Author’s Accepted Manuscript

Si-Bronze to 304 Stainless Steel GTA Weld Fusion

Zone Microstructure and Mechanical Properties

N. Switzner, H. Querioz, J. Deurst, Z. Yu

www.elsevier.com/locate/msea

PII: S0921-5093(17)31183-8
DOI: http://dx.doi.org/10.1016/j.msea.2017.09.025
Reference: MSA35492
To appear in: Materials Science & Engineering A
Received date: 13 July 2017
Revised date: 4 September 2017
Accepted date: 6 September 2017
Cite this article as: N. Switzner, H. Querioz, J. Deurst and Z. Yu, Si-Bronze to
304 Stainless Steel GTA Weld Fusion Zone Microstructure and Mechanical
P r o p e r t i e s , Materials Science & Engineering A,
http://dx.doi.org/10.1016/j.msea.2017.09.025
This is a PDF file of an unedited manuscript that has been accepted for
publication. As a service to our customers we are providing this early version of
the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting galley proof before it is published in its final citable form.
Please note that during the production process errors may be discovered which
could affect the content, and all legal disclaimers that apply to the journal pertain.
 Si-Bronze to 304 Stainless Steel GTA Weld Fusion Zone Microstructure and Mechanical Properties

Dissimilar gas tungsten arc (GTA) welds of Si-bronze to 304 stainless steel were characterized with
optical microscopy, microhardness, SEM, tensile testing, and fractography.

N. Switzner, H. Querioz, J. Deurst, and Z. Yu*

George S. Ansell Department of Metallurgical and Materials Engineering, Colorado School of Mines,
Golden, CO 80401

*Corresponding Author: zyu@mines.edu

Abstract

Si-bronze and 304 stainless steel were joined using gas tungsten arc (GTA) welding with two different
welding (filler) rods matching the base metals. Cross-sections of the welds were analyzed using light
optical microscopy (LOM), scanning electron microscopy (SEM), energy dispersive X-ray spectroscopy
(EDS), and hardness testing. Mechanical testing was performed for welds from each of the two filler
metals using cross-weld tensile bars. For (1) the Si-bronze filler weld, partial mixing resulted in (a)
copper-based regions with enhanced hardness due to iron-based islands and (b) iron-based regions with
elevated hardness due to Si enrichment. Joint strength was ~90% that of the base Si-bronze strength of
340 MPa, and elongation was ~48%, as compared to the base Si-bronze elongation of 55%. For (2) the
stainless steel filler rod (a) primary austenite solidification and (b) micro-fissures were identified in the
fusion zone. Thermodynamics calculation software was used to model solidification mode,
demonstrating that the Cu (contributed by the Si-bronze) stabilized primary austenite solidification,
which led to solidification micro-cracking or micro-fissuring. Micro-fissuring caused low ductility in the
transverse weld tensile specimens. Fractography was performed to characterize the fracture
mechanisms. For the stainless filler weld, a Cu film formed in interdendritic spaces between primary
austenite grains.

Keywords

Si-bronze, 304 stainless steel, dissimilar welding, gas tungsten arc (GTA) welding, fractography, Scheil
solidification modeling

1 Introduction

Copper and bronze to stainless steel joints are used in a variety of commercial applications such as
architecture, engine components, water and gas systems, pipe and pump weld and repair, and fittings
(1). Fusion welding of these two dissimilar metals presents a challenge because of the significant
differences in their thermophysical properties and low solid solubility. For instance, the melting range of
Si-bronze is 971-1027 °C, and the melting range of 304 stainless steel is 1399-1454 °C (2).

Several studies have compared dissimilar welds of combinations of various Cu alloys and stainless steels
with wide-ranging results. For Cu to 304L stainless steel electron beam welds, Magnabosco et al.
identified regions of Fe-based matrix with Cu-based islands and a Cu-based matrix with Fe-based islands
in the fusion zone. Segregation occurred during solidification due to the low solid solubility of Fe and Cu
(3). For Monel 400 (65Ni/30Cu) to 304 stainless steel GTA welds, Ramkumar et al. showed that stainless
steel (E309L) filler rod resulted in tensile fracture in the weld fusion zone whereas Monel (ENiCu-7) filler
rod resulted in fracture in the base metal (4). For Cu to alloy steel shielded metal arc (SMA) welds, Velu
and Bhat showed that Sn-bronze (93Cu7Sn) electrodes (filler) resulted in spatter and weld porosity,
likely due to the vaporization of Sn from the fusion zone. However, Ni-based electrodes
(68Ni/21Cr/6Mn) resulted in a weld tensile strength of 208 MPa and 12% elongation in comparison to
the base Cu 264 MPa strength and 33% elongation (5). For Cu to stainless steel SMA welding, Roy et al.
showed that Inconel (ENiCrMo-3) electrodes led to the highest strength of 320 MPa and E316L
electrodes led to the highest ductility of 16% (6). For Cu to stainless steel GTA welding, Shiri et al.
showed that Cu filler rod resulted in a joint tensile strength of 192 MPa, and ductility of 20%, whereas
the base Cu tensile strength and elongation were 200 MPa and 46% elongation, respectively (7).

In this study, GTA welds of Si-bronze (97Cu/3Si) with 304 stainless steel were made using two different
filler rods matching the base metals for comparison. Si-bronze and 304 stainless steel are structural
alloys commonly used in atmospheric environments (8). The microstructure and mechanical properties
of dissimilar welds of these alloys have not been reported in the literature. The GTA welding technique
is a versatile and widely available welding process. Thus the GTA welding technique was selected along
with two types of filler electrodes, matching the two base materials. The foci of this study included:
(i) establishing a welding parameter baseline for joining the dissimilar metals Si-bronze and 304 stainless
steel, (ii) comparing the influence of filler metal on the microstructural evolution and mechanical
properties of the dissimilar joints, and (iii) characterizing the failure mechanisms in the dissimilar welds.

2 Experimental Methods

Dissimilar 304 stainless steel and Si-bronze plates of 6 mm thickness were given a 2 mm thick bevel, wire
brushed, and joined using GTA welding. The base metal compositions from optical emission
spectroscopy (OES) are given in Table 1. The filler rods were 1.6 mm in diameter. Two types of filler rod
were used, ERCuSi-A and ER308, for which the composition ranges are provided in Table 1. The GTA
welding parameters for two trial welds are given in Table 2. Both the two-pass Si-bronze filler weld and
the one-pass stainless filler weld exhibited mixing in the fusion zone. Metallurgical samples were
sectioned across the weld for microstructure analysis. Samples were ground, hand-polished and
vibratory polished with 0.05 µm colloidal silica for one hour for SEM analysis. Microhardness testing was
conducted using a 500g load and a 12 s dwell time. Scanning electron microscopy (SEM) and energy-
dispersive X-ray spectroscopy (EDS) were performed using an FEI Quanta 600. In preparation for light
optical microscopy (LOM), samples were etched for 30 seconds by swabbing with cotton balls soaked
with glyceregia (10 ml HNO3, 20 ml glycerol and 30 ml HCl). Scheil simulation of phases formed during
weld solidification was conducted using Thermocalc™ software. Scheil solidification modeling can also
predict the solute redistribution during alloy solidification (9).

Table 1 Base metal compositions (wt. pct.)

Fe Cr Ni Mn Si C S P Cu
304 Stainless 71.42 18.12 8.17 1.20 0.42 0.01 0.004 0.02 0.17
ER308 weld rod bal. 19.5- 9.0- 1.0- 0.30- 0.03 0.03 0.03 0.75
(filler) 22.0 11.0 2.5 0.65 max max max max
Si-Bronze 0.03 - 0.01 0.83 3.01 0.02 0.001 0.01 95.84
ERCuSi-A weld rod 0.5 - - 1.5 2.8- - - - bal.
(filler) max max 4.0
Table 2 Parameters for the two-pass Si-bronze filler weld and the stainless steel stainless steel filler weld
Filler Rod Material Si-Bronze Stainless Steel
(Two Passes) (One Pass)
Welding Passes Pre-heat Add filler Add filler
Voltage (V) 14.0 13.6 14
Current (A) 175 138 160
Speed (mm/s) 5.6 4.4 3.4
Power (kJ/s) 2.5 1.9 2.3
Heat Input (kJ/mm) 0.43 0.42 0.67

Transverse tensile specimens with a 20 mm (0.8 in) gage length were extracted using electrical discharge
machining (EDM) across the welds as shown in the schematic in Figure 1(a). The tensile bars were
machined near the crown of the weld such that the cross-section of the tensile bar was entirely within
the weld deposit layer. Although micro-cracking was identified in the stainless steel filler welds, it was of
industrial interest to identify the resulting joint strength. The geometry of the specimens is illustrated in
Figure 1(b). The fusion zone was in the middle of each tensile bar with a 1 mm (parallel to normal
direction) by 2 mm (parallel to welding direction) cross-section area. The gage length included 5-8 mm
of base metal on each side of the weld. Digital image correlation (DIC) method was used to map the
strain distribution during tensile loading. Fracture surfaces of the dissimilar welds were examined using
SEM and EDS.

Figure 1 Tensile bars: (a) machined transverse to the weld from entirely joined weld section; (b) tensile
bar dimensions.

3 Results

3.1 Base metals

Figure 2 shows the starting microstructures of the (a) base 304 stainless steel and (b) Si-bronze plate
materials (etched with Glyceregia). Both the stainless steel and Si-bronze plates were wrought and
annealed, having an equiaxed microstructure with annealing twins. The annealing twin boundaries were
straight lines, indicating little or no cold work had been applied to materials after annealing. The 304
stainless steel was austenitic with a small fraction of ferrite stringers. Some areas of the 304 stainless
steel microstructure appeared dark due to formation of surface martensite from sample preparation
(10).

Figure 2 Base metal microstructural features; (a) stainless steel and (b) Si-bronze (Glyceregia etch).

3.2 Si-bronze filler: microstructure and hardness

For the weld with Si-bronze filler rod, first, a heating pass was used, wherein the heat input was
0.43 kJ/mm, with no filler rod added. Then, in the second pass, Si-bronze filler rod was added with a
heat input of 0.42 kJ/mm, resulting in melting and mixing of the stainless steel base metal with the Si-
bronze. The microstructure of the two-pass GTA weld with the Si-bronze filler is presented in Figure 3.
Both base metals and the fusion zone are shown in Figure 3(a). The coarse-grained solidification
structure of the bronze weld is visible near the crown of the weld and partial mixing (mottled Cu-based
and Fe-based regions) was evident around the weld root. Figure 3(b) shows an area of mixed Cu-based
matrix with dark Fe-based islands and an area of mixed Fe-based matrix adjacent to the stainless steel
base metal. Figure 3(c) shows an area of mixed Fe-based matrix area with Cu-based islands as-well-as a
region of re-melted base bronze where minimal mixing occurred. The sample was analyzed using SEM
and EDS prior to etching. The SEM image from the location labeled “SEM” is provided in Figure 4(a).

The SEM image in Figure 4(a), near the weld root, shows the mixed zones near the root of the weld with
Si-bronze filler. Regions of mixed Fe-based matrix and mixed Cu-based matrix are identified. Figure 4(b)
is an SEM image at greater magnification near the weld root to clarify the important features. Firstly,
there is a region of darkly shaded Fe-based matrix with light Cu-based islands in the upper right. This
region was representative of all of the Fe-based matrix regions with Cu-based islands in the fusion zone.
Secondly, there is a region of lightly shaded Cu-based matrix with dark Fe-based islands. This region was
representative of all of the Cu-based matrix regions in the fusion zone. Thirdly, there is a dark band of Si-
rich Fe in the base stainless steel in contact with the Cu-based matrix region. The shading of the band of
Si-rich Fe band matches that of the ferrites in the base stainless steel.

Figure 3 Dissimilar GTA weld of Si-bronze to 304 stainless steel using Si-bronze as the filler: (a) fusion
zone; (b) interface between fusion zone and base stainless steel; (c) interface between fusion zone and
Si-bronze (glyceregia etch)

Figure 4 SEM images for (a) the area highlighted by the blue frame in Figure 3, and (b) enlarged view of
the root region (no etch).

EDS scan results for representative features from the Si-bronze filler weld are provided in Table 3. The
composition totals do not add to 100 wt. pct. because Mn and minor alloying elements are omitted. As
listed in the table, several compositions were the results of averages of EDS scans at multiple locations
for similar features. The standard deviation of the averages was generally +1 wt. pct. Compositional EDS
scans for the base metals matched well with the base metal compositions in Table 1. The ferrites in the
base stainless steel had higher Cr and lower Ni (22Cr/6Ni) than the austenite (18Cr/8Ni), as expected
(11). The dark layer of the base stainless steel near the root exhibited elevated Si and Cu (7Si/6Cu).
Within the fusion zone, the Fe-based matrix also had elevated Si and Cu (6Si/7Cu). A representative Cu-
based island within the Fe-based matrix region contained some of the components of the stainless steel,
but was depleted of Si (1Si). The Cu-based matrix also contained some components of the stainless
steel, but was depleted of Si (1Si). The Fe-based islands within the Cu-based matrix exhibited elevated Si
and Cu (7Si/10Cu). The re-melted base bronze was also slightly depleted of Si (2Si). In general, note that
Si content was elevated in the mixed Fe-based—and diminished in the Cu-based—regions of the fusion
zone. These EDS scans provide evidence that the element, Si, was mobile in the mixed region of the
weld.

Table 3 EDS scan results for bronze filler weld (wt. pct.)
EDS scan location description Fe Cr Ni Si Cu
Base 304 stainless steel (austenite) 72 18 8 - -
Base 304 stainless steel (ferrite stringer) 70 22 6 - -
Fe-based, Si-enriched layer adjacent to fusion zone 66 16 5 7 6
Fusion zone: Fe-based matrix (4 locations) 64 16 6 6 7
Fusion zone: Cu-based island in Fe-based matrix 14 4 3 1 78
Fusion zone: Cu-based matrix (2 locations) 4 1 2 1 90
Fusion zone: Fe-based island in Cu-based matrix (4 locations) 62 14 6 7 10
Re-melted base Si bronze 4 1 2 2 90
Base Si-bronze - - - 3 96

Three hardness line scans were conducted for the dissimilar weld made with Si-bronze filler rod, and the
data are presented in Figure 5. There was no detectable weld heat softened zone in either the 304
stainless steel base material or the Si-bronze base material since both of the base materials were fully
annealed, and thus relatively soft prior to welding. The base 304 stainless steel hardness was ~200 HV,
and the base Si-bronze hardness was ~85 HV. Interestingly, Vickers microhardness readings of the fusion
zone showed some regions of elevated hardness. Some regions of the fusion zone with Fe-based matrix
exhibited readings >400 HV. Also, the fusion zone Cu-matrix region hardness was consistently ~135 HV,
which was higher than the base bronze hardness of ~85 HV.

3.3 Stainless steel filler: microstructure and hardness

For the weld with the stainless steel filler rod, the heat input was 0.67 kJ/mm. Both base metals melted
and mixed with the filler rod in the fusion zone using a single pass. Several LOM views of the GTA weld
made with stainless steel filler rod are presented in Figure 6. A Fe-based island ~200 μm in size was
observed in the re-melted bronze in Figure 6(a). A void (~0.5 by 1 mm) was observed at the root of the
weld between the two workpieces as shown in Figure 6(a). Small fissures, or cracks, were observed near
the center of the weld, as identified in Figure 6(a) and near the weld toe on the base stainless side as
shown in Figure 6(b). Several solidification morphologies were identified in the stainless steel in Figure
6(c), showing the progression from planar growth to cellular growth to dendritic growth (12). Figure 6(c)
also outlines the weld root region that was subsequently examined using EDS mapping in Figure 7.

Figure 5 Microhardness for the dissimilar GTA weld made with Si-bronze welding rod.

Figure 6 Dissimilar GTA weld of 304 stainless steel to Si-bronze using stainless steel filler rod: (a) fusion
zone and base alloys; (b) weld toe on the stainless steel side; (c) weld root (glyceregia etch).

Figure 7(a) shows an SEM image of the lower left portion of Figure 6(c), where molten Si-bronze was in
contact with both the Fe-based fusion zone (above) and the base stainless steel (below). Figure 7(d) is an
LOM image of the same location showing the Cu with its natural tone in contrast to the pale stainless
steel. Figure 7(b) is a map of Si concentration, revealing that (i) there was no Si in the unaffected base
stainless steel, (ii) there was a thin ~5 μm layer of stainless steel with elevated Si, (iii) the Cu-based
region was depleted of Si, and (iv) the Fe-based fusion zone also had elevated Si. Apparently, Si
partitioned from the Si-bronze into the stainless steel along the interfaces. The Fe-based regions are
highlighted by Figure 7(c), which correlated closely with the locations of elevated Cr revealed in Figure
7(f). Finally, Figure 7(e) provides a map of Cu distribution. There was no Cu in the unaffected base
stainless steel, but there was a slight indication of Cu in the Fe-based fusion zone.

Figure 7 Root region from Figure 6(c) of dissimilar weld with stainless filler including: (a) SEM image; (b)
EDS map for Si (c) EDS map for Fe; (d) LOM image; (e) EDS map for Cu; and (f) EDS map for Cr.

Figure 8 presents the Vickers microhardness readings across the dissimilar GTA weld made with stainless
steel welding rod. The base 304 stainless steel hardness was ~200 HV, and the base Si-bronze hardness
was ~85 HV. The Fe-based fusion zone exhibited slightly lower hardness (150-185 HV) than the base
stainless steel. Near the interface between the stainless filler region and the base bronze (labeled “Si-
enriched stainless layer”), however, the Fe-based hardness was slightly higher (~200 HV). Also, regions
of mixed Cu-matrix in the fusion zone were slightly higher in hardness than the base bronze.

Figure 8 Vickers microhardness variation along the dissimilar GTA weld with stainless steel weld rod.

3.4 Tensile behavior

The tensile testing results from both the base materials and the dissimilar welds are displayed in Figure
9. Three 304 stainless steel base metal tensile bars were tested. Tensile strength and elongation
exceeded 600 MPa and 43%, respectively, as shown in Figure 9 curves group (a). Two Si-bronze base
metal tensile bars were pulled, and the tensile strength and elongation exceeded 340 MPa and 55%,
respectively, as shown in Figure 9 curves group (c).

In the dissimilar weld cases, the bronze filler rod welds in Figure 9 curves group (d) had higher
elongation than the stainless steel filler rod welds in two out of three tests. Maxima of 32% and 48%
elongation were achieved in dissimilar welds using stainless steel and Si-bronze rods, respectively. One
photograph of a tested tensile bar for a weld made with bronze filler metal is shown for Figure 9 curves
group (d). The tensile bar failed in the Si-bronze base metal. The bronze color in the stainless steel
portion (left side) of the tensile bar reveals the mixing of the Si-bronze into the stainless steel.

The stainless steel filler rod welds in Figure 9 curves group (b) had higher strength than the bronze filler
rod welds in two out of three tests. All three of the stainless steel filler rod weld tensile bars failed in or
near the weld. During the tensile tests, cracks formed in the stainless filler rod samples. Figure 9(b)
shows two example photographs from stainless steel filler welds revealing expanded fissures in the
tensile bars. The two stainless filler samples that failed in the weld with 3% and 15% elongation
evidently failed due to the linkage of these micro-fissures.

Figure 9 Engineering stress-strain plots for base metals and dissimilar welds of Si-bronze to 304 stainless
steel: (a) base 304 stainless steel; (b) stainless filler weld; (c) base Si-bronze; and (d) Si-bronze filler weld.
3.5 Tensile bar fractography

Fractography was performed on the transverse tensile weld specimens. For the weld made with Si-
bronze filler, Figure 10 shows SEM images of the fracture surface of the tensile bar that exhibited 48%
elongation. The fracture occurred in the Si-bronze base metal, indicating adequate bonding in the fusion
zone and absence of a heat softened zone near the weld. As shown in Figure 10(a) the fracture face is
smaller than the original cross-sectional area because was significant reduction of area (high post-
uniform ductility). Figure 10(b) shows evidence of microvoid coalescence (ductile dimples), resulting in
ductile rupture.

For the weld made with stainless steel filler, Figure 11 is a photograph of the fracture surface for the
tensile bar failed at 3% elongation. A Cu film was evidenced by a reddish hue on much of the fracture
surface. SEM images of the opposing fracture face are shown in Figure 12. A variety of fracture features
were observed. At the top of the tensile bar in Figure 12(a), the columnar solidification structure was
visible, indicating that the fracture propagated along the Cu film between columnar stainless steel
grains. At the bottom of the tensile bar, the fracture appearance consisted of elongated layers. Figure
12(b) shows that the fracture evidently followed a continuous Cu film along some of the austenite
dendrites in the columnar dendritic region. Figure 12(c) reveals a location where the original machined
surface of the base stainless steel was visible, indicating that fracture had occurred by delamination.
This fracture surface morphology was due to interfacial separation between the Cu and the stainless
steel because the stainless steel did not melt and mix in the bottom area of the weld.

Figure 10 SEM images of tensile bar fracture surface from dissimilar weld using Si-bronze filler rod that
failed with 48% elongation showing (a) necking with significant reduction of area and (b) microvoid
coalescence.

Figure 11 Photograph of tensile bar fracture surface showing Cu film (reddish hue) for dissimilar GTA
weld with stainless steel filler rod, which failed after 3% elongation.

The results for the EDS locations in Figure 12(b) and (c) are given in Table 4 (note that the values do not
add to 100 pct. because not all elements are included). The composition for the location of a columnar
grain, EDS1, was similar to the stainless steel composition along with some Cu and Si. The Cu and Si was
due to the presence of a thin film over the columnar grain, such that the interaction volume was
primarily the underlying stainless steel. For EDS2, however, Cu was identified as the primary constituent.
The Cu layer was thicker in this location, resulting in a higher Cu peak. Si was slightly higher (1.7%) for
EDS1, a Fe-based region, than for EDS2 (1.1% Si), a Cu-based region. EDS3 was a location of
delamination, revealing mostly Fe and the components of the stainless steel. Elevated Si (6.3%) was
prominent in this Fe-based location. EDS4 was a location with ductile dimples, and was mostly Cu with
some of the stainless steel components. Note that the Si (1.5%) was lower than that of the base bronze
in Table 1.
Figure 12 SEM images of tensile bar fracture surface for dissimilar GTA weld made with stainless steel
welding rod, which failed after 3% elongation: (a) SEM overview of fracture regions showing
intergranular fracture of the columnar grains near the top of the weld and lack of fusion near the bottom
of the weld, (b) Cu film on intergranular fracture surface, and (c) regions of delamination (woody
appearance) and ductile dimples.

Table 4 EDS scan results for stainless filler weld fracture (wt. pct.)
Location Fe Cr Ni Si Cu
EDS1 columnar grain 58 19 9 1.7 9
EDS2 Cu-based film 26 9 7 1.1 54
EDS3 stainless delamination location 63 18 7 6.3 4
EDS4 ductile fracture location 7 2 2 1.5 87

4 Discussion

4.1 Bronze filler weld

For the dissimilar weld with made with Si-bronze filler rod shown in Figure 3, there was good wetting at
the weld toes due to a low wetting angle between liquid Si-bronze and both base metals (13). The
thermal energy and balance of forces present during GTA welding caused dilution and mixing of the
components of the two base alloys in the fusion zone (14). Partial mixing, as evidenced by the Fe-based
matrix regions in Figure 3(b) and (c), occurred due to the high melting temperature and sluggish mixing
of the stainless steel (2). Segregation during solidification resulted in Fe-based matrix regions with Cu-
based islands and Cu-based matrix regions with Fe-based islands as apparent in Figure 4. The Fe-based
regions solidified first due to their higher solidification temperature range (2) and much of the Cu and Si
were rejected toward the liquid due to non-equilibrium solidification. The segregated, globular
microstructures formed in the fusion zone because of the low solid solubility of the two elements (15)
and spinodal decomposition due to the Fe-Cu liquid miscibility gap (16). This microstructure was similar
to that produced by electron beam welding of Cu to stainless steel by Magnabosco et al. (3), except that
the Fe-based islands identified by Magnabosco et al. remained austenitic upon cooling.

The EDS results revealed Si enriched Fe-based features, as quantified by Table 3. These Si-rich Fe-based
features existed in the fusion zone as-well-as in an interfacial band of the stainless steel base metal as
identified in Figure 4, Figure 6, and Figure 7. The Si-enrichment of the stainless steel occurred during the
cooling and non-equilibrium solidification of the liquid bronze. Diffusion occurs down gradients in
chemical potential until equilibrium is achieved. In liquid Cu, the chemical potential for 4% Si at 1200 °C
was calculated (using thermal simulation software) to be -50 kJ/mol, and for 1% Si in Cu, -125 kJ/mol. In
austenite (Fe/18Cr/8Ni), the chemical potential at 1200 °C for 1% Si in was calculated to be -200 kJ/mol
and for 6.5% Si, -125 kJ/mol. Thus, the Si diffused from the Cu-based to the Fe-based regions until
chemical potential equilibrium was reached (at -125kJ/mol) at the interface at ~1% Si in the Cu-based
regions and ~6.5% Si in the Fe-based regions. The 6-7% Si enrichment value in a band of the stainless
steel in this study is similar to the 6% Si diffusion layer in 304 stainless steel obtained by pack-
cementation method (17). In this study, heating and cooling rates during welding were likely rapid, >100
°C/s for heating and 10-100 °C/s for cooling (18). The mobility of Si is greater than that of most other
substitutional atoms in Fe (19). Si is a ferrite stabilizer in stainless steel (20), and has a max solubility of
~12% in ferrite (15). Thus the Si-enriched Fe-based regions remained ferritic upon cooling.

For the welds made with Si-bronze filler, the hardness increase in the Cu-based matrix regions was due
to composite strengthening effect of the 5-50 µm size Fe-based islands. The Si-enriched, Fe-based
features in this study exhibited hardness readings of up to 400 HV, which is a hardness level comparable
to Fe with 6-12% Si (21). Tensile strength for the Si-bronze filler weld exceeded that reported for most
stainless steel to Cu-alloy welds in the literature. Specifically, the Si-bronze filler welds in this study had
higher strength and ductility than SMA welds using Sn-bronze (93Cu7Sn) and Ni-based (68Ni/21Cr/6Mn)
filler (5); SMA welds using Inconel (ENiCrMo-3), Monel (65Ni/30Cu), and E316L filler; (6) and GTA welds
using Cu filler (7). The Si-bronze base material in this study, though, had a higher strength than the Cu
base material used by some researchers (22). Nonetheless, the results of this study indicate that Si-
bronze filler rod is a good selection for dissimilar GTA welding of Si-bronze to 304 stainless steel when
accompanied by a pre-heating pass.

4.2 Stainless steel filler weld

For the dissimilar weld made with stainless steel filler shown in Figure 6, heat transfer from the liquid
steel was evidently sufficient to cause melting and re-solidification of the base bronze even after the
steel solidified, due to the much lower melting point of the Si-bronze. The solidification morphology in
the Fe-based region in Figure 6(c) progressed from planar to cellular to dendritic morphology due to a
decrease in the temperature gradient (23). Gross solidification cracking did not occur at the dissimilar
interface with the bronze base metal because the temperature gradient was quite high due to the high
thermal conductivity of the bronze. For comparison, the thermal conductivity of austenitic stainless steel
is ~25 J/(m·s·K) versus ~384 J/(m·s·K) for Cu (14). Rapid solidification precludes compositional
segregation in the liquid and therefore solidification cracking as well (24). Also, solidification cracking
often occurs in the last liquid to solidify, not at the onset of solidification (25).

The Fe-based island in the re-melted bronze in Figure 6(a) solidified prior to the surrounding Cu-based
metal (2). The Fe-based island solidified in a somewhat spherical shape to minimize surface energy in
contact with the liquid bronze (26). The density of solid 304 stainless steel (7.9 g/cm3) is typically lower
than that of solid Si-bronze (8.8 g/cm3)(27). However, the density of liquid bronze is 7.9 g/cm3.
Therefore the island did not float upward, but remained slightly below the stainless steel to bronze
interface (22).

The presence of the void in Figure 6(a) was due to the (i) melting temperature difference (2), (ii) the high
thermal contraction of the stainless steel (28), and (iii) the excellent wetting of Cu for stainless steel (29).
The molten stainless steel transferred heat into the bronze, which easily melted. Then surface tension
and excellent wetting drew the liquid bronze into the crevices and interdendritic spaces in the solidifying
austenite as will be discussed subsequently along with solidification cracking.

For the welds made with stainless steel filler, the hardness in Figure 8 decreased slightly from the base
stainless steel to the fusion zone due to the soft Cu-based islands. In the narrow (~5 μm) regions of
stainless steel adjacent to the base bronze, the hardness may have been slightly higher than the fusion
zone due to the previously described Si strengthening effect. In tensile testing, fissures formed from the
top (Figure 12a) of the test bars (corresponding to a location near the top of the weld) during tensile
straining. Metallography (Figure 6) and fractography indicated that these cracks propagated from
intergranular solidification cracks, which will be discussed subsequently. Weld penetration was shallow,
and near the bottom of the welds (Figure 12(a) and (c)) the type of bonding was similar to brazing,
wherein the base Si-bronze melted and bonded to the solid stainless steel surface (30). During cooling, Si
evidently diffused from the liquid Si-bronze into the stainless steel for the reasons discussed in Section
4.1. It is unclear whether this partitioning of the silicon from the bronze to the stainless steel was
detrimental to the overall bond strength. The fracture surface near the bottom of the weld was
characterized by alternating regions of interfacial delamination between the dissimilar metals and
ductile fracture in the Cu-based material.

4.3 Solidification cracking

Solidification cracking is a type of segregation cracking that occurs close to the solidus when only a small
volume of liquid remains (31). Solidification cracking occurred in the stainless steel filler weld due to a
crack-susceptible solidification structure and the liquid composition. Firstly, the addition of Cu in the
dissimilar fusion zone affected the primary solidification structure. Typically, 304 stainless steel is
considered weldable by the GTA process because the alloy solidifies as primary ferrite due to a
reasonably high Cr/Ni equivalency (CrEq/NiEq) (32). Ferrite has relatively high solubility for sulfur,
phosphorus, and silicon, thus precluding the formation of low melting compounds which cause
solidification cracking in the interdendritic regions (33). WRC-1992 states that the CrEq/NiEq ratio should
be >1.5 to avoid primary austenite solidification (34) and consequent solidification cracking (33). The
CrEq/NiEq ratio for the stainless steel for the present study was ~2.0, which should have been adequate to
avoid primary austenite solidification. However, mixing in the fusion zone affected the composition and
solidification structure. Scheil modeling showed that the addition of Cu to the stainless steel filler rod in
the fusion zone promoted the primary austenite solidification. The Scheil results in Figure 13 show that
for stainless steel (Fe/18%Cr/8%Ni), as Cu is increased from 0% to 2% the single phase BCC solidification
range decreases. As the Cu addition is further increased from 2% to 4%, the primary solidification phase
becomes FCC austenite. Note, however, that this Scheil prediction was based on a simplified stainless
steel composition with no O, C, S, P, or S. Nonetheless, Magnabosco also identified that the Fe-based
islands in Cu-based matrix solidified as austenite for electron beam welds of stainless steel to Cu (3).
Thus, the Cu enrichment in the stainless steel filler weld resulted in a crack-susceptible primary
austenite solidification mechanism.

Figure 13 ThermocalcTM Scheil solidification models indicating the effect of Cu on stainless steel
(Fe/18Cr/8Ni) solidification mode. (a) With no Cu, the stainless steel solidifies primarily as body-centered
cubic (BCC) ferrite, then two-phase solidification commences with mutual growth of ferrite and face-
centered cubic (FCC) austenite. (b) With 2% Cu, the BCC solidification mode is reduced to 25% of the
fraction of the solid. (c) With 4% Cu, FCC austenite is the primary solidification mode.

Regarding the intergranular Cu film for the stainless filler welds in this study, Noecker and DuPont
showed that for GTA weld overlays on steel, Cu contents of 5-52% led to cracking (16). Within 5-52%,
the Cu content was sufficient to result in terminal Cu upon solidification, which led to cracking. Above
52%, there was adequate terminal Cu to backfill solidification cracks (16). Similarly, in this study, for the
Si-bronze filler weld there was adequate Cu to backfill any solidification cracks that formed in the
stainless steel. However, for the stainless steel filler weld, the small amount of terminal liquid copper in
the intergranular regions led to crack formation during cooling. Thus Si-bronze welding rod is a superior
choice to ER308 stainless steel welding rod for GTA welding of Si-bronze to 304 stainless steel.

5 Conclusions

GTA welds were made between the Si-bronze and 304 stainless steel using two filler rods similar to the
base metals. For the Si-bronze filler weld, a two pass welding process was necessary to facilitate mixing
and melting. For the stainless steel filler weld, a single welding pass was sufficient for melting and
mixing of both base metals.

For the Si-bronze filler weld, the fusion zone consisted of Cu-based regions with Fe-based islands and Fe-
based regions with Cu-based islands due to low solid solubility and liquid spinodal decomposition. Upon
cooling, Si partitioned from the Si-bronze to the Fe-based regions, stabilizing ferrite with elevated
hardness. The hardness in the Cu-based fusion zone was higher than the base Si-bronze hardness
because of the dispersed Fe-based islands. In tensile tests, ductile fracture occurred away from the
fusion zone, indicating the formation of a strong metallurgical bond. The strength and ductility exceeded
those reported for most Cu-based to Fe-based weld joints in the literature.

The stainless steel filler weld exhibited fissuring during solidification and lack of bonding at the weld
root. The micro-fissures propagated during tensile testing and led to brittle failure in two out of three
tests. Delamination near the bottom of the weld tensile bar was due to poor weld penetration. A crack
susceptible Cu concentration in the weld led to primary austenite solidification, entrainment of a Cu film
at the columnar grain boundaries, and subsequent solidification cracking.

Acknowledgements

Tony Workman from Art Castings of Colorado, along with his excellent team of welders and technicians,
is to be thanked for producing the weld samples. We are grateful to Dr. Aaron Stebner and Garrison
Hommer for assisting with the tensile testing. The authors would also like to acknowledge Dr. Stephen
Tate’s help on the composition analyses. This research did not receive any specific grant from funding
agencies in the public, commercial, or not-for-profit sectors.

References

1. “The Uses of Bronze Rod,” 2012. [Online]. Available: flocast.wordpress.com/2012/09/01/the-

uses-of-bronze-rod. [Accessed: 01-Jan-2017].
2. American Welding Society, Brazing Handbook, 5th Edition. 2007.
3. I. Magnabosco, P. Ferro, F. Bonollo, and L. Arnberg, “An investigation of fusion zone
microstructures in electron beam welding of copper–stainless steel,” Mater. Sci. Eng. A, vol. 424,
no. 1–2, pp. 163–173, May 2006.
4. K. Ramkumar, N. Arivazhagan, S. Narayanan, K. Devendranath Ramkumar, N. Arivazhagan, and S.
Narayanan, “Effect of filler materials on the performance of gas tungsten arc welded AISI 304 and
Monel 400,” Mater. Des., vol. 40, no. 9, pp. 70–79, Sep. 2012.
5. M. Velu and S. Bhat, “Metallurgical and mechanical examinations of steel–copper joints arc
welded using bronze and nickel-base superalloy filler materials,” Mater. Des., vol. 47, pp. 793–
809, May 2013.
6. C. Roy et al., “Characterization of Metallurgical and Mechanical Properties of Commercially Pure
 Copper and AISI 304 Dissimilar Weldments,” Procedia Mater. Sci., vol. 5, pp. 2503–2512, 2014.
7. S. G. Shiri, M. Nazarzadeh, M. Sharifitabar, and M. S. Afarani, “Gas tungsten arc welding of CP-
copper to 304 stainless steel using different filler materials,” Trans. Nonferrous Met. Soc. China,
vol. 22, no. 12, pp. 2937–2942, Dec. 2012.
8. C. Powell and P. Webster, “Copper alloys for marine environments,” p. 33, 2011.
9. G. Krauss, Steels Processing, Structure, and Performance. Materials Park, OH: ASM International,
2005.
10. J. Rodelas, M. Maguire, and J. Michael, “Deformation-Induced Martensite in Austenitic Stainless
Steel Welds,” in Fabtech, 2014, pp. 9–10.
11. D. Olson, “Prediction of austenitic weld metal microstructure and properties,” Weld. J., vol. 74,
no. 10, pp. 281–295, 1985.
12. J. Elmer, S. Allen, and T. Eagar, “Microstructural development during solidification of stainless
steel alloys,” Metall. Trans. A, vol. 20, no. 10, pp. 2117–2131, 1989.
13. G. F. Deyev, Surface Phenomena in Fusion Welding Processes. CRC Press, 2005.
14. S. Kou, Welding Metallurgy, 2nd ed. Hoboken, NJ: John Wiley & Sons, 2003.
15. L. Swartzendruber, “Cu-Fe,” in ASM Handbook Vol 3: Alloy Phase Diagrams, H. Baker, Ed.
Materials Park, OH: ASM International, 1992, p. 2.168.
16. F. F. Noecker and J. N. DuPont, “Microstructural development and solidification cracking
susceptibility of Cu deposits on steel: Part I,” J. Mater. Sci., vol. 42, pp. 495–509, 2007.
17. M. Fukumoto, C. Tachikawame, Y. Matsuzaka, and M. Hara, “Formation of Si diffusion layer on
stainless steels and their high temperature corrosion resistance in molten salt,” Corros. Sci., vol.
56, pp. 105–113, 2012.
18. K. Poorhaydari, B. M. Patchett, and D. G. Ivey, “Estimation of Cooling Rate in the Welding of
Plates with Intermediate Thickness,” Weld. J., vol. 84, no. 10, pp. 149–155, 2005.
19. L. Li and R. W. Messler, “Segregation of phosphorus and sulfur in heat-affected zone hot cracking
of type 308 stainless steel,” Riv. Ital. della Saldatura(Italy), no. 5, pp. 78–84, 2002.
20. N. Baldwin and D. Ivey, “Iron Silicide Formation in Bulk Iron-Silicon Diffusion Couples,” J. Phase
Equilibria, vol. 16, no. 4, pp. 300–307, 1995.
21. J. Shin et al., “Ordering-disordering phenomena and micro-hardness characteristics of B2 phase
in Fe-(5-6.5%)Si alloys,” Mater. Sci. Eng. A, vol. 407, no. 1–2, pp. 282–290, 2005.
22. ASM Handbook Vol 2: Properties and Selection: Nonferrous Alloys and Special Purpose Materials.
Materials Park, OH: ASM International, 1990.
23. J. W. Fu, Y. S. Yang, J. J. Guo, and W. H. Tong, “Effect of cooling rate on solidification
microstructures in AISI 304 stainless steel,” Mater. Sci. Technol., vol. 24, no. 8, pp. 941–944, Aug.
2008.
24. M. Aziz, “Model for solute redistribution during rapid solidification,” J. Appl. Phys., vol. 53, no. 2,
p. 1158, 1982.
25. P. L. Sturgill, R. D. Campbell, and J. L. Henningsen, “Microstructural and solidification cracking
evaluation of electron beam welds in 304L,” 1991.
26. D. Porter, K. Easterling, and M. Sherif, Phase Transformations in Metals and Alloys. CRC Press,
 2009.
27. Metals Handbook Desk Edition. Materials Park, OH: ASM International, 1998.
28. J. Elmer, D. Olson, and D. Matlock, “The Thermal Expansion Characteristics of Stainless Steel
Weld Metal,” Weld. J., no. 9, pp. 293–301, 1982.
29. G. F. Deev and A. N. Goncharov, “Embrittlement Mechanism of Copper-Steel Welded Joints,” Fiz.
Mekhanika Mater., no. 1, pp. 68–71, 1993.
30. “An Introduction to Brazing: Fundamentals, Materials, and Processing,” Westbury, NY, 2011.
31. V. Kujanpaa, S. David, and C. White, “Formation of Hot Cracks in Austenitic Stainless Steel Welds
— Solidification Cracking,” Weld. Res. Suppl., no. Aug, pp. 203–212, 1986.
32. J. Elmer, J. Wong, and T. Ressler, “Direct Observation of Phase Transformations in Austenitic
Stainless Steel Welds Using In-Situ Spatially-Resolved and Time-Resolved X-ray Diffraction,” in
Sec. Int. Symp. on Joining of Advanced Mat., 1999, pp. 1–11.
33. V. Shankar, T. P. S. Gill, S. L. Mannan, and S. Sundaresan, “Solidification cracking in austenitic
stainless steel welds,” Sadhana, vol. 28, no. 8, pp. 359–382, 2003.
34. P. Korinko and S. Malene, “Considerations for the weldability of types 304L and 316L stainless
steel,” Pract. Fail. Anal., 2001.
Figure(s)
Figure(s)
Figure(s)
Figure(s)
Figure(s)
Figure(s)
Figure(s)
Figure(s)
Figure(s)
Figure(s)
Figure(s)
Figure(s)
Figure(s)