Bull. Mater. Sci., Vol. 19, No. 3, June 1996, pp. 549-557. :~? Printed in India.
Structure-property correlation on AA 2219 aluminium alloy weldments
P SRINIVASA RAO, K G SIVADASAN and P K BALASUBRAMANIAN
Materials and Metallurgy Group, Vikram Sarabhai Space Centre, Trivandrum 695 022, India
MS received 3 May 1995; revised 2 January 1996
Abstract. AA 2219 aluminium alloy is the material proposed for the construction of
tanks for liquid cryogenic fuels like liquid oxygen and liquid hydrogen, in Indian launch
vehicle programmes. To meet the material requirement, ISRO has developed the alloy
AA 2219 indigenously on an industrial scale. Process development and structure property
correlation of auto TIG weldments of AA 2219 AI alloy reported in this paper establishes that
postweld reaging of AA 2219 weldments in T62 condition enhances both strength and
ductility.
Keywords. AA 2219 aluminium alloy: cryogenic fuel tank: auto TIG welding; postweld
reaging; structure- property correlation.
1. Introduction
Aluminium alloy AA 2219 is the most used material for the construction of liquid
cryogenic rocket fuel tanks. It has a unique combination of properties such as its good
ability to weld, high strength to weight ratio and superior cryogenic properties. AA
2219 aluminium alloy has already been successfully used in various launch vehicle
systems such as Saturn V, Apollo, Space Shuttle etc. and is the material selected for
Ariane V and for substituting AFNOR 7020. The cryogenic tankage being supplied to
India by Russia is also using AA 2219 for the liquid hydrogen tank. Pressure vessels of
AA 2219 aluminium are usually re-heat treated after welding or used in the as welded
condition. The T81 and T87 tempers (table 1) are recommended for assemblies to be left
in the as welded conditions.
Assemblies to be re-heat treated after welding, may be welded in 'F' condition
(Robinson et al 1962). For better weld efficiency the most desirable weld/heat treating
sequence is solution treatment and aging after welding. The use of this sequence may be
prohibited for large tanks due to distortion and quenching problems which can occur
during solution treatment. Another methodology of weld sequence is a post-weld
aging. However, there are contradictory reports about ductility of weldments. Hence,
further studies are required to optimize the re-aging sequence for better ductility
(Material data handbook on AA 2219, 1972).
AA 2219 was neither produced nor used till recently in India and now has been
indigenously developed on an industrial scale. Plates and sheets of width up to 1"5 m
have already been produced in India. It proposes to use indigenously produced AA
2219 A1 alloy for cryogenic fuel tank.
A technological programme has been evaluated to establish weld parameters for
welding of AA 2219 alloy sheets and plates. Emphasis is on studying the effects of
post-weld reaging treatment on the properties of weldment. This paper deals with the
structure-property correlation on auto TIG weldments of AA 2219 3 mm A1 alloy
sheets. The welding technology has already been published elsewhere (Vinod Kumar
et al 1994).
549
550
P Srinivasa Rao, K G Sivadasan and P K Balasubramanian
Table 1. Tempers and aging treatments for AA 2219
alloy.
Temper
Description
F
T31
T37
T4
T62
T81
T87
As fabricated
Solution heat treat + stretch
Solution heat treat + 8% cold work
Solution heat treat
T4 aged 36h at 190°C
T31 aged 18h at 177°C
T37 aged 24h at 163"C
Table 2. Chemicalcomposition.
Elements (wt%)
2.
Material
Cu
Mn
Zr
V
Ti
AI
AA2219
AA2319
6"3
6-25
0"38
0-36
0.2
0-2
0"1
0-12
0.06
0.13
bal
bal
Materials and experimental methods
The chemical composition of 2219 sheet and 2319 filler wire as analysed by XRF
method is shown in table 2.
The 3 mm thick AA 2219 aluminium alloy sheets used in this study were produced
indigenously and heat treated to T62 temper involving solution treatment at 535°C and
water quenching and artificial aging for 36h at 190°C. Auto T I G welding was
accomplished using POLY S O U D E SAF F R A N C E - 360 A Welding Machine.
The weldments were sectioned, polished and etched using Keller's reagent for
metallographic examination. A Reichert MeF optical microscope was used to examine
the weld quality that included porosity, weld interface and grain structure. A Vicker's
microhardness tester was used to determine the hardness profile along the weld at
a load of 100 g. Tensile specimens were cut transverse to the weld, with the weld bead in
the centre. Tensile tests were carried out using Model I N S T R O N Universal Testing
Machine. Tensile fractures were observed using Stereo Scan 250 Mk3 Model scanning
electrolJ microscope (SEM) for fractographic analysis.
3.
Results and discussion
Mechanical properties and microhardness values of the as-welded samples are compared with parent metal properties in table 3.
Since the dominant factor governing all significant thermal events in welding is the
energy input which affects the peak temperature distribution and solidification time,
metallurgical considerations favour low energy heat input. It can be anticipated that
tensile properties of low heat input (condition D of table 3) would be highest due
to reduced total thermal exposure and thereby better metallurgical characteristics.
Structure-property correlation on AA 2219 aluminium alloy
551
Table 3. Efficiencyand microhardness in as welded condition.
Conditions
A
B
C
D
PM
2219
T62
Weld
Heat
speed
input
(mm/min) (J/ram)
150
200
250
300
--
800
625
575
510
--
Weldment
microhardness
Mechanical
properties
Weld
efficiency
(%)
LS
C
RS
UTS
YS
(MPa)
%El
UTS
YS
89
81
118
97
122
79
82
82
78
122
89
81
102
95
122
265
267
259
227
410
176
176
179
172
310
7
6
4.8
4.2
10min
65
66
64
56
57
58
58
56
LS, Left hand side; RS, right hand side; C, centre of weld: PM, parent metal.
Figure 1. Microstructure of condition D at the centre of the weld. Solute handing with
twinning is seen.
However, this was not the case as shown by the tensile test results (table 3), which
indicated the lowest weld efficiency of 56% for this weldment. The weld failure occurred
at the centre of the weld. The poorer properties can be attributed to generally observed
tear drop solidification geometry at higher weld speeds which in turn results in solute
banding as shown in figure I. As a general rule, higher welding speed results in heavy
solute banding as the rate of solidification fluctuates more cyclically above and below
a mean value of solidification rate (ASM Welding h a n d b o o k 1983). The solute band is
clearly evident at the centre of the weld which solidifies at the end. The fractograph of
the tensile fracture (figure 2) exhibits delamination of solute rich and solute p o o r bands,
as analysed by E D A X , with fine dimples of ductile m o d e fracture on band free structure.
The weld structure away from the centre of the weld is not banded but is more dendritic
(figure 3a). The weld bead heat affected zone (HAZ) interface shown in figure 4 exhibits
cellular structure with uniformly distributed CuA12 precipitates as confirmed by
552
P Srinivasa Rao, K G Sivadasan and P K Balasubramanian
Figure 2. Fractograph of condition D weldment.
Figure 3.
Structure-property correlation on AA 2219 aluminium alloy
553
Figure 3. Microstructure of wetdments: a. condition D, b. condition C, c. condition B and
d. condition A.
EDAX. Such a structure is preferred to brittle banded structure, for better mechanical
properties. In the case of condition C (table 3) the weld efficiency was marginally better
than condition D but elongation was again not significant. The tensile specimen failed
in HAZ region and not in the weldment unlike in the former case. The microstructure of
weldment (figure 3b) indicates fine dendritic structure with no banding. The microstructure of HAZ region as shown in figure 5 indicates partial grain boundary melting
causing brittle eutectic structure. Large porosity of about 30/~m size were also observed
in HAZ region. This could be the reason for the failure of the samples in HAZ region.
Pores have also contributed to loss of ductility. The fractograph shown in figure 6
indicates secondary cracks in the grain boundaries and a mixed mode of fracture.
Conditions A and B have shown similar mechanical properties and are superior to
the other two sequences with better ductility. Tensile specimen in both the conditions
failed in the weldment. The microstructure of condition B (figure 3c) is more fine and
uniform than condition A (figure 3d). Microhardness variations within the weldments
554
P Srinivasa Rao, K G Sivadasan and P K Balasubramanian
Figure 4.
Microstructure of condition D HAZ.
Figure 5.
Microstructure of condition C HAZ.
is minimum in condition B as compared to condition A. The fractographs of both the
conditions shown in figures 7 and 8 show ductile mode of failure.
From the above observations it is understood that condition B with weld speed of
200 mm/min is most optimum under present experimental conditions. Weld samples of
this condition have been subjected to various reaging treatments as shown in table 4.
The term reaging is used here to represent additional stage aging after welding of
already aged material (T62 condition). Reaging time was varied from 6 h to 30 h
at the aging temperature of 190°C. The properties of post-weld aged samples are
also shown in table 4. It was observed that post-weld aging for 18 h gives the highest
weld efficiencyof 80% UTS and 84% YS with no significant drop in ductility. However,
the tensile specimens failed within HAZ and not within the weldments after the reaging
treatment. This indicates that reaging treatment resulted in significant strengthening
of the weldments. The weld strength increases from 267 MPa to beyond 328 MPa by
Structure-property correlation on AA 2219 alurninium alloy
555
Figure 6. Fractographof HAZ/weldmentinterface condition C.
Figure 7. Fractographof condition B weldment.
reaging treatment. Reaging has not caused any deterioration of properties in the
parent metal.
The failure in HAZ region is due to the fact that both weldment and parent metal are
stronger than the HAZ. The weakness of the HAZ is quite possibly due to overaging
during welding, and reaging did not have any influence on HAZ. In a previous report
(Trazil and Hood 1969) it was mentioned that post-weld aging results in loss of
ductility, the weld sequence used was solution treatment-welding-aging. In this
sequence, structure in the solution treated condition, prior to welding, contains GP
zones and 0'. During welding, the GP zones and 0' revert to solid solution and,
subsequently, 0 precipitates out, preferentially atgrain boundaries in the HAZ region
(Dumolt et al 1981). This reduces the ductility. Further aging after welding does not
enhance properties of HAZ. The failure on these samples should have presumably
occurred within HAZ with overaged coarse precipitates at grain boundaries, resulting in
loss of ductility. Hence, it can be concluded that solution treatment-welding-post-weld
556
P Srinivasa Rao, K G Sivadasan and P K Balasuhramanian
Figure8. Fractograph of condition A weldment.
Table 4. Effectof reaging on weldments (condition B, table 3).
Mechanical
properties
Reaging
time
(h)
30
24
18
14
10
8
6
Weld efficiency
(%)
(wrt PM 2219 in T62}
Hardness
BHN
UTS
(MPa)
YS
(MPa)
%El
UTS
YS
12
13
17
15
11
10
I0
287
316
328
254
221
246
246
232
248
260
202
192
195
189
5
5
6
4
6
5
3.5
70
77
80
62
54
60
60
75
80
84
65
62
63
61
aging sequence is not recommended for AA 2219 alloy. However, partial reaging after
welding in T62 condition enhances strength without the loss of ductility.
4.
Conclusions
(i) A weld speed of 200 mm/min is optimum and gives weld efficiency of 65% on UTS,
in the as welded condition.
(ii) Reaging after welding in T62 condition improves weld efficiency by progressive
strengthening of weldment by aging.
(iii) Post-weld reaging for 18 h was found to be optimum with better weld efficiency of
80% on UTS but without any sacrifice in weld ductility.
Acknowledgements
The authors thank the staff members of Materials and Metallurgy Group, R P F F and
Q D C C for extending the facilities of heat-treatment, fabrication, welding, metallography,
Structure-property correlation on A A 2219 aluminium alloy
557
non-destructive testing and materials testing. The authors are also grateful to Shri
M C Mittal, Head, SMD, Dr K V Nagarajan, Group Director, MMG and Director,
Vikram Sarabhai Space Centre for their interest.
References
ASM Welding Handbook 6 1983 9th edition pp 28-35
Dumolt S D, Laughlin D E and Williams J C 1981 Proc. of the first international aluminum welding
conference, Cleveland, Ohio (eds) R A Kelsey and C R Felmley Jr. pp 115-135
Material Data Handbook Aluminum Alloy 2219 1972 2nd Edn Western Applied Research & Development
Inc.
Robinson I B, Collins F R and Dowd J D 1962 Welding J. 42 2215
Trazil J P and Hood D W 1969 Welding J. 48 395
Vinod Kumar C, Sivadasan K G, Kurup M S, Kothandaraman G and Balasubramanian P K 1994 National
welding seminar, Jamshedpur, India pp 5-11