Federal Highway Administration 100-Year Coating Study

CHAPTER 3. TEST RESULTS AND DISCUSSION

Initial coating characterization results and test results from ALT and outdoor exposure tests pertaining to color, gloss, DFT, pull-off adhesion, surface appearance, surface defects, and rust creepage are presented in this chapter with some supplementary digital images.

3.1 INITIAL CHARACTERIZATION OF COATING SYSTEMS

DFT

Table 5 and figure 14 show mean DFT, standard deviation, and CV of the coating systems measured on type I panels. DFT varied significantly from 7.5 to 17.8 mil depending on the coating system.

Table 5. DFT data for type I panels.

Figure 14. Graph. DFT data (type I panels).

None of the mean DFTs was less than 7 mil, and two of the two-coat systems (ZE/PU and Zn/PS) were between 7 and 8 mil. All three-coat systems (IOZ/E/PU, ZE/E/PU, and MCU/E/F) had mean DFTs between 13 and 15 mil. The remaining two-coat systems (TSZ/LE and ZnE/LE) had the highest DFT values (16–18 mil) and the largest variation of DFT data in terms of CV and standard deviation. The one-coat system (HRCSA) had a DFT of 8.3 mil, which was slightly thicker than two of two-coat systems, ZE/PU and Zn/PS.

Gloss

Figure 15 shows the initial mean gloss readings of type I and type II panels. The eight coating systems exhibited a wide range of initial gloss, with HRCSA having the lowest gloss reading of 9.5 and MCU/E/F, Zn/PS, TSZ/LE, and ZnE/LE having the high gloss readings greater than 70. The other coating systems, IOZ/E/PU, ZE/E/PU, and ZE/PU, possessed intermediate gloss. In most cases, the coating systems exhibited different gloss readings depending on the type of panel (i.e., type I versus type II). This observation suggests that absolute gloss readings may not be useful due to inherent variation of the measurement technique.

Figure 15. Graph. Initial mean gloss.

Color

Initial mean color readings are summarized in table 6. To calculate change of color (ΔE), color readings made at different times were necessary, as shown in the equation in figure 11.

Table 6. Initial mean color readings.

Adhesion Strength

Initial mean adhesion values are shown in table 7 and figure 16. Two of the two-coat systems (ZnE/LE and TSZ/LE) exhibited the highest baseline adhesion strengths of 2,110 and 1,834 psi (CV = 21 and 13 percent), respectively, while the one-coat system (HRCSA) demonstrated the lowest adhesion strength of 259 psi (CV = 45 percent). The poor adhesion strength of HRCSA was observed in a previous FHWA study.⁽²²⁾ The two control coating systems (IOZ/E/PU and ZE/E/PU) had the second and third lowest adhesion strengths of 1,119 and 1,173 psi (CV = 36 and 30 percent), respectively. All remaining three-coat and two-coat systems showed moderate adhesion strengths between 1,200 and 1,400 psi (CV = 17 to 26 percent).

Table 7. Initial pull-off adhesion strength of type I panels.

Figure 16. Graph. Initial pull-off adhesion strength data (type I panels).

3.2 PERFORMANCE DURING ALT AND OUTDOOR EXPOSURE TESTING

Performance of the coating systems in ALT, NW, NWS, and GGB was evaluated by the following parameters:

• Gloss reduction.

• Change of color.

• Change of adhesion strength.

• Development of surface defects.

• Growth of rust creepage at the scribe.

Due to premature coating failures of two of the two-coat systems (TSZ/LE and ZnE/LE) in ALT, the study was terminated after 10 cycles of ALT (3,600 h), 10 months of NW and NWS, and 6 months of GGB exposure. Unexpected early termination of the study led to no data collection from type II panels used for 6 months at GGB. Therefore, most outdoor performance is discussed with the NW and NWS data only.

Gloss Reduction

At the termination of the study, gloss data of each coating system were obtained from type I and type II panels exposed to ALT, NW, and NWS environments. Then, gloss reductions of the coating systems after 3,600 h of ALT and 10 months of NW and NWS testing were calculated with respect to their initial readings. Table 8 and figure 17 present the gloss reduction data. The largest gloss reduction was exhibited by two two-coat systems (TSZ/LE and ZnE/LE) and the one-coat system, HRCSA, as shown in figure 17. These coating systems experienced the most dramatic changes from the highest initial gloss readings (two-coat systems) and the lowest gloss reading (one-coat system). It should be noted that the two-coat systems failed prematurely, and both had the same linear epoxy top coat. Conversely, two other coating systems (MCU/E/F and Zn/PS), which also exhibited high initial gloss readings (see figure 15), exhibited the least amount of changes after testing. The three-coat controls (IOZ/E/PU and ZE/E/PU) and the two-coat system ZE/PU exhibited intermediate initial gloss readings and remained at less than 30 percent gloss reduction after tests. The type II panels in NW tended to show the largest gloss reduction among coating systems, with little to moderate overall gloss reductions.

Table 8. Mean gloss reduction (percent).

Figure 17. Graph. Mean gloss reduction.

Color Changes

Table 9 and figure 18 show mean color change data. HRCSA exhibited the largest color change, followed by two two-coat systems (TSZ/LE and ZnE/LE). These coating systems exhibited the largest gloss reductions. The others exhibited less than 2 percent color changes after testing. Because of scattered data among different test conditions, effects of salt spray and type of test panel on color change were inconclusive, except that most coating systems exhibited the least amount of color changes in ALT.

Table 9. Mean color changes (percent).

Figure 18. Graph. Mean color changes.

Changes in Adhesion Strength

Adhesion strength changes in type I panels after exposure in ALT, NW, and NWS are summarized in table 10, and the corresponding bar graph is shown in figure 19. All coating systems, except HRCSA and MCU/E/F, showed varying degrees of adhesion strength reduction at the end of testing in every test condition. No clear trend was observed between reduction of adhesion strength and test conditions. It is unclear at this time why HRCSA, which exhibited the weakest initial adhesion strength, had the largest strength gain during testing. However, it may be reasoned that because HRCSA had the weakest coating before and after testing, relatively small increases in post-test absolute adhesion strength resulted in significant increases in terms of percentage. Such a trend was more pronounced for NW and NWS panels than ALT panels. Adhesion strength data for TSZ/LE could not be obtained at the termination of ALT because the coating surface was so blistered that the adhesion dollies did not stick to the surface of the panels. Because ZnE/LE panels had peeling top coats at the end of testing, dollies were carefully placed in areas with intact top coats. ZnE/LE panels had the largest adhesion strength reduction. The two worst performing two-coat systems (TSZ/LE and ZnE/LE) in gloss and color also suffered from the largest adhesion reductions. The least amount of adhesion strength reduction was seen in both controls (IOZ/E/PU and ZE/E/PU) as well as ZE/PU. The other two-coat system, Zn/PS, had a moderate adhesion strength loss.

Table 10. Mean adhesion strength changes of type I panels.

Figure 19. Graph. Mean adhesion strength changes of type I panels.

Development of Surface Defects

ALT

The total number of cumulative surface defects that developed during ALT is listed in table 11, and the corresponding line-scatter graph is shown in figure 20. The exact number of defects could not be counted when an excessive amount was detected. As a result, an arbitrary number of 100 was entered in the data sheet. Excessive defects were observed on the surface of two two-coat system (TSZ/LE and ZnE/LE) test panels. Even though initial assessment of these coating systems showed no defects, including invisible defects (holidays), on the surface, a sudden increase in the number of defects during ALT led to severe surface deterioration as indicated by blistering, rusting, and/or cracking, which is discussed in the next section.

In the case of TSZ/LE, one of the panels developed four defects after 1,080 h of testing, and the number of defects increased excessively after 1,440 and 1,800 h of testing. These surface defects were then followed by excessive blistering and cracking of the surface. For ZnE/LE, no defects were detected until 2,880 h of testing. Progressive changes were observed leading to surface microcracks, which transformed to macrocracks. As shown in table 11, MCU/E/F did not develop any defects, while other coating systems only developed minimal coating defects upon termination of ALT at 3,600 h.

Table 11. Cumulative number of surface defects developed during ALT.

Figure 20. Graph. Cumulative number of surface defects developed on type I panels during ALT.

NW and NWS

Most of the type I panels did not develop any defects during outdoor exposure in NW and NWS. Exceptions were two of the two-coat systems, TSZ/LE and ZnE/LE, which performed poorly in other tests. TSZ/LE exhibited countless defects on every NW and NWS type I panel. For ZnE/LE, five defects on NW panels and seven defects on NWS panels were developed after 10 months of outdoor testing. All type I and type II test panels employed in NW and NWS developed many defects along the panel edges. Type II panels also had coating defects in areas such as nuts, bolts, underside of the T-attachment, and the wide-angle attachment due to less than ideal coating application conditions. As expected, these defects became rust spots. This finding confirms that it is difficult to avoid initial coating defects from coating applications on field bridge structures due to complex shapes of structural elements. These imperfect sites are prone to advanced coating failures and subsequent steel corrosion in service environments.

Table 12 lists coating defect development for type II panels. For the sake of simplicity in this report, only DFT-measured flat areas on all panels were studied for defect development. Zn/PS initially contained numerous coating defects in the low DFT areas, and none of the high and normal DFT areas developed additional defects during outdoor exposure testing. Although TSZ/LE started with a moderate number of coating defects in the low DFT and six in the normal DFT areas, they ended up with numerous defects on every panel regardless of DFT and exposure condition. Also, the low DFT area of the ZnE/LE coating system in NWS developed greater than 70 percent surface deterioration accompanied by surface cracking and coating failure. HRCSA started with 3 initial defects and developed 15 defects in the low DFT and 1 defect in the normal DFT areas in both NW and NWS. The 15 defects in the NW and NWS environments appeared to be related to mechanical damage through the soft coating. The other coating systems did not develop additional coating defects during the 10 months of exposure testing. Coating defect data were not available for the panels at the GGB exposure site.

Table 12. Number of surface defects developed on type II panels during NW and NWS exposure testing.

Rust Creepage

ALT

Table 13 and figure 21 show rust creepage data measured during 3,600 h of ALT. Because all of the TSZ/LE panels started to develop serious surface deterioration after only 1,080 h of ALT, rust creepage data from this coating system were not measured.

Based on the mean creepage values at the end of 3,600 h of ALT, the coating systems were ranked in the following order from highest to lowest rust creepage:

MCU/E/F >> ZE/E/PU ≈Zn/PS ≈ZE/PU ≈ZnE/LE > IOZ/E/PU > HRCSA

It should be noted that ZE/E/PU creepage data showed a steadily decreasing trend during the last three cycles, which was caused by measurement errors. Based on a rapid creepage growth of ZnE/LE observed from the sixth cycle (2,520 h), it can be projected that the coating system would have shown a very high rust creepage if ALT had continued for more cycles. As observed in previous a FHWA one-coat study, HRCSA had the best performance for resistance to rust creepage.⁽²²⁾

Table 13. Rust creepage growth during ALT (mm).

Figure 21. Graph. Rust creepage growth with time during ALT.

NW and NWS

None of the type I panels developed any rust creepage during NW and NWS outdoor exposure testing except for ZnE/LE, which had a rust creepage of 0.16 inches at the end of NWS exposure. After 10 months of exposure of type II panels in NW and NWS testing and 6 months of exposure at the GGB, only the ZnE/LE coating system exhibited recognizable rust creepage, as shown in table 14 and figure 22. High DFT areas of type II panels with ZnE/LE in NWS showed extremely high creepage of around 0.51 inches. Similarly, the nominal DFT area of the same panel showed creepage of about 0.23 inches. The rest of the type II coating systems did not show any rust creepage. It is interesting to observe that ZnE/LE, which was mediocre in ALT, showed the worst rust creepage performance in NW and NWS, whereas MCU/E/F, which performed the worst in ALT, performed well in NW and NWS without rust creepage.

Table 14. Rust creepage growth of ZnE/LE during outdoor exposure.

1 inch = 25.4 mm

Figure 22. Graph. Rust creepage growth of ZnE/LE (type II panels).

Physical Condition of Representative Test Panels

Figure 23 through figure 30 show typical digital photographs of all coating systems for type I panels during ALT at 0, 1,440, 2,520, and 3,600 h. These figures also show representative photographs of the type I panels in NW and NWS before and at the termination of testing at 10 months.

ALT

The two controls, IOZ/E/PU and ZE/E/PU (see figure 23 and figure 24), had the best surface retention properties in all exposure conditions despite the fact that ZE/E/PU exhibited the second highest rust creepage. The one-coat system, HRCSA, also retained good surface characteristics with the least rust creepage and minimal defect development despite the fact that it exhibited the third largest gloss reduction and the highest color change (see figure 30). As observed in a previous FHWA one-coat study, HRCSA remained very soft throughout the testing.⁽²²⁾ The third three-coat system, MCU/E/F did not develop defects, but it had the highest creepage at the termination of ALT (see figure 25). Two of the two-coat systems, ZE/PU and Zn/PS, had moderate defect development and rust creepage growth (see figure 26 and figure 27). The remaining two-coat systems, TSZ/LE and ZnE/LE, had the worst surface deterioration over time in ALT as indicated by the development of defects, surface blistering, coating peel-off, and/or surface cracking (see figure 28 and figure 29).

The visual observation of surface changes was followed up with digital microscopy of the test panel surfaces. Digital microscopy of TSZ/LE showed the various phases of progressive deterioration of the coating system at 0, 1,800, and 3,600 h (see figure 31). At 1,800 h, the surface of the coating started to develop blisters, which progressively grew to form a completely blistered coating surface at the termination of ALT, as shown in figure 31. Certain areas still had blisters intact, while some areas had half-peeled or detached coating from the surface. White residual zinc oxide formed on the surface where the coating had peeled off.

Digital microscopy of ZnE/LE during ALT yielded unexpected critical observations. The surface of the coating system showed microcracks all over the surface of a test panel, as shown in figure 32. The central points at which these microcracks originated were a few mils in size, and the cracks themselves were a few hundred mils in size. The microcracks developed into surface macrocracks at the termination of ALT (see figure 32). However, these test panels did not show any surface defects indicated by closed circuit shorting (beeping) during holiday testing, indicating that the cracks did not develop through the coating thickness and might have formed superficially near the surface. Comparison of digital microscopy images before and after exposure indicated that these crack-originating locations were present on the panel surface before testing. However, cracks that propagated from these locations appeared only after ALT. The density of these cracks did not seem to grow as ALT progressed.

NW and NWS

Figure 23 through figure 30 also show type I panels exposed to NW and NWS. In general, their physical conditions appeared to be better than their counterparts in ALT. TSZ/LE exposed to NWS formed distinctive white zinc oxide within the scribe. Additionally, ZnE/LE and HRCSA under NWS exposure exhibited heavy rust stains in some areas. Interestingly, none of the TSZ/LE and ZnE/LE type I panels, shown in figure 28 and figure 29, respectively, suffered from the same surface deteriorations of blistering (see figure 31) and cracking (see figure 32) observed on their ALT counterparts. Such a finding suggests that either different deterioration mechanisms worked in ALT and NW/NWS or that 10 months in outdoor exposure testing was not sufficient to introduce significant deterioration.

Figure 33 through figure 40 show digital photographs of all coating systems on type II panels before exposure and at 6 months of outdoor exposure testing at GGB. These figures also show conditions of panels before and after exposure in NW and NWS for 10 months. Although each type II panel consisted of low, high, and nominal DFT areas, rust was observed predominantly at hard-to-reach areas where coating was non-uniformly applied and/or where there were coating defects (bare areas). These areas included nuts, bolts, underside of the T-attachment, and the wide-angle attachment. TSZ/LE showed formation of white zinc oxide within the scribe on all DFT areas as well as along the edges of nuts and bolts and underneath the attachments (see figure 38). The ZnE/LE panel exhibited excessive rusting on all DFT areas and along the edges of nuts and bolts and underneath the attachments in NWS (see figure 39). The same panel also showed rust buildup within the V-notch section where salt water could be collected. This form of deterioration reveals that ZnE/LE was particularly venerable to prolonged exposure to salt water. The low DFT area of this coating system developed cracks all over the surface, which was similar to what was observed toward the end of ALT (see figure 32). The two controls did not show any signs of rusting within the three DFT areas (see figure 33 and figure 34). As observed in ALT, the next best performing coating system in comparison to the two three-coat controls was HRCSA, as indicated by negligible rusting and rust creepage (see figure 40).

Figure 23. Photo. Progressive changes of IOZ/E/PU—type I in ALT, NW, and NWS.

Figure 24. Photo. Progressive changes of ZE/E/PU—type I in ALT, NW, and NWS.

Figure 25. Photo. Progressive changes of MCU/E/F—type I in ALT, NW, and NWS.

Figure 26. Photo. Progressive changes of ZE/PU—type I in ALT, NW, and NWS.

Figure 27. Photo. Progressive changes of Zn/PS—type I in ALT, NW, and NWS

Figure 28. Photo. Progressive changes of TSZ/LE—type I in ALT, NW, and NWS.

Figure 29. Photo. Progressive changes of ZnE/LE—type I in ALT, NW, and NWS.

Figure 30. Photo. Progressive changes of HRCSA—type I in ALT, NW, and NWS.

Figure 31. Photo. Photomicrographs of progressive changes of TSZ/LE—type I in ALT.

Figure 32. Photo. Photomicrographs of progressive changes of ZnE/LE—type I in ALT.

Figure 33. Photo. Progressive changes of IOZ/E/PU—type II in NW, NWS, and GGB.

Figure 34. Photo. Progressive changes of ZE/E/PU—type II in NW, NWS, and GGB.

Figure 35. Photo. Progressive changes of MCU/E/F—type II in NW, NWS, and GGB.

Figure 36. Photo. Progressive changes of ZE/PU—type II in NW, NWS, and GGB.

Figure 37. Photo. Progressive changes of Zn/PS—type II in NW, NWS, and GGB.

Figure 38. Photo. Progressive changes of TSZ/LE—type II in NW, NWS, and GGB.

Figure 39. Photo. Progressive changes of ZnE/LE—type II in NW, NWS, and GGB.

Figure 40. Photo. Progressive changes of HRCSA—type II in NW, NWS, and GGB.