Comparison of the mechanical response of B400c and B450c dual phase steel bar categories, in long terms

In this work, the effects of chloride-induced corrosion on tempcore B400c and B450c steel grades are evaluated, in terms of corrosion resistance and mechanical characteristics, after the performance of Tensile and Low Cycle Fatigue Tests. Both steel categories, characterized by high strength and high ductility, have been used in existing structures, indicating that they demonstrate different performance against the ascribed corrosion environments. B450c steel grade seems to preserve higher energy reserves, ensuring higher expectancy to the corresponding reinforced concrete structures. Additionally, due to buckling and buckling reversal, both steel grades demonstrate limited ductility at 4%. Finally, when cyclically stresses occur, crack nucleation is taking place, at the areas where sulphides, FeS and MnS can be found, leading to sub-surface crack propagation, interacting with external pits.


INTRODUCTION
einforced concrete has been the most popular material for more than a century. During these years, in an effort to upgrade the reliability of the structures, various steel grades, characterized by high strength and ductility, have been used as reinforcement. However, there is a plethora of degradation factors that may affect the performance and the reliability, not only of the material, but also of the whole structure [1]. Corrosion of steel reinforcement constitutes a major problem for reinforced concrete structures, as far as their durability is concerned, as it has also been mentioned by Almusallam [2]. Corrosion degradation is responsible for several issues, such as deterioration of both durability and service life of structures, resulting in premature failure. Ιn recent years, the problem of the actual residual strength degradation of ageing reinforced concrete structures has attracted considerable attention, however it is far from being fully understood and even less resolved. According to Apostolopoulos et al. [3], Papadakis [4], Capazucca [5], Diamond [6] and Alvarez et al. [7], this scenario becomes even worse when coastal environment, which is rich in chlorides, is combined with seismic phenomena. Cyclic loading, which is due to seismic activity, in conjunction with the pre-existing downgrade, which results from corrosion effect, leads to prompt deterioration of useful lifespan of structures, that face durability issues. Furthermore, cyclic loading, leads to a nonlinear response of the structures that face durability issues. Therefore, to predict the time dependent damage of corroded reinforced concrete structures under seismic loading it is necessary to consider the effects of material degradation, due to corrosion [8].
In the present work, the effects of chloride-induced corrosion on two different steel bar categories, B400c and B450c are evaluated, in terms of corrosion resistance and mechanical characteristics, before and after the corrosion process. The method that was used for the accelerated corrosion tests, was salt spray chamber and the mechanical tests performed for both steel bar categories were tensile tests and Low Cycle Fatigue Tests. The goal of the study was to highlight the differences among the two steel categories, as far as their mechanical performance and their resistance to corrosion is concerned. Given that both steel grades have been used in existing civil engineering structures, it is important to gain knowledge of their mechanical behavior and their vulnerability to corrosion factor, so as to be able to estimate their life expectancy.

MATERIALS AND METHODS
or the goals of the present study, material of two different steel reinforcement categories B400c and B450c was used. The material, for both categories, was received from European factories and were produced by the same steel manufacturer using the "tempcore" method. The chemical composition, which is approximately the same for the two different steel bar categories. Specifically, both steel categories contain 0.22% C, 0.05% P, 0.05% S, 0.8% Cu and 0.012% N, while B450c has an additional 0.5% Ceq. Two series of specimens, 500mm and 250mm long, of 16mm nominal diameter, were organized for each steel grade. For comparison reasons, one more set of specimens of nominal diameter 12mm and 500mm long was prepared for B450c. Given that the goal of the present study is to evaluate the mechanical performance of both steel categories, before and after corrosion, salt spray chamber method was used for the performance of the corrosion tests. Exposure periods equal to 0 days, 45 days and 90 days were scheduled for each set of specimens. Consequently, each set was separated into three smaller groups, the first of which was only mechanically tested (Reference Specimens) while for the rest two groups corrosion took place prior to the mechanical tests. Before the corrosion process, specimens were properly prepared. They were primarily cleaned in order to remove all the unwanted impurities from their surface. The configuration of equal surface areas, that would be exposed to the corrosive media, was necessary for comparison reasons. The exposed length for each specimen was equal to 20mm, while the rest part of the metal was covered with wax, to prevent corrosion (Fig.1). The samples were exposed to the corrosion process and they were mechanically tested in Tensile and Low Cycle Fatigue (LCF) tests. Artificial corrosion tests were executed with the use of salt spray fog chamber -according to ASTM B117 [9].  or the corrosion process all the specimens were placed in a salt spray chamber (Fig.2). The corrosion tests were organized in accordance to ASTM B117-94 specification (directly exposed to the corrosive medium). The ASTM B117 [9] specification covers every aspect of the apparatus configuration, procedure and conditions required to create and maintain a salt spray (fog) testing environment. The selection of such a procedure for corroding the specimens, relies on the fact that the salt spray environment lies qualitatively closer to the natural coastal (rich in chlorides) conditions than any other accelerate laboratory corrosion test. In principle, the testing apparatus consists of a closed chamber in which a salted solution atomized by means of a nozzle, produces a corrosive environment of dense saline fog. In this particular study a special apparatus, model SF 450 (mode by Cand W. Specialist Equipment Ltd) was used. The salt solution was prepared by dissolving 5 parts by mass of sodium chloride (NaCl) into 95 parts of distilled water (pH range 6.5-7.2). The temperature inside the salt spray chamber was maintained at 35°C (+1.1-1.7°C). Additionally, in the present study, a severe exposure environment of wetting/drying (chloride ponding) was organized in order to achieve a better approach to the environmental conditions and to simulate the chloride exposures of marine structures under splash and tidal zones. Hence, it is widely known that that in reality, structures are subjected to wet and dry periods, rather than a constant relative humidity [10]. Chloride ponding application is in agreement with the simulating corrosion methods used in existing studies as well [11][12][13] and ensures conditions which are qualitatively closer to the natural. This is because during wetting, chloride solution penetrates a layer of the material; during the drying stage the evaporation front moves inwards and takes some of the chloride with it [10]. It is deducted in theory that atmospheric corrosion rate of metals can be accelerated by increasing the frequency of wet-dry cycling [12]. Consequently, the ponding cycles organized consisted of alternating 1.5 hour of exposure to wet conditions and 1.5 hour of dry conditions. Eight ponding cycles of wet/dry conditions were scheduled per day. By the end of each exposure time, specimens were removed from the corrosive environment, washed with clean running water, to remove ant salt deposits from their surfaces and air dried. The corrosion products were removed from the surface of the specimen by means of a brittle brush, according to ASTM G1 specification [14]. The specimens were then weighted and the mass loss due to corrosion exposure was calculated as: Where m 0 is the mass of uncorroded specimens and m c the reduced mass of the corroded specimens.

MECHANICAL TESTS
fter the completion of the corrosion process the mechanical tests were organized. In Table 1 is given in detail the number of the specimens used in each case. According to Table 1, a total of 18 tensile tests and 134 Low Cycle Fatigue Tests was conducted for the goals of the present study. The differentiated number of the LCF tests, per strain, per steel category is owed to the fact that there were some additional tests that were invalid, the results of which are not included in the present paper.    On the other hand, it is worth mentioning that LCF tests procedure for the cyclic mechanical behavior of steel bar is not included in the European production standards for reinforcements (EN 10080:2005). Only Spanish and Portuguese standards prescribe the execution of symmetrical tension/compression cycles for the production control of steel reinforcements, while the draft of new European standard for reinforcements (prEN 10080:2012) gives only some indications for the execution of LCF tests by Caprilli et al. [16][17].
In the present study, the LCF tests were performed using the above-mentioned servo-hydraulic MTS 250 kN machine. Two free lengths, equal to 6D and 8D respectively, were tested for each steel category. The frequency used was equal to 2.5Hz and strain imposed was equal to ±2.5% and ±4%, in reference to their free length (either 6D or 8D).

RESULTS AND DISCUSSION
lthough both B400c and B450c steel bar categories have been used in existing structures, several worth mentioning differences can be reported, as far as their mechanical performance and their corrosion resistance are concerned, in long term. In Tables 3 and 4 are presented the results concerning both the mass loss percentages and the mechanical properties of the long specimens (500mm), of 16mm nominal diameter. Taking into consideration the data of Tables 3 and 4, it is evident that both steel grades demonstrate similar vulnerability to the corrosion factor. As far as the mechanical characteristics are concerned, B400c recorded 6.5% drop of yield strength against 5.7% of B450c, after the 90 days of A exposure to the corrosive environment. On the contrary, for the same exposure period, elongation of B450c steel was diminished about 77.12% against 69.87% of B400c category. However, it is remarkable, that plastic deformation recorded for both steel categories, after the completion of the first 45 days, was lower than the value that Eurocode 2 requires, which is equal to 7.5%. The corresponding results of LCF tests are presented in Figs.3 Figs.3 and 4, it seems that although corrosion factor is responsible for similar mass loss percentages in both steel categories, however they demonstrate different performance under the cyclic testing. This fact is owed to the different thickness of the martensite layer of the two steel grades [18]. Precisely, B400c steel bar category recorded a great drop, of more than 60%, on the number of cycles up to failure until the first 45 days of exposure to the aggressive conditions, for both deformations imposed (Δε= ±2.5% and Δε= ±4%) and for both free lengths defined (6D and 8D). Whereas, consistency was recorded for the next 45 days, up to the completion of the exposure period of 90 days. On the other hand, B450c steel grade, recorded an initial drop on its cyclic life, the mean value of which was higher than 50% for both deformations (Δε= ±2.5% and Δε= ±4%) and for both free lengths defined (6D and 8D) and an extra dron higher than 12.5% up to the 90 days, for Δε= ±2.5% (for both free lengths 6D and 8D). Additionally, due to buckling and buckling reversal, the material's ductility especially at 4% is particularly limited. Explicitly, the free length of the specimen strongly influences the results of the experimental tests, since buckling phenomena due to compression axial loads can lead to premature unexpected failures of the rebars (Fig.5); in order to avoid this problem, actual standards for reinforced concrete constructions (EN 1998-1:2005, D.M. 14/01/2008) prescribe the adoption of opportune limits for the free length of reinforcements between stirrups, that shall be lower than 6 or 8 times the diameter for buildings respectively designed in high, medium or low ductility class. For comparison reasons, a few more LCF tests were performed on reference B450c steel bar specimens, of 12mm nominal diameter. The corresponding results are presented in Fig.6.

Energy stock-mass loss-6D
Ncycles-mass loss-6D Energy stock-mass loss-8D Ncycles-mass loss-8D Figure 6: A graphical depiction of energy stocks and number of cycles of the B450c (Φ12) specimens tested in LCF, in reference to their corresponding mass loss, before and after their exposure to corrosive conditions Comparing the results of Fig.4, with the corresponding results of Fig.6, where the same steel category-but with different nominal diameter-was used, it seems that specimens with 12mm diameter record lower mass loss percentages. This is owed to the fact that corrosion process, that is conducted in the salt spray chamber, is a method that mainly affects the external surface of the material. In the case of 16mm nominal diameter, the spread of the exposed surface is greater than in the case of 12mm, and it is a fact that results in a greater surface attack. This remark, in combination with the existing knowledge, related to the differential aeration corrosion phenomenon, can sufficiently explain that the rate of the electrons flow-and therefore the corrosion rate-depend on the volume of the exposed material as well as the volume of the protected (not exposed to the aggressive conditions) part of the steel reinforcement.
Additionally, specimens of 12mm nominal diameter record higher performance under LCF testing, than the corresponding specimens of 16mm nominal diameter. This annotation is closely associated with the lower corrosion damage (e.g. mass loss rate or/ and pitting severity) recorded.
Combining the results of the tensile and the LCF tests, a correlation with real structures can take place. Analytically, it is a fact, that during strong earthquakes, yielding structures are subjected to increased number of cycles into the inelastic range and the accumulated damage may significantly affect their overall performance. However, little attention has been devoted by the research community on the combined effect of corrosion and LCF on steel reinforcement, since each one of these factors affects the rebar durability and performance and shortens the life expectancy of structures [19][20][21][22][23][24]. It is known though, that besides external overt damage, more internal and subcutaneous phenomena are realized. At those areas, plethora of MnS and FeS compounds (potential sites of corrosion initiation or for the formation of corrosion paths) can be detected. Synergy of both internal and external corrosion factor, may result in premature and unexpected failure of the material [25].
In an effort to further investigate the degradation phenomena, which are related to the internal downgrading of the material, several SEM and EDX analyses took place. Prior to tensile testing, the cross sections of reference rebars were examined under scanning electron microscope, after having been polished to surface roughness of Ra < 0.8 μm.  Figs.9 and 10 illustrate the SEM and EDX analyses in a corroded B400C, Φ16 sample, after tensile test. Moreover, through X rays' diffusion, the appearance of pores was obvious at the external martensitic surface and at the same area high sulfide concentration was noticed. This fact increased the requirements for damage expansion at the inner area of the external surface (skin). Fig.11, which has been taken from study [26], that processes on the same rebar group, presents a SEM analysis of a longitudinal cut in the fracture region in a non -corroded B450c specimen after tensile test. Figure 9: SEM analysis in a corroded B400C, Φ16 sample, after tensile test.  Figs.12 and 13 present a SEM and an EDX analysis of a longitudinal cut in the fracture region in a noncorroded specimen after tensile test (B450c). According to these findings, the area depicted is rich in MnS and FeS compounds, phenomenon which gradually leads to a local decomposition, and inevitably to the failure of the material. Figure 12: SEM analysis of the fracture region in a noncorroded specimen after tensile test (B450c). It is very likely that the adsorption of Cl-ions in and around the strained regions of a MnS-Fe interface triggers the anodic dissolution of Fe2+ ions. It is important to reemphasize here that Cl-ions are not involved in redox reactions but catalyze the anodic processes by adsorbing on the surfaces around the MnS inclusions and by chasing away the conduction electrons of the strained matrix, which results in the anodic dissolution of iron from these regions as depicted in Fig.14 [27]. From a structural point of view, dissolved MnS sites and porosity represent high stress raisers, which upon loading can lead to the development of microcracks. These phenomena within the martensitic zone, near the location of surface pitting, result in local disruption of the material and a possible conjunction of extended imperfections with the surface pits. In these locations, the local disruption of the material is a crucial factor of its mechanical behavior since during the axial load an intense stress concentration is developed and the subsequent crack propagation is inevitable and rapid [8]. Rapid depletion of the ductility or even failure may occur in high strength and ductility dual phase steel bars, due to the combination of interior and exterior damage phenomena under strong stresses. Consequently, an additional purpose of this study is to highlight the significance of the corrosion factor on the mechanical performance of the dual phase steel bars and to estimate the degradation of its mechanical properties, which is the result of the cooperation of the internal and the external damage, under those hazardous circumstances [25].
The results of the mechanical tests confirmed that corrosion, due to chloride ions, is one of the main degradation factors of steel reinforcement.

CONCLUSIONS
he conclusions of the present study can be summarized as follows: • Steel bar category B450c recorded slightly higher mass loss percentages than B400c steel class, a fact that demonstrates that the first category is more vulnerable to corrosion. • In both steel categories, a slight mechanical degradation was recorded, in terms of yield and maximum strength, while ductility and dissipated energy were dramatically diminished. • Due to buckling and buckling reversal, materials demonstrate limited ductility especially at 4%. • When cyclically stresses occur, sulphides, FeS and MnS sites will host crack nucleation leading to sub-surface crack propagation, which interact with external pits. The case becomes very complex in terms of analysis, especially under fully reverse loading and high plastic strain levels leading to buckling • Among two specimens of the same steel class, the one with smaller nominal diameter records higher performance against LCF tests. • The areas with MnS compounds present a significant development under the presence of (Cl − ) ions, which combined with the occurrence of other impurities, as well as pits at the surface of steel, may inevitably induce mechanical stress concentration during the tensile loading • Therefore, the mechanical degradation of precorroded steels that were examined can result from the synergy of mass loss effect, external pitting, and a variety of inevitable side effects from regions with MnS compounds within the martensitic zone. • Examining the performance of both steel categories under LCF tests, it seems that reference B450c steel has grater life cycle and higher energy stocks than the reference B400c steel. However, when both categories record 5%-10% mass loss, under corrosion conditions, life cycles of both steel categories are approximately equal.