Astm stp 558 1974

CORROSION IN NATURAL ENVIRONMENTS Three symposia presented at the Seventy-sixth Annual Meeting AMERICAN SOCIETY FOR TESTING AND MATERIALS Philadelphia, Pa 24-29 June 1973 ASTM SPECIAL TECHNICAL PUBLICATION 558 List Price $29.75 04-558000-27 AMERICAN SOCIETY FOR TESTING AND MATERIALS 1916 Race Street, Philadelphia, Pa 19103 Copyright by ASTM Int'l (all rights reserved); Fri Aug 14 17:24:15 EDT 2015 Downloaded/printed by Southeast University (Southeast University) pursuant to License Agreement No further reproductions authorized 1974 Library of Congress Catalog Card Number: 74-77097 (~) BY AMERICAN SOCIETY FOR TESTING AND MATERIALS NOTE The Society is not responsible, as a body, for the statements and opinions advanced in this publication Printed in Lutherville-Timonium,Md August 1974 Copyright by ASTM Int'l (all rights reserved); Fri Aug 14 17:24:15 EDT 2015 Downloaded/printed by Southeast University (Southeast University) pursuant to License Agreement No further reproductions authorized Foreword The papers in this special technical publication were presented during three symposia at the Seventy-sixth Annual Meeting of the American Society for Testing and Materials held in Philadelphia, Pa., 24-29 June 1973 The three symposia were: Atmospheric Corrosion S W Dean, Jr., Olin Corporation, chairman V P Pearson, Inland Steel Company, cochairman Metal Corrosion in Seawater W H Ailor, Reynolds Metals Company, chairman Statistical Planning and Analysis of Corrosion Experiments F H Haynie, Environmental Protection Agency, chairman E H Jebe, Ann Arbor, Mich., cochairman These three symposia are included in this publication Copyright by ASTM Int'l (all rights reserved); Fri Aug 14 17:24:15 EDT 2015 Downloaded/printed by Southeast University (Southeast University) pursuant to License Agreement No further reproductions authorized Related ASTM Publications Localized Corrosion Cause of Metal Failure, STP 516 (1972), $22.50 (04-516000-27) Stress Corrosion Cracking of Metals A State of the Art, STP 518 (1972), $11.75 (04-518000-27) Manual of Industrial Corrosion Standards and Control, STP 534 (1973), $16.75 (04-534000-27) Copyright by ASTM Int'l (all rights reserved); Fri Aug 14 17:24:15 EDT 2015 Downloaded/printed by Southeast University (Southeast University) pursuant to License Agreement No further reproductions authorized Contents Introduction I TESTING IN NATURAL ATMOSPHERES Weather Factors Affecting Corrosion of Metals P J SEREDA Moisture Pollution Sulfur Dioxide Chlorides Corrosion Products Temperature Summary 13 13 16 17 19 21 Selecting Testing Conditions Representative of the Atmospheric Environment-J F, STANNERS Estimating the Distribution of Steel Utilization Environmental Factors to be Taken into Account Variability of the Selected Environmental Factors Selection of Conditions Required at Exposure Sites Feasibility of Selecting Sites Discussion Conclusions Correlation Between Corrosion Behavior of Steel and Atmospheric Pollution Data E H HAYNIE AND J B UPHAM Experimental Procedure Results and Analysis Discussion Conclusions 23 24 25 26 28 30 30 31 33 34 35 40 43 Corrosion Aggressivity of Model Regions of Czechoslovakia D KNOTKOV,~CERMAKOV/~, B BOSEK, AND J VLCKOV/~ Analysis of the Problem Tests Analysis of the Results Mathematical Model of Atmospheric Corrosion of Metals Conclusion Seven-Year Exposure at Point Reyes, California w H AILOR Alloys Test Program Short-Term Atmospheric Corrosion of Various Copper-Base Alloys Two- and Four-Year Results R s HERMAN AND A P CASTILLO Experimental Procedure Results Discussion Conclusions 52 53 57 59 66 72 75 75 76 82 84 86 92 96 Copyright by ASTM Int'l (all rights reserved); Fri Aug 14 17:24:15 EDT 2015 Downloaded/printed by Southeast University (Southeast University) pursuant to License Agreement No further reproductions authorized Technical Note: An Evaluation of Titanium Panels After Seven Years' Exposure in a Marine Atmosphere L c COVINGTON 97 Exfoliation Corrosion Testing of 7075 and 7178 Aluminum Alloys Interim Report on Atmospheric Exposure Tests D O SPROWLS,T J, SUMMERSON, 99 AND F E LOFTIN Test Materials Experimental Procedure Results and Discussion Summary Status 100 102 104 112 113 SEAWATER ENVIRONMENTS Ten-Year Seawater Tests on Aluminum w H AILOR Procedure Discussion Conclusions Corrosion of Copper Alloys in Hydrospace F M RE1NHART Results and Discussion Corrosion Stress Corrosion Corrosion Products Mechanical Properties Summary and Conclusions 117 119 122 132 135 139 139 164 164 167 167 Galvanic Corrosion of Ferritic Stainless Steels in Seawater R BABOIANAND (3 S HAYNES Theory Experimental Results Discussion Conclusions Corrosion Tests in the Gulf Floor J, s DI GREGORIOAND J P FRASER Corrosion Tests in the Gulf of Mexico Unprotected Specimens Conclusions Evaluation of Paint Coatings Tested in the Deep Atlantic and Pacific Oceans-J V RYNEWICZ Description of Test Specimens Description of Test Sites Description of Test Fixtures Evaluation of Recovered Test Specimens Comparison of Test Site Results Conclusion 171 172 176 178 180 183 185 188 196 206 209 211 216 216 216 231 234 Copyright by ASTM Int'l (all rights reserved); Fri Aug 14 17:24:15 EDT 2015 Downloaded/printed by Southeast University (Southeast University) pursuant to License Agreement No further reproductions authorized LABORATORY AND STATISTICAL TECHNIQUES Electrochemical Technique for Determination of the Instantaneous Rate of Atmospheric Corrosion v KU~ERAAND E MATTSSON Experimental Procedure Results Discussion Conclusions Accelerated Testing of Marine Grade S t e e i s - A Approach R R M JOHNSTON AND C P LLOYD 239 241 246 256 258 Localized Corrosion Preliminary Development Work Experimental Results and Discussion Summary and Conclusions 261 262 265 268 275 Design of a Laboratory Experiment to Identify the Effects of Environmental Pollutants on Materials j w SPENCEAND F H HAYNIE 279 Background Materials and Effects Measurement Techniques Laboratory-Controlled Simulation of Polluted Environments Evaluation of Environmental Chambers Experimental Design Summary 279 280 281 283 289 291 Prediction of Environmental Pitting and Corrosion Rates R N SHULER AND W H AILOR 292 Procedure Results Conclusions 293 295 305 Relationship of Accelerated Test Methods for Exfoliation Resistance in 7XXX Series Aluminum Alloys with Exposure to a Seacoast Atmosphere-n w LIFKA AND D O SPROWLS Alcoa Seacoast Weathering Station at Point Judith, Rhode Island Performance of Commercial Products in the Seacoast Atmosphere Comparative Performances in Seacoast Atmosphere and in Accelerated Test Media Conclusion 306 307 312 318 333 Effect of Alloy Composition on the Atmospheric Corrosion Behavior of Steels Based on a Statistical Analysis of the Larrabee-Coburn Data S e t - R A LEGAULT AND H P LECKIE Experimental Details Statistical Procedures Used Results Obtained from Statistical Procedures Analysis of the Results Summary 334 335 337 339 342 347 Copyright by ASTM Int'l (all rights reserved); Fri Aug 14 17:24:15 EDT 2015 Downloaded/printed by Southeast University (Southeast University) pursuant to License Agreement No further reproductions authorized STP558-EB/Aug 1974 Introduction As part of its on-going program for "collection of engineer-data and the development of methods of test," Committee G-1 on Corrosion of Metals sponsored three symposia in 1973 These symposia included papers on atmospheric corrosion, seawater corrosion, and statistical planning and analysis of corrosion experiments The papers in this book have been arranged into three groups to permit easier reference These groupings are: Part Testing in Natural Atmospheres Part Seawater Environments Part Laboratory and Statistical Techniques The papers concerned with atmospheric corrosion in this volume were part of the Symposium on Atmospheric Corrosion, which was organized to update the existing knowledge in the field The impetus for this effort was twofold The efforts of governments and industries to reduce atmospheric pollution, especially in urban areas, have changed the nature of atmospheric corrosion in these areas Also, many agencies are now collecting a range of atmospheric data, and this information is now available to correlate with atmospheric corrosion results Furthermore, a variety of new materials has been developed since the last ASTM symposium on this subject in 1967,1 and it was of interest to have at least early performance data on these materials Papers on atmospheric corrosion have been assembled covering a wide range of subjects Five of the papers are concerned with the effects of various weather factors on atmospheric corrosion These cover a range of topics, including estimating the effects of various weather factors in quantitative terms and selecting sites to give an accurate assessment of the performance of materials Other subjects of interest include the effects of various alloying elements in steel on its atmospheric corrosion resistance and a new electrochemical technique for measuring instantaneous atmospheric corrosion rates Four papers deal with the performance of specific materials in atmospheric sites These include some early results on new copper-base alloys and data on the exfoliation of aluminum alloys containing zinc, magnesium, and copper One paper is concerned with the correlation of Metal Corrosion in the Atmosphere, A S T M S T P 435, American Society for Testing and Materials, 1968 Copyright 1974 ASTM www.astm.org Copyright by ASTM Int'lby(all rightsInternational reserved); Fri Aug 14 17:24:15 EDT 2015 Downloaded/printed by Southeast University (Southeast University) pursuant to License Agreement No further reproductions authorized CORROSION IN NATURAL ENVIRONMENTS atmospheric corrosion with the performance of aluminum alloys in accelerated tests ASTM Committee G-1 is now responsible for the 20-year test program initiated in 1957 by its predecessor, Committee B-3 Many of the original seven-year exposure panels from Point Reyes, California were lost through vandalism, making it necessary to prepare a second set of panels for a seven-year exposure These panels have now been retrieved, and[ the results are given in this volume The Symposium on Seawater Corrosion represented a timely mix of several alloys and metal coatings exposed at surface seawater, deep ocean, and sea-floor locations Some galvanic corrosion data for ferritic stainless steels in seawater were also presented The Symposium on Statistical Planning and Analysis of Corrosion Experiments was organized to stimulate greater use of a valuable mathematical tool by corrosion researchers For this reason, two basic statistics educational lectures were presented during the symposium that were not appropriate for inclusion in this volume The remaining presentations represented examples of how some researchers are presently using statistics to plan their corrosion experiments and analyze their data Two of these papers dealt with corrosion in natural environments and are included in this volume The information in this book should be useful to engineers interested in the performance of materials in natural environments; to environmentalists interested in obtaining information on the effects of pollution factors on material performance; and to research workers who are developing new materials intended for service in natural environments Statistically designed experiments provide the researcher with a maximum amount of desired information from a set amount of work Decisions based on statistically analyzed data can be accepted with a measurable degree of confidence These papers should suggest to the reader how he may be able to enhance the results through statistical design and analysis of corrosion experiments W H Ailor, Jr Metallurgical Research Division, Reynolds Metals Company, Richmond, Va 23219; symposium chairman, Seawater Corrosion S W Dean, Jr Olin Corporation, New Haven, Conn 06504; symposium chairman, Atmospheric Corrosion Copyright by ASTM Int'l (all rights reserved); Fri Aug 14 17:24:15 EDT 2015 Downloaded/printed by Southeast University (Southeast University) pursuant to License Agreement No further reproductions authorized LIFKA AND SPROWLS ON EXFOLIATION RESISTANCE 333 service However, the available experience shows that, in order to ensure performance significantly better than that of T6 temper material, the C and D degrees of exfoliation should not be allowed Conclusion In review, it has been shown that commercially fabricated 7075-T73, 7075-T76, and 7178-T76 aluminum alloy products have excellent resistance to exfoliation corrosion when exposed to seacoast atmosphere at a relatively severe test site Thus, these products should provide m a n y years of trouble-free service in aircraft applications Limited evaluations indicate that both the E X C O test, and the acidified salt spray test for determining resistance to exfoliation correlate well with results obtained in seacoast atmosphere However, both tests can produce very slight exfoliation on material that as yet has not exfoliated in seacoast atmosphere and also exaggerate the tendency for undermining-pitting corrosion of material that is resistant to exfoliation Thus, it is concluded that a very slight tendency for exfoliation in these two accelerated tests, as represented by the A S T M visual standard for exfoliation of degree A, is of dubious concern for practical aircraft applications It is suggested that consideration be given to tolerating this degree of susceptibility in 7075-T76 and 7178-T76 materials so as to maintain the o p t i m u m balance between resistance to exfoliation and strength References [1] Dix, E H., Jr and Mears, R B in ASTM Sympsosium on Atmospheric Exposure Tests on Non-Ferrous Metals, American Society for Testing and Materials, 1946, p 57 [2l Walton, C J and King, W in Atmospheric Corrosion of Non-Ferrous Metals, ASTM STP 175, American Society for Testing and Materials, 1955, p 21 13] Lifka, B W and Sprowls, D O., Corrosion, Vol 22, No 1, 1966, p [4] Sprowls, D O., Summerson, T J., and Loftin, F E., this symposium, pp 99-113 15l Sprowls, D O., Walsh, J D., and Sbumaker, M B in Localized Corrosion Cause of Metal Failure, ASTM STP 516, American Society for Testing and Materials, 1972, p 38 [6] McGeary, F L., Englehart, E T., and Ging, P J., Materials Protection, Vol 6, No 6, 1967, p 33 [7] Walton, C J., Sprowls, D O., and Nock, J A., Jr., Corrosion, Vol 9, No 10, 1953, p 345 [8] Ketcham, S J and Jeffrey, P W in Localized Corrosion Cause of Metal Failure, ASTM STP 516, American Society for Testing and Materials, 1972, p 273 Copyright by ASTM Int'l (all rights reserved); Fri Aug 14 17:24:15 EDT 2015 Downloaded/printed by Southeast University (Southeast University) pursuant to License Agreement No further reproductions authorized R A Legault t and H P Leckie Effect of Alloy Composition on the Atmospheric Corrosion Behavior of Steels Based on a Statistical Analysis of the Larrabee-Coburn Data Set REFERENCE: Legault, R A and Leckie, H P., "Effect of Alloy Composition on the Atmospheric Corrosion Behavior of Steels Based on a Statistical Analysis of the Larrabee-Coburn Data Set," Corrosion in Natural Environments, A S T M STP 558, American Soicety for Testing and Materials, 1974, pp 334-347 ABSTRACT: In 1962 C P Larrabee and S K Coburn published an extensive collection of atmospheric exposure data showing the effect of variations in copper, nickel, chromium, silicon, and phosphorus content on the corrosion resistance of low-alloy steel Two hundred and seventy steel alloys of different composition were exposed for 151/2 years at three separate locations representing industrial, semirural, and marine atmospheres Among these experimental steel compositions there were five levels of copper concentration, two levels of nickel concentration, three levels of chromium concentration, three levels of silicon concentration, and three levels of phosphorus concentration represented A statistical analysis of these data has been conducted and the results of the analysis are presented and discussed KEY WORDS: corrosion, alloy steels, atmospheric corrosion, chromium steels, copper steels, corrosion statistics, industrial atmospheres, low-alloy steels, marine atmospheres, nickel steels, phosphorus steels, regression analysis, semirural atmospheres, silicon steels, statistical analysis, steel alloy development, steel compositions, predictions, weathering steels In 1962, Larrabee and Coburn published a very comprehensive collection of atmospheric exposure data showing the effect of variations in the content of copper, nickel, chromium, silicon, and phosphorus on the atmospheric corrosion resistance of low-alloy steel Two hundred and seventy low-alloy steels of different compositions were exposed for ~ years at three separate locations representing industrial, semirural, and marine atmospheres Although generalized conclusions were drawn from Senior research engineer and director, Coated Products Division, respectively, Inland Steel Research Laboratories, Inland Steel Company, East Chicago, Ind 46312 Larrabee, C P and Coburn, S K., First International Congress on Metallic Corrosion, Butterworths, London, p 276, 1962 334 Copyright 1974 by ASTM International Copyright by ASTM Int'l (all rights reserved); Fri Aug www.astm.org 14 17:24:15 EDT 2015 Downloaded/printed by Southeast University (Southeast University) pursuant to License Agreement No further reproductions authorized LEGAULT AND LECKIE ON ATMOSPHERIC CORROSION OF STEELS 335 the data in the initial publication of test results, no thorough statistical analysis of this remarkable data set has, to our know|edge, ever been published Such an analysis, designed to establish the contributions of main effects, binary interaction effects, and quadratic effects of the five alloying elements to the total effect on corrosion rate, has now been completed Reliable predictive equations were developed for corrosion rate as a function of composition for all three test site conditions Experimental Details All 270 experimental steels were reported to have been made in a 30-1b high-frequency induction furnace utilizing an aluminum deoxidation prior to addition of the alloying elements, and then east into by by 12-in slabs The slabs were reheated and cross-rolled to about a 1/~ in thickness, normalized at 1850~ (1010~ and pickled The sheets were sheared into by 6-in specimens and weighed Exposure was initiated in October and N o v e m b e r of 1942 and all of the specimens at each location were placed on the test racks during the same day During exposure, the specimens were supported on porcelain insulators mounted on metal frames placed at 30 deg to the horizontal facing south The exposure sites are described in Table TABLE Exposure sites Industrial: Semirural: Marine: Kearny, New Jersey (five miles west of Lower Manhattan) South Bend, Pennsylvania (thirty-six miles northeast of Pittsburgh) Kure Beach, North Carolina (800 It from the ocean surf) TABLE Relative corrosivity o f exposure sites Site State College, Pa (rural) South Bend, Pa (semirural Kure Beach, N C (marine Kearny, N J (industrial) I Year Years Years Years 1.0 1.5 2.0 3.3 1.0 1.5 2.5 1.0 1.6 3.5 1.0 1.7 5.8 2.7 2.5 2.6 The relative aggressweness or corrosivity of these three sites has been reported by an ASTM committee and is shown in Table It should be noted that the relative corrosivity decreases with time at Kearny, New Jersey; increases slowly with time at South Bend, Pennsylvania; and increases rapidly with time at Kure Beach, North Carolina These changes in site corrosivity with time are shown graphically in Fig I Although it might be intuitively felt on the basis of short-term comparisons that an industrial atmospheric site should be significantly more 3Proceedings A S T M , Report of Committee B-3, Vol 59, 1959 Copyright by ASTM Int'l (all rights reserved); Fri Aug 14 17:24:15 EDT 2015 Downloaded/printed by Southeast University (Southeast University) pursuant to License Agreement No further reproductions authorized 336 CORROSION IN NATURAL ENVIRONMENTS J Ix O ,Y, Z O i u~ O ~3 I O i.u > i.- -,- SOUTH B e~ N O ~ STATE COLLEGE,PaoIRURAL SITE) O E I I I I I 10 I 12 I 14 I 16 I 18 YEARS FIG Change in relative atmospheric corrosion test site corrosivity with time aggressive than a semirural site, the data shown in Fig indicate that after 15 ~ years' exposure little difference in aggressiveness exists between the two sites A marine site, on the other hand, can be expected to be significantly more corrosive than the other two types of sites when longterm exposure data are compared This is borne out in the LarrabeeCoburn 151~-year data as given in Table TABLE 15 89 Location South Bend, Pa Kearny, N J Kure Beach, N C corrosion rates for each exposure site Type Median of 260 Values (mils) Low Value High Value Semi-Rural Industrial Marine 905 3.965 213 1.3 1.6 3.1 7.9 10.4 14.3 Ten of the 270 experimental steels were omitted from this statistical analysis either because the data reported was estimated at one or more sites or because the data reported for one or more sites clearly represented statistical outliers In all of the steel compositions investigated, carbon was present at less than 0.1 percent, manganese at 0.25 to 0.40 percent and sulfur at less than 0.02 percent The experimental response reported, and the dependent Copyright by ASTM Int'l (all rights reserved); Fri Aug 14 17:24:15 EDT 2015 Downloaded/printed by Southeast University (Southeast University) pursuant to License Agreement No further reproductions authorize LEGAULT AND LECKIE ON ATMOSPHERIC CORROSION OF STEELS 337 variable used in this analysis, was the average reduction in thickness for the entire 151~-year exposure period expressed in mils and calculated from weight losses For convenience, this response is referred to herein as corrosion rate The experimental steels used in the study represented five concentration levels of copper, two levels of nickel, three levels of chromium, three levels of silicon and three levels of phosphorus Extreme compositional levels are shown in Table The smallest concentration range, that for silicon, still involves a factor of more than six, and the widest range, that for copper involves a factor of more than forty TABLE Compositional limits represented hi the data set Additive Low High Copper Nickel Chromium Silicon Phosphorus 0.012 0.05 l 0.1 0.01 0.51 1.1 1.3 0.64 0.12 Statistical Procedures Used The first step in the process of statistically analyzing the three sets of empirical data provided by Larrabee and Coburn was the curve fitting of each data set to obtain mathematical expressions which accurately describe the data Several, essentially linear, models were tried and discarded before the decision was made to settle on a quadratic model Such a model for a five-additive system involves an equation of twenty terms: five main effect terms, ten binary-interaction terms, and five squared main effect terms The curve fitting itself was accomplished using standard multiple-regression analysis techniques The procedure used was the "stepwise regression" recommended by Draper and Smith The basic steps in the procedure are as follows: A regression equation is computed by inserting variables in turn until the regression equation is satisfactory The order of insertion is determined by using the partial correlation coeffieient as a measure of the importance of variables not yet in the equation As each variable is entered into the regression, the following values are examined: first, R 2, the multiple correlation coefficient which is defined as the ratio of the sum of squares due to regression to the sum of squares about the mean; and second, the partial F-test value for the variable most recently entered, which shows whether that variable has taken up a sig4 Draper, N R and Smith, H., Applied Regression Analysis, Wiley, New York, 1966, p 171 Copyright by ASTM Int'l (all rights reserved); Fri Aug 14 17:24:15 EDT 2015 Downloaded/printed by Southeast University (Southeast University) pursuant to License Agreement No further reproductions authorized 338 CORROSION IN NATURAL ENVIRONMENTS nificant a m o u n t of v a r i a t i o n over that r e m o v e d by variables previously in the regression Variables which were i n c o r p o r a t e d into the m o d e l in previous stages are r e e x a m i n e d at every stage in the regression A variable which m a y have been the best single variable to enter at an early stage may, at a later stage, be superfluous because of the r e l a t i o n s h i p s between it a n d other variables n o w in the regression Sequential F-test values at each stage are c o m p a r e d with a preselected significance level a n d this provides the criterion for acceptance or rejection o f a p a r t i c u l a r variable This stepwise process is c o n t i n u e d until n o m o r e variables wilI be a d m i t t e d to the e q u a t i o n a n d n o m o r e are rejected F o r the d a t a sets in question, a sequential F statistic equal to 6.7 corr e s p o n d s to a 99 percent confidence level and this was the significance level selected as the criterion for b o t h the acceptance and the rejection of variables in the stepwise regression procedure used TABLE Quadratic model regression statistics for marble exposure Sequential-F to enter Sequential-F to remove Multiple correlation coefficient (R 2) Equation F statistic Constant term Copper Nickel Chromium Silicon Phosphorus Copper-nickel Copper-chromium Copper-silicon Nickel-silicon Nickel-phosphorus Chromium-silicon Silicon-phosphorus (Copper) (Chromium)-" (Silicon) 6.7 (9970 Confidence) 6.7 (9970 Confidence) 950 307 15.49 Coefficient Student-T - 16.30 -4.34 4.79 12.41 - 32.01 2.93 2.46 4.36 2.74 12.82 1.75 20.88 16.60 1.20 4.25 16.99 25.69 16.54 11.79 16.63 7.58 6.31 4.25 8.07 7.18 4.99 4.6I 10.47 6.42 2.89 In the more customary form, the equation would appear as follows : Corrosion rate (mils) = 15.49 - 16.30 (70 Cu) - 4.34 (70 Ni) 4.79 (70 Cr) - 12.41 (~o Si) 32.01 (70 P) + 2.93 (70 Cu 70 Ni) + 2.46 (70 Cu ~o Cr) + 4.36 (70 Cu 70 Si) + 2.74 (70 Ni 70 Si) q- 12.82 (70 Ni 70 P) + 1.75 (70 Si ~oP) q- 16.60 (70 Cu) -k- 1.20 (70 Cr) h- 4.25 (70 Si)2 Copyright by ASTM Int'l (all rights reserved); Fri Aug 14 17:24:15 EDT 2015 Downloaded/printed by Southeast University (Southeast University) pursuant to License Agreement No further reproductions authorized LEGAULT AND LECKIE ON ATMOSPHERIC CORROSION OF STEELS Results Obtained from Statistical 339 Procedures The equation obtained using these procedures together with pertinent statistical information on the data gathered at Kure Beach, N o r t h C a r o l i n a is shown in Table It should be noted that despite the rather severe criteria imposed for the inclusion o f variables in this equation, 15 of the 20 possible terms are represented here The values listed for R and for the F-statistic testify to the quality of the fit obtained It should also be noted that only the main effect terms tend to decrease the corrosion rate; all o f the interaction terms tend to increase the rate Similar information obtained from the d a t a gathered at South Bend, Pennsylvania is given in Table Only ten o f the possible twenty terms appear in this equation and it should be noted that two o f the additives, c h r o m i u m and silicon, appear only in interaction terms Every term which appears in this equation also appears in the marine exposure equation, suggesting that this set o f variables is a subset o f the g r o u p i n g present in the marine exposure equation This equation p r o d u c e d the only interaction terms with negative coefficients It must be pointed out that these terms involved only c h r o m i u m and silicon or both for which no negative m a i n effect terms appear A n a l o g o u s information obtained from the data gathered at Kearny, New Jersey is given in Table Here only eight of the possible twenty TABLE Quadratic model regression statistics J or semirural exposure Sequential-F to enter Sequential-Fto remove Multiple correlation coefficient (R 2) Equation F statistic Constant term Copper Nickel Phosphorus Copper-nickel Copper-silicon Nickel-phosphorus Chromium-silicon (Copper) (Chromium) (Silicon) z 6.7 (99% Confidence) 6.7 (99% Confidence) 0.911 256 8.50 Coefficient Student-T - 13.39 3.03 21.27 3.48 2.41 9.55 - 1.11 15.31 -0.65 -2.82 16.29 24.22 18.22 9.89 2.62 5.88 3.73 10.65 8.96 6.17 This equation can be written as follows: Corrosion rate(mils)= 8.50 13.39 ( ~ Cu) 3.03 ( ~ Ni) 21.27 ( P) + 3.48 (% Cu ~o Ni) + 2.41 (~o Cu ~ Si) q- 9.55 ( ~ Ni ~P) 1.11 (~o Cr ~o Si) q- 15.31 (% Cu) 0.65 (~o Cr) 2.82 (~o Si)2 Copyright by ASTM Int'l (all rights reserved); Fri Aug 14 17:24:15 EDT 2015 Downloaded/printed by Southeast University (Southeast University) pursuant to License Agreement No further reproductions authorized 340 CORROSION IN NATURAL ENVIRONMENTS TABLE Quadratic model regression statistics for #Mustrial exposure Sequential-F to enter Sequential-Fto remove Multiple Correlation Coefficient (R 2) Equation F statistic Constant term Copper Nickel Chromium Silicon Phosphorus Copper-nickel Nickel-phosphorus (Copper)~ 6.7 (99~o Confidence) 6.7 (99~o Confidence) 804 128 10.00 Coefficient Student-T - 26.01 - 3.88 - 1.20 - 1.49 - 17.28 7.29 9.10 33.39 17.63 15.76 9.97 5.22 7.52 10.78 2.91 12.07 This equation would be written as follows: Corrosion rate (mils) = 10.00 - 26.01 ( ~ Cu) - 3.88 (~o Ni) - 1.20 ( ~o Cr) 1.49 ( ~ Si) - 17.28 ( ~o P) + 7.29 ( ~ Cu ~ Ni) + 9.10 ( ~ Ni 70 P) + 33.39 ( ~ Cu) ~ t e r m s a p p e a r in the e q u a t i o n , with o n l y three i n t e r a c t i o n t e r m s r e p r e s e n t e d , e a c h o f which tends to increase the c o r r o s i o n rate A g a i n , every t e r m which a p p e a r s in this e q u a t i o n also a p p e a r s in the m a r i n e e x p o s u r e e q u a t i o n , suggesting t h a t this set o f v a r i a b l e s is also a subset o f the g r o u p i n g present in the m a r i n e e x p o s u r e e q u a t i o n T h e effect o f d r o p p i n g t e r m s n o t m e e t i n g the F > 6.7 c r i t e r i o n is s h o w n in T a b l e T h e h e a d i n g s l a b e l l e d 6.7, 6.7 i n d i c a t e t h a t the F - s t a t i s t i c c r i t e r i o n for e n t r y as well as for r e m o v a l was 6.7 ; the 0, h e a d i n g s i n d i c a t e t h a t all o f the v a r i a b l e s were a l l o w e d to enter a n d t h a t n o n e were r e m o v e d Decreases in the values o f the m u l t i p l e c o r r e l a t i o n coefficients as v a r i a b l e s are o m i t t e d , are insignificant F o r instance, in the N e w Jersey d a t a , alt h o u g h 12 o f the 20 v a r i a b l e s were o m i t t e d , the R value d e c r e a s e d b y o n l y 0.015 T h e 3~ (99 p e r c e n t ) c r i t e r i o n for a c c e p t a n c e a n d for r e j e c t i o n o f variables, t a k e n t o g e t h e r with the highly significant values o b t a i n e d for b o t h m u l t i p l e c o r r e l a t i o n coefficients a n d F-statistics, indicates t h a t a highly s a t i s f a c t o r y fit o f the d a t a has been o b t a i n e d in each case TABLE Effect of dropping equation terms Pennsylvania 0, M.C.C (R 2) F-Statistic No of terms 0.921 139 20 6.7, 6.7 0.911 256 10 New Jersey 0, 0.819 54 20 6.7, 6.7 0.804 128 North Carolina 0, 0.955 256 20 Copyright by ASTM Int'l (all rights reserved); Fri Aug 14 17:24:15 EDT 2015 Downloaded/printed by Southeast University (Southeast University) pursuant to License Agreement No further reproductions authorized 6.7, 6.7 0.950 307 15 LEGAULT AND LECKIE ON ATMOSPHERIC CORROSION OF STEELS 3~1 A measure of how reliable these predictive equations are is shown in Fig 2, in which the observed corrosion rates for the Kure Beach, North Carolina site are plotted against values calculated using the 3~ predictive equation Similar plots for the other two sites show that satisfactory fits have been obtained for all three sites, although there is s o m e w h a t more scatter evident in the N e w Jersey data than from the other two exposure sites Table shows a comparison of the equation coefficients obtained for the three types of exposures This table provides us with a measure of the relative effect of a given alloying element in each environment It should be remembered that the effect of a given additive cannot be obtained +12.7535 I ! J ) +11.7385 T +1o.7234 ~ "-2 >, i l > i ~E + I r I "~ + z i ~~ T +7.6781- i U i a +6.6630~ :~ T ,~ T -.I I ;:r - U 2 i U.I I- m +5.6480- ~::22 22 2::~ 2 ~::2;:: X" I ;i:3 2 :::,:: u +4.63291J I ~::;:,2 ~',, p[: 2 +3.6178', I ;2 3~2 I +2.6027s ::i +3.1000 +5.3400 _ +7.5800 +9.8200 OBSERVED CORROSION RATE (m.p.y.) FIG - - P l o t o f calculated corrosion rates versus observed corrosion rates at Kure Beach, North Carolina Copyright by ASTM Int'l (all rights reserved); Fri Aug 14 17:24:15 EDT 2015 Downloaded/printed by Southeast University (Southeast University) pursuant to License Agreement No further reproductions authorized 342 CORROSIONIN NAIl/RAt ENVIRONMENTS TABLE Comparison of coefficients obtained Jot each site Constant Cu coefficient Ni coefficient Cr coefficient Si coefficient P coefficient Cu Ni coefficient Cu Cr coefficient Cu Si coefficient Ni Si coefficient Ni P coefficient Cr Si coeff• Si P coefficient Cu coefficient Cr coefficient Si2 coefficient Pennsylvania New Jersey North Carolina 8.50 13.39 - 3.03 214i 10.00 - 26.01 - 3.88 - 1.20 - 1.49 - 17.28 7.29 9[5; - 1.11 91 i0 15[)i 0.65 - 2.82 33[;0 15.49 - 16.30 -4.34 -4.79 12.41 - 32.0l 2.93 2.46 4.36 2.74 12.82 1.75 20.88 16.60 1.20 4.25 - 2i[ 57 3.48 s i m p l y b y n o t i n g the a p p r o x i m a t e m a i n effect coefficient; every v a r i a b l e which involves t h a t a d d i t i v e m u s t be t a k e n into account It s h o u l d also be n o t e d t h a t the effect o f a s q u a r e d term coefficient b e i n g greater t h a n zero is to i m p o s e an u p p e r c o n c e n t r a t i o n limit on t h a t p a r t i c u l a r a d d i t i v e a b o v e which it will t e n d to increase the c o r r o s i o n rate r a t h e r than to lower it Analysis of the Results H a v i n g p r o d u c e d a s a t i s f a c t o r y e q u a t i o n for each o f the three e x p o s u r e sites, a t t e n t i o n was d i r e c t e d t o w a r d a d e t e r m i n a t i o n o f o p t i m u m a d d i t i v e levels b a s e d on these e q u a t i o n s O p t i m u m a d d i t i v e levels are defined here as those i n c l u d e d within the c o m p o s i t i o n a l levels e v a l u a t e d in the L a r r a b e e C o b u r n d a t a set; no a t t e m p t was m a d e to e x t r a p o l a t e to c o m p o s i t i o n a l levels o u t s i d e these values F o r e a c h o f the three c o r r o s i o n r a t e e q u a t i o n s , the p a r t i a l d e r i v a t i v e was t a k e n with r e s p e c t to each o f the five additives E a c h rate e q u a t i o n thus p r o d u c e d five p a r t i a l derivatives which were set equal to zero a n d then s o l v e d s i m u l t a n e o u s l y to o b t a i n the o p t i m u m c o n c e n t r a t i o n levels for t h a t e x p o s u r e site T h e o p t i m u m a d d i t i v e levels d e t e r m i n e d b y this p r o c e d u r e are s h o w n in T a b l e 10 T h e letter " m " signifies here t h a t the c o n c e n t r a t i o n level in q u e s t i o n is the m a x i m u m level r e p r e s e n t e d in the d a t a set F o r the K e a r n y , N e w Jersey a n d the S o u t h Bend, P e n n s y l v a n i a e x p o s u r e sites, o n l y c o p p e r shows an o p t i m u m level o t h e r t h a n the highest one r e p r e s e n t e d in the d a t a set A t the K u r e Beach, N o r t h C a r o l i n a exp o s u r e site, silicon j o i n s c o p p e r in s h o w i n g an o p t i m u m c o n c e n t r a t i o n at an i n t e r m e d i a t e level Copyright by ASTM Int'l (all rights reserved); Fri Aug 14 17:24:15 EDT 2015 Downloaded/printed by Southeast University (Southeast University) pursuant to License Agreement No further reproductions authorized LEGAULTAND LECKIEON ATMOSPHERICCORROSIONOF STEELS 343 TABLE lO Optimum additive concentration levels Pennsylvania New Jersey North Carolina 0.17 1.1 (m) 1.3 (m) 0.64 (m) 0.12 (m) 0.18 1.1 (m) 1.3 (m) 0.64 (m) 0.12 (m) 0.25 ~.I (m) 1.3 (m) 0.42 0.12 (m) Copper Nickel Chromium Silicon Phosphorus Accordingly, Fig shows the variation in corrosion rate at Kure Beach, North Carolina with changes in copper and silicon concentration when the other three additives are present at m a x i m u m levels for the data set; that is, nickel at 1.1 percent, phosphorus at 0.12 percent, and chromium at 1.3 percent This is a three-dimensional plot in which corrosion rate is plotted vertically, thus producing a surface which depicts corrosion rate changes It is interesting to note that whereas at the lowest level of copper shown (0.05 percent) the corrosion rate decreases with increasing silicon content, the opposite is true at the highest level of copper shown (0.4 percent) A minimum corrosion rate is produced, therefore, at an intermediate level of both copper and silicon This effect can be seen in Fig 4, in which a series of contour lines (iso-corrosion rate lines) for various corrosion rate levels at Kure Beach is depicted The minimum corrosion rate occurs at lower and lower silicon levels as the copper concentration is increased Ni = 1.1~ P = 0.12% / ~~ 2"8 ~x, ~ 2.7' = 2.6z O 2.42.3~ 2.2- ~ \ I ~ //~ / ~ / ~ / A" I II /_JUJ/ I 0.10 0.15 0.20 0.25 %Cu , , 0.30 0.35 I 0.50 ,/O46 I /*'0.44 ~.42 % Si I J ~ , , 0.05 "~ /0.40 0.40 FIG Three-dimensional plot showing the effect o f corrosion rate at a marine test site as both copper and silicon concentrations are varied Copyright by ASTM Int'l (all rights reserved); Fri Aug 14 17:24:15 EDT 2015 Downloaded/printed by Southeast University (Southeast University) pursuant to License Agreement No further reproductions authorized 344 CORROSIONIN NATURAL ENVIRONMENTS 0.50, 2.779 _2.519 0.48 0.46 784 !2.792 0.44 ~.804 0.42 0.40 !.818 2.837 2.519 ~ ~ 2.523 2.530 o 2.341 =2.247 2.236 2.307 2.462 2.699 2.337 o/0 2.290 4 2.673 o2"i37 2.554 2.234 o 2.213 C o 2.339 ~ ~I ~ ~/ 2.540 2.239~2.223 2.346 ~ 2.266 ~ iv ~ olol 2.234 ~/ 2.651 2.276 2.205 ~ I 2.259 2.206 2.255 02.4i o 632 2.396 2.616 2.388 2.604 0.38 858 ,2.5.72 =2.368 &2.247 V2.2098,2.2 55 ,2.38.4 2.595 0.05 0.10 0.1 O.2 0.40 0~25 0.30 0.35 %Cu FIG 1SO-corrosion rate lines f o r a marine test site as both copper and silicon concentrations are varied Figure shows the variation in corrosion rate with copper content at Kure Beach when the other four additives are at the highest levels represented in the data set (Ni = 1.1 percent, Cr = 1.3 percent, P = 0.12 percent and Si = 0.64 percent), and again when they are at the optimum levels represented in the data set (Ni = 1.1 percent, Cr = 1.3 percent, P 0.12 percent and Si = 0.42 percent) The advantage derived from using the lower silicon concentration is shown dearly in Fig A three-dimensional plot such as that shown in Fig makes it possible to examine interaction effects on corrosion rate in the concentration regions of greatest interest A large number of such plots have been constructed using each of the three equations derived in the present study Figure shows the effect of copper and silicon at an industrial site A comparison of Fig with Fig thus allows an examination of behavior differences noted between marine and industrial exposures The highest level of copper shown (0.4 percent) is clearly more effective in lowering corrosion rate at an industrial site than at a marine site Another example of a three-dimensional plot is shown in Fig in which the effect of copper and nickel concentration changes at an industrial site is depicted It is interesting to note that, at Kure Beach, for the highest Copyright by ASTM Int'l (all rights reserved); Fri Aug 14 17:24:15 EDT 2015 Downloaded/printed by Southeast University (Southeast University) pursuant to License Agreement No further reproductions authorized LEGAULT AND LECKIEON ATMOSPHERICCORROSION OF STEELS 345 2.8~ 2.7 2.6 Si=0.42% z 2.5 O O ,-, 2.4 O U 2.3 2.2- 2.1- HIGHEST LEVELSO Ni=1.1%, Cr=1.3%, P=0.12%, S i = % O P T I M U M LEVELSO Ni=1.1%, Cr=1.3%, P=0.12%, S i = % 2.0 0.05 I O.10 I 0.15 I 0.20 I 0.25 I 0.30 I 0.35 I 0.40 % COPPER FIG Effect o f copper concentration on corrosion rate at a marine test site comparison o f the optimum levels o f the other four additives with the highest levels represented in the data set level of copper shown (0.4 percent), the corrosion rate decreases with increasing nickel concentration The reverse effect was calculated for the Kearny, New Jersey site These considerations allow not only the selection of optimum concentration levels for a given type of exposure, but also provide information concerning how much the corrosion rate can be expected to change as the additive concentrations are varied from their optimum values Copyright by ASTM Int'l (all rights reserved); Fri Aug 14 17:24:15 EDT 2015 Downloaded/printed by Southeast University (Southeast University) pursuant to License Agreement No further reproductions authorized 346 CORROSION IN NATURAL ENVIRONMENTS I CORROSION /RATE Impy) / \ 1.5 , Zl o$ /1\1 I\I I ~ i/ 0.05 I " ,/ Z~ /I O.15 /1 ,,/4"[ z 0.20 0.25 %Cu ~ 0.30 Jo4o Z ~' ' /' /~' ~ / ~ / I / ~ I~ 0.10 ~i_-d::2 ~ I - ' o ~ ~' 1~4~ 0.35 0.40 F I G Three-dimensional plot showing the effect on corrosion rate at an industrial test site as both copper and silicon concentrations are varied CORROSION RATE (mpy) I / J ^^ I- z u ~ ~ Si=0.64% P =O.12% 0.2 0.4 0.6 1.0 0.05 0.10 0.15 0.20 0.25 % Cu 0.30 0.35 1.2 0.40 F I G Three-dimensional plot showing the effect on corrosion rate at an industrial test site as both copper and nickel concentrations are varied Copyright by ASTM Int'l (all rights reserved); Fri Aug 14 17:24:15 EDT 2015 Downloaded/printed by Southeast University (Southeast University) pursuant to License Agreement No further reproductions authorized LEGAULT AND LECKIE ON ATMOSPHERIC CORROSION OF STEELS 347 Summary A reliable predictive equation relating corrosion rate to composition has been obtained for each of the three exposure sites (marine, industrial, and semirural) used in the Larrabee-Coburn atmospheric corrosion study For both the industrial exposure site at Kearny, New Jersey and the semirural site at South Bend, Pennsylvania, only copper among the five additives studied produces a minimum in corrosion rate at a concentration level other than the highest level represented in the data set For the marine exposure site at Kure Beach, North Carolina both copper and silicon produce corrosion rate minima at intermediate concentrations Within the compositional limits of the data set, statistical analysis shows corrosion rates to vary over a wide range A statistical analysis of the type described here makes possible the delineation of optimum compositional ranges with respect to corrosion resistance under environmental conditions corresponding to the three exposure sites described Acknowledgment The authors wish to acknowledge the helpful discussions and suggestions of J Dalal and A G Preban of the Inland Steel Research Laboratory Copyright by ASTM Int'l (all rights reserved); Fri Aug 14 17:24:15 EDT 2015 Downloaded/printed by Southeast University (Southeast University) pursuant to License Agreement No further reproductions authoriz