How to Liquid Coat Carbon Fiber Composite Panels

Q. I am trying to liquid coat a carbon fiber composite panel. The panels look good hot out of the oven, but after cooling I am getting readthrough of the fibers. What’s causing this appearance loss?

A. Automotive manufacturers are using various lightweight materials to meet 2015 Corporate Average Fuel Economy (CAFE) mileage standards. These include aluminum, high-strength steel, magnesium and both glass fiber and carbon fiber plastic composites.

Many of these applications will be used for structural elements within the vehicle and will not be visible or require a Class A paintable surface. However, carbon fiber composites can provide a significant weight savings over both steel and aluminum for exterior body panels where a class A finish is required. Paint suppliers must be prepared to meet whatever challenges these new substrates might provide.

There are many different combinations of fiber length, types of plastic and molding processes available for these composites. Each has its own impacts on stiffness, material cost and process time. Types of plastic include both thermosets such as epoxy or polyurethane or thermoplastics such as polyamide. The types of process include compression molding, autoclave, injection molding and resin transfer molding (RTM), which is similar to injection.

The process type is the main determiner of process cycle time. For mass production of exterior panels, process times of less than five minutes are desired. RTM uses a continuous fiber weave and a thermoset resin transfer or injection and can approach the desired cycle time for some parts.

We evaluated an RTM carbon fiber composite substrate using several primer technologies for appearance after topcoat. The primers studied included both one-component and two-component primers at both high bakes (20 min. at 250°F) and at low bakes (20 min. at 210°F). After baking, a portion of the primer was sanded smooth. The panels were then topcoated with a water basecoat and a two-component polyisocyanate clearcoat and baked for 20 min. at 210°F. After cooling, all systems studied had a readthrough of the weave pattern from the carbon fiber in the substrate.

To be more specific, there were depressions in the topcoat surface between the bundles of carbon fiber that made up the weave. Even the areas that were sanded smooth prior to topcoat had some readthrough. This unusual result lead us to reheat the panels back to the bake temperature and observe their appearance directly out of the oven. Indeed, as you describe, there was no readthrough when the panels were hot.

In order to understand the changes in surface morphology, we measured the topography using an optical profilometer at room temperature and at the bake temperature. We did this for both the substrate alone and for the topcoated system. The topcoated system was smooth at the baking temperature but developed 2-micron deep grooves when it was cooled to room temperature.

When we measured the substrate alone, we found that these areas between the carbon fiber bundles had 3-micron deep grooves at the baking temperature of 100°C. But these became 5-microns deep after cooling to room temperature. The depth of the grooved area was increasing by 2 microns for both the substrate alone and for the topcoated substrate. Thus, the topcoat was flowing out and completely filling the texture in the substrate during the bake. However, this texture in the substrate is changing as the panel cools and the topcoat can no longer flow since it is already cross linked.

We then cross sectioned the substrate to inspect the area where the defect was occurring. In the cross section, we see that the carbon fiber bundles have a very high density of fibers and a low density of resin. Between these bundles, there are domains that are only resin with no carbon fiber. These domains are approximately 0.25 mm² in cross section.

Carbon fiber is known to have an extremely low coefficient of thermal expansion (CTE = 0-3 ppm/°C). On the composite part, there was a small section of resin only at the resin transfer gate area. We cut this section off and measured the coefficient of thermal expansion for the resin as 55 ppm/°C in the range between the baking temperature and room temperature. Therefore, as the panel cools, the resin-rich domains (within the red box) contract 55 ppm/°C for 75°C or 4,125 ppm.

At 250 microns of height, this is over 1 micron of total shrinkage in the resin-rich domain. The carbon fiber does not shrink. This differential thermal expansion (or contraction) then causes the surface of the substrate to change as it cools. At this point in the process, the topcoat is a solid and cannot flow out to level over this deeper substrate texture.

It was also noted that the readthrough of texture persisted even in the sanded areas. This can be explained by the diagram below. In the top right, we see the primer that has been sanded smooth at room temperature.

Here we see that the differential thermal expansion of the resin rich domains has caused the primer to raise upward over these domains. It is in this configuration that the topcoat flows, levels and solidifies. Upon cooling, the substrate and primer return to their original room temperature configuration which draws the clearcoat surface downward above these resin domains.

Donald Campbell is a chemist for automotive coatings at BASF. Visit basf.com.

 

---

Originally published in the August 2017 issue.