The Influence of Crystallographic Texture and Niobium Stabilisation on the Corrosion Resistance of Ferritic Stainless Steel

The objective of this work is to investigate the influence of the crystallographic texture on the corrosion resistance of 16% Cr ferritic stainless steel. Samples of ASTM S43000 ferritic stainless steel, both niobium-stabilised and non-stabilised, were used. The samples were subjected to crystallographic characterisation (EBSD) and analysed using an inverse pole figure (IPF) and a crystalline orientation distribution function (CODF). The samples also underwent anodic potentiodynamic polarisation tests (deaeration by high-purity argon gas) in 3.56% NaCl and 1N H2SO4 solutions, and the surface was examined by SEM after the tests. The results showed a clear influence of the crystallographic texture on the corrosion resistance. The niobium decreases the amount of the preferred orientation and thus the influence of the texture on the corrosion resistance, although the corrosion resistance of the stainless steel is increased as niobium carbides are formed.


Introduction
The global production of stainless steel in 2012 reached approximately 38.1 million tons, which is almost twice as much as ten years ago 1 .This fact justifies the need to obtain more information about this material, and its increasing production calls for the need to investigate more economical solutions.The demand for stainless steel continues to grow, and this includes the demand for improvements in the material.This evolution can only succeed if the phenomena (corrosion, abrasion, etc) affecting the properties of stainless steel and the interrelations between said phenomena are well-understood.The main feature of stainless steel is its high corrosion resistance, although it is not completely resistant to corrosion.Thus, it is quite important to obtain a wider knowledge about corrosion in stainless steel and the interactions (synergy) of different phenomena, for example, tribological phenomena.As a result, improvements in stainless steel research can help spur further studies.
Previous studies reported that after the hot and the cold rolling process stainless steel showed preferential crystallographic textures [2][3][4][5][6][7][8] .In particular, Raabe and Lücke, 4 showed that ferritic steels, non-stabilised or stabilised with Nb or Ti, had a texture gradient through the thickness of the samples.It was observed that the magnitude of the Goss shear texture varied with the position through the thickness.
In the same way, the texture oriented along {111} <112> is affected by recrystallization in Cr17Nb alloys 5 .It is believed that this orientation, {111} <112>, occurs favourably during the recrystallization as a result of nucleation associated with particles of M 23 C 6 , Nb(CN), TiC or TiN, and that this texture is favourable to deep drawing 4 .A more recent study showed that, for the ferritic steel AISI 430 with different Nb content, new crystallographic texture component appear and these are attributed to the formation of coarse-grains when niobium content varies 9 .
By relating these textures to the main characteristic of stainless steel, that is, the corrosion resistance, one realizes that a high packing density of the crystallographic planes and orientations provides a better resistance against chemical attack and an improved passivation resistance and repassivation characteristics 10 .Some early studies [11][12] analysed the influence of structural texture of carbon, low-alloyed and high alloyed steel rods in the anisotropy of corrosion resistance and showed a strong dependence by the orientation of metal fibres, and recent studies reinforce this fact: the influence of fibre orientations on corrosion 13,14 .More recently, it has been concluded that the orientations <111> and <110> show a better corrosion and pitting resistance because of their high atomic density, and conversely, one expects to find worse corrosion and pitting resistance associated with orientations showing a lower atomic density [10][11][12][13][14][15][16] .It was also concluded 17 that the crystallographic texture is the main cause of the anisotropy of electrochemical corrosion in hot rolled rods.This was attributed to different rates of dissolution of the crystals oriented with the {100} and {110} planes parallel to the surfaces of the samples cut out in the directions perpendicular to the axis of the rods and along this axis.
Since ferritic stainless steel is a more economical choice than austenitic and duplex steels, this work investigates the a Laboratório de Tribologia e Materiais, Universidade Federal de Uberlândia, Uberlândia, MG, Brazil influence of the crystallographic texture on the corrosion resistance of hot rolled ferritic stainless steel ASTM S43000, both non-stabilised and stabilised with niobium.

Materials and Methods
Two ferritic stainless steels were selected for this work: S43000 with and without niobium stabilisation (430A without Nb which presents lower cost than most austenitic stainless steels and is mainly used to the cutlery sector and 430E with Nb, which is mainly employed in the cutlery and stamping sectors).
All the stainless steel specimens were tested after industrial hot rolling and annealing (Table 1).Their chemical composition, which was evaluated by different techniques (infrared absorption, thermo conductivity, X-ray fluorescence spectrometry and optical emission spectrometry) is presented in Table 2.The sample dimensions were 25 mm × 25 mm.Three different thicknesses were taken for each of the samples (the original thickness was 4 mm) to obtain the preferential crystallographic textures as they differed through the sample thickness (Figure 1).The thicknesses were: a) 100% of the original thickness (4.0 mm), b) 50% of the original thickness (2.0 mm), and c) 90% of the original thickness 3.6 mm.The 2.0 and 3.6 mm samples were produced through conventional machining of the original sample (4.0 mm).In order to know if the conventional machining affected the sub superficial microstructure of 3.6mm and 2mm samples, cross-sectioned samples were observed after conventional machining.It was possible to determine that the microstructure was altered to a depth of less than 40 μm.In order to remove the layer altered by the machining, after machining, the samples were sanded (sandpaper # 220).By measuring 10 samples before and after sanding it was possible to determine that this process removed about 55.5 ± 4 μm i.e. more than the depth affected by the machining process, thus ensuring that the machining did not influence the results of the present work.
The samples had their surface sanded with #120, #220, #320 and #600 sandpaper.The crystallographic texture analyses were performed on a conventional "Electron Back Scattered Diffraction" (EBSD) attached to an SEM, with LaB 6 filaments (PHILIPS, XL-30).The data obtained by EBSD were analysed using the software OIM Analysis ® .
The samples for this analysis were polished, in addition to being sanded, by a DP-Plancom polishing cloth immersed in a solution of 9.3 and 1 μm diamond abrasives on a Struers Abramin-Bcom polishing machine.The final stage consisted of polishing using colloidal silica for approximately 2700 seconds in a Minimet 1000B polishing machine.The data were obtained for the longitudinal section through the thickness for the two stainless steel samples with 16% Cr, using an accelerated beam of 20 kV and a step range between 1 and 2 μm.
A qualitative description was obtained through an inverse pole figure (IPF), whereas a crystalline orientation distribution function (CODF) was used to provide a quantitative description.The textures produced by CODF will be grouped into fibre textures and shear textures.Fibre and shear textures are common in operations such as rolling and drawing and are produced by the axially symmetric deformation.These textures generally have rotational symmetry, and are parallel or almost parallel to the axis of deformation, these textures are called fibre textures [18][19][20] .On the opposite, the shear textures are perpendicular to the axis of rotation.
An SP150 potentiostat manufactured by BioLogic was used for the potentiodynamic polarisation experiments.Data acquisition and data processing were accomplished using the software EC Lab ® V10.18 and EC Lab Express ® V5.4.The tests were performed in a BioLogic flat electrochemical cell (EL-FLAT 3).An 80 mesh platinum gauze counter electrode was used, and a saturated calomel electrode was chosen as a reference electrode.Two different solutions were used: a solution composed of distilled water and NaCl at 3.56% by weight and 1N H 2 SO 4 in distilled water.All parameters and methodologies applied in the tests were based on ASTM G5-94 standards 21 : the ratio of the potential variation (E) of 50 mV per 5 min (0.167 mV/s), deaeration with highly pure argon at a flow rate of 150 cm 3 /min during the entire test, and a 45 min deaeration before the potential was applied.The tests were started at a potential 50 mV lower than the stabilised open circuit potential (OCP).
After the potentiodynamic tests, the samples were analysed by SEM as previously described.

Characterisation of the crystallographic texture
The inverse pole figures obtained for the EBSD texture analyses are given in Figure 2. It can be seen that the 430A samples show a larger texture gradient through their thickness compared to the 430E samples.On the surface, the preferential orientations are <101> and <111>, whereas in the centre, they are <001> and <111>.It was also inferred that the 430E steel has larger grains that the 430A steel.Figure 3 shows the distribution of mean grain size through the thickness of the steel samples 430A and 430E.The 430A steel has a smaller grain size average through the thickness compared to steel 430E.Furthermore, the 430A steel has a minimum variation of grain size while 430E steel has a greater variation.These differences could be explained by the Nb addition in steel 430E.Niobium additions affect the uniform grain growth after recrystallization and hamper the grain refinement 9 .
For a quantitative description of the crystalline orientations, a crystalline orientation distribution function (CODF) was used.The quantification of the preferential textures are presented in Figure 4 for the 430A samples and in Figure 5 for the 430E samples.In some thicknesses the sum of the  Comparing Figure 4 and 5, it can be seen that the 430A and 430E steels have similar textures in all thicknesses (4 mm, 3.6 mm and 2 mm), i.e., the predominant fibre and shear textures are the same for the two steels and the intensities of these textures are very close (5% of maximum variation).It is also observed that in the centre of the plate (2 mm) there are no relevant shear textures, all have intensities lower than 1%.
The 430E and 430A steels near the centre of the plate (2 mm) showed as preferential texture α fibre {100}<110>, in concordance with previous studies on the texture of ferritic (4 mm) and the bulk of the plate.For both materials the fibre textures was little affected by the position.Moreover, at the centre of the plate, no significant shear textures were observed, and the α fibre {100}<110> was predominant with about 50% intensity.As a consequence, we may suppose that variation in properties is governed by the presence of Goss texture with greater intensity at 3.6 mm.

Influence of the crystallographic texture on the corrosion resistance
To verify the influence of the crystallographic texture on corrosion of hot rolled ferritic stainless steel ASTM S43000, both non-stabilised and stabilised with niobium, the samples (4 mm, 3.6 mm and 2 mm samples taken from the same plate of each steel) were subjected to anodic potentiodynamic polarization tests.
Typical results of the anodic potentiodynamic polarization curves measured through the thickness for the 430A and 430E samples in a 3.56% NaCl solution showed that the largest corrosion resistance is associated with the position corresponding to 3,6 mm (Figure 6).For clarity, the pitting potentials were evaluated and are given in Figure 7. stainless steels [5][6]8 . Inaddition, in the surface (4 mm) and 3.6 mm of thickness, the fibre textures ζ {001} <110> [5][6]8 and ε {010} <110> was predominant for both materials with similar intensities (between 22% and 27% of intensity).The principal shear texture in 3.6 mm was Goss {011} <100> with intensity 4.9% for the 430A and 5.5% for the 430E.It can also be observed that the shear texture brass {011} <211> is more intense closer to the surface (4 mm).This is due to the recrystallisation taking place preferentially on the metal surface and can be explained by the plastic deformation generated during this phenomenon.This deformation is influenced by the presence of alloying elements such as niobium [4][5]9 , and explain the slightly high intensity of the brass texture on the steel 430E (5%) compared on the steel 430A (4%).The copper shear texture {112} <111> was presented in similar intensity on the surface (4 mm) and 3.6 mm, with approximately 2.50% for the 430A and 3.35% for the 430E.The fibre ε {010} <110> was presented in the centre of the samples (2 mm) but showed half intensity when compared to the fibre α {100} <110> also presented in this thickness.
The Goss texture also prevails in the 430A steel, but its highest intensity lies at the thickness of 3.2 mm (65% from the centre to the surface), although at 3.6 mm of the thickness, the values are also remarkable, which agrees with previously reported results 4,8 .The difference in the intensity through the thickness may be caused by the stabilisation with niobium [4][5]9 and also by the grain size 9 .
Summarizing, the shear texture Goss {011} <100> is the most important texture variation between the surface The 430E steel presented higher pitting resistance than the 430A steel.The carbon content of the 430A steel is higher than that of the 430E steel, and in this way, it is more susceptible to the formation of Cr carbides, mainly in grain boundaries.In addition, the 430E steel is Nb stabilized, i.e., the carbon The variation of grain size through the thickness (S n ) varied only 0.02 for each steel (430A and 430E).According to Gollapudi 28 this variation has little influence when Sn varies between 0 and 0.2.In addition, the grain size variation between the thicknesses was 4 μm (4x10 3 nm) maximum.According to the literature [28][29][30] these grain sizes have similar electrochemical behaviour.On the other hand, it is reasonable to suppose that the difference in grain sizes, between 430A and 430E steels, would have an influence on the result of potentiodynamic polarization; Aghuy et al 30 showed that grain size does not have a significant influence on the potential of pitting in stainless steel using the same electrolytic solution as in the present study (3.5% NaCl).These authors also stated that grain refinement decreases the frequency of occurrence of metastable pits (phase prior to pitting), that is, it decreases the probability of pitting apparition.Another study 31 using the same medium (NaCl) showed that the grain size does not appear to influence the pitting potential in pure Al.Accordingly, it is believed that the difference in grain size between 430A and 430E steels had no influence on the result of pitting potential in the present study.
By observing the EBSD texture analysis (Figure 4, Figure 5 and Figure 6), one realises that the orientation of the fibre textures <110> (α, ζ and ε fibres) was predominant for both steels (430A and 430E) through the thickness, whereas the shear texture (Goss) with orientation <100> appears predominantly at 3.6 mm.It is reasonable to suppose that the presence of the orientation <100> in the shear texture is at the origin of the improvement of the pitting resistance in ferritic steels which has higher atomic density intrinsic of the FCC systems (ferrite) 15,[32][33] .Previous studies 15,34 showed that the lower susceptibility to pits generation is related with increasing atomic density planes of the FCC system, this susceptibility decreasing in the following order: 110 > 100 > 111.The crystallographic planes with a high number of nearest neighbour atoms requires a higher total energy for the breaking of the bonds and the subsequent dissolution of atoms 35 .
A comparison between the behaviour on the surface (4 mm) and at the centre (2 mm) of the samples allows one to notice a slight difference in the pitting potential (approximately 15 mV).The surface (4 mm) seems to have a better pitting resistance.By comparing the orientations in both regions, one observes that in addition to the fibre textures oriented along <110>, the surface also exhibits a Brass shear texture (oriented along <211>), which, as previously stated, arises from the recrystallization in ferritic steels.The orientation <211> in the brass shear texture could provide some corrosion resistance, which is supported by this orientation also being found at 3.6 mm and having the second highest intensity after the Goss shear texture.In addition to this, it can be inferred that the fibres with a direction {100} and orientation <110> (α fibre), which are more intense at the centre of the sample, will preferentially form Nb carbide instead of Cr carbides thus decreasing the Cr depletion in the matrix 22 .Higher Cr free content in stainless steel benefits the formation of more stable passive films on the steel surface, which prevents the penetration of chloride ions and sulphate ions.Consequently, Cr enhances the pitting corrosion resistance and uniform corrosion resistance 23 .Moreover, Nb contributes to the corrosion resistance when it is added to the alloy [24][25][26][27] , the additions of this element on ferritic stainless steels changes the characteristics of the surface film of oxide semiconductor caused by Nb 5+ incorporation into the passive layer, this fact shifts the current to a more noble, and resulting in increased pitting corrosion resistance 24,27 .
The pitting potential increases at 3.6 mm and decreases at the centre of the sample (2 mm), i.e., the corrosion resistance is reduced at the centre.Figure 7 shows the influence of the crystallographic texture on the pitting potential, because the crystallographic texture changes throughout the thickness of steel in the same way as the pitting potential, and demonstrating that the pitting phenomenon has an slightly anisotropic behaviour [11][12][13][14][15] .This anisotropic behaviour occurs because carbon segregation 22 is more intense in the centre of the slab in non-stabilised steel.As the P430E steel is niobium-stabilised, this effect was not observed.In both steel types, the effect of the crystallographic texture alone can be clearly identified.
In addition, the distribution of grain size (Table 3) had almost no effect on pitting.tend to lower the pitting corrosion resistance because this location has the lowest pitting potentials for each of the steels.Figure 8 presents typical results of the anodic potentiodynamic polarisation through the thickness, in a solution of 1N H 2 SO 4 for both steel types 430A and 430E.In this environment, the steels have an active-passive behaviour, whereas in the 3.56% NaCl solution (Figure 6) the steels were completely passive.There is no large difference through the thickness for both steel types.
The evolution with thickness is quite similar to that seen for the pitting potential (Figure 7), i.e., the best corrosion behaviour occurs at 3.6 mm, where the lowest current density was found.As previously stated, at 3.6 mm, the orientation <100> of the Goss shear texture is the main difference compared to the other positions, i.e., the orientation <100> that occurs in the shear texture has a positive influence on i min .On the other hand, the minimum passive current was presented by the 430A steel that is not stabilized.This result contradicts a previous study for a steel with the Nb addition in an H 2 SO 4 environment which suggested for potentiodynamic and potentiostatic studies a higher Nb content decreases the passive current density within the corrosive environment 36 .It is believed that this behaviour is not related to the crystallographic textures or alloying elements and that it is produced by the variation of grain size.More refined grains can help the uniform passive layer grow, due to the higher free energy associated with grain boundaries 25,[37][38] .In addition grain size distribution (S n ) has also effects: the passive layer may vary from being very uniform and compact at some locations to being very non-uniform at other locations if the grain size distribution (S n ) is large 25 .
Table 4 presents a summary of the variation of texture through the thickness and its influence on the corrosion for 430A and 430E steels by comparing the higher intensity textures with the pitting potential and minimum passive current (i min ).
Figure 10 shows typical surfaces after the anodic potentiodynamic polarisation tests in 1N H 2 SO 4 .It is clear that, in the 430E steel samples, the severity of corrosion allowed a clean visualisation of niobium carbides, whereas in the 430A steel samples, intergranular corrosion is more prevalent, which agrees well with previous results regarding corrosion phenomena in 430E and 430A steels.
Niobium carbide precipitates were observed throughout the surface of the 430E steel samples as white spots.At 4 mm and 2 mm, a weak granular corrosion was identified (Figure 10 (a) and (e)), as well as a larger corrosion at the borders of the carbide, a phenomenon that was not visible at 3.6 mm (Figure 10 (c)).At this position, in addition to a small amount of intergranular corrosion, the slight presence of pits was observed, but it cannot be concluded that the corrosion is significant there.In the 430A steel samples, the intergranular corrosion was remarkable through the thickness, although on the surface (at 4 mm), some pits could be observed inside the grains (Figure 10 (b)).Compared to centre of the samples, i.e., at 2 mm (Figure 10 (f)), a large amount of intergranular corrosion was observed, as well as a generalised degradation on the surface that can be considered generalised corrosion.The average result of the minimum passivation current (i min ) in 1N H 2 SO 4 was calculated and the results are showed in the Figure 9.

Concluding Remarks
The local corrosion (pitting) exhibits an anisotropic behaviour in ferritic stainless steel and is influenced by the crystallographic texture, particularly with respect to the pitting potential.This anisotropic behaviour has shown that the presence of the Goss shear texture (oriented along <100>) has a positive influence on the corrosion resistance, whereas the fibre α {100} <110> has a negative influence on the pitting potential.The pitting potential through the thickness showed a trend in the region of the highest Goss texture intensity (orientation <100>), where it was worse near the centre of the sample.These results would be questionable if only the P430A steel (unstabilised) was used.
Nb stabilisation in stainless steel helps to protect against intergranular corrosion, as the grain borders becomes less reactive.This becomes more evident when comparing the P430A and P430E steels, the second one being more resistant to intergranular corrosion because of the Nb stabilisation.This element reduces the presence of preferential orientations, and thus the influence of the texture on the corrosion resistance.However, it increases the corrosion resistance in stainless steel by giving rise to niobium carbides.

Figure 1 .
Figure 1.Schematic showing the different thicknesses and the corrosion test surface location of the samples in reference to the original thickness of the plate.

Figure 3 .
Figure 3. Distribution of mean grain size through the thickness of the samples.differentcomponents adds up to more than 100%; this is because some textures are very close to each other and the software takes them into account twice; this is a recurring problem with this technique19 .

Figure 4 .
Figure 4. Texture intensity through the thickness of the 430A steel.(a) Fibre texture and (b) shear texture.

Figure 5 .
Figure 5. Texture intensity through the thickness of the 430E steel.(a) Fibre texture and (b) shear texture.

Figure 7 .
Figure 7. Influence of slab thickness on the pitting potential.

Figure 9 .
Figure 9. Minimum passivation current density (i min ) through the thickness in 1N H 2 SO 4 .

Table 1 .
Industrial hot rolling and annealing conditions.

Table 2 .
Chemical compositions of the specimens in weight percent, via X-ray fluorescence spectrometry.

Table 3 .
The approximate grain size and the standard deviation in a weighted grain size distribution (S n ) for the 430A and 430E steels.The Influence of Crystallographic Texture and Niobium Stabilisation on the Corrosion Resistance of Ferritic Stainless Steel

Table 4 .
Summary of the texture variation through the thickness for the P430A and P430E steels samples and their pitting, corrosion and passivation behaviour.