by P. J. Sereda Division of Building Research, National Research Council, Ottawa, Canada
I CORROSION A
W O R K B O O K
F E A T U R E
E C
Atmospheric Corrosion of Steel Measurement of temper ature, sulfur dioxide, and surface moisture has re sulted in a systematic evaluation of the atmos pheric corrosion of steel and should provide a basis for the rating of the inland exposure sites
laboratory experiments showed that temperature affects corrosion rate of steel (5), and that at — 25° C. corrosion rate decreases to a low value, relationship between at mospheric temperature and cor rosion rate has not been siudied. Similarly, moisture and sulfur di oxide air pollution rate to corrosion rate of steel are important as shown by laboratory and field experiments (7-6), but the exact relationship has not been established. A relationship can be established which shows how the rate of corro sion of low-carbon plain steel, in the presence of oxygen and water, is modified by the atmospheric factors of temperature and sulfur dioxide pollution rate. This relationship applies to particular type of steel samples when exposed for one month at 30 degrees to the herizontal. That temperature is important in atmospheric corrosion was not appar ent until the observed corrosion rate was reduced on the basis of days of wetness. The apparent corrosion rate of steel based on total time of exposure was always found to be higher in the winter than in the summer (6). However, when the time-of-wetness was also measured it was found that this, too, was much higher in the winter than in the sum mer. When the two were combined the result was a lower rate of corro^\LTHOUGH
Comparison of panel temperature with the corresponding air temperature
sion in the winter than in the sum mer. When monthly mean tempera tures for Ottawa obtained by the Meteorological Department of the Department of Transport are com pared with the monthly average tem peratures of the panel during the periods of wetness a surprising similarity is found. This relation ship (figure above) suggests that on a yearly basis it may be possible to use mean annual temperature as the temperature at which corrosion occurs. Atmospheric pollution is also an important factor especially when considered in relation to tempera ture. At 11° F., a fivefold increase in sulfur dioxide pollution rate in creased the rate of corrosion by a factor of 1.6 whereas at 39° F. a threefold increase in the pollution I/EC
rate increased the rate of corrosion by a factor of 1.9. Experimental
During exposure of the test sam ples, the period when moisture was present on the surface, defined as the time-of-wetness, was obtained from the record of the moisture-sensing element exposed in the same manner and location as the test samples. The method (7, 8) measures the potential developed by closely spaced electrodes of platinum and zinc when the surface film moisture provides the electrolyte. The time-of-wetness was obtained for both the skyward and the groundward exposures and the total time-of-wetness for any test sample was the average of the
W
ORKBOOK
FEATURES
79 Α
I/EC
CORROSION
.
A Workbook Feature
two values giving the corroding time as days of wetness, Dw. The groundward exposure showed the presence of moisture longer than the skyward exposure during periods when con densed water was present. This first direct evidence explains at least in part observations that steel cor rodes faster on the groundward side (4). T h e temperature of a corroding sample of steel was measured con tinuously by a copper constantan thermocouple and a Leeds and Northrup Speedomax recorder. Thirty-gage wire was looped on the groundward side of the panel and held to the metal surface by Scotch electrical tape No. 56. This average temperature was used in relating the effect of temperature on the rate of corrosion. Temperature of the sample dropped as much as 8° F. below air temperature at night when the sky was clear and rose as much as 30° F. above air temperature when the sun was shining. Except in the coastal areas where sea-water spray is present, there is little evidence that trace contami nants other than sulfur dioxide con tribute substantially to the corrosion of steel. Only the deposition rate of sulfur dioxide was measured using the lead peroxide method (3). Sulfur dioxide was measured at two sites, 800 yards apart, where steel samples were exposed. One of these sites was near the power house of the National Research Council property where the sulfur dioxide pollution rate was as much as four times higher than at the other site (one of eight Canadian corrosion sites used by the N R C Associate Committee on Corrosion Research and Prevention, A C C R P ) . This sulfur dioxide pollution rate was an average daily rate based on a monthly collection. It was assumed, for purposes of correlation with cor rosion rate, that the rate during periods when wetness occurred was the same as for dry periods. It is probable, however, that some dif ferences do occur because of wind direction associated with rainy weather. It is also assumed that the rate at which sulfur dioxide was collected on the lead peroxide cylin der bears a constant relationship to the rate at which it is deposited on the corroding samples of steel. 80 A
Samples of low-carbon plain steel measuring 4 by 6 inches by l/g inch were used. T h e composition of the steel, determined by the Steel Co. of Canada, is :
T h e analysis of variance indicates that there is a definite regression of temperature and sulfur dioxide pol lution rate with respect to the rate of corrosion.
% Carbon 0.052 Phosphorus 0.014 Sulfur 0.039 Manganese 0.350 Silicon 0.002 Aluminum 0.008 Copper 0.066 Molybdenum 0.007 Tin 0.011 Nickel 0.048 Surface of the samples was gritblasted using No. 30 aluminum oxide, then degreased with carbon tetrachloride. Samples were exposed for one month at 30 degrees to the horizontal on standard corrosion racks facing south. The corrosion products were removed by 5 0 % hydrochloric acid solution containing 2 % rodinc. A blank was run to check the amount of steel removed. This amount was always less than 1 % of the total weight loss. Samples were exposed in triplicate and agreement was good with differences between weight losses of less than 2 % . T h e weight loss was reduced to the corrosion rate in terms of milli grams per square decimeter per day of wetness by combining the weight loss for the month with the time-ofwetness for the same month. This corrosion rate was then plotted against the average temperature values and the corresponding pol lution rate in terms of milligrams of sulfur trioxidc per square decim eter per day, on three coordinate isometric graph paper. The rate of corrosion increased logarithmically with temperature and a replot of the values of log rate against tempera ture and sulfur trioxide gave a reasonable plane. The least-squares method was then used to determine the equation for the best fitting plane for the values. T h e equation for this relation is:
Observations
Y = 0.1315AT + 0.0180^ + 0 7873 where Y = log corrosion rate, MDDW (mg./sq. d m . / d a y of wetness) X = sulfur dioxide pollution rate, AIDD (mg. sulfur trioxide/sq. dm./day) Ζ = temperature ° F. (monthly average during the time-of-wetness)
INDUSTRIAL A N D ENGINEERING CHEMISTRY
When corrosion in coastal areas is considered, the effect of the salt carried as spray from the ocean must be included. The relationship given here should be modified to include the effect of salt and the measure ment of salt concentration. Temperature of the sample ap pears to affect the process of corro sion directly and it is also important in altering the time-of-wetness due to the temperature difference between the sample and the ambient condi tions. This involves the heat trans fer by radiation and convection between the sample and its environ ment. In this regard the size, geometry, and orientation of the sample are important (2) and these factors should be studied. Acknowledgment
T h e author gratefully acknowl edges the assistance of John Harris and Maurice Cohen in analyzing the data, S. E. Dodds and H. F. Slade in collecting the data, and J. S. Ride of the Steel Co. of Canada for steel analysis. This contribution is from the Division of Building Re search, National Research Council, Canada, and published with ap proval of the Division Director. Literature Cited (1) Copson, H. R., Am. Soc. Testing Mate rials, Proc. 45, 554 (1945). (2) Gciger, R., "The Climate near the Ground," Harvard Univ. Press, Cam bridge, Mass., 1950. (3) Great Britain, Dcpt. of Scientific and Industrial Research, Her Majesty's Sta tionary Office, London, "Atmospheric Pollution," 18th Rept., 1953. (4) Larrabee, C. P., Trans. Eleitrochem. Soc. 85, 297 (1944). (5) Preston, R, St. J., J. Iron Steel Inst. (London) 160, 286-94 (1948). (6) Schikorr, Gcrhart, Schikkor, Ina, Z. Metallic. 39, 175-81 (1943). (7) Sereda, P. J., ASTM Bull., No. 228, 53-5 (February 1958). (8) Sereda, P. .T., Ibid., No. 238, 61-3 (May 1959).
c Our authors like to hear from readers. If you have questions or comments, or both, send them via The Editor, l/EC, 1155 16th Street N.W., Washington 6, D. C. Letters will be forwarded and answered promptly.
