Structural Characterization of C-Mn Steel Laser Beam...
7
Click here to load reader
Transcript of Structural Characterization of C-Mn Steel Laser Beam...
ABSTRACT. The possibility of performingsuccessfully sound welds on steel usingthe laser beam welding (LBW) processwith powder filler metal has been used toinvestigate possible advantages regard-ing an easier control of the compositionof the welded zone. In this work, LBWjoints were carried out on C-Mn steelplates with prepared square and bevelededges. Ni filler powder was used as atracer of the powder penetration insidethe joints. The examination of welds in-cluded metallographic work by opticaland scanning electron microscopy(SEM), with microanalysis by energy-dispersive spectroscopy, to obtain indi-cations on powder distribution. Mechan-ical properties were evaluated by Vickersmicrohardness tests.
Introduction
Welding with powder filler metal,using conventional techniques such asarc welding (Ref. 1) and submerged arcwelding (Refs. 2–4), could enable en-hanced deposition rates. Moreover, thepowder addition can offer advantages inthe chemical composition of the fillermaterial, particularly where a controlledcomposition of the welded zone (WZ) isrequired or when welding materials arenot easily combinable. Powder fillermetal should allow an improvement ofprocess stability as compared to the risksof incomplete fusion and sticking whenusing a wire filler material.
In this work, the LBW process hasbeen used to verify the possibility of uti-lizing powder filler metal. The industrialapplication of this method could be of in-terest in the construction of vessels,equipment of large dimensions, ships,
etc., thanks to the peculiar characteristicsof laser beam welding, i.e., easy automa-tion of the process, reduction of the heat-affected zone and minimizing distortionsand residual stresses.
The results of metallurgical investiga-tions on C-Mn steel butt-joint welds, withtwo different preparations (square andbeveled edges) are given. Observationsby optical and scanning electron mi-croscopy with microanalysis, by energydispersive spectroscopy (EDS), were per-formed to evaluate the depth of powderpenetration and the contribution of pow-der to the formation of a melted pool. Themechanical properties were measured byVickers microhardness tests.
Materials and Methods
Materials
Laser beam welding tests were per-formed on low-alloy C-Mn steel plates,10 mm thick, specified as L24N (corre-sponding to a steel type AISI 15XX withan extra-low content of P and S impuri-ties), having the chemical compositionreported in Table 1. The plates were sup-plied in the normalized condition with aferritic-pearlitic microstructure.
Commercially pure Ni powder (parti-
cle sizes in the range 20–50 µm), trans-ported by an Ar flow, was employed asfiller material. The Ni powder was uti-lized as a tracer because it allows aneasy investigation into the distribution ofthe filler material and the evaluation ofthe penetration depth along the jointthickness.
Laser Beam Welding Procedure
The laser system was a continuousCO2 unit, type TLF12000, with 9 kW ofnet power on the workpiece. The focus-ing device was a water-cooled, copperparaboloid mirror. The high power den-sity, focused on a spot of dimension of theorder of 0.5 mm, leads to the keyholemechanism of energy transfer to theworkpiece.
A sketch of the process is shown inFig. 1. The system worked in robotic stylewith the plates placed in the flat positionon a moving table that, translating belowthe laser head, allows the bead formationin a single pass.
A nozzle with three tubes, arranged asshown in Fig. 1, was utilized for the var-ious gas flows. The welded zone protec-tion was obtained by a primary Ar flow(25 L/min) through tube 3; the powdertransportation by a secondary Ar flow (5 L/min) through tube 1; and the plasmacontrol was performed by a He flow (10 L/min) through tube 2.
Many preliminary experiments wereperformed to optimize the geometric pa-rameters that have a remarkable influ-ence on the powder penetration insidethe melted metal (Ref. 5). The nozzle wasdirected just above the interaction zoneof the laser beam and workpiece, in prox-imity of the melted pool, with an incli-nation angle of 37 deg. This value was as-sumed in the middle of the range 30–45deg in which plasma control and stabil-ity can be optimized. Moreover, the pow-der flow interaction with the workpiece
WELDING RESEARCH SUPPLEMENT | 317-s
RE
SE
AR
CH
/DE
VE
LO
PM
EN
T/R
ES
EA
RC
H/D
EV
EL
OP
ME
NT
/RE
SE
AR
CH
/DE
VE
LO
PM
EN
T/R
ES
EA
RC
H/D
EV
EL
OP
ME
NT
Structural Characterization of C-Mn Steel LaserBeam Welded Joints with Powder Filler Metal
BY S. MISSORI AND A. SILI
Weld quality of butt-joint welds with beveled and square edges is investigated
KEY WORDS
C-Mn SteelPowder Filler MetalLaser Beam Welds (LBW)Ni Powder TracingWelded Zone (WZ)Heat-Affected Zone (HAZ)S. MISSORI is with Department Mechanical
Engineering, University of Rome, Tor Vergata,Italy. A. SILI is with Department Industrial,Chemical and Materials Engineering, Univer-sity of Messina, Italy.
and the powder penetration inside themelted pool depend on nozzle and laserbeam focus positions. With reference toFig. 1, the powder trajectory, if approxi-mated by the axis of the smaller tube in-tersects the workpiece surface at distanceξ from the laser beam axis, has an opti-
mized value equal to 5.8 mm. The dis-tance z between the laser beam focusand workpiece surface has an optimizedvalue equal to 1 mm, and the distance xbetween the nozzle center and the laserbeam axis equal to 8 mm. In any case, ac-cording to the process sketched in Fig.1,an adjustment of nozzle positioning canbe done by a little translation along itsaxis that changes the distance x, butkeeps constant the distance ∆ξ betweenthe powder impingement point and thelaser beam axis.
Preliminary bead-on-plate tests werepermitted to optimize the process para-meters, such as the powder rates in therange 20–60 g/min or the welding speedin the range 0.5–1 m/min.
Butt-joint welds were performed onplates with square edges, with a rootopening in the range 0–0.4 mm (prepa-ration A), or beveled edge preparation,with bevel angle α = 3, 6, 9 deg and noopening at the root (preparation B), as re-ported in Fig. 2. The welding speed andpowder rate values are given in Table 2.
Experimental Work
The welded plates were cut trans-versely to the bead to obtain samples ofthe welded sections. The samples weremetallographically prepared by abrasive
paper, pol-ished by finealumina andthen etched by2% Nital.
The follow-ing experi-mental work
was performed on the metallographicsamples.
A) Visual and macrographic inspec-tions were performed. A minimum num-ber of three macrographic samples pereach procedure were cut and examined.The WZ area was measured for eachsample.
B) Optical/SEM observations and mi-croanalysis by EDS were performed onmetallographic samples taken from eachof the welding procedures in order to an-alyze WZ morphology and assess the Nicontent along the thickness of the welds.The Ni content was measured along thecenterline of the WZ by an electronprobe in a SEM type JSM-35 CF,equipped with an energy-dispersivespectrometer model EDAX 711, at an ac-celerating voltage of 15 kV. The depthand diameter of the interaction pair wereestimated at about 0.5 and 1 µm, re-spectively. The distance between twomicroanalysis measurements was equalto about 0.5 mm.
c) Vickers microhardness tests wereperformed to evaluate the mechanicalproperties in the welded joints. Measure-ments were extended to the entire thick-ness along the centerline of the WZ andto the heat-affected zone (HAZ). Vickerstests were performed with a load of 5 Nand a loading time of 10 s.
Results
Visual and Macrographic Inspection
The samples submitted to visual andmacrographic inspections confirmed the
possibility of having sound welds with agood repeatability (Figs. 3, 4) for bothjoint preparations. In almost all cases, thepenetration was complete along all theplate thickness, and the welds were freefrom cracks or other defects. However,occasional microcracks in the WZ wererevealed. The average values of the WZarea, measured in each sample, are givenin Table 3.
Samples welded with square edgepreparation show inhomogeneities in thesolidification structure, characterized byirregular banding in the WZ — Fig. 3.
Samples welded with beveled edgepreparation show the best appearance ofthe WZ at visual and macrographic ex-amination; the WZ is alloyed enough inNi to be only weakly etched by Nital —Fig. 4.
Generally, the size and shape of theferritic steel HAZ is regular and clearlydistinct, even if some mismatch (step) ofthe edges is present, thus demonstratinga good bridgeability of the process. In themacrographs of Figs. 3 and 4, for bothjoint preparations, the HAZ is clearly ob-servable. The width is gradually variedalong the thickness in the range of 1–2mm, being, in general, larger at the sideexposed to the laser beam, correspond-ing to a wider melted zone.
Microanalysis by EDS
The profiles of Ni concentration alongthe centerline of the welded sections,which characterize the powder penetra-tion, are reported in Fig. 5 for sampleswelded with both joint preparations.These diagrams report the best fit of theexperimental results and show a Ni pen-etration along all the weld thickness witha content that, in general, is decreasingalong the thickness, starting from thelaser beam side. Moreover, the Ni con-centration increases when the root open-ing or the bevel angle values increase.
318-s | NOVEMBER 2000
RE
SE
AR
CH
/DE
VE
LO
PM
EN
T/R
ES
EA
RC
H/D
EV
EL
OP
ME
NT
/RE
SE
AR
CH
/DE
VE
LO
PM
EN
T/R
ES
EA
RC
H/D
EV
EL
OP
ME
NT
Fig. 1 — A — Sketch of the LBW process; B — nozzle section.
Fig. 2 — Plate preparation of butt-joint welds: A — square edges; B— beveled edges.
root opening
For samples welded with joint prepa-ration A, the Ni content is in the range5–3% in weight (0–0.2 mm root opening)and in the range 14–6% in weight (0.4mm root opening), with higher values atthe outer surface (laser beam side).
The most satisfactory results, in termsof powder penetration, are obtained withjoint preparation B. In these cases, the Nicontent is substantially higher and betterdistributed along the weld thickness.With a bevel angle of 3 deg, the Ni con-tent profile decreases along the weldthickness from about 22% at the outersurface (laser beam side) to about 10% inweight on the opposite side. With higherbevel angles — 6 and 9 deg — the Nicontent is in the range from about 28% atthe outer surface (laser beam side) to20% in weight on the opposite side.
Optical / SEM Observations
A sample with beveled preparation (α = 3 deg) was taken as representative ofthe metallographic constituents ob-tained. The metallographic appearanceof the WZ at low magnification is shownby the optical micrograph in Fig. 6; de-tails of the solidification structures are
WELDING RESEARCH SUPPLEMENT | 319-s
RE
SE
AR
CH
/DE
VE
LO
PM
EN
T/R
ES
EA
RC
H/D
EV
EL
OP
ME
NT
/RE
SE
AR
CH
/DE
VE
LO
PM
EN
T/R
ES
EA
RC
H/D
EV
EL
OP
ME
NT
Table 1 — Base Material Composition
C Si Mn P S Cr Ni Ti V Nb Al
L24N 0.13 — 1.05 0.009 0.002 0.030 0.030 — — — 0.041
Fig. 3 — Macrograph of welds with square edge preparation: A — root opening = 0; B — root opening = 0.2 mm; C — root opening = 0.4 mm.
A B C
Table 2 — Welding Parameters
Joint Bevel Root opening Power rate Welding speedpreparation angle α (mm) (g/min) (mm/min)
A) square edges 0 20 1000A) square edges 0.2 20 1000A) square edges 0.4 35 1000B) beveled edges 3 deg 40 1000B) beveled edges 6 deg 60 700B) beveled edges 9 deg 60 500
Table 3 — Average Values of the WZ Area
Joint Bevel angle α Root opening WZ areapreparation (mm) (mm2)
A) square edges 0 21.2A) square edges 0.2 21.9A) square edges 0.4 19.4B) beveled edges 3 deg 21.4B) beveled edges 6 deg 28.8B) beveled edges 9 deg 43.4
320-s | NOVEMBER 2000
RE
SE
AR
CH
/DE
VE
LO
PM
EN
T/R
ES
EA
RC
H/D
EV
EL
OP
ME
NT
/RE
SE
AR
CH
/DE
VE
LO
PM
EN
T/R
ES
EA
RC
H/D
EV
EL
OP
ME
NT
Fig. 4 — Macrograph of welds with bevelededges preparation: A — α = 3 deg; B — α= 6 deg; C — α = 9 deg;
A
B
C
Fig. 5 — Profiles of Ni concentration along the centerline of the welded sections: A — square edgepreparation; B — beveled edge preparation.
Fig. 6 — Optical micrograph at low magnification of the WZ for a sample with beveled edgepreparation and α = 3 deg.
given in the SEM micrographs of Fig. 7.The dendritic solidification structures ofthe matrix grain are surrounded by an in-tergranular constituent, whose naturewas not identified.
Near the outer surface on the laserbeam side (Fig. 7A), the matrix is sup-posed to be austenitic, according to thehigh Ni content, which gradually de-creases from a value of around 22%, nearthe outer surface on the laser beam side,to a value of about 10%, on the oppositeside — Fig. 5.
Local microanalysis measurementson the intergranular constituent gave a Nicontent comparable to the one present inthe matrix. On this basis, the interpreta-tion of this constituent as the FeNi3 inter-metallic compound has to be excluded.Moreover, in this zone (Fig. 7B) some mi-cropores of dark appearance androunded elongated shape, having about1 mm size, are present. These microporeshave a morphology similar to that ob-served in materials produced by powdermetallurgy (Ref. 6).
At the opposite side (Fig. 7C), the ma-trix is characterized by an acicular struc-ture, suggesting the possible presence ofmartensite, according to the microhard-ness measurements. Here again, the in-tergranular constituent surrounding thedendritic matrix grains and some micro-pores are visible.
The same acicular structure can beobserved in the samples welded with thesquare edge preparation (root opening= 0.4 mm) in the region with the highestvalues of microhardness — Fig. 8.
Vickers Microhardness Test
For samples welded with joint prepara-tion A, the Vickers microhardness values,measured along the centerline of thewelded section, are given in Fig. 9. Thehardness fluctuations can be related to so-
lidification structureinhomogeneities, asresulting from thedistinct banding re-vealed by themacrographic in-spection — Fig 3.The higher hardnessvalues (above 450HV) recorded forsamples weldedwith 0.4 mm rootopenings confirm,on the basis of theNi content mea-surements and SEMobservations, thepresence of marten-sitic structures.
For joint prepa-ration B with thelower bevel angle(α = 3 deg), themeasured valueswere in the range200–480 HV —Fig. 10A. The rela-tively high micro-hardness values(average value 350HV) in the zonedistant 1–3 mmfrom the outer sur-face, laser beamside, are reason-able when takinginto account thehardening effectsof the high Ni con-tent — Fig. 5. Thepeak value of 480HV suggests thepresence ofmartensitic struc-tures, as supportedalso by the acicular morphology ob-served in the zone far from the laser beam
side — Fig. 7C.With the other values of bevel angle
WELDING RESEARCH SUPPLEMENT | 321-s
RE
SE
AR
CH
/DE
VE
LO
PM
EN
T/R
ES
EA
RC
H/D
EV
EL
OP
ME
NT
/RE
SE
AR
CH
/DE
VE
LO
PM
EN
T/R
ES
EA
RC
H/D
EV
EL
OP
ME
NT
Fig. 7 — SEM images of the WZ for a sample with beveled edge prepa-ration (α = 3 deg): near the outer surface at the laser beam side (a, b); atthe opposite side (c).
Fig. 8 — SEM images of the WZ for a sample with square edge prepara-tion and root opening = 0.4 mm.
(α = 6 and 9 deg), the hardness profilesalong the centerline of the welded sec-tions are characterized by values around200 HV and by the absence of relevantfluctuations — Fig. 10. Considering thehigh Ni content (a range of 20–28% inweight) recorded by microanalysis mea-surements, these low hardness values canbe ascribed to austenitic structures, whichare present along all the weld thickness.
322-s | NOVEMBER 2000
RE
SE
AR
CH
/DE
VE
LO
PM
EN
T/R
ES
EA
RC
H/D
EV
EL
OP
ME
NT
/RE
SE
AR
CH
/DE
VE
LO
PM
EN
T/R
ES
EA
RC
H/D
EV
EL
OP
ME
NT
Table 4 — Expected and Experimental Average Ni Content
Joint Bevel Root opening Expected Ni Average Experimentalpreparation angle α (mm) content (wt-%) Ni content (wt-%)
A) square edges 0 11 4A) square edges 0.2 11 4A) square edges 0.4 22 10B) beveled edges 3 deg 23 17B) beveled edges 6 deg 36 25B) beveled edges 9 deg 34 24
Fig. 9 — Vickers microhardness profiles along the centerline of thewelded sections for samples with square edges preparation: A — rootopening = 0; B — root opening = 0.2 mm; C — root opening = 0.4 mm.
Fig. 10 — Vickers microhardness profiles along the centerline of thewelded sections for samples with beveled edges preparation: A — α =3 deg; B — α = 6 deg; C — α = 9 deg.
Discussion
In order to estimate which is the bestgeometry to convey the powder flow intothe melted pool, the expected Ni con-tent, calculated on the basis of the weld-ing parameters and of the measured WZarea, was compared with the Ni contentalong the welded sections experimen-tally obtained by microanalysis.
As a result of microanalysis measure-ments, the contribution to the joint for-mation is given by the base metal, withdensity δFe=7.87 10-3 g/mm3 and concen-tration CFe (weight fraction), and by the Nipowder, with density δNi=8.9 10-3 g/mm3
and concentration CNi (weight fraction).So, for a welded volume having as basethe WZ area AS (mm2) and a height of 1mm, considering the powder rate p(g/min) and the welding speed vS
(mm/min), the expected Ni content isgiven by the following expression:
The average value of the Ni content ex-perimental profile along the welded section, obtained by microanalysis measurements, was considered repre-sentative of the Ni content along thewelded sections.
In Table 4, expected and experimen-tal average Ni content are compared forboth joint preparations. The experimen-tal average values are, in any case, lowerthan the expected ones, due to probablepowder loss. From data reported in Table4, the ratio of the experimental averageNi content to the expected Ni content re-sults are about 0.4 for joint preparationA and about 0.7 for joint preparation B.So, the best results appear to be thoseobtained in the samples welded with abeveled joint, in which the experimentalaverage Ni content shows the best ap-proximation to the expected one.
With the beveled joint preparation, the
powder conveyed by the gas flow is “fun-neled” and forced to melt into the spacebetween the edges. On the contrary,when welding using square edges withlittle or no root opening, the possibility ofthe powder penetrating between theedges is drastically reduced and part ofthe filler powder is blown away with theprocess gas.
Generally, samples butt-joint weldedwith square edges gave rise to a low con-tribution of the powder filler metal to theformation of the WZ, revealed by thelean content of nickel and, occasionally,with presence of cracks observed onmacrographs and micrographs.
In some welded sections of joints withbeveled edge preparation, the presenceof rounded and elongated micropores(representative size of 1 mm) were ob-served. These micropores are similar tothose often found in materials producedby powder metallurgy.
Conclusions
The experimental work contributed tothe setup of the laser beam weldingprocess with a single pass by utilizationof filler metal powder. The obtainabledistribution of the filler metal and its alloyelements along the thickness of butt-jointwelds was considered for two differentpreparations — square and bevelededges.
• The powder penetration along theweld thickness decreased starting fromthe outer surface at the laser beam side.
• For square edge preparation, the Nicontent increased at the increasing of theroot opening between plate edges, but itis always very low in respect to the ex-pected value.
• The highest contribution of the fillerpowder to the formation of the weld zonewas obtained with the beveled jointpreparation, reaching about 70% of theNi expected content.
• In some welded sections, particu-
larly those for joints with beveled edgepreparation, the presence of roundedand elongated micropores, similar tothose often found in materials producedby powder metallurgy, were observed.
Acknowledgments
This work was supported by theHuman Capital and Mobility Program ofthe European Community within theframework of the project European LaserCenter at the Fraunhofer Institut fürLasertechnik in Aachen, Germany. Theauthor wishes to thank Dipl. Ing. W.Frohn of the Laser Welding Departmentof ILT-Aachen for his extensive support insetting up welding procedures and tech-nical assistance. Grateful thanks are dueto the management of ILT.
References
1. Campbell, C., and Johnson, W. C. 1967.Granular metal filler metals for arc welding.Welding Journal 46(3): 200–208.
2. Troyer, W., and Mikurak, J. 1974. Highdeposition submerged arc welding with ironpowder joint fill. Welding Journal 53(8):494–300.
3. Rodgers, J., and Lochhead, J. C. 1987.Submerged arc welding with metal powderadditions, productivity and properties. Weld-ing Journal 66(10): 21–27.
4. Bailey, N. 1991. Submerged arc weld-ing ferritic steels with alloyed metal powder.Welding Journal 70(8): 187–194.
5. Sili, A. 1998. Laser beam welding ofHSLA steels with utilization of powder fillermetal. CEC Project — Final report, User Se-lection Panel Meeting. Fraunhofer Institut fürLasertechnik—ILT, Aachen, Germany.
6. Metallography and Microstructures.1985. Metals Handbook, Ninth Ed., Vol. 9.ASM, Materials Park, Ohio, p. 504–530.
CNi p NiCNi FeCFe vs As= + × ×[( ) ] ( )δ δ 1
WELDING RESEARCH SUPPLEMENT | 323-s
RE
SE
AR
CH
/DE
VE
LO
PM
EN
T/R
ES
EA
RC
H/D
EV
EL
OP
ME
NT
/RE
SE
AR
CH
/DE
VE
LO
PM
EN
T/R
ES
EA
RC
H/D
EV
EL
OP
ME
NT
