Investigation of laser-arc hybrid welding utilizing CMT, effects upon oxygen and nitrogen contents as well as the weld stability due to oxygen contents of the shielding gas and gap oxides
Jan Frostevarg, Pan Qinglong, Masami Mizutani, Yousuke Kawahito, Seiji Katayama
Abstract In this study, the effects upon weld arc stability and contents of oxygen and nitrogen inside the weld is investigated for laser-arc hybrid welding (LAHW) using the CMT arc mode. CMT is a controlled short arc technique that is here expected to be less sensitive to oxygen levels in the shielding gas. Having oxides or nitrogen inside the weld region is also known to alter the fluid mechanics such as surface tension, but also a cause for porosity. The gap preparation prior to welding is typically carried out by thermal cutting, water jet cutting or grinding. These edges are then placed to form the joint to be welded. In the case of thermal cutting, oxides are usually formed on the cut edge, adding oxygen into the melt pool when welding is conducted. The effects are studied by macroscopy, HSI and chemical analysis. The arc stability and oxygen and nitrogen levels in the welds are more affected by the oxides remaining from the laser cut edges than the oxygen in the shielding gas when the cutting oxides are removed.
Keywords Laser hybrid, CMT, Oxides, Laser cut, Stability
1 Introduction
Laser-arc hybrid welding (LAHW), combines a focused high power laser beam with an electric arc within the same processing zone [1-4], illustrated in Fig. 1. Compared to autonomous laser welding, LAHW provides filler wire to fill gaps and, aided by the arc, shape the weld surface. Gas Metal Arc (GMA) welding can be carried out by many different techniques. The most common arc modes supplied by all manufacturers are the Standard and the Pulsed mode. In LAHW the most commonly used arc mode is the GMA Pulsed arc mode, where one drop per arc pulse is transferred towards the melt pool in a semi-controlled flight [5,6].
Recently, another, even more controllable Pulsed short arc mode technique was developed, the Cold Metal Transfer (CMT) [7]. It utilizes controlled wire feeding and surface tension drop transfer. The wire is pulled back and forth instead of using a constant wire feed. If the reduced wire feed rate is deemed acceptable, the technique offers smooth drop transfer and less electrical power needed by to melt the wire. For LAHW, the CMT arc mode has also been compared with the Pulsed and Standard arc mode in recent studies [8-13], looking at weld stability and metallurgy.
J. Frostevarg
Luleå University of Technology, Dept. TVM, SE-971 87 Luleå, Sweden
e-mail: jan.frostevarg@ltu.se
Osaka University, Joining and Welding Research Institute
11-1 Mihogaoka, Ibaraki, Osaka 567-0047, Japan e-mail: qlp@jwri.osaka-u.ac.jp
mizutani@jwri.osaka-u.ac.jp kawahito@jwri.osaka-u.ac.jp katayama@jwri.osaka-u.ac.jp
Fig. 1 Sketch of the laser-arc hybrid welding setup, including geometrical parameters
The quality and strength of welds is significantly determined by surface geometry [14,15], that results from the complex fluid flow mechanics caused by the electric arc and the laser keyhole [16]. The process parameter combination needs to be adapted depending on chosen arc mode and work piece conditions [17,18]. The welding process may otherwise become unstable, resulting in an uneven surface geometry or inner imperfections [9,19- 21]. High Speed Imaging (HSI) enables study of drop transfer and keyhole conditions [9,17,22,23]. The presence of a gap influences the metal flow and the arc [17].
Apart from having an impact on surface geometry [1], unexpectedly increased levels of oxygen and nitrogen contents within the weld is unfavourable. The increased amounts degrade the weld metal mechanical properties, especially concerning low temperature toughness. However, sometimes in certain weld metals the presence of oxides shows improved tendency to form acicular ferrite, improving the mechanical properties. It is also well known that absorption of nitrogen into the weld pool during welding causes porosity as the weld melt solidifies as the dissolved nitrogen is released as gas again. Due to this, it has been said that for general GMA welding, the involvement of air surrounding the arc into the melt pool should be prohibited and welding in strong wind should be prohibited. Different edge preparations have effect upon LAHW, and laser cutting seems to be preferred over welding with milled edges. Laser cutting results in oxides and striations on the cut surface. These oxides are believed to be harmful for weld performance due to pre-mentioned reasons. The oxide layers on the surface formed by thermal cutting processes should be removed or at least controlled regarding evaluation of residual oxygen content. Furthermore, it has been reported that oxygen contents greatly affects the weld bead formation [24,25]. In the world of laser and laser arc hybrid welding, the lowest possible levels of oxygen and nitrogen are required for achieving high quality weld caps with good morphology and mechanical properties.
It is expected that the advantages of CMT should not only reduce the heat input into the work piece, but also offers a more stable deposition of weld wire that should not be very sensitive to the effects of the amount of CO2
in the shielding gas. The involvement of oxygen accompanied by decomposition of the shielding gas is usually necessary in other GMA welding processes. In this paper, the CMT-technique is used with LAHW in 10 mm
thick steel sheets concentrating on arc stability and oxygen and nitrogen contents of the final welds when welding with and without oxygen in the shielding gas and the thermally laser cut edges.
2 Methodology
2.1 Welding equipment
An illustration of the LAHW setup is shown Fig. 1, with the geometrical setup parameters listed in Table 1.
The laser used was a 15 kW Yb:fibre laser (manufacturer IPG Laser GmbH, type YLR-15000 (fibre core diameter; 200 µm, beam parameter product; 10.3 mm·mrad, wavelength; 1070 nm). The laser was operated in the continuous wave (cw) mode, focused at the surface by 300 mm focal length optics to a spot size of 400 µm diameter (Rayleigh length ±4 mm). To prevent back reflections damaging the optical fiber a slight tilting of the laser was applied. The GMA torch was applied in a tilted leading position. The GMA welding equipment was a Fronius MAG power source TPS4000 VMT Remote. The wire feeder is a combination of a continuous feeding unit VR7000 with a Robacta Drive unit (from Fronius) that carries out the back and forth motion of the wire tip which enables the CMT-process used in the experiments. The welds were carried out using an articulated robot (Motoman).
Table 1 Constant geometrical weld setup parameters, as presented in Fig. 1 Parameter (unit) 𝑠 (mm) 𝑑𝐿𝐴 (mm) 𝛼𝐿 (°) 𝛼𝑇 (°) 𝑧0 (mm) f (mm)
Value 18 2.5 7 -30 -8 300
2.2 Experimental procedure
Most of the parameters could not be freely chosen as they were preset by the system at different wire feed rates for a chosen synergy curve. From these presets, some adjustments are allowed. The filler wire used was Lincoln SupraMIG Ultra (AWS A5.18), a steel-based wire with a diameter of ∅=1.2 mm. The plates welded were 10 mm thick steel Domex 420 MC (S420 MC), laser cut into pieces 50 mm wide and 200 mm long, clamped tight in a butt joint configuration with I-gap. The mill scale on the surface was sandblasted off prior to cutting, providing improved wetting and the avoidance of oxide inclusions and lack of fusion at the weld surface [1]. The material composition of the wire and the steel plates are shown in Table 2.
The fixed weld parameters used in the experiments are seen in Table 3 and the variable weld parameters are seen in Table 4. The applied shielding gas was Argon mixed with a corresponding specified amount of CO2. The laser power and welding speeds where also varied to compare differences of the weld shape. In the performed experiments as specified, the oxides remaining on the cut surface was occasionally brushed off with a steel brush, without damaging the structure of the surface made by the laser cutting in order to investigate the influence of the oxides instead of surface structure.
Table 2 Material composition (in wt%) of the steel plates and filler wire
Name C Si Mn P S Al Nb V Ti Si Fe*
S420 MC 0.10 0.03 1.50 0.025 0.01 0.015 0.09 0.20 0.15 98.01
AWS A5.18 0.08 1.70 0.85 97.37
* Displayed values are maximum, except for Fe, which is minimum.
Table 3 Fixed (preset) weld parameter settings Parameter
(unit)
Welding speed 𝑣 (m/min)
Wire feed rate 𝑣𝑤 (m/min)
Shield gas flow
𝐹 (L/min) Arc voltage
𝑈 (V) Arc current
𝐼 (A) Arc power 𝑃𝐴 (kW)
Value 1.5 3 20 13.3 115 1.53
Table 4 Variable weld parameter settings Sample Protection gas
CO2 content (%) Welding speed
v (m/min) Average laser power
𝑃 (kW) Joint cutting oxides brushed off
1 18 1.5 11 No
2 0 1.5 11 No
3 18 1.5 11 Yes
4 0 1.5 11 Yes
5 18 1.5 10 Yes
6 0 1.5 10 Yes
7 8 0
0 1.8
2.0 11
12 Yes
Yes 2.3 Analysis
The resulting welds were analyzed by different methods of the samples, in addition to ocular surface observation and macrographs of the cross sections of the welded samples.
2.3.1 High speed imaging
To evaluate the weld experiments, the top surfaces where scanned prior to and after welding. High Speed Imaging (HSI) [1,3,4,9,23] at 5000 frames/s was also used during the welding process for each experiment to observe the melt flow and GMA arc behaviour. Figure 2 shows a typical HSI frame with indicated regions of interest, using the GMA in Pulsed mode with globular drop transfer.
Figure 2. a) High Speed Image (HSI) of laser arc hybrid weld pool (top view, 30° inclined) 2.3.2 Experimental procedure of oxygen and nitrogen analysis
The surface of welded joint was firstly removed by milling machine. Then samples were cut from the central part of the weld zone with an electric discharge machining. The dimension was 10×1×1 mm. Three samples were prepared at each welding condition for the oxygen and nitrogen analysis. Before the measurement, all faces of the sample were grinded using coated abrasives to remove the oxide film layer and the sample was kept in the absolute ethanol to prevent the oxidization. Then the precise measurement with an oxygen and nitrogen analyzer, Horiba EMGA-520 was carried out.
3 Results and discussion
3.1 High speed imaging
In Fig. 3, short image sequences of the first four welds are shown, for each specific case (corresponding to the sample numbers shown in Table 4) concerning influence of oxides. Case 1 and 2 have the oxides form cutting remaining, while Case 1 has oxygen (CO2) in the shielding gas and Case 2 does not. Case 3 and 4 have the oxides brushed off with intact structure remaining from laser cutting. Case 3 has oxygen (CO2) in the shielding gas while Case 4 doesn’t. Note that for the CMT process, the arc length varies during the weld, and the frequency is coupled to the retraction and extrusion of the wire so the pulse period length therefore varies.
Arc length varies, therefore also frequency (also varies during weld)
Figure 3. HSI for the selected samples 1-4 from Table 4
Sample 1: Figure 3a-i shows a stable arc and a gouge forming in front of the keyhole. The arc is later turned off and the wire is pushed into the melt where the formed droplet is transferred to the melt pool by surface tensional forces, Fig. 3a-ii. In Fig. 3a-iii wire is already pulled back and the arc is again ignited and shows the same stable behaviour as the first pulse. The melt flow the keyhole is calm and the solidification edges have a proper v-shaped form.
Sample 2: In Figs. 3b-i to iii, there is no oxygen in the shielding gas so no oxides form on the surface to lower electrical resistance, so the arc gets wider and fluctuates over the surface but is kept in the vicinity of the gap. The gouge in front of the keyhole is also much smaller, barely formed at all due to the defocused arc and the droplet is placed right in front of the keyhole. The melt behind the keyhole is similar as for sample 1.
Sample 3: In Figs. 3c-i to iii the arc fluctuates, but less than as for sample 2. The oxides forming on the surface helps to focus the arc so a gouge forms, but not as deep as for Sample 1. The melt width behind the keyhole is larger than when the oxides remain. This is probably due to the wider arc preheats the material and the keyhole confers the extra required heat by conduction in the material. The melt also flows downward behind the keyhole, more so than for the previous cases. This melt flow behaviour causes roothang on the weld root and a cavity on the weld cap. The increased with and decreased viscosity due to oxides in melt might be the cause for the change of inner melt flow direction (e.g. Marangoni effects changing the melt-flow direction from outward to inward).
Sample 4: For Figs. 3d-i to iii, the arc is wide and shows high fluctuation. Since there is no gouge formed from the arc, the immediate surface is instead melted in a large area beneath it. The bad surface melting also causes the solid wire to hit the solid surface, temporarily bending the wire while leaving the droplet before it is pulled back up. The melt behaviour behind the keyhole is the same as for Sample 3, except that large dropout is formed instead of roothang.
3.2 Macroscopy
For the eight welds, the resulting weld cap and root appearances are shown in Fig. 4a and one cross section from each are shown in Fig. 4b. All the welds show some degree of roothang/dropout at the weld root. The weld cap appearance is best when the oxides from laser cutting remains (Samples 1-2), ie. no underfill, near net-shape.
For Sample 4 when no oxygen is inside the shielding atmosphere or the gap, the dropouts are larger. Sample 4 and 5 show the largest dropout, where Sample 5 seems to penetrate after half the weld, supposedly barely if results from [26] are taken into account. While not penetrating fully, the weld cap has the best appearance. When speed and power is increased, the droplets on the root are more elongated. It seems that the presence of CO2 does not counteract the dropout as much as the remaining oxides in the gap do. In Fig. 4b it is clear that the remaining oxides in the gap help to form less wide weld caps and root exits. It is unclear how the inner melt is affected by these oxides, but having a wider root exit seems to provoke dropout.
a) b)
Figure 4. a) top and root surface appearances and b) cross section appearances for the eight welds
3.3 Oxygen and nitrogen analysis
The results of the nitrogen and oxygen content measurements are shown in Fig. 5 for all samples, including the base material (BM). For all welds, the levels of nitrogen and oxygen are higher than for the BM, even though the oxides are brushed off. Since there is supposedly no nitrogen in the shielding gas and there is no difference between remaining or removed oxides from laser cutting, the added nitrogen has to be from the surrounding air.
There are much higher values of oxygen in the welds where the cutting oxides remained in the gap prior to welding. The variation of content in the brushed samples could be imperfect removal of cutting oxides. When only looking at the weld material, the increased values of Sample 1 and 2 are expected to yield lower mechanical properties than the other samples.
Figure 5. Measured oxygen and nitrogen content of the eight samples and the base material (BM)
4 Conclusions
The morphological effects observed when having the oxides form laser cutting remaining or removed (cutting structure intact) prior to welding for LAHW using CMT are as follows:
(i) Having the oxides from laser cutting seems to counteract dropout when laser welding (ii) The cutting oxides also help to stabilize the arc towards the gap
(iii) Removing the cutting oxides results in wider weld cap and weld root
(iv) Having the cutting oxides remaining results in much higher oxygen contents inside the final weld
Concerning the shielding gas influences;
(v) Having CO2 in the shielding gas helps to stabilize the arc, but the difference is not as big as for other commonly used arc modes
(vi) Having CO2 present in the shielding gas or not does not seem to have a major impact on the final weld result on neither morphology or oxygen or nitrogen levels inside the weld
(vii) The surrounding air seems to feed nitrogen into the melt pool, even though the process is shielded
Future studies of this topic includes achieving a stable LAHW process insusceptible to oxygen and nitrogen contents. This may include proper configuration of torch and laser to prevent involvement of surrounding air, as well as having a second back shielding gas nozzle.
References
[1] J. Karlsson, P. Norman, A.F.H. Kaplan, P. Rubin, J. Lamas, A. Yañez, Observation of the mechanisms causing two kinds of undercut during laser hybrid arc welding, Applied Surface Science. 257 (2011) pp. 7501-6.
[2] H. Engström, K. Nilsson, J. Flinkfeldt, A. Skirfors, B. Gustavsson, Laser Hybrid Welding of High Strength Steels, 20th ICALEO Conference, Jacksonville, FL, USA (2001).
[3] J. Powell, J. Lamas, J. Karlsson, P. Norman, A. Kaplan F. H., The sensitivity of hybrid laser welding to variations in workpiece position, LIM Conference. Munich, Germany (2011).
[4] A. Kaplan F. H., J. Lamas, J. Karlsson, P. Norman, A. Yañez, Scanner analysis of the topology of laser hybrid welds depending on the joint edge tolerances, 12th NOLAMP Conference, Trondheim, Norway (2011).
[5] A. Scotti, V. Ponomarev, W. Lucas, A scientific application oriented classification for metal transfer modes in GMA welding, Journal of Materials Processing Technology. 212 (2012) pp. 1406-13.
[6] G. Campana, A. Fortunato, A. Ascari, G. Tani, L. Tomesani, The influence of arc transfer mode in hybrid laser-mig welding, Journal of Materials Processing Technology. 191 (2007) pp. 111-3.
[7] C.G. Pickin, K. Young, Evaluation of cold metal transfer (CMT) process for welding aluminium alloy, Science and Technology of Welding and Joining. 11 (2006) pp. 583-5.
[8] J. Karlsson, A.F.H. Kaplan, Comprehensive monitoring and control of laser hybrid arc welding in industrial production, 31st ICALEO Conference. Anaheim, California, USA (2012).
[9] J. Frostevarg, A.F. Kaplan, J. Lamas, Comparison of CMT with other arc modes for laser-arc hybrid welding of steel, Welding in the World. 58 (2014) pp. 649-60.
[10] J. Lamas, J. Frostevarg, A.F. Kaplan, Gap bridging for two modes of laser arc hybrid welding, Journal of Materials Processing Technology. (2015).
[11] M. Song, J. Kim, G. Myung, J. Oh, A Study on Laser-Arc Hybrid Weldability of Aluminium Alloy with Arc Welding Mode, Science of Advanced Materials. 7 (2015) pp. 91-6.
[12] G. Li, C. Zhang, M. Gao, X. Zeng, Role of arc mode in laser-metal active gas arc hybrid welding of mild steel, Materials & Design. 61 (2014) pp. 239-50.
[13] C. Zhang, M. Gao, X. Zeng, A comprehensive study of laser, CMT, laser-pulse MIG hybrid and laser-CMT hybrid welded aluminium alloy, 31st ICALEO conference. Aneheim, California, USA (2012).
[14] M.M. Alam, Z. Barsoum, P. Jonsén, A.F.H. Kaplan, H.Å Häggblad, The influence of surface geometry and topography on the fatigue cracking behaviour of laser hybrid welded eccentric fillet joints, Applied Surface Science. 256 (2010) pp. 1936-45.
[15] B. Ribic, T.A. Palmer, T. DebRoy, Problems and issues in laser-arc hybrid welding, International Materials Reviews. 54 (2009) pp. 223-44.
[16] J.H. Cho, S.J. Na, Three-dimensional analysis of molten pool in GMA-Laser hybrid welding, Welding Journal. 88 (2009) pp. 35-43.
[17] Y.-. Kim, N. Alam, H.-. Bang, H.-. Bang, Observation of hybrid (cw Nd:YAG laser + MIG) welding phenomenon in AA 5083 butt joints with different gap condition, Science and Technology of Welding &
Joining. 11 (2006) pp. 295-307.
[18] P.F. Mendez, T.W. Eagar, Penetration and defect formation in high current arc welding, Welding Journal.
82 (2003) pp. 296-306.
[19] M. Moradi, M. Ghoreishi, J. Frostevarg, A.F.H. Kaplan, An investigation on stability of laser hybrid arc welding, Optics and Lasers in Engineering. 51 (2013) pp. 481-7.
[20] J. Lamas, J. Karlsson, J. Powell, A. Kaplan F. H., A. Yañez, The effect of fit-up geometry on melt flow and weld quality in laser hybrid welding, Journal of Laser Applications. 25 (2013) pp. 7.
[21] J. Frostevarg, A.F. Kaplan, Undercut suppression in laser-arc hybrid welding by melt pool tailoring, Journal of laser applications. 26 (2014) pp. 031501.
[22] A. Fellman, A. Salminen, Observation of the phenomena of CO2 laser-MAG hybrid welding, ICALEO.
LIA, (2006).
[23] I. Eriksson, P. Gren, J. Powell, A.F.H. Kaplan, New high-speed photography technique for observation of fluid flow in laser welding, Optical Engineering. 49 (2010).
[24] K. Ono, K. Adachi, I. Miyamoto, T. Inoue, Influence of oxide film on weld characteristics of mild steel in CO2 laser welding, Journal of laser applications. 14 (2002) pp. 73-7.
[25] T. Onishi, M. Mizutani, Y. Kawahito, S. Katayama, High-Power Fiber Laser Butt Welding of Thick High- Strength Steel Plate Using Sensing System with Hot Wire, ICALEO 2009. Orlando, FL, U.S.A. (2009).
[26] V. Romminger, P. Haug, N. Speker, M. Holzer, High‐power Full Penetration Welding Behavior, Laser Technik Journal. 10 (2013) pp. 36-40.
