1-s2.0-S1044580308003410-main

M A T E R I A L S C H A R A C T E R I Z A T I O N 6 0 ( 2 0 0 9 ) 1 5 0 – 1 5 5

Microstructural characteristics and corrosion behavior of asuper duplex stainless steel casting

Marcelo Martinsa,⁎, Luiz Carlos Castelettib

aIndustrial Manager of SULZER BRASIL S/A and Professor of the São Paulo Salesian University Center (UNISAL),Americana Division, SP, BrazilbDepartment of Materials, Aeronautical and Automotive Engineering, São Carlos School of Engineering,University of São Paulo (USP), São Carlos, SP Brazil

A R T I C L E D A T A

⁎ Corresponding author. Av. Eng. João Fernand+55 11 45892102.

E-mail address: [email protected]

1044-5803/$ – see front matter © 2008 Elsevidoi:10.1016/j.matchar.2008.12.010

A B S T R A C T

Article history:Received 11 December 2007Received in revised form10 June 2008Accepted 8 December 2008

The machining of super duplex stainless steel castings is usually complicated by thedifficulty involved in maintaining the dimensional tolerances required for givenapplications. Internal stresses originating from the solidification process and fromsubsequent heat treatments reach levels that exceed the material's yield strength,promoting plastic strain. Stress relief heat treatments at 520 °C for 2 h are an interestingoption to solve this problem, but because these materials present a thermodynamicallymetastable condition, a few precautions should be taken.Themain objective of this work was to demonstrate that, after solution annealing at 1130 °Cand water quenching, stress relief at 520 °C for 2 h did not alter the duplexmicrostructure orimpair the pitting corrosion resistance of ASTM A890/A890M Grade 6A steel. This findingwas confirmed by microstructural characterization techniques, including light optical andscanning electron microscopy, and X-ray diffraction. Corrosion potential measurements insynthetic sea water containing 20,000 ppm of chloride ions were also conducted at threetemperatures: 5 °C, 25 °C and 60 °C.

© 2008 Elsevier Inc. All rights reserved.

Keywords:Sigma phaseX-ray diffractionCorrosion resistance

1. Introduction

Super duplex stainless steels (SDSS) have a two-phasestructure containing ferrite and austenite in almost equalproportions, and a pitting corrosion resistance of over 40.These materials are important for use in offshore platforms inapplications that involve pumping of produced water, i.e., seawater containing high concentrations of chloride ions (Cl−),CO2, hydrosulfuric gas (H2S), and HS− and S−2 ions, amongothers, at temperatures ranging from 40 °C to 80 °C [1].

The production of cast components for centrifugal pumpswith wall thicknesses exceeding 125 mm (5 in.) in duplex andsuper duplex stainless steels has become a complex task dueto the low cooling rates during the solidification process.

es G. Molina, 905 – Distrit

m (M. Martins).

er Inc. All rights reserved

Precipitation of intermetallic and carbide phases are commonand include the sigma phase, which appears in the highestproportion, sometimes reaching values of close to 20% [2,3].Solution annealing heat treatments followed by waterquenching dissolves these precipitates and keep the alloyingelements in solid solutions.

Another common problem with these materials is thedifficulty of maintaining the dimensional tolerances duringand after the machining process of cast components, regard-less of the size and thickness of the components [3]. Stressrelief heat treatments at temperatures close to 520 °C allowthe dimensional stability to bemaintained without promotingprecipitation of secondary phases or destroying the meta-stable equilibrium of these SDSS alloys.

o Industrial, Jundiaí 13213-080, SP, Brazil. Tel.: +55 11 45892020; fax:

.

Fig. 1 –Absorbed energy in impact test vs volumetricconcentration of sigma phase.

Table 1 – Composition of the chemical elements in superduplex stainless steel casting (wt.%)

C 0.016 Nb 0.014Cr 25.69 Ti 0.005Mo 3.80 Al) 0.016Ni 7.18 V 0.049Si 0.74 Zr 0.065Mn 0.52 Co 0.055Cu 0.716 Sn 0.007W 0.736 Pb 0.002N 0.22 S 0.008P 0.027 Fe 60.16 (by difference)

Fig. 2 –Micrographs characterizing the as cast state.Murakami reagent. Magnifications: (a) 100× and (b) 200×.

151M A T E R I A L S C H A R A C T E R I Z A T I O N 6 0 ( 2 0 0 9 ) 1 5 0 – 1 5 5

Because they are thermodynamically metastable in thesolution annealed condition, super duplex stainless steels aresubject tomicrostructural changeswhen exposed to heat, be itthrough heat treatments at low temperatures aimed atrelieving internal stresses or through the welding process.This is especially the case in the welding repair of cast com-ponents that normally present some type of foundry defectafter detection in machining operations.

The tendency of these steels to reach a thermodynamicallystable condition implies the precipitation of intermetallic and/or carbide phases rich mainly in chromium andmolybdenum.The effect of this precipitation is to form regions adjacent tothe particles that are depleted of these elements. Concentra-tion can fall below the minimum 12 wt.% to 13 wt.% of chro-mium required to ensure the passivation process [2].

Regions with quantities below the minimum Cr rangeundergo active dissolution and usually corrode at rates closeto those of carbon steels and pure iron. Stainless steels withhigh chromium content, as is the case of duplex and superduplex stainless steels, are relatively passive and corrode atmuch lower rates than do alloys with low chromium content[4].

In alloys containing chromium, the passive region isinitiated at lower potentials than in alloys that do not containthis element, such as carbon steels and pure iron. The trans-passive regions of alloys containing chromium originate fromthe easy oxidation of the chromium ion fromCr+3 to Cr+6whichforms the passive film [4].

Chloride ions cause aqueous corrosion of stainless steelsbecause they increase the electrical conductivity of thesolution and penetrate easily into the protective oxide film,thereby breaking its passivity. As the concentration of theseions increases, so does the current density, because chlorideions present high charge densities [4].

The best mechanical and chemical characteristics of superduplex stainless steels are obtained after solution heattreatments, in which all the secondary phases – carbide andintermetallic – originating from the as cast microstructure aredissolved at high temperatures. Water quenching preventsthem from precipitating again, keeping large quantities ofsolute atoms in solid solution.

One of the phases precipitated during solidification coolingis the sigma phase, which contains about 30 wt.% of chro-mium and 8 wt.% of molybdenum [3]. The presence of morethan 3% in volume of sigma phase in the microstructurereduces the impact toughness of these materials from 220 J to20 J at room temperature, Fig. 1 [3].

It is impossible to avoid the precipitation of sigma phaseduring cooling in the solidification process, in castings withthick sections (N75 mm), in regions close to feeder heads andexothermal “gloves” used for feeding during contraction andfor the components that are cooled inside silica sand molds,where thermal conductivity is very low. The volumetricfraction of this precipitate can be minimized by controllingthe chemical composition and the cooling rate in the solidstate.

The solution heat treatment itself introduces a certain levelof internal stresses in cast components, because solute atomssuch as chromium, nickel and molybdenum are larger thanthe iron atom. The bcc crystal lattice of ferrite and fcc structureof austenite are slightly deformed and this deformationstrains the entire lattice.

Another factor that contributes to increase in the level ofinternal stresses is nitrogen, which lodges in the octahedraland tetrahedral interstices of the austenite, causing micro-deformations in the crystal lattice. The use of this element insuper duplex stainless steels is justified because it providesgreater pitting corrosion resistance, because it is a strongstabilizer of austenite, contributing to the final equilibrium ofthe volumetric fraction of austenite in the microstructure.

Fig. 3 –Micrograph of the material solution annealed at1130 °C and water quenched. Beraha II reagent.Magnification: 200×.

Fig. 5 –Diffractogram of the material in the solution annealedand stress relieved condition. CuKα1, λ=1.5406 Å.

152 M A T E R I A L S C H A R A C T E R I Z A T I O N 6 0 ( 2 0 0 9 ) 1 5 0 – 1 5 5

Some alloying elements have specific functions in superduplex stainless steels, but because they are usually largerthan iron atoms, they also cause internal stresses. This phe-nomenon promotes a certain dimensional instability duringand after the machining process, especially when largequantities of materials are removed.

Oneway tominimize this effect is through stress relief heattreatments at sufficiently low temperatures to prevent theprecipitation of secondary phases. Carrying out this procedurerequires good knowledge of the metallurgical aspects of thematerial. In the solution annealed condition, these steelsconstitute thermodynamicallymetastable systems. If the heatinput is not well controlled, undesirable phases will appear inthe microstructure.

In some kinds of duplex and super duplex stainless steelsothers chemical elements are used to promote specificproperties. Among them, tungsten and copper are the mostcommons. The presence of tungsten in these materials delaysprecipitation of the sigma phase during the cooling from thesolidification process, thereby reducing the tendency forembrittlement [5].

Fig. 4 –SEM micrograph after solution annealing heattreatment at 1130 °C and stress relief at 520 °C. Beraha IIreagent. Magnification: 3000×.

On the other hand, copper in solid solution increases erosion–corrosion resistance, especially in components of centrifugalpumps where there are continuous flows of aggressive fluids [6].

Fig. 6 –Anodic polarization curves of the materials solutionannealing heat-treated and solution annealed heat-treated+stress relieved.

Fig. 7 –Variation of the pitting potential as a function of testtemperature.

153M A T E R I A L S C H A R A C T E R I Z A T I O N 6 0 ( 2 0 0 9 ) 1 5 0 – 1 5 5

2. Experimental Procedure

The steel studied here was produced in a vacuum inductionfurnace with an electric power frequency (60 Hz) and a maxi-mum power of 400 kW.

The first step consisted of drawing up the foundry design ofthe 25 mm diameter, 300 mm long test specimens, followed bysimulation of the solidificationprocess using a specific softwareprogram. The material's chemical composition was analyzedusing an optical emission spectrometer equipped with 47 dif-ferent channels.

The solution annealing heat treatments recommended forthese materials were carried out in an electric furnace with aheating capacity of up to 1300 °C, and the temperature adoptedfor this kind of heat treatment was 1130 °C. After the solutionheat treatment, several samples were stress relieved at 520 °Cfor 2 h.

X-ray diffraction analyses were conducted in a diffract-ometer coupled to a multipurpose camera, using copper(CuKα1) radiation at a wavelength of 1.5406 Å, a 50 kV source,100mAand a scanning velocity of 1°/min. The pitting potentialwas determined in a synthetic sea water solution containing20,000 ppm of chloride ions, at three different testingtemperatures: 5 °C, 25 °C and 60 °C. The reason for thesechoices is that the temperatures in deep waters (close to3000 m) reach 5 °C, and the temperature of 60 °C served tosimulate the pumping of produced sea water, which is thewater injected into the oil well to increase its pressure anddragthe crude oil from the marine subsoil up to the platform.Because the temperature of petroleum deposits ranges from140 °C to 160 °C, thewatermoved upwith the oil (which is thenseparated from it and reinjected into the well), reaches atemperature of close to 60 °C.

To obtain good resolution of the polarization curves, thescanning rate adopted was 1mV/s, and the immersion time inthe open circuit was 1 h prior to initiating the scan [7]. All thescans were initiated starting from the corrosion potentialestablished during the first hour of immersion and endedwhen they reached a current density of 1×10−2 A/cm2.

The electrochemical cell consisted of a platinum counterelectrode. A saturated calomel electrode (SCE) was used asreference. The synthetic sea water was prepared according to

Table 2 – Pitting potentials for aerated and CO2 saturatedsolutions, for solution annealed at 1130 °C and solutionannealed at 1130 °C+stress relieved at 520 °C

Sample Solution Testtemperature

(°C)

Pittingpotential

(mV)

Solution annealedat 1130 °C

Aerated 5 107060 900

CO2 saturated 5 103060 880

Solution annealedat 1130 °C+stressrelieved at 520 °C

Aerated 5 102060 910

CO2 saturated 5 107060 880

Test temperatures: 5 °C and 60 °C.

the ASTM D 1141 Standard, using prepared stock solutions forthe respective chloride ion concentrations in aerated and CO2

saturated conditions.The alloying elements that determine the pitting corrosion

properties of stainless steels are chromium, molybdenum andnitrogen. To determine quantitatively the effect of these ele-ments on the corrosion resistance of stainless steels, a pit-ting resistance equivalent, PREN coefficient was introduced.Expression 1 gives the most commonly used formula.Different multiplying factors for the nitrogen contents, vary-ing from 10 to 30, are sometimes used for different types ofstainless steels. For duplex stainless steels, for example, thefollowing formula is given in the literature [8]:

PREN = wt:k Cr½ � + 3;3ð Þ wt:k Mo½ � + 16ð Þ wt:k N½ � ð1Þ

Other alloying elements such as tungsten are also some-times taken into account [9]:

PREW = wt:k Cr½ � + 3; 3ð Þ wt:k Moþ 1=2ð Þwt:k W½ � + 16ð Þ wt:k N½ �ð2Þ

In duplex stainless steels, the pitting resistance of the twophases must be considered separately [9].

3. Results and Discusssion

The chemical composition of the material used, ASTM A890/A890M Grade 6A [10], determined by means of opticalemission spectrometry, is described in Table 1. The volumetricconcentrations of the phases, determined experimentally byhand counting on a grid of points following the ASTM E562standard [11], were 43% and 57%, respectively, for the ferriteand the austenite. The concentration of ferrite can beincreased by up to 10% with the same chemical compositiononly by increasing the temperature of the solution annealingheat treatment.

Fig. 2 illustrates themicrostructure of thematerial in the ascast state, after solidification in a silica sand mold

154 M A T E R I A L S C H A R A C T E R I Z A T I O N 6 0 ( 2 0 0 9 ) 1 5 0 – 1 5 5

agglomerated with organic urethane-phenol resin. Thesigma phase nucleates at the ferrite/austenite interfaces andgrows towards the ferrite phase, which is the phase thatsupplies it with stabilizing elements (Cr and Mo). Duringcooling inside the silica sand mold, the solubility of alloyingelements such as chromium and molybdenum diminishesin the ferrite, favoring high concentrations at the interfaces,which are also regions of high energy, promoting thesubsequent nucleation of the sigma phase and secondaryaustenite. Precipitation of the sigma phase during coolingin the solidification process is difficult to avoid in this kindof material (SDSS), because the chemical composition fa-vors its formation. However, the quantity of precipitatedsigma phase can be minimized by controlling the chemi-cal composition itself. This is achieved by limiting theconcentration range of each element, and by controlling thecooling rate in the solidification process, maximizing it bymeans of forced air or water. Once it is formed, the sigmaphase can only be dissolved by means of heat treatment at ahigh temperature of approximately 1060 °C followed bywater quenching [3]. The soak dwell time at this temperaturemust be sufficient to allow the entire material to reach anisothermal condition.

Fig. 3 depicts the microstructure of the material in thesolution annealing condition (1130 °C, followed by waterquenching). The matrix contains the precipitated austenitephase, with rounded and elongated morphologies. Note theabsence of intermetallic or carbide phases precipitated at theferrite/austenite interfaces or precipitated intergranularly inthe ferrite. The black spots indicate inclusions originatingfrom the deoxidation process.

The heat treatment promoted the total dissolution of thesigma phase that had precipitated during solidification cool-ing, leaving the microstructure composed solely of ferrite andaustenite. After solution annealing and water quenching,some of the samples were stress relieved at 520 °C for 2 h. Thematerial's microstructure shows only the two constituentphases of the super duplex structure, i.e. ferrite and austenite,as shown in Fig. 4.

Although the material was in a thermodynamic conditionof metastability after the solution annealing heat treatment,the heat input from the stress relief was not enough toovercome the energy gap between the metastable and stableconditions in which intermetallic phases, mainly sigmaphase, would nucleate and grow in the microstructure. Theferrite/austenite interfaces, as well as the inside of the ferritegrains are devoid of precipitates, and confirm that thematerial is still in the metastable condition.

The phases in the samples that were solution annealed at1130 °C, water quenched, and stress relieved at 520 °C werecharacterized by X-ray diffraction, producing a diffractogramshown in Fig. 5. The figure shows only a few peakscorresponding to reflections of atomic planes belonging toferrite and austenite, and no reflections of crystalline planes ofany intermetallic phase, indicating that the stress relief heattreatment at 520 °C did not cause the precipitation ofundesirable secondary phases.

The anodic polarization curves of the samples solutionannealed at 1130 °C and for the samples treated at the sametemperature and then stress relieved at 520 °C for 2 h are

depicted in Fig. 6. Note that the test temperaturewas 25 °C, buttests for 5 °C and 60 °C were also made and the results areshowed at Table 2. All these temperatures tried to simulate theconditions as close as possible to the situations of pumping onoffshore platforms.

The polarization curves resulting from the aerated solutionand the CO2 saturated solution were very similar, withoutmajor deviations in shape or position in relation to theCartesian E× I system. The samples solution annealed at1130 °C and stress relieved at 520 °C generally presentedmore positive pitting potentials than the samples onlysolution annealed at 1130 °C. The region of passivation onthe curve corresponding to the solution annealed and stressrelieved samples was found to shift to higher currents only attwo of the test temperatures (5 °C and 60 °C, for CO2 saturatedsolutions), with the curves appearing practically superim-posed in the other conditions.

The anodic current oscillations (“serrated” regions of theE× I curve) observed below the pitting potential are related tothe formation and repassivation of microscopic pits, ormicropits [12]. These metastable pits, are very small, growingand repassivating in a few seconds below the pitting potentialand during the induction time for the formation of stable pits.Fig. 7 illustrates the behavior of the pitting potential as afunction of the test temperature for aerated solution (the CO2

saturated solution presented the same values). Generallyspeaking, the stress relief heat treatment at 520 °C for 2 hdid not significantly alter the pitting potential at any of thetemperatures tested here, in view of the fact that curves for allpractical purposes were superimposed.

The pitting potential decreased linearly as the test tem-perature increased, in both the solution annealed samples andthe solution annealed and stress relieved samples. Indepen-dent of the test solution (aerated or CO2 saturated), thebehavior of the straight lines was quite similar. The valuesof the pitting potentials in both aerated and CO2 saturatedsolutions were practically identical at the different testtemperatures, indicating that the stress relief heat treatmentdid not produce differences in this variable.

4. Conclusions

The sigma phase is always present in the as cast micro-structure of super duplex stainless steels. It precipitates at theferrite/austenite interfaces, occupying a major part of theferritic phase. The quantity of sigma phase can be minimizedin the as cast microstructure only by controlling the chemicalcomposition and the cooling process.

The solution annealing heat treatment followed by waterquenching dissolved the sigma phase that precipitated duringthe solidification process, leaving the microstructure com-posed solely of ferrite and austenite.

The stress relief heat treatment at 520 °C for 2 h did notalter the material's microstructure, i.e., it did not promote theprecipitation of any intermetallic phase.

From the standpoint of corrosion, the stress relief heattreatment hardly altered the pitting potential at all, at any ofthe tested temperatures in either the aerated or CO2 saturatedsolutions.

155M A T E R I A L S C H A R A C T E R I Z A T I O N 6 0 ( 2 0 0 9 ) 1 5 0 – 1 5 5

R E F E R E N C E S

[1] Weber, J. Materials for sea water pumps and related systems.SULZER Brothers Limited; Winterthur; Switzerland, p. 1–11.

[2] Maehara Y, Ohmori Y, Murayama J, Fujino N, Kunitake T.Effects of alloying elements on σ phase precipitation in δ–γduplex stainless steels. Met Sci November 1983;17:541–7.

[3] Martins, M. Caracterização microestrutural-mecânica eresistência à corrosão do aço inoxidável super duplex ASTMA890/A890M Grau 6A, Doctoral Thesis, Universidade de SãoPaulo, Interunidades EESC-IFSC-IQSC.

[4] Laycock NJ, Newman RC. Localized dissolution kinetics, saltfilms and pitting potentials. Corros Sci 1997;39(10–11):1771–9.

[5] Kim SB, Paik KW, Kim YG. Effect of Mo substitution by W onhigh temperature embrittlement characteristics in duplexstainless steels. Mater Sci Eng 1998;A247:67–74.

[6] Garfias-Messias LF, Sykes JM. Effect of copper on activedissolution and pitting corrosion of 25% Cr duplex stainlesssteels. Corros Sci 1998;54(1):40–7.

[7] Alonso, N. Análise do método potenciodinâmico dedeterminação do potencial de pite. Doctoral Thesis (Materials

Science and Engineering) — Escola Politécnica daUniversidade de São Paulo — USP — São Paulo, 1992.

[8] Nilsson JO. Overview: super duplex stainless steel. Mater SciTechnol 1992;8:685–700.

[9] Okamoto H. The effect of tungsten and molybdenum on theperformance of super duplex stainless steels, applications ofstainless steels. In: Nordberg H, Björklund J, editors. Proceedingof international conference on stainless steels. KristianstadBoktryckeri ABSweden: Stockholm; 1992. p. 360–9.

[10] American Society For Testing and Materials — ASTM A890/A890M-91. Standard practice for castings,iron–chromium–nickel–molybdenum corrosion resistant,duplex (austenitic/ferritic) for general application. Annualbook of ASTM standards. V.01.02. Ferrous castings;ferroalloys; 1999. p. 556–69.

[11] American Society for Testing and Materials — ASTM E 562.Standard practice for determining volume fraction bysystematic manual point count. Annual book of ASTMstandards. V.03.01; 1988. p. 575–9.

[12] Szklarska-Smialowska Z. Mechanism of pit nucleation byelectrical breakdown of the passive film. Corros Sci2002;44:1143–9.