Int J Adv Manuf Technol

DOI 10.1007/s00170-015-7876-6

ORIGINAL ARTICLE

Experimental investigations on the SiO2 flux-assisted

GTA welding of super-austenitic stainless steels
K. Devendranath Ramkumar 1 & Jelli Lakshmi Narasimha Varma 1 &
Gangineni Chaitanya 1 & S. Logesh 1 & Madhav Krishnan 1 & N. Arivazhagan 1 &
N. Siva Shanmugam 2

Received: 1 June 2015 / Accepted: 21 September 2015

# Springer-Verlag London 2015

Abstract This study addressed the use of activated flux SiO2 for building, petroleum, nuclear, aerospace and power generation
the gas tungsten arc (GTA) welding of super-austenitic stainless industries for joining different metals and alloys. Tseng et al.
steel plates of 5 mm thickness. Trials were carried out to investi- [1] reported that the autogenous GTA welding without the use
gate the effect of flux and the welding current to determine the of filler wire could be applied for thin sections; however, the
depth of penetration. Bead-on-plate trials were carried out on the low deposition rate and shallow joint penetration result in
AISI 904L with and without flux. It was inferred from the mac- lower productivity. Higher thickness workpieces could be suc-
rostructure studies that with a welding current of 180 A, complete cessfully joined using GTA welding process with the use of
penetration could be achieved using flux-assisted GTA welding. filler material. Also, the selection of filler metal should be
The optimal process parameters were validated by conducting the made carefully in order to obtain enhanced properties. Several
experimental investigations to ascertain the structure–property re- researchers reported that the improper selection of filler results
lationships of flux-assisted GTA weldment. Experimental results in several negative metallurgical and mechanical effects in the
corroborated that the average tensile strength and impact tough- weldments. Joint preparation is required for welding plates of
ness of SiO2 flux-assisted GTA weldments of AISI 904L were thickness >3 mm, and multi-pass welding was usually pre-
observed as 553 MPa and 49.3 J. It was inferred from the studies ferred. It was evident from the various studies that multi-
that defect-free welds of super-austenitic stainless steel could be pass welding significantly increases the heat input that causes
obtained on employing flux-assisted GTA welding process. the formation of secondary phases and grain coarsening effect
at the heat-affected zone (HAZ). Further, the formation of
Keywords Flux-assisted gas tungsten arc welding . inter-metallics and/or non-uniform distribution of elements
Super-austenitic stainless steel . Microstructure . Mechanical were noticed in the fusion zone during multi-pass welding.
property Activated flux tungsten inert gas (A-TIG) or flux-assisted
GTA welding receives major attention nowadays due to the
improvement in depth and consistency of the welds. This
welding process involves the use of a thin layer of refractory
1 Introduction
oxide (activated flux) applied in the form of thin paste to the
workpiece surface prior to welding. The heat developed dur-
Gas tungsten arc welding (GTAW) is one of the versatile and
ing the process melts and vaporizes part of the flux as arc
reliable arc welding processes widely adopted in ship
passes over the flux which results in deeper penetration.
Tseng et al. [2] and Berthier et al. [3] investigated that the
* K. Devendranath Ramkumar welding pool shape could be changed by the presence of micro-
deva@vit.ac.in elements, such as oxygen and sulphur. Similarly Tsann-Shyi
Chern et al. [4] reported that the use of SiO2, MoO3 and
1
School of Mechanical & Building Sciences, VIT University, Cr2O3 fluxes not only significantly increased the penetration
Vellore 632014, India capability but also improved the mechanical strength of the
2
Department of Mechanical Engineering, NIT, stainless steel welds compared with conventional TIG welds.
Tiruchirappalli 620002, India Shyu et al. [5] investigated the performance of 304 SS obtained
 Int J Adv Manuf Technol

from activated flux TIG welding employing Al2O3, Cr2O3, that Mo segregation could be reduced or controlled with the
TiO2, SiO2 and CaO fluxes. The authors observed the better use of Mo-rich filler wires. Also, they have recommended the
depth-to-width ratio, improved penetration and reduced angular use of faster cooling rates for controlling the same.
distortion while employing the A-TIG welding process. It is evident from the open literature that the flux-assisted or
Morisada et al. [6] reported that the depth/width ratio was A-TIG welding has potential applications in the industries. Sev-
affected by the shielding gas flow rate during the A-TIG eral researchers reported the viability of carrying out activated
welding of SUS 304 stainless steel. Kuo et al. [7] investigated flux TIG welding of stainless steels. However, no such detailed
the joining of dissimilar metals involving G3131 mild steel and study has been reported on the joining of super-austenitic stain-
SUS 316L stainless steel using activated flux TIG welding less steel AISI 904L. This study addressed the effect of SiO2
process using different fluxes such as CaO, Fe2O3, Cr2O3 and flux and welding current on the weldability, microstructure and
SiO2. The authors concluded that welding with SiO2 powder mechanical properties of 5-mm-thick super-austenitic stainless
resulted in an increase of weld depth-to-width ratio and a higher steels. The outcomes of the study will be beneficial to the in-
degree of energy concentration. Li et al. [8] reported that the arc dustries employing this grade of stainless steel.
shrinkage is the main aspect to increase the weld penetration on
using SiO2 flux compared to TiO2 flux. Similarly Tseng et al.
[1] compared the TiO2 and SiO2 flux-assisted TIG welding 2 Experimental procedure
processes for joining AISI 316L and concluded that the energy
density of SiO2 flux-assisted TIG welding is higher than that of The chemical composition of the base metal, AISI 904L, is
TiO2-flux-assisted TIG welding. Sakthivel et al. [9] compared ascertained using dry spectroscopic method and represented in
the creep behaviour of multi-pass TIG and activated flux TIG Table 1. The mechanical properties of the base metal inferred
welding process of 316L (N) stainless steel. The authors ob- as 647 MPa and 61 J as average tensile strength and impact
served lower hardness at the fusion zone of A-TIG weldments toughness.
compared to multi-pass TIG weldments owing to the presence Bead-on-plate welding was carried out with and without
of columnar dendrites, whereas the equiaxed dendrites domi- flux using the various welding currents on the 5-mm-thick
nated in the A-TIG fusion zone. plate of AISI 904L. Table 2 shows the data representing the
AISI 904L is a super-austenitic stainless steel, widely used depth of penetration, bead width and heat input developed
in marine, petroleum and nuclear industries owing to their during the autogenous GTA welding of super-austenitic stain-
combined mechanical properties such as strength and high less steel without the use of flux. The activated flux used in the
temperature corrosion resistance. Also, the presence of suffi- study is commercially pure SiO2 powder. Before carrying out
cient amount of Mo enhances the hot cracking resistance of welding, the SiO2 flux powder was uniformly mixed with
this stainless steel. It was reported that autogenous GTAW of methanol which is used as carrier solvent to form a paint-
super-austenitic steel with shielding gases and arc lengths like consistency. This was manually applied using a paint-
yielding nitrogen content in excess of 0.2 %, resulted in po- brush so as a sufficient layer thick to prevent visual observa-
rosity [10]. Zambon et al. [11] investigated the CO2 laser tion of the base metal underneath. Bead on trials was carried
welding of 3-mm-thick plates of super-austenitic stainless out on the plates that were coated with flux. A 15-mm-wide
steel. The authors observed an imperceptible HAZ which flux coating was applied throughout the length of the plates,
allowed the retention of high toughness. One of the major with joint line at the centre. The flux quantity applied on the
problems associated with welding of AISI 904L was Mo seg- plate was 4 to 5 mg/cm2 as the coating density has significant
regation. Banovic et al. [12] reported that during solidification impact on the depth of penetration. Further, the cross-section
of the weld, Mo segregates preferentially to the liquid, owing macrostructure studies were carried out to infer the role
to the low solubility of Mo in the austenite phase (γ), and welding current and the flux in achieving the aforementioned
leaves the first solid to form depleted in Mo. Further, the process parameters and are shown in Table 3. Ensued to the
low diffusion rate of Mo in austenite phase does not allow results from Table 3, the A-TIG welding was carried out on
for Mo to diffuse back towards the dendrite cores to eliminate the plates of 170 mm×55 mm×5 mm which were obtained
the concentration gradient and can lead to poor corrosion re- from wire cut electrical discharge machining (WEDM)
sistance of the weld metal. The researchers have addressed process.

Table 1 Chemical composition of base metal, AISI 904L

Chemical composition (by wt%)

Base metal C Cr Ni Mo Fe Cu Other elements

AISI 904L 0.017 19.64 24.82 4.02 Rem. 1.27 P—0.018; S—0.003; N—0.037; Si—0.374; Mn—1.33
Int J Adv Manuf Technol

Table 2 Geometric characteristics of autogenous GTA welding of AISI 904L—without flux

11.9 Current, 180 A
Depth of penetration, 1.57 mm
1.57 Bead width, 11.9 mm
Heat input, 0.481 kJ/mm

14.0 Current, 200 A

Depth of penetration, 2.29 mm
2.29 Bead width, 14.0 mm
Heat input, 0.563 kJ/mm

11.4 Current, 220 A

Depth of penetration, 2.60 mm
2.60 Bead width, 11.4 mm
Heat input, 0.457 kJ/mm

Current, 240 A
14.19
Depth of penetration, 1.85 mm
Bead width, 14.19 mm
1.85
Heat input, 0.434 kJ/mm

Current, 260 A
12.61
Depth of penetration, 2.79 mm
2.79 Bead width, 12.61 mm
Heat input, 0.554 kJ/mm

Square butt configuration was employed on the plates, and a standard 2 % thoriated tungsten electrode was used for welding
single-pass, flux-assisted autogenous GTA welding was carried trials using Kemppi MasterTIG welding machine. Argon of
out along the centreline of test specimen to produce the weld. A 99.99 % purity, at a constant flow rate of 18 lpm, was used as
 Int J Adv Manuf Technol

Table 3 Geometric characteristics of SiO2 flux-assisted GTA welding of AISI 904L

6.47 Current, 140 A
Depth of penetration, 4.02 mm
Bead width, 6.47 mm
4.02 Heat input, 0.375 kJ/mm

7.91 Current, 160 A

Depth of penetration, 4.63 mm
Bead width, 7.91 mm
4.63 Heat input, 0.423 kJ/mm

6.23 Current, 180 A

Depth of penetration : full penetration (5 mm)
Bead width, 6.23 mm
5.23 Heat input, 0.465 kJ/mm

6.23
8.74 Current, 200 A
Depth of penetration, 4.93 mm
Bead width, 8.74 mm
4.93 Heat input, 0.553 kJ/mm

the shielding gas and a constant travel speed was maintained as 15 mm×5 mm. Standard metallographic procedures were
140 mm/min. The process parameters employed for complete adopted to reveal the microstructure of the weldments. Elec-
welding of 170-mm-long plate of AISI 904L plates using SiO2 trolytic etching (10 % oxalic acid, 6 V DC supply, current
flux-assisted GTA welding are shown in Table 4. density 1.6 A/cm2) was employed to reveal the grain structure
Non-destructive testing (NDT) analysis was carried out of the different zones of these weldments. Microstructure ex-
using gamma ray radiography to determine for any flaws at amination was carried out using optical microscopy (OM) and
the surface or sub-surface of the weldments. Ensued to the scanning electron microscopy (SEM) techniques. SEM/
NDT results, the coupons were cut to different dimensions energy-dispersive X-ray analysis (EDAX) point analysis and
according to the standards for establishing the metallurgical line mapping analysis were also performed to understand the
and mechanical properties. Macro, microstructure studies and movement of elements across the weldments. To ascertain the
hardness measurements were carried out on the cross- mechanical properties, the weldments were subjected to dif-
sectioned coupons whose dimensions were 28 mm × ferent mechanical tests and were outlined.
Int J Adv Manuf Technol

Table 4 Process parameters employed for flux-assisted GTA welding the same. The cross-section macrographs of GTA weldments
of AISI 904L plates
without flux addition shown in Table 2 revealed that the maxi-
Process parameter Unit Value mum penetration achieved was 2.79 mm for the current of
260 A. It was also inferred that the width of the bead was wid-
Current A 180 ened in all the welding trials.
Voltage V 9.9–11.7 Followed by this, the bead on plate GTA welding trials were
Travel speed mm/min 140 performed on the AISI 904L plates with the addition of SiO2 flux
Shielding gas – Argon by varying the welding current from 140 to 200 A. The cross-
Shielding gas flow rate lpm 18 sectional macrographs of flux-assisted GTA welding of AISI
Heat input kJ/mm 0.511 904L plates are represented in Table 3. It is vividly seen that
the depth of penetration increases with an increase in welding
current for 140, 160 and 180 A. The penetration depth was
Hardness measurements were carried out on the cross- slightly lowered on increasing the current value to 200 A. Pen-
sectioned coupon of the weldment vis-à-vis top, middle and etration depths are also inversely proportional to travel speed of
bottom locations using Vickers micro-harness tester. A standard the torch at a particular weld current. Similar observations were
load of 500 gf was applied for a dwell time of 10 s, and the also reported by [1, 2]. The results showed that a complete pen-
readings were measured at regular intervals of 0.25 mm. Further, etration could be achieved for the current of 180 A.
the tensile studies were performed on the coupons fabricated as The macrographs of the flux-assisted GTA welding showed
per the ASTM: E8/8M standard. Tensile test was carried out that the weld beads were deeper and narrower for all the current
using Instron Universal Testing Machine employing a constant values except for the weld bead obtained from 200 A, offered a
cross-head velocity of 2 mm/min to induce a strain rate of 3.3× wider and broader bead profile. Based on the trial studies, it is
10−4 s−1. Charpy V-notch studies were carried out on the cou- opined that the SiO2 flux-assisted GTA welding resulted in
pons prepared as per ASTM: E23-12c standard to conjecture the higher penetration depth for a current of 180 A. Further, the
response of the A-TIG weldments towards impact loading. flux-assisted GTA welding was carried out on the AISI 904L
plates based on the primitive studies, and the process parameters
employed are represented in Table 4.
3 Results
3.2 Metallurgical characterization of flux-assisted GTA
3.1 Pre-study on the penetration depth of autogenous GTA weldment of AISI 904L
welding with and without flux
Macrostructure of the butt-welded AISI 904L plates obtained
Bead-on-plate trials were carried out using GTA welding tech- from flux-assisted GTA welding process is shown in Fig. 1.
nique by keeping constant welding speed employing with and The microstructure at the different locations of the weldment is
without flux in order to establish the welding current for achiev- also represented in Fig. 1a–e. Interface microstructure shown in
ing the maximum depth of penetration. The trials were conduct- Fig. 2 revealed the absence of heat-affected zone or the presence
ed by maintaining the welding current of 140 and 160 A without of any secondary phases. Weld microstructures at top, middle
flux addition. Since the depth of penetration was very lower, the and bottom are shown in Fig. 3a–c. Both cellular and columnar
current values were varied to 180, 200, 220, 240 and 260 A for dendritic growth was observed at the top and middle regions of

Fig. 1 a–e Macrostructure of

SiO2 flux-assisted GTA-welded
AISI 904L showing the complete
fusion on employing current of (a) (b) (c) (d) (e)
180 A

(a) (b) (c) (d) (e)

 Int J Adv Manuf Technol

was noticed from the SEM/EDAX point analysis that the weight
percentage of Si was slightly higher in the fusion zone than the
parent metal (Fig. 5d).

AISI 904L
Fusion zone 3.3 Mechanical characterization of flux-assisted GTA
weldment of AISI 904L

Hardness measurement was carried out on the cross-sectioned

coupon of the AISI 904L weldment at top, middle and bottom
regions. Figure 6 shows the hardness profile of the flux-assisted
GTA weldment of AISI 904L. The average hardness at the top,
middle and bottom regions of the fusion zone was observed as
151, 155 and 153 HV, respectively. It was inferred that the hard-
Fig. 2 Interface microstructure of the SiO2 flux-assisted GTA weldments ness trend was uniform across the weldment. However, there
of AISI 904L were some variations in the hardness values noticed at the fusion
zone and parent metal.
the fusion zone, whereas the fine, equi-axed cellular growth was Tensile studies carried out on the flux-assisted AISI 904L
revealed at the bottom region of the fusion zone. Line mapping GTA weldments clearly showed that the fracture was experi-
analysis shown in Fig. 4 indicated clearly that there was no much enced at the fusion zone for all the trials. Table 5 shows the
elemental movement across the weldment. SEM/EDAX point average tensile properties of the flux-assisted GTA weldments
analysis at the fusion zone of the flux-assisted GTA-welded AISI of AISI 904L. It was observed from the visual examination
904L is shown in Fig. 5a–c. The line spectra (Fig. 5a) indicated (Fig. 7a) that the weldments experienced higher amount of plas-
the presence of elements such as Fe, Cr, Ni, Cu and Mo at the tic deformation before rupture. The average tensile strength and
grain body of the fusion zone. The weight percentage of the 0.2 % offset yield strength were observed as 553 and
elements in the grain body was observed to be almost closer to 274.56 MPa, respectively. The average percentage elongation
the base metal’s composition. However, the presence of tiny at the break point was inferred to be 36.01 % for the flux-
secondary phases observed at the dendritic arms of the fusion assisted GTA-welded 904L plates (Table 6). SEM fractograph
zone (Fig. 5b, c) which are enriched with Mo, Cr, Ni and Fe. It of the tensile tested sample showed the presence of macro/micro-

Fig. 3 Fusion zone Top Middle

microstructure of the SiO2 flux- (a) (b)
assisted GTA weldments of AISI Equiaxed
904L showing a top, b middle
and c bottom regions

Equiaxed

Columnar dendritic

Bottom
(c)

Equiaxed
Int J Adv Manuf Technol

AISI 904L
Weld zone

Ni Cr Fe Mn Mo

Fig. 4 Line mapping analysis at the interface of SiO2 flux-assisted GTA-welded AISI 904L

voids with the ridge pattern. Also, the secondary phases were reported by Vasantharaja et al. [14]. The effect of surface tension
embedded in the few of the void spaces (Fig. 7b). also plays a main role in enhancing the depth of penetration of
Charpy V-notch studies demonstrated the deformation oc- the welds. According to Huang et al. [15], the surface tension
curred at the notch zone without undergoing complete rupture would be higher near the centre region of the weld pool, and the
upon impact loading (Fig. 8a). The average impact toughness of fluid flow would be inwards along the surface of the weld pool
the flux-assisted GTA weldments was 49.3 J (Table 7). SEM towards the centre and then down, thereby increasing the weld
analysis on the impact tested sample revealed the presence of penetration in case of flux-assisted welding.
scarce voids and the rock sand appearance as the topographic The primitive studies showed that a current of 180 A could
features. suffice to produce full penetration of 5.23 mm, and hence, this
current value was set for joining AISI 904L plates using SiO2
flux. The macrostructure of butt-welded AISI 904L plates using
4 Discussion flux-assisted GTA welding showed that complete fusion oc-
curred in single pass exhibiting shallow bead profile (Fig. 1).
This study inferred clearly that the successful welds of AISI This clearly elucidated that a lower arc heat would be sufficient
904L could be obtained by SiO2 flux-assisted GTA welding for obtaining successful joints of super-austenitic stainless with
technique. It can be inferred from Table 2 that the bead width the aid of SiO2 flux. As SiO2 increases the fluidity of the mate-
was also higher. Similarly the trials were carried out to establish rial, a complete fusion with the nominal bead width has resulted.
the optimal weld current in achieving maximum depth of pene- Similar observations were reported elsewhere [1]. As reported
tration during SiO2 flux-assisted GTA welding. The results by Kuo et al. [7], the shallow and deep welds could be achieved
showed that the depth of penetration increases with an increase by GTAwelding assisted with the Fe2O3, Cr2O3 and SiO2 fluxes.
in the welding current. As reported by Tseng [13], the penetra- Several mechanisms were addressed by various researchers to
tion depth is inversely proportional to welding speed at a partic- reason out the higher depth of penetration achieved via A-TIG
ular weld current. Further, the author reported that beyond the welding process. The flux used in the welding process shall
optimal coating density, the penetration depth remains approxi- serve as an insulation barrier to the arc current. As reported by
mately constant before decreasing. The reduction in the penetra- Lowke et al. [16], the arc was constricted due to this insulating
tion depth could be attributed to the higher arc heat energy con- effect of flux leading to higher temperatures at the centre of the
sumption required to melt the thick flux barrier. It was observed weld pool, which would be sufficient to melt the flux so that the
that the weld bead width was reduced by almost half for the same electric current could penetrate the flux to the weld pool and
welding current with the use of activated flux which shall be workpiece. Moreover, for a given current, the current density
attributed to arc constriction effect. Similar observations were at the centre of the weld pool would be increased and led to
 Int J Adv Manuf Technol

Fig. 5 a–c SEM/EDAX point (a)

mapping analysis at the fusion
Si K 0.49%
zone of SiO2 flux-assisted GTA- Cr K 20.15%
welded AISI 904L. d Area Mn K 1.23%
mapping analysis on the parent Fe K 49.97%
metal of AISI 904L Ni K 23.19%
Cu K 1.01%
Mo L 3.95%

Grain

(b) Si K 0.70%
Cr K 21.31%
Mn K 1.56%
Fe K 43.65%
Ni K 22.97%
Cu K 1.29%
Mo L 8.52%

(c)
Si K 0.66%
Cr K 21.45%
Mn K 1.86%
Fe K 42.10%
Ni K 22.49%
Cu K 1.38%
Mo L 10.06%

(d)
Si K 0.49%
Cr K 20.82%
Mn K 1.61%
Fe K 47.28%
Ni K 23.42%
Cu K 1.34%
Mo L 5.04%

increase the magnetic pinch forces and pressure in the weld pool The microstructure at different locations of the weldment is
which would result in strong convective flow downwards in the shown in Fig. 1a–e. Microstructure at the weld interface clear-
weld pool and an increased weld depth [16]. ly indicated the absence of secondary phases or unmixed zone,
It is evident from the present study that there was no evi- which shall be attributed to lower heat input and the faster
dence of hot cracking tendency in the fully austenitic welds of cooling rate and also due to the constriction of arc focussed
AISI 904L as the penetration was achieved in single pass with at the weld pool, enunciated with flux-assisted welding. Weld
the use of flux. This is well in agreement with Lundin et al. microstructure depicted the presence of coarse columnar, cel-
[17]. According to these authors, avoiding the multiple ther- lular and equiaxed dendritic growth prevailing in different
mal cycles and the enrichment of Mo and Mn at the weld and locations of the fusion zone. Specifically, the bottom region
HAZ enhanced the hot cracking tendency resistance. of the fusion zone was witnessed to have equiaxed-dendritic
Int J Adv Manuf Technol

Top

Middle

Bottom

Fig. 6 Hardness measurement across the different zones of the flux-assisted GTA-welded AISI 904L

microstructure. Specifically, the centre of the fusion zone was noticed at the dendritic arms of the fusion zone. The EDAX
witnessed to have equiaxed-dendritic microstructure. As re- point analysis showed that these secondary phases were richer
ported by Norman et al. [18], the formation of an equiaxed- with the elements Cr, Mo, Ni and Si. This could be due to
dendritic microstructure in the centre of autogenous welds micro-segregation of these elements in weld fusion zones. The
shall be due to the development of an undercooled liquid present study also attested the scarce amount of Mo segrega-
ahead of the moving solid–liquid interface, which provides tion at the inter-dendritic arms of the fusion zone which could
the correct thermal conditions for the nucleation and growth be attributed to the moderate cooling rates in the flux-assisted
of new grains. The equiaxed growth predominant in the fusion TIG welding process. Due to the differences in the Mo at the
zone could be attributed to the faster cooling experienced in inter-dendritic arms and the matrix, the corrosion potential
flux-assisted GTA welding process. As reported earlier, the would be slightly different and probably reduces the corrosion
flux-assisted GTA welding results only in re-melting and so- resistance compared to the parent metal. Further studies are
lidification, and the changes in the microstructure alone were required to confirm the same.
experienced in the fusion zone. Hardness studies showcased that the average hardness of
Similar observations were depicted by Sakthivel et al. [9]. the fusion zone of flux-assisted GTA weldment was observed
Also, the weld zone was found to be complete austenitic struc- to be 155, 151 and 153 HV at top, middle and bottom loca-
ture even though the rate of cooling is faster. This is because tions. The study inferred that the average weld hardness was
the super-austenitic stainless steel has enriched amount of slightly impoverished compared to the hardness of the parent
austenite-stabilizing elements Ni and N. The presence of metal (163–165 HV). As reported earlier, the microstructure
higher amounts of Ni in the weld zone was once again con- of the fusion zone was predominantly observed with equiaxed
firmed from the SEM/EDAX point analysis which depicted microstructure which may be reasoned for lower hardness.
that Ni is almost uniform both in the grain body and boundary. This is well in agreement with the work carried out by other
The ferrite analysis showed that the ferrite number was zero researcher [9]. The authors reported that columnar structure
which clearly nurtured that the fusion zone was completely displayed higher hardness than the equiaxed structure because
austenitic. No much elemental variations were observed dur- of high dislocation density in the columnar structure. Tensile
ing flux-assisted GTA welding as indicated from the line map- results are well in agreement with the hardness data such that
ping analysis (Fig. 4). Scarce tiny precipitates were also tensile fracture experienced at the fusion zone for all the trials
owing to lower hardness values. The average tensile strength
Table 5 Hardness measurement on the flux-assisted GTA weldments of the SiO2 flux-assisted GTA weldments of AISI 904L was
of AISI 904L reported as 553 MPa which is lower than the strength of the
parent metal (647 MPa). The joint efficiency [defined as
Description Vickers micro-hardness value (HV)
(UTSweld joint)/(UTSbase metal)×100] of the flux-assisted AISI
Top Middle Bottom 904L weldments was ascertained to be 85.47 %. The average
ductility of the weldment was reported as 36.01 %. The tensile
Average hardness of 160 157 159 results obtained from the current study were greater than that
the A-TIG weldment
of the earlier works carried out by the authors using compound
Average hardness at the 155 151 153
fusion zone flux. However, the average tensile strength of SiO2 flux weld-
ments was slightly lower than that of the autogenous GTA
 Int J Adv Manuf Technol

Fig. 7 a Tensile photographs of (a)

the flux-assisted GTA weldments
of AISI 904L before and after test.
AISI 904L
b SEM fractograph of the tensile-
tested AISI 904L weldment

AISI 904L

AISI 904L

Fracture observed at the weld zone

(b)

Ridges

Voids

Secondary phases

weldment without flux [19]. SEM fractographs elucidated the scope of research can be extended by studying the effect of
mode of fracture as ductile which could be confirmed by the welding speed, flux and the coating thickness of the flux on
presence of macro/micro-voids in the fractured surface. Also, the depth to width ratio, depth of penetration, weldability, and
Charpy V-notch test trials inferred the average impact tough- structure–property relationships of AISI 904L.
ness of the flux-assisted weldments as 49.3 J which was slight-
ly lower than the toughness of the base metal (61 J). The lower
toughness could be attributed to the complete austenitic struc-
ture prevailing in the weld zone. In addition to that, sometimes 5 Conclusions
the oxides could be trapped inside during the A-TIG welding
process due to the improper melting of the flux. This will tend This comprehensive study addressed the feasibility of joining
to deteriorate the impact toughness and corrosion resistance of AISI 904L super-austenitic stainless steel by SiO2 flux-
the weld metal. The present investigation is well in agreement assisted GTA welding. The following are the conclusions de-
with Vasantharaja et al. [14]. duced from the present study.
SEM fractograph of the impact tested sample corroborated
that the presence of scarce voids and the beach sand markings Table 6 Tensile properties of the flux-assisted GTA weldments of AISI
revealed the quasi-cleavage mode of fracture; however, the 904L
visual examination of SiO2 flux-assisted GTA weldments of
AISI 904L reported the V-notch deformation in the fusion Mechanical property Unit Trial 1 Trial 2 Trial 3 Average
zone without undergoing complete rupture. Max. load kN 14.72 15.82 15.40 15.32
In a nutshell, this study reported the detailed study on the 0.2 % yield strength MPa 271.63 279.84 272.23 274.56
effect of SiO2 flux on the weldability, microstructure and me- Ultimate tensile strength MPa 546.0 563.0 548.0 553.0
chanical properties of flux-assisted GTA welding of 904L
% Elongation at break % 35.38 37.45 35.33 36.01
plates. It is observed that the welding current plays a signifi-
Fracture zone – Weld Weld Weld –
cant role in obtaining the depth of penetration. Similarly the
Int J Adv Manuf Technol

(a) Table 7 Impact test results of flux-assisted AISI 904L weldments

Mechanical property Unit Trial 1 Trial 2 Trial 3 Average

V-notch
deformation Impact toughness J 48 48 52 49.3

5. Tensile studies corroborated the fracture experienced at the

fusion zone in all the trials, and the average tensile strength
was reported to be 553 MPa and the joint efficiency of
85.42 % could be ascertained during flux-assisted GTA
welding. Further SEM fractograph results confirmed the duc-
tile mode of fracture by showing micro-voids and dimple
facets.

Acknowledgments The authors convey their sincere thanks to the De-

partment of Science and Technology–Science and Engineering Research
Board (DST–SERB) for funding this research work (File No. YSS/2014/
000706). The authors also wish to convey their sincere thanks to DST for
equipping our university with Universal Testing Machine under FIST
scheme. The support rendered by Sathyabama University for carrying
out SEM/EDAX analysis is duly acknowledged.
(b)
References

1. Tseng KH, Chen KL (2012) Comparisons between TiO2‐ and

SiO2‐flux assisted TIG welding processes. J Nanosci
Nanotechnol 12:6359–6367
2. Tseng KH, Hsu CY (2011) Performance of activated TIG process in
austenitic stainless steel welds. J Mater Prod Technol 211:503–12
3. Berthier A, Paillard P, Carin M, Valensi F, Pellerin S (2012) TIG
and A-TIG welding experimental investigations and comparison to
simulation. Part 1: identification of Marangoni effect. Sci Technol
Weld Join 17:609–15
4. Chern T-S, Tseng K-H, Tsai H-L (2011) Study of the characteristics
of duplex stainless steel activated tungsten inert gas welds. Mater
Des 32:255–263
5. Shyu SW, Huang HY, Tseng KH, Chou CP (2008) Study of the
performance of stainless steel A-TIG welds. J Mater Eng Perform
Fig. 8 a Impact test carried out on the flux-assisted GTA weldments of
17:193–201
AISI 904L before and after test. b SEM fractograph of the impact tested
AISI 904L weldment 6. Morisada Y, Fujii H, Xukun N (2014) Development of simplified
active flux tungsten inert gas welding for deep penetration. Mater
Des 54:526–530
7. Kuo C-H, Tseng K-H, Chou C-P (2011) Effect of activated TIG
1. Successful welds of AISI 904L could be obtained by flux on performance of dissimilar welds between mild steel and
employing SiO2 flux-assisted GTA welding process. stainless steel. Key Eng Mater 479:74–80
8. Qing-ming L, Wang X-h, Zeng-da Z, Wu J (2007) Effect of acti-
2. Primitive studies showed that for the same current of 180 A,
vating flux on arc shape and arc voltage in tungsten inert gas
the depth of penetration was more than twice for the flux- welding. Trans Non-ferrous Met Soc China 17:486–490
assisted GTA welding process compared to autogenous 9. Sakthivel T, Vasudevan M, Laha K, Parameswaran P, Chandravathi
welding. KS, Mathew MD, Bhaduri AK (2011) Comparison of creep rupture
3. There were no unmixed zone or secondary phases were behaviour of type 316L (N) austenitic stainless steel joints welded
by TIG and activated TIG welding processes. Mater Sci Eng A 528:
observed at the weld interface of flux-assisted GTA weldment 6971–6980
of AISI 904L. Both equiaxed and columnar dendrites were 10. Hertzman S, Pettersson RJ, Blom R, Eriksson EKJ (1996) Influence
observed in the fusion zone. However, the equiaxed dendrites of shielding gas composition and welding parameters on the N-
dominated in the centre of the fusion zone. content and corrosion properties of welds in N-alloyed stainless
steel grades. ISIJ Int 36(7):968–976
4. The presence of dense, fine equiaxed dendritic microstruc- 11. Zambon A, Ferro P, Bonollo F (2006) Microstructural, composi-
ture could be reasoned for the lower hardness at the fusion tional and residual stress evaluation of CO2 laser welded super-
zone of flux-assisted GTA weldment of AISI 904L austenitic AISI 904L stainless steel. Mater Sci Eng A 424:117–127
 Int J Adv Manuf Technol

12. Banovic SW, DuPont JN, Marder AR (2002) Dilution and 16. Lowke JJ, Tanaka M, Ushio M (2005) Mechanisms giving in-
microsegregation in dissimilar metal welds between super austen- creased weld depth due to a flux. J Phys D Appl Phys 38:3438–
itic stainless steel and nickel base alloys. Sci Technol Weld Join 3445
7(6):374–383 17. Lundin CD, Chou CPD, Sullivan CJ, (1980) Hot cracking resis-
13. Tseng K-H (2013) Development and application of oxide-based tance of austenitic stainless steel weld metals. Welding Research
flux powder for tungsten inert gas welding of austenitic stainless Supplement, American Welding Society: 226-s–232-s
steels. Powder Technol 233:72–79 18. Norman AF, Drazhner V, Prangnell PB (1999) Effect of welding
14. Vasantharaja P, Vasudevan M (2012) Studies on A-TIG welding of parameters on the solidification microstructure of autogenous TIG
low activation ferritic/martensitic (LAFM) steel. J Nucl Mater 421: welds in an Al–Cu–Mg–Mn alloy. Mater Sci Eng A259:53–64
117–123 19. Devendranath Ramkumar K, Varma JLN, Chaitanya G, Choudhary
15. Huang HY, Shyu SW, Tseng KH, Chou CP (2005) Evaluation of A, Arivazhagan N, Narayanan S (2015) Effect of autogeneous GTA
TIG flux welding on the characteristics of stainless steel. Sci welding with and without flux addition on the microstructure and
Technol Weld Join 19(5):566–573 mechanical properties of AISI904L joints. Mater Sci Eng A 636:1–9