Evaluation of the atmospheric corrosion resistance of AISI A-36 steel painted with coatings based on epoxy and poly(urethane) resins using semi-accelerated testing.

Abstract The influence of temperature, humidity, salinity, and the number of we-dry cycles in the atmospheric corrosion resistance of the AISI A-36 steel, painted with coatings based on epoxy and poly(urethane) resins was studied. The aim of this work is to present an alternative to the accelerated and field tests developing a semi-accelerated test for evaluating that is applied to a specific of a protection system that is applied to a specific environment for one year. The samples were exposed in a corrosion cyclic test chamber for 2000 h with fixed parameters of temperature, relative moisture of air, salinity, and a number of dry-wet cycles. Mass loss of substrate in the scratch area was measured after each 250 h of testing. The second stage of research was the sensibility analysis of the corrosion rate with the change of the reference test parameters. The third stage was the testing of the samples using the cyclic test for 750 h, and the transfer of samples to a marine atmosphere exposure for one year. The mass loss was measured after each 60-day period of testing. Material characterization involved scanning electron microscopy, X-ray diffraction, energy dispersive spectroscopy, Fourier transform infrared spectroscopy, and X-ray fluorescence spectroscopy. After polynomial adjustment of the mass loss function associated with the sensibility analysis of the parameter alteration, the estimated life span of the samples ranged from 30 to 38 months.

Keywords Atmospheric corrosion, Epoxy, Poly(urethane), Field test, Cyclic corrosion test, Carbon steel

Introduction

Atmospheric corrosion is a complex phenomenon that depends on several factors, such as temperature, relative humidity, salinity, and number of wet-dry cycles. (1) Incompatibility between the conditions of the laboratory tests and the conditions of the service environment can generate serious mistakes of paint system specification and of maintenance plans, producing failure of the corrosion protection system.

Laboratory tests are important due to the short test period, but the obtained results are insufficient for the engineers to make decisions about material selection and material specification. Literature (2) reports the use of semi-accelerated tests for the atmospheric corrosion resistance evaluation of steel and its painting system, as well as for the civil construction material selection. Lins et al. (2) reported that the results of corrosion resistance of steels that were phosphatized and painted before exposure to an industrial atmosphere for a 24-month period, and with an application of a saline solution spray once a week, were similar to those of the results of GM 9540P/B cyclic testing with 40 cycles.

The aim of this work is to present an alternative to the accelerated and field tests, developing a semi-accelerated test to evaluate the resistance and durability of a protection system applied to a specific environment for one year. The experiment starts in the laboratory and ends as a field test in the marine atmosphere. The material studied was carbon steel painted with coalings based on poly(amide) epoxy and aliphatic poly(urethane) resins.

As cure agents of epoxy resins, the poly(amides) show paint film with better flexibility and water resistance, and chemical and thermal resistance than poly(amines). The epoxy resins, or epoxy-based coatings, (3-6) are used in paint systems as a corrosion protector coating. Literature (7) cites electrophorctic coating (E-coating) using epoxy resin and spraying of polyure-thane coating with the addition of metallic pigments as two coatings that represent the state-of-the-art technology for protection in the heat, ventilation, and air conditioning/refrigerating field (HVAC/R).

Poly(urethane) is a thermoset polymer that has been used in the coaling industry over the last 45 years. (8.9) One disadvantage of polyurethanes is the potential risk of isocyanates used as cure agents, but this is being overcome.

A sensibility study of the variation of corrosion rates in the functions of salinity, temperature, relative humidity, and frequency of wet-dry cycles was performed. Sensibility analysis is a systematic process used by exploring how an optimal solution (in the paretian sense) is able to react under the changes that have been introduced to the initial conditions. (10), (11) The analysis is based on the initial assumption that optimization is the most important and desirable instance, by taking uncertainty as a potentially detrimental factor. The objective of sensibility analysis is to analyze and discover the sensibility strength of the optimal solution with the changes introduced in the essential factors. (10), (11) The estimated and obtained curves of the corrosion resistance behavior were created. The error analysis was made. The knowledge of the corrosion rate of the paint system can help manufacturers of equipment to estimate the life span and maintenance plans of the products.

The atmospheric corrosion resistance of a structural steel AISI A-36, painted with iron oxide poly (amide) epoxy and aliphatic poly(urethane), was evaluated by using accelerated, field, and semi-accelerated corrosion tests. The sensibility curves (10.11) of the corrosion rate of painted steel in the function of salinity, temperature, relative humidity, and the frequency of wet-dry cycles were obtained. The life span of the painted steel was estimated.

Understanding the atmospheric corrosion mechanism requires study of the influence of environment on corrosion rates and study of the formation of corrosion products by using selected analytical techniques. (12) The accelerated tests develop specific mechanisms that can be different from those observed in service. The cyclic tests were more aggressive than the field tests in some atmospheres. The phase identification of the corrosion product is an important tool in the mechanism determination. The presence of a great amount of goethite can indicate a great number of wet-dry cycles. (12) New mathematical models are being developed that consider ion diffusion through the corrosion film, and the dissolution of the corrosion product layer due to the interaction with the environment. (13) A model for the marine atmospheric corrosion was proposed by Melchers. (14)

Park el al. (15) studied the water absorption in a double-layer paint film consisting of epoxy and urethane by using electrochemical impedance spectroscopy.

Methodology

The substrate used was AISI A-36 steel, blasted and painted with iron oxide poly(amide) epoxy and aliphatic poly(urethane) in a gray color. The steel samples were cut from a 3 mm x 1200 mm x 3000 mm sheet. The samples were produced in two shapes, shown in Fig. 1. One type has dimensions of 150 mm x 150 mm, and the other has an irregular shape.

[FIGURE 1 OMITTED]

The blasting with steel shot was performed according to the SSPC-SP-10 NACE 2 Standard and the SA 21/2 standard. The granulometry of the steel shot and the roughness of the substrate were measured. The steel roughness was 37 [+ or -] 2 [micro]m.

Painting of the steel was performed after 8 h of blasting. Iron oxide poly(amide) epoxy resin (Starflex 339, Star Chemicals) was applied by using spray at a relative air humidity of 65%. The viscosity of paint was measured according lo the ASTM D1200-94 (2005) Standard, and was 40 [+ or -] 2 s, using Ford cup number 4. The dry paint thickness was measured in accordance with the ASTM E 376/89 Standard, and was 80--0/ + 10 [micro]m over of peaks of blasting, totaling 115-130 (129) [micro]m of dry film. The cure was performed in the air for 12 h.

The second layer applied on a steel surface was aliphatic poly(urethane) (Stardur 470. Star Chemicals). The application was performed with 55 [+ or -] 2 s viscosity, using the Ford cup number 4, according to the ASTM D 1200-94(2005) Standard. The dry paint thickness was measured in accordance with the ASTM E 376/89 Standard, and was 80--0/+10 [micro]m. After a period of 10 days of cure, the paint film adherence was evaluated using the tape test method according to the ASTM D 3359/97 Standard, in dry conditions. A longitudinal cut was performed to expose the substrate in an edge of 0.4 mm in width. The substrate region exposed to environmental conditions is called the "scratch area."

A reference accelerated test was performed using a salt spray chamber, USC Model of BASS Industry, and a stove manufactured by Toshiba T & D of Brazil. The instruments used--such as thermometers and hygrometers--were calibrated. The samples were tested for 2000 h at a temperature of 35[degrees]C, relative humidity of 100%, saline solution spray with 5% (w/w) of NaCl, and 360 wet-dry cycles per year. The cycle was 23 h per day of exposure in a salt spray chamber, and 1 h per day of exposure in a stove at 35[degrees]C.

The penetration of the corrosion product under the paint film was evaluated in accordance with the ASTM D1654-05 Standard. The blistering was evaluated in accordance with the ASTM D714 Standard.

The characterization of the corrosion products was performed by using scanning electron microscopy (SEM) and energy dispersive spectroscopy (EDS). The equipment used was Jcol (JSM-35 model), 25 kV of accelerating potential, and lateral resolution of 3 urn. X-ray fluorescence and X-ray diffraction (XRD) were performed using the Philips spectrometer (PW 2400 model) and Philips diffraetometer (X'Pert-APD system, PW 3710/31 control). Fourier transform infrared (FTIR) analysis was done using the Perkin-Elmer spectrometer (1760X model).

A Shimadzu thermal analysis system was used for this study, and thermogravimetric analysis (TGA-50 module) was the technique employed. TGA was used to measure the weight loss and degradation temperatures of the samples. All the measurements were made in a synthetic air atmosphere--a mixture of O2 (20 [+or-] 0.5% v/v) and N2--with a 50 mL/min flux. TGA was made using 8.65 mg of poly(urelhane) and 11.73 mg of epoxy resin at temperatures ranging from 40 to 900[degrees]C, with a heating rate of 20[degrees]C/min.

The samples were weighed after time periods of 250 h, using a Ohyo Jupter analytical balance (C3-200D model) and a resolution of 0.0001 g. The cleaning of samples was done after the corrosion tests in accordance with the ASTM Gl-03 Standard.

A polynomial approximation was done using the least squares minimization method to obtain a function of mass loss vs time for the reference test. The method used in the modeling of results was the method reported by Hoffman. (16) Corrosion tests were performed by changing one parameter in each experiment in relation to the reference test. The studied values of temperature, relative humidity, salinity, and the number of wet-dry cycles per year are shown in Table 1. The sensibility (10), (11) for the changed parameter was evaluated and a new function of mass loss vs time was obtained.

Table 1: Values studies of parameters

Temperature    Relative        NaCl content    Number of
([degrees]C)   humidity (%)    (%w/w)          wet-dry cycles
                                               per year

               60                              72
               70                              90
               80              3.5             120
25             90              4.2             180
35 (reference  100 (reference  5.0 (reference  360 (reference
test)          test)           test)           test)
45
55

Another series of samples was tested in a laboratory for 750 h and then transferred from the laboratory to the field test in a marine atmosphere for one year. The samples were exposed in an atmospheric station in accordance with the ASTM D1014-02 Standard. The mass loss was measured after each 60-day period of testing. Samples that were not submitted to the accelerated test were also exposed in a field test for one year. Temperature, relative humidity, and the number of rainy days were registered. The salinity of this test was considered as 3.5% (w/w), considered to be the average content of salt in oceans. (3)

Results and discussion

The substrate used was 0.16%C-0.14%Si-0.6%Mn-0.01%Cr-0.03%Al-0.01%Cu (w/w) AISI A-36 carbon steel. The substrate was blasted with steel shot, whose average diameter was 0.9 mm. The sample diameter was below 2.38 mm.

The painted steel submitted to 1500 h of the reference accelerated testing showed a maximum detachment of 1.5 mm in the scratch area, and the permissible detachment was 2 mm in accordance with NBR 11388 Standard. Blistering started after 1600 h of accelerated testing. Blistering was classified as d2/t3 according to the NBR 5841 Standard. The maximum corrosion resistance was evaluated at 1500 h. At this time, blistering and failure of adhesion occurred. Total testing time of accelerated corrosion test was 2000 h.

Thermal analysis of aliphatic poly(urethane) resin indicated a thermal stability up to 250[degrees]C. The rate of mass loss was higher at 441 [degrees]C The iron oxide epoxy has a thermal stability up to 100[degrees]C. The maximum temperature studied in this work was 55[degrees]C and at this temperature the paints were thermal stable.

The chemical structure of epoxy resin is shown in Fig. 2. FTIR analysis-identified bands related to the following groups: OH of carboxyl, CH, C[H.sub.2], C-O, C-H of benzene, and (C[H.sub.2])n which are present in the chemical structure of epoxy resin. FTIR analysis of poly(urethane) identified the bands related to the groups of OH of polyol, C=0 of urethane linkage. -RNHCOOR'-. and C[H.sub.2]. X-ray fluorescence technique identified the following elements in higher amounts in epoxy resin: zinc, iron, calcium, silicon, nitrogen, chromium, sulfur, phosphorous, magnesium, and oxygen. The poly (urethane) paint showed titanium, oxygen, barium, silicon, iron, and sulfur as main elements.

[FIGURE 2 OMITTED]

Figure 3 shows a surface of painted steel after 306 h of the reference test. The semi-quantitative chemical analysis (EDS analysis) of the paint surface identified the elements of Ti, Ca, CI, Si, Al. Mg, Na, and oxygen (O), which are derived from the polymers, fillers, or pigment materials, or from the salt spray. In the scratch area, iron and oxygen were the main elements presented in higher amounts. Chlorine is derived from the epichlorohydrin of epoxy resin and from the salt spray. SEM results showed that the width of scratch increased from 1.50 mm for the sample submitted to 306 h of cyclic test (Fig. 3) up to 8.74 mm for the sample tested for 1500 h.

[FIGURE 3 OMITTED]

X-ray diffraction results identified the akaganeite and magnetite phases in the corrosion product of samples submitted for 1000 h of reference accelerated testing. Akaganeite (17), (18) is an iron(III) oxide-hydroxide/ chloride mineral with the formula: [beta] [Fe.sup.3]+O(OH,Cl; [beta]-FeO(OH). (18) The mass loss in function of testing time for the reference test was obtained by using polynomial approximation (square minimum deviation), as in equation (1).

m(t) = -1.41 x [10.sup.-4] [t.sup.2] + 0.45t + 10.12 [R.sup.2] - 99.44%. (1)

Since the function was of the second degree of lime, the corrosion rate was a function of the first degree. The corrosion rate decreased with the increase of testing time due to formation of a corrosion product layer, which acted as a barrier. According to Melchers, (14) the mass loss of steel exposed to marine atmosphere is described by equation (2):

c(t) = [[2(At + B)].sup.1/2] (2)

The A and B constants were calculated using the results obtained for times of 250 and 500 h. A comparison between the results of mass loss obtained in this work and the values obtained using the function proposed by Melchers (14) showed a difference between --6.0% up to +0.6%. According Benarie,(19) atmospheric corrosion can be described by equation (3):

[delta]m = [kt.sup.(1-n)] (3)

A comparison between the results of mass loss obtained in this work and the values of mass loss obtained using the function proposed by Benarie (19) showed a difference between +0.3% and +10.9%.

Sensibility to alteration of parameters

The corrosion rate was changed, due to the decrease in temperature, from 35 to 25[degrees]C, determining a new function of mass loss. Considering that the new function has the same behavior, the same order, and a constant ratio of the both corrosion rates, the sensibility to the alteration of one parameter can be represented by equation (4):

Sensibility = ([partial derivative][m.sub.(1)] (txyzw))/[partial derivative]/([partial derivative][m.sub.(2) (tx'yzw))/[partial derivative]t = f1(t)/f2(t) (4)

where [m.sub.(1)] is corroded mass for a time function of r, x, v, z e w; [m.sub.(2)] is corroded mass for a time function of f, x', y, z e w; x is temperature (T); v is air relative humidity (RH); z is salinity (SLN); w is number of wet-dry cycles (CDW); and x' is temperature for the f2(f) function.

The x, y, z, w e x' variable were constant in each test, and the derivative can be changed from a partial derivative to a total derivative.

([dm.sub.(1)] (txyzw))/dt/([dm.sub.(2)] (tx'yzw))/dt = f1(t)/f2(t) (5)

The following equations are obtained from equation (5).

([dm.sub.(2)] (tx'yzw) = [dm.sub.(1)] (txyzw)/f1(t)/f2(1) (6)

([dm.sub.(2)] (tx'yzw) = f2(t)/f1(t) [dm.sub.(1) (txyzw) (7)

[integral] ([dm.sub.(2)] (tx'yzw) = [integral] f2(t)/f1(t) ([dm.sub.(1)] (txyzw) (8)

The f2(t)/f1(t) term is a constant and can be moved outside the integral.

[integral] [dm.sub.(2)] (tx'yzw) = f2(t)/f1(t) [dm.sub.(1) (txyzw) (9)

[m.sub.(2)] (t) = f2(t)/f1(t) [m.sub.(1)] (t) (10)

f2(t) = [m.sub.(2)] (t)/[m.sub.(1)] (t) f1(t) (11)

The [m.sub.(2) (t = 250)/[m.sub.(1)] (t = 250) ratio is known because the corroded mass at 250 h of test was measured, and so the f2(t) function can be calculated. The last term of the equation of f1(t) function can be calculated. The last term of the equation of f1(t) can be excluded, which is not a function of time.

The same procedure was adopted for the study of variation of relative humidity, salinity, and the number of cycles. The results are shown in Figs. 4-7.

Temperature has a great influence on corrosion rate, and the obtained ratio of [m.sub.(2)] (t)/[m.sub.(1) (t)--a sensibility factor--varies significantly as a function of temperature (see Fig. 4). The curves shown in Fig. 4 have an error due to the exclusion of the third member of equation (1), which is not a function of time. For 250 h of exposure, the error was 9.64%, and decreased for higher times. The corroded mass increased with the increase in temperature, but the corrosion rate decreased. The barrier effect of the corrosion product inhibits the development of corrosion.

[FIGURE 4 OMITTED]

A curve of corroded mass was obtained by using the least square method and was adjusted to second degree polynomial;

m(UR) = 0.02 x [(UR).sup.2] - 2.75 x (UR) + 76.77

[R.sup.2] = 99.79% (12)

The curve of the reference test can be corrected for other values of relative humidity with the data of the m(UR) function, and the curves in Fig. 5 were obtained.

[FIGURE 5 OMITTED]

The relative humidity was the factor of greatest influence in the increase of corrosion rate. The volume of the corrosion product and the corrosion rate increased as the relative humidity increased. The error due to the exclusion of the third member of equation (1) was 23.9% and decreased for higher testing times. The equation was adjusted with the results.

m(SLN) = -0.40 x [(SLN).sup.2] + 1.85 x (SLN) + 133.03 [R.sup.2] = 1 (13)

The sodium chloride solution is a strong electrolyte, which increases electrical conductivity. However, the increase in the NaCl content from 3% to 5% decreased the corrosivity due to the decrease of the oxygen solubility. Gentil (20) reported that the mass loss increased with the increase of the NaCl concentration up to 3%, and the increase of the NaCl content from 3% to 26% decreases the corrosion rate. The mass loss has a minor alteration in the range from 3% to 5% NaCl, and the corrections to the reference curve were not performed.

The adjustment of the curve was performed for the number of wet-dry cycles and the result is the following equation:

m(CSU) = -0.0006 x ([CSU).sup.2] + 0.74 x (CSU) + 13.37 [R.sup.2] = 99.89% (14)

The curve of the reference test can be corrected for the other values of the number of cycles, and the curves obtained are shown in Fig. 7. The variation of the number of wet-dry cycles has great influence on the corrosion rate. The error in the calculation of curves was -5.3% for 500 h of exposure in the test, and decreased for higher testing times.

[FIGURE 6 OMITTED]

[FIGURE 7 OMITTED]

Semi-accelerated test

The curves shown in Fig. 8 represent the data obtained by using laboratory and field tests. The corrosion rate is lower for the field corrosion test. The curve of mass loss in a function of time was projected for the corrosive process in the field environment based on the reference test and the sensibility analysis, considering the alteration of temperature, relative humidity, salinity, and the number of wet-dry cycles. The curves for maximum corrosion and minimum corrosion were projected. The projection of the corrosive process as a function of the sensibility of the studied parameters generated a life span of 34.3 months. The higher difference in mass change between the obtained curve and the evaluated curve was 10.2 mg for 10 months of exposure in the field environment. Variables such as wind system, chemical pollutants, and rain effects contributed to generate the difference between the values of mass loss.

[FIGURE 8 OMITTED]

The error analysis was performed statically considering the experimental error and the error in the polynomial adjustment derived from the transformation of discrete points in a continuous function. Moreover, the error due to the other factors that influence the atmospheric corrosion, which were not taken into account in this research, were attributed to the value of the difference between the projected function and the function with real data. The partial standard errors, after combination and expansion, are [+ or -]40.6 mg, or 10.8%, in relation to the highest value of mass loss (374.9 mg for 1500 h of testing).

The life span of the painted steel was 34 [+ or -] months.

Conclusions

After exposure of steel painted with coatings based on epoxy and poly (urethane) resins for one year in a marine atmosphere, the paint system did not show blistering or failure of coating/substrate adhesion. After 1600 h of accelerated testing, the steel painted with epoxy and poly (urethane) resins did show blistering. The magnetic and akaganeite phases were identified in the corrosion product formed after the laboratory tests. The corrosion product showed amorphous and crystalline phases.

The aliphatic poly(urethane) resin and iron oxide epoxy resin showed thermal stability up to 250 [degrees] C and up to 100 [degrees] C, respectively.

The sensibility curves of the corrosion rate in relation to changes of salinity, temperature, relative humidity, and the number of wet-dry cycles were obtained. The corrosivity of the accelerated test was 16.5 times more than that of the marine atmosphere field test. The parameters that had more influence on corrosion rate were the relative humidity (52%), temperature (24.5%), and the number of wet-dry cycles (23.5%). The change of salinity in the range of 3% to 5% (w/w) did not alter the corrosion rate.

The life span of the steel painted with epoxy and poly (urethane) resins was 34 [+ or -] 4 months. The error of the estimated life span of the painted steel was [+ or -] 10.8%.

References

(1.) Katayama, H, Noda, K, Masuda, H, Nagasawa, M, Itagaki, M, Watanabe, K, "Corrosion Simulation of Carbon Steels in Atmospheric Environment." Corros. Sci., 47 (10) 2599-2606 (2005)

(2.) Lins, VFC, Paranhos, RMS, Alvarenga, EA, "Behavior or the Electrogalvanized and Painted Carbon Steel and Low Cu and Cr Carbon Steel During Cyclic an Field Corrosion Tests." J. Mater. Sci, 42 5094-5104 (2007)

(3.) Roberge, PR, Handbook of Corrosion Engineering. McGraw Hill, New York (1999)

(4.) True, TA, Hang, TTX, Oanh, VK, Dantras, E, Lacabanne, C, Oquab, D, Pebere, N, "Incorporation of an Indole-3 Butyric Acid Modified Clay in Epoxy Resin for Corrosion Protection of Carbon Steel." Surf. Coat. Technol., 202 (20) 4945-4951 (2008)

(5.) Garcia, SJ, Suay, J, "Influence on the Anticorrosive Properties of the Use of Erbium (III) Trifluoromethanesulfonate as Initiator in an Epoxy Powder Clearcoat." Corros. Sci., 49 (8) 3256-3275 (2007)

(6.) Zhang, S-Y, Ding, Y-F, Li, S-J, Luo, X-W, Zhou, W-F, "Effect of Polymeric Structure on the Corrosion Protection of Epoxy Coatings." Corros. Sci., 44 (4) 861-869 (2002)

(7.) Fedrizzi, L, Andreatta, F, Paussa, L, Deflorian, F, Maschio, S, "Heat Exchangers Corrosion Protection by Using Organic Coatings." Prog. Org. Coat., in press, available online 7 March (2008)

(8.) Gonzalez-Garcia, Y, Gonzalez, S, Souto, RM, "Electrochemical and Structural Properties of a Polyurethane Coating on Steel Substrates for Corrosion Protection." Corros. Sci., 49 (9) 3514-3526 (2007)

(9.) Sangaj, NS, Malshe, VC, "Premeability of Polymers in Protective Organic Coatings." Prog. Org. Cout., 50 28-39 (2004)

(10.) Vincke, P, "Robust Solutions and Methods in Decision Aid." J. Multicrit. Decis. Anal., 8 181-187 (1999)

(11.) Rios-Insua, D, Ruggeri, F, Robust Bayesian Analysis. Springer Verlag, New York (2000)

(12.) Marco, JF, Gracia, M, Gancedo, JR, Martin-Luengo, MA, Joseph, G, "Characterization of the Corrosion Products Formed on Carbon Steel After Exposure to the Open Atmosphere in the Antartic and Easter Island." Corros. Sci., 42 753-771 (2000)

(13.) Corvo, F, Minotas, J, Delgado, J, Arroyave, C, "Changes in Atmosphere Corrosion Rate Caused by Ions Depending on Rain Ragime." Corros. Sci., 47 883-892 (2005)

(14.) Melchers, RE, "Mathematical Modeling of the Diffusion Controlled Phase in Marine Immersion Corrosion of Mild Steel." Corros. Sci., 45 923-940 (2003)

(15.) Park, JH, Lee, GD, Ooshige, H, Nishikata, A, Tsuru, T, "Monitoring of Water Uptake in Organic Coating Under Cyclic Wet-Dry condition." Corros. Sci., 45 1881-1894 (2003)

(16.) Hoffman, JD, Numerical Methods for Engineers and Scientists. Congress Cataloging-in-Publication Data, New York (1992)

(17.) Garcia, KE, Barrero, CA, Morales, AL, Greneche, JM, "Lost Iron and Iron Converted into Rust in Steels Submitted to Dry-Wet Corrosion Process." Corros. Sci., 50 (3) 763-772 (2008)

(18.) Remazeilles, C, Refait, Ph, "On the Formation of [beta]-FeOOH (Akaganeite) in Chloride-Containing Environments." Corros. Sci, 49 (2) 844-857 (2007)

(19.) Benarie, M, Lipfert, FL, "A General Corrosion Function in Terms of Atmospheric Pollutant Concentrations and Rain pH." Atmos. Environ., 20 1947-1958 (1986)

(20.) Gentil, V, Corrosao. Livros Tecnicos e Cientificos Editora, Rio de Janeiro (2003)

E. M. Oliveira

Production Engineering Department, Toshiba Transmissao e Distribuicao do Brasil Ltda, km 1,5, Fernao Dias highway, 32240-090 Contagem, Minas Gerais, Brazil

e-mail: [e-mail]

J. R. G. Carneiro

Mechanical Engineering Department, Catholic University of Minas Gerais, Av. Dom Jose Gasper, 500, Coracao Eucaristico, 30535-610 Belo Horizonte, Minas Gerais, Brazil

e-mail: [e-mail]

V. F. Cunha Lins

Corrosion and Surface Engineering Laboratory, Chemical Engineering Department, Engineering School, Federal University of Minas Gerais, 35 Espirito Santo Street, 30160-030 Belo Horizonte, Minas Gerais, Brazil

e-mail: [e-mail]

[C] FSCT and OCCA 2008