Abstract
There is great interest in the steel industry of incorporating new products that go beyond the properties of existing ones. Among these properties, corrosion resistance is extremely important for countries that, like Chile, have an extensive ocean coastline. In costal zones, the chloride ions in air produce corrosion that reduces the service life of structures. For this reason, it is of utmost importance to study the influence of different alloying elements such as nickel, which lead to improve steel's resistance to marine corrosion. In this context the development of new types of steel is hindered if the evaluation of their corrosion resistance takes very long times. This paper presents a methodology based on accelerated wet–dry cycle corrosion experiences to simulate the behaviour of steel over time in a marine environment. The results of the proposed methodology allow adequate prediction of corrosion thickness in ASTM A242 and A588 steels exposed for years to a specific marine environment.
Introduction
Structural weathering steels are low alloyed steel containing elements such as Cu, Cr, Ni and Mo of no more than 3-5%1. They are highly corrosion resistant due to the generation of dense adherent oxides that prevent progress of damage produced by the corrosive environment. The protective corrosion products are composed of a double layer of oxides and oxyhydroxides with an upper layer composed mainly of lepidocrocite (γ-FeOOH), goethite (α-FeOOH), akaganeite (β-FeOOH), ferroxihite (δ-FeOOH), maghemite (γ-Fe2O3), magnetite (Fe3O4) and ferrihydrite (Fe5HO8√4H2O), while the lower layer, closest to the metal surface, is mainly goethite, with some crystalline Fe3O4 2–4. These compounds can coexist as partially crystalline and amorphous structures1. Protection capacity is directly related to the alloying elements present in steel. Nickel has been studied particularly in situations of exposure to atmospherical corrosion in a marine environment 4 , steels with higher content of this alloying element showing higher resistance to corrosion. This phenomenon is accompanied by a proportional increase in the oxide crust. However, to assess this steel adequately, years of exposure to this atmosphere are required. In this context, several studies5–12 have tried to reproduce corrosion phenomena on new steel grades, forming protective rust layers during long term exposure to wet and dry cycles. However, none of the authors correlate these accelerated tests to extrapolation validated with reality.
This study aims to correlate, at a laboratory scale, an accelerated wet–dry cycle corrosion test, with real time exposure data reported under ASTM G101-10 13 standard of ASTM A242-13 14 and A588-10 15 weathering steels in marine environment. Because the above standard is specifically referred to these two types of steel, this study is precisely focused on these materials.
Experimental
The three parts dividing this study and applied methodology are presented below.
Manufacturing studied steel
The experimental steels were manufactured at a laboratory scale to have close chemical compositions to those reported in ASTM G101-10. Melting was prepared in a 25 kg induction furnace and cast into a 100 × 100 × 320 mm billet. Chemical compositions are shown in Table 1, and compared with the aimed chemistry. The billet was forged with a hydraulic ram at a temperature of 1·100°C reaching a thickness of 17 mm. Specimens measuring 51 mm wide and 138 mm long were cut and subject to a homogenisation heat treatment at 1250°C for 30 min. Then, the specimens were hot rolled, reaching a final thickness of 5 mm (70% thickness reduction), taking care that temperature did not fall < 850°C. Temperature was controlled both by lased pyrometer and an imbedded thermocouple. Immediately after the last rolling pass, the specimens were quenched in a salt bath furnace to 550°C and held for 120 s, and then cooled to room temperature inside a furnace, this way simulating industrial production conditions 16 . The specimens were analysed for microstructure and mechanical properties and compared to the expected requirements.
Chemical analysis obtained by optical emission spectrometry
Steel used in ASTM G101 standard.
ASTM A36-12 standard (maximums values).
Accelerated corrosion cycles of studied steel
For accelerated wet–dry cycle corrosion tests on the different manufactured steel grades, 50 coupons of 50 × 30 × 5 mm in size were prepared. Similar coupons of ASTM A36-12 17 steel of common structural use were included, in order to compare them with weathering steel. Coupon preparation was carried out with abrasive paper according to ASTM G1 19 , then washed with acetone and alcohol, dried, measured for size and weighed.
The accelerated corrosion testing incorporated wet and dry cycling, totaling 24 h per cycle, as follows:
3 h residence in salt spray chamber, as per ASTM B117-11
18
2 h drying 17 h residence in salt spray chamber, as per ASTM B117-11, followed by strong water washing 2 h drying.
Both drying time and the strong water washing stages are fundamental, as they simulate real time weathering action on the steel, including environmental erosion, lowering NaCl concentration considerably and enabling the protective patina to form.
The specimens were subject to 40 cycles (960 h), extracting coupons every five cycles. The thickness was calculated from weight loss (gravimetric method) and corrosion rates through equation (1), as specified by ASTM G1.
19
The extracted coupons were oxide stripped with hydrochloric 33% v/v acid solution, using an industrial inhibitor (HENKEL RODINE 185) for a 20 min time lapse to assure complete oxide removal. Trials for longer times showed additional removal off < 1%.
Corrosion products were analysed through X-ray diffraction (XRD). The oxides were removed and pulverised. The XRD analysis was carried out with Cu K α radiation in Theta-2Theta mode, with a scanning speed of 2° min− 1 between 10 and 80°. The obtained difractogram was evaluated with X'Pert HighScore Plus software. Quality analysis of the corrosion products was carried out using the same software and the Rietveld method, using aluminium oxide as internal standard at 30%. 20
Evaluation methodology of results
Corrosion thicknesses were compared to results shown in ASTM G101-10 for ASTM A242-13 and A588-10 steels in marine environment, data corresponding to the coasts of Kwa Zulu, in South Africa, latitude 32°S, 16 years exposition at 300 m from the sea shore, where the average temperature and rainfall are quite similar to those in the central coastal area in Chile. This as reaffirmed by comparing the ISO Corrosivity Category of the central coastal area in Chile, Valparaiso, latitude 33°S2′39·63″ at 200 m from the sea shore is C3 21 , with data from literature1,13,22 that indicate that Kwa Zulu coastal area, in South Africa, is also C3.
The graph presented in ASTM G101-10
13
shows the kinetic equation parameters for atmospheric corrosion1, equation (2), for both of the studied steels on the Kwa Zulu coasts of South Africa. Parameters A and n are characteristic of each type of steel and the Geographic location.
The laboratory results obtained in the salt spray chamber for the studied steels, ASTM A242-13 and A588-10 are correlated with those presented in the ASTM G101-10 standard over more than a decade. If the obtained correlations between the corrosion thickness and exposure time are the same in both types of steel exposed to the marine environment, it would be possible to validate the extrapolation of accelerated corrosion tests to real time exposure to outdoor marine environment.
Results
The main results obtained during the development of this investigation are presented below.
Manufacturing of studied steel
The chemical compositions of the manufactured steel grades are quite similar to the specified steels in ASTM G101-10 and are shown in Table 1. The microstructures are, as expected, ferritic–perlitic, and illustrated in Fig. 1. Mechanical properties, shown in Table 2, comply with their respective specification standards (ASTM A242-13 and A588-10).
Micrographs of ASTM A242-13 and ASTM A588-10 steels manufactured in our laboratories
Mechanical properties
Minimum values established by the respective standards.
Study of specimens subject to accelerated wet–dry cycle corrosion test
Specimens subject to accelerated wet–dry cycle corrosion tests as described previously were evaluated for weight loss after every five cycles. Figure 2 shows corrosion thickness, with their respective standard deviations, versus cycle time. It is observed that ASTM A36-12 presents corrosion rates considerably higher than weathering steels. It is also observed that ASTM A242-13 steel presents better response to corrosion than ASTM A588-10 steel. Standard deviations are based on four samples, being always < 5%.
Corrosion thickness (μm) versus accelerated wet–dry cycling test time (hours); ASTM A242-13, ASTM A588-10 and ASTM A36-12 steels
Fig. 3 shows corrosion speed versus exposure time during the accelerated wet–dry corrosion test in the study according to the equation (1). The standard deviations are also shown. In the curves, it is clearly seen that after 400 h testing, corrosion speed tends to be constant, which is consistent with ASTM G101-13 real time exposure time results. It is also noteworthy that specimens left without wet–dry cycling, that is submitted only to salt spray chamber exposure, showed higher corrosion rates for similar wetting times.
Corrosion rate (mpy) versus accelerated wet–dry cycling test time (hours); ASTM A242-13, ASTM A588-10 and ASTM A36-12 steels
To confirm the similarity of the corrosion phenomena in the accelerated wet–dry corrosion tests presented in this work with real time corrosion, goethite formation must be clearly present in higher quantities in weathering steel than in structural steel. Table 3 shows the amount of different corrosion products after the last cycle, which were quantified through XRD, showing much higher presence of goethite (α-FeOOH) in weathering steel.
Quantitative analysis by diffraction of X-rays of corrosion species in percentages
Correlation of steel exposed to accelerated corrosion test and real time exposure
The experimental data of this work show that the best performance in marine environment is presented by ASTM A242-13 steel grade, being consistent with literature and the real time data presented in the ASTM G101-11 standard, which is shown in Fig. 4, correlating with the Kwa Zulu coasts of South Africa.
Corrosion thickness (μm) versus exposure time (years); ASTM A242-13 and ASTM A588-10 steels in South African marine environment (reproduced from ASTM G101-11)
The kinetic equations arising from ASTM G101-11 standard are as follows, where equation (3) corresponds to ASTM A242-13 steel and the equation (4) to ASTM A588-10 steel
Correlation between accelerated wet–dry cycle corrosion test time and real exposure time
Corrosion thickness experienced in the accelerated wet–dry cycle corrosion test.
Real steel exposure time (years), according to ASTM G101-11, using equations (2) and (3).
By contrasting the behaviour of the steel between cycling hours and the years of exposure in a marine environment, Fig. 5, according to the results exposed in Table 4, it is possible to reach a unique expression, equation (4), which can correlate the results of cycling hours with real time exposure of ASTM A242-13 and ASTM A588-10, where t r corresponds to real time of exposure of the steel in years and t c corresponds to the hours of exposure of the steel to the accelerated wet–dry cycle corrosion test.
Correlation of real exposure time to marine environment (years) of accelerated wet–dry cycle corrosion test time (hours) of ASTM A242-13 and ASTM A588-10 steels, using equation (5)
Discussion
While it is true that the steel used in this study is commercially available, the ranges of chemical composition specified in the ASTM standards are very broad. For this reason, in order to perform this study, it was necessary to manufacture ASTM A242-13 and ASTM A588-10 steels in the laboratory to get chemical compositions as close as possible to the steel that was specifically used in the tests corresponding to the ASTM G101-11 standard. The chemical compositions obtained are fairly similar to those aimed to, and both microstructures and mechanical properties complied well with expected values of hot rolled weathering steels.
The results obtained in the accelerated wet–dry cycle corrosion test are as expected and are consistent with literature1. Common structural steel ASTM A36-12 has high corrosion speed in comparison with weathering steel such as ASTM A242-10 and ASTM A588-13. The presence of copper in steel promotes the transformation of Lepidocrocite to Goethite, a very compact and adherent oxyhydroxide, which gives the weathering steel's qualities3. As expected the higher nickel contents of ASTM A242-13 steel improves the performance of this type of steel in marine environments 5 , as compared to ASTM A588-13 steel.
Corrosion resistance is directly related to the amount of goethite present, due to the excellent adherence of this oxyhydroxide to the base metal, thereby preventing the progress of corrosive damage. In this study, no great difference was observed in the amount of goethite among the studied weathering steel grades, while it is observed that its presence is much lower in the ASTM A36 structural steel. Although the presence of different types of corrosion products observed are not determined by the alloying elements in the steel, its proportion depends on them. This should not exclude the presence of hematite in ASTM A36-12 steel, but nevertheless, its presence is minimal compared to other oxides. It is also important to note the absence of akaganeite, which is expected due to the high concentration of chlorine in the salt chamber. It should be noted that akaganeite is very brittle and not adherent, and therefore is lost from the sample during the strong water washing between cycles, as well as during the manipulation of the coupons. The presence of maghemite/magnetite is directly associated to the inner layer of the double protective layer of oxides/oxyhydroxides and is present in weathering steel and more so in ASTM A36-12 steel. Greater presence cannot be ruled out in ASTM A242-13 steel because it may be present in its amorphous condition, which also has corrosion protection capacities in this type of steel 23 .
With the real time corrosion data extracted from ASTM G101-11 standard, the equations of the corrosion kinetics of ASTM A242-13 and ASTM A588-10 steel grades (equations (3) and (4)), the constant values of A and n, are consistent with those mentioned in the literature1.
By correlating laboratory with real time exposure results, it is found that they can be represented in equation (5), which is common to both types of tested weathering steels. This verifies that the laboratory experience of accelerated wet–dry cycle corrosion test is valid to simulate the behaviour in marine environment for >20 years of exposure of weathering steel.
Having validated this test, the development of new steel grades may be undertaken, with the advantage that in a short experimental laboratory time term, real time predictions of corrosion behaviour may be established.
Conclusions
A methodology capable of predicting the corrosion behaviour in two types of weathering steel when exposed to real marine environment was established. The methodology was based on extrapolating the performance of both types of steel in accelerated wet–dry cycling corrosion tests with the real behaviour obtained through the data in the ASTM G101-11 standard.
A correlation is proposed between test time and equivalent real time of exposure to marine environment, solving the problem of performing valid quantitative approaches of the corrosion behaviour of two types of steel from the results obtained in the accelerated corrosion cycles.
Although this methodology has been proposed for steel exposed on the coasts of South Africa, the same methodology may correlate the data from other parts of the world, based on the zones analysed in the ASTM G101-11 standard.
Acknowledgements
The authors wish to thank the Department of Scientific and Technological Research, DICYT, of the Universidad de Santiago, Chile, as well as the National Fund for Scientific and Technological Development FONDECYT, project no. 1120537.