GTA Weld Cracking - Alloy 625 to 304L

3041 rich welds and GTA W-P exacerbate an already

sensitive hot cracking problem

BY R. A. PATTERSON AND J. O. MILEWSKI

ABSTRACT. Autogenous gas tungsten arc input versus weld area graph —Fig. 1. which was observed as a qualitative
welds joining alloy 625 and 304L stainless Increasing the frequency of weld current decrease in both the number and length
steel were found to be susceptible to pulsation to nearly 10 Hz tended to of cracks within a 10 cm (3.9 in.) linear
weld solidification cracking. Utilization of decrease the cracking susceptibility weld. At high pulsation frequencies (~ 10
pulsed current GTA welding produced a
higher sensitivity to solidification cracks 304L Stainless Steel To Inconel 625
than continuous current welding. Spot
Varestraint tests show that the sensitivity
of this dissimilar metal combination to 30 -
cracking exists over the entire range of
dilutions while the greatest sensitivity is in oo
304L stainless steel rich compositions.
Auger electron spectroscopy indicates
that segregation of sulfur and phospho-
rous to the interdendritic phase promotes 25
the hot cracking.
u
Introduction C
'O
3
Dissimilar metal welding often C
produces difficulties which may not be a
obvious when evaluating the system
I 20 c
u
from a design viewpoint. One case in u
3 u
point is the design of nuclear fuel ele- O 0
ments for the TREAT reactor (Transient -3
I-
Reactor Test Facility operated by the 2
3 u
Argonne National Laboratory, Idaho Falls, a.
c
Idaho). A high temperature-oxidation £
resistant alloy (Alloy 625 —commercial- o
a
ly designated Inconel 625A Trademark of c
the Inco Family of Companies. The equiv- u
>
alent SAE/ASTM unified number is u
N06625.) was specified for the fuel rod C
cladding, while austenitic stainless steel -*-.
10 3
(type 304L) was used for the cladding end C
caps due to the lower exposure temper- a
<t
ature. Initially, this combination did not u
indicate a major concern from a weldabil- tf
• Pulsed
u
ity viewpoint. However, experimentation
revealed a significant susceptibility to o Continuous
a
solidification cracking for autogenous gas ^>
tungsten arc (GTA) welds. h
2
Figure 1 shows that over a relatively u,
broad range of weld heat input, pulsed
S
a>
current GTA welds joining Alloy 625 to u.
C
304L exhibit severe centerline cracking. C
-^
Table 1 lists the range of welding vari- X
ables used to generate the weld heat 5 10 15 20 25 30 35 40 45 50
2 4 c
Weld Area ( m m xlO )
<
R. A. PA TTERSON AND /. O. MILEWSKI are Fig. 1 — Graph of weld heat input versus weld cross-sectional area for 304L stainless steel to Alloy LU
05
with Los Alamos National Laboratory, Los Ala- 625 gas tungsten arc welds. Note that the pulsed current GTA weld photomicrographs exhibit LLI
mos, N. Mex. pronounced centerline cracking a

WELDING RESEARCH SUPPLEMENT | 227-s

Table 1—Welding Variable Range Used to
Produce Heat Input Versus Weld Area
Graph

Parameter Value Range

Weld Current 30-100

(A)
Arc Potential' 3 ' 10-13.5
(V)
Weld Speed 1.8-3.6
(mm 's) Spot Weld
Pulse 1.5-10
Frequency (Hz) Top View Wedge
Low Pulse 25"-. of weld
Current (A) current Specimen After Bend GTAW Torch Attached To
Pulse Duty 50",,
Cycle Specimen Before Bend Die Block Support Block
Torch Gas Pure Argon To Move With It
Electrode 1.5 mm dia. W-2"„
Th
Electrode Tip 30 included angle
Geometry to a point
= = r—I <3rips
(a)
Arc voltage was a dependent variable ol the weld
current and was set lo concur with acceptable welding
practice.

Hz), the surface appearance of the weld

became macroscopically similar to a con-
tinuous current weld and as such no
effort was made to evaluate the pulsed
current welding at higher frequencies.
Essentially no cracking susceptibility was
observed when utilizing continuous cur-
rent GTA welding —Fig. 1. This result Wedge
spawned t w o questions: Why is pulsed
current GTA welding producing an
increased cracking susceptibility? What
\^S<^Z> "-Stop Blocks
mechanism promotes the cracking in this
dissimilar metal combination?
Pulsed current welding has in the past Front View
been touted due to reductions in the Fig 2 —Schematic oi the Spot Varestraint test apparatus showing a weld specimen after
average heat input and compositional bending
segregation which results in an assumed
decrease in weld cracking susceptibility
(Refs. 1-3). Reduced average heat input
does produce greater uniformity in weld nickel alloys is promoted by the segrega- over a range of nearly 100°o 304L to
geometry and penetration by reducing tion of S, P, Ti and Nb with a suggested 100% Inconel 625. Cracking susceptibility
the effect of preheat. A second benefit is Laves Phase (NbFe) formed in the nickel was monitored using the Spot Varestraint
a reduction in post-weld residual stress alloys. Savage, et al. (Ref. 5) reported that technique since this technique would
(Ref. 1). However, the assumption of segregation of P and S caused hot crack- demonstrate the cracking susceptibility of
reduced compositional segregation is ing in Inconel 600. Ogawa, et al. (Ref. 6) a single spot weld and perhaps best
probably not as accurate as stating that reported auger electron spectroscopy illustrate the controlling mechanisms in
the segregated solute is redistributed in a (AES) results indicating that P, S and Nb pulsed current GTA welding.
finer second phase due to the higher promote cracking in austenitic stainless
solidification rates resulting from steeper steel welds and that increased C concen-
temperature gradients. Unfortunately, trations appear to counteract the Nb Experiment
these steeper temperature gradients may effect. The literature, thus, presents a The Spot Varestraint Test (Refs. 5 and
increase the susceptibility to cracking fairly uniform opinion that segregation of 9) was developed to evaluate the relative
through an increase in the stress gradient P, S, and Nb in austenitic welds promotes hot-cracking sensitivity of base materials.
associated with solidification. Proof of cracking. Figure 2 shows schematically the Spot
this hypothesis is not presented here, but Varestraint Test which incorporates a
The basis for experimentation here
rather an experimental examination of stationary spot welded specimen that is
was to examine the role of weld metal
the cracking mechanism from a weld deformed by a pneumatically driven radi-
dilution on the cracking mechanism. An
dilution point of view. us block during solidification. Crack sus-
assumed difference in the threshold con-
Previous investigators have reported centration of elements which promote ceptibility is gauged by measuring the
that fully austenitic welds are susceptible cracking (S, P, Nb) for either Alloy 625 or total length of weld surface crack versus
to solidification cracking (Refs. 4-8). Rob- 304L is the primary point of discussion. the amount of augmented strain intro-
inson and Scott (Ref. 4) reported that An experiment was designed which duced during solidification. Augmented
cracking in austenitic stainless steels and allowed weld metal dilution to be varied strain is calculated from the relationship:

228-s | AUGUST 1985

 Table 2—Weld Variable Schedule
— Alloy 625
Time Current
I evel (Seconds) (Amperes)

Prepurge 30
— 304L S.S. Arc Start 1 15
Upslope 1 15 to W.l.<a>
Weld 5 W.I.
Downslope 1 W.I. to 20
Postpurge 30
Gos Tungsten Arc Weld
Ca)
W.I. = Weld current, range 45 to 70 A in 5 A intervals
1. Torch Gas = Pure argon al 35 cfh
2. Arc Length = 1.5 mm
3. Pneumatic Ram Initiation at downslope

errors into the relative dilution versus

crack susceptibility d e t e r m i n a t i o n .
Crack length measurements w e r e per-
f o r m e d using m a c r o g r a p h s of the w e l d
t o p surface (10X) as s h o w n in Fig. 4A.
Crater cracks w e r e included in the total
Electron Beom Electron Beam
crack length d e t e r m i n a t i o n (Fig. 4B), a n d
Weld Weld
represent that p o r t i o n o f crack length
p r o d u c e d w i t h o u t a u g m e n t e d strain, i.e.,
cracks resulting f r o m solidification strain.
Crack length w a s n o r m a l i z e d w i t h the
measured w e l d cross-sectional area t o
c o m p e n s a t e f o r changing w e l d p o o l v o l -
u m e . This p r o d u c e d a cracking index (Cl)
equal t o the total length o f crack d i v i d e d
b y the w e l d area.
Table 3 lists the chemical c o m p o s i t i o n s
f o r each alloy used. N o t i c e that each alloy
has a significantly different quantity o f P,
S, and N b . T h e r e f o r e , v a r y i n g the d e g r e e
Fig. 3 —Description of the modified test specimen used in this experiment. Note the use of
explosively bonded clad metal which was electron beam welded to stainless steel tabs. The GTA
spot weld was made with the Alloy 625 facing the welding electrode as shown in the top
photomicrograph. Magnification: Explosive bond I00X, electron beam welds and GTA weld
I5X

t s ; t/2R
W e l d metal dilution w a s varied by
w h e r e t h e tangential strain o n t h e speci- changing t h e w e l d p e n e t r a t i o n t o p r o -
m e n t o p surface, t, is g i v e n as a f u n c t i o n gressively melt greater a m o u n t s of 304L,
of specimen thickness, t, a n d the b e n d i n g w i t h the Alloy 625 always placed t o w a r d
radius, R. t h e w e l d i n g arc. Table 2 s h o w s t h e w e l d
Specimens w e r e cut f r o m sheet stock schedule used t o vary p e n e t r a t i o n f r o m
a n d measured a p p r o x i m a t e l y 20 m m a p p r o x i m a t e l y 0.64 m m (0.025 in.) (pene-
w i d e b y 150 m m l o n g b y 3.8 m m thick t r a t i o n t o the A l l o y 625-304L interface) t o
(0.8 X 5.9 X 0.15 in.). A m o d i f i e d speci- a p p r o x i m a t e l y 2.9 m m (0.114 in.) (full
m e n w a s used f o r this e x p e r i m e n t t o penetration). Sequencing o f t h e w e l d i n g
a l l o w m e a s u r e m e n t of crack susceptibility variables w a s p e r f o r m e d w i t h an A m p -
(total length of crack) versus w e l d metal t r a k * m i c r o p r o c e s s o r based controller
c o m p o s i t i o n (dilution of Alloy 625 a n d i n t e r f a c e d t o a M e r r i c k Engineering Incor-
304L). Figure 3 s h o w s the specimen p o r a t e d TXR100 w e l d p o w e r supply.
w h i c h consisted o f an explosively b o n d - W e l d dilution (defined in this study as
e d clad metal o f Alloy 625 t o 304L the p e r c e n t o f fusion z o n e w h i c h w a s
stainless steel (center section), a n d elec- d e r i v e d f r o m the 304L) w a s calculated b y
t r o n b e a m w e l d e d stainless steel e x t e n -
v o l u m e using the w e l d centerline cross-
sions t o create t h e desired 150 m m (5.9
sectional area a n d assuming c o m p l e t e
in.) length and yet c o n s e r v e cladding
mixing a n d circumferential w e l d p o o l
material. Explosively b o n d e d specimens
s y m m e t r y . W e l d p o o l assymetry did n o t
w e r e used t o eliminate t h e effects o f a Fig. 4 —Photographs of representative GTA
appear t o be prevalent a n d v o l u m e cal- weld surfaces showing the solidification crack-
c o n t a c t interface b e t w e e n the dissimilar
culations should n o t i n t r o d u c e significant ing: A — Cracking appearance after bending
metals and p r o d u c e specimens w h i c h
(25X); B-Cracking which resulted from the
r e s p o n d in a consistent m a n n e r t o t h e
*Amptrak - is a registered trademark of Mer- solidification strain without additional bending
a u g m e n t e d strain.
rick Engineering Incorporated. deformation (25X)

W E L D I N G RESEARCH S U P P L E M E N T 1229-s
Table 3—Alloy Compositions (%)

Fe Ni Cr Mo Nb Ta Mn Al Co
Alloy 625 3.51 60.1 21.5 9.15 3.48 0.044 0.33 0.15 0.024 0.002 0.005 0.20 0.39 0.10
304L ss Bal. 8.89 18.7 0.3 0.04 0.10 0.66 1.14 0.023 0.019 0.025 0.004 0.0014 0.10

Results of analyses performed at the Los Alamos National Laboratorv.

Alloy 6 2 5

Fig. 7 — Photomicrographs of the electron

beam welds used to join the explosively bond-
ed clad metal to stainless steel tabs. Note the
solidification crack extending from the 304L
Volume % 304L Stainless Steel fusion boundary into the weld fusion zone,
Fig. 5 —Graph of cracking index (total length of weld surface crack/weld cross-sectional area) indicating increased susceptibility to cracking in
versus calculated weld dilution in volume percent of 304L 304L rich weld compositions. Magnification:
Top 25X, Bottom 50X

ol intermixing should greatly affect the f u r t h e r substantiated the o b s e r v a t i o n that

susceptibility t o solidification cracking. cracking is m o s t p r o n o u n c e d in the 304L
rich w e l d s . Figure 7 ( b o t t o m ) s h o w s a
typical crack in t h e EBW w h i c h occurs
Results
adjacent t o t h e 304L fusion b o u n d a r y .
Figure 5 is a graph of cracking index Figures 8 A and B s h o w typical scanning
versus w e l d metal dilution f o r 0 a n d 2°o electron fractographs of the GTA w e l d
a u g m e n t e d strain. Cracking susceptibility cracks resulting f r o m the Spot Varestraint
resulting f r o m solidification strain ( 0 % test. P r o n o u n c e d solidification structures
a u g m e n t e d strain) appears t o be most o n the crack face are indicative o f a w e a k
p r o n o u n c e d in the 304L rich w e l d s . W i t h interdendritic c o n d i t i o n at t h e m o m e n t
increased w e l d strain ( 2 % a u g m e n t e d that the a u g m e n t e d strain w a s i n t r o -
strain curve), t h e peak in cracking suscep- duced.
tibility shifts slightly t o w a r d 50 p e r c e n t Borland (Ref. 10) has p r o p o s e d a g e n -
dilution. These o b s e r v a t i o n s indicate that eralized t h e o r y t o explain solidification
a high solidification cracking susceptibility cracking in terms of segregation a n d
exists f o r G T A w e l d s joining 304L and d e n d r i t e f o r m a t i o n . T h e t h e o r y states
Alloy 6 2 5 , since measurable cracking that at s o m e p o i n t d u r i n g solidification
occurs w i t h n o r m a l solidification strains the p r i m a r y dendrites b e c o m e partially
(0% a u g m e n t e d ) . Also, t h e fairly s y m m e t - i n t e r l o c k e d p r o d u c i n g a semirigid struc-
ric c u r v e shape (Fig. 5) indicates that ture w i t h l o w melting liquid dispersed
c o m p e n s a t i o n f o r cracking susceptibility a m o n g the i n t e r l o c k e d dendrites. This
is n o t possible w i t h simple joint design structure is n o t able t o a c c o m m o d a t e t h e
modifications; i.e., decreases in cracking applied solidification strain a n d cracking
Fig. 6 — Transverse cross section photomicro-
graphs of the solidification structures in GTA susceptibility require v e r y l o w dilutions of results.
welds joining 304L and Alloy 625: Top — 304L a n d , thus, impact e f f e c t i v e w e l d
T h e distribution o f the s e c o n d phase is
Photomicrograph of the pulsed current GTA penetration.
e x t r e m e l y crucial in d e t e r m i n i n g the
weld microstructure exhibiting a cellular den-
W e l d m e t a l l o g r a p h y revealed that nor- cracking susceptibility (Ref. 11). Liquid
dritic solidification mode and cracking in the
minor phase network (lOOx); Bottom — Photo- mal cellular dendritic solidification is d o m - distribution is c o n t r o l l e d b y t h e relative
micrograph of the continuous current GTA inant w i t h crack p r o p a g a t i o n along inter- free energies of grain b o u n d a r i e s ( s o l i d /
weld exhibiting a similar microstructure to the dendritic paths —Fig. 6 (top). Examination solid) and o f interphase b o u n d a r i e s (sol-
pulsed current weld (100X) o f the e l e c t r o n b e a m w e l d s (Fig. 7), i d / l i q u i d ) , w h i c h are related t o the d i h e -

230-s I AUGUST 1985

 constituents, Cr, Ni and Fe, were not 3. Weld metal compositions which are
significantly different from the bulk weld rich in the 304L constituent exhibit the
composition. Niobium enrichments were highest susceptibility to cracking as
assumed to result from precipitates since shown most conclusively by the 0% aug-
extremely large variations in Nb analysis mented strain curve in Fig. 5.
from one point to the next occurred. 4. Pulsed current gas tungsten arc
welding of 304L to Alloy 625 produces a
Summary greater tendency for cracking than the
The results indicate that a significant continuous current technique.
susceptibility to hot cracking exists in GTA 5. A modified specimen design has
welds joining Alloy 625 and 304L stainless been developed for the Spot Varestraint
steel. The cracking susceptibility was test which allows the evaluation of dis-
shown to coincide with the segregation similar metal welding.
of S, P and perhaps Nb to the interden-
dritic phase, with S the constituent Acknowledgments
present in the greatest quantity. These
The authors acknowledge the valuable
observations are in agreement with the
assistance of the following Los Alamos
cited literature and in particular the phe-
National Laboratory personnel: D. L.
nomenology proposed by Borland (Ref.
Marts for metallography and K. D.
10).
Fig. 8 —Scanning electron micrographs of rep- Rachocki for AES. Work was performed
resentative crack surfaces showing the pro- The correlation between cracking under the United States Department of
nounced solidification structure appearance index and calculated weld dilution Energy contract W7405-Eng 36. W e
which is indicative oi interdendritic cracking revealed that cracking is more likely in thank Gloria A. Romero for typing the
assisted by a weak near liquid phase. Magnifi- 304L rich welds, which is substantiated by manuscript.
cation: Top 250X, Bottom 500X the 0% augmented strain data and the EB
weld metallography. These observations
dral angle as follows: References
indicate that weld joint designs which
5SL . 1 promote Alloy 625 rich welds would be 1. Ecar, G. M „ and Brody, H. D, 1982. Effect
preferred. This criteria is counterproduc- of pulsing variables on weld bead size and
oss' 2 cos 0 / 2
tive to the present project, since the shape in GTAW of Fe-26", Ni sheets. Trends in
where o7,L is the interphase energy, 5SS is achievable weld penetration and weld Welding Research in the United States, ASM,
the grain-boundary energy and 0 is the geometry are significantly impacted by Nov. 16-18, pp. 399-417.
dihedral angle. As the dihedral angle reductions in the amount of 304L melted. 2. Ecar, G. M , Tzaveras, A., Gokhale, A.,
approaches zero, the liquid forms a con- Also, there is sufficient susceptibility to and Brody, H. D. 1982. Weld pool fluid motion
tinuous film over the grain faces (Ref. 12). cracking over the entire weld metal com- and ripple formation in pulse-current GTAW.
This occurs when the ratio 5SL/5ss is Trends in Welding Research in the United
position range to indicate unacceptable States, ASM, Nov. 16-18, pp. 419-442.
slightly greater than 0.5. weld reliability for nuclear fuel element 3. Birman, U. I., and Petrov, A. V. 1971.
Decohesion of the interdendritic fabrication. Influence of the weld metal solidification pat-
regions is possible when the high stresses tern on hot cracking during pulsed-arc tung-
resulting from solidification strain are Finally, the observation that pulsed cur-
rent GTA welds exhibited a greater sus- sten-electrode welding. SVAR Proiz, No. 6. pp.
applied directly to a weak and possibly 14-17.
liquid interdendritic phase which is evenly ceptibility to cracking than continuous
4. Robinson, |. L, and Scott, M. H. 1980.
dispersed between the dendrites. Con- current GTA welds (Fig. 1) is unexplained.
Liquation cracking during the welding of aus-
versely, if the second phase is segmented Solidification structures resulting from tenitic stainless steels and nickel alloys. Phil.
or confined to grain triple points, the both techniques appear identical —Fig. 6 Trans. R. Soc. Land., A 295, pp. 105-117.
strength of the remaining dendrite (top and bottom). The compositional seg- 5. Savage, W. F., Nippes, E. F., and Good-
boundaries could redistribute the resul- regation in pulsed current welding is suf- win, G. M. 1977. Effect of minor elements on
tant solidification strain and cracking ficient to promote cracking, but the seg- hot-cracking tendencies of Inconel 600. Weld-
regation occurring in the continuous cur- ing journal 56 (8):245-s to 253-s.
would not occur.
rent welds should not be significantly 6. Ogawa, T., and Tsunetomi, E. 1982. Hot
Sulfur, and to a lesser extent phospho- cracking susceptibility of austenitic stainless
different. Therefore, the previously pos-
rous, have been shown to promote steels. We/ding journal 61 (3):82-s to 93-s.
tulated increase in solidification stress
cracking (Refs. 5-8). Cracking has also 7. Linnert, G. E. 1967. Weldability of austen-
resulting from steeper temperature gradi-
been reported to be caused by the itic stainless steels as affected by residual
ents and an additional stress imposed by
formation of continuous interdendritic elements, ASTM STP 418, American Society
overlapping spot melting events must be for Testing and Materials, pp. 105-119.
phases, which, due to interfacial tension
sufficient to cause cracking with the 8. Lippold, ). C. 1983. An investigation of
forces, wet the entire boundary region
pulsed current technique. This is especial- heat affected zone hot cracking in alloy 800.
(Ref. 11). Review of the solidification
ly true if one spot melting event is only Welding journal'62(1): 1-s to 11-s.
structure resulting from GTA welds
partially solidified during the next melting 9. Turner, P. W „ and Lundin, C. D. 1970.
between Alloy 625 and 304L (Fig. 6) Effect of iron on the hot cracking of uranium
event.
reveals that a eutectic-like structure is weld metal — part I. Welding journal 49
present with the second phase evenly (12):579-s to 587-s.
dispersed around the dendrites. Conclusions
10. Borland, I. C, and Younger, R. N. 1960.
Auger electron spectroscopic (AES) 1. Spot Varestraint tests indicate that Some aspects of cracking in welded Cr-Ni
analysis performed on the fracture sur- gas tungsten arc welds between Alloy austenitic steels. Brit. Welding j. 7 (1):22-59.
faces of representative Spot Varestraint 625 and 304L exhibit a high sensitivity to 11. Kammer, P. A., Masubuchi, K., and
specimens indicated that a high degree of solidification cracking. Monroe, R. E. 1964. Cracking in high-strength
steel weldments —a critical review. DMIC
S and P segregation had occurred. Alloy 2. Auger electron spectroscopy per- Report 197, February 7.
constituents which were enriched on the formed on the crack faces indicated that 12. Smith. C. S. 1948. Grains, phases and
crack surface in descending amounts S and P segregation is promoting this interfaces: an interpretation of microstructure.
were S, Nb, P and C while the major alloy cracking sensitivity. Trans AIME 175:15-52.

WELDING RESEARCH SUPPLEMENT 1231-s