SUPPLEMENT TO THE WELDING JOURNAL, MARCH 1975

Sponsored by the American Welding Society and the Welding Research Council (if J^Lj

Welding of an Advanced High

Strength Titanium Alloy

Metallurgical investigation points the way to solution of

high strength titanium welding problems

BY M. A. GREENFIELD AND D. S. DUVALL

ABSTRACT. Many recently devel- imens were utilized to more fully study airframe and engine components in
oped high strength titanium alloys do this weldment region. These tests order to reduce costs, simplify con-
not possess the inherent good weld- revealed that the weld and near heat- struction, or achieve superior perfor-
ability of previous alloys in this mate- affected zone regions were extreme- mance. It is well known that porosity,
rial system. Difficulties have been e n - ly hard and brittle as welded and after c o n t a m i n a t i o n , and e m b r i t t l e m e n t
countered in achieving satisfactory a n o r m a l p o s t w e l d stress relief. problems can be encountered when
weldment mechanical properties in Studies showed that this was caused welding these chemically active
several of these alloys. The present by rapid auto-aging of an o r t h o r h o m - alloys. However, these difficulties can
investigation was conducted to study bic martensite phase formed during be p r e v e n t e d by p r o p e r p r e w e l d
the metallurgical response of one of the weld thermal cycle. Solution heat cleaning methods and adequate pro-
these materials, Ti-6246, to arc weld- treating after welding removed the tection of the material during welding.
ing. The goal was to identify the orthorhombic martensite but failed to Furthermore, because of the inherent
causes for inferior properties in a d - substantially improve weld ductility. good weldability of many of today's
vanced titanium alloy weldments and Films of alpha phase precipitated at titanium alloys, weldment cracking is
to provide guidelines for improved prior beta grain boundaries during usually not a problem during fabrica-
p r o c e s s i n g p r o c e d u r e s to c o r r e c t the solution heat treatment were re- tion and near base metal mechanical
welding difficulties. sponsible for the low weld ductility in properties can be achieved without
Gas tungsten-arc welds made on this condition. Based upon these the need for unique filler metal c o m -
0.125 in. thick sheet were evaluated results, a new postweld heat treat- positions, weld energy input controls
metallographically and with hard- ment was established which suc- and postweld heat treatment cycles.
ness, b e n d , and tensile tests to c e s s f u l increased weld ductility by The ever increasing performance
characterize their properties in the as- favorably altering the weld and heat- demand of advanced aircraft and
welded condition and after various affected zone alpha precipitation. engine systems has led to the need
postweld heat treatments. Gleeble This approach and others suggested for t i t a n i u m a l l o y s w i t h g r e a t e r
synthetic heat-affected zone spec- by this investigation appear to be strength and more heat resistance,
promising methods for improving the and several new alloys have been de-
weldability of a d v a n c e d t i t a n i u m veloped. Many applications require
M. A. GREENFIELD is Technical Manager alloys. that they be employed in welded con-
lor Joining, Air Force Materials struction. Unfortunately, initial inves-
Laboratory, Metals and Ceramics Divi- tigations suggest that some of these
sion, Wright-Patterson AFB, Ohio 45433. Introduction alloys do not have the easy weldabil-
D. S. DUVALL is Research Metallurgist,
Structural components of titanium ity inherent in previous alloys.
Advanced Materials R&D Laboratory, Pratt
& Whitney Aircraft, Middletown, Conn. and titanium alloys are widely utilized The increased amounts of beta
06458. in both airframe and gas turbine stabilizing elements (e.g.,
Paper was presented at the 55th A WS engine construction because of their molybdenum, iron, vanadium, and
Annual Meeting held at Houston, Texas, high strength-to-density ratios. others) in advanced titanium alloys
during May 6-10, 1974. Welding is often employed on these has tended to reduce weldability. The

WELDING RESEARCH SUPPLEMENT! 73-»

 'alpha-lean beta" Ti-6AI-2Sn-4Zr- considerably stronger at elevated surface c o n t a m i n a t i o n . Gas t u n g -
2Mo* alloy exhibits inconsistent arc t e m p e r a t u r e s than c o n t e m p o r a r y sten-arc welding was conducted in an
welded mechanical properties due to alpha-beta titanium alloys. At the time evacuable c o n t r o l l e d a t m o s p h e r e
sensitivity to small variations in weld this study was initiated, little expe- chamber with argon as the shielding
cooling rates (Ref. 1). The a l p h a + b e t a rience existed on joining the Ti-6246 gas. Full penetration bead-on-plate
Ti-6AI-6V-2Sn alloy has poor ductil- alloy. However, because of its high autogeneous welds were made using
ity (Refs. 2, 3) and toughness (Ref. 4) hardener content and quantity of beta 10 V dcsp, 150 A, and 5.5 ipm trav-
when conventionally fusion welded stabilizing elements, it was felt that el speed (16.4 k J / i n . energy input).
and heat treated. Likewise, unsatis- the Ti-6246 alloy would present more Current and travel speed parameters
f a c t o r y d u c t i l i t y has been e n - of an arc welding problem than other were altered on a few welds to vary
countered in fusion weldments in the advanced titanium materials in- energy input over a range of 10-20
metastable beta Ti-8Mo-8V-2Fe-3AI c l u d i n g t h e T i - 6 A I - 6 V - 2 S n alloy kJ/in.
alloy (Ref. 5). previously discussed. In order to study the response of
These studies have shown that it Studies on wrought Ti-6246 have the alloy to welding in more detail, a
can be difficult to achieve satisfac- shown that the parent alloy can ex- series of synthetic heat-affected
tory properties when welding these hibit wide ranging strengths, ductil- zones were created using a Gleeble.
high strength titanium alloys. It has ities, and f r a c t u r e t o u g h n e s s d e - Sheet specimens 3 in. long X 0.5 in.
become necessary to develop unique pending upon heat treatment (Ref. wide x 0.125 in. thick were heated to
weld and heat treat procedures to 10). Cooling rates from high solution various peak t e m p e r a t u r e s along
control the alpha-beta phase trans- heat treatment temperatures appear weld thermal cycles. The synthetical-
formations and precipitation reac- to be especially critical in relation to ly created heat-affected zone regions
tions responsible for the deleterious plane strain fracture toughness and in the center portions of the samples
weld properties in these materials. tensile p r o p e r t i e s . This behavior were subsequently examined metal-
For instance, it has been shown that s u g g e s t e d that the Ti-6246 alloy lographically or tested mechanically.
control of weld heating and cooling would be very sensitive to weld ther- Bend tests, hardness measure-
rates by preheating could improve the mal cycles. ments and tensile tests were used to
toughness of Ti-6AI-6V-2Sn welds In the initial phase of this work, the characterize the mechanical behavior
(Ref. 4). Experiments have also been microstructural and mechanical of the welded sheet specimens and
conducted to match fusion zone p r o p - property characteristics of Ti-6246 Gleeble samples. Bend tests were
erties to those of the base metal by gas tungsten-arc welds were exam- performed at room temperature using
weld chemistry control with dissim- ined in the as-welded condition and a 3-point guided bend fixture with a
ilar filler metals (Ref. 6). However, this after conventional heat treatments. 0.25 in. (2T) radius plunger. The bend
a p p r o a c h c a n n o t solve attendant These results identified deficiencies specimens were approximately 2 in.
heat-affected zone difficulties. in heat-affected zone and weld metal long, 0.5 in. wide and 0.125 in. thick.
A thorough understanding of the properties. A more detailed study was Welded specimens were bent with the
metallurgical response to welding of then undertaken, including use of welds in the longitudinal direction
these advanced titanium alloys is re- Gleeble synthetic s p e c i m e n t e c h - (perpendicular to the bend axis) and
quired if the causes of inferior heat- niques, to d e t e r m i n e m i c r o s t r u c - the face (top surface) of the weld in
affected zone and weld metal p r o p - ture/property relationships which tension. Prior to bending, the spec-
erties are to be identified and cor- were responsible for the weldment imens were mechanically ground and
rected. Some recent studies have properties. Based on these results, polished to remove the weld bead
successfully used this approach. Gas various solutions were evaluated for reinforcement and then pickled to
tungsten-arc weldment ductilities in improving the weldability of this a d - eliminate any surface contamination.
the Ti-6AI-6V-2Sn and Ti-8Mo-8V- vanced titanium alloy. Bend tests were terminated at the
2Fe-3AI alloys were significantly im- onset of cracking, when possible, to
proved by unique postweld heat treat- Material and Procedures allow identification of crack initiation
ments (Refs. 3, 7). These heat treat- sites. Hardness measurements were
ments produced heat-affected zone The chemical analysis of the heat of made on metallographic specimens
alpha phase morphologies which are Ti-6246 alloy employed in this in- using a Vickers hardness tester with a
considerably more fracture resistant vestigation is listed in Table 1. All 5 kg load. Room temperature tensile
than the microstructures obtained in material was taken from 36 in. wide by tests were conducted with the test
the as-welded or conventionally post- 0.125 in. thick sheet supplied in an specimen geometries illustrated in
weld heat treated conditions. A s i m - annealed c o n d i t i o n (1600 F/15 Fig. 2. In the welded specimens, the
ilar approach was used to increase min/air cool). Prior to welding or longitudinal test direction was paral-
poor heat-affected zone impact other processing, all specimens were lel to the weld axis. The miniature
strengths in 60 mm thick electroslag solution heat treated at 1700 F (25 F specimen design (Fig. 2b) was e m -
welds of an experimental Ti-AI-V-Mo- below the beta transus for this par- ployed for testing fusion zone mate-
Cr-Fe-Zr alloy (Ref. 8). In addition, ticular heat) for 1 hour in argon fol- rial and synthetic heat-affected zones.
thermomechanical processing in- lowed by an air cool. Welded samples Specimens for light metallography
volving postweld rolling has been e m - were evaluated in the as-welded c o n - were prepared by polishing through
ployed to favorably alter intragran- dition and after various postweld heat 600A paper followed by fine polish-
ular alpha precipitation in Ti-8Mo-8V- t r e a t m e n t s . Welds given a direct ing with alumina slurries. The s a m -
2Fe-3AI weld metal (Ref. 9). age/stress relief were heat treated in ples were then swab etched in a 50
The present investigation was c o n - argon at 1100 F/8 h followed by an part lactic acid, 30 part HN0 3 , 2 part
d u c t e d to g a i n a b e t t e r u n d e r - air cool. Other welded samples were HF solution. Thin foils for transmis-
standing of the welding behavior of given a full heat treatment by solu- sion electron microscopy were pre-
advanced titanium alloys by studying tioning at 1700 F/1 h/air cool and ag- pared by the Air Force Materials Lab-
the welding metallurgy of the alpha- ing at 1100 F/8 h/air cool. The micro- oratory and at Polytechnic Institute of
beta Ti-6AI-2Sn-4Zr-6Mo (Ti-6246) structure of the base metal given the New York (PINY) by Dr. Ernest Levine
alloy. This high strength material is full heat treatment (solution plus age) using standard solutions (Ref. 11) but
is shown in Fig. 1. with a potentiostatic power source. An
To prepare for welding, the heat x-ray diffractometer with pulse height
"This nomenclature indicates nominal treated specimens were pickled in a analyzer was used to measure the
compositions in weight percent. HNO3-HF acid solution to remove any degree of phase orthorhombicity of

74-s I M A R C H 1975
each zone of the weldment at PINY. treatment (Figs. 5a, 5b). In addition, Table 1 — Chemical Analysis of Ti-6246
Scanning electron microscopy was large amounts of alpha phase pre- Sheet Used in This Study, wt. %
employed for fractographic analyses cipitated along the prior beta grain
of broken bend and tensile spec- boundaries. Titanium Bal.
imens. The middle and lower temperature Aluminum 5.9
far-heat-affected zone regions, Figs. Tin 2.0
Results 5c, 5 d , contained a heterogeneous Zirconium 4.1
Metallographic examination of as- distribution of coarse alpha platelets Molybdenum 5.8
welded Ti-6246 bead-on-plate welds in a fine structured matrix. In the m i d - Iron 0.11
dle heat-affected zone, alpha plate- Carbon 0.026
showed that a wide range of micro-
lets in the Widmansfatten relation- Nitrogen 0.009
structures existed across t h e weld Hydrogen 0.004
zone, Fig. 3. These heat-affected and ship also formed at the sites of former
Oxygen 0.11
fusion zone microstructures reflect- equiaxed alpha particles which were
ed the effects of rapid heating and dissolved during welding. This was
cooling weld thermal cycles on an also evident in the lower temperature
allotropic material. In heat-affected
zone regions farthest from the fusion
zone, the equiaxed primary alpha
particles (white etching phase in Fig. t
3) had been partially dissolved by y/////////2^y//y////.t **<?////, v///////////////.
;
exposure to temperatures slightly
below the beta transus*. These par- \. 1 r •
tially solutioned alpha particles were
surrounded by a matrix of fine trans-
formed beta phase. The higher peak
temperatures reached in t h e middle
heat-affected zone p r o d u c e d i n -
creased dissolution of the base metal —•] 0 500"
-
microstructure. In these regions, the 0125"

structure was characterized by ghost

islands of affected alpha particles in a 2 500"-
transformed beta matrix.
All remnants of the prior equiaxed
alpha phase were erased in the high Fig. 1 — Microstructure of fully heat Fig. 2 — Specimens used for tensile test-
temperature heat-affected zone loca- treated Ti-6246 sheet employed in this ing, (a) Full-size specimen for welded sam-
tions nearest to t h e fusion zone. program. X1000, not reduced ples. (b> Miniature specimen tor fusion-
zone and synthetic heat-affected zone
These regions exhibited a fine trans-
tests
formed beta structure in large prior
beta grains. T h e fusion zone c o n -
tained a similar structure, although
NEAR H A Z
the transformed beta morphology
was more heterogeneous as a result
FUSION ZONE
of solute segregation during solid-
ification. The grain size of the prior
beta grains in the heat-affected zone
increased in t h e classical fashion as
the fusion zone was approached.
Likewise, the fusion zone contained
large c o l u m n a r p r i o r beta grains
typical of a solidification structure.
Welded specimens given a direct
age/stress relief of 1100 F/8 h after
welding exhibited the microstruc-
tures shown in Fig. 4. It was apparent
that some additional metallurgical
changes had taken place in the heat-
affected zones during this heat treat-
ment. However, it was difficult to
ascertain the exact nature of these
changes with the light microscope.
A more dramatic effect (Fig. 5) was
produced by fully heat treating the
welded specimens at 1700 F/1 h
(solution) plus 1100 F/8 h (age).
Coarse alpha platelets were formed
throughout the prior beta grains in the
higher temperature near-heat-affect-
ed zone and fusion zone regions dur-
ing the postweld solution part of t h e

"The beta transus for this heat of material

was experimentally determined to be
1725± 5 F. Fig. 3 — As-welded microstructure of gas-tungsten arc weld in Ti-6246 alloy, reduced 29%

W E L D I N G R E S E A R C H S U P P L E M E N T ! 75-s
 Table 2 — Ti-6246 Fusion Zone Properties
Room-temperature
tensile tests

Postweld Y.S., T.S., El.,

condition ksi ksi %
As welded — 187 0.4
Direct age/
Stress Relief
1100 F/8 h/AC — 184 0.3
Full H.T.
1700 F/1 h/AC +
1100 F/8 h/AC 159 165 2.4

heat-affected zone, where the alpha

platelets nucleated and grew from
f o r m e r e q u i a x e d a l p h a particles
which were only partially dissolved
during the weld thermal cycle.
It was anticipated that the variety of
microstructural changes induced by
welding and postweld heat treatment
w o u l d strongly affect m e c h a n i c a l
behavior in this alloy. A series of
microhardness traverses were taken
across Ti-6246 weldments which were
in the above conditions. The results
were quite different from what is gen-
erally e x p e r i e n c e d in arc w e l d e d
alpha-beta titanium alloys such as Ti-
6AI-4V (Ref. 12) and Ti-6AI-6V-2Sn
(Ref. 3). As shown in Fig. 6, the weld
Fig. 4 — Microstructure of Ti-6246 weldment given a direct age/stress relief at 1100 F/8 h
zone and near-heat-affected zone in
AC. (a) Fusion zone, (b) Near heat-atfected zone, (c) Middle heat-affected zone, (d) Far
heat affected zone. X1000, not reduced the as-welded condition were ex-
tremely hard when c o m p a r e d to the
base metal (460 VHN vs 360 VHN).
Moreover, the direct age/stress relief
not only failed to lower these hard-
nesses but e x t e n d e d this region
farther into the heat-affected zone.
Only after a full solution and age heat
treatment did the hardness values
level out across the weld.
Longitudinal bend tests conducted
on welded specimens in these three
conditions also demonstrated that
mechanical properties were strong-
ly altered. As-welded specimens
cracked with essentially no macro-
scopic deformation at a bend angle of
0 deg. Specimens given the direct age
also failed in a brittle manner at an
average bend angle of about 10 deg.
The fully heat treated weld spec-
imens exhibited an average bend
angle of about 20 deg although each
showed intergranular fracture across
most of the weld affected region. Fully
heat treated base metal bend spec-
imens (with the same sheet orienta-
tion) failed at angles of about 30 deg
in similar bend tests. Examination of
the fractured bend specimens in-
dicated that cracking initiated in the
fusion zone or high temperature por-
tion of the heat-affected zone in all
welded samples tested.
In order to ascertain the meaning of
these hardness and bend data and to
quantitatively determine the actual
Fig. 5 — Microstructure of Ti-6246 weldment given a full heat treatment of 1700 F/1 h/AC properties of each weldment zone, a
+ 1100 F/8 h/AC. (a) Fusion zone, (b) Near heat-atfected zone, (c) Middle heat-affected series of subsize fusion zone and
zone, (d) Far heat-affected zone. X1000, not reduced Gleeble simulated heat-affected zone

76-B | M A R C H 1975
 JPARENT
WELD ZONE - 4 - —HAZ [ IdETAl

* • - • * - . . DIRECT AGE
A ^ v STRESS RELIEF

HARDNESS
VHN
FULL HEAT
TREATMENT

AS W E L D E D \ 4 .

0 0.1 0.2 0.3 0.4 0.5

DISTANCE FROM CENTER OF WELD -INCHES

Fig. 6 — Microhardness traverses across

Ti-6246 welds

,j»»;
specimens was prepared and tensile
^mmkm
tested at room temperature. The ten- Fig. 7 — Ti-6246 fusion zone fracture surfaces showing brittle transgranular failure
sile results of the fusion zone spec- mode, (a) As welded, (b) Direct age/stress relief at 1100 F/8 h/AC. X500, reduced 33%
imens are shown in Table 2. It can be
seen that in both the as-welded and
direct aged/stress relieved c o n d i -
tion the tensile strength was extreme-
ly high with unacceptably low duc-
tility. In the fully heat treated c o n d i -
tion, the fusion zone had base metal
strength but also possessed low duc-
tility.
Examination of the fracture sur-
faces of these specimens with a scan-
ning electron microsope revealed that
the low ductility fractures could be at-
tributed to the failure mode. In the as-
welded and direct aged/stress re-
lieved conditions, the fracture sur-
faces showed a transgranular cleav- JMx
age type and a brittle precipitate fail-
ure respectively, Fig. 7. In contrast, Fig. 8 — Ti-6246 fusion zone fracture in
fully heat treated fusion zone spec- v
solution heat treated and aged specimen \
imens exhibited an intergranular fail- showing intergranular failure with ductile .)*>
ure with ductile dimpling at the grain dimpling at grain boundaries. X500, re-
boundary surfaces, Fig. 8. Subse- duced 35%
quent metallography demonstrated
that the change in fracture mode and
the low tensile elongation in the fusion heated above the beta transus were
zone were related to microstructural c h a r a c t e r i z e d by e x t r e m e l y high
phase distributions. This will be dis- strength and low ductility. This corre-
cussed later in detail. lated with the trends apparent in the
Figure 9 shows the difference in microhardness traverses and the o b -
microstructures which was found in servation that the higher temperature
synthetically created heat-affected heat-affected zone was a location for
zones as a result of different peak crack initiation in bend tests.
temperatures. These Gleeble struc- Fig. 9 — Ti-6246 synthetic heat affected
tures closely resembled the actual Discussion zone specimens, (a) Cycled to 1700 F indi-
heat-affected zone of the as-welded cated peak temperature, (b) Cycled to
sheet. Peak t e m p e r a t u r e s below Individual testing of the fusion 2000 F indicated peak temperature.
1725 F were characterized by par- zones and simulated heat-affected X1000, not reduced
tially dissolved primary alpha par- zones showed that the high hardness
ticles s u r r o u n d e d by t r a n s f o r m e d profile in Fig. 6 is related to the peak
beta, Fig. 9a. In cases where the beta temperature attained in a given area
transus was exceeded, such as in Fig. during welding. Areas which were
9b, a structure of transformed beta heated above the beta transus during
phase resulted. Consistent with pre- welding exhibited abnormally high
strength levels.
vious investigations (Ref. 13), x-ray
diffraction analysis showed no re- In order to determine the cause of
tained beta phase. these high strength levels, fusion zone
The results of tensile tests on and near-heat-affected zone samples
various Gleeble specimens are were studied by electron microscopy
1700 IBOO 1900
plotted in Fig. 10. It was observed that and x-ray diffraction. Examination PEAK TEMPERATURE OF SIMULATED
the higher the peak temperature of r e v e a l e d t h a t an o r t h o r h o m b i c HEAT-AFFECTED ZONE —F

the heat-affected zone, the higher the martensite phase (a") was formed as Fig. 10 — Room temperature tensile prop-
strength level and the lower the duc- the beta transformation product upon erties ot Ti-6246 synthetic heat affected
tility. In particular, these specimens cooling following Welding. The dark zone specimens

WELDING R E S E A R C H S U P P L E M E N T ! 77-S
 age/stress relief treatment was not
suitable for overcoming this dele-
terious a" microstructure produced
by welding. This is not unexpected
from the above discussion and the
aging response shown in Fig. 12. The
direct age/stress relief heat treat-
ment does nothing more than further
the aging reaction of the structure.
Figure 12 shows that eight hours is
not sufficient time for significant over-
aging to occur in the a" rich fusion
and near-heat-affected zone. In addi-
Fig. 11 — Electron micrograph of Ti-6246
synthetic heat-affected zone. Dark field tion, this treatment resulted in an in-
on 1011 n reflection showing twinned or- crease in hardness, Fig. 6, of the far-
thorhombic martensite. X18.000, reduced heat-affected zone and the base
62% , " ' • : * * ' ' '" * , _ - • material. This is a result of the con-
ventional a precipitation aging of the
Fig. 13 — Void nucleation and growth at
grain boundary alpha in the near heat-af- preweld solution treated material.
MIL!
fected zone of a Ti-6246 weldment ten- Since a range of peak temper-
sile specimen. X250, reduced 21% atures and cooling rates occurs dur-
450
ing the welding process, it appeared
VICKERS that a controllable solution and age
HARDNESS 4oo
cycle after welding would be bene-
NUMBER characteristic of GTA welding of Vs in. ficial. It would not only result in the
350 plate) resulted in a hardness of 445 elimination of orthorhombic marten-
VHN. This also is similar to the Fig. 12 site and achievement of base metal
300' y AS QUENCHED
curve data at 10 s e c o n d s aging. strengths but also the elimination of
1 0.1 1 10 100 so
Hence, the high hardness of the as- strength variation across the weld-
welded near-heat-affected and fusion ment. In fact, the m i c r o h a r d n e s s
Fig. 12 — 1100 F isothermal aging curve zones can be attributed to the forma- traverse in Fig. 6 for the fully heat
for beta quenched Ti-6246 material tion of orthorhombic martensite and treated material showed just that.
(From Ret. 13) its subsequent aging on cooling.
Gleeble specimens heated to a However, as with the Table 2 fusion
peak temperature in the alpha-beta zone results, fully heat treated bead-
field electron micrograph in Fig. 11 i l - phase field, 1700 F, exhibited only on-plate weldments possessed low
lustrates internally twinned ortho- m o d e r a t e h a r d n e s s in t h e " a s - ductility, e x h i b i t i n g less than 4 %
r h o m b i c m a r t e n s i t e platelets o b - w e l d e d " condition. This structure (Fig. elongation and less than 9% RA in a
served in a Gleeble heat-affected 9) was characterized as a combina- longitudinal tensile specimen. The
zone s p e c i m e n . This martensite tion of conventionally aged beta reason for the low ductility can be at-
phase occurs in the Ti-6246 and is u n - phase (strengthening by alpha pre- tributed to the formation of the grain
like the normal hexagonal martensite cipitate through nucleation and boundary alpha phase in the near
found in other titanium systems (Ref. growth) and some aged orthor- heat-affected zone and fusion zone
14). The a" martensite is soft in the hombic martensite. In fact, the entire during the 1700 F solution treat-
unaged condition as are all titanium hardness curve of Fig. 6 in the heat- ment. At this temperature, only 25 F
martensites. However, what makes affected zone can be related to the below the beta transus, the driving
this martensite unique in the case of a m o u n t of a g e d orthorhombic force for alpha precipitation was low
welding is its ability to rapidly age to martensite present. The near-heat- and hence the beta grain boundaries
extremely high strength levels. Aging affected zone was heated above the represented the most favored sites for
occurs by the precipitation of beta beta transus and consisted of all aged nucleation.
phase in the highly twinned marten- a" with resultant high hardness. As As has been shown in studies on
site, thereby effectively locking slip the peak temperature d r o p p e d below wrought alpha-beta titanium alloys
and inhibiting deformation (Refs. 13, the beta transus with increasing dis- (Ref. 17), this continuous network of a
15). This results in the material with tance from the fusion zone, there was soft alpha phase, Fig. 5a, 5b, severely
very high yield strength. Figure 12 a c o r r e s p o n d i n g d e c r e a s e in the degrades the ductility of the material.
(reproduced from Ref. 13) d e m o n - amount of or" and an increase in aged This is a consequence of the ability of
strates that aging at 1100 F for as little beta. Since the conventional beta ag- this alpha phase to readily deform by
as 10 seconds increases the beta ing reaction is less effective and slip due to its Burgers' relationship
quenched alloy's hardness from 310 slower than the aging of the a", these with the surrounding beta matrix, i.e.,
VHN to 450 VHN. It is therefore not regions had lower hardness. { I 1 0 } / 3 | | ( 0 0 0 1 ) a . < 1 1 1 > l 3 || < 1 1 2 0 > a
unreasonable to expect that a similar Finally, in the unaffected base and form voids at the alpha/beta in-
effect would be encountered due to metal, the hardness and structure terface. These voids can then grow at
auto aging on cooling after welding. were indicative of the preweld alpha low stress levels by progressive slip at
In order to determine if this indeed plus beta solution t r e a t m e n t and the void tip until they reach a critical
was the case, Gleeble specimens slow-cool consisting of a structure of size for fracture.
were heated to 2000 F and 1700 F precipitated alpha in a beta matrix. Figure 13 is a photomicrograph of
and cooled at various cooling rates. This was in agreement with Cias (Ref. the heat-affected zone of a weldment
The data are presented in Table 3. It 16) who showed that cooling rates unloaded just prior to failure. A large
can be seen that the s p e c i m e n s below approximately 10 F/s in the Ti- void is clearly seen at the grain b o u n -
heated above the beta transus and 6246 alloy avoided the formation of dary alpha phase/beta matrix inter-
cooled at 1000 F/s possessed low orthorhombic martensite; the beta face. Rapid void growth early in the
hardness, 295 VHN, equivalent to the phase transformed by nucleation and straining process promotes inter-
as-quenched hardness in Fig. 12. growth of alpha precipitates only. granular fracture (e.g., Fig. 8) and
However, cooling at 60 F/s (more A standard titanium postweld direct hence limits the achievable ductility. A

78-S | M A R C H 1975
similar phenomenon has been metal levels were achieved in welds preparing, testing, and helping to analyze
reported in metastable beta titanium given this intermediate postweld heat the experimental specimens used in this
alloys (Ref. 7). treatment, Table 4. investigation. In addition, we appreciate
These results are encouraging and those portions of the electron microscopy
and x-ray diffraction analyses conducted
Methods for demonstrate the benefits which can during this program by Drs. E. Levine and
Improving Weldability be achieved by manipulation of weld- M. Young at the Polytechnic Institute of
This investigation identified the ment microstructures. Further work is New York.
metallurgical causes of the poor weld- needed to completely optimize post-
ability of the Ti-6246 alloy. This alloy weld heat treatments for the Ti-6246 References
in t h e a s - w e l d e d and d i r e c t alloy and to fully match these cycles to 1. Mitchell, D. R., and Tucker, T. J.,
aged/stress relieved condition is un- base metal requirements. Likewise, "The Properties and Transformation
acceptable due to brittleness from the experimentation to explore the effects Characteristics in Welds in Ti-6AI-2Sn-
of weld process variations (e.g., pre- 4Zr-2Mo Titanium Alloy," Welding Journal,
autotempering of the orthorhombic 48 (1), Jan. 1969, Res. Suppl., 23-s to 33s.
martensite. A solution and age after heat and postheat control) would be
2. Schwenk, W., Kaehler, W. A., and
welding does eliminate the marten- useful to determine possible benefits
Kennedy, J. R., "Weldability of Titanium
site problem but creates another from this approach. Nevertheless, the Alloy Sheets 6AI-6V-2Sn and 8AI-1M0-
problem by the formation of a grain data from this study suggest that it will 1V," Welding Journal, 46 (2), Feb. 1967,
boundary alpha network prone to be possible to weld many of the ad- Res. Suppl., 64-s to 73-s.
subsequent low ductility intergranular vanced, high strength titanium alloys 3. Simpson, R. P., and Wu, K. C ,
failure. provided weldment microstruc- "Microstructure — Property Control With
ture/property relationships are un- Postweld Heat Treatment in Ti-6AI-6V-
The high strength level and low derstood and controlled. 2Sn," Welding Journal, 53 (1), Jan. 1974,
ductility of the fusion zone can be Res. Suppl., 13-s to 18-s.
altered by choosing a suitable filler 4. Lewis, R. E., and Wu, K. C , "A Study
metal and therefore is tractable. What Conclusions of W e l d Heat-Affected Z o n e s in the
must be improved is the ductility of 1. Gas tungsten-arc welds in the Titanium 6AI-6V-2Sn Alloy," Welding Jour-
the heat-affected zone. There are two nal, 42 (6), June 1963, Res. Suppl., 241-s
Ti-6246 alloy were extremely hard to 249-s.
possible approaches: (1) preheat and and brittle in the as-welded condition 5. Hatch, W., "Development of Welding
post heat control to achieve cooling and f o l l o w i n g n o r m a l d i r e c t Practices for Titanium Alloy 8Mo-8V-2Fe-
rates which will eliminate the forma- age/stress relief heat treatments. A 3AI," Technical Report to A r m y Materials
tion of orthorhombic martensite or (2) full solution + age postweld heat and Mechanics Research Agency, Water-
an improved postweld heat treat- treatment reduced hardness but t o w n , Ma., A M M R C PTR-73-4, M a r c h
ment at some intermediate temper- failed to substantially improve duc- 1973.
ature at which grain boundary alpha tility. 6. Mitchell, D. R., and Fiege, N. G.,
formation is limited. The second ap- "Welding of Alpha-Beta Titanium Alloy in
2. The cause for the hard, brittle
proach is more desirable in that it is One Inch Plate," Welding Journal, 46 (5),
condition of these welds was the for-
not sensitive to part geometry and May 1967, Res. Suppl., 193-s to 202-s.
mation and rapid auto-aging of
was therefore examined in this study. 7. Greenfield, M. A., and Pierce, C. M.,
orthorhombic martensite in the weld
A postweld time-temperature cycle and heat-affected zone on cooling
was sought which would effectively from welding. A normal direct
eliminate the detrimental a" without age/stress relief postweld heat treat- Table 3 — Effect ol Cooling Rate on Heat-
concurrently precipitating the un- ment only extended this reaction. Affected-Zone Hardness
desirable grain boundary alpha 3. Solution heat treatments after
phase. The temperature for such a cy- welding removed the deleterious Gleeble
cle had to be far enough below the orthorhombic martensite. However, peak Cooling
beta transus to insure adequate they caused heterogeneous precip- temp., rate
supersaturation for uniform intra- itation of continuous alpha films along F F/s VHN
granular alpha precipitation. This prior beta grain boundaries in weld 2000 1000 295
would minimize heterogeneous alpha heat-affected regions. 2000 60 445
formation at prior beta grain boun- 4. An intermediate postweld heat 2000 25 450
daries. However, the heat treatment treatment was developed which was
temperature also had to be high 1700 60 400
successful in improving the weld-
enough to either overage or dissolve ability of the Ti-6246 alloy. This heat
the undesirable a" martensite. treatment eliminated the orthor-
To find the desired postweld heat hombic martensite while producing
treatment, welded specimens were uniform alpha precipitation through- K- • ' t x * T * § ••

heat treated for various times over a out the weld and heat-affected zone.
range of temperatures from 1300 F to
1600 F, air cooled, and then aged at Acknowledgments
The authors wish to thank Messrs. M.
1100 F. The most successful post-
Murphy, C. M. Wickstrand, and A. R.
weld heat treatment was at 1400 F/8 Geary of Pratt & Whitney Aircraft and Mr.
h + 1100 F/8 h. This cycle yielded no N. L. Harruff of the Air Force Materials
preferential grain boundary alpha in Laboratory for their assiduous efforts in
either the heat-affected zone or fusion
zone. The microstructure, shown in
Fig. 14, is characterized by Widman- Table 4 — Intermediate Postweld Heat
statten alpha in aged beta. At this Treatment for Ti-6246 Weldments
temperature, the driving force for pre-
1400 F/8 h + 1100 F / 8 h
cipitation is sufficiently great to cause
this type of uniform precipitation in- Y.S., T.S., EL, R.A.
stead of the more heterogeneous ksi ksi % % Fig. 14 — Wear heat-affected zone region
grain boundary alpha phase. Since Weld 154 163 7.3 16.5
after intermediate postweld heat treat-
there is no continuous path for failure, Base metal 162 172 12.5 18.8
ment at 1400 F/8 h/AC * 1100 F/8 h/AC.
tensile ductilities approaching base X400, not reduced

W E L D I N G R E S E A R C H S U P P L E M E N T ! 79-s
"Postweld Aging of a Metastable Beta 4Zr-6Mo Alloy," Mat'l Science Eng. Vol. 9, Trans., Vol. 1, p 2648, (1970).
Titanium Alloy," Welding Journal, 52 (11), p 197, (1972). 15. Young, M., Levine, E., and Margolin,
Nov. 1973, Res. Suppl., 524-s to 527-s. 11. Williams, J. C , and Blackburn, M. H., "Deformation Behavior of Alpha, Beta
8. Grabin, V. R., and Cerkasov, N. I., J., "The Preparation of Thin Foils of Titan- and Martensite in Ti-6246 Alloy," present-
"Structure and Properties of the Heat ium Alloys," Trans. AIME, 239, p 287, 1967. ed at A S M / T M S - A I M E "Material Science
Affected Zone in the Electroslag Welding 12. Banas, C. M., " E l e c t r o n B e a m , S y m p o s i u m , " Detroit, Oct. 2 1 , 1974.
of a High Strength Titanium Alloy," Avt. Laser Beam and Plasma Arc Welding 16. Cias, W. W., "Phase Transforma-
Svarka, No. 9, 20-23, (1973). Studies," NASA Report CR-132386, NASA tional Kinetics, M i c r o s t r u c t u r e s and
9. Greenfield, M. A., and Haggard, D. Langley Research Center, March, (1974). Hardenability of the Ti-6246 Alloy," Cli-
K., "Thermomechanical Processing of a 13. Young, M., Levine, E., and Mar- max M o l y b d e n u m Report RP-27-71-02,
Welded Metastable Beta Titanium Alloy," golin, H., "The Aging Behavior of Ortho- March 6, 1972.
Welding Journal, 53 (8), A u g . 1974, Res. r h o m b i c M a r t e n s i t e in T i - 6 2 4 6 , " Met. 17. Greenfield, M. A., and Margolin, H.,
Suppl., 339-s to 342-s. Trans. Vol. 5, p 1891, (1974). "The Mechanism of Void Formation, Void
10. Hall, J. A., Pierce, C. M., Ruckle, D. 14. Williams, J. C , and Hickman, B. S., Growth and Tensile Fracture in an Alloy
L., and Sprague, R. A., "Property-Micro- "Tempering Behavior of Orthorhombic Consisting of Two Ductile Phases," Met.
structure Relationship in the Ti-6AI-2Sn- M a r t e n s i t e in T i t a n i u m A l l o y s , " Met. Trans., Vol. 3, p 2649, (1972).

WRC Bulletin No. 198

SEPT. 1974
"Secondary Stress Indices for Integral
Structural Attachments to Straight Pipe"

by W. G. Dodge
"Stress Indices at Lug Supports on
Piping Systems"

b y E. C. R o d a b a u g h , W . D. D o d g e a n d S . E.
Moore

This report presents a simplified method for calculating the stresses induced
in straight pipe by thrust and moment loadings applied to lugs and other integral
attachments. Following the philosophy of the nuclear power piping portion of Sec-
tion III of the ASME Boiler and Pressure Vessel Code, appropriate secondary
stress indices are defined. A simple and conservative formula for computing the
stress indices is developed using analytical results as a guide. A comparison is
made between experimental stress indices and those obtained using the simpli-
fied analysis procedure developed here as well as the more complex analysis pro-
cedures of Welding Research Council Bulletin 107 (WRC-107 method). The method
is extended to attachments having a variety of cross sections.
Stress indices and the appropriate simplified design formulas are developed
for analyzing integral lug attachments on straight pipe according to the philos-
ophy of Section III of the ASME Boiler and Pressure Vessel Code for Class I Pip-
ing Systems. Indices are developed for the evaluation of primary stresses, pri-
mary-plus-secondary stresses, and peak stresses due to internal pressure in the
pipe for radial thrust and transverse shear forces and torsional and bending mo-
ment loads acting on the lug; and for a thermal gradient between the pipe and the
lug. The indices for thrust and bending moment loads are based on an extensive
parameter study and are represented by simple formulas that may be used di-
rectly by designers and/or incorporated into codes and standards. From compari-
sons with other methods of analysis these formulas are considered to be more ac-
curate and easier to use. Indices for the other loadings are based in part on
strength-of-materials theory and information in the literature. Specific recommen-
dations are made for incorporating the stress indices and design formulas into
the ASME Code. As an example, a simple pipe support design is analyzed using
the recommended formulas.
Publication of these papers was sponsored by the Pressure Vessel Research
Committee of the Welding Research Council. The price of WRC Bulletin 198 is
$6.00. Orders should be sent to the Welding Research Council, 345 East 47th
Street, New York, N.Y. 10017.

80-8 I MARCH 1975