Effect of Micro Plasma Arc Welding Technique on the Microstructure and Pitting Corrosion of AISI 316L Stainles Steels in LiBr Brines

Corrosion Science 53 (2011) 2598–2610

Contents lists available at ScienceDirect

Corrosion Science

journal homepage: www.elsevier .com/locate /corsc i

Effect of the micro-plasma arc welding technique on the microstructure andpitting corrosion of AISI 316L stainless steels in heavy LiBr brines

R. Sánchez-Tovar, M.T. Montañés, J. García-Antón ⇑Ingeniería Electroquímica y Corrosión, Departamento de Ingeniería Química y Nuclear, Universitat Politècnica de València, Camino de Vera s/n, 46022 Valencia, Spain

a r t i c l e i n f o

Article history:Received 4 February 2011Accepted 23 April 2011Available online 29 April 2011

Keywords:A. Stainless steelB. PolarizationC. Pitting corrosionC. Welding

0010-938X/$ - see front matter � 2011 Elsevier Ltd.doi:10.1016/j.corsci.2011.04.019

⇑ Corresponding author. Tel.: +34 963877632; fax:E-mail address: [email protected] (J. García-Antó

a b s t r a c t

The effects of the micro-plasma arc welding technique on the microstructure and pitting corrosion ofdifferent zones of an AISI 316L stainless steel were studied using different microscopy and electrochemicaltechniques. Galvanodynamic measurements and laser scanning confocal microscope were used to evalu-ate the corrosion evolution in situ. Results show, in general, the worst corrosion behaviour for the heataffected zone. Furthermore, there is a relation between the effects of the micro-plasma arc welding pro-cess on the materials microstructure and their pitting corrosion resistance. The weld zone was alwaysin the cathodic position of the possible galvanic pairs.

� 2011 Elsevier Ltd. All rights reserved.

1. Introduction

Absorption cooling systems are a suitable alternative to refriger-ation compression systems because the use of chlorofluorocarbons(CFCs) has been banned (Montreal Protocol [1], 1987) and their sub-stitutes, i.e. hydrochlorofluorocarbons, are submitted to severe reg-ulations (Kyoto Protocol [2], 1997), since they are responsible for theozone layer depletion and climate change. Lithium bromide (LiBr)heavy brines are one of the most widely used absorbents in heattransformers because LiBr possesses favourable thermo-physicalproperties [3,4]. Nevertheless, LiBr contains bromides which areaggressive ions and, thus, they can cause serious corrosion problemsin the materials which form these machines [5,6].

Stainless steels (SS) and their welded forms are widely used asstructural elements in LiBr absorption machines [7–9]. Austeniticstainless steels possess excellent resistance to general corrosion[10], however, they are susceptible to localised corrosive attacks,such as pitting corrosion, intergranular corrosion and stress corro-sion cracking in highly electrical conducting media, such as heavyLiBr brines [11–14]. Furthermore, welding procedures often wor-sen this situation by introducing residual stresses and metallurgi-cal changes [15–19]. Moreover, these changes in materials increasethe dissimilarity of the base metal-weld metal pair causing gal-vanic corrosion [20]. Pitting corrosion is directly related to thethermal energy associated with welding processes [21,22].

Weldments are indispensable in the manufacture of most com-ponents [23]. Austenitic stainless steels can be readily weldedthrough various arc welding processes [24]. Plasma arc welding

All rights reserved.

+34 963877639.n).

(PAW) can be defined as a gas-shielded arc welding process wherethe coalescence of metals is achieved via the heat transferred by anarc that is created between a tungsten electrode and a workpiece.The plasma is formed through the ionisation of a portion of theplasma gas. The PAW process uses three current modes: micro-plasma, medium current plasma and keyhole plasma. This categor-isation is primarily based on the level of welding current. Themicro-plasma mode is usually defined in the current range from0.1 to 20 A. The advantages of the PAW process are primarilyintrinsic to the keyhole mode of operation, because greater thick-nesses of the metal can be penetrated in a single step, comparedwith other processes, such as gas tungsten arc welding (GTAW).This greater penetration reduces preparation time. In some materi-als, for example, a square-grooved butt joint preparation can beused for thicknesses up to 12 mm. The PAW process can producehigh weld integrity (similar to GTAW) while minimising weld stepsand, hence, welding times and labour costs [25]. Disadvantages in-clude the higher equipment costs, when compared with the GTAWprocess [26]. Several authors have studied the micro-plasma arcwelding (MPAW) technology. Specifically, Karimzadeh et al.[27,28] studied the effect of MPAW process parameters on graingrowth and porosity distribution of a Ti6Al4V alloy weldment. Onthe other hand, metallurgical examination of AISI 304 micro-plas-ma arc welded alloys have been used in aerospace industries[29,30]. However, there is little research concerning the influenceof the micro-plasma arc welding technique on the corrosion resis-tance of stainless steels [31]. In particular, no works were foundthat study the effect of MPAW on the corrosion behaviour of thedifferent zones of a welded alloy.

The objective of this work is to study how the micro-plasma arcwelding procedure affects the corrosion behaviour of an AISI 316L

Table 2Welding parameters.

Welding process Micro-plasma arc welding (MPAW)Process type ManualBacking gas Argon (99.9%)

Flow rate 2.5 L/minPlasma gas flow rate 6.5 L/minNumber of passes 2

Step 1 Current 11.3 AVoltage 20 VWelding speed 2.6 mm/s

Step 2 Current 13 AVoltage 20 VWelding speed 2.6 mm/s

R. Sánchez-Tovar et al. / Corrosion Science 53 (2011) 2598–2610 2599

austenitic stainless steel in a heavy LiBr brine by means ofmicroscopy and electrochemical techniques. In particular, galvano-dynamic measurements together with the laser scanning confocalmicroscope were used to evaluate corrosion evolution in situ. Inthis way, three distinct zones were studied: weld zone (WZ), heataffected zone (HAZ) and base zone (BZ). Pitting corrosion mecha-nisms and the galvanic corrosion among the different parts of themicro-plasma arc welded alloy were also studied. Furthermore,the mechanical properties of the different parts of the AISI 316Lmicro-plasma arc welded SS were analysed by means of microh-ardness tests.

2. Materials and methods

2.1. Materials

The materials studied in this work were made of micro-plasmaarc welded AISI 316L SS. They were made from tubes 14 mm and16 mm in inner and external diameter, respectively, and 10 mmin length. Thus, a tube 20 mm in length was obtained. Both thebase material and the filler alloy were AISI 316L SS; Table 1 pre-sents its chemical composition, as given by the manufacturer.The composition of the filler alloy is more noble than that of thebase material. In spite of this, the compositions of both the baseand the filler alloy are similar in order to avoid future galvanicproblems. Table 2 shows the welding parameters used in the mi-cro-plasma arc welding procedure. The argon backing gas was usedto protect the inner surface of the materials during welding [32].

2.2. Microstructure analysis

The aim of the microstructural analysis was to study the effectof micro-plasma arc welding on the microstructure of AISI 316L SS.In order to estimate the possible microstructural variations pro-duced by MPAW, the welded alloy was cut lengthwise and coveredin cold mounting acrylic resin for the embedding of specimens.Then, the samples were wet abraded from 220 silicon carbide(SiC) grit to 4000 SiC grit (in several steps; i.e. 220, 500, 1000,2500 and 4000). Afterwards, the mounted samples were polishedwith 1 and 0.3 lm alumina and were rinsed with distilled water,followed by ethanol. Once the samples were polished, metallo-graphic etching was carried out according to ASM International[33]. Samples were immersed in the etching solution during 90 sand then rinsed with distilled water, followed by ethanol. The etch-ant composition consisted of 10 mL of nitric acid (65 wt.%), 10 mLof acetic acid (99–100 wt.%), 15 mL of hydrochloric acid(37–38 wt.%) and 5 mL of glycerine (100 wt.%). Once etched, thematerials were examined by light microscopy (LM) and scanningelectron microscopy (SEM) to estimate possible microstructuralvariations during the MPAW procedure. Energy dispersive X-rayanalyses (EDX) were carried out to trace the changes in the compo-sition of the welded materials. Then, the laser scanning confocalmicroscope was used to analyse the corrosion process in situ.

2.3. Microhardness tests

In order to know how the micro-plasma arc welding process af-fects the different zones of the welded alloy, Vickers microhardness

Table 1Composition (wt.%) of the AISI 316L SS base material and filler alloy used in this work acc

Material Cr Ni Mn Mo S

Base 16.957 10.171 1.337 2.298 0.004Filler alloy 18.160 12.100 1.860 2.540 0.007

measurements were taken with a microhardness tester (StruersDuramin) with a diamond pyramid indenter at a load of 300 g for15 s [34]. Hardness values were obtained as the mean of sixreadings.

2.4. Corrosion analysis

2.4.1. Corrosion experiments set upAn electrochemical mini-cell designed by the research group

was used to perform the corrosion analysis [35]. Fig. 1 shows ascheme of the parts of the mini-cell. The cell was made of glassand it was composed of two pieces. One of them has the functionof supporting the micro-plasma arc welds (Fig. 1c) and the otherpiece (Fig. 1a and b) was located at the top. The latter piece pos-sesses the inlets and the outlets of the cell. In particular, it hastwo inlets: one for the reference electrode (RE), which is a refer-ence mini-electrode of Ag/AgCl 3 M KCl, and another for takingout the electrical connection of the working electrode (WE, the mi-cro-plasma arc weld). The inlets of the cell also help to introducethe solution into the system. The counter electrode (CE) consistsof two platinum filaments that pass through the glass and are con-nected outside the cell to the potentiostat.

To conduct the electrochemical tests, the samples were cutlengthwise and covered with an epoxy resin; then, they werewet abraded from 220 silicon carbide (SiC) grit to 4000 SiC grit(in several steps; i.e. 220, 500, 1000, 2500 and 4000) and finally,they were rinsed with distilled water, followed by ethanol. Theelectrochemical connection to the potentiostat was done by meansof a conductor wire. Furthermore, an insulating lacquer was usedto study a specific zone of the stainless steel. In this way, an areaof 1.5 mm2 of each different part of the welded alloy was exposedto the LiBr solution (the area of the samples was determined forevery test by image analysis). In this way, any morphologicalchange could be detected in detail.

Micro-plasma arc welded materials were tested in an 850 g/LLiBr solution (a heavy brine like the commercial solutions com-monly used in absorption machines). Nitrogen was bubbled intothe solution for 20 min, prior to introducing the solution into thecell. Then, the mini-cell was tight closed to keep these conditions.Experiments were performed at 25 �C.

2.4.2. Electrochemical testsTwo different electrochemical techniques were used in the

corrosion studies: cyclic polarisation and galvanodynamic

ording to the inspection certificate supplied by the manufacturer.

Si P C Cu N Fe

0.368 0.030 0.022 – – Bal.0.760 0.018 0.010 0.080 0.050 Bal.

(a) (b)

(c)

WEconnection oulet

REinlet

CE

Sample location

Fig. 1. Minicell scheme: (a) main view of the top of the cell, (b) cross view of the top of the cell, and (c) cross view of the base of the cell (all the measurements are inmillimetres). RE: reference electrode; WE: working electrode; CE: counter electrode.

2600 R. Sánchez-Tovar et al. / Corrosion Science 53 (2011) 2598–2610

measurements. To perform the tests, a potentiostat (Solartron 1285provided with the Corrware software) was used. All the tests wererepeated at least three times for reproducibility.

2.4.2.1. Cyclic polarisation measurements. Before obtaining the cyc-lic potentiodynamic curves, the open circuit potential (OCP) wasrecorded for one hour. After the OCP test, the potential was re-duced progressively to �1000 mVAg/AgCl; then, the working elec-trode potential was scanned from �1000 mVAg/AgCl to the anodicdirection until the current density reached 10 mA/cm2, where thepotential scan was reversed. A scan rate of 0.5 mV/s was used.

As mentioned above in the introduction section, the micro-plasma arc welding procedure introduces changes in the materials,creating different parts with diverse microstructures in the alloys,which could cause galvanic corrosion problems. In order to evalu-ate the galvanic corrosion caused by micro-plasma arc welding, themixed potential theory (MPT) was used [36]. The mixed potentialtheory was evaluated from the cyclic polarisation curves, by super-imposing the potentiodynamic curves of the different parts of themicro-plasma arc welded material. The predicted coupled poten-tial (Ecoup) and the coupled current density (icoup) of the possiblepairs were estimated.

2.4.2.2. Galvanodynamic measurements. Before the galvanodynamicmeasurements, the open circuit potential (OCP) was recorded for1 h. After the OCP test, the current density was reduced to 0 mA/cm2 in order to start all the experiments with the same conditions.Then, the galvanodynamic curves were scanned from 0 mA/cm2 topositive current densities using a scan rate of 10�5 mA/s.

2.4.3. In situ corrosion analysis by LSCMThe electrochemical mini-cell makes it possible to use laser

scanning confocal microscopy to analyse the corrosion processin situ. Furthermore, the galvanodynamic tests permit controlling

the current applied. In this way, the corrosion process can be ana-lysed in situ by obtaining images of the surface of the differentparts of the micro-plasma arc welded materials with the laserscanning confocal microscope. The laser scanning confocal micro-scope is an Olympus LEXT OLS3100 microscope, which uses LEXTOLS 6.0.3 software. The LSCM uses a Laser Diode with a wavelengthof 408 nm, an outstanding horizontal resolution of 0.22 lm and avertical resolution of 0.01 lm (z-axis).

3. Results and discussion

3.1. Effect of MPAW on AISI 316L microstructure

As indicated in Section 2.2 of the experimental procedure, thematerials were etched in order to observe their microstructureafter the micro-plasma arc welding. Fig. 2 shows a LM image ofthe lengthwise surface of the micro-plasma arc welded alloy. Thisfigure also shows the microstructure of the main areas of thewelded material obtained by scanning electron microscopy. Inspite of the fact that the micro-plasma arc welding procedurewas performed along two small AISI 316L SS tubes (each tube10 mm in length) and all the tube can be considered heat affectedzone, three major areas were differentiated: the weld zone, WZ,(Fig. 2a), the zone which is closer to the weld, referred to as heataffected zone, HAZ, (Fig. 2b) and the zone farther from the weld,called base zone, BZ (Fig. 2c).

Fig. 2 shows that the microstructure of AISI 316L changes dueto the micro-plasma arc welding process. As Fig. 2a shows, themicrostructure of the weld zone possesses a vermicular mor-phology characterised by columnar grains with delta-ferrite[37]. This fact is expectable since fast cooling in welding doesnot last enough time to complete phase transformation. As aresult, a large portion of delta-ferrite is retained in the WZ ofthe micro-plasma arc welded alloy and an incomplete

100 m

a

100 m100 µm

a

100 m

b

100 µm

b

1 mm

a

b

c

c

b

1 mm1 mm

a

b

c

c

b

100 m

c

100 µm

c

Fig. 2. Length ring surface of the MPA welded alloy obtained by means of LM. Images of the weld zone (a), HAZ (b) and base zone (c) using SEM.

R. Sánchez-Tovar et al. / Corrosion Science 53 (2011) 2598–2610 2601

transformation results in the retention of the skeletal dendriticdelta-phase within the austenite matrix [38]. A small amountof delta-ferrite is necessary to avoid the problem of hot crackingduring weld solidification [39,40]. In fact, the weld zone ofaustenitic stainless steels has a cast structure with 2–10 wt.%delta-ferrite in the austenite matrix [10]. However, delta-ferritecontents could lead to many Cr-depleted zones which are likelyto form due to micro-segregation and, consequently, to a reduc-tion in corrosion resistance due to the formation of a less stablepassive film [41].

The HAZ which is closer to the weld zone presents an austeniticmicrostructure with the typical recrystallisation and grain growth[42]. In general, the HAZ is characterised by a heterogeneousmicrostructure relative to grain size. In Fig. 2b and c the differencesin grain size due to the micro-plasma arc welding process can beobserved. The zone which is farther from the weld zone also hasan austenitic microstructure but the grain does not undergo con-siderable enlargement of the grain size; thus, this zone will be re-ferred to as base zone. BZ microstructure also has polygonalaustenitic grains where twins are visible.

2602 R. Sánchez-Tovar et al. / Corrosion Science 53 (2011) 2598–2610

An energy dispersive X-ray analysis (EDX) was performed in or-der to observe variations in the composition of stainless steel dueto the micro-plasma arc welding procedure. In this way, the maincomponents of the alloy (i.e. Cr, Mo and Ni) were analysed. Table 3shows that the compositions of HAZ and BZ are similar, with Crvalues of 17.43–17.32 wt.%, Mo contents of 2.38 wt.% and Ni valuesfrom 10.57 to 10.45 wt.%. Table 3 also shows higher contents of allthese elements (Cr, Mo and Ni) in WZ. This could be due to the factthat the composition of the filler alloy presents larger amounts ofCr, Mo and Ni. A composition analysis was performed on WZ to dis-tinguish between the delta-ferrite grains and the austenite matrixcomposition. Fig. 3 shows that accumulations of Ni were found inthe austenite matrix (A-zone) while the Ni content is lower in thedelta-ferrite zone (F-zone) [40]. On the other hand, the delta-ferriteregions present higher Cr values. Using transmission electronmicroscopy, Brooks et al. [43] proved that during cooling transfor-mation there is a separation of Cr to ferrite and Ni to austenite.

Table 3EDX (wt.%) analysis on the MPA welded zones: WZ, HAZ and BZ.

Material Cr Mo Ni

WZ 18.05 2.54 11.49HAZ 17.43 2.38 10.57BZ 17.32 2.38 10.45

A-zone

F-zone

50 m

A-zone

F-zone

A-zone

F-zone

50 µm

0

5

10

15

20

25

Elements

ED

X (

wt.

%)

F-zone A-zone

Cr Mo Ni

a

b

Fig. 3. Images of the WZ: austenite and ferrite zones (A-zone = austenite zone;F-zone = ferrite zone) (a) and EDX (wt.%) composition of the WZ of the MPA weldedmaterial (b).

3.2. Effect of MPAW on AISI 316L microhardness

Fig. 4 shows the microhardness values obtained for the differentzones of the micro-plasma arc welded AISI 316L SS. The base zonepresents the highest microhardness value. The microhardness val-ues of HAZ and WZ are similar (174.33 and 181.67 HV300g, respec-tively). The different microhardness values between the weld andbase zone could be due to the presence of delta-ferrite in the for-mer [44].

As it can be observed, the micro-plasma arc welding processleads to different microhardness values depending on the zone.In general, there is a positive correlation between hardness andstrength: the higher the hardness, the higher the strength. There-fore, the weld strength is weaker in HAZ and WZ than in BZ.

3.3. Effect of MPAW on AISI 316L corrosion

3.3.1. Cyclic polarisation measurements3.3.1.1. Open circuit potentials. Previous to the polarisation curvesof the different zones of the micro-plasma arc welded material,an open circuit potential measurement was performed. The OCPvalues, obtained as the arithmetic mean of the last 5 min valuesof open circuit potential measurements [45], for WZ, HAZ and BZin the LiBr solution are �173 ± 21, �189 ± 29, �81 ± 4 mVAg/AgCl,respectively. According to these results, the most noble open cir-cuit potential is obtained for the base zone whereas the most activeone for HAZ. However, the OCP values for HAZ and WZ are similar.The most positive open circuit potential values of the base zonecould be related to the better Cr2O3 protective film formed onthe metal surface due to its homogeneous austenitic microstruc-ture [24].

Fig. 5 shows the cyclic potentiodynamic curves of BZ, HAZ andWZ of the micro-plasma arc welded alloy in the LiBr solution. Astests were reproducible, the curves shown in Fig. 5 illustrate oneof the recorded measurements. Fig. 5 shows that the obtainedcurves are typical of passivable materials with a current plateauin the anodic branch; thus, micro-plasma arc welded materialspassivate in 850 g/L LiBr solutions.

On the other hand, as it can be observed in Fig. 5, the OCP valuesare in the passivity range for the three micro-plasma arc weldedzones of AISI 316L stainless steel. This means that, at the open cir-cuit potential, the MPA welded alloy spontaneously passivates.

3.3.1.2. Corrosion parameters. Table 4 shows the typical corrosionparameters (corrosion potential: Ecorr and corrosion current

181.67

223.44

174.33

0

50

100

150

200

250

300

BZ HAZ WZ

Material zone

HV

300

g

Fig. 4. Microhardness values of the different zones of the MPA welded alloy.

0.0001

0.01

1

100

-1 -0.5 0 0.5

E (VAg/AgCl)

IiI (

mA

/cm

)2

BZ HAZ WZ

Fig. 5. Cyclic polarisation curves of the different zones of the MPA welded stainlesssteels in 850 g/L LiBr solutions.

R. Sánchez-Tovar et al. / Corrosion Science 53 (2011) 2598–2610 2603

density: icorr) determined from the polarisation curves shown inFig. 5. Ecorr was obtained as the potential at which the net currentdensity (sum of the anodic partial current density and the cathodicpartial current density) was equal to zero. Corrosion current densi-ties were determined by the extrapolation of the cathodic apparentTafel slopes to the point that yields the corrosion potential. Extrap-olation started over about 50 mV away from Ecorr and the samerange of potential values was used throughout the tests. The mostprecise determination of the corrosion current density values byTafel extrapolation is when both the anodic and cathodic branchesshow linearity. However, it is also possible to make an accurateevaluation of the corrosion current densities if one of the branchesof the polarisation curves displays a sufficiently long linear ten-dency around the corrosion potential [46–49]. In fact, two rulesof thumb should be applied when using Tafel extrapolation. First,at least one of the branches of the polarisation curve should exhibitTafel behaviour (i.e., linear on semilogarithmic scale) over aboutone decade of current density. Second, extrapolation should startat least 50–100 mV away from Ecorr. These two rules improve theaccuracy of manual extrapolation [50]. This allows a precise evalu-ation of the effect of micro-plasma arc welding on the corrosionprocess of the alloy. Previous works also used the cathodic Tafelslope to evaluate icorr in passivable materials [51,52].

Regarding the corrosion potentials, Table 4 shows that the mostactive values (more negative) are obtained for the base and heat af-fected zones (�902 and �784 mVAg/AgCl, respectively). The corro-sion potentials obtained from the potentiodynamic curves aremore negative than the OCP values; this is due to the polarisationapplied during the potentiodynamic sweep [53–55]. During theopen circuit measurements a protective film could be formed onthe metal surface, shifting the potential to more positive values.However, at the beginning of the potentiodynamic sweep the po-tential diminished to �1 VAg/AgCl. At this negative potential the me-tal surface is modified (the surface of the samples is activated bythe cathodic polarisation) and the corresponding Ecorr calculatedfrom the potentiodynamic curve is more negative than the poten-tial obtained under open circuit conditions (OCP value). Underthese conditions, the weld zone presents the most noble corrosion

Table 4Parameters of the MPA welded materials obtained from the cyclic polarisation curves in 8

Material Ecorr (mVAg/AgCl) icorr (lA/cm2) Ep (mVAg/

WZ �676 ± 31 5.8 ± 1.1 288 ± 3HAZ �784 ± 42 7.0 ± 1.2 366 ± 18BZ �902 ± 18 4.3 ± 1.7 593 ± 37

potential values due to its higher Cr, Ni and Mo content. In fact,Pardo et al. [13] proved that Mo additions shifted the corrosion po-tential to more noble values.

Table 4 also shows that HAZ has the highest corrosion currentdensity and the BZ the lowest. Since the corrosion current densityis a parameter related to the corrosion rate, higher values involvegreater corrosion rates. Thus, the heat affected zone of the micro-plasma arc welded stainless steels possesses the worst corrosionresistance in terms of corrosion rate values.

3.3.1.3. Pitting susceptibility and passivation parameters. Cycliccurves also provide information about the pitting, passivationand repassivation behaviour of the materials with parameters suchas pitting potential (Ep) and passivation current density (ip), whichare also shown in Table 4.

Pitting potential (Ep) is the potential at which current densityreaches 100 lA/cm2 [56]. Pitting potential determines the pittingcorrosion susceptibility of the micro-plasma arc welded zone, thatis, susceptibility to local breakdown and pit initiation. All the mi-cro-plasma arc welded zones of AISI 316L SS are susceptible to pit-ting corrosion since bromides are very aggressive ions thatpromote passive film breakdown of stainless steels [56–58]. Table 4shows that BZ presents the most positive Ep value and WZ the mostnegative value. Ep value for the HAZ lies between the ones obtainedfor the base and the weld zones. In fact, Luo [24] observed that thecoarse grains in HAZ have a negative influence on corrosion resis-tance. By contrast BZ is more resistant to pitting corrosion and,consequently, the breakdown of the passive film and the initiationof pits can occur at more positive potentials. The residual stressesinduced by MPAW will make WZ more sensitive to localised corro-sion [10]. According to Lee et al. [21], a fully austenitic structurehas the lowest susceptibility to pitting corrosion attack since thepassive film on such homogenous microstructure is more resistantto pitting corrosion. According to the literature [59–61], the pres-ence of chromium-depleted regions is related to pit nucleationand growth. As shown in Fig. 3, the delta-ferrite microstructureof WZ promotes the formation of Cr-depleted regions (in the aus-tenite matrix). Furthermore, delta-ferrite has been proved to bedetrimental because of its susceptibility to be attacked in chloridemedia [62,63] and the same effect may occur in the presence ofbromide ions. This could lead to a decrease in pitting potential,which could explain the most negative OCP values of the weldzone.

Passivation current densities (ip) were obtained from the cyclicpotentiodynamic curves as the mean values where current densityremains stable when potential shifts to the anodic direction. The ipvalues determine the corrosion rates of the different zones of theMPA welded alloy in the passivity range. Table 4 shows that passiv-ation current densities are higher than their corresponding corro-sion current densities in all the zones studied. This fact meansthat only partial passivation is possible due to the formation of apassive layer in the MPA welded SS zones in LiBr solutions. It isgenerally accepted that Cr2O3-based products form barrier layersand they are responsible for the superior corrosion resistance ofstainless steels [64]. Cr2O3-based barrier layers tend to form onCr-containing iron based alloys if the Cr content of the alloy ex-ceeds 12–15 wt.%. Bromide ions can be adsorbed by the film andreplace oxygen, forming a soluble metal bromide, destroying the

50 g/L LiBr solutions.

AgCl) ip (lA/cm2) irp (mA/cm2) Erp (mVAg/AgCl)

13.9 ± 3.2 37.1 ± 10.6 �79 ± 325.1 ± 7.3 60.1 ± 16.4 �89 ± 1419.2 ± 0.3 35.8 ± 1.6 �100 ± 1

-1000 -400 200 800

E (mVAg/AgCl)

WZ

BZ

HAZ

E p - E corr = 964 mV

E p - E corr = 1150 mV

E p - E corr = 1495 mV

E corr E p

-1000 -400 200 800

E (mVAg/AgCl)

WZ

BZ

HAZ

E rp - E corr = 597 mV

E rp - E corr = 694 mV

E rp - E corr = 802 mV

E rpE corr

a. Passivity range

b. Perfect passivity region

c. Imperfect passivity region

-1000 -400 200 800

E (mVAg/AgCl)

WZ

BZ

HAZ

E c - E corr = 367 mV

E c - E corr = 456 mV

E c - E corr = 693 mV

E pE rp

Fig. 6. Passivity ranges (a), perfect passivity region (b) and imperfect passivityregion (c) of the MPA welded zones in 850 g/L LiBr solutions.

2604 R. Sánchez-Tovar et al. / Corrosion Science 53 (2011) 2598–2610

film and increasing corrosion rates [65]. Additionally, alterations inthe microstructure and composition of the welded stainless steelcan affect the structure of the passive film. HAZ has the highestpassivation current densities (ip values 1.8 and 1.3 higher thanfor WZ and BZ, respectively) and WZ the lowest due to its higherCr, Mo and Ni content (Table 4).

3.3.1.4. Repassivation parameters. The parameters related to therepassivation characteristics of the materials, repassivation currentdensity (irp) and repassivation potential (Erp), were also determinedfrom the cyclic polarisation curves of the different zones in the mi-cro-plasma arc welded materials (Table 4). irp values represent themaximum current reached since the current does not begin to de-crease immediately after scan reversal [22]. Repassivation currentdensity is an inverse measure of the ability of materials to repass-ivate and, hence, of the extent of propagation once corrosion hasinitiated. Table 4 shows that the highest repassivation current den-sity is obtained for HAZ, with the worst ability to repassivate. Thisis related with the high susceptibility to pitting corrosion and highpassivation current density of the heat affected zone. Garcia et al.[38] also determined a sharp increase in the repassivation currentdensity value of HAZ.

Erp was recorded at the crossing point between the backwardand forward scans. This parameter refers to the limit below whichthe metal remains passive and active pits repassivate. If Erp is morepositive than the corrosion potential, the material is able to regen-erate an eventual breakdown of the passive film; however, if Erp ismore negative than the corrosion potential, the material is not ableto completely repassivate the pits. The Erp values obtained aremore positive than the corrosion potential in all cases; thus, thematerial can repassivate the pits. The repassivation potentials areof the same order of magnitude in all stainless steel zones (�79,�90 and �100 mVAg/ACl for WZ, HAZ and BZ, respectively), thoughthey are slightly more positive for WZ, which exhibits betterrepassivation characteristics. This could be attributed to its highnickel [57] and molybdenum [66] content, which are elements thatpromote pit repassivation.

Fig. 6 shows the different zones of the passivity range. Thisrange is a measure of the tendency to nucleation pitting. The great-er the difference, the higher the resistance of the material to pittingcorrosion [65]. The repassivation potential could differentiate twozones in the passivity range of the materials (Ep–Ecorr) which weredefined by Bellezze et al. [67], i.e. a perfect or stable passivity re-gion (Erp–Ecorr) between the corrosion potential and Erp (where pit-ting corrosion cannot initiate and existing pits cannot propagate)and an imperfect or unstable passivity region (Ep–Erp) betweenErp and Ep, (where pits cannot initiate but existing pits can propa-gate). In addition, the last region (Ep–Erp) also defines the hystere-sis loop. The narrower the hysteresis loop, the easier it becomes torepassivate the pit. Furthermore, the reasons why the hysteresisloop narrows must be considered since it is associated with a shiftof the Erp to more positive values as indicative of good repassiva-tion properties. In this study Fig. 5 shows that the hysteresis loopnarrows due to the welding process (in WZ and HAZ); however,the Erp values are similar in all zones studied (Table 4). Hence, thisfact cannot be associated with a less difficult repassivation of pitsin BZ. Indeed, the widest imperfect passivity region of BZ is dueto its high positive Ep value.

Fig. 6a also shows that the highest passivity range is obtained inthe base zone of the micro-plasma arc welded alloys. The fine crys-talline grains facilitate the formation of the Cr2O3 passive film [24].HAZ and WZ microstructural changes decrease the resistance ofAISI 316L SS to pitting corrosion, in particular, the weld zone,which possesses the smallest passivity range. Fig. 6b and c showthe perfect or stable passivity region and the imperfect or unstablepassivity region, respectively. As it can be seen there, the size of the

perfect and imperfect passivity regions is similar in all MPA weldedzones, the former being slightly larger.

3.3.1.5. Evaluation of galvanic corrosion. The results show that themicro-plasma arc welding procedure alters the microstructure ofthe materials causing local variations in the composition and struc-

R. Sánchez-Tovar et al. / Corrosion Science 53 (2011) 2598–2610 2605

ture of the alloy. Changes in the parameters determined from thepolarisation curves have also been analysed. The dissimilaritiesin properties and microstructural morphologies of the regions ofa welded alloy such as MPA welded AISI 316L SS, can result in dif-ferent corrosion sensitivities and may promote the formation ofgalvanic pairs [39].

Open circuit potentials of the different samples play an impor-tant role in galvanic interactions. However, the magnitude of thegalvanic current is only in part a function of the open circuit poten-tials of these samples. The anodic and cathodic behaviours havestrong influences on this interaction. As a consequence, it is neces-sary to consider the cathodic and anodic polarisation curves ineach particular case, which provides a better predictive techniquefor the galvanic corrosion rate [68,69]. The mixed potential theoryhas been widely used to study the galvanic corrosion [70–73] and,in particular, to analyse the galvanic corrosion behaviour betweennon-welded/welded couples [74].

Fig. 5 shows that the base zone is located in the anodic positionwith respect to both HAZ and WZ. The galvanic pairs are BZ/HAZand BZ/WZ, where HAZ and WZ remain protected and BZ corrodes.The heat affected zone has a more active corrosion potential thanthe weld zone; thus, the formation of the galvanic pair HAZ/WZis also possible, where HAZ corrodes and WZ remains protected.This fact could be attributed to the higher chromium, molybdenumand nickel content of the filler alloy (18.16 wt.%, 2.54 wt.% and12.10 wt.%, respectively) used in the micro-plasma arc weldingwith respect to the base metal composition (16.96 wt.%,2.30 wt.% and 10.17 wt.%, respectively). Dadfar et al. [62] workedwith AISI 316L SS in saline solutions and their study revealed that,when the base and weld zones were adjacently placed together inan electrolyte, the base metal was the anodic member of the pair.The same result was obtained in Lee et al. [21] for AISI 304 weldsand base material in NaCl solutions. The galvanic pile formed atthe interface will not endanger the weld with other forms of corro-sion (crevice, pitting, or intergranular corrosion) resulting from thegalvanic coupling [74].

Table 5 shows the different galvanic parameters of the pairsaccording to the mixed potential theory [36]. The predicted cou-pled potential (Ecoup) and the galvanic current density (icoup) ofthe pair were estimated from the intersection point between theanodic branch of the less noble material and the cathodic branchof the most noble material of the polarisation curves (Fig. 5). Let-ters A and C in Table 5 refer to ‘‘the anode’’ and ‘‘the cathode’’,respectively, and the position of the micro-plasma arc weldedzones in the galvanic pair. Other parameters of the importance inthe galvanic pair were also analysed. Minimal differences of 100–130 mV between the corrosion potential of the cathode and the an-ode of the pair (Ecorr_C–Ecorr_A) are necessary to consider the gal-vanic effect significant [75]. According to Mansfeld and Kendel[76] the relative increase in the corrosion rate of the anode ofthe galvanic pair can be expressed by the ratio icoup/icorr, where icorr

is the corrosion current density of the uncoupled anode. The mag-nitude of this ratio may be used as a guide to reflect the severity ofthe galvanic effect and it was suggested that a icoup/icorr value lowerthan 5 implies compatibility of the members in a galvanic pair [77].

Table 5Galvanic corrosion parameters of the possible pairs among the different MPA weldedzones.

Couple(A/C)

Ecoup

(mVAg/AgCl)icoup

(lA/cm2)Ecorr_C–Ecorr_A

(mV)icoup/icorr

BZ/HAZ �822 ± 21 10.7 ± 2.8 118 2.5BZ/WZ �766 ± 29 15.9 ± 2.0 226 3.7HAZ/WZ �731 ± 11 11.5 ± 2.1 108 1.6

Table 5 shows that the most noble coupled potentials were ob-tained for the HAZ/WZ pair whereas the most active Ecoup valueswere obtained for the BZ/HAZ pair. Regarding the coupled currentdensities, all pairs present higher values than the respective corro-sion current density obtained for the uncoupled anode of the pair.BZ duplicates and triplicates its corrosion current density when itis coupled with HAZ and WZ, respectively (from 4.3 lA/cm2 to10.7 and 15.9 lA/cm2, respectively). The heat affected zone in-creased its corrosion current density by 64%. The values of the dif-ference between the corrosion potentials of the cathode and theanode indicate a significant galvanic effect in the BZ/WZ pairwhereas the galvanic pairs BZ/HAZ and HAZ/WZ are in the limitto be considered significant. As for the severity of the galvanic pair,icoup/icorr values indicate compatibility between the members ofthe pair in all cases, although, the BZ/WZ galvanic pair presentsthe highest icoup/icorr value (3.7). This is in agreement with the re-sults obtained for the minimal differences (Ecorr_C–Ecorr_A) to con-sider the galvanic effect significant.

Consequently, the pair with the worst galvanic behaviour interms of highest coupled current density and severity of the pairis the BZ/WZ galvanic pair, which also presents the highest icoup/icorr ratio. This could be due to the fact that the base and the weldzones have the greatest differences in terms of composition andmicrostructure.

3.3.2. Galvanodynamic measurementsThe best approach to study corrosion evolution is the in situ

observation technology. As it was previously mentioned in theexperimental procedure, galvanodynamic tests were performedin order to analyse the corrosion process in situ by means of anelectrochemical mini-cell connected to a laser scanning confocalmicroscope since these experiments make it possible to controlthe applied current. Like with the cyclic potentiodynamic curves,prior to the galvanodynamic tests, an open circuit potential wasregistered for 1 h. The mean values of the OCP, obtained as thearithmetic mean of the last 5 min values of open circuit potentialmeasurements [45], are �159 ± 5, �162 ± 17 and �62 ± 5 mVAg/AgCl

for WZ, HAZ and BZ, respectively in the LiBr solution. The OCP val-ues follow the same tendency as that of the potentiodynamiccurves, i.e. the most noble open circuit potential is obtained forthe base zone whereas more active open circuit potentials are ob-tained for HAZ and WZ.

Fig. 7 shows the galvanodynamic scans obtained for the differ-ent micro-plasma arc welded zones studied: WZ, HAZ and BZ. Ta-ble 6 shows the parameters obtained from the galvanodynamictest of Fig. 7. The corrosion potential (Ecorr) was calculated as themean value when current density was equal to zero (before start-ing the test and after the open circuit potential recordings). The po-tential and current density at which the first pit was observedin situ were named first pitting potential (Efp) and first pitting cur-rent density (ifp). The values shown in Table 6 are in accordancewith those obtained from the polarisation curves (see Table 4). Cor-rosion potentials follow the same tendency as in the polarisationmeasurements, that is, the most active value is obtained for BZwhereas WZ presents the most noble corrosion potential. The cor-rosion potentials are more negative than the OCP values due to theapplied polarisation. On the other hand, ifp was very close to100 lA/cm2 in all cases (Table 6), which is the current density va-lue that determined the pitting potential from the cyclic polarisa-tion measurements. Thus, the first pit observed on the surfaces ofthe materials occurs close to the pitting potential. The trend ofthe Efp values is the same as that of the pitting potentials in thecyclic polarisation tests (see Table 4), that is, the most positiveEfp is obtained for BZ, followed by HAZ and the most negative cor-responds to WZ.

0

500

1000

-0.3 0 0.3 0.6

E (VAg/AgCl)

IiI (µ

A/c

m2 )

BZ

HAZ

WZ

100

Fig. 7. Galvanodynamic curves of the different zones of the MPA welded stainlesssteel in 850 g/L LiBr solutions.

Table 6Parameters of the MPA welded materials obtained from the galvanodynamic curves in850 g/L LiBr solutions.

Material Ecorr (mVAg/AgCl) Efp (mVAg/AgCl) ifp (lA/cm2)

WZ �665 ± 26 324 ± 46 102.2 ± 4.5HAZ �788 ± 34 424 ± 12 105.4 ± 20.8BZ �941 ± 21 520 ± 17 102.9 ± 11.6

2606 R. Sánchez-Tovar et al. / Corrosion Science 53 (2011) 2598–2610

Since the OCP values of the different zones of the MPA weldedalloy were more negative than their corresponding first pittingpotentials, pits were not observed during the OCP measurements.

Galvanodynamic tests together with LSCM allow the in situanalysis of the corrosion process by obtaining images of the surface

160 m160 m160 µm160 160 m160 m160 µm160

a. A = 1.4·103 µm2 b. A = 3.1·103 µm2 c.

80 m80 m80 m80 µm 80 m80 m80 µm80 m

e. A = 6.2·103 µm2 f. A = 8.3·103 µm2 g

80 m80 m80 µm80 m 160 m160 m160 µm160 m

i. A = 37.4·103 µm2 j. A = 42.9·103 µm2 k

Fig. 8. Images of a pit evolution in the WZ obtained by means of LSCM during the galvaordered. A = surface area of the pit.

of the different parts of the micro-plasma arc welded materials(Figs. 8–11). These Figures show that the corrosion mechanism ofAISI 316L SS in LiBr solutions is typical of pitting corrosion sincestainless steels in the presence of aggressive ions (such as bro-mides) are susceptible to pitting corrosion [11,13]. Susceptibilityto pitting corrosion increases due to the micro-plasma arc weldingprocedure as reported in the literature [21,22,78] and proved inthis work (see Tables 4 and 6). Despite the fact that pits propagateat very high current densities, the surface area of an initial pitgrowing is very small and even a high current density can resultowing to low values of the absolute current flowing out of thepit [79]. Thus, galvanodynamic tests are an interesting techniqueto analyse pitting corrosion susceptibility because they allow thecontrol of the current density values. Pitting corrosion phenomenaare favoured by the anodic dissolution of the metal with metalliccations inside the pit. As a consequence of the hydrolysis of themetallic ions, the local pH is reduced and Br- ions migrate to thepit due to electroneutrality [80–82]. This effect generates moreaggressive local conditions and the metal dissolution continues.The release of aggressive anions from pits and their diffusion overthe electrode surface may cause weakening of the protective pas-sive film and each active pit is more likely to create further pitsin its vicinity [83–85]. The next Sections (3.3.2.1–3.3.2.3) describethe different processes of pitting corrosion which occur in the dif-ferent parts of the micro-plasma arc welded zones.

3.3.2.1. WZ corrosion. Figs. 8 and 9 show the corrosion processwhich occurs in the weld zone of the micro-plasma arc weldedmaterials. The WZ galvanodynamic curve in Fig. 7 shows that thepotential first shifts to more positive values increasing currentdensity and when the current density reaches a certain value thepotential remains constant (Efp). Then, the potential shifts to morenegative values and some potential peaks in the galvanodynamicregister are observed. Efp is related to the formation of pits onthe WZ surface. The potential peaks that appear after the shift of

80 m80 m80 m80 m80 m80 m80 µm80 80 m80 m80 µm80

A = 3.4·103 µm2 d. A = 4.4·103 µm2

80 m80 m80 µm80 m 80 m80 m80 µm80 m

. A = 11.5·103 µm2 h. A = 16.3·103 µm2

160 m160 m160 µm160 m 160 m160 m160 µm160 m

. A = 183.6·103 µm2 m. A = 268.5 ·103 µm2

nodynamic test in 850 g/L LiBr solutions after reaching Efp. Images are temporarily

a (t = 220 s) b (t = 490 s)

160 m160 m160 µm 160 m160 m160 µm

c (t = 880 s) d (t = 1280 s)

160 m160 m160 µm 160 m160 m160 µm

Fig. 9. Images of the corrosion process of the WZ obtained by means of LSCM during the galvanodynamic test in 850 g/L LiBr solutions after reaching Efp. Images aretemporarily ordered.

30 m30 m30 µm

Fig. 10. Image of the corrosion process of the HAZ during the galvanodynamic testin 850 g/L LiBr solutions after reaching Efp obtained using the LSCM.

80 m80 m80 µm

(a)

(b)

Fig. 11. Images of the corrosion process of the BZ obtained by means of LSCMduring the galvanodynamic test in 850 g/L LiBr solutions after reaching Efp. 2-Dimage (a) and 3-D image (b).

R. Sánchez-Tovar et al. / Corrosion Science 53 (2011) 2598–2610 2607

the potential to more negative values are due to the formation ofnew pits and the spread of the corrosion products. Fig. 8 showsthe evolution of a pit (where more than one pit is shown, the evo-lution refers to the pit highlighted with a circle) from Efp (Fig. 8a)until the end of the test (Fig. 8m). The pit grows uniformly in alldirections from the initiating defect as the current density appliedis higher (the area values of the pit evolution are also indicated inFig. 8). The surface of the material shown in Fig. 9 presents multi-corrosion attack where small pits cover the surface of the alloy (aninterdendritic attack of the WZ surface). This fact was also ob-served in the work of Lin et al. [86], where a great number of pitswere obtained in the WZ. First (Fig. 9a) all pits have the same sizeand their surface increases with time (Fig. 9b–d). A delta-ferritemicrostructure could be associated with a large number of inter-faces where austenite particles can nucleate and grow [22]. Someauthors [10,62,87] agree that while delta-ferrite increases during

welding many chromium-depleted zones may form because ofmicrosegregation of Cr at the ferrite-austenite interphase bound-

2608 R. Sánchez-Tovar et al. / Corrosion Science 53 (2011) 2598–2610

aries, leading to the formation of a large number of pits throughoutthe WZ of the MPA welded alloy. Kwok et al. [88] also observedthat austenitic stainless steel regions, which were depleted in Cr,can act as nucleation sites for pits in chloride media.

3.3.2.2. HAZ corrosion. Fig. 10 shows the surface of the heat affectedzone after the galvanodynamic test. Contrarily to what occurs in theweld zone, the multicorrosion process is not observed in HAZ. Fig. 7shows that potential shifts to more positive values with current den-sity until a sharp shift to more negative potential values is observed,indicating the formation of a pit on the HAZ surface. Then, the poten-tial remains almost constant. From the first pit, corrosion productsspread out continuously around the initial damage. Corrosion prod-ucts form a new hemisphere and, subsequently, corrosion affects alarge area of the HAZ surface. HAZ microstructure composed onlyof austenite grains leads to the formation of single pits instead ofmulticorrosion processes. Fig. 10 shows a photograph of the pit ob-tained by LSCM. Pits may be crystallographic in nature with flatwalls and etched interior surfaces (formed at relative low potentials)or could be approximately hemispherical with polished interiors(formed at higher potentials) [12]. According to Fig. 10, this mightbe a polishing pit. Sato [89] showed that pits initiated above Ep (inthis case above Efp), grew as hemispherical polishing state pits, butif the potential is then reduced, the pits either repassivate or propa-gate as active (etching) pits.

3.3.2.3. BZ corrosion. Fig. 11 shows the corrosion process affectingthe base zone. Like in HAZ, no signal of multicorrosion wasdetected. This is in agreement with the galvanodynamic curve ob-tained for BZ, which shows a single peak corresponding to the firstpitting potential (see Fig. 7). BZ has a homogeneous austeniticmicrostructure with fewer defects associated with pit initiationsites. As it happened in HAZ, potential shifts to more positive val-ues with increasing current density until the potential reaches aconstant value followed by a drop in potential (pit formation).Then, the potential remains constant. Fig. 11 shows that BZ surfacepresents two pits after the test in LiBr solutions; both pit surfacesincrease with time and are dish-shaped. Large pits are often foundto be dish-shaped rather than perfectly hemispherical becausewithout a pit cover, the pit edges have lower associated solutionor diffusional resistances and dissolve more rapidly [12]. Pistoriusand Burstein [90] believed that all pits grew under diffusional con-trol with a salt on their surface, even before the cover was lost.Fig. 11b shows a three dimensional (3-D) image of the pits. Thisimage shows that the pits generate an accumulation of corrosionproducts on the BZ pit surface (volume is higher where corrosionbegins).

4. Conclusions

In this work the effect of the micro-plasma arc welding tech-nique on the microstructure and pitting corrosion susceptibilityof AISI 316L stainless steels in a heavy LiBr brine has been studiedby means of microscopy and electrochemical techniques. The mainconclusions obtained from this work are presented below:

(1) The microstructure of AISI 316L stainless steel changes dueto the micro-plasma arc welding process. The weld zone(WZ) possesses vermicular morphology characterised bycolumnar grains containing delta-ferrite. Both the HAZ andbase zone (BZ) present an austenitic microstructure butthe former has typical recrystallisation and grain growth.

(2) X-ray analysis reveals similar compositions in HAZ and BZwhereas WZ presents higher contents of chromium, nickeland molybdenum.

(3) According to Vickers’ microhardness values, the joiningstrength is weaker in the HAZ and WZ than in BZ.

(4) BZ presents the best overall corrosion behaviour comparedto the other micro-plasma arc welded zones. However,HAZ shows the highest corrosion, passivation and repassiva-tion current densities, i.e. greater corrosion rates and aworse ability to repassivation.

(5) The galvanic pairs BZ/HAZ, BZ/WZ and HAZ/WZ could beformed according to the Mixed Potential Theory, where theWZ remains protected. BZ doubles and triples its corrosioncurrent density when coupled with HAZ and WZ, respec-tively and HAZ increases its corrosion current density by64%.

(6) The pair which presents the worst galvanic behaviour interms of highest coupled current density and severity isthe BZ/WZ galvanic pair, which also presents the highesticoup/icorr ratio.

(7) Corrosion potentials follow the same tendency regardless ofthe electrochemical technique used. ifp was very close to100 lA/cm2 in all cases indicating that the first pit observedon the surfaces of the materials appears close to the pittingpotential. The trend of the Efp values is the same as that ofthe pitting potentials.

(8) All the micro-plasma arc welded zones (WZ, HAZ and BZ) ofAISI 316L SS are susceptible to pitting corrosion. A relationbetween the pitting corrosion behaviour and the microstruc-ture of the different MPA welded materials was found. Mul-ticorrosion attack occurs in the weld zone (due tomicrosegregation of Cr at the ferrite-austenite interphasebondaries); however, no evidence of multicorrosion wasdetected in HAZ and BZ, which are zones with austeniticmicrostructure.

Acknowledgements

The authors would like to express their gratitude to the MICINNfor the financial support (project CTQ2009-07518), to the FPUgrant given to Rita Sánchez Tovar, to FEDER, to the GeneralitatValenciana for its help in the LSCM acquisition (MY08/ISIRM/S/100) and to Dr. Asuncion Jaime for her translation assistance.

References

[1] Council Decision of 14 October 1988 concerning the conclusion of theVienna Convention for the protection of the ozone layer and the MontrealProtocol on substances that deplete the ozone layer, Official J. L 297 (1988)8–28.

[2] Council Decision of 25 April 2002 concerning the approval, on behalf of theEuropean Community, of the Kyoto Protocol to the United FrameworkConvention on Climate Change and the joint fulfilment of commitmentsthereunder, Official J. L 130 (2002) 1–20.

[3] J.W. Furlong, The Air Pollution Consultant 11/12 (1994) 1.12–1.14.[4] K. Tanno, M. Itoh, H. Sekiya, H. Yashiro, N. Kumagai, The corrosion inhibition of

carbon steel in lithium bromide solution by hydroxide and molybdate atmoderate temperatures, Corros. Sci. 34 (1993) 1453–1461.

[5] V. Oleinik, I.K. Yu, A.R. Vartapetyan, Corrosion inhibition of steel in lithiumbromide brines, Prot. Met. 39 (2003) 12–18.

[6] E. Sarmiento, J.G. González-Rodriguez, J. Uruchurtu, O. Sarmiento, M.Menchaca, Fractal analysis of the corrosion inhibition of carbon steel in abromide solution by lithium chromate, Int. J. Electrochem. Sci. 4 (2009) 144–155.

[7] E. Samiento-Bustos, J.G. González-Rodriguez, J. Uruchurtu, V.M. Salinas-Bravo,Corrosion behavior of iron-based alloys in the LiBr + ethylene glycol + H2Omixture, Corros. Sci. 51 (2009) 1107–1114.

[8] D. Itzhak, O. Elias, Behavior of type-304 and type-316 austenitic stainless-steels in 55-percent lithium bromide heavy brine environments, Corrosion 50(1994) 131–137.

[9] S. Zor, M. Soncu, L. Çapan, Corrosion behavior of G-X CrNiMoNb 18–10austenitic stainless steel in acidic solutions, J. Alloy Compd. 480 (2009) 885–888.

R. Sánchez-Tovar et al. / Corrosion Science 53 (2011) 2598–2610 2609

[10] B.T. Lu, Z.K. Chen, J.L. Luo, B.M. Patchett, Z.H. Xu, Pitting and stress corrosioncracking behavior in welded austenitic stainless steel, Electrochim. Acta 50(2005) 1391–1403.

[11] B. Maier, G.S. Frankel, Pitting corrosion of bare stainless steel 304 underchloride solution droplets, J. Electrochem. Soc. 157 (2010) C302–C312.

[12] N.J. Laycock, R.C. Newman, Localised dissolution kinetics, salt films and pittingpotentials, Corros. Sci. 39 (1997) 1771–1790.

[13] A. Pardo, M.C. Merino, A.E. Coy, F. Viejo, R. Arrabal, E. Matykina, Pittingcorrosion behaviour of austenitic stainless steels – combining effects of Mnand Mo additions, Corros. Sci. 50 (2008) 1796–1806.

[14] F.E.-T. Heakal, M.M. Hefny, A.M.A. El-Tawab, Electrochemical behavior of 304Lstainless steel in high saline and sulphate solutions containing alga Dunaliellasalina and [beta]-carotene, J. Alloy Compd. 491 (2010) 636–642.

[15] J.F. Lancaster, Metallurgy of Welding, sixth ed., Woodhead Publishing Limited,Cambridge, 1999.

[16] S. Lamb, Stainless Steels and Nickel Alloys, second ed., Casti Publishing Inc.,Canada, 2002.

[17] O. Martín, P. De Tiedra, M. López, M. San-Juan, C. García, F. Martín, Y. Blanco,Quality prediction of resistance spot welding joints of 304 austenitic stainlesssteel, Mater. Des. 30 (2009) 68–77.

[18] D. Özyürek, An effect of weld current and weld atmosphere on the resistancespot weldability of 304L austenitic stainless steel, Mater. Des. 29 (2008) 597–603.

[19] F. KarcI, R. Kaçar, S. Gündüz, The effect of process parameter on the propertiesof spot welded cold deformed AISI304 grade austenitic stainless steel, J. Mater.Process. Technol. 209 (2009) 4011–4019.

[20] T. Hemmingsen, H. Hovdan, P. Sanni, N.O. Aagotnes, The influence ofelectrolyte reduction potential on weld corrosion, Electrochim. Acta 47(2002) 3949–3955.

[21] D.J. Lee, K.H. Jung, J.H. Sung, Y.H. Kim, K.H. Lee, J.U. Park, Y.T. Shin, H.W. Lee,Pitting corrosion behavior on crack property in AISI 304L weld metals withvarying Cr/Ni equivalent ratio, Mater. Des. 30 (2009) 3269–3273.

[22] P.D. Bilmes, C.L. Llorente, C.M. Méndez, C.A. Gervasi, Microstructure, heattreatment and pitting corrosion of 13CrNiMo plate and weld metals, Corros.Sci. 51 (2009) 876–881.

[23] R. Kumar, V.K. Tewari, S. Prakash, Oxidation behavior of base metal, weldmetal and HAZ regions of SMAW weldment in ASTM SA210 GrA1 steel, J. AlloyCompd. 479 (2009) 432–435.

[24] W. Luo, The corrosion resistance of 0Cr19Ni9 stainless steel arc welding jointswith and without arc surface melting, Mater. Sci. Eng. A – Struct. 345 (2003) 1–7.

[25] E.O. Correa, S.C. Costa, J.N. Santos, Weldability of iron-based powder metalmaterials using pulsed plasma arc welding process, J. Mater. Process. Technol.198 (2008) 323–329.

[26] I.D. Harris, Plasma arc welding, in: D.L. Olson, T.A. Siewert, S. Liu, G.R. Edwards(Eds.), ASM Handbook Volume 6: Welding, Brazing and Soldering, ASMInternational, USA, 1993, pp. 195–199.

[27] F. Karimzadeh, M. Salehi, A. Saatchi, M. Meratian, Effect of microplasma arcwelding process parameters on grain growth and porosity distribution of thinsheet Ti–6Al–4V alloy weldment, Mater. Manuf. Process. 20 (2005) 205–219.

[28] F. Karimzadeh, A. Ebnonnasir, A. Foroughi, Artificial neural network modelingfor evaluating of epitaxial growth of Ti6Al4V weldment, Mater. Sci. Eng. A –Struct. 432 (2006) 184–190.

[29] F. Javidrad, H. Farghadani, A. Asgharzadeh, Microplasma arc welding of AISI304 stainless steel for use in aerospace industries, Modaress J. Sci. Eng. 31(2008) 67–77.

[30] F. Javidrad, R. Rahmati, An integrated re-engineering plan for themanufacturing of aerospace components, Mater. Des. 30 (2009) 1524–1532.

[31] W. Luo, Effect of micro-plasma arc melting on the corrosion resistance of a0Cr19Ni9 stainless steel SAW joint, Mater. Lett. 55 (2002) 290–295.

[32] L. Jeffus, Welding. Principles and Applications, fifth ed., Delmar ThomsonLearning, USA, 1997.

[33] G.F. Vander Voort, Metallography. Principles and Practice, second ed., ASMInternational, USA, 1999.

[34] ASTM E-384, Standard Test Method for Microindentation Hardness ofMaterials, ASTM International, Pennsylvania, 1999.

[35] R. Leiva-García, J. García-Antón, M.J. Muñoz-Portero, Contribution to theelucidation of corrosion initiation through confocal laser scanning microscopy(CLSM), Corros. Sci. 52 (2010) 2133–2142.

[36] C. Wagner, W. Traud, Über die Deutung von Korrosionsvorgängen durchÜberlagerung von elektrochemischen Teilvorgängen und über diePotentialbildung an Mischelektroden, Z.f. Elektroch. 44 (1938) 391–402.

[37] M.G. Pujar, N. Parvathavarthini, R.K. Dayal, Influence of solution-annealing andstress-relieving on the pitting corrosion resistance of modified 316N SS weldmetals: a study using EN technique, Mater. Chem. Phys. 123 (2010) 407–416.

[38] C. Garcia, F. Martin, P. De Tiedra, Y. Blanco, M. Lopez, Pitting corrosion ofwelded joints of austenitic stainless steels studied by using an electrochemicalminicell, Corros. Sci. 50 (2008) 1184–1194.

[39] Y. Cui, C.D. Lundin, Evaluation of initial corrosion location in E316L austeniticstainless steel weld metals, Mater. Lett. 59 (2005) 1542–1546.

[40] I.H. Lo, W.T. Tsai, Effect of heat treatment on the precipitation and pittingcorrosion behavior of 347 SS weld overlay, Mater. Sci. Eng. A – Struct. 355(2003) 137–143.

[41] M.G. Pujar, R.K. Dayal, T.P.S. Gill, S.N. Malhortra, Role of delta-ferrite in thedissolution of passive films on the austenitic stainless-steel weld metals, J.Mater. Sci. Lett. 18 (1999) 823–826.

[42] G. Conde y Santiago, Aceros inoxidables, refractarios y criogenicos, first ed.,Interciencia, 1971.

[43] J.A. Brooks, J.C. Williams, A.W. Thompson, Microstructural origin of the skeletalferrite morphology of austenitic stainless steel welds, Metall. Mater. Trans. A14 A (1983) 1271–1281.

[44] K. Rajasekhar, C.S. Harendranath, R. Raman, S.D. Kulkarni, Microstructuralevolution during solidification of austenitic stainless steel weld metals: a colormetallographic and electron microprobe analysis study, Mater. Charact. 38(1997) 53–65.

[45] ASTM G-5, Test Method for Making Potenciostatic and PotenciodynamicAnodic Polarization Measurements, ASTM International, Pennsylvania, 2004.

[46] E. McCafferty, Validation of corrosion rates measured by the Tafelextrapolation method, Corros. Sci. 47 (2005) 3202–3215.

[47] M.A. Amin, S.S.A. El Rehim, H.T.M. Abdel-Fatah, Electrochemical frequencymodulation and inductively coupled plasma atomic emission spectroscopymethods for monitoring corrosion rates and inhibition of low alloy steelcorrosion in HCl solutions and a test for validity of the Tafel extrapolationmethod, Corros. Sci. 51 (2009) 882–894.

[48] M.A. Amin, K.F. Khaled, S.A. Fadl-Allah, Testing validity of the Tafelextrapolation method for monitoring corrosion of cold rolled steel in HClsolutions – experimental and theoretical studies, Corros. Sci. 52 (2010) 140–151.

[49] E. Bardal, Corrosion and Protection, Engineering Materials and Processes,fourth ed., Springer, USA, 2004.

[50] R.G. Kelly, J.R. Scully, D.W. Shoesmith, R.G. Buchheit, ElectrochemicalTechniques in Corrosion Science and Engineering, first ed., Marcel Dekker,New York, 2003.

[51] M.A. Ameer, A.M. Fekry, F.E.-T. Heakal, Electrochemical behaviour of passivefilms on molybdenum-containing austenitic stainless steels in aqueoussolutions, Electrochim. Acta 50 (2004) 43–49.

[52] A.B. Silva, S.M.L. Agostinho, O.E. Barcia, G.G.O. Cordeiro, E. D’Elia, The effect ofcysteine on the corrosion of 304L stainless steel in sulphuric acid, Corros. Sci.48 (2006) 3668–3674.

[53] C. Lavos-Valereto, I. Costa, S. Wolynec, The electrochemical behavior of Ti–6Al–7Nb alloy with and without plasma-sprayed hydroxyapatite coating inHank’s solution, J. Biomed. Mater. Res. A 63 (2002) 664–670.

[54] S.L. De Assis, S. Wolynec, I. Costa, Corrosion characterization of titanium alloysby electrochemical techniques, Electrochim. Acta 51 (2006) 1815–1819.

[55] B. Jegdic, D.M. Drazic, J.P. Popic, Open circuit potentials of metallic chromiumand austenitic 304 stainless steel in aqueous sulphuric acid solution and theinfluence of chloride ions on them, Corros. Sci. 50 (2008) 1235–1244.

[56] M. Kaneko, H.S. Isaacs, Pitting of stainless steel in bromide, chloride andbromide/chloride solutions, Corros. Sci. 42 (2000) 67–78.

[57] M. Kaneko, H.S. Isaacs, Effects of molybdenum on the pitting of ferritic- andaustenitic-stainless steels in bromide and chloride solutions, Corros. Sci. 44(2002) 1825–1834.

[58] E.A. Abd El Meguid, N.A. Mahmoud, Inhibition of bromide – pitting corrosion oftype 904L stainless steel, Corrosion 59 (2003) 104–111.

[59] G.T. Burstein, S.P. Vines, Repetitive nucleation of corrosion pits on stainlesssteel and the effects of surface roughness, J. Electrochem. Soc. 148 (2001)B504–B516.

[60] C.-O.A. Olsson, D. Landolt, Passive films on stainless steels – chemistry,structure and growth, Electrochim. Acta 48 (2003) 1093–1104.

[61] M.P. Ryan, D.E. Williams, R.J. Chater, B.M. Hutton, D.S. McPhail, Why stainlesssteel corrodes, Nature 415 (2002) 770–774.

[62] M. Dadfar, M.H. Fathi, F. Karimzadeh, M.R. Dadfar, A. Saatchi, Effect of TIGwelding on corrosion behavior of 316L stainless steel, Mater. Lett. 61 (2007)2343–2346.

[63] W.E. White, Observations of the influence of microstructure on corrosion ofwelded conventional and stainless steels, Mater. Charact. 28 (1992) 349–358.

[64] T. Laitinen, M. Bojinov, I. Betova, K. Mäkelä, T. Saario, The Properties of andTransport Phenomena in Oxide Films on Iron, Nickel, Chromium and theirAlloys in Aqueous Environments, STUK-YTO-TR 150.STUK, Helsinki, 1999.

[65] A. Igual-Muñoz, J. García-Antón, J.L. Guiñón, V. Pérez-Herranz, Effects ofsolution temperature on localized corrosion of high nickel content stainlesssteels and nickel in chromated LiBr solution, Corros. Sci. 48 (2006) 3349–3374.

[66] A. Al-Odwani, J. Carew, M. Al-Tabtabaei, A. Al-Hijji, Materials performance inSWRO desalination plant at KISR research and development program,Desalination 135 (2001) 99–110.

[67] T. Bellezze, G. Roventi, R. Fratesi, Electrochemical characterization of threecorrosion-resistant alloys after processing for heating-element sheathing,Electrochim. Acta 49 (2004) 3005–3014.

[68] M.M. Aghamirian, W.T. Yen, Mechanisms of galvanic interactions betweengold and sulfide minerals in cyanide solution, Miner. Eng. 18 (2005) 393–407.

[69] G.O. Davis, J. Kolts, N. Sridhar, Polarization effects in galvanic corrosion,Corrosion 42 (1986) 329–336.

[70] L. Reclaru, R. Lerf, P.Y. Eschler, A. Blatter, J.M. Meyer, Pitting, crevice andgalvanic corrosion of REX stainless-steel/CoCr orthopedic implant material,Biomaterials 23 (2002) 3479–3485.

[71] F. Mansfeld, Z. Sun, C.H. Hsu, A. Nagiub, Concerning trend removal inelectrochemical noise measurements, Corros. Sci. 43 (2001) 341–352.

[72] R. Akid, D.J. Mills, A comparison between conventional macroscopic and novelmicroscopic scanning electrochemical methods to evaluate galvanic corrosion,Corros. Sci. 43 (2001) 1203–1216.

2610 R. Sánchez-Tovar et al. / Corrosion Science 53 (2011) 2598–2610

[73] G. Kear, B.D. Barker, K.R. Stokes, F.C. Walsh, Corrosion and impressed currentcathodic protection of copper-based materials using a bimetallic rotatingcylinder electrode (BRCE), Corros. Sci. 47 (2005) 1694–1705.

[74] L. Reclaru, R. Lerf, P.Y. Eschler, J.M. Meyer, Corrosion behavior of a weldedstainless-steel orthopedic implant, Biomaterials 22 (2001) 269–279.

[75] E. Otero Huerta, Corrosión y degradación de materiales, first ed., Síntesis,Madrid, 1997.

[76] F. Mansfeld, J.V. Kendel, Laboratory studies of Galvanic corrosion of aluminiumalloys, in: R. Baboian, W.D. France Jr., L.C. Roew, J.F. Rynewicz (Eds.), Galvanicand Pitting Corrosion-Field and Laboratory Studies, ASTM STP 576, ASTM,Philadelphia, 1976, pp. 20–47.

[77] F.T. Cheng, K.H. Lo, H.C. Man, NiTi cladding on stainless steel by TIG surfacingprocess Part II. Corrosion behavior, Surf. Coat. Technol. 172 (2003) 316–321.

[78] C.T. Kwok, S.L. Fong, F.T. Cheng, H.C. Man, Pitting and galvanic corrosionbehavior of laser-welded stainless steels, J. Mater. Process. Technol. 176 (2006)168–178.

[79] G.T. Burstein, C. Liu, R.M. Souto, S.P. Vines, Origins of pitting corrosion, Corros.Eng. Sci. Technol. 39 (2004) 25–30.

[80] G.T. Burstein, Electrochemistry of pit formation and growth, Electrochem. Soc.Proc. 19 (2004) 1–12.

[81] J. Mankowski, Z. Szklarska-Smialowska, Studies on accumulation of chlorideions in pits growing during anodic polarization, Corros. Sci. 15 (1975) 493–501.

[82] J.R. Galvele, Transport processes and the mechanism of pitting of metals, J.Electrochem. Soc. 123 (1976) 464–474.

[83] C. Punckt, M. Bolscher, H.H. Rotermund, A.S. Mikhailov, L. Organ, N. Budiansky,J.R. Scully, J.L. Hudson, Sudden onset of pitting corrosion on stainless steel as acritical phenomenon, Science 305 (2004) 1133–1136.

[84] M. Dornhege, C. Punckt, J.L. Hudson, H.H. Rotermund, Spreading of corrosionon stainless steel. simultaneous observation of metastable pits and oxide film,J. Electrochem. Soc. 154 (2007) C24–C27.

[85] A.S. Mikhailov, J.R. Scully, J.L. Hudson, Nonequilibrium collective phenomenain the onset of pitting corrosion, Surf. Sci. 603 (2009) 1912–1921.

[86] C. Lin, X. Li, C. Dong, Pitting and galvanic corrosion behaviour of stainless steelwith weld in wet–dry environment containing Cl�, J. Univ. Sci. Technol. Beijing14 (2007) 517–522.

[87] B.T. Timofeev, G.P. Karzov, A.A. Gorbakony, Y. Nikolaev, Corrosion andmechanical strength of welded joints of downcomers for RBMK reactors, Int.J. Pres. Ves. Pip. 76 (1999) 299–307.

[88] C.T. Kwok, K.H. Lo, F.T. Cheng, H.C. Man, Effect of processing conditions on thecorrosion performance of laser surface-melted AISI 440C martensitic stainlesssteel, Surf. Coat. Technol. 166 (2003) 221–230.

[89] N. Sato, The stability of localized corrosion, Corros. Sci. 37 (1995)1947–1967.

[90] P.C. Pistorius, G.T. Burstein, Metastable pitting corrosion of stainless steel andthe transition to stability, Philos. Trans. R. Soc. A 341 (1992) 531–559.