Corrosion Science. 1969, Vol. 9. pp. 179 to 183. Pergamon Press.

Printed in Great Britain

THE CORROSION RATES OF STRUCTURAL METALS IN

SEA-WATER, FRESH WATER AND TROPICAL
ATMOSPHERES*
SUMMARY OF A SIXTEEN-YEAR EXPOSURE STUDY

CHARLES R. SOUTHWELL
Naval Research Laboratory, Washington, D.C., 20390

METALS exposed to natural environments gradually build up surface films. These

accumulating films markedly change the rate of corrosion of the basematerial. Each
alloy forms different corrosion products and has distinct corrosion-time functions in
each specific environment. Normally this function is curvilinear during the first few
years, but after sufficient film build-up equilibrium conditions are approached and a
practically constant stabilized corrosion rate is [Link] rates determined
during the changing curvilinear period have limited significance. Only the final steady
state rate or one approaching it can provide an effective value for [Link] of
different metals and environments or for extrapolating service life to extended periods
of exposure.
Generally, stabilized corrosion rates can be determined only by measuring
corrosion lossesof an adequate number of continuously exposed samplesat suitably
spacedintervals over a long span of time. Except for certain atmospheric studies, little
quantitative data of this type have been collected for natural environments. For
example, even with the present high and increasing interest in the effects of undersea
environments on structural materials, some of the most common structural metals
have not been effectively studied in sea-water.
A long-term exposure study of metal corrosion in the Panama Canal Zone has
now beencompleted. Results of this study are providing considerable new information
from which stabilized corrosion rates in underwater and marine atmospheric environ-
ments can be established. This 16y programme was initiated by the Panama Canal
Company and the U.S. Army in the late 1940’sand was continued to completion by
the U.S. Naval Research Laboratory, which took charge of the project in 1952.
Exposures were provided for 52 structural metalsand alloys in five uncontaminated
natural environments. Bi-metallic couples as well as single plates were included. Each
sample was exposed to Pacific sea-water, both in continuous immersion and at mean
tide elevation; Gatun Lake fresh water; and marine and inland tropical atmospheres.
A sufficient number of samplesof each metal was provided so that duplicate specimens
could be removed from each environment at intervals of 1, 2, 4, 8 and 16~; a total of
13,000 test pieceswere exposed. After removal and chemical cleaning of specimens,
corrosion effects were determined by measuring weight lossand depth of pitting, after
*Manuscript received 11 July 1968.
179
180 &ARLES R. SOUTHWELL

which loss in tensile strength was determined on representative strips cut from the
exposed plates.
Nine reports presenting results of this study have now been prepared by the U.S.
Naval Research Laboratory: Part 1 (1) details the background and procedures of the
investigation and gives results through 8y for six unalloyed metals. Parts 2, 3, 4 and
5 (2-5) are interim reports concerned with various wrought ferrous metals. Part 6-
Aluminium and Magnesium,s Part 7-Copper and Copper Alloys,’ Part 8-Nickel
and Nickel-Copper Alloy~,~ and Part 9-Sea and Fresh Water Corrosion of Wrought
and Cast Ferrous Metals,s are final reports for the complete 16y exposure.
In the few instances where comparable exposure data are available from temperate
latitude environments, comparison with tropical corrosion is being included. These
comparisons indicate that tropical water (continuous immersion) and atmospheric
conditions generally produce time-corrosion relations of similar pattern but of slightly
greater magnitude than temperature environments. The tropical data should prove to
be most useful, as they provide approximate upper limits of corrosion in uncontami-
nated natural environments. As such, they should, in addition to their general academic
interest, have considerable practical significance for corrosion engineers and designers
of undersea installations.
A brief summary of the results of the four final reports follows :

Al and Mg
Part 6 presents completed long-term corrosion data for three alloys of Al and
two of [Link]. 1100 (99O/bAl) and alloy 6061T (0.8Mg, 0*6Si) were studied in both
sea-water and fresh water. Each alloy was extremely resistant to corrosion in the sea-
water environments, with the only damage to the panels being mild etching around
attached fouling [Link] 16y exposure the average of the 20 deepestpoints of
penetration in theseetcheswas lessthan 18mil for both alloys. Weight lossdetermina-
tions showed that general corrosion rates in sea-waterdecreasedwith time. Between 8
and 16y practically stabilized rates of lessthan 0*04mil/y were obtained. In the warm
fresh water of Gatun Lake, the two Al alloys were both deeply pitted. After 16y,
average depths of the 20 deepestpits were 96 and 98mil. In the marine and inland
tropical atmospheresthese sametwo alloys and a 2024 (4*2%Cu, 1.7Mg, 0.6Si) alloy
were all extremely resistant and about equal. No measurablepits were found on any
specimen,and general corrosion rates in the atmosphereswere very low and constant
throughout the exposure period, averaging lessthan O*Olmil/y.
Mg alloy AZ31X (3*OAl, 1.5Zn) was exposed to three underwater environments.
As expected, this metal was rapidly attacked in sea-water, with perforations of the
ain. panels occurring during the first 6 months’ exposure. At mean tide stone-boring
molluscs (Lithophaga) entered the magnesiumpanels. The borers developed slowly,
but after 2y a few extended transversely through the in. thick plates. By 8y there were
a large number of perforations occupied by these mytilids.
In the tropical atmospheresMg alloy AZ6lX (7%Al) was also tested and showed
slightly better resistance than AZ3lX. Mg in both inland and marine atmospheres
corroded at a constant rate throughout the exposure period. In the marine atmosphere
the rates were 0.7 and O*Bmil/y, respectively, for the 61X and 31X alloys.
 The corrosion rates of structural metals 181

Cu and Cu alloJu
Corrosion of Cu and nine wrought Cu alloys during 16y exposure to all five
environments is presented in Part 7. The alloys included three high Cu bronzes
(Si-, Al-, and P-); three alpha-phase Cu-Zn alloys (9&10, 80-20, and 70-30); and
three high-Zn brasses(Mn-, Naval-, and Muntz f 0*2°%,As). Weight-loss curves
for the Cu alloys showed considerable variability for the several alloy-environment
combinations; some exhibited mild downward curvature throughout the exposure
period, while others showed marked early rate decreasewith relatively rapid estab-
ment of a linear relation. Cu and the high Cu alloys were resistant to all environ-
ments, but two-phase brasses,including the As-fortified Muntz metal, were heavily
dezincified in marine exposures. For Muntz and Naval brassestensile tests revealed
the occurrence of moderate dezincification in fresh water also. Of the various alloys
studied, 5%Al bronze showed the highest overall corrosion resistance; its 16y loss
in sea-water was only l/5 that of Cu and the other two bronzes.
As corrosion product films increasedwith exposure, fouling resistanceof all copper
alloys [Link] 16y all were moderately to heavily fouled, including plates of
pure Cu.
Results for a few galvanic couples containing Cu alloys are given in the report
and show that the corrosion of brassand bronze are appreciably acceleratedby contact
with stainlesssteel. On the other hand, C-steel anodes effectively protected bronzes
and high-Zn brassesover long periods of immersion in sea-water.

Ni and Ni-Cu alloys

Part 8 covers the corrosion of five Ni-Cu metals through 16y exposure in the five
environments. Additional data are presented which compare four other non-ferrous
metals with Ni-Cu and Cu-Ni alloys in sea-water.
The weight loss vs. time curves for the Ni-base metals were normally curvilinear
relations, but considerable variation in direction and magnitude was observed for the
different metals and environments. No single Ni alloy exhibited superior resistance
in all environments. For example, pure Ni had the lowest lossesin fresh water and in
the atmosphere, but was the most corroded in tropical sea-water, where it suffered
severe localized pitting. “Monel” (67Ni, 3OCu, 2Fe) was less resistant in sea-water
than expected; it was subject to a rapid high-incidence pitting attack. However, in
contrast with weight loss, which continued to increase with time, deepening of pits in
“Monel” almost ceasedafter the first 2-3~. The terminal pit depths were about 50mils.
Comparison of Ni-Cu alloys with a few of the more sea-resistantmaterials from
other metal groups showed the resistance to quiet sea-water to be highest for 5%Al
bronze, closely followed in order of resistanceby 6061T Al and 70-3CCu-Ni, then by
Pb, “Monel” and Zn.
A few results for galvanic corrosion of Ni alloys which are presented in the report
show the long-term effectiveness of C-steel anodes for protecting Ni-Cu metals in
sea-water. Additional galvanic data presentevidence that considerableanodic corrosion
will be induced in normally sea-water-resistantmetals if theseare coupled with certain
Ni alloys in an underseaenvironment.
The Ni alloys were extremely resistant in the marine and inland tropical atmo-
spheres. The slight differences observed between the various metals indicated that
182 CHARLES R. SOUTHWELL

resistance increased with Ni content. Pure Ni was one of the most resistant of the 52
metals exposed to the tropical atmosphere; its stabilized rate was < [Link]/y.

Underwater corrosion of cast and wrought structural ferrous metals

Sea-water and freshwater exposures of structural steels and wrought and cast irons
are treated in Part 9. Included are fourteen different ferrous alloys consisting of two
mild C-steels (0.08% and O*35o/0Cu), two Ni steels (2% and 5%Ni), two Cr-steels
(3% and 5%Cr), four proprietary high-tensile low-alloy steels, an Aston process
wrought Fe, a grey cast Fe, a cast steel and an austenitic cast Fe (18 %Ni). Five of the
metals were exposed with more than one surface condition, i.e. pickled, mill scale, and
machined. Discussion is focused on mild C-steel in the three environments, and alloy
steels and cast metals in sea-water. However, complete tabulated data are presented
for all alloy-water combinations.
Freshwater corrosion of mild C-steel exhibited an almost parabolic relation with
time. The general corrosion rate (weight loss) changed from 7*4mil/y average for the
first year to a practically stabilized rate of 0*7mil/y for the interval from 8 to 16~. In
sea-water the same steel showed considerable rate change during 1 and 2y exposure
following which a linear relation was approached. The stabilized rate for general
corrosion under sea-water was approximately 2*7mil/y. At mean tide the corrosion
rate stabilized only after 4y exposure, but the final constant rate was again about
2*7mil/y.
Pitting penetration of mild steel for these three environments showed much higher
earlier penetration, 2.5-5 times deeper than general corrosion, but by 8y pits on all
were deepening at about the same rate as general corrosion. Sea-water pitting changed
from 25-35mil/y for the first two years to a final stabilized rate of approximately
3mil/y.
Ni-alloy steels (2% and 5%Ni) in sea-water were practically equal to mild steel
on the basis of weight loss, but depth of pitting was considerably deep:r in the Ni
alloys. Comparing averages of the 20 deepest pits at 16y, the difference was 115 and
139mils for 5% and 2%Ni-steel vs. 90mil for C-steel.
Cr-steels (3 y. and 5 %Cr) in sea-water were unexpectedly inferior to C-steel on the
basis of metal lost over long periods of exposure. Their corrosion rates stabilized at
3.5 and 3*8mil/y vs. 2.7 for C-steel. For 1 and 2y the weight losses were appreciably
lower for both Cr-steels, but by 4y the advantage had reversed and by 16y 3% and
5%Cr-steels had lost 45 and 22 per cent more metal respectively than C-steel.
Four high-strength low-alloy steels developed corrosion losses in sea-water equal
to or greater than plain C-steel. Maximum corrosion rates for these metals were found
for the two containing small percentages of Cr. Their stabilized rates were 4-8 and
5*0mil/y vs. 2.7mil/y for C-steel. Pitting also was more severe in three of the low-alloy
metals, although one containing Cu, Ni, Mn and MO had a lower pitting rate than
mild steel.
Aston process wrought Fe developed a slight weight-loss advantage over mild C-
steel in sea-water. Comparative stabilized rates were 2.4 vs. 2*7mil/[Link],
however, was about equal for the two.
Grey cast Fe had the highest corrosion rate of any of the materials in sea-water.
 The corrosion rates of structural metals 183

Its stabilized rate was 58mil/y which was more than double that of mild steel and over
five times the l*lmil/y steady-state rate which developed for the 180/Ni cast Fe.
The final long-term results for the 52 metals have repeatedly emphasized the value
of conducting such lengthy tests. Analysis of the completed data reveals that many
miscalculations and gross errors in corrosion rates would have resulted if evaluations
had been based solely on the early periods of exposure. Additional final reports of the
completed 16y data will be published in the near future. These will concern Structural
Ferrous Metals in Marine and Inland Tropical Atmospheres, Non-ferrous Metals
and Metallic Coatings, Galvanic Corrosion in Sea-water, and Bi-metallic Couples in
Tropical Atmospheres.
Copies of final reports can be obtained from the Naval Research Laboratory,
Washington, D.C., 20390 (Attn. author) or from the U.S. Department of Commerce,
Clearinghouse, Springfield, Virginia, 22151.

REFERENCES
1. A. L. ALEXANDER, B. W. FORGESON, H. W. M~NDT, C. R. SOUTHWELL and L. J. THOMPSON,
Corrosion of Metals in Tropical Environnrenrs, Parr I-Test Methods Used and Results Obtained
for Pure Metals and a Sfrucfural Steel. NRL Report4929, June 1957; Corrosion 14, 73t (1958).
2. C. R. SOUTHWELL, B. W. FORGESON and A. L. ALEXANDER, Corrosion of Metals in Tropical
Environments, Part 2-Atmospheric Corrosion of Ten Structural Steels. NRL Report 5002, Dec.
19.57; Corrosion 14, 435t (1958).
3. B. W. FORGESON, C. R. SOUTHWELL and A. L. ALEXANDER, Corrosiorz of Metals in Tropical
Environmenfs, Part 3-Underwater Corrosion of Ten Slrucltrral Steels. NRL Report 5153, Aug.
1958; Corrosion 16, 105t (1960).
4. C. R. SOIJTHWELL, B. W. FORGESON and A. L. ALEXANDER, Corrosion of Merals ii Tropical
Environmenfs, Part &Wrough/ Iron. NRL Report 5370, Oct. 1959; Corrosion 16, 512t (1960).
5. B. W. FORGESON, C. R. SOUTHWELL and A. L. ALU(ANDER, Corrosion of Metals in Tropical
Environments, Part 5-Stainless Steels. NRL Report 5517, Sept. 1960; Corrosion 17, 345t (1961).
6. C. R. SO~THWELL, C. W. HUMMER, JR. and A. L. ALEXANDER, Corrosion of Metals in Tropical
Environments, Part 6-Aluminium and Magnesium, NRL Report 6105, Dec. 1964; Maferials
Protection 4 (No. 12). 30-35 (1965).
7. C. R. SOIJTHWELL, C. W. HUMMER, JR. and A. L. ALEXANDER, Corrosion of Metals in Tropical
Environments, Part I-Copper and Copper Alloys-Sixteen Years’ Exposure. NRL Report 6452,
Oct. 1966.
8. C. R. SOUTHWELL and A. L. ALEXANDER, Corrosion of Metals in Tropical Environments, Part 8-
Nickel and Nickel-Coooer Allovs-Sixteen Years’ Exoosure. NRL Report 6592, Oct. 1967.
9. C. R. SOLJTEIWELL and :4. L. ALEXANDER, Corrosion df Metals in Tropical Environments, Parr 9-
Structural Ferrous Metals-Sixteen Years’ Exposure to Sea and Fresh Water. NRL Report.