Article pubs.acs.org/Langmuir
Synergy between Galvanic Protection and Self-Healing Paints Lance Michael Baird, Marcia W. Patchan, Melanie Morris, Adam J. Maisano, Terry E. Phillips, Jason J. Benkoski,* and Rengaswamy Srinivasan* The Johns Hopkins University, Applied Physics Laboratory, Laurel, Maryland 20723-6099, United States S Supporting Information *
ABSTRACT: Painting is a cost-effective technique to delay the onset of corrosion in metals. However, the protection is only temporary, as corrosion begins once the coating becomes scratched. Thus, an increasingly common practice is to add microencapsulated chemical agents to paint in order to confer self-healing capabilities. The additive’s ability to protect the exposed surface from corrosion depends upon (i) how long the chemical agent takes to spread across the exposed metal; (ii) how long the agent takes to form an effective barrier layer; and (iii) what happens to the metal surface before the first two steps are complete. To understand this process, we first synthesized 23 ± 10 μm polyurea microcapsules filled with octadecyltrimethoxysilane (OTS), a liquid self-healing agent, and added them to a primer rich in zinc, a cathodic protection agent. In response to coating damage, the microcapsules release OTS into the scratch and initiate the self-healing process. By combining electrochemical impedance spectroscopy, chronoamperometry, and linear polarization techniques, we monitored the progress of self-healing. The results demonstrate how on-demand chemical passivation works synergistically with the cathodic protection: zinc preserves the surface long enough for self-healing by OTS to reach completion, and OTS prolongs the lifetime of cathodic protection.
1. INTRODUCTION Coating damage remains unavoidable in paints that are used as protective coatings on metal surfaces to form a barrier layer against water and salt to prevent corrosion. When the coating is damaged, corrosion proceeds unabated. A new class of coatings with the ability to repair themselves promises protection even when the coating is damaged.1−8 They employ a wide range of strategies toward this goal, including external stimuli such as photoinduced healing,9 reversible hydrogen bonding,10 nanoparticle-mediated healing,11 and thermally activated remodeling of polymer cross-links.12 Another strategy relies upon a small reserve of liquid corrosion inhibitors entrained within microcapsules that are dispersed evenly throughout the coating.13 Microcapsules (MCs) containing materials such as linseed oil, hexamethylene diisocyanate, 2-mercaptobenzothiazole, and isophorone diisocyante have previous been investigated for their use as corrosion inhibition additives.14−21 When damaged, the microcapsules release healing agent into the scratch and coat the exposed surface with a thin film. In order for the selfhealed barrier layer to be effective, the released chemical must form a protective layer on the exposed nascent metal surface and not on the corrosion byproducts created at the site of the damage. In a race to occupy the metal surface, large corrosioninhibiting molecules must compete with rapidly reacting © XXXX American Chemical Society
corrosive molecules such as oxygen, water, and chloride ions. Compared to the rate of adsorption of organic monomers or surface coupling agents on iron, aluminum, or titanium,22 the rate of the oxygen reduction reaction (ORR) on those metal surfaces is almost instantaneous.23 Therefore, releasing corrosion inhibitor in sufficient quantity to coat the entire metal surface is essential but not sufficient to achieve selfhealing. Corrosion inhibitors are less effective once corrosion has begun, since corrosion products can block the inhibitor from adsorbing on the metal. The kinetics cannot be tilted to favor self-healing by infusing larger concentrations of MCs into the paint. Doing so has unintended consequences for the coating integrity. Mechanically, each MC acts as a spherical void, effectively creating a stress concentration.24 Peterson’s analytical solutions shows that the stress near a spherical void is more than double that of the applied uniaxial tension, effectively reducing the damage threshold. Each MC is therefore a potential crack initiation site. More importantly, the bulk modulus (k) of the coating decreases with the volume fraction of voids (ϕ) as follows Received: June 12, 2015 Revised: August 28, 2015
A
DOI: 10.1021/acs.langmuir.5b02115 Langmuir XXXX, XXX, XXX−XXX
Article
Langmuir Scheme 1Schematics of the Four Sample Typesa
−1 ⎛ 3(1 − v0) ⎞ ϕ⎟ k = k 0 ⎜1 + 2(1 − 2v0) ⎠ ⎝
where k0 is the bulk modulus of the bulk material and ν0 is Poisson’s ratio. Thus, larger quantities of MC additives in the paint would compromise the barrier properties, appearance, and abrasion resistance of the paint. The key to preserving the desirable properties of the paint is to achieve effective selfhealing with the minimum loading of MCs. Despite these tradeoffs, MCs with encapsulated corrosion inhibitors have been in use for more than a decade.25−32 2. Defining Self-Healing. Four key features for designing a practical microencapsulation-based self-healing coating are (i) the MCs should not degrade the mechanical integrity of the coating by means of excessive size, high degree of loading, or percolation; (ii) when the coating is damaged, the self-healing action should be autonomous, i.e., a protective barrier layer should form without an externally supplied stimulus such as light, heat, or catalyst; (iii) the self-healing action should be completed prior to the onset of corrosion; and (iv) following the self-healing action, the healed location should remain corrosion-free for an extended period of time. Of these four features, (i) and (ii) are relatively easy to achieve; (iii) and (iv) are much harder. The latter two challenges must be fully resolved in order to enable widespread adoption of self-healing paints. The most crucial interlude that governs the effectiveness of the self-healing action of a coating is related to feature (iii): the time it takes to self-heal while the metal surface remains pristine and unaffected by corrosion. The effectiveness depends upon three simple steps: (a) how long it takes for the agent to wet the exposed metal; (b) how long it takes the agent to form an effective barrier layer; and (c) what happens to the surface before the first two steps are complete. With these three steps in mind, we have arrived at a new paradigm: combine the selfhealing activity of MCs with the galvanic protection of a sacrificial pigment. 2.1. Zinc-OTS Synergy. Galvanic protection works synergistically with the self-healing action by protecting exposed metal from corrosion, while the self-healing chemical from the MC wets the metal surface and undergoes necessary chemical reactions to form a durable and effective barrier layer. The sacrificial pigment in our paint is a zinc microparticle. Our self-healing chemical is a moisture-curable octadecyltrimethoxysilane (OTS). On a steel substrate, the quantity of zinc powder that we use is the same as that in zinc-rich paint. The duration of cathodic protection offered by that zinc is much longer in the presence of OTS that tends to diminish the rate of zinc oxidation without affecting the degree of cathodic protection. As a result, cathodic protection lasts for several hours more than it takes for OTS to form barrier layer on steel. These key features are easily traced in real time using three electrochemical techniques, namely, chronoamperometry (CA), linear polarization (LP), and impedance spectroscopy (EIS).
a
(A) Standard CARC, (B) Zn-rich CARC, and (C) standard CARC with MC, and (D) Zn-rich CARC with MC.
layers: a 100-μm-thick one-part zinc-rich primer with a moisturecurable polyurethane matrix was applied to the steel panel, followed by a 37-μm-thick seal coat of MIL-DTL-53022 epoxy-polyamide primer and a 50-μm-thick MIL-DTL-53030 CARC topcoat. Paint layers in the panels in Scheme 1A,B did not contain the microcapsules. Two more sets of panels, shown under Scheme 1C,D, had microcapsules in one of the paint layers, but only one of them had zinc. Scheme 1C was standard CARC with MC, a variant of standard CARC, with 1% (w/w) microcapsules added to the MIL-DTL-53022 epoxy-polyimide primer prior to spraying and the topcoat identical to the one in Scheme 1A. Scheme 1D is a variant of the Zn-rich CARC, with 1% (w/w) microcapsules added to the zinc-rich primer prior to spraying. The microcapsules themselves were 23 ± 10 μm in diameter with polyurea shells formed from the interfacial polymerization of isophorone diisocyanate, diethylenetriamine, and polyethylenimine. The self-healing agent, octadecyltrimethoxysilane (OTS), comprised 63 ± 2% of the microcapsules by mass. Further details of the MC preparation and characterization are described elsewhere.33 All samples were spray painted with a high-volume, low-pressure (HVLP) spray gun in accordance with the CARC paint specification. Additional details of the coating layup and drying times are provided in the Supporting Information. 3.2. Silanized Samples. We also prepared a separate set of test panels to isolate the effects of OTS on the surface properties of steel and zinc in the absence of the complex CARC coating system. An ASTM A1008 steel panel and a galvanized steel panel were treated directly in a 1% (v/v) solution of OTS in toluene for 72 h. The long incubation time was chosen to achieve thorough coverage of OTS. The silanized panels were then compared against untreated controls to verify that OTS could form passivation layers on each metal. Procedures for the silanized metal coupons preparation are also provided as Supporting Information. 3.3. Electrochemical Measurement of Corrosion Resistance. All samples were tested at room temperature (23 °C) using the ASTM G-42 electrolyte, which is an aqueous solution of 1% (w/w) each of sodium chloride, sodium sulfate, and sodium carbonate. Before the coating was damaged and between the moment the coating was scratched, exposing the steel substrate to ASTM G-42 electrolyte, and the time when the self-healing action is complete, we monitored the galvanic protection by zinc, the kinetics of OTS self-assembly, and the corrosion protection by OTS on the steel substrate using electrochemical impedance spectroscopy (EIS). We measured the influence of OTS on the kinetics of zinc oxidation (associated with galvanic protection) and oxygen reduction reaction (ORR), which is the main conjugate corrosion reaction for steel using chronoamperometry (CA) and linear polarization (LP). Two types of electrodes were used in this study. One type was for painted steel panels. It was used to evaluate the self-healing action of OTS-containing MCs in the presence and absence of zinc in the paint. In this type, an 11.6 cm2 area of the flat surface of the painted panel was used in the testing. The edges of the painted coupon were treated carefully to avoid contact with the test solution. The paint on the tested surface of the coupon was free of any damage, except when deliberately scratched to test the self-healing effect of the MCs. The impedance of an undamaged painted surface can be large, >109 Ω·cm2; to measure it accurately, it is necessary to have a large area of the painted surface in contact with the electrolyte and eliminate all
3. EXPERIMENTAL SETUP 3.1. Paint Sample Scheme. As shown in Scheme 1, ASTM A1008 steel panels were coated with either two or three layers as described in the Chemical Agent Resistant Coatings (CARC) specification. Standard CARC consisted of a 37-μm-thick seal coat of MIL-DTL53022 epoxy-polyamide primer and a 50-μm-thick MIL-DTL-53030 CARC topcoat (Scheme 1A), and it was used as a control sample. In Scheme 1B, the second control, Zn-Rich CARC, was applied in three B
DOI: 10.1021/acs.langmuir.5b02115 Langmuir XXXX, XXX, XXX−XXX
Article
Langmuir possibilities of the electrolyte contacting the edges of the coupon. By clamping a cylindrical tube with O-ring on the top of the painted surface, we were able to achieve both objectives. The cylindrical electrochemical cell also contained a high-surface-area carbon counter electrode and a Ag/AgCl reference electrode. Over a 96 h contact period with the electrolyte at room temperature, the impedance of the coating dropped by no more than a factor of 2; the drop was presumably due to the penetration of the electrolyte into the paint. The surface was then scratched over a 12 mm length with a 1.6 mm scriber, exposing a 0.2 cm2 area of steel. The time-dependent evolution of the self-healing action was followed over the next 200 h through changes in the impedance of the coating. The derivation of the corrosion resistance from the impedance data is described in more detail in section S1.7 in the Supporting Information. 3.4. Electrochemical Measurements of Zinc and Steel Oxidation. For the steel and zinc samples treated directly with OTS in the absence of CARC paint, we performed chronoamperometry (CA) and linear polarization (LP) measurements to investigate the influence of OTS on the kinetics of zinc oxidation (associated with galvanic protection) and the oxygen reduction reaction, which is the main conjugate corrosion reaction for steel. For this sample type, most of the surface was coated and masked with epoxy paint, except for a 6.5 cm2 area of the flat surface. Each coupon was placed in a separate electrochemical cell containing the electrolyte that was constantly saturated with air using a bubbler. The cell also contained a platinum counter electrode and an Ag/AgCl reference electrode. On each coupon, all measurements were completed within 1 h of immersion. Additional details are described in section S1.8 in the Supporting Information.
4. RESULTS AND DISCUSSIONS 4.1. Visual Observation. Gross visual inspection at 5× magnification revealed that corrosion occurred along the entire length of the scratch only in the two control samples, namely, standard CARC and zinc-rich CARC (not shown). In standard CARC with MC and no zinc, corrosion was present but less prevalent than for the two controls. In zinc-rich CARC with MC, corrosion was virtually absent along the entire length of the scratch except in three spots shown in Figure 1. When it did appear, the rust was confined to the scratch. No corrosion was observed in the regions of undamaged paint. Microscopic visual inspection at 20× magnification in Figure 1 shows only a small segment of the scratch in which the corrosion products can be easily identified as dots of different shapes and sizes. It is evident from Figure 1 that steel at the scratched locations in all four coating schemes corroded after 24 h of exposure to the ASTM-G42 electrolyte. However, the degree of corrosion varied among the different types of coatings. The growth and emergence of new corrosion spots occurred at the highest rate in the standard CARC and zinc-rich CARC controls. Corrosion appeared to proceed at a slower rate in standard CARC with MC and no zinc. After the first 24 h, no new corrosion spots occurred and old ones grew little in the zinc-rich CARC with MC sample. These observations are consistent with the impedance data described next. 4.2. Electrochemical Impedance of Self-Healing Paint. The four sample types described in Scheme 1A−D are designed to isolate the effects of self-healing and galvanic protection on the prevention of corrosion within a scratch. Scratch- and pinhole-free panels from each scheme were first exposed for 96 h to the ASTM G-42 electrolyte. Then the coating from each scheme was damaged to create a 12-mm-long, 1.6-mm-wide scratch exposing the steel substrate. Before, during, and after the scratch, multifrequency impedance responses of the coatings were continually monitored. The corrosion resistance
Figure 1. Optical micrographs of selected segments of the scratch locations for coatings from all four schemes (Scheme 1A−D) obtained at two different times of exposure to the AGTM-G42 electrolyte. The spotted locations seen in the micrographs are presumably corrosion products. Each white bar in the micrographs represents a 0.4 mm length.
(Rcorr) and the corrosion potential (Ecorr) data derived from the impedance are shown in Table 1 and Figures 2 and 3. Ecorr values are shown only for the postscratch period to highlight the role of zinc in protecting the freshly exposed steel from corrosion. Figure S2 in the Supporting Information provides the fitting protocol used in deriving Rcorr from multifrequency EIS data to an electrochemically equivalent circuit. Optical micrographs of a small segment of the scratch in each panel are shown in Figures 2 and 3. It is evident from the data in Figure 2 that for all different types of coatings, Rcorr remained high at around 109 Ω·cm2 prior to scratching, suggesting equally good protection over the 50 to 96 h exposure to the ASTM G-42 electrolyte. The more important observation for this discussion is the temporal changes in Rcorr after the scratch. Scheme 1A with the standard CARC coating (no zinc or MC) and Scheme 1B with zinc-rich CARC (no MC) both registered the smallest Rcorr after the scratch, at around 105 Ω·cm2. These panels experienced the highest rate of corrosion, and their Rcorr changed very little with time. The corresponding Ecorr values carried an important clue as to the extent of cathodic protection that was available for C
DOI: 10.1021/acs.langmuir.5b02115 Langmuir XXXX, XXX, XXX−XXX
Article
Langmuir
Table 1. Electrochemical Impedance Spectroscopy Tracks the Synergistic Effect of Sacrificial Zinc Anode and OTS-Filled Microcapsules (MC) in Self-Healing Paint additives coating type Scheme Scheme Scheme Scheme
1A: standard CARC 1B: zinc-rich CARC 1C: standard CARC with MC 1D: zinc-rich CARC with MC
zinc no yes no yes
steady-state values hours after scratch
MC
Rcorr (Ω·cm2)
no no yes yes
0.99 × 10 (Figure 2A) 0.5 × 104 (Figure 2A) 0.1 × 108 to 0.7 × 106 (Figure 2A) 2 × 108 to 4 × 108 (Figure 3A) 4
Ecorr (V) −0.63 −0.53 −0.58 −0.25
self-healing action
(Figure 2B) (Figure 2B) (Figure 2B) ± 0.05 (Figure 3B)
no no yes (limited) yes
Figure 2. Time-dependent changes in the corrosion resistance (A) and corrosion potential (B) for coated steel panels exposed to ASTM-G42 electrolyte at room temperature. In the graph, the 0 h corresponds to the occurrence of the scratch; negative time corresponds to before the scratch. The three color-coded graphs are standard CARC with no zinc or MC (black), zinc-rich CARC with zinc and no MC (red), and standard CARC with MC and no zinc (blue). (For data on standard CARC with MC and zinc, see Figure 3.) The micrographs show images of a small section of a panel from Scheme 1C (standard CARC with MC), its surface freshly scratched (left) and after 210 h of exposure to the electrolyte; each white bar in the micrograph is 1.6 mm long. The OTS from the MCs managed to slow down corrosion along much of the scratch. Corrosion proceeded unabatedly on the scratched areas of the panels from Scheme 1A,B that did not have the MC. See also Figure 1 for scratch locations with corrosion.
steel from zinc. The Rcorr data for the panels in Scheme 1A (containing no zinc or MC) is shown with black curves in Figure 2. Its Ecorr remained between −0.6 and −0.65 V, which is a normal range for iron in aqueous salt solution of neutral pH. Scheme 1B with zinc-rich CARC (red curve) registered an Ecorr more negative than −0.85 V less than 20 h after the scratch, followed by a step change toward −0.5 V, indicating that zinc was no longer available to protect steel; a concomitant step change is also seen in Rcorr.
Scheme 1C, standard CARC with MC and no zinc (blue curve), registered an Rcorr of about 106 Ω·cm2 immediately after the scratch. An increase in corrosion resistance by a factor of 10 relative to the controls is arguably due to the presence of OTS. Later, Rcorr rose to 108 Ω·cm2 after 60 h, and after 160 h, it fell back to 106 Ω·cm2. Clearly, the OTS was attempting to self-heal the damaged paint, but the barrier layer did not last indefinitely. The drop in Rcorr from 108 to 106 Ω·cm2 suggests shedding of the self-healed OTS layer at least in some areas within the bare metal surface exposed by the scratch. In this scheme, Ecorr D
DOI: 10.1021/acs.langmuir.5b02115 Langmuir XXXX, XXX, XXX−XXX
Article
Langmuir
Figure 3. Evolving self-healing action and the absence of corrosion is best observed in panels coated with zinc-rich CARC containing MC (described in Scheme 1D). (A) Corrosion resistance that dropped immediately after the scratch recovered and remained high, nearly matching the resistance of the coating before the scratch. (B) Ecorr variations with time indicate synergy between zinc and OTS, with zinc providing robust cathodic protection, allowing the slow-acting OTS to bond fully with the uncorroded steel surface. (The synergy between zinc and OTS is further seen in the Figure 7 data.) Micrographs show the absence of corrosion at much of the scratch location, before (left) and after (right) exposure of the panel for 160 h to the electrolyte; each white bar in the micrograph is 1.6 mm long. See also Figure 1 for scratch locations with corrosion.
quickly rose up to and remained around −0.55 V, a property indicative of direct contact between bare steel and the electrolyte and the absence of a self-healed OTS layer within some areas of the scratch. Optical micrographs of the scratch from the Scheme 1C panel showed evidence of corrosion in some locations (Figure 1), although much of the scratched area remained free of corrosion (Figure 2), suggesting limited corrosion activity that is consistent with the large Rcorr of 106 to 108 Ω·cm2; compare this with
