Characterization of Welded Joints (MIG and SAW) on LDX 2101 Duplex SS

Ubertalli Graziano; Donato Firrao; Gianmarco Taveri

CrossMark

Available online at www.sciencedirect.com

ScienceDirect

Procedía Engineering 109 (2015) 484 - 491

Procedía Engineering

www.elsevier.com/locate/procedia

XXIII Italian Group ofFracture Meeting, IGFXXIII

Characterization of weldedjoints (MIG and SAW) on LDX 2101

Duplex SS

Ubertalli Graziano*, Donato Firrao, GianmarcoTaveri

DISATdpt., Politecnico di Torino, Corso Duca degli Abruzzi 24, Torino 10129 Torino, Italy

Abstract

LDX 2101 (21,5%Cr, 5%Mn, l,5%Ni and 0,3%Mo) duplex stainless steel 255x123mm sheets, 10mm thick, have been welded by using either MIG (Metal Inert Gas) or SAW (Submerged Arc Welding) processes with appropriate filler material. It has been performed, in base material (BM), heat affected zone (HAZ) and fusion zone (FZ), metallographic and micrographic characterization (WDS image mapping, EDS chemical analysis, image analysis), evaluation of mechanical and physical properties (Vickers micro-hardness, tensile and impact testing) followed by SEM analysis and observation of fracture profile. WDS mapping images evidence alloying elements partition between ferritic and austenitic phases in the different zone of the welded sheets. The mechanical properties are even guaranteed in case of welding process, while the resilience values evidenced scattered values, however always above 40 J. No peculiar differences have been detected between the two employed welding processes, i.e. MIG and SAW.

(http://creativecommons.org/licenses/by-nc-nd/4.0/).

Peer-review under responsibility of the Gruppo Italiano Frattura (IGF)

Keywords: LDX-2101; Dulpex Stainless Steel; DSS Welding; FZ fracture morphology; Resilience test.

1. Introduction

Duplex stainless steels (DSS) are widely employed in industrial applications of mechanical and structural items, as well as for production of heat exchangers. As compared to the most classical and used stainless steels (SS), such as the austenitic ones, they offer a good general corrosion resistance added to the stress corrosion resistance and localized

* Corresponding author. E-mail address: graziano.ubertalli@polito.it

(http://creativecommons.Org/licenses/by-nc-nd/4.0/).

Peer-review under responsibility of the Gruppo Italiano Frattura (IGF)

doi:10.1016/j.proeng.2015.06.253

one[l], due to the simultaneous presence of ferrite together with austenite phases in microstructure. Remarkable corrosion properties had been noted for pitting and crevice corrosion type [2]. From this point of view, the PREN (Pitting Resistance Equivalent Number) value is commonly adopted, which gives an idea of the corrosion resistance level of the alloy (even though this parameter refers to pitting corrosion, it also suites for crevice). It is strictly related to the chemical composition of the material, since it is defined as follow [3]:

Those three alloying elements play a predominant role in corrosion resistance; indeed, the right percentage of those elements in melt could permit to optimize the Critical Pitting Temperature (CPT) and Critical Crevice Temperature (CCT), which are the lowest temperatures at which pitting and crevice corrosions take over respectively. In general, for these DSS, CCT turns out to be lower than the CPT, as supported from many researches [3]. In fact, some DSS and SDSS alloys, such as S31254 and S32750 (UNS grade designation), have the maximum CPT temperature, around 80°C, much higher than CCT temperature, that is 50°C. This difference is not that much remarkable for DSS alloys, as N08904 and S31803 (UNS grade designation), for which the difference rates around only a dozen of grades.

However, the outstanding feature which makes DSS far more competitive with respect to the austenitic SS are their mechanical properties (in terms of oy and ou), even triple of those related to ferritic and austenitic SS [4]. In case of welding processes, it is of relevant importance to properly balance the ferrite-austenite stabilizing alloying elements amount; mostly in DSS, these elements influence the microstructure of the welded joints in a way to guarantee the right content of ferrite and austenite phases in order to reach the required corrosion resistance and mechanical properties of the welded joint. Chemical composition defines the ferrite and austenite-stabilizing elements amount as follows:

The Welding Research Council (WRC), in 1992, issued a working diagram that permits to define the relative percentage of the two phases, after welding process, as a function of the heating temperature and the Creq/ Nieq ratio, at which corresponds different DSS alloys. By means of this diagram, it has been detected a considerable increase in ferrite content at 1050°C, whereas part of that transforms in austenite during cooling. Therefore, DSS are often heat treated at high temperature (usually solution heat treated between 1000°C and 1200°C) so as to transform the ferrite in austenite, according to a proper formation kinetic, strictly dependent upon the cooling rate [6].

It has been proved that a slightly higher concentration of nickel in welding bead confers better cryogenic properties: in particular, researches emphasize a lowering of the ductile to brittle transition temperature (DBTT) for DSS LDX 2101 in case of 5 or 6 wt% of nickel in melt, whereas the same material, with nickel content as low as 1.3 wt %, resulted in a higher DBTT temperature [5]. In welding it ought to pay attention to avoid the precipitation of undesired stable o, and metastable %, R, n and Cr2N phases, which normally transform into stable o-phase at room temperature; what could induce a drop of the material ductility. Researches besides this problem[6] reveals that the most dangerous temperature range to be avoided is 650 - 800°C, with a minimum in ductility of 37 J at 700°C. In order to avoid a long soaking time at this temperature range, the welding passes to provide for adequate cooling in HAZ is suggested to be below 150 °C. There are many research papers in the literature supporting investigation on mechanical properties and welding procedures for these DSS, with little insights concerning the comparison between microstructure and welding process from one side and mechanical properties of LDX 2101 from the other side. To investigate deeply in the argument, the research work is aimed to analyze the mechanical and metallographic characteristics of LDX 2101 Duplex SS plates welded with different technologies.

2. Experimental methods and results

LDX 2101 DSS plates (255x123mm and 10mm thick) were joined along the longer side by welding by the following techniques: MIG welding for plates 1 and 2, SAW welding for the third one instead. The base material and

PREN = (%Cr) + 3.3x(%Mo) +16x (%N)

Creq = (%Cr) + (%Mo) + 1.5x (%Si) + 0.7x(%Nb)

Nieq = (%Ni) + 35X (%C) + 20X (%N) + 0.5X (%Mn) + 0.25x (%Cu)

(2) (3)

filled material, analyzed with optical emission spectroscopy (OEM) show the chemical composition in wt.% reported in Table 1 and 2, respectively.

Table 1. Base material chemical composition

%c %Si %Mn %P %s %Cr %Mo %Ni %A1 %Co %Cu

0.062 0.716 5.272 0.016 0.003 21.14 0.167 1.35 <0.001 0.034 0.24

%Nb %Ti %V %w %Pb %Sn %As %B %N %Fe

0.008 0.009 0.059 0.013 <0.001 0.01 <0.001 0.002 0.236 Remainder

Table 2. Filler material chemical composition.

%Si %Mn %Cr %Ni %Mo %n"

~04 0^5 23^0 7^0 <0^5 oU

Filler material contains a higher percentage of nickel, an austenite-stabilizer element, in respect to the base material, in order to guarantee the right austenite content after the solidification of the weld bead. The weld plates (base material, HAZ and FZ) were tested by Vickers hardness, tensile and resilience tests as well as metallographic observation; the etched samples underwent optical and scanning electronic microscope analysis, X-Ray diffraction and EDS microprobe analysis together with WDS mapping images. Tensile and impact specimens have been cut off from the plates. Optical microscopy of different transverse section of grinded and polished unetched metallographic samples shows that the welding beads have no defects.

The welding transverse microstructure evidences a Heat Affected Zone (HAZ) characterized by a roundness of austenite grains with thickness ranging from 0.7 a 1.3 mm, depending on surface distance (figure lc). The fusion zone (FZ) shows a dendrite morphology with the primary arms oriented mostly along the heat gradient direction (that is perpendicular to the weld border). The dendritic microstructure is finer on the bead border where the cooling rate was higher, and getting coarser moving toward the core ofthe bead, where the solidification rate is slower.

Ferrite and austenite contents in the three zones have been calculated by using proper image analysis software (Leica Qwin®). The values for every plate and zone of measurement (average of 10 images) are reported in Table 3. Image binary analysis have been conducted in BM, HAZ and FZ.

Regarding the FZ, in Table 3 a set of values which represent categories of maximum and minimum results have been reported, and not only the average values, due to the breadth variability ofthe XRD peaks.

In fact, it results that there are some zones in FZ where concentration of ferrite is pretty lower in respect to the BM, in which ferrite and austenite are equally present, because ofthe higher concentration of austenite stabilizers in filler material. While HAZ is differently influenced by the heat flows coming from the weld bead, in relation to the welding technologies: plate 1, characterized by a MIG welding with one front and one back bead, shows a slight higher percentage of ferrite phase, due to the higher heat supply on HAZ followed by a fast cooling rate that transforms some

of the austenite in ferrite (see the WRC phase diagram). HAZ values of plates 2 and 3 do not vary that much tojustify this variation, providing that the difference falls in the admissible error in area integration of the images.

Table 3. Average values offerrite for the different welding zones and welding conditions.

Sample Base material HAZ max FZ min back

l°plate 50.3 60 43.4 \ \

2°plate 47 42.5 33.5 45

3°plate 46 36.7 28.5 47

XRD analysis and diffractograms were performed on the plates, by means of Rigaku D-Max machine with Co anticathode (Ka = 0.1789 nm). Thanks to that, it has been possible to quantify the ferrite percentage, i.e. around 49.5%: what pretty well confirmed what the image analysis demonstrated before on the base material. Likewise, XRD analysis conducted on transverse section of the welding zone did not evidence any precipitate phases k etc...),

attributed to heat treatments effects. Precipitates do not appear neither in metallographic observation of microstructures.

Vickers micro-hardness evaluation has been conducted on ferrite and austenite phases in all the main zones of each plate, at three different planes within the plate thickness. Vickers hardness load was 25g in order to obtain indentations with maximum dimension falling within a single-phase crystal (figure 2). For each plate, six Vickers micro-hardness trials were performed and the mean value was calculated and reported in table 4.

O : C?

" < - ■ ■ >/*

Y -r *.. r u<

Fig. 2. a) Ferrite; b) Austenite Table 4. Average micro-hardness values offerrite and austenite in the different zones of the welded plates

Base material

Ferrite Austenite Ferrite Austenite Ferrite Austenite

Plate 1 229 293 262 314 239 282

Plate 2 220 295 226 230 259 306

Plate 3 225 303 263 319 264 302

The austenite is generally harder with regard to the ferrite phase everywhere. In fact, austenite micro-hardness values rate around 300 HV0.025, whereas about 250HVo.o25 has been measured for ferrite. Ferrite micro-hardness results are higher in FZ and HAZ, whereas those of austenite turns out to be slightly higher only in HAZ, with the only exception of HAZ of the second plate where austenite and ferrite have the same hardness, with a value far lower than the HAZ pertaining to the other two samples.

Micro-analysis results conduced on transverse section of the plates allowed to get phase compositional insights and

elements distribution and partition in ferrite and austenite, that is represented by the R coefficient, R= % Creq / % Nieq, namely the ratio between the ferrite stabilizers content and the austenite stabilizer content. Mappings of chromium, nickel and manganese have been performed and reported in Figure 3 for BM, likewise for the HAZ zone and FZ.

a b c d

ïfSg Ils»

Fig. 3. Electron microscope image of the etched base material a) and the corresponding WDS image mapping of b) chromium, c) nickel and d) manganese element distribution.

Higher concentration of a given alloying element is denoted by a clearer coloration of the WDS mapping image. Irrespective of real percentage of each element, chromium counts are more intense in ferrite phase, whereas nickel and manganese take place more in austenite phase. Figure 3 shows a predominant correspondence between the secondary electrons image (figure 3a) and the corresponding elements mapping images (figures 5b, c and d), which neatly distinguish the different phases they stabilize.

Table 5 Alloying elements spot chemical detection points.

Base material Welding zone

Phase austenite ferrite austenite ferrite

R coefficient 4.9 6.5 3.8 4.8

Microprobe analysis conducted on the different phases on different zone (BM, FZ) yields the chemical compositions that were adopted to calculate the R coefficient, for austenite and ferrite, as explained above, Table 5. The R coefficient results always to be lower for austenite phase than ferrite one, for both BM and FZ, since higher amount of Ni makes Nieq higher in value. In BM, the R coefficients are higher than those of FZ because of lower Ni% not properly compensated by the Mn content. It must be stated that calculated R coefficient have been assessed without considering N and C elements, because sizable error in wt.% calculation of these elements induces a wide range of uncertainity and however these element are both not considered in the two phases and zones.

Tensile tests were effected to evaluate the mechanical and physical properties of the welded LDX 2101 DSS plates. The results of the tensile tests conducted on base material and on dumbbell shape with a round junctions tensile test specimens are reported in Table 6 in terms ofstrength, yield strength and elongation properties. The given results are average values and the scattering of the obtained data is narrow either among specimens of the same type and welding conditions or among the different welded plates. The elongation data reveal for the welded material, a wider scattering ofvalues in respect to the base material instead.

Metallographic samples, transverse to the fracture profile of the tensile specimens, were prepared in order to reveal crack onset occurrence; it generally starts in FZ and sometimes propagates in BM, also evidencing a considerable necking. The optical analysis of the fracture profiles show that the samples with lower elongation correspond to crack fully propagation in FZ, while the higher values are detected in samples with crack propagation mainly in BM. Micrographies at higher magnification evidence that, wherever crack propagated through the BM, fracture shows an indented profile, sometimes it follows the austenite and ferrite grains orientation, parallel to plate surfaces, sometimes it propagates transversely, with a sharper deformation of the austenite phase. In case of crack propagation through FZ,

Ubertalli Graziano et al. /Procedia Engineering 109 (2015) 484 - 491 a straighter fracture profile is observed, which follows the finer, primary and secondary, dendrite grains.

Table 6. Base material and welded material properties.

Mechanical properties Base material Welded components

= 750 MPa 730 - 750 MPa

°y = 550 MPa 510 - 525 MPa

E% = 29 - 32% 20% - 32%

Young's Modulus E= 190 GPa /

The fracture morphology analyzed by electron microscope (SEM) (Figure 4), shows ductile morphology of fracture and some grains with brittle behavior (fig.4a). The crack propagation is always transgranular. It is quite predominant the presence of dimples everywhere on the fracture surface (fig.4c). Smooth zones accompany dimples in morphologies (fig.4b), owing to plastic shear deformation.

Fig. 4 Fracture morphology of plate 1 -A specimen after tensile test in different zones of the fracture surface with mainly ductile morphology

Impact testing, performed at room temperature, involved BM, either tested in longitudinal or in transvers thickness impact direction in respect to the plate, and the welding zone of the different plates. The results of impact tests conducted on Charpy V-notched standard samples are reported in Table 7.

It has been observed that BM evidence high resilience values, though they result to be different in dependence upon the direction of notch incision, because of the preferential orientations of the austenite and ferrite grains, due to the lamination process.

Table 7: Resilience testing results.

Specimen Impact Load (J)

BM Longitudinal 150-166

Transversal 82-94

FZ 41-85

These results are confirmed by transverse micrographies of the fracture profile, relative to BM (fig.5a) and FZ (fig.7b), which evidenced such a "delamination" of the longitudinal tested specimens. Impact testing related to the welded material experienced a wider scattering ofthe data, and the crack always initiated in FZ, in the V-notch edge,

and in some cases, it propagated within HAZ first and then in BM. The observation of transverse fracture metallographic specimens show a quite linear crack propagation, with low plastic deformation of grains in BM resilience testing; just small plastic deformation involves the austenite grains in welded material impact test; fracture profile seems to follow the ferrite-austenite grain interfaces. The length of fracture profile is connected with impact values with a good agreement.

IPS, ft/ ft > 1

Tirr mm

mm- ww

Fig.5. Points where spot chemical analysis has been performed.

SEM observations effected on fracture surfaces of impact welded samples of plate 1 show a mixed ductile and brittle fracture morphology (Fig. 5, as example) in almost all samples. On some samples, spot micro-analysis were conducted in different areas (indicated as small square numbered in Fig. 5) that evidenced ductile or brittle behavior to calculate the R coefficient (as previously defined) and therefore to know the phase that evidence a certain fracture morphology. Effectively, in case of ductile fracture morphology the R coefficient values range between Rmin=4.50 and Rmax=4.83, whereas for brittle ones the mentioned parameter ranges between Rmm=5.32 and Rmax=5.75. The average values ofR coefficient detected for many ductile and brittle morphologies are reported in Table 8.

Table 8. Spot chemical analysis results.

Ductile morphology Brittle morphology

R coefficient 4.65 5.50

3. Conclusions

Giving a rough overview, MIG and SAW welding are both valid techniques to weld with appreciable properties LDX 2101 DSS plates, but it seems that there is no likely difference in mechanical properties between MIG welding (one or more beads) and SAW welding: tensile and resilience test results do not give any hint about this argument. Image analysis does instead, since FZ of third plate, jointed by SAW welding, revealed a lack of ferrite-austenite balance of content, giving that the cooling rate is slightly slower; the same thing can be stated for the second plate (multi beads MIG welding). Mechanical properties evidences do not show any difference whether testing base material or welded components. The latter broke predominantly in FZ while tensile testing, showing a relevant necking right before the failure. Metallographic observations evidenced a higher deformation of the austenite phase around the fracture zone, and crack mainly propagates in a ductile manner along the ferrite phase. Impact test confirm the lower ductility of the ferrite phase, involving at times a brittle morphology.

References

[1] R. N. Gunn, Duplex Stainless Steels - Microstructure, properties and applications, Abington Publishing, Cambridge, England, 1997.

[2] S. Bemhardsson, Duplex stainless steel91, Beune, France, Les édition de physique, 185.

[3] F. Bonollo, A. Tiziani, P. Ferro, Evoluzione Microstrutturale di Acciai Duplex e Superduplex in Relazione ai Processi di Saldatura, DTG-Università di Padova, Metallurgia Italiana, (2005) 27-38.

[4] Z. Wei, J. Laizhu, H. Jincheng, S. Hongmei, Study ofMechanical and Corrosion Properties ofa LDX 2101 DSS, Baosteel Research Institute, Shangai, China, Elsevier, Materials Science &Engineering A. 497 (2008) 501-504.

[5] J. Philagen, R. Sandstrom, Influence ofNickel on the Toughness ofLean Duplex Stainless Steel Welds, Royal Institute ofTechnology, Stocholm, Sweden, Elsevier, Materials Science &Engineering A. 602 (2013) 49-57.

[6] Z. Wei, J. Laizhu, H. Jincheng, S. Hongmei, Effect of Ageing on Precipitation and Impact Energy of2101 Economical Duplex Stainless Steel, Shangai Jiaotong University, Shanghai, China, Elsevier, Materials Characterization. 60 (2009) 50-55.

Characterization of Welded Joints (MIG and SAW) on LDX 2101 Duplex SS Academic research paper on "Materials engineering"

Abstract of research paper on Materials engineering, author of scientific article — Ubertalli Graziano, Donato Firrao, Gianmarco Taveri

Similar topics of scientific paper in Materials engineering , author of scholarly article — Ubertalli Graziano, Donato Firrao, Gianmarco Taveri

Academic research paper on topic "Characterization of Welded Joints (MIG and SAW) on LDX 2101 Duplex SS"