, and basecoat stabilization.

An additional challenge for the coating manufacturer who wishes to predict the performance of light stabilizers by determining their migration characteristics is the availability of analytical techniques for the determination of light stabilizer location and levels. This is especially true for reactable light stabilizers, which cannot be easily extracted and analyzed, as well as for HALS in general, which undergo chemical transformations as part of their stabilizing activities during weathering. (19) This article discusses the development of several relatively new analytical techniques used to determine the levels and location of both reactable and extractable stabilizers. The analytical data obtained by these methods establishes a direct correlation between stabilizer properties, their extent of migration, and, ultimately, performance in the coating systems.

EXPERIMENTAL

Coating Materials 

COATING SYSTEMS:

Study #1 -- Coatings were supplied by Dupont Performance Coatings. The clearcoat was a 2K flexible urethane designed for application on plastic substrates. The basecoat was a medium blue metallic 1K acrylic-melamine flexible basecoat. Substrates used were cold-rolled steel with electrocoat (supplied by Advanced Coatings Technology (ACT) to which a conventional 1K primer-surfacer was applied and baked for 25 min at 127[degrees]C (part temperature) prior to topcoating, and Montell Hifax [R] TPO (CA287AC) primed with chlorinated polyolefin adhesion promoter and air-dried prior to topcoat application followed by a 25-min bake at 127[degrees]C (part temperature).

Study #2 -- The clearcoat was a flexibilized two-component urethane. The basecoat was a light tan metallic 1K acrylic-melamine flexible basecoat. Substrates used were cold-rolled steel with electrocoat and primer surfacer (supplied by ACT) and Solvay Sequel[R] 1440 TPO. The TPO was first primed with a black conductive adhesion promoter and flashed for five minutes before application of the basecoat.

[FIGURE 3 OMITTED]

LIGHT STABILIZERS: The light stabilizers were supplied by Ciba Specialty Chemicals. Their structures are shown in Figure 2.

UV Absorbers -- 2,4-bis(2,4-dimethylphenyl)-6-(2-hydroxy-4-iso-octyloxyphenyl)-s-triazine (TZ 2 UR), 2-[4-[2-Hydroxy-3-dodecylcloxypropyl)oxy]-2-hydroxyphenyl]-4,6-bis (2,4 dimethylphenyl)-1,3,5-triazine (TZ 1 R), [beta]-[3-(2-H-Benzotriazole-2-yl)-4-hydroxy-5-tert.butyl-phenyl]-propionic acid poly(ethylene glycol)300 ester and Bis{[beta]-[3-(2-H-Benzotriazole-2-yl)-4-hydroxy-5-tert.butylphenyl]-propionic acid}-poly(ethylene glycol)300-ester (BZT 1R), 2-(2H-Benzotriazol-2-yl)-6-(1- methyl-1-phenylethyl)-4-(1,1,3,3-tetramethylbutyl) phenol (BZT 2 UR) (Ciba). Structure of BZT 3 R is confidential. Figure 3 shows the absorbance spectra at equal weights as supplied and at equal weight on active UVA.

Hindered Amine Light Stabilizers -- 2,4-bis[N-Butyl-N-(1-cyclohexyloxy-2,2,6,6 tetra methyl piperidin-4-yl) amino]-6-(2-hydroxyethylamine)-1,3,5-triazine (HALS 1 R), bis(1-octyloxy-2,2,6,6-tetramethyl-4-piperidinyl) sebacate (HALS 2 UR). Structure of HALS 3 UR is confidential.

Most of the compounds evaluated were commercially available. Consequently, the stabilizers were not necessarily direct analogs of each other, or 100% reactable. For example, TZ 1 R and BZT 1 R are, respectively, 72% and 36% reactable in system #2 (see Table 4 in Results section). Two noncommercial (experimental) compounds were included in study #2 in order to make the set of variables more complete. These were BZT 3 R, a 100% functional primary hydroxyl functional UVA, and HALS 3 UR, a highly polar and compatible noninteracting HALS. In addition to the defined variables of UV absorber type, UV absorber reactability, and HALS react-ability, the stabilizers used in this study differed in several other aspects as follows:

BZT 3 R: Experimental (noncommercial) hydroxyphenyl benzotriazole UV absorber fully functionalized with primary hydroxyl, expected to show the highest retention in the topcoat. This compound was not part of the complete factorial experimental design (see Study Design section), but was added into study #2 to determine the effect of full reactability of a Benzotriazole versus Triazine UVA.

[FIGURE 4 OMITTED]

BZT 1 R: Partially (50%) primary hydroxyl functionalized Benzotriazole UV absorber, expected to exhibit 50% retention in topcoat on TPO. Also, its lower specific extinction coefficient was expected to contribute to somewhat lower performance than its triazine analog, TZ 1 R.

BZT 2 UR: Nonreactable BZT UV absorber. Somewhat lower in molecular weight and specific extinction versus its triazine counterpart, TZ 2 UR.

TZ 1 R: Fully functionalized secondary hydroxyl Triazine UVA. Expected to show high degree of reactability with isocyanate.

TZ 2 UR: Nonreactable TZ UVA. Its lower molecular weight versus TZ 1 R gives it a high specific extinction versus both TZ 1 R and BZT 2 UR.

HALS 1 R: 100% primary hydroxy functional, alkoxy amine (noninteracting) HALS. Expected to show full reactability with isocyanate crosslinker, and hence high retention in topcoat.

HALS 3 UR: Experimental (noncommercial) alkoxy amine (noninteracting) HALS. It is somewhat more polar and compatible than HALS 2 UR, and was expected to show higher retention than HALS 2 UR but lower retention and performance than HALS 1 R in the urethane topcoat. This HALS 3 UR was the nonreactable HALS used in the study #2 design, in order to eliminate the variable of incompatibility.

[FIGURE 5 OMITTED]

HALS 2 UR: Commercial alkoxy amine HALS. It was expected to show poor retention in the topcoat when applied on TPO. This is partially due to its nonreactability, but also a result of its nonpolar structure. (20) It was not included in the study #2 design matrix because of slight incompatibility in this clearcoat system, but was evaluated in add-on formulations.

The above differences among the stabilizers will account for some of the differences observed in performance beyond those defined in the experimental design, as well as allow for some evaluation of the effects of these variables, such as degree of reactability, extinction coefficient, compatibility, and molecular weight.

Formulation and Application

STUDY #1: Light stabilizers were incorporated at concentrations based on the nominal solids content of the clearcoat. The stabilizers were incorporated as supplied; hence, if the stabilizer contained solvent, levels were not adjusted to compensate for the solvent. This was the case with both TZ 1 R (supplied as 85% active in methoxypropyl acetate) and TZ 2 UR (supplied as 65% active in xylene). Results are also reported on an as-supplied basis. The coatings were spray-applied to a targeted basecoat film thickness of 18 microns and clearcoat film thickness of 45 microns, then baked at 127[degrees]C for 30 min.

STUDY #2: Light stabilizers were incorporated at concentrations based on the resin solids content of the clearcoat and basecoat. The stabilizers were incorporated on a solids basis; hence, if the stabilizer contained solvent, levels were adjusted to compensate. This was the case with both TZ 1 R (supplied as 85% active in methoxypropyl acetate) and TZ 2 UR (supplied as 65% active in xylene). The coatings were spray-applied to a targeted basecoat film thickness of 25 microns and a clearcoat film thickness of 40 microns. Basecoat and clearcoat were applied wet-on-wet with five minutes of flash between coatings at film builds indicated in Table 1, then baked at 122[degrees]C part temperature for 30 min.

Study Design

STUDY #1: The study design is shown in Table 2. Fully reactable versus fully nonreactable systems were compared on steel and on TPO. Stabilizers were incorporated into the clearcoat only (basecoat unstabilized).

STUDY #2: A full two-level, three-variable factorial design (A) (21) was used to explore the effects of UV absorber type (Hydroxy phenyl benzotriazole versus Hydroxy phenyl s-triazine), UV absorber reactability, and HALS reactability. A second full factorial design (B) was also included to evaluate the UV absorber type, stabilizer package reactability, and basecoat stabilization factors. In addition, a D-optimal design which allows for the evaluation of all four factors could be devised by the inclusion of several additional formulations. Table 3 shows the study design and the variables which are being evaluated in each part of the study.

The complete design of study #2 was applied on Solvay Sequel 1440 TPO, primed with a black conductive adhesion promoter for TPO at both normal film thickness and thin film. A subset of the above formulations (Factorial Design B) was also applied on steel.

All the light stabilizers were formulated based on active stabilizer and total resin solids (polyol + isocyanate). Assuming equilibration of the UV absorber between the basecoat and clearcoat, the level of 2.8% UV absorber in the clearcoat and 2.5% in the basecoat would have resulted in a level of 2.7% in both layers. The choice of equal levels instead of equal absorbance for the UV absorbers proved to be a somewhat complicating factor, since both triazines have a higher absorbance at equal weight levels versus their benzotriazole UV absorber analogs, making it difficult to distinguish between the effects of UV absorber type versus UVA absorbance.

Exposure

STUDY #1: Panels 2.5 X 5 in. were exposed in an Atlas Ci65 Xenon Weather-Ometer equipped with borosilicate (S)-type inner and outer filters, as per SAE J1960 cycle. Panels were exposed in triplicate initially, with sets being removed at 1500 and 3000 hr for microtoming and subsequent analysis. Gloss measurements at 20[degrees] were obtained on a Byk micro TRI-gloss[R] meter after every 300 hr of exposure.

STUDY #2: Panels 6.5 X 13 cm were exposed in an Atlas Ci65 Xenon Weather-Ometer using borosilicate (S)-type inner and outer filters, and 0.55 W/[m.sup.2] irradiance, with an SAE J1960 cycle. Duplicate panels were exposed initially, with one set being removed at 3000 hr for microtoming and analysis. Triplicate panels were also exposed at 5[degrees] South Florida, as well as in Sun 10[R] concentrated solar irradiance devices. Twenty degree gloss was measured every 500 hr on a Byk micro TRI-gloss meter.

[FIGURE 6 OMITTED]

[FIGURE 7 OMITTED]

[FIGURE 8 OMITTED]

[FIGURE 9 OMITTED]

Analytical Test Methods for Light Stabilizers

A number of techniques, some of them recently developed, were used to determine the UV absorber and HALS content in exposed and unexposed clearcoats, basecoats, and TPO substrate. Different methods were required for the analysis of UV absorbers versus HALS, and for reactable versus nonreactable light stabilizers. These methods are described in detail below.

MICROTOME: Most of the UV absorber and HALS quantification studies were performed on seven micron-thick microtomed coating sections. Sections were obtained using a Polycut[R] slab microtome.

EXTRACTION: Approximately 10-15 mg of scraped or microtomed clearcoat or basecoat were weighed accurately and swelled with tetrahydrofuran (THF). The mixture was sonicated, rediluted to a specific volume, and filtered. The concentration of extracted UV absorber was determined by comparing long wavelength peak height to that of known concentration standards of the same UVA. This method can be used only for nonreactable UV absorbers.

[FIGURE 10 OMITTED]

UV MICROSCOPY: Steel samples were bent until the paint system popped free from the metallic substrate. TPO samples were cut with a razor blade to include the paint system along with 1 or 2 mm of substrate material. The paint systems were then sandwiched between two pieces of reference film and glued together using a two-part, solvent-free, high molecular weight polyurethane adhesive (Ashland Chemical Pliogrip[TM] 7779/100). The solvent-free nature of the adhesive ensured that additives did not migrate while the coating was being analyzed. The reference film used was an acrylic/melamine clearcoat film with known UV absorber concentration and known UV absorbance. The entire sandwich was placed in an RMC Model MT990 rotary microtome and a 10-[micro]m thick cross-section was cut using a diamond blade. The cross-section was then mounted on a quartz microscope slide in a solution of 10% glycerin in water and a quartz coverslip was placed over the specimen. UV spectra were collected using a Carl Zeiss, Inc., Axiotron[TM] MPM 800 microscope/spectrometer system with programmable stage. A 5-[micro]m wide X 10-[micro]m long aperture was set and positioned at the top surface of the clearcoat. A spectrum was collected from 250-500 nm and the stage was programmed to move in 5-[micro]m increments from the surface of the clearcoat, across the basecoat, and then into the primer and substrate (for the TPO sample). UV spectra were collected at each 5-[micro]m step across the sample, as shown in Figure 4. Since this is a transmission technique, there was some scattering in all the pigmented layers, making it difficult, if not impossible, to quantify UV absorber levels in these layers. Finally, the aperture was focused on the reference film and a UV spectrum was collected. Since the reference film has a known UV absorber concentration, it served as a way to measure actual cross-section thickness and allowed all data to be normalized to a thickness of exactly 10 [micro]m.

UV microscopy (22-25) has both advantages and disadvantages in comparison with extraction techniques. A major advantage is that, since the method requires no extraction of the UVA, reactable UV absorbers can be detected and quantified as easily as nonreactable UVAs. It is also more precise and reproducible than the extraction-plus-hydroysis technique, which is required for the analysis of reactable UVAs. The only disadvantage of this technique is that since basecoats and TPO are not transparent, it cannot be used to quantify UVAs in these layers. Table 5, in the Results section, compares the results obtained by UV microscopy with those obtained by the extraction of microtomed layers (results in parentheses) for some of the nonreactable UV absorbers.

OXIDATION AND QUANTIFICATION OF TOTAL HALS CONTENT: A recently developed technique used to accurately determine the HALS content for N-H and NOR functional HALS was used. This technique, developed by Gerlock, Kucherov, and Matheson (26,27) involves in-situ conversion of all HALS into nitroxide, then quantification in the coating by electron paramagnetic resonance (EPR). This is the only method currently available which has the ability to analyze for nonreactable HALS, reactable HALS, and HALS in weathered coatings.

Approximately 8-12 mg of the coating sample containing HALS 2 UR were placed in a quartz EPR tube (3 mm ID), and enough methylene chloride was added (approximately 300-350 mg) to create a sample height of approximately 25 mm. The tube was capped, and the suspension sonicated for approximately 15 min to ensure swelling of the sample by the methylene chloride.

EPR spectra were obtained on a Bruker EMX X-band spectrometer equipped with a 4102ST resonator. The double integral of the first derivative spectrum was used to quantify the nitroxide signal. After obtaining the unoxidized steady state nitroxide content, approximately 1-2 mg of para nitro-perbenzoic acid were then added to the solvent-swelled coating suspension. This quickly converted all the NOR and NH HALS groups into nitroxide, which could be quantified by EPR. The swelled coating and acid were mixed thoroughly both mechanically and via sonication. The EPR spectra were monitored at short time intervals until the nitroxide signal was observed to decay, indicating that a peak absorbance had been obtained. The peak signal was double integrated, and the maximum integrated value was used to calculate nitroxide concentration versus a calibration curve obtained with a nitroxide standard in methylene chloride.

CALIBRATION STANDARDS: Nitroxide radical standard; solutions of nitroxide radical (4-benzoyl-2,2,6,6-tetramethyl-piperidine-1-yloxy) in a range of concentrations from 0.15-6.5 micromoles/[g.sub.solvent] in methylene chloride were used to establish calibration curves for nitroxide.

HALS 2 UR AND HALS 1 R COATING STANDARDS: To determine the efficacy of the swelling and oxidation technique in cured coating systems, HALS 1 R and HALS 2 UR were incorporated into the 2K urethane clearcoat and the 1K acrylic/melamine basecoat formulation at a concentration of 1% based on binder solids. The coatings were applied onto primed steel panels via drawdown, cured for 30 min at 122[degrees]C, then microtomed into eight-micron thick layers through the clearcoat. Several composite samples comprised of a portion of each layer were made and used as replicate samples. Approximately 93% recovery of HALS 2 UR was obtained, with a 90% confidence limit of [+ or -] 2.4%. Similar recovery was obtained for HALS 3 UR. For the basecoat, recovery was lower, but consistent. An average of 53% recovery was obtained for both HALS 1 R and HALS 2 UR, and 63% for HALS 3 UR. The recovery rates were used to normalize calculations for the retention of HALS in the basecoat and clearcoat.

There are two major benefits for this technique over extraction (for nonreactable HALS) and extraction/hydrolysis (for reactable HALS): the ability to quantify active HALS content even after weathering for both nonreactable and reactable HALS, and better accuracy and reproducibility than extraction/hydrolysis for reactable HALS analysis. Since the HALS is analyzed in-situ, there is also no time-consuming extraction or hydrolysis step. Gerlock (26) and Cliff (5) have previously shown the accuracy and efficacy of this technique for quantification of HALS in coatings.

[FIGURE 11 OMITTED]

[FIGURE 12 OMITTED]

RESULTS AND DISCUSSION

Study #1

Determination of Effect of Light Stabilizer Reactability on Performance

The poorer performance of conventional (nonreactable) light stabilizer systems on TPO than on steel is shown in Figure 5A for system #1--2K urethane clearcoat/medium blue metallic basecoat. Figure 5B shows the improvement in performance on TPO resulting from the replacement of the nonreactable UVA and HALS with hydroxy functional reactable light stabilizers. These results support previous findings that the nonreactable light stabilizers migrate out of the topcoat either before or during exposure, leaving the topcoat minimally stabilized or unstabilized at an early stage of exposure. Reference 6 reports analytical results for the same stabilizer systems incorporated into a 1K siloxane system applied over the same basecoat, which correlate with these exposure results.

[FIGURE 13A OMITTED]

Although study #1 clearly shows the benefits of reactable light stabilizer systems for coatings on TPO, it does not differentiate between the effects of reactable UVA and reactable HALS. Its restriction to the use of commercially available stabilizers also does not allow for the differentiation between the effects of HALS polarity and reactability, in that HALS 2 UR is significantly less polar than HALS 1 R, and should have a greater affinity for the TPO than other more polar HALS. In order to achieve a better understanding of the effects of these coating variables, study #2 included experimental compounds which would allow for the differentiation between the effects of polarity and reactability. It included formulations which would differentiate between the effects of reactable UVA versus reactable HALS on performance, and it included a noncommercial UVA which would allow determination of effect of primary hydroxyl functionality on UVA reactability and performance.

[FIGURE 13B OMITTED]

Study #2

Determination of Reactability of Light Stabilizers In 2K Urethane

The stabilizers were incorporated into the clearcoat at the designated levels. Drawdowns on glass at approximately 50 microns were made and cured at 122[degrees]C for 30 min. After ascertaining complete cure by IR, the coatings were removed from the substrate, cryoground, and extracted using the extraction technique described in the Experimental section. Extractable UV absorber was quantified using UV-Vis spectroscopy and HALS was quantified using HPLC analysis. The results are reported in Table 4.

As expected, none of the primary hydroxyl-functional BZT 3 R and HALS 1 R were extractable, indicating 100% reaction with the isocyanate crosslinker. Approximately 72% of the secondary hydroxyl functional TZ 2 R and only 36% of the partially-functional BZT 1 R reacted.

System #2 Analytical Results--Effect of UV Absorber Reactability and Type

ANALYTICAL QUANTIFICATION OF UVA IN WEATHERED AND UNWEATHERED COATINGS: Based on one's knowledge of coating chemistry, as well as on previously reported results, one would expect to observe the least migration for the fully reactable BZT 3 R, followed by the secondary hydroxyl functional TZ 1 R, then the partially reactable BZT 1 R, and finally the nonreactable UVAs TZ 2 UR and BZT 2 UR. Also, one would expect to see little migration from coatings on steel and significant migration on TPO.

Analysis of reactable and nonreactable UV absorbers in the clearcoat, both after initial application/cure and after 3000 hours of exposure, was done by UV microscopy on selected formulations. Examples of the profiles obtained by UV microscopy are given in Figures 6-8. In addition, several of the formulations which contained nonreactable UV absorbers were also analyzed by microtoming and extraction, which provided a comparison of the two techniques and allowed for the quantification of nonreactable UV absorber in the basecoat and deep into the TPO substrate. Examples of these results are given in Figures 9 and 10. The quantitative results for UV absorber retention are given in Table 5.

UV MICROSCOPY: Comparison of Figures 6 and 7 shows that little migration of the nonreactable UV absorber BZT 2 UR occurred from the clearcoat after initial bake on either the steel or TPO substrate. Calculations based on these spectra (Table 5) indicate that before exposure, the clearcoat applied on TPO contained only 10% less of the UV absorber BZT 2 UR than the coating applied on steel. Similar results can be observed for TZ 2 UR--i.e., that little or no migration of the nonreactable UV absorbers occurred from the clearcoat during application and cure. In fact, very close to the theoretical amount was found in the clearcoat when applied on both steel and TPO.

After 3000 hr of exposure, approximately 65% of the BZT 2 UR in the clearcoat on TPO was gone, leaving only 35% of the original amount, or approximately 0.9% BZT 2 UR in this clearcoat. TZ 2 UR also exhibited significant loss (70%). In contrast, the corresponding clearcoats applied on steel showed little loss of nonreactable UV absorber during exposure. BZT 2 UR exhibited approximately 85% of the original amount of UVA, or 2.3%. Similar retention was observed for TZ 2 UR. It can also be seen from the profiles in Figure 7 that any loss of UV absorber from the coating on steel had occurred from the surface of the coating, and not by migration into the primer.

[FIGURE 14 OMITTED]

As expected, all the reactable UV absorbers also exhibited excellent retention before exposure on both TPO and steel. After 3000 hours of exposure, the best retention was observed for the primary hydroxy functional BZT 3 R, followed by the secondary hydroxy functional TZ 1 R, and finally the 36% reactable BZT 1 R. These results are in line with those predicted based on the observed reactability in this coating as reported in Table 3. Figure 8 shows the UV microscopy profile for the reactable TZ 1 R and illustrates the excellent retention on TPO both before and after 3000 hours of exposure.

UV ABSORBER EXTRACTION STUDIES: The samples used for the UV microscopy analyses were complete basecoat/clearcoat systems. However, only the UV absorber in the clearcoat could be accurately analyzed by this method. In contrast, the solvent extraction technique can be used on microtomed layers of basecoat and TPO, and yields UV absorber content for these layers as well as for the clearcoat. As indicated in Table 5, for BZT 2 UR and TZ 2 UR on TPO, the data obtained from both UV microscopy and microtomy/extraction are in good agreement for most of the samples analyzed.

[FIGURE 15 OMITTED]

Figures 9 and 10 show the results of microtomy + solvent extraction performed on systems containing the nonreactable UV absorbers BZT 2 UR and TZ 2 UR. Figure 9 contrasts the distribution of nonreactable BZT 2 UR throughout the systems applied on steel and on TPO after 3000 hours of xenon exposure. In addition to showing the loss of the UV absorber from the topcoat applied on TPO, this data shows that BZT 2 UR is found as deep as 600 microns into the TPO. No UV absorber was found in an uncoated sample of TPO, confirming that all of the UV absorber found in the TPO had migrated there from the topcoat. The data in Figure 9 also confirm very little migration of UV absorber from the topcoat into the primer layers on steel.

Figure 10 shows the distribution of nonreactable TZ 2 UR on TPO before and after exposure. Little migration of the UV absorber from the clearcoat had occurred before exposure, but some migration from the basecoat into the TPO appears to have occurred. After exposure, the migration was extensive, and this UV absorber was also found to have migrated deep into the TPO.

System #2 Performance in Xenon--Effect of UV Absorber Type and Reactability

PERFORMANCE ON TPO: In some cases, differentiation in 20[degrees] gloss has not yet occurred between formulations in the normal film build series, especially those fully stabilized in basecoat and clearcoat. However, it is anticipated that similar trends to those seen in the thin film set, and in those formulations receiving only clearcoat stabilization will eventually be observed. Therefore, in some cases, data from the differentiated sets are shown graphically and discussed to the exclusion of those which have not differentiated.

Figure 11 illustrates the dramatic effect of UV absorber reactability on gloss retention and time to failure observed in the thin-film study. It compares the gloss versus xenon weathering results for reactable and nonreactable benzotriazole and triazine UV absorbers in combination with the reactable HALS 1 R. The same trend is observed when the UV absorbers are used in combination with the nonreactable HALS 3 UR (Figure 12A). The data in Figure 11 and 12A show that the gloss retention at 5000 hr, as well as time-to-failure in the thin-film application, correlates with UV absorber reactability. The 100% reactable primary hydroxyl functional BZT 3 R outperformed the secondary hydroxyl functional 72% reactable TZ 1 R, followed by the 36% reactable BZT 1 R, and finally the nonreactable TZ 2 UR and BZT 2 UR.

A similar performance improvement is seen in the normal film build study (Figure 12B) for reactable versus nonreactable benzotriazole UVAs, but no differentiation in gloss has yet occurred between the fully reactable BZT 3 R and partially reactable BZT 1 R, or between the reactable TZ 1 R and nonreactable TZ 2 R. The higher absorbance of TZ 2 UR compared with TZ 1 R may be delaying the differentiation in the normal film build study.

The statistical (DOE) analysis of the complete data set of Design A for the thin-film study indicates that both the factors of UV absorber type and UV absorber reactability have a significant effect on performance, with reactability (Factor B) having the greatest effect. It does not indicate any interaction between the two variables--i.e., the degree of improvement imparted by UV absorber reactability is the same for benzotriazole and triazine UV absorbers. Figure 13A shows the half normal plot for the analysis of 20[degrees] gloss at 5000 hours of xenon exposure. The data points in this graph represent the sum of squares analysis of the effects versus the probability of the result occurring if it were due to normal scatter in the data. Any points that are off the straight line are not considered to be part of a normal distribution curve, and hence are due to real "effects." (21)

Figure 13B shows an interaction graph for the two factors of UV absorber type and reactability. It indicates that gloss retention after xenon exposure is improved by the use of both reactable and triazine UV absorbers. The parallel performance improvements for both factors indicate that there is no synergistic interaction between the two factors of reactability and UVA type.

Statistical analysis of the complete data set for the normal film build study (Figure 14) also indicates that both the factors of UV absorber type and UV absorber reactability have a significant effect on performance, with reactability having the greatest effect. It also indicates an interaction between UV absorber type and reactability--i.e., the effect of reactability on gloss is seen with the benzotriazole but not, so far, with the triazine UV absorber. A possible explanation for the seemingly better performance of the nonreactable triazine TZ 2 UR versus the nonreactable BZT 2 UR is its higher absorbance at equal weight levels (higher specific extinction coefficient). As seen in Figure 3, at equal levels TZ 2 UR had an absorbance of approximately 28% higher than that of TZ 1 R and BZT 2 UR at their long wavelength peaks. So even though approximately equal weight levels of TZ 2 UR and BZT 2 UR are retained, the TZ 2 UR-stabilized clearcoat will have approximately 28% higher absorbance through the film.

Figure 14 also shows that the gloss of the nonreactable TZ 2 UR-containing formulations is lower than that of the TZ 1 R-stabilized formulations, but the difference is not yet statistically significant. It is anticipated that further exposure will result in significant differentiation between the reactable and nonreactable triazine UV absorber.

[FIGURE 16 OMITTED]

[FIGURE 17 OMITTED]

[FIGURE 18 OMITTED]

PERFORMANCE ON STEEL: On steel, the experimental design was "Factorial B" (Table 3), where the variables were UV absorber type (Factor A), overall stabilizer reactability (Factor B), and basecoat stabilization (Factor C). The design of this series consisted of combinations of reactable UVA + reactable HALS versus nonreactable UVA + nonreactable HALS. The 20[degrees] gloss data on steel shows no differentiation due to any of the variables, except a slightly greater gloss loss for BZT 2 UR (Figure 15). Statistical DOE analysis also shows no significant effects on gloss after 10,500 hours of xenon exposure.

System #2 Effect of HALS Reactability and Structure

ANALYTICAL QUANTIFICATION OF HALS IN WEATHERED AND UNWEATHERED COATINGS: The HALS oxidation technique (26,27) was used to quantify active HALS in both unexposed and exposed coatings. Formulations containing each of the HALS in combination with the same UV absorber (TZ 2 UR) were analyzed before and after exposure. Samples were microtomed and approximately equal parts of the layers were combined to form composite samples of clearcoat and basecoat. The HALS level was calculated on resin solids for the basecoat and clearcoat, but on total solids for the TPO, which has an unknown ratio of resin to pigment. The results of the analysis of the composite samples are tabulated in Table 6.

[FIGURE 19 OMITTED]

After application and cure, similar levels of all the HALS were found in the clearcoats, indicating only minor loss of any of the HALS from the clearcoat of systems applied on TPO. There was some loss of both nonreactable and reactable HALS during cure from the basecoats, with greater loss occurring for the nonreactable HALS, especially HALS 2 UR. HALS 3 UR migrates less than HALS 2 UR, presumably due to its more highly polar structure, which makes it less soluble in the TPO and more compatible with the coating.

After 3000 hours of xenon exposure, significant loss of HALS 2 UR occurred from both the clearcoat and basecoat of the coatings on TPO. Only 20% of HALS 2 UR remained in the clearcoat, and only 10% was detected in the basecoat. In this study, there was no corresponding system containing HALS 2 UR on steel; however, previous analytical studies (6) found that HALS 2 UR exhibited excellent longevity in coatings applied on steel substrates, and that the loss of HALS 2 UR from the topcoat was due to migration into the TPO, and not due to consumption of the HALS during exposure.

HALS 3 UR exhibited much better retention than HALS 2 UR in both basecoat and clearcoat on TPO; however, its retention was still not as good as that of HALS 1 R, which had retention of 60% in the clearcoat and 60% in the basecoat after 3000 hr of xenon exposure.

In addition to the analysis of the composite clearcoat and basecoat samples, concentration versus depth profiles were obtained on some samples by analysis of the individual microtomed layers. Figure 16 shows a concentration versus depth profile for HALS 1 R and HALS 2 UR in combination with TZ 2 on TPO after 3000 hours of xenon exposure. All the formulations were stabilized in basecoat and clearcoat. These results indicate that reactable HALS 1 R remained in the topcoat during exposure, while nonreactable HALS 2 UR migrated deep into the TPO.

[FIGURE 20 OMITTED]

PERFORMANCE IN XENON--EFFECT OF HALS REACTABILITY: The most dramatic effect of HALS reactability on weathering performance was seen in the systems applied on TPO which contain no basecoat stabilization (Figure 17A). Formulations stabilized with reactable HALS 1 R in combination with either the reactable BZT 1 R or BZT 2 UR showed improved gloss retention versus the same UV absorbers with HALS 3 UR stabilization. As expected, the same formulations on steel exhibited a difference in performance between the reactable and nonreactable systems. (Figure 17B).

In the fully basecoat + clearcoat-stabilized systems, the performance of the reactable HALS 1 R versus the nonreactable HALS 3 UR was not yet as clearly differentiated; however, the performance of the nonreactable HALS 2 UR was clearly inferior to the other two HALS (Figure 18). These results correlate with the analytical results (Table 6), which show that HALS 2 UR had migrated out of the topcoat to a much greater extent than either HALS 1 R or HALS 3 UR. It is anticipated that differentiation between HALS 1 R and HALS 3 UR will occur upon further exposure. The improved retention of HALS 3 UR reflects the fact that HALS 3 UR is more compatible in this polar coating, and presumably less soluble in the TPO, than HALS 2 UR, which has a very hydrophobic and nonpolar structure. The statistical analysis also confirmed no significant difference between HALS 1 R and HALS 3 UR in gloss retention in basecoat + clearcoat-stabilized systems.

The practical application of these findings is illustrated in Figure 19. Here, the performance of the reactable stabilizer systems without basecoat stabilization is compared with the nonreactable stabilizer systems with both basecoat and clearcoat stabilization. It can be seen that the improved retention of reactable stabilizers both before and after exposure allows for more efficient use of stabilizers--i.e., rather than having to "overstabilize" with nonreactables in both basecoat and clearcoat in order to compensate for migration, more realistic levels of reactable stabilizers will perform well.

PERFORMANCE IN XENON EXPOSURE--EFFECT OF BASECOAT STABILIZATION: In the thin-film study, a benefit for basecoat stabilization was seen for almost all the formulations tested (Figure 20A). This effect was borne out by the DOE analysis, which indicated a significant benefit for all three factors--basecoat stabilization, UVA type, and stabilizer reactability (Figures 21 A and B). In the normal film build study on TPO, a significant benefit of basecoat stabilization was seen in the systems containing the nonreactable benzotriazole UV absorber BZT 2 UR in combination with the nonreactable HALS 3 UR, followed by BZT 2 UR with the reactable HALS 1 R (Figure 20B). The combination of reactable BZT 1 R and nonreactable HALS 3 UR also benefits from basecoat stabilization. The combinations of either reactable or nonreactable triazine UV absorbers with either HALS has not yet exhibited a performance advantage for basecoat stabilization in the normal film build study. There was no significant effect of basecoat stabilization on steel for any of the combinations.

The combined results of the normal film build and thin-film studies again correlate with the migration behavior and absorbance characteristics of the UV absorbers and HALS on TPO. The completely nonreactable system showed the greatest benefit from basecoat stabilization, followed by the systems where either the UV absorber or HALS was reactable. The completely reactable systems showed the least difference in performance between the basecoat stabilized and unstabilized. Presumably, in the nonreactable systems, the presence of the stabilizer in the basecoat may provide a "buffer" for migration from the clearcoat for a certain amount of time, allowing some stabilizer to remain in the clearcoat and provide protection.

[FIGURE 21A OMITTED]

[FIGURE 21B OMITTED]

CONCLUSIONS

Through the use of appropriate experimental design, it was possible to determine which variables affect the performance of coatings on plastic substrates. It was also possible to quantify the magnitude of these effects and to determine interactions between them.

The variables evaluated in this study were UV absorber type, UV absorber reactability, HALS reactability, and basecoat stabilization. Of these variables, it was found that UV absorber reactability had the most profound effect on coating performance versus xenon exposure for this system applied on TPO. Hydroxyl functional reactable UV absorbers significantly outperformed nonreactable UV absorbers. It was also found that a primary hydroxyl functional, 100% reactable UV absorber gave the best performance, followed by a secondary hydroxyl functional UVA. These both outperformed a 50% reactable hydroxyl functional UV absorber. It was also observed that a co-condensible functionality, such as hydroxy, reacted sufficiently rapidly with the crosslinker to effectively immobilize the stabilizer in the topcoat before it had a chance to migrate out of the topcoat.

In study #2, the UV absorber type appeared to have a significant effect on the performance of the nonreactable UV absorbers, with the triazine TZ 2 UR outperforming its BZT 2 UR counterpart. However, this effect is somewhat confounded by the fact that the TZ 2 UR UV absorber also had a significantly higher starting absorbance than BZT 2 UR when it was incorporated at equal levels.

HALS reactability alone also had a significant positive effect on performance in systems without basecoat stabilization. It is anticipated that the effect will also be significant after longer exposure to the basecoat-stabilized systems. In addition, the combined effects of HALS reactability and polarity characteristics significantly affected both the performance and migration behaviors. The nonpolar HALS 2 UR performed poorly even in the basecoat-stabilized systems relative to both the reactable HALS 1 R and the nonreactable, but polar, HALS 3 UR. The performance also correlated with the extensive migration of HALS 2 UR from the topcoat into the TPO substrate.

These results also support the conclusions of previous studies (6) that very little migration of stabilizers from topcoat to TPO substrate occurs during initial cure, and that stabilizer migration occurs chiefly during subsequent exposure. They also support the assumption that loss of stabilizer from the topcoat correlates with loss of weathering performance, and that there is a direct relationship between performance, extent of migration, and degree of reactability of the stabilizers.

Finally, although the "ideal" reactable stabilizers system would consist of a fully reactable primary hydroxyl functional UV absorber and HALS, the commercially available secondary hydroxyl UV absorbers (TZ 1 R) and primary hydroxyl functional HALS 1 R provide outstanding resistance to migration in coatings on plastics. Hence, they also provide the associated significant advantages in weathering performance on nonreactable light stabilizers.

Table 1 -- Application Conditions

Series            Substrate  Primer                   Basecoat DFT

Normal DFT        TPO        Conductive AdPro         23-28 microns
Low DFT           TPO        Conductive AdPro         23-28 microns
Steel Normal DFT  Steel      E-coat, Primer-surfacer  23-28 microns

Series            Clearcoat DFT

Normal DFT        38-43 microns
Low DFT           20-23 microns
Steel Normal DFT  38-43 microns

Table 2 -- Design of Study #1 2K Flexible Urethane Clearcoat with Medium
Blue Metallic Basecoat

Substrate      3% UVA Clearcoat  1.8% HALS Clearcoat

1T  TPO        TZ 1 R            HALS 1 R
2T  TPO        BZT 1 R           HALS 1 R
3T  TPO        TZ 2 UR           HALS 2 UR
4T  TPO        BZT 2 UR          HALS 2 UR
5T  TPO        Unstabilized
6T  TPO        BZT 1 R           None
1S  Steel      TZ 1 R            HALS 1 R
2S  Steel      BZT 1 R           HALS 1 R
3S  Steel      TZ 2 UR           HALS 2 UR
4S  Steel      BZT 2 UR          HALS 2 UR
5S  Steel      Unstabilized
6S  Steel      BZT 1 R           None

Substrate      Comment

1T  TPO        Reactable TZ/Reactable HALS
2T  TPO        Reactable BZT/Reactable HALS
3T  TPO        Nonreactable TZ/Nonreactable HALS
4T  TPO        Nonreactable BZT/Nonreactable HALS
5T  TPO
6T  TPO
1S  Steel      Reactable TZ/Reactable HALS
2S  Steel      Reactable BZT/Reactable HALS
3S  Steel      Nonreactable TZ/Nonreactable HALS
4S  Steel      Nonreactable BZT/Nonreactable HALS
5S  Steel
6S  Steel

Table 3 -- Design of Study #2 2K Flexible Urethane Clearcoat with Light
Tan Metallic Basecoat

              Factorial Design A
      2.8%        1.0%         2.5%        1.0%
     UVA CC      HALS CC      UVA BC      HALS BC

 1   BZT 1 R     HALS 1 R     BZT 1 R     HALS 1 R
 2   TZ 1 R      HALS 1 R     TZ 1 R      HALS 1 R
 3   BZT 2 UR    HALS 1 R     BZT 2 UR    HALS 1 R
 4   TZ 2 UR     HALS 1 R     TZ 2 UR     HALS 1 R
 5   BZT 1 R     HALS 3 UR    BZT 1 R     HALS 3 UR
 6   TZ 1 R      HALS 3 UR    TZ 1 R      HALS 3 UR
 7   BZT 2 UR    HALS 3 UR    BZT 2 UR    HALS 3 UR
 8   TZ 2 UR     HALS 3 UR    TZ 2 UR     HALS 3 UR

              Factorial Design B

 1  1,2,7,8 are same formulations as above--used in
 2  Factorial Design B
 7
 8
 9   BZT 1 R     HALS 1 R     None        None
10   TZ 1 R      HALS 1 R     None        None
11   BZT 2 UR    HALS 3 UR    None        None
12   TZ 2 UR     HALS 3 UR    None        None
13   BZT 2 UR    HALS 1 R     None        None
14   BZT 1 R     HALS 3 UR    None        None

              Additional Formulations
15   TZ 2 UR     HALS 2 UR    TZ 2 UR     HALS 2 UR
16   BZT 2 UR    HALS 2 UR    BZT 2 UR    HALS 2 UR
17   BZT 3 R     HALS 1 R     BZT 3 R     HALS 1 R
18   None        None         None        None

    Factor 1  Factor 2      Factor 3
      UVA       UVA          HALS
      Type    Reactivity    Reactivity

 1  BZT            R             R
 2  TZ             R             R
 3  BZT            U             R
 4  TZ             U             R
 5  BZT            R             U
 6  TZ             R             U
 7  BZT            U             U
 8  TZ             U             U

    UVA         Stabilizer    Basecoat
    Type      Reactability  Stabilization

 1  BZT            R             +
 2  TZ             R             +
 7  BZT            U             +
 8  TZ             U             +
 9  BZT            R             -
10  TZ             R             -
11  BZT            U             -
12  TZ             U             -
13            Additional formulations complete
              a 4 factor d-optimal design with UVA
              type, UVA react., HALS react, bc stab. factors
14

                   Comments
15            Additional formulations to test HALS 2
              UR versus HALS 3 UR
16
17            Primary OH functionalized BZT 3 R
18            UNSTABILIZED CONTROL

DesignExpert [R] Software, from Statease, was used both for the
generation of the design and analysis of the results.

Table 4 -- Determination of Reactability of Light Stabilizers

Stabilizer     Extracted Level   Percent Retained

BZT 1 R         64                36
TZ 1 R          28                72
BZT 3 R          0               100
BZT 2 UR        92                 8
TZ 2 UR        100                 0
HALS 1 R         0               100
HALS 2 UR       94                 6

Table 5 -- Calculation of UV Absorber Levels in Clearcoat of a Basecoat/
Clearcoat System Applied on TPO and Steel. Results Obtained by UV
Microscopy and Extraction

                                Initial              3000 hr Xenon
                                UV Microscopy        UV Microscopy
                     Method:    (Extraction)         (Extraction)
UVA                  Substrate  % UVA in Clearcoat   % UVA in Clearcoat

Basecoat Stabilized
2.7% BZT 3 R          TPO           2.8                  3.1
2.7% BZT 1 R          TPO           2.5                  1.9
2.7% TZ 1 R           TPO           2.4                  2.0
2.7% BZT 2 UR         TPO           2.4 (2.3)            0.90 (0.45)
2.7% TZ 2 UR          TPO           2.3 (2.6)            0.72 (0.78)
2.7% BZT 2 UR         Steel         2.7                  2.3
2.7% TZ 1 R           Steel         2.3                  1.9
BC Unstabilized
2.7% TZ 1 R           TPO           1.6 (a)              1.4 (a)
2.7% BZT 2 UR         TPO           1.8 (a)              0.4 (a)

(a) In coatings where basecoat is unstabilized, equilibration between
basecoat and clearcoat results in theoretical level of 1.73% in the
clearcoat and 1.73% in the basecoat.

Table 6 -- Results of HALS Analysis Before and After Xenon Exposure in
Systems on TPO and Steel Fully Stabilized with 2.7% UV Absorber + 1%
HALS in Basecoat and Clearcoat

                   HALS Level--Percent on Coating Resins Solids
              Clearcoat          Basecoat                 TPO
           Initial  3000 hr  Initial  3000 hr  Initial  3000 hr  3000 hr

                                                        Top 10   100-200
                                                        microns  microns
HALS 1 R     1.0      1.0      1.0      0.9     0.1       0.2      0
HALS 3 UR    1.2      0.7      1.4      0.8     0.03      --       0.02
HALS 2 UR    0.9      0.2      0.8      0.1     0.2       0.1      0.1

References

(1) Ryntz, R., Adhesion to Plastics, Polymer Surfaces and Interfaces, Urban, M. (Ed.) Global Press, Chicago, IL, 1998.

(2) Light, F., Microtome Technology--Distribution of Stabilizers in Multilayer Coatings and Plastic Coatings, Proc. of Addcoat 2001, Orlando, FL, 2001.

(3) Haacke, G., Andrawes, F.F., and Campbell, B.H., "Migration of Light Stabilizers in Acrylic/Melamine Clearcoats," JOURNAL OF COATINGS TECHNOLOGY, 68, No. 855, 57 (1996).

(4) Andrawes, F., Valcarcal, T., Haacke, G., and Brinen, J., "Depth Distribution of Light Stabilizers in Coatings Analyzed by Supercritical Fluid Extraction-Gas Chromatography and Time-of-Flight Secondary Ion Mass Spectrometry," Anal. Chem., 70, p. 3762 (1998).

(5) Cliff, N., Yaneff, P., and Adamsons, K., "Reactable Light Stabilizers for Automotive Coatings Over Plastic Substrates," in Proc. of Coatings for Plastics Symposium, Troy, MI, 2001.

(6) Yaneff, P., Adamsons, K., Cliff, N., and Kanouni, M., "Migration of Reactable UVAs and HALS in Automotive Plastic Coatings and Their Impact on Durability," Presented at the 30th Annual International Waterborne High-Solids and Powder Coatings Symposium, New Orleans, LA, 2003.

(7) Malik, J., Hrivik, A., and Tuan, D.Q., "Diffusion and Solubility of Hindered Amine Light Stabilizers in Polyolefins--Influence of Stabilization Efficiency and Implications for Polymer-Bound Stabilizers," in Polymer Durability, American Chemical Society, Washington, D.C., p. 4555-4671, 1996.

(8) Ligner, G. and Malik, J., "How to Increase HALS Persistence in Polyolefins Without Jeopardizing Stabilization Efficiency," in Proc. of Addcon 96: Worldwide Additive Polymer Modification Conference, 1996.

(9) Callais, P., Kamath, R., and Sargent, J., "Synthesis of Acrylic HSC Resins with Peroxides Containing Hindered Amine Light Stabilizers," Mod. Paint Coat., p. 41-47 (1988).

(10) Avar, L. and Bechtold, K., "Studies on the Interaction of Photoreactive Light Stabilizers and UV-absorbers," Prog. Org. Coat., 35 (1999).

(11) Galbo, J., Ravichandran, R., and Holt, M., "Non-Migrating 1-Hydroxycarbyloxy-2,2,6,6-Tetramethylpiperidine 1,3,5-Triazine Derivatives," Ciba, U.S. (1993).

(12) Ohrbom, W., Aubin, D., Boisseau, J., Oberg, P., and Rehfuss, J., "Durability Enhancing Agents, Method Therefore and Cured Coating Compositions Containing the Same," BASF, U.S., 2000.

(13) Dawson, W. and Gupta, G., "UV Curable Coatings Having Improved Weatherability," The Sherwin Williams Company, U.S., 1996.

(14) Elan, G., Ludwig, B., and Andrus, M., "Polymeric Compositions," in U.S. 6251521, 3M, U.S., 2001.

(15) Ludwig, B. and Elan, G., "UV-Absorbing Core/Shell Particles," in U.S. 6312807 B1, 3M, U.S., 2001.

(16) Pickett, J., "Compounds Based on Dibenzoylresorcinol and Related Compositions and Articles," General Electric Company, U.S., 2002.

(17) Pickett, J., "Silylated Agents Useful for Absorbing Ultraviolet Light," General Electric Company, U.S., 1995.

(18) Ashby, B. and Schroeter, S., "Ultraviolet Light Absorbing Agents and Compositions and Articles Containing Same," General Electric Company, U.S., 1981.

(19) Turro, N.J., Step, E.N., Gande, M.E., and Klemchuk, P.P., "Model Studies on the Mechanism of HALS Stabilization," Die Angewandte Makromolekulare Chemie, 232, p. 165-83, 1995.

(20) Cliff, N., "Reactable Hindered Amine Light Stabilizers for Coatings," in Advanced Coatings Technology, Detroit, MI, 1998.

(21) Montgomery, D., Design and Analysis of Experiments, New York, John Wiley and Sons, 2001.

(22) Smith, C., Gerlock, J.L., and Carter, R.O., Poly. Degrad. Stab., 72, p. 89-97 (2000).

(23) Stoecklein, W. and Fujiwara, H., "The Examination of UV-Absorbers in 2-Coat Metallic and Non-Metallic Automotive Paints.," Science and Justice, 39(3), p. 188-195 (1999).

(24) Dudler, V. and Muinos, C., ACS Advances in Chemistry Series, No. 249, Clough, R.L., Billingham, N.C., and Gillen, K.T. (Eds.), American Chemical Society, Washington, D.C., p. 441-453, 1996.

(25) Frank, H.P. and Lehner, H., J. Poly. Sci.,: Part C, 31, p. 193-203 (1970).

(26) Gerlock, J., Kucherov, A., and Matheson, R., "Determination of Active HALS in Weathered Automotive Paint Systems I: Development of ESR Based Analytical Techniques," Poly. Degrad. Stab., 69, p. 1-9 (2000).

(27) Gerlock, J., Kucherov, A., and Smith, C., "Determination of Active HALS in Automotive Paint Systems II: HALS Distribution in Weathered Clearcoat/Basecoat Paint Systems," Poly. Degrad. Stab., 73, p. 201-210 (2001).

Nancy Cliff and Mouhcine Kanouni -- Ciba Specialty Chemicals Corporation*

Cindy Peters -- Ford Research Laboratories ([dagger])

Phillip V. Yaneff** and Karlis Adamsons ([double dagger]) -- Dupont Performance Coatings

* 540 White Plains Rd., Tarrytown, NY 10591.

[dagger] Scientific Research Laboratory, 20000 Rotunda Dr., P.O. Box 2053, MD 3182, Dearborn, MI 48121.

** 377 Fairall St., Ajax, Ont. L1S 1R7, Canada.

[double dagger] 3401 Grays Ferry Ave., Philadelphia, PA 19146.