Concurrent physical aging and degradation of crosslinked coating systems in accelerated weathering.

Abstract Coating degradation is a combination of both chemical and physical processes; however, physical processes have not received much attention. Physical aging has a non-negligible effect on coatings' mechanical properties and permeability etc. through the densification that continues as a polymer approaches its thermodynamic equilibrium below the glass transition temperature, [T.sub.g]. Observations in recent work showed that physical aging affects coatings' mechanical property response during accelerated weathering and is, itself, affected by the associated chemical degradation. Two crosslinked coating systems were studied in order to compare different chemical compositions, their [T.sub.g], and their thermal response in accelerated weathering. During thermal cycling, physical aging measured by enthalpy recovery exhibited different trends in the two coatings. A "rejuvenation" mechanism was observed in the coating with a [T.sub.g] between the top and bottom limits of the exposure cycle; continued aging was observed for the coating with a high [T.sub.g]. Stress relaxation tests detected aging and "memory" behavior over periods comparable with accelerated weathering cycles. Both thermal and mechanical responses changed in complicated and different ways as the coatings degraded. Different degrees of coating thickness reduction were observed in both isothermal relaxation and degradation. When various coatings are evaluated, simply judging their performance under the same weathering environment is not reliable since polymer relaxation behavior depends on the relationship between the exposure temperatures and the [T.sub.g] of each polymer.

Keywords Physical aging, Degradation, Thermal cycling, Relaxation, Accelerated weathering

Introduction

Signification 

Protective attributes of coatings, such as scratch resistance, corrosion protection, etc., are related to the coating integrity and thus to mechanical properties. Changes in mechanical properties are significant in their own right but they also share, with other protective properties, sensitivity to cracks, crosslink density, temperature, swelling by water, etc. Studies on coatings' protective properties and their durability have most often been concerned with ultraviolet irradiation spectra and ultraviolet absorption, but moisture absorption and temperature variation can be very important also. (1), (2)

Structural relaxation, known as physical aging, is a kinetic phenomenon wherein glassy polymers tend to relax since they are not in an equilibrium situation when stored at a temperature lower than the glass transition temperature ([T.sub.g]). It is well known that physical aging is always accompanied with densification and reduction in fractional free volume, and reduction of segmental mobility. Perera (3) has claimed that physical aging phenomena have an impact on the formulation, application, and the service life behavior of organic coatings in many domains. Decreasing gas and water vapor permeability (4-6) with physical aging was observed in crosslinked polymer films. Studies on physical aging reported in the literature (3-11) are mainly concerned with changes in the thermal and mechanical properties with continuous exposure to simple isothermal conditions. If temperature varies, polymers age physically more when the temperature is decreasing than when it is increasing. A negative temperature step induces both a slowing down of the physical aging and a thinning of the spectrum of relaxation times; a positive temperature jump above [T.sub.g] results in a "rejuvenation" followed by acceleration of physical aging. (12), (13) By referring to the consequences of physical aging for polymers, the possible effects upon coatings' long-term properties were listed in Fig. 1. Compared to the sensitivity of coatings' durability to details of the ultraviolet irradiation spectrum, physical aging may not be the most important phenomenon in coatings' field usage, but the illustration in Fig. 1 reminds us that many properties depend on it and thus it cannot be ignored.

[FIGURE 1 OMITTED]

Degradation

A common idea about physical aging is that it is separated from curing or degradation phenomenon (chemical aging, damage, etc.) due to reversibility. (14) Physical aging may be desirable due to the diminishing of permeability and increase of strength, but it is undesirable as coating films get brittle with aging. However, what really happens in nature is more complex. Recently, some combined effects with physical aging drew researchers' attention. Morancho and Salla (15) noticed how physical aging and curing of epoxy are linked; Zheng and McKenna (16), (17) have studied the effect of moisture content on the physical aging responses of glassy polymers; McCaig (18) also reported that crosslinking slowed the rate of physical aging significantly but did not completely stop the aging process. Until now, very few studies have addressed the complexity of physical aging with degradation of the crosslinked structure.

The presumption of this study, agreeing with some of the above observations, is that physical aging phenomena depend on the network composition and thus depend on the chemical structure and concentration, and, inevitably, depend on the order in which polymer segments are removed during the degradation processes. When exposed to weathering, the loss of neighboring polymer chain segments due to photodegradation may either permit faster aging relaxation temporarily, due to diminished molecular packing, or it may effectively drive the system further from structural equilibrium and thus prolong aging relaxation. Vice versa, possibly spontaneous chain rearrangements and associations occur while approaching equilibrium under a given exposure temperature, which might affect the rate of coatings degradation. (19) This study investigates how physical aging affects crosslinked polymeric coatings when degraded by irradiation and thermal exposure, which complicates the link between composition and properties during long-term exposure.

Accelerated and natural weathering

Accelerated weathering has been developed as a rapid substitute for natural weathering. Several processes, such as irradiation, temperature and humidity cycles in weathering, as claimed above and elsewhere (20) might, or might not, be independent of each other and result in combined effects on coatings, property changes; for example, ASTM D5894 has reported that increasing fog/dry cycle time produced slower degradation. Figure 2 demonstrated the temperature difference of natural and accelerated weathering. As a comparison, the [T.sub.g] of some industrial coatings were collected from online and printed literature; [T.sub.g] of an acrylic-melamine coating is 45[degrees]C (21); for a polyester-triglycidyl isocyanurate coating it is 58-69[degrees]C (22); a few thermosetting powder coatings based on triglycidyl isocyanurate have [T.sub.g] in the range 50-70[degrees]C (23); and a methacrylic copolymer coating has a value of 110[degrees]C (24), etc. With such a wide range of glass transition temperatures, we suspect that accelerated and natural weathering might allow different relaxation and therefore degradation mechanisms depending on the difference between exposure temperature, the rate of change in temperature, and the coating's [T.sub.g]. Even the same formulation used in different locations will experience different aging as well as degradation. The study of physical aging would help to understand how physical aging proceeds in cyclic testing, all of which will be valuable in relating the accelerated weathering tests to performance in the field. ASTM D6695, a widely used standard accelerated weathering procedure, was used in this study to provide an environment with combined thermal cycling and light irradiation.

[FIGURE 2 OMITTED]

Polyester-urethane and epoxy-polyamide are currently used as aerospace top and base coatings, both of which have [T.sub.g] within the exposure temperature cycles, and they have different sensitivity to water, etc. Though epoxy is not durable against ultraviolet exposure, the two materials were compared in this study to examine the different crosslinking structures, [T.sub.g], their behavior under thermal cycling, and concurrent degradation environment. No additive was included because our aim was to get insight about the basic crosslinked coating's behavior in weathering.

Experimental details

Specimen preparation

Polyurethane samples were obtained by reacting polyester with a HDI trimer in the presence of 0.05% dibutyltin dilaurate catalyst, as described elsewhere(5) the epoxy sample was formulated by mixing Epon 828 with EPIKURE 3115 curing agent at a 1:1.1 stoichiometry ratio. Coatings were applied onto substrates covered with DuPont Tedlar [R] film, from which a free film could be peeled easily. A dry film thickness of 60-80[micro]m was targeted. Polyurethane was cured at ambient temperatures for 1 week followed by a force-dry at 85[degrees]C for 1 h; epoxy was oven cured at 85[degrees]C for 45 min immediately after application. Both coatings were subsequently baked at 120[degrees]C for 30 min to remove entrapped solvent entrapped and to achieve complete cure. This ensured that [T.sub.g] reached a maximum (polyurethane 54.6[degrees]C and epoxy 77.4[degrees]C), no residual cure exotherm could be observed by differential scanning calorimetry (DSC), and no weight change by thermogravimetric analysis (TGA). Heating above a coating's [T.sub.g], then quenching, defines the beginning of the aging process(1),(14). IR spectroscopy confirmed that coatings were fully cured before any thermal treatment or weathering exposure.

Exposure

For simple isothermal aged samples, after being annealed far above [T.sub.g], samples were quenched to the required aging temperature and tested after various aging times. Isothermal aging was practiced in an oven and desiccator to avoid water absorption from the ambient environment. Q-Sun[R] chamber exposure with xenon light irradiation for 4 h at 58[degrees]C (air temperature) followed with 4 h at 25[degrees]C without water spray (to avoid the effect from moisture uptake) was used as accelerated weathering exposure. Thermocouple measurements showed that the coating temperature during the hot cycle was 55[degrees]C while the air temperature reached 58[degrees]C.

Characterization

Thermal analysis 

DSC was employed to study the effects of physical aging quantitatively by measuring the enthalpy changes in the glassy phase, with specimen size of 5-8 mg and scanning from -25[degrees]C to 130[degrees]C at 5[degrees]C/min heating rate. The same heating and cooling rate was maintained for all the testing (1),(8). As one should expect, aging is seen to occur even at room temperature and this must be taken into account in making enthalpy change measurements. Dynamic mechanical thermal analysis (DMTA) was used to measure and calculate the crosslink density from the storage modulus at 130[degrees]C(5) and TGA was used to measure the amount of water absorbed during exposure.

Mechanical properties

Stress relaxation was measured, as degradation proceeded, in an Instron 5545 tensile testing machine with an environmental chamber. Samples free of bubbles or other visible defects were cut to dumbbell shape with a die (ASTM 412 type C), with an elastic effective gauge length of 53.2 mm, and were uniaxially deformed to 1% strain and the stress measured as a function of time. The crosshead speed was 53.2 mm/min, so the strain was applied in 0.6 s. All the tests were replicated three to four times. Relaxation under repeated thermal cycling was examined on specimens with dimensions of approximately20 x 5 mm wide and thickness of 0.08-0.1 mmin a dynamic mechanical analyzer since temperature control was superior. After annealing at 130[degrees]C, cycling between 25 and 55[degrees]C was done at 4 h intervals to simulate accelerated weathering while the stress decay, at a fixed 0.5% strain, was recorded simultaneously.

Results and discussion

Crosslink density and modulus

The thermal and mechanical behavior of the two coating systems was examined after different accelerated weathering periods. As can be observed in Fig. 3, [T.sub.g] (by DSC) and the crosslink density (by DMA) kept decreasing with weathering exposure, while simultaneously the tensile modulus measured at 25[degrees]C increased, shown in Fig. 4. This complicated phenomenon is one of the motivations for starting this study, to investigate molecular relaxation changes occurring in a polymer during weathering as an explanation to supplement chemical degradation for these observations.

[FIGURE 3 OMITTED]

Enthalpy relaxation

"Enthalpy relaxation" (or "recovery") is the difference in enthalpy between an aged glass and a glass newly equilibrated above [T.sub.g] and is a common quantification of the overall extent of physical aging. This enthalpy difference can be easily quantified by doing a heat-cool-heat run on DSC.

[FIGURE 4 OMITTED]

Temperature dependence of structure relaxation in isothermal aging

Figure 5 shows the nonlinear temperature dependence of enthalpy found for polyurethane and epoxy aged for 8 h at varying temperatures. Commonly, mechanical and thermal properties of polymers are much more nonlinear than other properties. The location and intensity of the endothermic peaks varied with the aging temperature. Enthalpy relaxation rate increases as the temperature approaches [T.sub.g] from below but it disappears at temperatures very close to or above [T.sub.g], since the glass is then close to thermodynamic equilibrium. At very low temperatures, values of enthalpy recovery are very low since molecular motion becomes very limited. This has been noted by several authors (25), particularly when studying crosslinked polymers. Enthalpy relaxation of the epoxy reached a maximum at around 55[degrees]C ([T.sub.g]-[T.sub.aging] [approximately equal to] 20[degrees]C) and polyurethane exhibits pronounced enthalpy relaxation at 25 and 35[degrees]C ([T.sub.g]-[T.sub.aging] [approximately equal to] 20-30[degrees]C); these results agree with most studies (3-10), where characterization of physical aging was performed in the neighborhood of [T.sub.g]

[FIGURE 5 OMITTED]

Referring to the temperature variation of accelerated weathering in Fig. 2. we predict that different coating systems, such as the two in this study, may have different relaxation rate and variable performance when exposed to the cycling of weathering.

Enthalpy relaxation of polyurethane and epoxy samples was studied after more extended continuous thermal aging at 25 and 55[degrees]C. One can observe by comparing the results in Fig. 6 that PU exhibits continuous enthalpy recovery at 25[degrees]C, as does epoxy at 55[degrees]C; however, little relaxation at 55[degrees]C for PU and 25[degrees]C occured for epoxy even after being thermally aged for over 1000 h. This result is consistent with the results in Fig. 5, since relaxation will diminish as the polymer is aged closer to its [T.sub.g] (polymer is closer to being in equilibrium), as for the polyurethane, and it will also diminish further away from its [T.sub.g] (since the mobility is much diminished further below [T.sub.g]), as in the case of the epoxy.

[FIGURE 6 OMITTED]

Thermal cycling effect

In this experiment the two coating systems were cycled, at 4 h intervals, between 25 and 55[degrees]C a number of times. It was found that the two crosslinked coatings responded in different ways (Fig. 7). Epoxy has a continuous aging accumulation with cycle number. For polyurethane, an aging/recovery mechanism was observed in the thermal cycling, that is, material relaxes at 25[degrees]C and when temperature reaches above [T.sub.g] in hot cycle, the system recovers almost to equilibrium.

[FIGURE 7 OMITTED]

Overall trends after weathering (including degradation)

Decreasing enthalpy relaxation, as the crosslink density increased, in model epoxy thermosets was reported by Skaja(7) and Cook, (26) which was explained by the increasing topological restrictions on the process of segmental re-organization. Systems examined in this study had crosslinking reduced by degradation, which damaged the network and resulted in less inter-chain connection. Figure 8 describes the overall trend in enthalpy relaxation after some periods of accelerated weathering. Enthalpy recovery of polyurethane decreases with long-term exposure, which can be explained by the aging/recovery mechanism since the upper temperature is close to [T.sub.g]. Enthalpy relaxation in the epoxy increases continuously with exposure since it is never close enough to [T.sub.g] for rejuvenation.

[FIGURE 8 OMITTED]

No obvious enthalpy changes were observed in the polyurethane when exposed to a dry or the standard exposure with water spray due to its more hydrophobic nature. In the case of a more hydrophilic coating, such as epoxy, analysis of enthalpy relaxation in weathering becomes more complicated due to the involvement of water as an effective plasticizer.(27) For epoxy exposed to a weathering environment with water, enthalpy recovery experiments showed a very broad separation between the aged and the rejuvenated calorimetric curves. TGA tests showed that epoxy would take up almost 2% water during a 2-week exposure period and the broad separation in the calorimetric traces was due to the first curve including the effect of plasticization and the heat required to raise the temperature of water and volatilize it. However, in the accelerated exposure cycling without moisture, the difference in the curves was much more typical of enthalpy recovery seen in other polymers.

Crosslink density effect

As opposed to the previous section, which showed the overall thermal history of coatings when exposed to accelerated weathering, this section is to investigate the aging potential at specific temperatures.

As shown in Fig. 3, accelerated weathering results in decreasing crosslink density with exposure time. Each chemical structure has its own thermodynamic equilibrium state, toward which it ages. Coating samples with different degrees of crosslink density (different exposure times) were aged and examined. The thermal history of each sample was removed by holding at 130[degrees]C for 20 min in the DSC. Then enthalpy relaxation was measured after the same length of thermal aging at the labeled temperatures.

As shown in Fig.9, the endothermic peak becomes wider and the area of the peak increases with longer exposure time, that is, diminishing crosslink density. This result suggests that, with degradation, coatings have more potential to relax at a particular exposure temperature, although there is less inter-connection between molecules as exposure progresses. This result is consistent with one of our hypotheses, that is, degradation that results in polymer chains being broken results in a looser physical structure, which has more relaxation potential.

[FIGURE 9 OMITTED]

Stress relaxation

The stress relaxation of crosslinked polymers is a field that is still not well understood and is currently attracting significant experimental and theoretical interest. In principle, the time dependence of relaxation contains much information about the internal physics of random polymer networks.(28) In stress relaxation tests, a constant strain is applied quickly and the stress in the polymer is measured as a function of time while the deformation is maintained. The decreasing of relaxation rate reflects a retarding response by thermal motion within the structure. By observing the relative position of relaxation curves, one can identify the microchange in the system.

Thermal cycling effect

Figure 10 demonstrates the relaxation curves of polyurethane and epoxy under two different thermal cycles, simulating accelerated weathering (25-55[degrees]C cycling as in ASTM D6695, 25-35[degrees]C cycling as in ASTM B 117), which is the very first trial that has been done to track the structure relaxation under thermal cycling in DMA.

[FIGURE 10 OMITTED]

In Fig. 10a, polyurethane in the low temperature cycles, there are small changes of relaxation rate after the first cycle. The increase of initial stress with cycling indicates that the structure still gets stiffer even being annealed/recovered at high temperature repeatedly. At 55[degrees]C cycles, relaxation curves are overlapped and are at very low stress, which indicates that the material reaches a close-to-equilibrium state at the high temperature (near [T.sub.g]) and gets soft, thus the material essentially starts over (but not completely) after each hot cycle. Figure 10b presents cycling with both temperatures lower than [T.sub.g] of polyurethane, and continued change in relaxation was observed with cycling number.

For epoxy in Figs. 10c and 10d, it can be seen that the relaxation rate decreases in successive cycles at 25, 55, and 35[degrees]C and the initial stress increases. The structure has responded significantly as the cycle number increases.

It is clear that the structure relaxation in each previous cycle has an effect on coatings' response in the current cycle. Secondly, most of the tests show that the first cycle is most distinctive from the following; this phenomenon is consistent with the exponential pattern of most relaxation phenomena; that is, greatest change at the start and then more gradual change with time. If we take the first cycle as the initial history that the coating experienced, this phenomenon confirms the idea that the temperature changes in the initial service life have an effect on the long-term behavior (durability) of the coating, as Perera and Eynde (29), (30) have discussed in their papers.

The comparison between the enthalpy relaxation (Fig. 7) and stress relaxation tests showed that the relaxation increase by physical aging followed a similar trend, but the aging rates measured from the two methods were different. For example, enthalpy relaxation was not as sensitive to changes in relaxation with cycle number as the mechanical tests. One possible explanation is that the mechanical tests examine the network in different ways than the enthalpy measurements, which is a topic that will be touched on again later.

Weathering effect

A decrease of crosslink density might reflect a loose and collapsed structure with weathering. Generally, relaxation time decreases with reducing crosslink density due to less chain bonding and molecular interactions. However, a different pattern was observed in the stress relaxation test of accelerated weathered coatings, which indicates the combined effect of photodegradation and structure relaxation, as exhibited in Fig. 11. For both polyurethane and epoxy, for the first few exposure periods seen in the data, the stress at any time is higher than in the unexposed sample; also the time required achieving equilibrium increases with exposure time. Increasing stress and relaxation time reflect toughening of the structure despite crosslink density diminishing. After further weathering, scission of the molecular chains dominates and we see the large decay in mechanical properties, polyurethane after 1000 h exposure, and epoxy after around 300 h exposure.

[FIGURE 11 OMITTED]

Note in Fig. 11 that relaxation at room temperature of materials with different degree of degradation/aging is more than 10 h, that is. much longer than the typical cycle length (1 or 4 h) of accelerated weathering cycles. Thus, in combination with the results in Fig. 10, we believe that in accelerated weathering the memory of one cycle will be carried over into the next. Thus the response of a polymer coating to accelerated weathering cycles will be different to its response in the slower cycles typical of natural weathering.

Modeling by KWW equation

Kohlrausch--Williams--Watts (KWW) model (31)-(34) is a very successful, semiempirical model for the time-dependence of glass properties. It accounts for a structural relaxation time distribution with a stretched-exponential parameter (or, nonexponentiality parameter). [beta], such that the time dependence can be described as:

[PHI](t) = exp { - [(t/[tau]).sup.[beta]]}

where [PHI](t) is often used to represent the changes of any physical properties with time following a change in temperature, stress, or electric field.

Since the application of KWW equation to a complete correlation with the microscopic dynamics in a real system is as yet lacking, (35) our target was to quantify the effects seen in the data and test the prediction of a modified KWW equation. as shown below, by stress relaxation testing on weathered coatings.

[sigma](t) = [[sigma].sub.0]exp [ - [(t/[tau]).sup.[beta]]] + [[sigma].sub.[infinity]]

where [sigma](t) = stress, as it relaxes with time, t; [[sigma].sub.0] = Pre-exponential stress parameter; [[sigma].sub.[infinity]] = value of stress supported by the completely relaxed material; [[sigma].sub.0] + [[sigma].sub.[infinity]] = initial value of stress; [tau] = relaxation time; and [beta] = nonexponential time parameter, lying between 0 and 1.

The trends in the parameters for the polyurethane are shown in Fig. 12a; as degradation proceeded, the value of [beta] was reduced. The change in these parameters for the epoxy system, seen in Fig. 12b, would indicate that relaxation becomes lengthier after exposure. However, one should point out that the epoxy aged physically at both the lower and upper temperature of the exposure cycle and the literature shows that it is common for relaxation times to increase as physical aging progresses. (36-38)

[FIGURE 12 OMITTED]

In contrast, the polyurethane had a [T.sub.g] very close to the upper temperature in the cycle so it may be "rejuvenated" in each cycle, so that we see changes in KWW parameters that are determined more by degradation than physical aging.

Variation in [T.sub.g] with exposure and measurement technique

Clearly, the data here show how physical aging changes the response of the polymer network and how the difference between the exposure temperature and [T.sub.g] is crucial. The [T.sub.g] of a coating and its surface temperature will depend on any pigment it contains (via pigment polymer interactions and heat adsorption), substrate it is attached to, environment, etc. The implications of different temperature scenarios to the aging process will be discussed later.

It is accepted that [T.sub.g] diminishes with decreasing crosslink density. It is interesting that if one calculates what the reduction should be, according to accepted theory, (39)-(41) the variation in [T.sub.g] is much greater than found here. This will be explored in the future.

If, as is commonly done, one measures [T.sub.g] using dynamic mechanical thermal analysis, using either the peak temperature of the tan [delta] curve or the onset temperature for the drop in storage modulus, there is a different trend with exposure (see Fig. 13).

[FIGURE 13 OMITTED]

Depending on whether the measurement of [T.sub.g] is done mechanically or calorimetrically leads to an expectation of the network gaining crosslinks or losing them. In addition, there are substantial differences in values, depending on the method, that may mislead expectations concerning physical aging and what is happening in other ways to the coating polymer. Results presented here suggest that changes in mechanical properties and physical aging can be rationalized using the calorimetric value of [T.sub.g], which may not be possible with the value from tan [delta], for example.

The standard definition of [T.sub.g] is not in terms of mechanical properties but in terms of enthalpy, volume, or entropy, which depend on the configuration and motion of each and every moiety on the polymer, large and small. Conventional mechanical properties, around the glass transition, depend on larger scale correlations over the network and do not examine the molecular structure in the same way as the properties that define the glass transition. Understanding how changes in mechanical properties relate more exactly to the glass transition remains an active area of research elsewhere.

Summary and conclusions

This research investigated the processes of polymer physical aging under circumstances where the environment allows relaxation by thermal cycling and simultaneously affects the polymer network by the action of photodegradation. Model polyester-urethane and epoxy-polyamide were used as representatives of crosslinked coatings with low and high [T.sub.g]. Here [T.sub.g] and coating temperature are based on free films of clear coatings to avoid the complication, at this stage, of any effect from pigments, substrate, etc. Mechanical and thermal tests were applied on coatings that were isothermal aged as well as exposed to accelerated weathering. From the experimental data, the following conclusions were drawn:

1. Thermal and mechanical behavior of polymeric coatings during weathering involves physical phenomena as well as chemical degradation. Physical aging is an explanation for the increase in mechanical stiffness that occurs when crosslink density is reduced during degradation.

2. Physical aging in weathering is dependent on materials' [T.sub.g] and how [T.sub.g] is related to the temperature of thermal cycling. A relaxation/rejuvenation mechanism was observed in low [T.sub.g] coatings, since a hot cycle with temperature higher than [T.sub.g] brings the system to equilibrium. In this situation, physical aging is interrupted periodically, but remains to affect mechanical properties. Other high-performance coatings with [T.sub.g] significantly above the usage temperature will be expected to experience a greater and inevitable effect from physical aging.

3. Sensitive small-strain stress relaxation tests have tracked structure changes with exposure cycle number: behavior during the first cycle is always different from the following cycles. The relaxation time of samples with any degree of aging is longer than the typical cycle length of accelerated weathering, thus the memory of one cycle will be carried over into the next. Relaxation may be faster in accelerated weathering than in natural weathering, which also contributes to the discrepancy of two weathering methods.

4. The juxtaposition of [T.sub.g] with respect to the exposed temperature is an important factor in physical aging. However, the mechanical characterization of [T.sub.g] can be significantly different to that characterized by calorimetry. Therefore it is important to know the differences between the exposed temperature and the [T.sub.g] from different methods, depending on the final application of a polymer.

5. Structure changes due to physical aging can lead to important variations between short-term data used to characterize the material and its long-term performance. Evaluating multiple coating materials with a single exposure cycle is not reliable. Relaxation behavior and its effects upon degradation is an important aspect that needs to be included when correlating accelerated weathering with natural weathering.

Acknowledgment This work was supported by the Air Force Office of Scientific Research, contract number FA9599-04-1-0368.

References

(1.) Fernando, BMD, Shi, XD, Croll, SG, "Molecular Relaxation Phenomena During Accelerated Weathering of a Polyurethane Coatings." J. Coatings Tech. Res. (2006) (in press)

(2.) Ochs, H, Vogelsang, J, Meyer, G, "Enhanced Surface Roughness of Organic Coatings due to UV-Degradation: An Unknown Source of EIS-Artifacts." Prog. Org. Coat., 46 182-190 (2003)

(3.) Perera, DY, "Physical Ageing of Organic Coatings." Prog. Org. Coat., 47 61-76 (2003)

(4.) Huang, Y, Paul, DR, "Experimental Methods for Tracking Physical Aging of Thin Glassy Polymer Films by Gas Permeation." J. Memb. Sci., 244 167-178 (2004)

(5.) Huang, Y, Wang, X, Paul, DR, "Physical Aging of Thin Glassy Polymer Films: Free Volume Interpretation." J. Memb. Sci., 277 219-229 (2006)

(6.) Kim. JH, Koros, WJ, Paul, DR, "Effects of C[O.sub.2] Exposure and Physical Aging on the Gas Permeability of Thin 6FDA- based Polyimide Membranes: Part 2. With Crosslinking." J. Memb. Sci., 282 32-43 (2006)

(7.) Skaja, A, Fernando, D, Croll, SG, "Mechanical Property Changes and Degradation During Accelerated Weathering of Polyester-urethane Coatings." J. Coat. Technol., 341-51 (2006)

(8.) Leveque, D, Schieffer, A, Mavel, A, Maire, JF, "Analysis of How Thermal Aging Affects the Long-term Mechanical Behavior and Strength of Polymer-Matrix Composites." Compos. Sci. Technol., 65 395-401 (2005)

(9.) Lee, A, Mckenna, GB, "Effect of Crosslink Density on Physical Ageing of Epoxy Networks." Polymer, 29 1812-1817 (1988)

(10.) Barral, L. Cano, J, Lopez, J, Lopez-Bueno, I, Nogueira, P, Abad, MJ, Ramirez, C, "Physical Aging of an Epoxy/Cycloaliphatic Amine Resin."Eur. Polym. J., 35403-411 (1999)

(11.) Ho, CH, Vu-Khanh, T, "Physical Aging and Time-Temperature Behavior Concerning Fracture Performance of Polycarbonate." Theor. Appl. Fract. Mech., 41 103-114 (2004)

(12.) Montes, H, Viasnoff, V, Jurine, S, Lequeux, F, "Ageing in Glassy Polymers Under Various Thermal Histories." J. Stat. Mech.: Theory Exp. March (2006)

(13.) Perera, DY, "Effect of Thermal and Hygroscopic History on Physical Ageing of Organic Coatings." Prog. Org, Coat., 44 55-62 (2002)

(14.) Bradshaw, RD, "Physical Aging in Polymers and Polymer Composites: An Analysis and Method for Time-Aging Time Superposition." Polym. Eng. Sci., 37 31-44 (1997)

(15.) Morancho, JM, Salla, JM, "Relaxation in Partially Cured Samples of an Epoxy Resin and of the Same Resin Modified with a Carboxyl-terminated Rubber." Polymer, 40 2821-2828 (1999)

(16.) Zheng, Y, McKenna, GB, "Structural Recovery in a Model Epoxy: Comparison of Responses After Temperature and Relative Humidity Jumps." Macromolecules, 36 2387-2396 (2003)

(17.) Zheng, Y, Priestley. RD, McKenna, GB, "Physical Aging of an Epoxy Subsequent to Relative Humidity Jumps Through the Glass Transition." J. Polym. Sci., Part B: Polym. Phys., 42 2107-2121 (2004)

(18.) McCaig, MS, Paul, DR, "Effect of UV Crosslinking and Physical Aging on the Gas Permeability of Thin Glassy Polyacrylate Films." Polymer, 40 7209-7225 (1999)

(19.) Feller, RL, Curran, M, Colaluca, V, Bogaard, J, Bailie, C, "Photochemical Deterioration of Poly (vinylbutyral) in the Range of Wavelengths from Middle Ultraviolet to the Visible." Polym. Degrad, Stab., 92 920-931 (2007)

(20.) Gulmine, JV, Akcelrud, L, "Correlations Between Structure and Accelerated Artificial Ageing of XLPE." Eur. Polym. J., 42 553-562 (2006)

(21.) Nguyen, T et al., "Relating Laboratory and Outdoor Exposure of Coatings: IV. Mode and Mechanism for Hydrolytic Degradation of Acrylic-Melamine Coatings Exposed to Water Vapor in the Absence of UV Light." J. Coat. Technol., 75 37-50 (2003)

(22.) Linde, RV et al., "Effect of Physical Aging and Thermal Stress on the Behavior of Polyester/TGIC Powder Coatings." Prog. Org. Coat., 40 215-224 (2000)

(23.) Montserrat, S, Calventus, Y, Hutchinson, JM, "Physical Aging of Thermosetting Powder Coatings." Prog. Org. Coat., 55 35-42 (2006)

(24.) Hasegawa, Y, Murakami, T, Asami, K, "(Meth)acrylic Copolymer Resin and Coating Film Thereof," United States Patent 7,186,788, Issued on March 6, 2007

(25.) Fraga, F, Castro-Diaz, C, Rodriguez-Nunez, E, MartinezAgeitos, JM, "Physical Aging for an Epoxy Network Diglycidyl Ether of Bisphenol A/m-xylylenediamine." Polymer, 44 5779-5784 (2003)

(26.) Cook, WD, Mehrabi, M, Edward, GH, "Ageing and Yielding in Model Epoxy Thermosets." Polymer, 40 1209-1218 (1999)

(27.) Blasi, P et al., "Plasticizing Effect of Water on Poly(lactideco-glycolide)." J. Control. Release, 108 1-9 (2005)

(28.) Texas A&M University, A.S.D., Available online http://www.met.tamu.edu/class/ATM0203/tut/rad/actdicyl.html

(29.) Perera, DY, Eynde, DE, "Internal Stress in Pigmented Thermoplastic Coatings." J. Coat. Technol. 53 40 (1981)

(30.) Perera, DY, Eynde, DE, "Solvent Influence on the Development of Internet Stress in a Thermoplastic Coating." J. Coat. Technol. 55 37 (1983)

(31.) Hotta, A, Clarke, SM, Terentjev, EM, "Stress Relaxation in Transition Networks of Symmetric Triblock Styrene-Iso-prene-Styrene Copolymer." Macromolecules, 35 271-277 (2002)

(32.) Hodge, IM, "Enthalpy Relaxation and Recovery in Amorphous Materials." J. Non-Cryst. Solids, 169 211-266 (1994)

(33.) Calventus, Y, Montserrat, S, Hutchinson, JM, "Enthalpy Relaxation of Non-Stoichiometric Epoxy-Amine Resins." Polymer, 42 7081-7093 (2001)

(34.) Lu, H, Nutt, S, "Enthalpy Relaxation of Layered Silicate-Epoxy Nanocomposites." Macromol. Chem. Phys., 204 1832-1841 (2003)

(35.) Bartolotta, A, Di Marco, G, Lanza, M. "Elastic Behavior of Polymers: Anharmonic and Relaxational Effects." Phys. Rev. B, 48 10137-10142 (1993)

(36.) Hutchinson, JM, "Physical Aging of Polymers." Prog. Polym. Sci., 20 (4) 703-760 (1995)

(37.) Angell, CA, Ngai, KL, McKenna, GB, McMillan, PF, Martin, SW, "Relaxation in Glassforming Amorphous Solids." J. Appl. Phys., 88(6) 3113-3157 (2000)

(38.) Struik, LCE, Physical Aging in Amorphous Polymers and Other Materials. Elsevier Scientific Publications Co., Amsterdam (1978)

(39.) DiMarzio, EA, "On the Second-Order Transition of a Rubber." J. Res. Natl. Bur. Stds, A. Phys. Chem., 68A (6) 611-617 (1964)

(40.) Devanne, T, Bry, A, Audouin, L, Verdu, J, "Radiochemical Ageing of an Amine Cured Epoxy Network. Part I: Change of Physical Properties." Polymer, 46 229-236 (2005)

(41.) Devanne, T, Bry, A, Raguin, N, Sebban, M, Palmas, P, Audouin, L, Verdu, J, "Radiochemical Ageing of an Amine Cured Epoxy Network. Part II: Kinetic Modeling." Polymer, 46 237-241 (2005)

[C] FSCT and OCCA 2008

This paper was awarded First Place in the 2007 Gordon Awards technical paper competition, held as part of the Future Coat! conference, sponsored by the Federation of Societies for Coatings Technology, in Toronto, ON, Cannada, on October 3-5. 2007.

X. Shi, B. M. D. Fernando, S. G. Cross (??)

Department of Coatings and Polymeric Materials. North Dakota State University, Fargo, ND 58105, USA

e-mail: [e-mail]