D J Mills*, S Mabbutt* and R Akid+
*School of Technology and Design, University College Northampton, Northampton NN2 6JD E-Mail :
+Centre for Corrosion Technology, Materials Research Institute, Sheffield Hallam, University, Sheffield, S1 1WB E-mail :
Two standard coatings containing inhibitive pigments were scribed and the degree of protection afforded was investigated when the specimens were immersed in 10% Harrison's solution. This was assessed both visually, using ferricyanide solution and electrochemically. The Electrochemical Noise Method was used to monitor corrosion and it was found that the higher the value of Rn the greater the degree of protection afforded. The Scanning Reference Electrode Technique was also used on the same scribed coatings. These results which provided an indication of the amount of anodic activity along the scribe correlated well with the Electrochemical Noise results. The effect of a pre-soak in de-mineralised water was also investigated. It is suggested that the two methods when used together can provide the means by which paint manufacturers can quickly and easily assess the effectiveness of their systems in preventing corrosion at a break in the coating.
Keywords Electrochemical Noise, Scanning Reference Electrode Technique (SRET), anti-corrosive coatings, scribe protection
At University College Northampton there has been a programme of work investigating electrochemical techniques to examine intact organic coatings on steel. Many researchers have tried to use the AC impedance method for this . Some of the fairly recent papers using this method are summarised in a review by Murray . Although this method is of undoubted use it produces a lot of information which is sometimes difficult to interpret and from an experimental disadvantage point of view, is electrically intrusive. In the search for a relatively simple technique to replace the successful DC resistance method (also electrically intrusive), the Electrochemical Noise Method (ENM) was selected. This has now been quite extensively investigated . This followed on from previous work by Skerry [4,5] at Sherwin-Williams, Cleveland and by Bierwagen's group [6,7] at NDSU in Fargo, North Dakota. Recent work by two of the present authors has examined detached and waterborne coatings on steel and aluminium [3,8]. The study has now progressed to examine defects in coatings - deliberately created either using a scribe or by laser ablation. Some interesting electrical responses have been obtained and there is indirect evidence of processes occurring at the scribed area such as protective film formation. However what was lacking, particularly in the early stages of exposure, was direct evidence of microscopic and local electrochemical activity as a function of time.
At Sheffield and elsewhere SRET (Scanning Reference Electrode Technique) has been used and modified to study a variety of corrosion processes [9-12]. With the ability to give localised, quantitative information about anodic and cathodic processes it has proved a very useful tool. Particularly useful for coatings evaluation is the rotational configured SRET as application is easy and the problem of height variation is minimised. A project in early 1998 looked at a Polyaniline coating on steel using SRET. As expected from previous work with intact coatings, the high resistance of the coating made it difficult to "pick up" anodic and cathodic areas, at least during the initial stages of exposure. However when used to examine deliberately created circular defects results proved quite promising.
The general aim of the work described in this paper is to identify a fast laboratory test method to assess the effectiveness of inhibitive materials within a coating to protect the substrate at a break. Previous studies looking at point defects or scribes in coatings include work by De Wit using AC Impedance  and work by Mills using the DC resistance method . A further objective of the work is to gain a mechanistic understanding of the coating behaviour where protection takes place. Defects are being investigated in several solvent-borne and water-borne primers on steel using the Electrochemical Noise Method. Reported here are results for two proprietary primers, an epoxy (CM) and an alkyd (RH). Red Lead in linseed oil has also been investigated along with a control (water-borne unpigmented epoxy). All apart from the control contain inhibitive pigments in the coatings. Defects were produced mainly by scribing. Resistance (Rn) measurements were monitored as a function of time in either de-mineralised water or 10% Harrison's solution. It was found that none of the inhibitive coatings were successful in preventing corrosion upon immediate immersion in 10% Harrison�s solution. So coatings were given a pre-soak by immersing for a time of 24 or 96 h in de-mineralised water. SRET has also been used to examine scribed (deliberately damaged) areas in the same solutions over a similar period of time. The expectation is that SRET would be able to provide useful information about the initial stages of passivity, or of substrate damage i.e. corrosion, at the (initially) bare metal area.
The substrate used was steel Q panels or steel circular (22mm diameter) SRET specimens. The two solvent-borne primers CM (epoxy) and RH (alkyd) were expected to contain inhibitive pigments based on zinc phosphate and zinc tetroxy chromate respectively. The solvent for the amine portion in the unpigmented epoxy (IC 152) was water. This coating would not be expected to contain any components that would lead to effective inhibition at a scribe. The red lead was in linseed oil. It was used because it was expected to be effective. Coatings were applied by spreader bar to Q panels and by dipping for SRET specimens. Harrison�s solution (3.5% ammonium sulphate, 0.5% NaCl) was used unless indicated otherwise at 10% dilution for the tests reported here.
Most of the Electrochemical Noise results for CM and RH coatings were obtained from the same samples used for SRET (i.e. cylinders). The beaker method was employed. This is shown in Figure 1. Fresh scribes were made either with or without re-coating before SRET measurements. Noise results from the red lead and IC 152 were obtained from Q panel samples using the bridge method . Blanking was achieved using a 75 / 25 mix of beeswax and colophony resin. The scratch area was typically 0.02cm2 (10mm x 0.2mm). The scribes were produced by a novel mechanical removal technique using a specially designed tool that removes the area of coating rather than displacing it by plastic/elastic deformation. The cutting edge of the blade resembles a small lathe (part-off tool) when viewed under magnification. This lifts the coating and shears the edges of the cut to give a square section scribe with an easily calculable area and leads to more reproducible results. Potassium ferricyanide indicator was used throughout this work to establish if corrosion was actually taking place at a particular point at a particular time. The indicator is very sensitive to Fe++ ions, a blue colour can be detected even when only traces of iron ions are present.
Electrochemical Noise was measured using a zero resistance ammeter and data logger with dedicated software supplied by ACM. Measurements are generally made every half-second for a period of 300 s. The current between the two working electrodes is monitored and at the same time the voltage of the pair is measured with respect to the reference electrode. From the standard deviations of the resultant 300 data points, the voltage noise and current noise are calculated. By dividing the former by the latter, the parameter Noise Resistance (Rn) is derived. Further details can be found elsewhere [3,8].
SRET results were obtained by mapping over an area of 1cm by 1.5cm that included the scribe. The specimen was rotated at 100rpm and the measurements were taken over a period of up to about thirty hours. A diagram showing the experimental set-up is shown in Figure 2. Further details on the experimental set-up can be found [9,10].
Although not given here intact coatings were also investigated which gave Rn values that were considerably higher than the scribed sample results. Thus all values here are attributed to the scribe. Figure 3 shows the comparison between the clear epoxy and the CM coating after just 2 h in solution. It can be seen that the Rn value of the CM coating is approximately an order of magnitude higher than the unpigmented epoxy. Reproducibility among the four different pairs was very good. Figure 4 shows the effect of time over a 24 h period for both CM and RH coatings in 10% Harrison�s solution. Initially the CM coating provided a level of inhibition (corresponding to an increase in the Rn value). However within 24h both coatings were failing to inhibit although the nature of the corrosion product was somewhat different in the two cases. Nevertheless some degree of reduction in corrosion rate, based on Rn beyond that shown by the unpigmented epoxy, was evident. Results for the red lead coating are shown in Figure 5. When immersed in 10% Harrison�s solution this system was not very effectively in providing inhibition, as shown by the ferricyanide test and low Rn values. However when pre-soaked for 96 h in water and then transferred to Harrison�s solution, it maintained a sustained high value of Rn and no visible corrosion was seen at the scribe. This graph also shows that when the Harrison�s solution was increased to 55% of full strength this overwhelmed the inhibitor and visible corrosion ensued along with a drop in the Rn value to the levels of the non-inhibiting situation. Figure 6 shows a comparison of Rn value between red lead and the CM and RH coatings. After a 24h soak in de-mineralised water the latter two primers failed to afford much long-term (more than 24 h) protection at the scribe. However (as already seen in figure 5) the red lead maintained a high Rn value.
SRET is able to provide a spatially resolved measurement of electrochemical activity at pre-selected intervals over a given period of immersion. Note that although results given here are only for single specimens, the effects described were reproducible i.e. very similar results were obtained when they were rescribed and reimmersed and also from second, nominally identical specimens. The region of the scratch is shown in Figure 7. Some qualitative results obtained for the CM coating in 10% Harrison's are shown in Figure 8. The green areas indicate anodic activity. It can be seen that there is initial anodic activity in the centre of the scribe but this decreases within 4h. If the specimen is re-scribed (after 24h) this anodic activity recommences along the whole scribe but then falls away to almost zero after 10 days (by this time there is some red corrosion product visible in the base of the scribe). Results for RH coating are similar (see Figure 9). However there is a greater degree of electrochemical activity overall which agrees with the ENM results. Although after 10 days this has diminished, there was a quite voluminous yellow corrosion product visible in the scribe of the RH coated specimen.
The effect of a short pre-soak for the CM coating was also investigated and the results are shown in a timed sequence over a period of some 24 h in Figure 10 and 11. These results are presented in black and white where the shade of grey corresponds to a probe output value in mV. Note the darker the shade the higher the current density.
Although both sequences were obtained with CM specimens in 10% Harrison's solution, the results in Figure 10 were obtained after preliminary immersion in de-mineralised (DM) water for 2-3 h whereas those shown in Figure 11 were obtained after immediate immersion in Harrison�s solution. Corrosion can be seen to increase/decrease over 24h and corresponds approximately with the Rn results for the CM coatings as shown in Figure 2. However note that there is somewhat decreased activity overall in the pre-soaked sample showing that (as with red lead) there is some advantage in a pre-soak with the CM coating. It should be noted that the electrochemical noise and ferricyanide test results indicated that pre-immersion did not apparently cause any long-term decrease in corrosion. However the greater sensitivity of the SRET shows there is some effect at least in the early stages.
 J. Murray Progress in Organic Coatings, June, (1997)
 H. Xiao and F. Mansfeld J. Electrochem. Soc. 141 9, p 2332-2337 (1994)
 S. Mabbutt and D. J. Mills, Surface Coatings International, 80,1, p18-25, (1997)  B. S. Skerry and D. A. Eden, Prog. Org. Coat., 15, p269-285, (1987)
 C. T. Chen and B. S. Skerry, Corrosion 47, p598-61, (1991)
 G. P. Bierwagen, D. J. Mills, D.E. Tallman , B. S. Skerry, Proc. Conf. Electrochemical Noise for Corrosion Applications, May 1994 Montreal pub. ASTM STP No. 1277, (1996)
 D.J. Mills, G. P. Bierwagen, B. Skerry and D. Tallman, Materials Performance 34, 5, p33-38, (1995)
 D. J. Mills and S. Mabbutt, EFC Publication No 20, pub Institute of Materials (Ed L Fedrizzi and P L Bonora) p83-93, (1997)
 R. Akid, Materials World, November, p 522-525, (1995),
 R. Akid, Corrosion Management, July/August, p 14-16, (1999)
 H S Isaacs and B Vyas, in Electrochemical Corrosion Testing, F Mansfeld and U Bertocci (Eds), ASTM 1981, p3-33
 G Pallos and G Wallwork Corrosion, 38, 1982, p 305-312
 F. M. Geenen, J. H. W. De Wit and E. P. M. Van Westing. Progress in Organic Coatings,18, p299-312, (1990)
 D J Mills and P J Boden, Corrosion Science 35, p1311-1318, (1993)
 RC Bacon, J Smith and Rugg, Ind. Eng. Chem. 40, p (1948)
Figure 1. Experimental set up for EN measurements.
Figure 2. SRET experimental arrangement
Figure 3. Noise Resistance (Rn) values for scribed CM(1-4) and IC 152 (5,6) coatings after 2h in 10 % Harrison�s solution
Figure 4. Effect of time on Rn, Scribed CM epoxy coating and RH alkyd coating in 10% Harrison's Solution.
Figure 5. Effect of time on Rn for scribed red lead samples in 10% Harrison's solution. (Non-inhibited) immediate immersion. (Soaked) 0-192h in 10% Harrison's solution, 192-350h in 55% Harrison's solution.
Figure 6. Effect of time on Rn for CM, RH and Pre-soaked Red Lead Scribed Coatings.
Figure 7. SRET map showing corrosion at the scratch location after 4 hours immersion
|(a) CM 0.5 hr (A1)||(b) CM 4 hr (B2)|
|(c) CM 27 hr (cleaned) (1)||(d) CM 10 days|
Figure 8. CM specimens in 10% Harrison�s solution. Note coating re-scribed after 27h. (click for a larger image)
|(a) RH 0.5 hr||(b) RH 2 hr|
|(c) RH 26 hr (+20 mV)||(d) RH 10 days|
Figure 9. RH specimens in 10% Harrison's solution. Note imposed potential of +20 mV after 26h. (click for a larger image)
a) 15 min
b) 5 hr
c) 9 hr
Figure 10 CM Coating after 2 hr pre-soak in demineralised water, followed by immersion in 10% Harrison�s; scribe mark indicated by arrow
a) 2 hr
b) 3 hr
c) 25 hr
Figure 11 CM Coating after immersion in 10% Harrison�s; scribe mark indicated by arrow