Professional Documents
Culture Documents
Microstructure/Property Relationships in
Dissimilar Welds between Duplex Stainless
Steels and Carbon Steels
The effect of weld metal microstructure on toughness and pitting corrosion
resistance is evaluated for both a duplex stainless steel and Ni-based filler metal
BY E. J. B A R N H O U S E A N D J. C. LIPPOLD
ABSTRACT. The metallurgical character- chanical properties and corrosion char- This study was designed to provide
istics, toughness and corrosion resistance acteristics relative to other stainless steels some insight into the microstructure/
of dissimilar welds between duplex stain- and structural steels. Although the joining property relationships in dissimilar fu-
less steel Alloy 2205 and carbon steel of duplex stainless steels to themselves sion welds with duplex stainless steels.
A36 have been evaluated. Both duplex has been studied extensively, the in- The dissimilar materials selected for the
stainless steel ER2209 and Ni-based creased application of these steels will re- overall study included a plain carbon
Alloy 625 filler metals were used to join quire a better understanding of the issues structural steel (A36), an austenitic stain-
this combination using a multipass, gas associated with welds to dissimilar met- less steel (Type 304L) and a martensitic
tungsten arc welding (GTAW) process. als. The joining of dissimilar materials is stainless steel (Type 410). This report fo-
Defect-free welds were made with each generally more challenging than that of cuses on the dissimilar combination of
filler metal. The tough hess of both the 625 similar materials because of differences in Alloy 2205 and A36. A future paper will
and 2209 deposits were acceptable, re- the physical, mechanical and metallurgi- report the results of studies conducted on
gardless of heat input. A narrow marten- cal properties of the base metals to be the other combinations.
sitic region with high hardness was ob- joined. These differences may also com-
served along the A36/2209 fusion plicate the selection of filler metals com- Stainless Steel/Carbon Steel Dissimilar Joints
boundary. A similar region was not ob- patible to both base metals. Therefore,
served in welds made with the 625 filler filler metal selection is often a compro- Early investigations on the joining of
metal. The corrosion resistance of the mise between the two dissimilar metals. dissimilar metals were primarily devoted
welds made with 2209 filler metal im- There are few guidelines for dissimilar to ferrous alloys; however, much of the
proved with increasing heat input, prob- metal joining and, in most cases, predict- emphasis was placed on the prevention of
ably due to higher levels of austenite and ing the microstructure and resultant prop- weld metal liquation cracking (often re-
reduced chromium nitride precipitation. erties of the weld deposit can be difficult. ferred to as microfissuring), heat-affected
Welds made with 625 exhibited severe at- zone cracking, carbon migration and
tack in the root pass, while the bulk of the oxide penetration, as discussed by Pattee,
weld was resistant. This investigation has etal. (Ref. 1). In the 1940s, Schaeffler pro-
shown that both filler metals can be used posed a diagram for the selection of elec-
to join carbon steel to duplex stainless KEY WORDS
trodes for the dissimilar joining of plain
steels, but that special precautions may be carbon and stainless steels that related the
necessary in corrosive environments. Duplex Stainless
Alloy 625 microstructural constitution of the weld
Dissimilar Metals deposit to its composition, as dictated by
Introduction
Corrosion Resistance the relative proportion of filler metals and
Constitutions Diagrams base metals (Ref. 2). This diagram (Fig. 1),
Duplex stainless steels have become
Ferrite Number commonly referred to as the Schaeffler
increasingly attractive to a number of in-
Filler Metal Constitution Diagram, can be used as a
dustry sectors due to their superior me-
Carbon Steel means of predicting the weld metal mi-
E+J. BARNHOUSE is with Weirton Steel Corp., crostructure of dissimilar metal welds in a
Weirton, W. Va. J. C. LIPPOLD is with The select group of alloys. By plotting the Cr-
Ohio State University, Columbus, Ohio. and Ni-equivalents for the materials on
i
24 ~k
~sF-
,3'~1
I I/li=~V,~
20 I > Y X/'Y//~,~ UII
cl ~ b -
o
K.~'~z2"/Y/)f/,/-/-/~,'~, ~ I
16 z ur
r.4,',~/~/xz. UI]
o ~I--- r/////.'~, III
12 __ V'C/";F/VK/V/~~I I I II
G ol /
~6
~4
0 4 8 12 16 20 24 28 32 36 40
0 i 6 ~ 9~0 ~2 14 ~5 1
Cr~ = Cr + Mo + 1.5Si + 0.5Nb Cr~ ;r • MO • 0,7 N b
Fig. 1 - - Schaeffler Diagram (Ref. 2) showing the dissimilar combina- Fig. 2 - - WRC- 1992 diagram, extended version (Ref. 5).
tions used in this investigation.
Table 1 - - Fusion Zone Mechanical Table 2 --Chemical Composition and Cr~ and Ni~ of the Base and Filler Metals (wt-%)
Properties of Dissimilar Joints with Alloy
2507 (Ref. 6) Element Alloy 2205 A36 ER2209 ERNiCrMo-3
Cr 22.16 22.71 21.84
2507/ Ni 5.64 8.23 64.46
Property Carbon Steel 2507/316 Mo 3.04 3.18 8.91
Tensile Strength, 61.8 85.1 Mn 1.46 0.627 1.64 0.02
ksi Si 0.48 0.236 0.52 0.03
Yield Strength, 43.4 54.5 C 0.014 0.088 0.015 0.030
ksi N 0.18 0.160
Elongation, % 23 45 S 0.001 0.025 0.001 < 0.001
P 0.028 0.005 0.020 0.003
Fe bal. bal. bal. bal.
the diagram and connecting the base Cr,,qc,'~ 25.20 25.89 33.15
metals by a tie line, the deposit mi- Ni~q~b) 9.73 m 11.98 65.51
Creq/Nieq 2.59 2.16 0.51
crostructure can be estimated by con- 1.66 0.27
Q factor .... 2.19
necting a point along that tie line (se- PRE~ 35.07 m 35.76 51.24
lected as the midpoint in Fig. 1) to a tie
line to the filler metal composition. The (a) Cr + M o + 0.7Nb, WRC-1992.
{b) Ni + 20N + 35C + 0.25Cu, WRC-1992.
weld metal constitution then lies along (c) (Cr + 1.5Mo + 2Mn + 0.25Si)/(2Ni + 12C + 12N).
the line between the filler metal and base (d) PRE (Pitting Resistance Equivalent) = Cr + 3.3(Mo + 0.5W) + 16N.
metal midpoint as dictated by the level of
dilution. Mn, and FN (Ferrite Number) predictions other constitution regimes, as in the
The ability to predict microstructure of highly alloyed compositions such as Schaeffler Diagram.
using the Schaeffler Diagram provided a 309 stainless steel were found to be in-
valuable tool for the selection of filler accurate (Ref. 4). Furthermore, its limited Duplex Stainless Steel to Carbon Steel
metals and the determination of the ef- composition range made it difficult to uti-
fects of base metal dilution. The diagram lize for dissimilar metal welding. Recently, Odegard, et al. (Ref. 6),
was particularly useful for predicting the Siewert, et al. (Ref. 4), developed a studied the joining of duplex Alloy SAF
ferrite content in austenitic stainless steel modified prediction diagram called the 2507 to carbon steels with respect to fu-
deposits and determining the constitu- WRC-1988 diagram. This diagram mod- sion zone mechanical properties. They
tion of dissimilar combinations of carbon ified and greatly simplified the Cr- and reported that the phase stability and the
steels and austenitic stainless steels. Ni-equivalent formulae and corrected overall properties of the fusion zone were
The Schaeffler Diagram does not have the overestimation of FN for higher al- influenced by the welding parameters
a specific weighting factor for nitrogen, loyed weld metals. Recently, Kotecki, et and that low heat inputs were necessary
however, and as a result, a diagram was al. (Ref. 5), had shown Cu to influence to ensure structural integrity and solidifi-
proposed by DeLong in 1974 that incor- the austenite formation and therefore cation cracking resistance. Furthermore,
porated N in the Ni-equivalent formula added a Cu factor in the Ni-equivalent they noted that a highly ferritic fusion
(Ref. 3). This diagram allowed for more formula. This change resulted in the zone resulted from high dilution by the
accurate estimation of ferrite content WRC-1992 diagram, which is essentially carbon steel, making the weld metal mi-
over a narrower composition range than identical to the WRC-1988 with the ad- crostructure susceptible to secondary
the Schaeffler Diagram, improving and dition of a Cu factor in Ni-equivalent for- austenite formation in multipass welding
correcting the limitations associated with mula. An extended version of this dia- due to reheating of the deposited weld
the Schaeffler Diagram. The DeLong Di- gram (Fig. 2) allows FN estimation in metal by subsequent passes. High heat
agram was later found to misrepresent dissimilar welds but does not contain input welding schedules were reported to
Experimental Procedure
1 2 3 4 S 6 7 8 9 10 11 12 13 14 iS 16 17 18 19 20 21 22 23 24 25 ~ 27 a 29 30 3T
C r R . C¢ + MO * O,;INb
Materials
Welding trials were designed to study Fig. 4 - - WRC- 1992 diagram for A 3 6 / 2 2 0 9 / 2 2 0 5 combination.
the dissimilar joining of duplex stainless
steel Alloy 2205 to A36 carbon steel. This
dissimilar combination was selected
based on a survey of industrial users. lected and used throughout the study, as ployed to provide better resolution be-
Both duplex stainless steel and Niobased listed in Table 3. The welding heat inputs tween the ferrite and austenite. These
filler metals were selected to join these used were 1.57 kJ/mm (39.8 kJ/in.) and techniques included a two-step modified
metals. The chemical compositions for 2.60 kJ/mm (66 kJ/in.), with the only ex- Murakami's etch (Ref. 7), which produces
the base and filler metals are listed in ception being the root passes in each com- a composition-sensitive film resulting in
Table 2. The base materials were sup- bination. Due to the difference in material a variety of colors on the ferrite phase,
plied in the form of 12.5-mm (0.5-in.) properties (mainly thermal conductivity) while the austenite remains white (un-
plate. These plates were cut into 5 x 20 the heat flow differed between the two colored). A ferro-fluid colloidal suspen-
cm (2 x 7.8 in.) coupons with a 37-deg filler metals, thus the root pass heat inputs sion of Fe]O 4 was also used to reveal the
bevel on each plate to provide a 74-deg were different. The 2205/2209/A36 com- duplex microstructure (Ref. 7). Residual
groove angle for a single-V-groove butt bination utilized a root pass heat input of magnetism attracts Fe304 particles to the
joint configuration. The root face was 1.6 1.05 kJ/mm (26.7 kJ/in.), while that of the ferrite phase to provide a color of brown
to 2.4 mm (0.062 to 0.094 in.) with a root 2205/625/A36 combination was 1.90 or dark blue in stark contrast to the
opening of 2.4 mm. kJ/mm (48.3 kJ/in.). austenite, which is white.
All welding was performed using the ODtical Metallo~raDhv Hardness testing was carried out on a
automatic gas tungsten arc welding Leco Model M-400 microhardness unit
(GTAW) process with cold wire feed of Weld cross sections were removed using a 1-kg load. Surveys were con-
1.14-ram (0.045-in.) diameter welding from each of the combinations and pol- ducted both across the weld and from
wire. ER2209 and Alloy 625 filler metals ished through 0.05-micron alumina. The root to cover pass within the weld de-
were used in the multipass welds to join samples were electrolytically etched in posit. The distance between indents was
duplex Alloy 2205 to A36 carbon steel. 10% oxalic acid at 5-7 V for 20-25 s. determined by the weld dimensions in
Two heat input levels were initially se- Also, color etching techniques were em- order to obtain appropriate data for each
A Z~
• &
•A
•~ 30
~ 20
u. 10
0
0 2 4 6 8 10 12 14 16
Through Thickness Distance from Root Pass (ram)
pass. Generally, the distance between in- Charpy Impact Toughness Testing mm, standard 25 x 50 mm (1 x 2 in.) spec-
dents was 0.5 mm (0.02 in.) in the lower imens could not be used, and samples of
passes of the weld, 1.0 mm (0.04 in.) in Charpy V-notch (CVN) specimens 12.5 x 33.0 mm (0.5 x 1.3 in.) were used.
the central portions of the weld metal and were machined from the welded The specimens were ground through 600-
2.0 mm (0.12 in.) in the cover pass of the coupons. Samples were prepared in the L- grit SiC and weighed prior to immersion
weld metal. Hardness was also measured T orientation as per ASTM E-23. The L-T in the solution. The pitting corrosion so-
along the fusion boundary and across the orientation represents a sample transverse lution was 6% FeCI3.6H20. The speci-
heat-affected zone (HAZ) into the base to the welding direction with the notch lo- mens were immersed in this solution for
metal. cated such that testing occurs through the 72 h and were removed and observed
Hardness testing was also conducted thickness of the weld from the root to the every 12 h to note any changes. Follow-
at reduced loads (25 and 50 g) along the cover passes. All notches were located in ing the 72-h test period, the samples were
2209/A36 and 625/A36 fusion bound- the center of the weld deposit. Charpy V- carefully scrubbed under running water
aries. This was done to more closely ex- notch testing was performed at tempera- and then ultrasonically cleaned in
amine the hardness in the microstruc- tures from -196°C (-320.8°F) to more methanol for approximately 30 min.
tural transition region between the weld than 100°C (212°F) in order to develop a Evaluation of pitting was determined
fusion zone and HAZ. complete ductile-to-brittle transition tem- by optical microscopy. The pits were
iiI~ ¸ perature (DBTT) curve for each combina- counted at magnifications of 50X and
I=h~ Ferrite Number Measurement tion. Scanning electron microscopy (SEM) measured at 400X. Pit density was deter-
was utilized to examine the fracture sur- mined by dividing the number of pits per
m
iliM~ The ferrite number (FN) in each weld faces of the CVN impact samples. unit area. A pit density was determined
layer was determined using a Magne- for each weld layer so a comparison
Gage calibrated per AWS A4.2-91 ac- Pitting Corrosion Testing could be made between the different base
u cording to the method developed by material combinations and the heat input
Kotecki (Ref. 8). All measurements were Pitting corrosion tests were performed variations.
obtained on longitudinal sections in the at 50°C (122°F) as described in ASTM G-
fusion zone rather than transverse sec- 48. Two types of samples were tested. The Results
tions in order to better quantify variation first type consisted of an entire weldment,
within a given pass. Four measurements including both base materials and the fu- 2205/2209/A36 Weldment
were taken at four different locations sion zone. The second type was an all-
within each pass. These measurements weld-metal sample machined out of the Microstructure
were then averaged to obtain the FN for fusion zone through the thickness. Since
each pass. the thickness of the material was 12.5 The dissimilar 2205/A36 combination
300 1
25O
2OO
'~°I .F ° 150
'°°I li,'~ tO0
50
°oT .,, o<,,,,, 0
-250 -200 -150 -100 -50 0 50 100
Temper~ure(C)
was welded with duplex filler metal nitrogen than does the
ER2209. Figure 3A shows the root pass ferrite phase. The other
microstructure of this combination re- major difference be-
200. PHOTO- 58
sulting from a heat input of 1.05 kJ/mm tween the two heat input
and consists of an austenite matrix levels was the level of
(white) with both skeletal and acicular secondary austenite (%2).
ferrite (dark etching). Ferrite measure- Both welds contained
ments indicated 20-25 FN in the root significant Y2 in the fill
pass for both welding heat inputs of 1.57 passes beneath the
KJ/mm and 2.60 KJ/mm. Based on this FN cover pass. This Y2 for-
level, the WRC-1992 diagram (Fig. 4) mation occurs due to re-
predicts that the dilution in the root pass heating of weld metal
was approximately 33%, which corre- during subsequent weld
sponds well with the 35-40% dilution passes. This reheating
that was estimated metallographically. and precipitation of Y2
Subsequent passes in these weldments alters the ferrite/austen-
exhibited increased FN from the root to ite volume fraction rela-
the cover pass, ultimately reaching 92 tive to that of the initial
FN. A micrograph of the cover pass is weld deposit.
shown in Fig. 3B. The FN transition from
the root to the cover pass is shown in Fig. Fusion Zone Hardness.
5. Note that the FN differential between
the root pass and cover pass is roughly 70 The resultant weld
FN and that a significant increase in FN deposit hardness of this Fig. 8 - - C V N fracture b e h a v i o r for 2 2 0 5 / 2 2 0 9 / A 3 6 w e l d m e n t : A
- - L o w e r shelf," B - - fracture transition range.
occurs between the last fill pass and the combination varied
cover pass for both heat inputs. with heat input. The
The filler metal composition, dilution lower heat input weld
and cooling rate are the principal factors exhibited higher hardness in the weld weld. Average hardness levels within and
that influence the weld metal metal and along the A36 fusion bound- at the fusion boundary of the
ferrite/austenite balance in duplex stain- ary than the high heat input weld. The A36/2209/2205 weldments are summa-
less steels. The lower heat input (1.57 hardness in the root pass varied from 261 rized in Table 4.
kJ/mm) weld passes undoubtedly resulted to 287 DPH. The remainder of the weld The 2.60-kJ/mm weld exhibited a
in higher cooling rates than the higher metal exhibited hardness levels between lower average hardness, with root pass
heat input (2.60 kJ/mm) weld. However, 229 and 279 DPH. An average hardness values from 249 to 256 DPH and an av-
as shown in Fig. 5, the difference in heat of 260 DPH was found in the cover pass. erage hardness of 250 DPH (Table 4). The
input and cooling rate did not signifi- Lower hardness was found in the center remainder of the weld metal exhibited
cantly affect the ferrite/austenite balance. portions of the fusion zone (midway hardness levels between 223 and 265
The major microstructural difference be- through the thickness), as average hard- DPH. An average hardness of 245 DPH
tween these welds was that the lower heat ness was 250 DPH. The hardness along was measured in the cover pass. Lower
input conditions resulted in the precipita- the fusion boundary of the carbon steel hardness was found in the center por-
tion of more chromium nitride, presum- was found to be much higher than the tions of the fusion zone (midway through
ably Cr2N, within the ferrite phase. This bulk fusion zone hardness and ranged the thickness), with an average hardness
precipitation was most pronounced in the from 253-416 DPH. Generally, the hard- of 230 DPH. The hardness along the
cover pass. The layers beneath the cover ness along the 2209/A36 fusion bound- 2209/A36 fusion boundary in this weld
pass exhibited greater volume fractions of ary was above 300 DPH, with average was also significantly lower than the
austenite, which would therefore reduce levels reaching 410 DPH near the root lower heat input weld. The fusion bound-
the tendency for Cr2N precipitation be- pass. Softer regions along this boundary ary region associated with the root pass
cause austenite has a higher solubility for were found to exist in the top third of the exhibited the highest hardness, 300
A
Weld Layer
1.00
-.><.....---
the 2209/A36 fusion boundary region is the 2205/2209/A36
shown in Fig. 6. This micrograph clearly weld combination at 0.00
shows an acicular martensitic structure, both heat input levels 0 1 3 4 5 6
which results in significantly greater are shown in Fig. 7. Weld Layer
hardness along the 2209/A36 fusion The lower shelf im-
boundary than along the 2205 fusion pact energy was B
boundary. The carbon steel average base roughly 7.5 ft-lb at
metal hardness (155 DPH) and the fusion temperatures below
Fig. 1 0 - - Pit density vs. weld layer for the 2205/2209/A36 pitting sam-
boundary hardness (250-410 DPH) tran- -140°C (-220°F).
ples: A - - Weldment samples; B - - all-weld-metal samples.
sitions rapidly within a short distance The upper shelf of the
from the fusion boundary. Type II grain DBTT curve was
boundaries (Ref. 13) are also present roughly 240 ft-lb
along this fusion boundary. Type II (325 J) for the 1.57
boundaries have been previously re- kJ/mm heat input and 250 ft-lb (339 J) for Fracture surfaces from upper and
ported by Wu and Patchett (Ref. 13) in the 2.60 KJ/mm heat input. These DBTT lower shelf and transition region speci-
nickel-based alloy cladding on Cr-Mo curves also show the transition impact en- mens were evaluated in the SEM. The
steels. Carbide precipitation decorates ergy, midpoint between the upper and upper shelf fracture surface exhibited
these boundaries resulting from carbon lower shelf, to be 125-130 ft-lb (169-176 ductile fracture, as exemplified by a dim-
migration out of the carbon steel into the J) and the corresponding transition tem- ple-type fracture morphology. The lower
more highly alloyed, but lower carbon, perature to be between -45 to -50°C ( 4 9 shelf exhibited a cleavage-type fracture
fusion zone. to -58°F) for both heat input levels. showing a relatively flat fracture surface
100. 140
12o
80 ¸
A
.@
100|
>. 60
E= -,so
-- 40
:,eo~
a.
I •2.60KJ/mm
i
_s ..o
2O
.20
0 1.57 KJ/ram ~ . . ~ ., ~ . ~.m,,=
0 +----------'-,~- --.0
-225 -175 -125 -75 @5 25
T e m p e r a t u r e (C)
Fig. 12 - - CVN impact toughness vs. temperature for the Fig. 13 --SEM fractograph of 2205/625/A36 CVN specimen produced
2205/625/A36 weld combination. at 1.57 kJ/mm and tested at-196°C.
Fig. 15 - - Pit density vs. weld layer for 2205/625/A36 combination: A - - Weldment sample; B - - all weld metal.
W E L D I N G RESEARCH SUPPLEMENT I 4 8 5 - s
weld deposits. Because of the L-T orien- Cr2N precipitation within the ferrite are similar, being roughly 110 ksi and
tation of the test samples, the toughness phase of the fusion zone. Another bene- 30-40% elongation. Therefore, a balance
values represent a composite of the en- ficial effect of lower cooling rates on pit- between toughness and corrosion resis-
tire weld deposit and are not indicative of ting resistance is the healing of tance will determine the use of either the
local variations due to microstructure. chromium-depleted regions around any duplex filler metal ER2209 or the nickel-
For example, lower toughness might be precipitates. Ume (Ref. 18) reported that based Alloy 625. The recommendation
expected in the higher FN cover pass. the number of initiation sites decrease for cryogenic temperatures would be to
It should be pointed out that the with higher heat inputs, thereby support- utilize Alloy 625 over the duplex filler
toughness of the 2205/2209/A36 may be ing the data collected in the current study metal due the increased toughness at
significantly affected by the formation of of the weldment samples. these temperatures. However, dilution
martensite along the A36 fusion bound- The all-weld-metal samples exhibited should be minimized in the root pass to
ary region. This narrow band of marten- increased pit density and depth for the limit the carbon steel dilution in the fu-
site would undoubtedly result in reduced higher heat input relative to the 1.57 sion zone to prevent the reduction in cor-
toughness along this particular region kJ/mm weld. However, the all-weld- rosion resistance. On the other hand, in
relative to the fusion zone. However, no metal corrosion data may be erroneous service environment temperatures greater
fusion boundary CVN testing was per- due to the collapse of the surface in many than -50°C, the duplex filler metal
formed to investigate any toughness of these samples, thereby changing the ER2209 should be utilized due to higher
anomaly in this region. kinetics of the corrosion testing relative toughness over the nickel-base Alloy 625.
As expected, the fully austenitic fu- to the weldment samples. The heat input should be as high as pos-
sion zone of the 2205/625/A36 combi- Table 8 compares the weight loss due sible in order to gain the maximum cor-
nation exhibited good impact toughness to corrosion of the all-weld-metal sam- rosion resistance of the fusion zone.
over a wide range of temperatures. This ples. The percentage weight loss illus-
microstructure does not exhibit trates that increased corrosion resistance Conclusions
upper/lower shelf transition behavior is obtained with increased heat input for
such as ferritic (or high ferrite) weld met- the 2205/2209/A36 combination. Srid- 1) The fusion zone microstructures of
als, and, instead, the impact toughness har, et al. (Ref. 19), showed that pitting dissimilar weld combination
exhibits a gradual decrease with temper- corrosion resistance, expressed as a per- 2205/2209/A36 resulted in a general in-
ature. It is interesting that the upper shelf centage of weight loss, increases with crease in ferrite number with each sub-
toughness of the duplex 2209 deposit far higher heat inputs. They claimed that sequent pass and a large increase in FN
exceeds that of the Alloy 625 deposit and slower cooling rates resulting in in- occurred between the last fill pass and
that the duplex filler material provides creased austenite and the distribution of the cover pass. The measured FN in the
comparable toughness at temperatures the various elements between the ferrite cover pass was greater than that pre-
down to -60°C. and austenite are the reasons for the in- dicted by the WRC-1992 diagram.
Table 6 compares fusion zone impact creased pitting resistance. 2) The ferrite number and hardness of
energies obtained in this study to previous The 2205/625/A36 combination re- the 2205/2209/A36 combination were
work from duplex stainless base/filler suited in the localized attack of the root similar for both the 1.57 and 2.60 kJ/mm
metal combinations performed by pass in each of the corrosion samples heat inputs, suggesting that heat input is
Kotecki (Ref. 15) and Bonnefois, et al. tested. The root pass of the all weld metal a secondary factor relative to weld metal
(Ref. 16). The impact toughness data re- corrosion samples was completely dis- composition in controlling the
ported here are higher than from this pre- solved by the ferric-chloride solution. ferrite/austenite phase balance.
vious work and may be attributed to This is apparently the result of higher di- 3) The 2205/2209/A36 combination
higher austenite contents (lower FN) in lution of the filler metal by the carbon formed a narrow martensitic band adja-
the root and fill passes of the dissimilar steel, as indicated by a variation in mi- cent to the A36 fusion boundary along
welds. Most importantly, these compar- crostructure relative to the subsequent the entire thickness of the weld. No
isons suggest that dilution from a dissim- corrosion-resistant fill passes. This sug- martensite was observed in welds made
ilar carbon steel base metal does not com- gests that a critical composition change with the Alloy 625 filler metal, possibly
promise the toughness of the fusion zone occurs between the root pass and the re- due to the smaller composition range
relative to a similar metal combination. maining fill and cover passes. over which martensite forms and lower
carbon migration rates in Ni-based vs.
CorrosionBehavior Summary and Recommendations Fe-based alloys.
4) The 2205/2209/A36 fusion zone
The two types of corrosion samples The 2205/A36 base metal combina- exhibited similar upper shelf energies for
(weldment and all weld metal) and their tion has been successfully joined with the two heat inputs utilized, suggesting
pitting corrosion data are summarized in duplex stainless steel ER2209 and Ni- that heat input and dilution from the dis-
Table 7. As shown for the 2205/2209/ based Alloy 625 filler metals using mul- similar base metal had little or no effect
A36 material combination, the weld- tipass GTAW. Heat input had only a on the mechanical properties.
ment sample shows that for increasing minor effect on the microstructure and 5) The 2205/625/A36 fusion zone
heat input, the pitting corrosion resis- toughness for the ER2209 combination. toughness behavior was typical of face-
tance increased when considering the pit However, the corrosion behavior showed centered cubic (FCC) materials as impact
density. The average pit depth, on the a marked improvement for higher heat energies decreased 10-15 ft-lb (14-20 J)
other hand, increased with increasing input welding parameters relative to the over a temperature range from 25 ° to
heat input and may be due to higher con- lower heat input. -196°C.
centration on fewer initiated pits. Yasuda The choice of filler metals in joining 6) Pitting corrosion resistance of the
(Ref. 17) and Ume (Ref. 18) claim that pit- 2205 to A36 will be primarily dependent weldment samples utilizing 2209 filler
ting corrosion decreases with increased on the service requirements needed. The metal showed increased resistance as the
heat input due to slower cooling rates strength and ductility of nickel-based heat input increased from 1.57 kJ/mm to
and the formation of austenite rather than Alloy 625 and duplex filler metal ER2209 2.60 kJ/mm. The all weld metal samples
The seminar subcommittee will inform you by May 1, 1999 about the acceptance of your paper.
The final paper has to be sent to the chairman by September 1, 1999 by mail, fax or email (bernie@weld.tu-graz.ac.at).
Bernhard Schaffernak
Institut fiJr Werkstoffkunde, Schweigtechnik
und Spanlose Formgebungsverfahren
A-8010 GRAZ, Kipernikusgasse 24
Tel +43-316/873-7182, Fax +43-316/8737187
email: bernie@weld.tu-graz.ac.at