Our exploration of Wenzel’s theory in Part 1 of this article helps us to understand that surface roughness and adhesion are not always related to each other in a unique fashion. In other words, increasing the surface roughness depth of a blasted surface might not always improve coating adhesion.

With a high-viscosity or low-flow liquid coating, or a powder fusion bonded epoxy (FBE) pipeline coating applied at too-low-pipe-per-heating temperatures, it is very possible that no distinctive relationship can be found between the roughness depth of the substrate and the coating adhesion strength value. This is because the liquid droplets might have difficulty in penetrating and displacing air entrapped in the valleys of a surface profile.

Why the Industry Needs Better Methods

The intensity of the adhesion of a coating to an abrasive-blasted substrate does not depend on any length parameter of surface roughness, regardless of whether the roughness parameter is a profile linear height of Ra, Ry and Rz, or peak count (peaks per linear length). Instead, good adhesion depends not only on the energy that is used to create the contact, but also on the interaction existing in the contact interface.

As we stated in Part 1, the critical question on ensuring a good coating adhesion is not whether the surface profile height of the blasted profile is a minimum 1.5 or 2.5 mils or between 2.0–4.0 mils, but whether the level of the blasted surface preparation has created a rough surface that has a higher substrate surface energy and a lower contact angle in order to provide the adequate wettability for the coating.

Thus, using common surface profile methods, the roughness of a three-dimensional surface cannot be accurately characterized by simply measuring a single roughness parameter in linear length.

Parameters that characterize surface profiles Ra, Ry and Rz, or peak count, are two-dimensional parameters. Although these measures are widely utilized in different applications, they are not really able to provide the full information on the three-dimensional surfaces.

Most importantly, these linear length parameters do not form a linear relationship by themselves to surface area or to surface energy. In other words, the coating industry needs a different and better way of measuring surface profile from the current profile height or peak count.

A new parameter (for example, the roughness ratio, r, could be a potential candidate) or a set of combined parameters (such as peak count/density and developed surface area Sdr), shall be directly related to surface energy or contact angles as a result of the abrasive surface preparation.

It is important to understand that mechanical adhesion in tension differs significantly from mechanical adhesion in shear. This understanding matters in making the correct choice of coating adhesion strength test methods and determining specification requirements for given test parameters.

The differential volume changes between the substrate and coating, resulting from temperature gradients or shrinkage, cause both shear and tensile stresses at the interface.

In structural design, tensile stresses perpendicular to the interface are rare. By contrast, interface shear stresses occur frequently in the coated system.

Current standards and specifications for a corrosion protective coating define bond strength commonly in relation to tensile (pull-off) strength alone. In consideration of the mechanic adhesion information above, this appears problematic. When specifying and/or evaluating coating adhesion strength methods and values for a coating project, it is important to consider the dominant interface stress condition experienced by the actual coated structure.

Coating and Substrate-Specific Requirements

Wenzel’s theory helps us to understand that surface roughness and adhesion are not always related to each other in a unique fashion. This means that increasing the surface roughness of a blasted surface might not always improve the adhesion.

For certain high-viscosity and powder-type coatings, it is possible that no distinctive relationship can be found between the roughness of the substrate and the coating adhesion strength value. This is because the liquid droplets might have difficulty in penetrating the bottom of the surface profile (valleys) and displacing entrapped air. This has been proven in research done in 2005 by Hugh Roper et al1, who examined the effect of the peak count of surface roughness on coating performance of six coating systems. They concluded that peak count had less of an effect on the adhesion performance of two high-viscosity and less-wetting phenolic coatings.

Therefore, there will be an optimized surface roughness range for a specific coating system and specific substrate in order to achieve the proper adhesion, depending on:

  • Viscosity
  • Particle size
  • Coating liquidity (flow)
  • Surface temperature
  • Coating chemistry
  • Substrate chemistry

The industry’s current surface profile height requirements may not fit every project and every type of coating and substrate.

A Step in the Right Direction

The recent development by DeFelsko of a 3-D replica tape measurement on surface profile is heading in the right direction.2 However, further research is needed to quantify the relationships among the 3-D replica tape measurement parameters of the surface profile, degree of surface cleaning, the specific coating, exposure environment, mode of failure and expected lifetime of a coating system.

In summary, it is generally accepted that the nature of an abrasive blast-cleaned steel surface is predictive of long-term coating performance. The corrosion and coating industry has not yet fully understood the dynamics of this complex problem and the true relationship between the surface profile of the blasted surface and the long-term coating performance. Commonly held industry beliefs would suggest that increasing several of these surface profile parameters will improve long-term coating performance. But empirical data suggests it is not that simple.

More Testing Needed

Simply increasing profile parameters does not address all of the theories and questions related to the effect of surface profile on coating performance. The author’s vast experience in this field has led to other relationships or theories that still need to be verified by testing.

Additionally, other evidence indicates that these theories are valid for many common conditions, but not necessarily all conditions. It is the author’s experienced opinion that until a better way to measure and define a properly blasted steel surface profile for the expected coating performance is developed by the industry, a good coating standard/specification shall at least require all of the following surface profile parameters:

  • Minimum profile height
  • Minimum peak count/density
  • Minimum ISO comparator rating

A single and linear surface profile parameter is not capable of defining a proper surface preparation of coating application for long-term corrosion protection.

***

More in the "Questioning Current Methods in Defining Proper Surface Profile" series:

Questioning Current Methods in Defining Proper Surface Profile (Part 1)

Questioning Current Methods in Defining Proper Surface Profile (Part 2)

***
References:

  1. Hugh J. Roper, Raymond E.F. Weaver, and Joseph H. Brandon, The Effect of Peak Count of Surface Roughness on Coating Performance, JPCL, June 2005, pp. 52-64
  2. David Beamish, Replica Tape Unlocking Hidden Information, JPCL/paintsquare.com, July 2015, p1-6