Abstract
Adhesion of coatings is critical to residential, commercial, and defense applications. In particular, coatings used for corrosion mitigation in industrial and transportation infrastructure are needed over a range of scales and material systems, with surface preparation being crucial to the final adhesion performance. At this time, highly labor-intensive methods such as grit blasting (GB), needle gunning, and chemical scraping have been the standard approach for cleaning large infrastructure, all of which present additional issues associated with occupational and environmental hazards. Recently, the Virginia Department of Transportation (VDOT) has investigated several coating removal technologies and their effects on recoating performance. The first technology uses a high-power pulsed laser system and will be referred to as laser ablation coating removal (or LACR). The second technology, induction coating removal (or ICR), utilizes a high-frequency AC current to heat up the surface of the underlying metal substrate to thermally debond coatings. ICR and LACR may also be used in tandem to exploit the productivity of the former and higher efficacy of the latter.
In this research, focus was placed on comparing the adhesion of reapplied coatings on LACR or ICR cleaned surfaces with the main incumbent method, grit blasting (GB). In this dissertation, the research is divided into two main parts: 1) effects of different prior surface treatments and their resulting profiles on the adhesion of coatings on exemplar bridge beam sections and 2) exploration of a fracture mechanics-based adhesion testing method using laboratory-prepared samples. In part one, it was observed that the LACR process removed coating layers down to a pre-existing mill scale layer (ranging in thickness from 10 to 100 μm), which remained on the steel surface of legacy bridge sections, in contrast with GB, which fully removes the oxide layer. In comparison, the ICR-processed beams remained covered with remnants of an original lead primer beneath the paint, highlighting the lower degree of cleanliness achievable by this relatively high-productivity approach. Stylus profilometry and laser confocal microscopy were both used to measure the surface topography and an average roughness, Ra, of 4 μm was determined over multiple measurements of ICR + LACR samples. On the other hand, samples processed by GB had an average roughness of 12 μm, over three times greater than the roughness of the LACR-cleaned surfaces.
After coating removal, two standard corrosion coating systems were applied and compared with one another. An organic zinc-rich (OZ) primer, which consisted of an epoxy matrix loaded with zinc metal powder for corrosion protection, and an inorganic zinc-rich (IOZ) primer, with a sol-gel silica-based matrix. Tests performed with a pneumatic adhesion tensile testing instrument, or PATTI tester, showed equivalent adhesion between LACR, ICR + LACR, and traditionally prepared GB steel surfaces for both the OZ and IOZ coated surfaces, despite the 3-fold decrease in surface roughness on the LACR samples. These samples all averaged over 1700 psi adhesion strength for the OZ coating and over 900 psi with the IOZ coating. Of the two coating systems studied in this portion of the project, it was found that OZ primers (epoxy resin based) failed cohesively by one of two modes denoted C1 and C2. The C1 failures leave bulk zinc primer on the substrate surface post-testing, whereas C2 failures result in a thin, ~1 μm epoxy adhesive film layer that continuously covers the plate surface with no remaining zinc. C1 failures were associated with rough regions in the mill scale layer of the LACR surfaces. Detailed image analysis of the PATTI test sites on the substrates showed ~30% C2 type failures, matching the area over which smooth mill scale regions exist. The same two cohesive failure modes were observed for the OZ coated grit blasted samples as well. On the other hand, IOZ primers exhibit some level of adhesive failures on both GB and LACR treated steel samples. Complimentary SEM and EDS analysis of the PATTI stub fracture surfaces helped to determine whether adhesive or cohesive failure was occurring.
In part two, thin A36 steel sheet samples were prepared with a range of surface conditions. The sheet samples were recoated with OZ zinc-rich epoxy-based primer and were again tested with both PATTI tests as well as a tensile testing approach based the notched coating adhesion (NCA) test. This test consists of creating a through thickness scribe on the coated sample, which is then pulled in tension until the coating delaminates from the surface. The delamination is tracked in real time using digital image correlation (DIC). The surface roughness Ra of each of the NCA samples was measured to be 1.3 μm, 1.1 μm, 4.7 μm, and 4.8 μm for the degreased, laser surface modified (LSM), GB, and LACR samples respectively. Testing results showed that the adhesion from LACR processed samples was equivalent or superior to that of GB surfaces.
Surprisingly, the PATTI adhesion results were greatest for the smoothest sample, the LSM, with an average pull off strength of 329 ± 66 psi, followed by the LACR samples with an average adhesion strength of 318 ± 54 psi, and then the GB and degreased samples with adhesion values of 274 ± 28 and 199 ± 51 psi respectively. The LACR and LSM were shown to fail in the same C1 and C2 failure modes as the steel plate samples, and the GB and degreased samples showed a mix of adhesive failure and cohesive failure revealed using spatial X-ray photoelectron spectroscopy (XPS) maps.
Regardless of the adhesion testing method or adhesion failure mode observed, LACR samples were determined to be at least as good, if not better than grit blasted surfaces with regards to adhesion, even when the substrate profile is much less after LACR. Furthermore, ICR is found to effectively compliment LACR by removing thick layers of bulk coatings leaving only small amounts of primer on the surface which can be removed by LACR to ensure a clean and well-prepared surface for recoating. When combined together (ICR + LACR in tandem), the coating removal rate is significantly increased compared to LACR processing alone. Overall, LACR is shown to be a viable alternative to grit blasting for niche applications such as bridge beam end coating repairs and other jobs that require small areas of coating work, for both legacy mill-scale containing surfaces as well as previously grit-blasted (mill scale free) surfaces.
