Melting Characteristics of C-Type Filler Metal in GTAW
Melting Characteristics of C-Type Filler Metal in GTAW
Melting Characteristics of C-Type Filler Metal in GTAW
RESEARCH
Abstract Keywords
A C-type filler metal was developed to overcome ■ Gas Tungsten Arc Welding (GTAW)
the low deposition rate (DR) of gas tungsten arc ■ Filler Metal
welding (GTAW). The present study investigated ■ C-Type Filler Metal
the maximum DR for a novel C-type filler metal ■ Welding Wire
and compared it to conventional circular welding ■ Deposition Rate (DR)
wires during GTAW using an Alloy 625 filler metal. ■ Fusion Depth/Penetration Depth
For comparison with conventional circular welding
■ Metal Transfer
wires, a ø1.2-mm (0.047-in.) welding wire, which
■ Continuous Bridging Transfer
■ Welding Current
is most widely used in practice, and a ø2.4-mm ■ Filler Metal Current
(0.094-in.) welding wire, which has almost the same
sectional area as the novel C-type filler metal, were
used in this research. An industrial robot was utilized Introduction
to produce bead-on-plate welds in the flat position.
Excellent welding quality can be obtained via automatic gas
The results revealed that at the same 200-A welding
tungsten arc welding (GTAW) with a filler metal if continuous
current, the DR of the C-type filler metal was higher bridging transfer is implemented. However, the productivity
than the conventional circular welding wires by 1.17 of automatic GTAW with a conventional circular welding wire
to 1.4 times according to the sectional area of the is low. Therefore, many technologies have been developed to
circular welding wires. At a high welding current of increase the deposition rate (DR) of GTAW (Ref. 1). A higher
500 A, it was impossible for the ø1.2-mm welding DR requires a higher welding current and speed. However,
wire to deposit quality welds, and the acceptable humping beads may appear in the high-current region of
range of the DR for the ø2.4-mm welding wire was GTAW (Ref. 2).
narrow (i.e., 7–8 kg/h [15.4–17.6 lb/h]). However, Therefore, a C-type filler metal was developed as the filler
the acceptable range of the DR for the C-type filler metal of GTAW to increase the DR in the high-current region
metal was as broad as 5.04–12.1 kg/h (11.1–26.6 (Ref. 3). The wide and unique shape of the C-type filler metal
lb/h) under the high welding current of 500 A. could prevent high arc pressure on the weld pool when used in
high-current GTAW. Additionally, a high DR could be obtained
The maximum DR of the C-type filler metal was
even though a cold filler metal was used (Refs. 4–6). Hence,
1.51 times that of the ø2.4-mm welding wire. The
GTAW with a C-type filler metal is more effective for weld
mechanism of obtaining a high DR using the C-type overlay and wire arc additive manufacturing (WAAM) (Refs. 7,
filler metal was analyzed from the viewpoint of the 8). Obtaining a small and stable fusion depth is important in
continuous bridging transfer at the melting edge of the overlay and WAAM. For the purpose of this paper, fusion
the C-type filler metal. The ability of the C-type filler depth shall be referred to as penetration depth. This penetra-
metal to achieve high DRs at high-current regions tion depth may be affected by metal transfer from the melting
was superior to the conventional ø1.2- and ø2.4-mm edge of the filler metal to the weld pool (Ref. 9). However,
welding wires. before this research, a direct comparison between C-type
filler metals and circular welding wires in terms of metal
transfer and penetration depth had not been performed.
https://doi.org/10.29391/2023.102.016
SEPTEMBER 2023 | 201-s
Fig. 1 — Cross-sectional area of a newly developed
C-type filler metal. Fig. 2 — Definition and measurement of the bead
geometry. Note: penetration depth = P, bead width =
W, and reinforcement height = H.
A B
Fig. 3 — Schematic of GTAW with a C-type filler metal: A — Front view; B — side view.
In general, arc radiation may partially melt the filler metal conical angle was 40 deg and the truncated tip size was 1.5
during GTAW, but the plasma stream of the arc can melt it mm (0.059 in.) for 500 A. A mixed gas of Ar + 7%H2 was used
primarily. We considered the plasma stream to obtain a high as the shielding gas at a flow rate of 20 L/min (5 gal/min).
DR of the C-type filler metal (Refs. 6, 7, 10). The stand-off distance between the electrode and the base
This study aimed to establish a model for metal transfer and metal was 5 mm, which was kept before welding. Before arc
penetration depth using GTAW at a current of 200 and 500 starting, the filler metal height between the filler tip bottom
A with C-type filler metals and conventional welding wires. and the substrate surface was set to 0 mm, and the filler metal
In addition, the maximum DRs at each welding current were distance between the filler tip and the electrode center was
investigated for the C-type filler metals and circular welding set to 0 mm. The experiments were conducted using the
wires. The melting edge and bridge of the filler metal in the standard procedure described in ISO/TR 18491, Welding and
arc were observed using a welding camera. allied processes — Guidelines for measurement of welding
energies (Refs. 11–13). An industrial six-axis robot was used to
produce the beads. The waveforms of the welding arc voltage
Experimental Procedure were monitored using an arc monitoring system for GTAW.
The in-situ observations on the metal transfer behavior of
An ASTM A283 plate (t = 10 mm [0.39 in.]) was used as a the three welding wires/filler metals in various conditions
substrate for the bead-on plate welds using GTAW with Alloy were captured using a welding camera at a frame rate of 50
625 (ERNiCrMo-3) filler metal. Three types of filler metals/ fps. The welds were cross sectioned and polished for mac-
welding wires were used in the experiments. These were a rosection analysis. The penetration depth was measured
ø1.2-mm (0.047-in.) welding wire (usually used in automatic using software after etching the cross sections.
GTAW), a ø2.4-mm (0.094-in.) welding wire (the cross- Figure 2 shows the measured geometry of the welds: pen-
sectional area is the same as a C-type filler metal), and a etration depth, bead width, and reinforcement height. The
newly developed C-type filler metal of 5 mm2 (width: 5 mm deepest penetration depth of the bead was considered the
[0.196 in.]; thickness: 1 mm [0.039 in.]) — Fig. 1. penetration depth. The penetration depth was calculated by
A tungsten electrode (W-electrode) with a diameter of 4 taking the average of the three cross sections. Penetration
mm (0.157 in.) was used. The conical angle was 30 deg and depths greater than 0.3 mm (0.011 in.) and less than 3.0 mm
the truncated tip size was 0.5 mm (0.019 in.) for 200 A. The (0.118 in.) were considered. If the penetration depth is smaller
than 0.3 mm, welding defects such as insufficient penetra- metal feed speed (mm/s), U is the welding speed (mm/s), the
tion depth and lack of penetration depth may appear in any DR is measured in kg/h, ρ F is the density of the filler metal
weld. If the penetration depth is larger than 3.0 mm, dilution (g/mm3 [the density of Alloy 625 is 0.0084 g/mm3]), NHI
increases in overlay welding and worsens the wall resolution is the net heat input (J/mm), ηa is the arc efficiency, I is the
in WAAM. The penetration depth data of the welds, which welding current (A), E is the voltage (V), and NHIR is the net
included tunnel defects, undercuts, and humping beads, were heat input ratio (J/mm3).
excluded. The process setup with the C-type filler metal is
shown in Fig. 3.
The welding conditions used in the experiments are
Results and Discussion
listed in the Welding Conditions section of Tables 1 and 2.
The deposition area (DA) and DR were calculated using the Bead Appearance and Macrosection
cross-sectional area of the filler metal, filler metal feed speed, of the Welds
and welding speed, as given in Equations 1 and 2.
Figure 4A shows the typical bead appearance of the welds
made with a ø1.2-mm welding wire, ø2.4-mm welding wire,
DA = (AF × FFS) / U (1) and C-type filler metal at a welding current of 200 A. The bead
appearances were almost similar at a DR of 1 kg/h (2.2 lb/h)
for the two welding wires in the presence of uniform bead
ripples formed by intermittent bridging transfer. In the case
of the C-type filler metal, the bead surface was smooth, owing
DR = (A! × FFS × ρ! ) × 3600/1000 (2) to the continuous bridging transfer. However, when the DR
was increased to 2.5 kg/h (5.5 lb/h), the ø1.2-mm welding
wire showed signs of slight sticking, whereas the ø2.4-mm
welding wire and C-type filler metal showed sound beads.
Further increasing the DR resulted in more sticking at the DR
NHI = η! × (I × E) / U (3)
of 2.7 kg/h (5.9 lb/h) for the ø1.2-mm welding wire. On the
contrary, for the ø2.4-mm welding wire, sticking appeared
at the DR of 3.5 kg/h (7.7 lb/h).
In this study, only sound beads without sticking were con-
NHIR = NHI/DA = η! × (I × E)/(A" × FFS) (4) sidered to determine the maximum DR. Sticking and humping
beads were considered defects, as seen in Tables 1 and 2.
DuPont and Marder (Ref. 14) investigated the maximum DR
where the DA of the bead cross section is in mm2, AF is the of GTAW and plasma arc welding. The maximum volumetric
cross-sectional area of the filler metal (mm2), FFS is the filler filler metal (ø1.2-mm welding wire) feeding speed limit was
130 mm3/s (3.5 kg/h), owing to its low thermal efficiency to 12.1 kg/h (11 to 26.6 lb/h). The sticking started to form at a
and far wire movement from the arc column by large lateral DR of 13 kg/h (28.6 lb/h) for the C-type filler metal whereas
vibrations at 400 A. In contrast, in this study of the welding the sticking occurred at a DR of 9 kg/h (19.8 lb/h) for the
current at 200 A, the bead appearance of the C-type filler ø2.4-mm welding wire.
metal was sound at the same DR of 3.5 kg/h without any Figure 5 exhibits the macrosection of welds for the
sticking. Furthermore, at the same 200-A current for the ø1.2-mm welding wire, ø2.4-mm welding wire, and C-type
C-type filler metal, sticking occurred when the DR increased filler metal at currents of 200 and 500 A. The penetration
to 4 kg/h (8.8 lb/h). depths were measured using macrosections. The penetration
Figure 4B shows the bead appearance at 500 A. The ø1.2- depths gradually decreased with the DR of all types of filler
and ø2.4-mm welding wires formed humping beads at a DR metals. Also, the reinforcement height gradually increased
of 5 kg/h (11 lb/h). Shirali et al. (Ref. 15) mentioned that the with the DR.
weld bead geometry changed according to the welding cur-
rent and caused humping and undercut defects during GTAW
without the addition of filler metal. In this study, humping
Effect of Process Parameters and Filler
also appeared in the beads with the addition of welding wires Metal Shape on Penetration Depth
because the welding current was very high at 500 A. Mean-
while, the bead appearances of the C-type filler metal were The results data on penetration depth and width/height
sound without any imperfections at DRs ranging from 5.04 (W/H) are listed in the Results section in Tables 1 and 2.
Fig. 7 — Relation between the NHIR and penetration Fig. 8 — A — Definition of plasma stream inflow angle
depth of the ø1.2-mm welding wire, ø2.4-mm welding θ; B — plasma stream inflow to the filler metal’s top
wire, and C-type filler metal at different welding surface of the ø2.4-mm welding wire and C-type filler
currents. metal.
Figure 6 shows the relationship between penetration 7–8 kg/h (15.3–17.6 lb/h) for the ø2.4-mm welding wire was
depth and the DR for the ø1.2-mm welding wire, ø2.4-mm narrow. However, the acceptable DR range for the C-type
welding wire, and C-type filler metal at 200 and 500 A. In filler metal was as broad as 5.04–12.1 kg/h. At 500 A, the
these two levels of welding currents, penetration depth maximum DR for the C-type filler metal was 1.51 times that of
decreased as the DR increased. Penetration depth by the the ø2.4-mm welding wire. For the same kind of filler metals,
C-type filler metal was lower than the ø1.2- and ø2.4-mm the maximum DR with 500 A was higher than with 200 A
welding wires under the same welding current and DR. Addi- because the energy density of 500 A was higher than that
tionally, penetration depth was not deep even though the of 200 A (Ref. 16). At the same current of 200 A, penetration
welding current was high at 500 A for the C-type filler metal. depth was deepest in the ø1.2-mm welding wire followed by
The effect of filler metal feed speed on penetration depth the ø2.4-mm welding wire. The C-type filler metal had the
was similar to the DR in that penetration depth decreased lowest penetration depth. At 500 A, the penetration depth of
as the filler metal feed speed increased. This was because the C-type filler metal was lower than the ø2.4-mm welding
the DR was proportional to the filler metal feed speed for wire. Lower penetration depths are needed for some appli-
the same filler metal, as given in Equation 2. The acceptable cations with high DRs, such as overlay welding and WAAM.
range, which is shown in Fig. 6, was considered based on the Figure 7 shows the relationship between the net heat input
sound welds without humping and sticking. ratio (NHIR) and penetration depth for the ø1.2- and ø2.4-mm
The maximum DR at a welding current of 200 A was 2.5 welding wires and the C-type filler metal. Penetration depth
kg/h for a ø1.2-mm welding wire, 3.0 kg/h (6.6 lb/h) for a was linearly proportional to the NHIR for the ø1.2-mm
ø2.4-mm welding wire, and 3.5 kg/h for a C-type filler metal. welding wire, ø2.4-mm welding wire, and C-type filler metal. In
At the same welding current of 200 A, the maximum DR of the the authors’ previous research on C-type filler metals, pene-
C-type filler metal was 1.4 times that of the ø1.2-mm welding tration depth could be estimated using the NHIR during GTAW
wire and 1.17 times that of the ø2.4-mm welding wire. At a (Ref. 4). At the same NHIR, penetration depth was deepest for
welding current of 500 A, the ø1.2-mm welding wire could the ø1.2-mm welding wire, medium for the ø2.4-mm welding
not deposit the sound bead, and the acceptable DR range of wire, and shallowest for the C-type filler metal at the same
Fig. 9 — A — Comparison of metal transfer modes between the welding wires and C-type filler metal at a DR
of 1 kg/h (I = 200 A, U = 20 cm/min); B — in-situ welding camera images of the filler metal’s melting edge and
molten bridge for the welding wires and C-type filler metal at a DR of 1 kg/h (I = 200 A, U = 20 cm/min).
welding current. Because of their varying cross-sectional penetration depth at the same welding current and shape of
areas, the filler metal feeding speeds for the three distinct the filler metal. The penetration depth estimation equation
filler metals/welding wires were not the same. At the same for each welding wire was developed by modifying the C-type
filler metal feed speed, the penetration depth of the ø1.2-mm filler metal’s equation using a wide range of experimental
welding wire was deeper than the others. Additionally, the data. Based on the aforementioned results, it can be said
ø2.4-mm welding wire was much deeper than the C-type that penetration depth is governed by the NHIR, welding
filler metal. The filler metal feed speed or the DR effected current, and shape of the filler metal.
Fig. 11 — Change in weld pool current and filler metal current during the separation (A), bridging (B), and
waveforms (C) of the arc voltage and filler metal voltage of the C-type filler metal.
bridge between the welding wire/filler metal and the weld the C-type filler metal waveform was linear and stable with
pool because the bridging suitability was weak. On the other no voltage fluctuations at the same DR. Such a stable voltage
hand, the C-type filler metal showed continuous bridging waveform indicates a continuous bridging metal transfer for
transfer because the bridging sustainability was strong. the C-type filler metal. As a result, the metal transfer char-
Figure 11 shows the change in weld pool current and filler acteristics of the welds made with the ø1.2- and ø2.4-mm
metal current during the separation (Fig. 11A), bridging (Fig. welding wires differed from those of the C-type filler metal,
11B), and waveforms (Fig. 11C) of the arc voltage and filler resulting in the maximum DR for the C-type filler metal.
metal voltage of the C-type filler metal (Ref. 5). The total Figure 12B shows dynamic series images of the melting
arc current equaled the sum of the filler metal current and process in the welding wires and C-type filler metal at a DR
weld pool current. Based on what is shown in Fig. 11C, during of 3.5 kg/h. The dynamic series of images revealed sticking
separation, the filler voltage was higher than the bridging, of the ø1.2-mm welding wire with indications of an unmelted
so the filler metal resistance was higher than the bridging. welding wire behind the electrode center, which progressively
Therefore, during bridging, the filler metal current could be created a ring shape. During sticking, the bended wire was
high and the weld pool current could be low. While bridging close to the electrode, which lowered the voltage potential.
transfer happened in the C-type filler metal, penetration The welding wire melted quickly from the plasma stream
depth was low and a humping bead could be prevented due then moved backward, which raised the voltage potential
to the low weld pool current — Fig. 11B. Additionally, the max- between the electrode tip and the weld pool, as shown by
imum DR of the C-type filler metal obtained was very high the voltage waveforms. Similarly, the ø2.4-mm welding wire
due to the high filler metal current. exhibited comparable sticking behavior, with the exception of
Figures 12A and B show the arc voltage signals and in-situ the presence of the extended unmelted welding wire behind
images for a DR of 3.5 kg/h (I = 200 A, U = 20 cm/min). It the electrode center. Examination of the dynamic series
was identified that welding wire sticking arose in the ø1.2- images, however, indicated that there was no difference in
and ø2.4-mm circular welding wires, whereas continuous the continuous bridging transfer in the case of the C-type
bridging arose in the C-type filler metal. Interrupted voltage filler metal. As mentioned earlier, this was due to the high filler
signals were observed in the waveform for the welds made current and plasma stream inflow of the C-type filler metal
with both the ø1.2- and ø2.4-mm welding wires, indicating advancing melting efficiency compared to the welding wires.
that the unmelted welding wire tip stuck to the weld bead and Figures 13A and B show the arc voltage signals and in-situ
interfered with arc stability and bead formation. In contrast, images for a DR of 5 kg/h, I = 500 A, U = 60 cm/min [23.6
Fig. 12 — A — Comparison of the metal transfer modes between the welding wires and C-type filler metal at a
DR of 3.5 kg/h (I = 200 A, U = 20 cm/min); B — dynamic series images of the melting process in the welding
wires and C-type filler metal at a DR of 3.5 kg/h (I = 200 A, U = 20 cm/min).
in./min]). For the autogenous GTAW, the waveform was char- Hence, the melting edge of the wire was situated in front of
acterized by significant voltage fluctuations. This was due the arc, increasing the distance between it and the weld pool.
to the high arc pressure, which caused deep and shallow When welding with the ø2.4-mm welding wire, the intensity
depressions in the weld bead. Furthermore, deep and shallow of the arc pressure was slightly reduced, and a minor number
depressions resulted in the generation of weld beads with of deep and shallow depressions were generated. The wave-
humps and valleys. With waveform voltage fluctuations, a form showed that the voltage fluctuated less for the ø2.4-mm
similar phenomenon was observed for the ø1.2-mm welding welding wire than for the ø1.2-mm welding wire. When identify-
wire. However, the voltage variation during deep and shal- ing the waveform of the C-type filler metal, the voltage signals
low depressions was smaller than during autogenous GTAW. were linear and stable with no variations. Therefore, the C-type
Therefore, even though the ø1.2-mm welding wire was fed filler metal can withstand a high arc pressure and maintain a
into the arc column, the effect of arc pressure on the weld continuous bridging transfer while also maintaining the link
pool was still greater, causing humping beads. Consequently, between the filler metal and the weld pool. As mentioned earlier,
the size or cross-sectional area of the ø1.2-mm welding wire varying deep depressions caused by the ø1.2-mm welding wire,
could not withstand the arc pressure and construct a strong ø2.4-mm welding wire, and C-type filler metal are apparent in
metal transfer bridge between the filler metal and weld pool. the still images — Fig. 13B. The weld pool depression size of
Fig. 13 — A — Comparison of metal transfer modes between the welding wires and C-type filler metal at a DR
of 5 kg/h (I = 500 A, U = 60 cm/min); B — in-situ welding camera images of the filler metal’s melting edge and
molten bridge for the welding wires and C-type filler metal at a DR of 5 kg/h (I = 500 A, U = 60 cm/min).
the ø1.2-mm welding wire was bigger than the ø2.4-mm weld- the C-type filler metal was located at the center of the elec-
ing wire followed by the C-type filler metal. The lower voltage trode, where the plasma stream was stronger and enabled the
occurred due to the shallow depression, and the higher voltage C-type filler metal to melt efficiently at higher DRs. However,
occurred due to the deep deformation. The shallow depression there was a difference in the metal transfer bridge connec-
occurred through a bridging transfer. Deep depression occurred tivity in that the C-type filler metal had a thick bridge as well
due to the separation of the bridge — Fig. 13A. as a high bridge connectivity. Simultaneously, the wires had
Figures 14A and B show the arc voltage signals and in-situ slender bridges, and their bridge connectivity was low.
images for a DR of 9 kg/h (I = 500 A, U = 60 cm/min). Inter- However, when the DR was low, which meant the filler
rupted voltage signals were observed in the waveform for the metal feeding speed was slow, the melting edge of the weld-
welds made with the ø2.4-mm welding wire due to the wire ing wires were away from the electrode center. The plasma
sticking. In contrast, the waveform of the C-type filler metal was stream melted the welding wires by generating a big drop,
stable, indicating that a continuous bridging transfer occurred. which detached from the welding wires’ tips and was trans-
The observations of in-situ images revealed sticking of the ported to the weld pool when they became large enough.
ø2.4-mm welding wire’s melting edge to the bead and minimal Because of the broken bridge with each droplet, the cre-
depression of the weld pool for the C-type filler metal — Fig. 14B. ation of a continuous metal transfer bridge was difficult for
Figure 15 shows the metal transfer characteristics of the the welding wires and filler metal. At a lower DR, an elongated
welding wires and C-type filler metal at high and low DRs. bridge was discovered for the C-type filler metal. Although
At the high DR, the melting edge of the welding wires were both the welding wires and C-type filler metal deposited at
behind the electrode center and the unmelted welding wires the same rate (1 kg/h), the mode of metal transfer was dif-
stuck to the bead surface. In contrast, the melting edge of ferent. The inflow energy of the plasma stream and the filler
Fig. 14 — A — Comparison of metal transfer modes between the ø2.4-mm welding wire and C-type filler metal at
a DR of 9 kg/h (I = 500 A, U = 60 cm/min); B — in-situ welding camera images of the filler metal’s melting edge
and molten bridge for the welding wires and C-type filler metal at a DR of 9 kg/h (I = 500 A, U = 60 cm/min).
Fig. 15 — Schematic diagram of the metal transfer mode in the circular welding wires and C-type filler metal.