Corrosion Performance of Aluminum-containing Zinc Coatings

ISIJ International, Vol. 50 (2010), No. 3, pp. 455–462

1. Introduction

A development program investigating various Zn–Al al-loys for batch galvanizing was initiated several years ago atthe University of Cincinnati, Cincinnati, Ohio, USA in col-laboration with the Weert Groep, a galvanizing companybased in the Netherlands.1) In 2003, at the request of Prof.Wim van Ooij, Teck Metals Ltd. became a member of thedevelopment team and has contributed significantly to thedevelopment program ever since.2) The original objective ofthe development program was to improve the corrosion re-sistance of hot-dipped galvanized coatings, thereby reduc-ing zinc run-off from galvanized structural units such asguard rails, light poles, transmission towers, railings andfasteners. Work at the University of Cincinnati initially fo-cused on an alloy composition of Zn–23%Al. However,subsequent literature surveys and in-house corrosion testscarried out at Teck’s Product Technology Centre (PTC) sug-gested that the Zn–23%Al alloy was not an optimal compo-sition for either corrosion resistance or processability inhot-dip galvanizing, especially with respect to bath temper-ature requirements and coatability. In 2008, systematic cor-rosion tests were carried out at PTC, and the outcome ofthese tests suggested that the corrosion resistance of coat-ings could be further improved with a reduction in the alu-minum content in the coating alloy. In light of these find-ings, work has since been carried out at PTC to develop al-loys with an improved corrosion resistance and betterprocessability in hot-dip galvanizing. A thorough under-standing of the corrosion mechanisms of Zn–Al coatings iscritical to the success of the alloy development program atTeck. As a result, a detailed study of the corrosion ofZn–Al coatings was carried out recently, and the results are

reported here.

2. Experimental

Coating samples were produced in PTC’s laboratory.Hot-rolled low-carbon steel coupons, 150 mm�100 mm�3mm, were degreased in a caustic solution for 1 h, rinsedwith tap water and pickled in an 18% hydrochloric acid so-lution for 40 min, and rinsed with tap water again. Thecoupons were then immersed in a proprietary flux bath for45 s. The fluxed samples were dried for 4 min in a conven-tional oven set at 150°C; the temperature of the samples,however, was only about 100°C. The samples were thenhot-dip galvanized individually in baths containing the de-sirable aluminum levels of 5%, 7%, 9%, 12%, 15% and23%. The bath temperature was individually adjusted to30°C above the liquidus temperature of the contained alloy,and the dipping time was invariably 4 min. The sampleswere withdrawn from the bath at a speed of 35 mm/s usingan automatic dipping device.

In the first test, the 15% Al and 23% Al coatings (here-after referred to as 15Al and 23Al) were compared to othercommercial coatings such as galvanneal (GA), Zn–5%Alcontaining misch metals (GF), and Zn–55%Al–1.5%Si(GL). Coupons with a hot-dip galvanized (HDG) pure zinccoating were also included as the benchmark. A modifiedsalt-spray test3) was used to evaluate the coatings. One bareuncoated steel coupon was tested for validation of the test.

The test is cyclic in nature, with each cycle lasting45 min, including a spray of salty fog (0.01 M NaCl aque-ous solution) for 15 min followed by a 30 min drying periodat a temperature of 35°C. Only one side of the test coupons,100 mm�150 mm in size, faced the corrosive fog. The

Corrosion Performance of Aluminum-containing Zinc Coatings

Nai-Yong TANG and Yihui LIU

Product Technology Centre, Teck Metals Ltd., 2380 Speakman Drive, Mississauga, Ontario, Canada.E-mail: [email protected], [email protected]

(Received on March 9, 2009; accepted on November 25, 2009)

Corrosion performance of zinc coatings containing aluminum in excess of 5% was assessed in modifiedsalt-spray tests. X-ray diffraction studies indicated that the corrosion products consisted mainly of simonkolleite Zn5(OH)8Cl2·H2O, hydrozincite Zn5(OH)6(CO3)2, and zinc aluminum carbonate hydroxide hy-drate Zn6Al2(OH)16CO3·4(H2O). The content of Zn6Al2(OH)16CO3·4(H2O) in the corrosion product increasedas the aluminum content in the coatings increased. SEM-EDX analyses revealed that the microstructuralfeatures formed by the primary aluminum-rich a -phase frequently corroded first and at a faster pace thanthe zinc-rich b -phase in these coatings. The volume fraction and morphology of the zinc-rich b -phase exist-ing in the coatings as degenerated eutectic are the two main factors which determine the corrosion resist-ance of Zn–Al coatings under development. The corrosion resistance of coatings peaked at about 12% Al inthis study.

KEY WORDS: corrosion; hot-dip galvanizing; microstructure; coating; zinc; aluminum.

455 © 2010 ISIJ

other side of the coupon was almost non-corroded at theend of the test. This greatly facilitated the comparison ofthe morphology and microstructure of the corroded and thenon-corroded coatings. The mass of each coupon wasmeasured before the test with a high-precision scale, andthe mass of the corroded coupon was periodically moni-tored after pre-determined numbers of test cycles. Aftereach round of weight measurement, the locations of thetested coupons were randomized using a computer pro-gram. Five galvanized coupons were prepared for eachcoating alloy and tested concurrently. Among them, fourcoupons were used in the destructive measurement of coat-ing mass loss at the end of the four-week corrosion testwhile the remaining coupon was examined using X-ray Dif-fractometry (XRD) and Scanning Electron Microscopy(SEM) coupled with Energy Dispersive Spectrometry(EDS). To measure the mass loss, the corrosion productswere removed initially with a saturated solution of glycineat room temperature, a procedure used by the French Cor-rosion Institute4) in a contract research project prior to thisstudy. It was found in the first corrosion test that the glycinesolution removed the corrosion products effectively only forcoatings containing up to 5% Al. Coatings with an alu-minum content higher than 15% were then pickled again ina 20% chromium trioxide water solution (i.e., chromicacid) at a temperature of 70 to 80°C for one minute per theASTM G1 procedure. The ASTM procedure was used forall coatings in the second corrosion test. The X-ray diffrac-tion patterns for the corrosion layers on the coupon wereobtained using a Rigaku SA-HF3 XRD operated at 50 kVand 40 mA by Cu Ka radiation.

3. Results

It is known that the corrosion rate of bare uncoated steelis about ten to twenty times higher than that of HDG zinccoating in atmospheric exposures. Such a ratio was repro-duced in the current test, suggesting this accelerated labora-tory test can quite realistically rate the corrosion perform-ance of a coating in service.

The results of the first corrosion test are presented graph-ically in Figs. 1(a) and 1(b). Since coating thickness is knownto affect the corrosion performance of coated coupons, theweights of various types of coatings are also superimposedin Fig. 1(b). The results shown in Fig. 1 indicate that the15Al coating performed significantly better than the 23Alcoating and also outperformed all other coatings except theGL coating.

Encouraged by the results of the first test, the 15Al coat-ing was further tested against coatings containing 7% Al,9% Al and 12% Al in an attempt to identify an optimalcoating composition for the best corrosion resistance. Theretest of 15Al coating facilitated direct comparisons be-tween the results of two corrosion tests. The mass gain andmass loss of the coatings in the second test are shown inFig. 2. The mass gain of 12Al coating was clearly the low-est among all coatings tested, as displayed in Fig. 2(a). Themass losses of 12Al and 15Al were similar and were mar-ginally lower than other coatings, as shown in Fig. 2(b).

Metallurgical examinations of the corroded coatings in-dicated that the corrosion was not uniform over the entire

coating surface, presumably due to the non-uniform distri-bution of the salty fog. The corrosion was frequently moreadvanced in some locations, as shown in Fig. 3. However,following a thorough examination of all corroded coatings,the corrosion mode of these Zn–Al coatings was found tobe uniform in general, i.e., the corrosion front advancedmore or less uniformly along the coating thickness althoughthe microstructural features formed by the aluminum-richa-phase (fcc) frequently corroded first and slightly fasterthan the zinc-rich b-phase (hcp) in these coatings. The mi-crostructures of the non-corroded and corroded 15Al coat-ing are shown in Figs. 4(a) and 4(b); the corresponding mi-crostructures for the 12Al coating are shown in Figs. 5(a)and 5(b).

The compositions of the corroded coatings were ana-lyzed using the SEM-EDS technique. A composition mapfor the 15Al coating is shown in Fig. 6.

X-ray diffraction studies of the corrosion products werecarried out on coatings containing 5% Al, 15% Al, and23% Al. A corroded hot-dip galvanized coating containingonly a trace amount of aluminum was also subjected to theX-ray diffraction study for comparison. Analyses of the pat-terns indicated that the corrosion product consisted ofmainly simonkolleite Zn5(OH)8Cl2· (H2O), hydrozinciteZn5(OH)6(CO3)2, and zinc aluminum carbonate hydroxide

ISIJ International, Vol. 50 (2010), No. 3

456© 2010 ISIJ

Fig. 1(a). Standardized mass gains of alloy coatings in g/m2 as afunction of test time in hours. The 15Al alloy coatingoutperformed all coatings, except that of the Zn–55Al–1.5Si alloy, tested concurrently. Only the fogspraying time is counted in the plot.

Fig. 1(b). Standardized mass losses of various alloy coatings ing/m2 after the four-week test with a total fog spraytime of 221 h. The 15Al coating noticeably outper-formed the 23Al coating in the test.

hydrate Zn6Al2(OH)16CO3·4(H2O). The X-ray patterns gen-erated on these coatings evolved systematically. The por-tion of the Zn6Al2(OH)16CO3·4(H2O) increased with an increase in aluminum content in the coatings. It was thedominant corrosion product on the 15Al and 23Al coatings.Note that Zn6Al2(OH)16CO3·4(H2O) is isomorphic to Zn0.56

Al0.44(OH)2(CO3)0.22·x(H2O). The latter, with higher alu-minum content, has a slightly smaller cell size and, hence,

slightly larger 2q angles for corresponding peaks. It wasfound that the Zn6Al2(OH)16CO3·4(H2O) peaks for 23Alshifted slightly to the right, i.e., larger 2q angles when com-pared to the corresponding peaks for 15Al. Apparently, theample supply of aluminum from the coating favored theformation of a corrosion product with higher aluminumcontent. Shown in Figs. 7(a) to 7(d) are X-ray diffractionpatterns generated from corroded HDG, 5Al, 12Al and


457 © 2010 ISIJ

Fig. 2(a). Standardized mass gains of alloy coatings in g/m2 as afunction of test time in hours. Only the fog sprayingtime is recorded here.

Fig. 2(b). Standardized mass losses of various alloy coatings ing/m2 after the four-week test with a total fog spraytime of 221 h. The 12Al and 15Al alloys performedbest in the test.

Fig. 3. Cross-sectional views of the 12Al coating after the corrosion test. Corrosion developed into a more advancedstage in the location shown in the left micrograph than that shown on the right. Some pockets of the primary a-phase were totally corroded.

Fig. 4(a). Grain structure of the 15Al coating on the non-cor-roded side of the coupon. The primary a-phase hasdecomposed through the eutectoid reaction.

Fig. 4(b). Microstructure of the corroded 15Al coating. The de-composed aluminum-rich primary a-phase corrodedfaster than the degenerated pure zinc b-phase.

Fig. 5(a). Non-corroded grain structure of the 12Al coating. Fig. 5(b). Microstructure of the corroded 12Al coating. Zinc indegenerated eutectic is much more corrosion resistant.

23Al coatings, respectively. Peak positions for all relevantcorrosion compounds were calculated using their crystallo-graphic data, and these calculated positions are identifiedand marked in the patterns. It was found that the actual positions of all major peaks situated within 0.2° of theircalculated 2q values, except for Zn6Al2(OH)16CO3·4(H2O)

peaks for 23Al due to the isomorphic structure ofZn0.56Al0.44(OH)2(CO3)0.22·x(H2O), as discussed earlier.

Following the positive identification of the corrosionproducts, further analyses of the mass losses and gains ofall coatings were carried out. The data are tabulated inTable 1. The ratio of the gain to the loss was calculated. On


458© 2010 ISIJ

Fig. 6. Elemental mapping for a corroded 15Al coating. The corrosion product covering the coating surface is enrichedin zinc and chlorine but was depleted in aluminum. The darker corroded areas within the primary a-phase pock-ets in the normal image of the coating are associated with strong aluminum and chlorine signals and weak zincsignals. The chlorine was mostly enriched along the corrosion front.

Fig. 7. X-ray diffraction patterns generated from (a) HDG coating containing 0.003% Al (b) 5Al, (c) 15Al and (d) 23Al.The corrosion product on the 23Al coating was mainly the zinc aluminum carbonate hydroxide hydrate.

perusing the data, a trend is self-evident: the ratio increaseswith increasing aluminum content in the coating. The massgain is largely attributable to the functional groups and thecrystallized water in the corrosion products whereas themass loss is directly related to the metallic components inthe corrosion products. The corresponding stoichiometricmass gain/loss ratios for various corrosion products foundon HDG, 5Al, 15Al and 23Al coatings in the first corrosiontest are listed in Table 2. It is obvious that the aluminum-bearing corrosion products have higher gain/loss ratiosmainly because the atomic weight of aluminum is muchless than that of zinc. The higher the aluminum content inthe corrosion product, the higher the gain/loss ratio be-comes. The gain/loss ratios of the coatings listed in Table 1are closely related to the ratios for corrosion products listedin Table 2. The mass gain/loss ratio for GL in Table 1 wasskewed to a large value due to a very low mass loss and anadditional mass gain by the pickup of NaCl. The analysis ofthe change in the mass gain/loss ratio with coating alu-minum provided supplemental evidence, in addition to theXRD results, of the evolution in corrosion products with increasing coating aluminum content.

4. Discussion

Aluminum additions to a galvanizing bath can enhancecoating appearance and reduce zinc consumption by reduc-ing the oxidation of the bath metal. Added in sufficientamounts, aluminum also affords the control of silicon reac-tivity5) and can greatly reduce the rate of dross generationin galvanizing or totally change the nature of the dross par-ticles from Zn–Fe compounds to iron aluminides. The for-mer, being heavier than the molten zinc, sink to the bottomof the kettle while the latter is lighter than the molten Zn–Al alloy and floats to the top of the bath, thereby facilitatingits removal from the bath, provided that the aluminum con-tent of the alloy is below 30%.

The addition of aluminum to zinc-based alloys is knownto improve the corrosion resistance of a coating. For exam-

ple, coatings of Zn–5%Al alloy containing misch metalshave been shown to outlast HDG pure zinc coatings in vari-ous tests or services by a factor of about two6); a coating ofZn–55%Al–1.5%Si alloy has been proven to possess thebest corrosion resistance among all zinc-based coatings.However, the corrosion resistance of a coating does not im-prove monotonically with the aluminum content of thecoating in all environments. Atmospheric exposure of coat-ings containing aluminum levels from 0 to 70% suggestedthat only in severe marine environments does the aluminumaddition improve the corrosion resistance of the coatingbased on the time elapsed to the first appearance of the redrust.7) In this comparative study, the corrosion resistance ofthe tested coatings peaked at 4 to 7% Al.7,8) With a furtherincrease in coating aluminum content, the corrosion resist-ance declined and a large increase in the coating durabilityoccurred only when the coating aluminum content ex-ceeded 40%.7) Test results published by Humayun,8) inwhich the corrosion resistance is measured by weight loss,support this observation. On the other hand, data generatedin other long-term atmospheric exposures and in laboratoryaccelerated corrosion tests failed to substantiate these ob-servations. These tests suggested that the aluminum addi-tion improved the coating corrosion resistance in general,although the effect peaked at 4 to 7% Al.9) In this study, thecorrosion performance of Zn–Al coatings peaked at 12%Al instead. Such a discrepancy is believed to stem from anumber of factors: first, the coating sample productionmethods were quite different. Samples used in Townsendand Borzillo’s7) tests were produced on a pilot-scale contin-uous galvanizing line and coating thickness was controlledat 20 to 25 mm using coating rolls. Such a wiping actioncould significantly alter the coating microstructure, therebyaffecting corrosion performance as well. On the other hand,the thickness of the coating samples in this study was natu-rally achieved. No effort was exercised to achieve a certaincoating thickness. Secondly, the types of corrosion testswere different. Townsend and Borzillo conducted a long-term atmospheric exposure test whereas the present authorsconducted an accelerated laboratory test. The rating crite-rion of corrosion performance was also different. Townsendand Borzillo judged the coating performance by time periodwhen the first red rust appeared. Consequently, coatingthickness uniformity could significantly affect the test out-come. In the present study, the coating performance wasjudged by weight loss. Unfortunately, no analysis of coatingmicrostructure was carried out by those authors, and thecorrosion mechanism was not explored. As such, a mean-ingful discussion of the discrepancy would be difficult.

The corrosion performance of coatings could be affectedby numerous factors, such as galvanizing parameters, coat-ing surface treatment prior to test, sample handling, samplearrangements at the test site or in the environmental cham-ber, etc. As a result, a clear understanding of the corrosionmechanism is critical for the interpretation of these seem-ingly conflicting observations. Regrettably, a brief literaturesurvey carried out by the present authors revealed that thecorrosion data were frequently reported without any explo-ration of the related corrosion mechanisms.7,8) On the otherhand, detailed studies of corrosion mechanisms are fre-quently focused on one (5%) or two (5% and 55%) alu-


459 © 2010 ISIJ

Table 1. The mass gain/loss of various coatings after the firstcorrosion test.

Table 2. The related mass gain/loss ratio for various corrosionproducts.

minum compositions only,10) and the researchers believedthat the corrosion mechanisms discovered in the studieswere universally applicable to all aluminum-containingcoatings.

It is generally believed that the addition of aluminum tozinc will invariably improve the corrosion resistance of zinccoatings because pure aluminum possesses an extraordinarycorrosion resistance. It is suggested that when Zn–Al coat-ings are subjected to a corrosive environment, the zinc pref-erentially dissolves, leaving an aluminum-rich porous struc-ture surviving in the coatings.11) Because aluminum in thisporous structure could be passivated and, hence, would becathodic with respect to the steel substrate, red rust wouldoccur. Indeed aluminum owes its excellent corrosion resist-ance to its ability to form an oxide film that is bondedstrongly to its surface and that, if damaged, reforms imme-diately in most environments. The oxide film that developsin normal atmosphere is much thicker than 1 nm and is com-posed of two sublayers. The inner sublayer is a compactamorphous barrier layer.12) Such a capability is preserved inZn–55%Al–1.5%Si coatings, but not in zinc coatings con-taining aluminum less than 40%. In these coatings, the alu-minum-rich a-phase could contain over 70% Zn. As such,it does not have the capability of forming the barrier oxidefilm when it coexists with the zinc-rich b-phase in corro-sive environments. Also, when the coating solidifies, thealuminum-rich a-phase forms a primary dendritic skeletonwithin the coating in the hypereutectic composition region(�5% Al). The a-phase is further decomposed through theeutectoid reaction into finely dispersed zinc-rich and alu-minum-rich zones at ambient temperatures due to the highdiffusivity of zinc at such temperatures. Regardless ofwhich zones corroded first, a porous aluminum-rich struc-ture would not likely exist. Indeed, experimental work andcase studies carried out at the authors’ laboratory failed tolocate such a porous structure in corroded Zn–Al coatings.

The standard electrode potential of pure aluminum ismuch more negative than that of pure zinc. As a result,when the two metals are electrochemically coupled, the alu-minum will be the anode if it is not covered by a protectiveoxide film. In the present case, aluminum existed in thecoatings as the solvent of the solid solution a-phase. At anelevated temperature, the a-phase could contain zinc in ex-cess of 80% and yet maintain its face-centered cubic crys-tallographic structure. When fully decomposed through theeutectoid reaction, it could still contain over 1% Zn. Theelectrolytic solution potential of such a decomposed a-phase is about �1.06 V using a 0.1 N calomel electrode,13)

only slightly more anodic than that for zinc at �1.01 V. Itis, therefore, expected than the a-phase will corrode firstand slightly faster than the b-Zn in decomposed primary a-phase pockets.

Due to the fast solidification process in coatings contain-ing aluminum in excess of 5%, a dendritic frame of alu-minum-rich a-phase coexists with a degenerated eutecticfrequently consisting of b-Zn only. The b-Zn envelops thedendritic a-phase which appears as pockets in a cross-sec-tion, as shown in Fig. 4. The volume fraction of the primarya-phase in the coating is a function of the coating alu-minum content and the cooling rate to a lesser degree. If acoating containing aluminum over 18% is cooled down at

an extremely low rate and obtains an equilibrium microstruc-ture, the coating will consist entirely of the primary a-phase. However, commercial coatings can never obtain anequilibrium microstructure due to the relatively high cool-ing rate achieved in a production environment. Calculationsbased on Scheil’s14) assumption suggest that there will stillbe over 10% (molar fraction) of eutectic in a coating con-taining 50% Al, as shown in Fig. 8. Due to the relativelylow melting point of zinc, its diffusivity is significant whena coating solidifies. In a Zn–Al coating, the eutectic is always degenerated whereby the residual aluminum in thesolidifying liquid attaches to the exiting primary a dendritearms and the zinc forms separate grains nearby. These zincgrains appear to possess a much better corrosion resistancethan the aluminum-rich a pockets. No fundamental studyon this phenomenon is known to the present authors. How-ever, it is well known that the corrosion performance ofzinc can be significantly affected by the iron content in thezinc.15,16) Teel and Anderson16) found that a zinc anodefunctioned well only when the iron content was maintainedat a very low level in the high purity zinc anodes. They rec-ommended that the iron content be kept below 0.0015% inotherwise pure zinc. It is also known that the iron content inzinc powder for alkaline cells is kept at below 5 ppm be-cause iron possesses a low hydrogen overvoltage.17) Theiron solubility in liquid zinc is about 0.03% at the galvaniz-ing temperature of 450°C. Excess iron will be largely en-trapped in the coating as it solidifies and will subsequentlyprecipitate out as tiny particles. These particles will act ascathodes in corrosion galvanic cells, thereby adversely affecting the corrosion resistance of the coating. In a zincbath containing a high aluminum level, the iron dissolved inthe liquid alloy is almost completely scavenged by alu-minum through the formation of iron aluminides duringcoating solidification. When the aluminum level in a coat-ing alloy exceeds 4%, the iron solubility in the molten alloyis below 1 ppm. Iron entrapped in the degenerated eutecticb-Zn in coatings is even lower in concentration. This couldbe a contributing factor to the better corrosion resistance ofb-Zn. To verify this suggestion, the microstructure of a cor-


460© 2010 ISIJ

Fig. 8. Volume fractions of the primary a-phase in a coating as afunction of the aluminum content of the coating alloy.Solid dots are associated with an equilibrium solidifica-tion process whereas open circles are associated withScheil’s assumption of no solid-state diffusion during afast solidification process.14)

roded Zn–5%Al coating sample was examined: a represen-tative micrograph of the corroded grain structure is shownin Fig. 9. It can be seen that the primary b-Zn grains pos-sessed better corrosion resistance than the eutectic colonies.This finding supports the suggestion that the corrosion resistance of zinc was much improved when the iron solu-bility of the alloy dropped to below 1 ppm. Consequently,the coating corrosion performance will generally improvewith an increase in the aluminum content in the coatinguntil the volume fraction of the degenerated eutectic b-Znin the coating is reduced to an insignificant amount.

The existence of a critical volume fraction of the degen-erated eutectic in Zn–Al alloy coatings is likely an impor-tant factor affecting its corrosion performance because themorphology of the b-Zn could be a contributing factor tothe corrosion resistance of the coatings. If the dendritic aframe in the coating is totally enveloped by the degeneratedeutectic b-Zn, the corrosion resistance of the coating can befurther improved. This metallurgical variable is probablyone of the main reasons why the corrosion resistance ofcoatings peaked at about 12% Al. The microstructure of acoating can be easily affected by a number of factors, andcorrosion performance of a coating can also be influencedby numerous factors. As a result, the aluminum content ofthe coating which exhibited the best corrosion resistance ina test could vary around 12% Al. In this test, the 12Al coat-ing performed the best. However, the difference in corro-sion performance of 9Al, 12Al and 15Al coatings was mar-ginal.

Corrosion products change with the aluminum content inthe coatings. This signals an evolution of the corrosionmechanism. The corrosion products consist mainly of thesimonkolleite Zn5(OH)8Cl2·H2O, the hydrozincite Zn5

(OH)6(CO3)2, and the zinc aluminum carbonate hydroxidehydrate Zn6Al2(OH)16CO3·4(H2O). The Zn6Al2(OH)16CO3·4(H2O) content of the corrosion product increased with anincrease in the aluminum content in the coatings; it becamethe dominant corrosion product for the 23Al coating. Thisfinding is consistent with the observations of Nishihara et al.10) By means of photoemission spectroscopy usingsynchrotron radiation and m-FT-IR spectroscopy, these researchers elegantly demonstrated that aluminum in zinccoatings retards the formation of hydrozincite. They alsofound that the nature of the corrosion products forming inan early stage of the corrosion test are similar on Zn–5%Al

and Zn–55%Al coatings. Based on the above observations,these researchers attributed the improvement in corrosionresistance of these two types of coatings to the fact that thecorrosion products on the coatings are insulators which restrained the reduction of oxygen, thereby retarding theanodic dissolution. However, the corrosion performances ofZn–5%Al and Zn–55%Al coatings were quite different inindustrial applications and in laboratory tests. Furthermore,the dominance of Zn6Al2(OH)16CO3·4(H2O) in the corro-sion product accompanied a decline in the coating corro-sion resistance in the present study. As such, the improve-ment in corrosion performance brought about by the addi-tion of aluminum through the above mechanism is probablylimited. If this mechanism dominates, corrosion perform-ance of the Zn–Al coating would monotonically improvewith an increase in the coating aluminum content. However,the results obtained in the present study indicate that thecorrosion resistance of Zn–Al coatings decreased when thealuminum content reached 23%. Atmospheric exposure ofaluminum-containing zinc coatings also suggests the exis-tence of a minimum in corrosion resistance at an aluminumcontent of about 35%.7) As mentioned earlier, coatings con-taining aluminum in excess of 45% regain the capability offorming a protective alumina film on their surfaces, therebyushering in a new corrosion protection mechanism for thistype of coating.

5. Conclusions

The corrosion performance of zinc coatings containingaluminum up to 23% was assessed in modified salt-spraytests. The highest corrosion resistance was exhibited bycoatings containing about 12% Al. The corrosion productsconsisted mainly of simonkolleite Zn5(OH)8Cl2·H2O, hy-drozincite Zn5(OH)6(CO3)2, and zinc aluminum carbonatehydroxide hydrate Zn6Al2(OH)16CO3·4(H2O). The Zn6Al2

(OH)16CO3·4(H2O) content of the corrosion product in-creased with an increase in aluminum content in the coat-ings. It is the dominant corrosion product for the 23Al coat-ing. SEM-EDX analyses revealed that the microstructuralfeatures formed by the primary aluminum-rich a-phase fre-quently corroded first and faster than the zinc-rich b-phasein these coatings. However, the corrosion mode of thesecoatings is generally uniform corrosion, i.e., the corrosionfront advanced more or less uniformly along the coatingthickness. It appears that the corrosion resistance of thezinc component in the coatings increases with an increasein the aluminum content of the coatings. The b-phase of thedegenerated eutectic possesses a much better corrosion re-sistance than other microstructural features in the coatings.A microstructure consisting of the primary a dendrite en-veloped by the b-phase is better than an all eutectic mi-crostructure, making the corrosion resistance of 7Al, 9Al,12Al and 15Al better than that of 5Al.

In summary, the volume fraction and the morphology ofthe degenerated eutectic b-phase are the two main contribu-tors to the existence of a maximum corrosion resistance forzinc coatings containing approximately 12Al.

Acknowledgements

The corrosion tests, including the procurement and


461 © 2010 ISIJ

Fig. 9. Cross-sectional view of the corroded grain structure of aZn–5%Al with misch metals sample. The primary b-Zngrains (white ovals) evidently possessed better corrosionresistance than the eutectic colonies.

preparation of the coating samples, data collection and met-allurgical section preparation were carried out by SteveMurray, Fred Coady, Neil Gao and Daniel Jochim. The au-thors would like to express their thanks to Fred Coady forconstructive discussions of corrosion mechanisms. The au-thors also thank Teck for allowing the publication of the testresults and acknowledge the assistance of Angeline M.Prskalo in the manuscript preparation.

REFERENCES

1) M. Ranjan, E. Klerks and W. J. van Ooij: Proc. of Intergalva’03, Am-sterdam, Session: Sustaining Product Development (3), EuropeanGeneral Galvanizers Association, Surrey, United Kingdom, (2003).

2) W. J. van Ooij, M. Ranjan, A. Joshi, J. Verstappen, J. Zervoudis andG. W. Poag: Proc. of Intergalva’06, Naples, Italy, Session: The Met-allurgy of Galvanizing (2). European General Galvanizers Associa-tion, Surrey, United Kingdom, (2006).

3) X. G. Zhang, J. Hwang and W. K. Wu: Proc. of Galvatech’98, Ed. byN. Masuko, ISIJ, Tokyo, Japan, (1998), 410.

4) N. Larche and T. Prosek: Contract Research Report IC 89465,French Corrosion Institute, 220 rue Pierre Rivoalon, 29200 Brest,

France, (2008), 1.5) N.-Y. Tang: J. Phase Equilib. Diffus., 29 (2008), 333.6) J. L. Hostetler: Corrosion Resistance of Unpainted Zn–5%Al with

Misch Metals Steel, ILZRO Publication, Research Triangle Park,NC, USA, (1992).

7) H. E. Townsend and A. R. Borzillo: Proc. of Galvatech’95, The Useand Manufacture of Zinc and Zinc Alloy Coated Sheel Steel Productinto the 21st Century, AISI, Chicago, IL, USA, (1995), 171.

8) A. Humayun: Proc. of National Conf. on Coated Steel Technol., Ap-plications & Markets, New Delhi, India, (1994), 22.

9) A. Hannick: Proc. of Intergalva’79, Paris, France, (1979), 42.10) K. Nishihara et al.: Proc. of Galvatech’07, ISIJ, Tokyo, Japan,

(2007), 665.11) ASM Handbook, Vol. 13B, Corrosion: Materials, (2005), 406.12) ASM Handbook, Vol. 13B, Corrosion: Materials, (2005), 95.13) R. H. Brown, W. L. Fink and M. S. Hunter: Trans. AIME, 143

(1941), 115.14) E. Scheil: Z. Metallkd., 34 (1942), 70.15) F. C. Porter: Corrosion Resistance of Zinc and Zinc Alloys, Marcel

Dekker Inc., New York, (1994), 334.16) R. B. Teel and D. B. Anderson: Corrosion, 12 (1956), 343.17) F. C. Porter: Zinc Handbook, Marcel Dekker Inc., New York, (1991),

418.


462© 2010 ISIJ

Documents

Corrosion Performance of Aluminum-containing Zinc Coatings