Marine Atmospheric Corrosion Museum Report
on the Performance of Thermal Spray Coatings
on Steel

By: Robert M. Kain, Corrosion Scientist
and Earl A. Baker, Senior Research Technologist

Abstract
Corrosion control is usually implemented by alloy upgrading or material substitution, environmental modification or providing a protective barrier or sacrificial surface or both. Historically, the coatings approach has been the predominant first choice based on initial economics. Current and emerging coating technologies are continuing to expand the methodology and perfection of metallic and inorganic coatings.

Plated metal, chemical conversion coatings, and flame sprayed zinc and aluminum have been used for many years to protect steel and other metals. The thermal spraying approach is well suited for larger structures as coatings may be applied in the shop or field. Newer processes, such as vapor deposition and ion implantation, for now at least, are likely to be limited to providing improved corrosion resistance for small, critical components.

Regardless of the type or method of coating, anticipated performance over the lifetime of the equipment is based on laboratory testing and field trials. In this regard, long-term exposures in a variety of industrial and marine environments have been a recognized approach to characterizing coating performance.

In the early 1950's a number of zinc-aluminum compositions were applied to carbon steel using then- current practice and exposed at several test sites in England and the United States. The performance of these test panels was described by Hoar and Radovici following 10 ½ years of exposure. In the ensuing years, a number of the original panels remained on exposure at Kure Beach, NC, in the LaQue Center for Corrosion Technology, Inc., marine atmospheric exposure site.

This paper describes the protection afforded by the various coatings after 34 years of continuous exposure in the moderately aggressive marine location situated 250 m from the Atlantic Ocean. Several electrochemical methods presently in use are reviewed in a reflection of how accurate they might be in forecasting the effectiveness of these coatings.

Key Words
corrosion resistance, marine atmospheres, thermal spray coatings, single element powder, multi- metal powder, mixed powder, alloy powder, steels, electrochemical behavior.

Steel remains the economic choice of materials for construction of large structures. Over the years, numerous protection schemes involving organic, inorganic, metallic coatings, and combinations thereof have been successfully developed to extend the service life of these structures in corrosive environments. The present acceptance of thermal spray coatings of zinc and aluminum, for example, can be attributed to many years of documented field trials and service histories. Much of this experience has been gained from testing under natural atmospheric conditions of varying degrees of corrosivity. Among the often cited marine locations are the atmospheric test sites maintained by the LaQue Center for Corrosion Technology, Inc., at Kure Beach, NC. A unique feature of this facility is the presence of an outdoor corrosion museum containing numerous examples of classic corrosion behavior for a broad range of coated and uncoated materials. In many cases, specimens have been retained on testing long after their originally intended exposure period. This paper reports on a series of thermal spray coated steel panels that have been retained at the 250-m museum site since Feb. 1952 [1].

In addition, present day laboratory techniques were utilized to characterize the electrochemical behavior of 34-year exposed materials and new thermal sprayed aluminum and zinc wire coated materials recently reported elsewhere [2].

Background
The subject materials were originally part of an extensive matrix developed at Cambridge University, England, in 1951 to optimize coating composition and evaluate application methods. Under the direction of T. P. Hoar, the test program considered 27 coatings of single and multi-metal compositions. While recognizing pre-alloyed wire technology, mixed powders (MP) and alloy powders (AP) were applied with two types of guns using propane-oxygen-air flames. One of the gun models (identified herein as gun #2), later reported to be commercially obsolete, applied the coating via gas conveying and air conveying methods. In all, approximately 1650 panels with the thermal spray coatings identified in Table 1 were prepared.

Coatings were applied to grit blasted, 14-gage mild steel panels measuring 5 by 7.5 cm (2 by 3 in.) (all original dimensions and coating thickness measurements were in English units). Edges were precoated before the application of the desired thickness of 0.08 mm (3 mil). In some cases, thicker coatings, as indicated in the text, were intentionally deposited for zinc, aluminum, and manganese and dual layer Zn + Al and Al + Zn at the time of preparation. It is not known how actual coating thicknesses were quantified.

In order to ascertain a coating performance over a range of atmospheric conditions, Hoar chose two marine sites, one described as mild (Frinton on Sea, Essex, England) and the other moderately severe (250-m lot at Kure Beach, NC), and three industrial sites in England ranging from mild (Cambridge) to severe, acid conditions (Stratford, London).

Table 1 - Identification of various thermal spray coated steel panels in marine atmospheric corrosion museum at Kure Beach, NC

An earlier report by Hoar and Radovici described the performance of 20 zinc-aluminum coatings after exposure periods up to 12 years [1]. Although subsequent inspections were performed for these and the balance of the coatings identified in Table 1, it has not been determined if further documentation has been reported in the open literature. Performance described herein is based on visual inspection of the remaining specimen of each coating conducted in Dec. 1985.

Atmospheric Exposure
Test panels on exposure in the 250-m lot at Kure Beach were positioned on wood frames supported at a 30o incline from horizontal facing due south. Panels were attached to the frames with case- hardened, sherardized (a process which produces a coating of iron-zinc intermetallic compounds) and varnished steel screws with an insulating, fiber washer. These were subsequently replaced with insulated brass screws after about 20 years. The severity of the atmospheric conditions at Kure Beach can be affected by a number of environmental factors, including: prevailing wind direction, airborne chloride content, and time of wetness/relative humidity [3]. Figure 1 (all figures available upon request) documents the yearly corrosion rate for uncoated wrought iron and the chloride level determined by the wet candle method for the period covering most of the exposure time for the thermal spray coated panels.

Laboratory Tests
Following the 34-year inspection, several selected panels were sectioned and prepared for metallographic examination (light-optical and scanning electronmicroscopy) and for electrochemical testing. Test electrodes were cut from the same relative position of the selected panels regardless of coating condition. Remaining portions were edge sealed with epoxy and returned to the museum site at Kure Beach.

Previously exposed atmospheric test materials and newly prepared specimens coated with thermal sprayed zinc wire and aluminum wire were subjected to potentiodynamic polarization tests and to linear polarization resistance tests. Both experiments were conducted in natural, filtered seawater at 25o C utilizing electrodes with cut edges and groundward surfaces masked with epoxy. Before testing, the electrodes were rinsed with distilled water, then alcohol and air dried.

Electrodes were first tested for polarization resistance since this involved minimum alteration of the electrode potential, that is, -10 to +10 mV, form a 1-h ECORR value. An EG&G, PAR Model 173 potentiostat was used for this study.

Data acquisition and calculation of polarization resistance Rp values in terms of k-ohms/cm2 were performed with PARC Soft CORR-332 (trademark of Princeton Applied Research Corporation) corrosion software. Figure 2 schematically shows a linear plot of current versus potential for determining Rp.

Following the polarization resistance test, potentiodynamic scans were developed for the same electrodes using a potential sweep rate of 0.6 V/h in the anodic direction in accordance with ASTM Practice for Standard Reference Method for Making Potentiostatic and Potentiodynamic Anodic Polarization Measurements (G5) [4]. For these tests, the PAR 173 potentiostat was interfaced with an Apple computer. Hardcopy plots were prepared following transfer of the digital data to a Tektronix computer/plotter.

Results and Discussion
34-Year Inspection Results
Performance of the various coating schemes described herein is based on visual estimates of the percent surface area with base metal rust or yellow rust stain or both. Yellow rust stain is the appearance produced by substrate corrosion bleeding through porosity in an otherwise intact coating. As will become apparent, considerable variation in the extent of corrosion protection was afforded by these coatings.

Table 2 - Identification of protective thermal spray coated steel panels after 34-years exposure in the 250-m atmospheric lot at Kure Beach, NC

*May have up to 5% area covered with yellow rust stain.
NOTE:  MP = mixed powders, AP = alloy powder, SEP = single element powder


Performance Ranking
Coatings are ranked by one of the following three categories of performance: protective (Table 2), that is, panels exhibiting 0% base metal rust or up to 5% yellow rust stain; marginally protective (Table 3), i.e., those exhibiting up to 30% bare metal rust plus yellow rust stain; and finally non- protective, that is, panels with 50 to 100% base metal rust and yellow rust stain. Examples of each of the above are provided by the photographs in Fig. 3. In a number of cases, those coatings identified as nonprotective after 34 years exhibited substantial degradation much earlier. The 22% Zn-78% Sn coated panel shown in Fig. 3, for example, had already exhibited 100% base metal rust within 14 years. In contrast, many of those presently identified as marginally protective, for example, 80% Zn + 20% Al(AP), exhibited little or no rust or staining at that time.

Table 3 - Identification of marginally protective thermal spray coated on steel panels after 34-years exposure in 250-m atmospheric lot at Kure Beach, NC

Application Effects
A review of Tables 2 through 4 will show that the numerous coatings were influenced by the application method and type of powder used, that is, mixed or alloy (different metal powders were mechanically mixed or powder from pre-alloyed material was used). For the most part, performance of 100% zinc and 100% aluminum coatings, applied singularly or as dual layers, were affected least by differences in the gun applicator.

Table 4 - Identification of nonprotective thermal spray coated steel panels after 34-years exposure in 250-m atmospheric lot at Kure Beach, NC

As shown in Fig. 4, however, the 0.08-mm zinc coating was found to be sensitive to the application method. Also shown in the variation in behavior exhibited by the 50% zinc-50% aluminum mixed powder coating. While both of the above performed well after only an application with gun #2, the 0.08-mm aluminum coating was consistently protective regardless of application.

Single Element and Dual Layer Coatings
With the exception of the previously mentioned 0.08-mm zinc coating, all other 100% zinc (0.15 mm) and 100% aluminum (0.08 and 0.15 mm) coatings and dual layer coatings of aluminum over zinc and zinc over aluminum (0.15 mm combined) are identified in Table 2 as protective. The appearance of those prepared with the standard gun and exposed at the Kure Beach site for 34 years is shown in Figs. 5 and 6. Although performance rankings are based on base metal rust and staining, it was found that some deterioration of these protective coatings had occurred in the formation of aluminum and zinc surface nodules.

Single element manganese coatings were ranked as either marginally protective or nonprotective depending on applicator and thickness. Figure 7 compares the percent area affected for the manganese and other single element coatings (including dual layer Zn/Al and Al/Zn). As can be seen, increasing the manganese coating thickness from 0.08 or 0.13 mm to about 0.20 mm greatly improved performance when used in conjunction with the standard gun applicator. In contrast, even greater protection was afforded by increasing the zinc coating thickness to only 0.15 mm.

Mg Containing Coatings - From Table 2 it can be seen that several aluminum base magnesium containing coatings (mixed powder) were found to be protective. Figure 8 shows greater resistance for aluminum-magnesium coatings in comparison to zinc-magnesium coatings with the former being more tolerant of higher levels of magnesium. While the mixed tri-metal powder coating containing 60% Al-20% Zn-20% Mg is shown to be marginally protective (Table 3), the alloy powder version of the above the both versions of the aluminum base zinc-magnesium coating were found to be nonprotective. The application method had minimal impact on the behavior of magnesium containing bi-metal and tri-metal coatings.

22% Zn-78% Sn Mixed Powder Coating - The performance of the 22% Zn-78% Sn mixed powder coating is compared with several other bi-metal coatings shown in Fig. 8. As indicated earlier, however, 100% base metal rust was observed within 14 years and accounts for the severe corrosion shown in Fig. 3. All six panels were similarly affected.

Zn-Al and Al-Zn Coatings - At least one version, that is, mixed or alloy powder, of each the Zn-Al and Al-Zn coatings was found to be protective. Figure 9, for example, shows the appearance of a 90% Zn-10% Al mixed powder coating and an equally resistant 10% Zn-90% Al alloy powder coating. From Tables 2 and 3, it can be seen that coating performance was influenced by powder type and application method. Overall, the higher aluminum containing coatings, for example, 70 and 80% aluminum, exhibited the least variability in this regard. Most of the more protective coatings, regardless of powder type and applicator, contain 50% or greater aluminum. From this bi- metal series only six coatings, all containing 50 to 90% zinc, are identified in Table 4 as being nonprotective. Figure 10 compares the percent area affected by base metal rust and yellow stain for mixed powder coatings applied by the then standard gun technology. Of this group, only the 50% Zn-50% Al coating is classified in the non-protective category. While the 90% Al-10% Zn coated panel exhibited about 5% base metal rust, the corresponding alloy powder version was rust free after 34 years. This performance and that exhibited by the 80% Al-10% Zn coating has practical significance in that the above approximate the composition of present day 15% Zn-85% Al alloy wire sprayed coatings [2].

Table 5 - Corrosion potential for various thermal spray coated steel electrodes in seawater

(a) Powder coatings (all with standard gun applicator).
(b) Wire coatings from Shaw-Moran study.


Metallography
Figure 11 shows scanning electron micrographs of the surfaces of the 0.15-mm zinc and 0.15-mm aluminum coated panels and dual layer Al + Zn and the Zn + Al coated panels exposed for 34 years in the 250-m atmospheric lot. For both the Al and the Zn + Al coating, the splatted powders are clearly visible with minimal interference from aluminum corrosion product. The surface condition of these aluminum powder coatings appears quite similar to that recently described by Shaw and Moran [2] for new, unexposed coatings deposited by thermally sprayed aluminum wire. In the case of the Zn and the Al + Zn dual layer coating, the products of zinc corrosion are evident among the splatted particles.

Cross-sectional photomicrographs of the 0.08- and 0.15-mm zinc, and 0.08- and 0.15-mm aluminum are shown in Fig. 12 while similar views of the dual layered Zn + Al and Al + Zn coated panels are shown in Fig. 13.

From the respective cross sections of the 0.08- and 0.15-mm zinc coatings it is easy to explain the observed differences in base metal rust coverage indicated earlier in Fig. 7. Conversely, both the 0.08- and 0.15-mm aluminum coatings provided adequate protection. It can be seen, however, that the 0.08-mm aluminum coating appears to lack the cohesion between splat particle layers exhibited by the 0.15-mm aluminum coating.

Overall, the cross-sectional appearance of the thicker coatings more closely approximates the structure described by Shaw and Moran [2]. In the case of the dual layer coatings (Fig. 11), the appearance of the aluminum layers, regardless of lower or upper deposition, is similar. On the other hand, the zinc layer closest to the steel substrate appears to have deteriorated.

Examination of the photomicrographs indicate greater overall thickness than that originally cited by Hoar. Tinsley gage measurements made after 14-years exposure also indicated greater average thickness. It is perhaps possible that some of this difference in measured thickness can be attributed to swelling of the coating. This might be related to disbondment between splatted particles or forces exerted by entrapped corrosion product.

Table 6 - Polarization resistance values for various thermal spray coated steel electrodes in seawater

(a) Powder coatings.
(b) % base metal rust/% yellow rust stain.

Electrochemical Studies
Electrochemical measurements typically require the presence of a bulk electrolyte. Recognizing the limits of obtaining such data on materials in the atmosphere, natural seawater was selected as a convenient conductive medium. At some time of atmospheric wetness, however, conductivity of a water film on test panels could conceivably approach that of seawater. Nonetheless, the purpose of these experiments was to obtain relative delineation of post exposure (atmospheric) corrosion behavior of the various coated materials. As such, the following results are not intended for predicting seawater immersion resistance where morphology and protective properties of surface films and corrosion products might differ over prolonged periods.

Table 5 identifies a number of the 34-year-old exposed panels coated by the standard gun application and several new materials selected for electrochemical testing. Also provided are data giving their 1-h corrosion potentials in 25 degrees Celsius seawater established before linear polarization and subsequent potentiodynamic scanning. As can be seen, several of the electrodes from panels which exhibited little or no base metal rust exhibit potentials approximating those of the new zinc wire and aluminum wire coated electrodes. Although aluminum is often cited as an active metal in the galvanic series in seawater [5], the development of aluminum oxides by reaction with water imparts certain passive characteristics. As such, more noble potentials can be exhibited by aluminum with time, without accelerated corrosion. For example, the potentials of the 0.08- and 0.15-mm aluminum electrodes, shown in Table 5, are similar to those reported by Shaw and Moran [2] for sprayed aluminum wire coated steel exposed to artificial seawater for 20 days.

Fresh steel exhibited a 1-h ECORR of approximately -0.6 V saturated calomel electrode (SCE). In an aerated environment the potential of steel would also become more noble with time. The relatively noble potentials for the 80% Zn + 20% Al mixed powders and the 0.08-mm zinc coated steel are indicative of a corroded substrate with little sacrificial influence imparted from the zinc or aluminum coating.

Polarization Resistance - Tables 6 and 7 list the polarization resistance Rp values calculated by the Soft CORR program from the linear polarization data. Materials are listed in descending order of Rp. Since resistance is inversely proportional to corrosion current, coatings exhibiting higher Rp values should provide the greatest protection. The actual Rp values obtained reflect some measure of the respective coating and substrate corrosion resistance. This may vary depending on the extent of coating porosity and any substrate area exposed because of atmospheric deterioration.

From Table 6 it can be seen that the electrodes representing the various coated panels exposed for 34 years exhibited Rp values between that of the new aluminum wire sprayed steel and the new zinc wire sprayed steel. The dual layer Zn + Al coated material exhibited the closest Rp value to the new aluminum coating. The top five materials listed in Table 6 were all completely free of base metal rust and all contain aluminum as major or equal constituents with respect to zinc. The next five coatings listed are zinc or consist of 80 to 90% Zn either in mixed powder or alloy powder form. Except for the 90% Zn + 10% Al mixed powder coating, the others exhibited some degree of base metal rust or yellow rust stain.

Of the eleven original materials identified in Table 6, only the 0.08-mm aluminum coating does not appear to fit the limited but nonetheless apparent correlation between Rp value and actual performance in the atmospheric test. This difference may be attributable in part to the microstructural differences such as those associated with the voids between splat particles of the 0.08 mm coating shown in Fig. 13.

The extent to which these data can be, or should be, used to predict the remaining years of protection is certainly intriguing. It is known that the 0.08-mm zinc coated panel exhibited 10% base metal rust within 14 years (0% after 8 years). At the same inspection the mixed powder and alloy powder coatings of 80% Zn + 20% Al were totally free of attack. Twenty years later the same panels exhibited about 1% base metal rust and 5% yellow rust stain versus 10% base metal rust for the mixed powder and alloy powder. Likewise, the 0.15-mm zinc coating exhibited only 2% yellow rust stain, which developed between 14 and 34 years. This would suggest that the original materials that exhibited the highest Rp values after 34-year atmospheric exposure would be expected to provide adequate protection for many years to come. Furthermore, the newer thermal spray aluminum wire coating should be at least equally, if not substantially more, resistant.

Table 7 - Polarization values for new steel and thermal sprayed Al and Zn wire coatings (a)

It must be remembered that these coatings were tested without sealer as is present in common practice. In this regard, thermal spray coated zinc provides good sacrificial base coats beneath conventional paint-type top coatings.

Potentiodynamic Polarization
Figure 14 shows the polarization curves for new steel and the thermal sprayed aluminum and zinc wire coated steel specimens. While some deviation was noted for duplicate tests of the aluminum coating, excellent reproducibility was achieved for the uncoated steel and zinc coating. These curves, developed in natural seawater, are in quite good agreement with those described by Shaw and Moran for artificial seawater tests [2].

Figure 15 compares the curves for the new aluminum coated material and the 0.08-mm Zn + 0.08- mm Al coating. While the 34-year-old exposed material lacks the exact passive characteristic of the new material, the potential-current domains are in close agreement. Polarization characteristics for three materials exhibiting the highest Rp values (Table 6) are compared in Fig. 16.

Curves for the 0.08- and 0.15-mm zinc coated materials from the atmospheric test are compared in Fig. 17. The shift in the curves away from the domain for the new zinc coated material is indicative of the influence on the exposed substrate of the 0.08-mm zinc electrode and the combined influence of some substrate and zinc corrosion product (Fig. 5) for the 0.15-mm electrode.

Figure 18 compares curves for the 0.08- and 0.15-mm aluminum exposed materials with the new sprayed aluminum wire coating. As indicated earlier, ECORR values at test initiation were more noble as compared to the new aluminum coating. Although the Rp values of these materials were quite different, the broad similarities in the anodic, polarization behavior (particularly at high over potentials) are consistent with similarities in physical appearance, that is, 0% yellow rust stain.

Figure 19 shows the polarization behavior of the 10% Zn-90% Al alloy powder coated material. As can be seen, the polarization characteristics of this material after 34-year exposure is between that of the new aluminum wire and zinc wire sprayed material; favoring somewhat that of zinc with respect to active behavior.

Difference in polarization curves for several high zinc mixed powder and alloy powder coatings are shown in Fig. 20. The behavior of the more resistant 80% Zn + 20% Al alloy powders and the 90% Zn + 10% Al mixed powder coatings more closely approximate the new material at low over potentials, but with some limiting current characteristics likely caused by zinc or aluminum corrosion product. In contrast, the behavior of the partially failed 80% Zn-20% Al mixed powder coating more closely approximates the curve shown earlier in Fig. 18 for the failed 0.08-mm zinc coated panel.

Summary
The performance of a number of thermal sprayed mixed powder and alloy powder coatings have been evaluated following 34-years exposure to the moderately severe marine atmospheric conditions in the 250-m lot at Kure Beach, NC. In an effort to further distinguish possible differences between coatings that provided virtually complete protection to the steel substrate, laboratory electrochemical tests were performed for the exposed materials and unexposed sprayed wire coated steel.

Examination of the specimens for detectable base metal rust and rust stain through the coating revealed a broad range in the degree of protection. This varied considerably with alloy content, type of powder, and method of application.

Although a number of zinc coatings and high zinc containing, mixed and alloy powder, coatings performed well, the greatest number of resistant coatings were generally those of high aluminum content. The latter materials typically exhibited less variability as a function of the applicator gun.

Metallographic examination supported earlier indications that the coating thickness was greater than originally cited at the test initiation. It is not known if this difference was due to swelling of the coating or to other explanations. In contrast to present day sprayed wire coating, the remaining coating on the test panels is significantly less dense (that is, between splatted particles).

Electrochemical tests in natural seawater indicated significant differences in polarization resistance and potentiodynamic polarization behavior for the various coatings. Comparisons of the polarization behavior of the exposed materials with protective coating of zinc and aluminum (and various combinations thereof) followed closely the behavior of newly coated materials. In contrast, poor performance coatings with exposed substrate exhibited polarization behavior more consistent with that of steel.

Other investigators are presently studying the protection characteristics of thermal sprayed alloy wire coatings with compositions approximating those which have performed well in 34-year tests.

Acknowledgments
The authors wish to thank B. Shaw for the samples of aluminum and zinc wire sprayed electrodes, R. M. McGowan who conducted the electrochemical tests, and those members of the LaQue Center staff who contributed further to this paper, in particular B. S. Phull.

References

1. Hoar, T. P. and Radovici, O., "Zinc-Aluminum Sprayed Coatings," Proceedings of the 6th International Conference on Electrodeposition and Metal Finishing, London, International Society of Metal Finishing Transactions, Vol. 42, 1964.
2. Shaw, B. A. and Moran, P. J., "Characteristics of the Behavior of Zinc-Aluminum Thermal Spray Coatings," Materials Performance, Vol. 24, Nov. 1985, p. 27.
3. LaQue, F. L., "Marine Corrosion - Causes and Prevention," Environmental Factors, Chapter 4, John Wiley and Sons, New York, 1975, p. 7.
4. Lee, T. S. and Baker, E. A., "Calibration of Atmospheric Corrosion Test Sites," Atmospheric Corrosion of Metals, STP 767, American Society for Testing and Materials, Philadelphia, 1982, pp. 250-285.
5. LaQue, F. L., "Marine Corrosion - Causes and Prevention," Environmental Factors, John Wiley and Sons, New York, p. 179.

This site maintained by International Metalizing Corporation

U.S. Representative:
International Metalizing Corporation
PO Box 23279
Houston, Tx 77228 USA
TOLL FREE 888-INT-METL