Visco-elastic visco-plastic analysis of scratch resistance of organic coatings.

The sensitivity of automotive coatings to scratching, particularly coatings for plastics, has been a growing concern among automakers. Scratching may result from such predelivery events as polishing minor defects embedded in the paint, or postdelivery events such as car wash bristles, dirt embedded under a cloth utilized in polishing the car, tree branches, and the like. Warranty cannot separate out which event is the more prevalent (e.g., predelivery or postdelivery to the customer) form of damage on plastics. Data available on coatings for metal, however, does suggest that isocyanate-based crosslinked systems perform more poorly than their melamine-based crosslinked counterparts when exposed to in plant (predelivery) handling. Coatings on plastics, while lower in modulus than coatings on steel, are still subject to scratch events, albeit they have a greater tendency to "self-heal" once scratching events have occurred. This work attempts to correlate the scratch resistance behavior of three different one-component hydroxyl-functional acrylic or polyester-acrylic-melamine-crosslinked systems. Relationships between the tensile and indentation properties of the coatings were found to be related to the initial warranty data of the coating systems. Most importantly, it was apparent that the critical depth to fracture of the flexible coatings may be the relevant indicator of field performance for coatings on plastics. For metals, the relevant indicator to field performance is often only critical load to fracture, negating the influence of indentation depth. This variance is important to consider since coatings on plastic are often softer and more resilient to surface abrasions than their counterparts on metal.

Keywords: Scratch, hardness, melamine-formaldehyde coatings, plastics, solvent-based, mechanical properties, thermoplastic olefins, automotive-OEM, high-solids, thermoset, modelling

**********

Scratch durability of clear coatings in automotive applications has been ranked as one of the most important criteria for life performance of paint systems. Yet, no characterization methodology for scratch resistance has been accepted by the entire scientific and industrial communities. Scratch and mar resistance has been the subject of numerous research efforts in the past few years, both by research institutions and manufacturing industries. This body of research has established a series of correlations between certain scratch parameters and field performance survey results. Several researchers have attempted to relate the variation of modulus, (1) toughness, (2) and hardness (3) of the coating to a scratch resistance rating. Conventional test methods utilized to quantify the scratch resistance were measurements such as: reduction in gloss of a coating after being subjected to a traversing tangential load of known particle size abrasive (crockmeter test); measurement of damage depth and recovery of damage depth of a coating after exposure to a single point indentor (nanoindentor (4,5)); or quantitative measurement of the fraction of elastic, viscoelastic creep, and fracture response of a coating after exposure to a single point indentor (scanning probe microscope (6)).

Courter (7) evaluated several acrylic/melamine formulations for automotive sheet metal and proposed that a marresistant coating should possess a low modulus, which would translate to low yield stress and high toughness. She attributed the higher mar resistance of such coatings to their ability to resist cracking when scratched. However, unexplained contradictions appeared in other studies in trying to relate the scratch/abrasion test results to coating material attributes analyzed. These contradictions imply that further understanding of the rheology of the materials needs to be established in order to first understand these results and design a better, more general test methodology that would gain consensus from the entire automotive paint community.

Most scratch methodologies used today have been developed for metals, and the analysis of such scratch data only considers the behavior of metals. Polymers have a very different set of properties and they have to be taken into account in the definition of scratch parameters. A number of research studies have recently focused on the scratch behavior of polymers, taking into account the more complex rheology of these materials. In particular, when testing polymers, the effects of probe geometry, scratch speed, and sample temperature can be very important to the actual scratch mechanisms undergone by the surface. Also, the parameters that have been used to characterize scratch resistance may lose their significance for polymers because of the interaction of various mechanisms. First, the quantitative analysis of plastic deformation becomes more difficult as relaxation effects take place, and can reduce or erase any ductile deformation that may have occurred during the scratching process. Second, the transition between ductile behavior and brittle damage has been extensively studied in the literature, as fracture mechanisms usually generate the most visible features on a surface. It has usually been characterized via the load applied on the probe at the point where the first fracture event is occurring. For specific cases, the use of the "critical load" is sufficient and can be related to field performance. Recent results at various scratch speeds, however, indicate that the critical load does not solely characterize the fracture toughness of the material. Jardret et al. reported that for PMMA, depending on the shape of the conical indenter (various values of cone-included angle and tip radius), the critical load could increase or decrease with the scratch speed. The authors showed that visco-plastic behavior was interacting with the toughness of the material to determine the load under which the material would fracture.

[FIGURE 1 OMITTED]

[FIGURE 2 OMITTED]

In this study, a detailed analysis of the effects of polymer rheology on scratch mechanisms is summarized. This analysis was then applied to three paint systems deposited on TPO substrates. The experimental data generated on these samples were defined in order to gather enough information on the rheology of the materials to identify the most significant parameters to describe the resistance to fracture under scratch testing of these kinds of paint systems.

Effect of Polymer Rheology on Scratch Behavior

The particularity of polymeric materials resides in the evolution of their behavior with the velocity, the temperature, the amount of deformation (strain), and the history of this deformation. Also, a penetrating probe can create a complex distribution of stresses and strains around its contact with a surface, depending on the size and shape of the probe and the rheological properties of the material. It is, therefore, important to identify the influence of all these parameters in order to understand the origin of the scratch mechanisms.

Influence of the Shape and Size of the Scratch Probe

The influence of the shape of the scratch probe on the scratch behavior is now well established. One of the most significant contributions by Briscoe (8) on this subject provided a series of scratch maps for polymers (Figure 1).

[FIGURE 3 OMITTED]

These maps show the various abrasion mechanisms generated as a function of the attack angle of the probe in combination with the load applied on this probe. This analysis gave rise to more quantitative studies on the effect of specific probe geometry on specific materials. Jardret et al. (9) established the importance of the radius of curvature on the measurement of the load, at which the first fracture event occured during the scratch test.

In addition, to better understand the effect of the contact size and geometry on the scratch mechanisms, recent finite element simulations have allowed for the computation of the distribution of the stresses and strains around the contact, and highlighted the evolution of these stresses with the size and geometry of the contact (in particular, the included angle of the indenter and tip radius). (10) Figure 2 shows that for two different included angles, the level of strains applied by the indenter during an indentation test is very different in magnitude and distribution.

J.L. Bucaille established that for a given indenter geometry, the strain level applied during scratching is about 1.7 times the strain level applied during a normal indentation experiment.

Bucaille (11) also showed the extent of the tensile stresses at the rear of the contact in scratch testing, while under indentation, most stresses are compressive. It is also interesting to note that the distribution of stresses around the contact is not symmetrical. As shown in Figure 3, the material undergoes compression stresses in front of the indenter, and is then subjected to tensile stresses when in the back of the contact. Finally, the maximum plastic strains are found in the back of the contact where the tensile stresses are applied (Figure 4).

These results indicate that the tensile behavior of the material under large strain seems to have critical importance on the scratch mechanisms. In addition, it is generally accepted that most fracture events occur in the back of the indenter. Therefore, the tensile stresses generated under large plastic strains have a significant role in the initiation and propagation of fracture. The relation between fracture resistance and tensile behavior has already been studied. Jardret et al. (12) showed that the critical load for fracture of PMMA with a conical indenter increased as a function of temperature, and presented a transition around 50[degrees]C. They have related this transition to the transition existing in uniaxial tensile behavior between elastic brittle behavior and elastic-plastic brittle behavior at the same temperature. At that temperature it is shown that the strain at fracture increased significantly while the stress at fracture decreased.

[FIGURE 4 OMITTED]

[FIGURE 5 OMITTED]

In this work, Jardret et al. also showed that the critical load was not a good indication of the fracture resistance under scratch, as it was influenced by other rheological properties. Indeed, using two different conical indenters with spherical tips, they recorded the critical load as a function of scratch speed. The results they obtained indicated that with one indenter the critical load increased with the speed, and with the other, it decreased. The authors proposed a definition for a characteristic strain applied by the indenter, which would be more suitable for determining the fracture resistance of polymeric surfaces. The characteristic strain, inspired by the definition of the strain in indentation for a spherical and conical indenter, returned the same values as a function of strain rate for both indenters (Figure 5).

[FIGURE 6 OMITTED]

Effect of Visco-Elasticity and Visco-Plasticity On Scratch Behavior

One important characteristic of polymeric materials comes from the dependence of their behavior on the rate at which the strains are applied. The effect of the velocity of the scratch probe on the scratch mechanism has been well illustrated by Gauthier et al. (13) (Figure 6). He demonstrated that the visco-elastic characteristics of the material greatly influenced how the material deformed under the probe contact, and how it recovered after the scratch process. This behavior affects the actual shape and size of the contact area. Figure 6 shows that by changing the velocity of the probe, under a constant load, the behavior can evolve from a visco-plastic grooving to a visco-elastic sliding, without any groove being created after the experiment.

[FIGURE 7 OMITTED]

The visco-plastic behavior of polymers under scratch testing has been studied more quantitatively by Loubet et al. (14) Loubet has focused on the evolution of the visco-plastic strength (e.g., hardening and increase in the yield strength) during scratch and indentation as a function of temperature and strain rate, and identified that the activation energy associated with the formation of the groove via a visco-plastic process was related to the secondary relaxation and secondary transition temperature (Figure 7). On the other hand, the relaxation process that is taking place after the scratch process is related to the primary transition temperature and the primary relaxation mechanism.

The literature review indicates the importance of the stress-strain behavior, and the dynamic response of a polymer coating to its scratch resistance. The purpose of the experimental study is to compare the tensile behavior of three clear coatings with their scratch behavior. In particular, attention will be focused on the strain at fracture under tensile experiments and the depth under which the first fracture events occur during a scratch process.

EXPERIMENTAL AND ANALYSIS

All solventborne flexibilized (for application onto flexible plastics) coatings utilized in this work were spray applied:

(1) As a clearcoat, black basecoat, or black basecoat/clearcoat at nominal dry film build (25 to 40 microns or 20 microns for each clear-coat and basecoat layer, respectively) over untreated thermoplastic olefin (TPO) substrate supplied by Solvay Engineered Polymers (Sequel 1440), which was subsequently baked for 30 min at 121[degrees]C (for free film tensile specimens), or

[FIGURE 8 OMITTED]

(2) As a composite consisting of chlorinated polyolefin adhesion promoter (dry film build of 8 microns)/ black one-component melamine crosslinked basecoat (dry film build of 20 microns)/clear-coat (dry film build of 25 to 40 microns) over thermoplastic olefin substrate (Sequel 1440), which was applied wet-on-wet and subsequently baked for 30 min at 121[degrees]C.

The flexible coatings were one-component hydroxyl-functional acrylic or polyester (1K) melamine-crosslinked systems. Coatings were obtained from the commercial paint suppliers utilized in the automotive components business (e.g., BASF Coatings, PPG Coatings, or DuPont Automotive Coatings).

Tensile Behavior of Thin Coatings

In order to describe the mechanical behavior of the coatings using uniaxial tensile experiments, free films were manufactured of the clearcoats only, and of the basecoat/clearcoat systems. The samples were 25.4 mm wide with a thickness of 35 to 50 microns. The thickness of each sample had been measured prior to the experiment. The original samples length was 50.8 mm. Five duplicates were used to ensure the statistical significance of the results.

The uniaxial tensile experiments were performed at three different temperatures: 0[degrees]C, 25[degrees]C, and 60[degrees]C. The speed used for all these tests was 10 mm/min, which corresponded to a strain rate of 0.03 [s.sup.-1]. During the tests, the displacement rate was maintained until the sample broke. The load and the elongation of the sample were recorded at the point of break.

[FIGURE 9 OMITTED]

These uniaxial tensile experiments were used to describe the rigidity of the samples, as well as their fracture toughness. Various parameters were used to describe their resistance to fracture, the stress at break, the strain at break, and the energy to break.

Figure 8 presents the elastic modulus calculated from the uniaxial tensile tests performed on free films for the clear coatings only, and for the clear coatings on the basecoats.

It is interesting to note that the clearcoats were, in general, more rigid than the system clearcoat + basecoat. This indicated that the properties of the basecoats were lower than that of the clearcoats.

The energy to break was calculated from the area under the load-displacement curve until the breaking point. The energy to break indicated that coating A was a very brittle film in comparsion to the others at 0[degrees]C. Coating C's energy to break was very low as well from 25[degrees]C and up (Figure 9).

The stress at break (Figure 10) shows that coating A was fractured under higher stresses than the others, even though the energy at break was smaller. For all three samples, the stress at break decreased as a function of temperature, in the same manner as the modulus.

The strain at break (Figure 11) for coating A was smaller than the others. The figure shows a fairly typical behavior of a rigid but brittle material. One can see that coating C had a large strain at break at room temperature.

The evolution of the stress at break (Figure 10) from 0[degrees]C to 60[degrees]C indicated a decrease in the strength of the sample with increasing temperature, which is expected for most polymer materials. Although what is interesting is that the strain at break increased between 0[degrees]C and 25[degrees]C, but it decreased above 25[degrees]C. This is an indication of the nature of the crosslinking in the polymer structure. The strain at break would continue to increase for amorphous polymers.

[FIGURE 10 OMITTED]

[FIGURE 11 OMITTED]

[FIGURE 12 OMITTED]

[FIGURE 13 OMITTED]

Indentation and Scratch Behavior of the Coatings

The microhardness experiments were performed by Visteon Corporation. Microhardness measurements on the clearcoat only or basecoat/clearcoat composite structures were made with a Fischer Microhardness H-100 apparatus equipped with a Vickers indentor and a 100-mN load. The load rate was applied in 60 steps, with one second between steps. Unloading was accomplished after seven seconds of creep in 60 steps, with one second between steps. All the indentation and scratch experiments were performed at room temperature and were averaged after five replicates. The scratch experiments were performed on a Nano Indenter XP (MTS Systems) at NIST. A conical indenter of 90[degrees]C included angle and a 2-micron radius was used. Figure 12 summarizes the data for the hardness and elastic modulus.

As shown in the uniaxial tensile results given in Figure 12, the basecoats seemed to lower the properties over those of clearcoat only. Coating A showed a higher indentation modulus than tensile modulus. These results correlated with those obtained on free films in the uniaxial tensile experiments, shown in Figure 8. Results obtained in Figure 12, related to the microhardness experiments, seem to confound the tensile results, however. Microhardness testing was performed in compression, and measured only the "top tens of microns" of the coating. Tensile results, however, tested the bulk properties of the coating. Combining the findings of Figures 12a and 12b therefore seem to dictate that bulk coating properties, as measured in tension, can vary significantly from compression surface properties, as measured with indentation. From the microhardness indentation results it appeared that clearcoat C was as hard as clearcoat A, while in tension the properties of clearcoat C fell to roughly half those of clearcoat A. In most cases, the microhardness results for the basecoat/clearcoat systems indicated that basecoat "softens" the overall basecoat/ clearcoat system, with basecoat/clearcoat A being the exception.

Figure 13 shows the critical load for fracture under scratch, and the corresponding penetration depth, for the three samples at room temperature. These results indicated that while coatings A and B display a similar critical load to fracture, coating B allowed a deeper penetration of the probe before fracture occurred than coating A. From the uniaxial results, coating A was expected to present a brittle behavior and low scratch resistance. The depth at fracture correlated well with the strain at break under uniaxial tension (Figure 11). These results also seem to parallel warranty indicators found after three, six, and 12 months in service (MIS) for the bumper/fascia only on vehicles of similar platform size (albeit color in such data is not limited to black) (Figure 14). The categories binned in the "things gone wrong" (TGW) categories are chip and scratch. Since damage resistance in TGW data are labeled as chip or scratch damage, it is difficult to separate out only scratch. These two warranty indicators may confound the correlation.

[FIGURE 14 OMITTED]

DISCUSSION

In this work, the tensile results were compared to the scratch results. As mentioned in the first part of this article, it is very important to compare the results under the same conditions of strain rate and temperature. The strain rate applied during the tensile tests was 0.03 [s.sup.-1]. It is then important to determine what strain rate was applied during the scratch experiments at the point of fracture. The strain rate during a scratch process was defined by Hochstetter et al. as the ratio [dot.[epsilon]] = V/a, where V is the scratch speed and a is the scratch width. (15) The strain rates applied during the scratch experiments at the point of fracture were between 0.5 [s.sup.-1] and 0.7 [s.sup.-1], as all materials did not fracture at the same depth. A factor of 20 exists between the strain rate in scratch and in tensile experiments. Following the analysis from Hochstetter et al. presented previously, a test performed at 0[degrees]C at 0.03 [s.sup.-1] is equivalent to a test performed at 20[degrees]C at 0.6 [s.sup.-1]. Using this superposition principle, the scratch tests at 25[degrees]C should be compared to the tensile experiments performed at 0[degrees]C rather than 25[degrees]C.

Comparing these data, Figure 15 shows the correlation between the strain at fracture under tension and the penetration depth for fracture under scratch condition.

[FIGURE 15 OMITTED]

CONCLUSIONS

This work was an attempt to correlate the scratch resistance behavior of three different one-component hydroxyl-functional acrylic or polyester-acrylic-melamine crosslinked systems to tensile and indentation properties of the systems. Initial results seemed to indicate a relationship between initial field warranty data and the tensile and scratch properties of the coatings.

It was apparent that the critical depth to fracture of the flexible coatings may be a relevant indicator of performance for coatings on plastics. Higher elastic compliance and higher energy and percent strain at break afforded better scratch resistance. These values were found to fluctuate as a result of temperature variations, and although only one strain rate was considered, frequency effects should not be overlooked.

Correlation of material attributes to a coating's scratch resistance may be predicated on the substrate being painted. Coatings on plastics are more flexible than those on steel. The scratch resistance for flexible coatings may be correlated to the critical depth to fracture; in essence, how the stress is translated through the coating prior to fracture. Coatings on steel, which are often harder than those on plastic, may correlate more with the critical load to failure, negating the influence of penetration depth, since fracture is often limited to the top few microns of the coating. This variance is important to consider since the coatings on plastics, while being softer, are often more resilient to surface abrasions than their counterparts on metal.

An indepth study of the clearcoat subsurface effects as a result of basecoat is an area ripe for future research. This body of data seems to indicate that bulk tensile properties can vary from surface properties of the coatings as a function of the coating system studied. Little work has been done as a function of basecoat color, application conditions, or other variables that can potentially vary surface properties. The study of coating "robustness" as a function of depth and stress dissipation in the coating of plastics is now underway in our laboratories.

ACKNOWLEDGMENTS

The authors want to express a special appreciation to the staff of the National Institute of Standards and Technology for their help in performing the scratch experiments.

References

(1) Shinto, K., Beseche, S., and Sato, S., Proc. PRA Fourth Asia-Pacific Conference: "Advances in Coatings, Inks, and Adhesives," Hong Kong, China, Paint Research Association International Center for Coatings Technology, paper 4, 1994.

(2) Chen, M.J., Osterholtz, F.D., Phol, E.R., and Ramdatt, P.E., Proc. 23rd Intl. Waterborne, High-Solids, and Powder Coatings Symposium, New Orleans, LA, 23, 222, 1996.

(3) Tonge, J.S., Blizzard, J.D., Schmidt, S.M., and Washer, T.R., Proc. 22nd Intl. Waterborne, High-Solids, and Powder Coatings Symposium, New Orleans, LA, 22, 432, 1995.

(4) Lin, L. and Blackman, G., Proc. Adv. Coatings Technology Conference, International Body Engineering Conference, SAE/ESD, Detroit, MI, 1998.

(5) Jardret, V., Lucas, B.N., and Oliver, W., "Scratch Durability of Automotive Clear Coatings: A Quantitataive, Reliable and Robust Methodology," JOURNAL OF COATINGS TECHNOLOGY, 72, No. 907, 79 (2000).

(6) Shen, W., Ji, C., Jones, F.N., Everson, M.P., and Ryntz, R.A., Polym. Mater. Sci, Eng., 74, 346 (1996).

(7) Courter, J.L. and Kamenetzky, E.A., "Micro- and Nano-Indentation and Scratching for Evaluating the Mar Resistance of Automotive Clearcoats," presented at the 5th Nurnburg Congress on Creating Advances in Coating Technology, Nurnburg, Germany, April 12, 1999.

(8) Briscoe, B.J., Evans, P.D., Pelillo, E., and Sinha, S.K., "Scratching Maps for Polymers," Wear, 47, 200 and 137 (1996).

(9) Jardret, V. and Oliver, W., MRS Symp. Proc., Vol. 594, 2000.

(10) Bucaille, J.L., Felder, E., and Hochstetter, G., "Identification of the Viscoplastic Behavior of a Polycarbonate Based on Experiments and Numerical Modeling of the Nanoindentation Test," J. Mater. Sci., 37, 3999-4011 (2002).

(11) Bucaille, J.L., Ph.D. Thesis, "Simulation Numerique de l'indentation et de la rayure des verres organiques," These de Doctorat en Sciences et Genie des materiaux, Ecole Nationale Superieure des Mines de Paris, France 2001.

(12) Jardret, V. and Morel, P., Proc. 2002 Athens Conference on Coatings Science and Technology, Vouliagmeny (Athens), Greece, July 28, 2002.

(13) Gauthier, C., Lafaye, S., and Schirrer, R., Tribology International, 34, 469-479 (2001).

(14) Bertrand Lambotte, P., Loubet, J.L., Verpy, C., and Pavan, S., Thin Solid Films, 398-399: 306-312 (2001).

(15) Hochstetter, G., Jimenez, A., and Loubet, J.L., J. Macromol. Sci.-Phys., B38 (5-6): 681-692 (1999).

Vincent Jardret -- Tribometrix, Inc.*

Rose Ryntz -- Visteon Corporation ([dagger])

* 4132 W. Beaver Creek Dr., Powell, TN 37849.

([dagger]) 401 Southfield Rd., Dearborn, MI 48120.