Protective coatings are often erroneously thought of as perfect environmental barriers for metal substrates; however, a host of corrosion inducing environmental contaminants permeate through defect-free coatings. Carbon nanotubes are high aspect ratio nanofillers with unique mechanical, electrical, and polymer interaction properties with well-established yet, for practical reasons, often unrealized potential. The research objective was to quantify and understand the influential effects and relationships between low concentration levels of multiwall carbon nanotubes (MWCNT) dispersed into epoxy-amine matrix materials and the different water hydrogen bonding interactions on corrosion rates of steel substrates. We hypothesize that when water directly hydrogen bonds with polymer, substrate and/or MWCNTS, the localized water's capacity to transfer environmental contaminants through the coating, i.e., to and from the substrate, diminishes due to a reduced potential to contribute to the formation of water hydration shells and therefore aid in diminishing the corrosion rate. We measured the absolute pre-exposure water content, and monitored to delineate between the ratio and shifting ratio of in situ free versus bound water hydrogen bonding interactions at the coating/air interface using ATR-FTIR spectroscopy in a 5% NaCl fog environment in an attempt to correlate these differences with experimental corrosion rates. Free water content was reduced from ∼20% to <1% of the total water concentration when 1.0 wt % MWCNTs was dispersed into the parent polymer network. Concurrently, the bound water content was measured to shift from ∼2% to >80% with the same MWCNT concentration. The MWCNT bound water resulted in 25% less corrosion for the same steel substrates albeit the measured water vapor diffusivity was the same for each material combination evaluated. Interestingly, the measured pre-exposure bound water content was predictive of which material would corrode slowest and fastest, i.e., the ratio of starting water states seems to be mechanistically related to the corrosion process and the values have potential to predict corrosion rates for a variety of samples evaluated.
Read full abstract