HAL Id: jpa-00226557
https://hal.archives-ouvertes.fr/jpa-00226557
Submitted on 1 Jan 1987
HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are pub- lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers.
L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.
DEVELOPMENT OF SUPERPLASTIC 8090 AND 8091 SHEET
R. Grimes, W. Miller, R. Butler
To cite this version:
R. Grimes, W. Miller, R. Butler. DEVELOPMENT OF SUPERPLASTIC 8090 AND 8091 SHEET.
Journal de Physique Colloques, 1987, 48 (C3), pp.C3-239-C3-249. �10.1051/jphyscol:1987327�. �jpa-
00226557�
JOURNAL DE PHYSIQUE
Colloque C3, suppl6ment au n09, Tome 48, septembre 1987
DEVELOPMENT OF SUPERPLASTIC 8090 AND 8091 SHEET
R. GRIMES, W.S. MILLER* and R.G. BUTLER**
British Alcan Aluminium plc, c/o Alcan International Limited, Southam Road, Banbury, GB-Oxon OX16 7SP, Great-Britain
" ~ l c a n International Limited, Southam Road, Banbury, GB-Oxon OX16 7SP, Great-Britain
* * Superform Metals Limited, P.O. Box 150, GB-Worcester WR3 8TR Great-Britain
Abstract
The paper considers the superplastic forming of the alloys Lital 8090 and Lital 8091 after processing by a route designed to confer superplastic properties.
Examples of the components that can be fabricated by superplastic forming are shown and the properties that can be achieved in such components after different heat treatment procedures are given.
Lital 8091 possesses somewhat better forming characteristics and is capable of achieving appreciably higher strengths in the formed product.
However, the alloy is relatively quench sensitive and reasonably rapid cooling is necessary to achieve the maximum properties. Lital 8090 has somewhat reduced forming capabilities and can only achieve lower strength. However, it is less quench sensitive and near to full strength properties can be achieved by air cooling from the solution treatment temperature.
Introducrion
Superplastically formed sheet metal components made from the aluminium- lithium based alloy, 8090, were first reported in 1982(l) and were exhibited at the Farnborough International Air Show in the same year. Subsequently, it was shown(2) that the basic mechanism of deformation involved in the manufacture of these components was of dynamic recrystallisation of an initially wrought structure. In this respect the behaviour of the alloy appeared very similar to the deforoiation of the, so-called, Supral alloys(3) except that, unlike the Supral alloys, the 8090 did not contain a large amount of zirconium. These early results have been followed by a spate of papers (4,5) from which it has become clear that superplastic behaviour can be obtained in aluminium~lithium based alloys covering a wide range of compositions and a wide range of manufacturing routes. Indeed, there have been suggestions that aluminiumA lithium based alloys are inherently superplastic and that any kind of special processing to induce superplastic behaviour is unnecessary.
This paper considers the alternatives that exist for developing superplastic behaviour in aluminium~lithium based alloys pointing out the advantages and disadvantages of the alternatives and also giving indication of the forming behaviour and properties of the preferred route. It should be borne in mind that the "ideal" superplastic sheet material should possess the following properties:
adequate superplastic strain capability with reasonably isotropic
Article published online by EDP Sciences and available at http://dx.doi.org/10.1051/jphyscol:1987327
C3-240 J O U R N A L DE PHYSIQUE
- superplastic properties low flow stress
- low susceptibility to cavitation during forming
- forming at temperatures within the solution treatment rdLl,c.
- low quench sensitivity
- give good service properties after precipitation treatment.
Attainment, or otherwise, of the final two points on this list will be, very largely, determined by the chemistry of the alloy and the evidence already available makes it clear that significant compositional modification is not going to be necessary for superplastic behaviour. However, all of the other properties listed can be significantly influenced by the manufacturing route used for the sheet.
Development of Optimum Superplastic Quality Sheet
The possibilities for development a superplastic quality aluminium- lithium based alloy sheet include such measures as employing a powder v'etallurgy route(5). However, if such extreme and expensive options are discounted the following three realistic options seem to exist;
1. "Rockwell route"
Processing to produce a fine recrystallised grain structure by thermal rr~echanical processing before superplastic deformation begins. Essentially this would apply the type of technology pioneered for 7XXX alloys by Rockwell Tnternational to aluminium-lithium based alloys. It has already been shown(4) that, on a laboratory scale, this route can be made to work. Nevertheless, just as the patent literature demonstrates that translation of the basic Rockwell concepts into practical factory operations, is, by no means, straight- forward so application of these principles to aluminium-lithium based alloys would be difficult in the factory production environment. In particular, in order to achieve the very fine grain sizes very large cold rolling reductions are needed in alloys that, generically, tend to have limited cold working capability. Additionally, very rapid heating rates are required for final annealing and access to suitable equipment, such as large salt baths, is increasingly difficult.
Sheet produced by this route should have the advantage of considerable stability in prolonged exposure to elevated temperatures so might be the ideal structure if used, in due course, in conjunction with diffusion bonding.
2. High zirconium-dynamic recrystallisation "Supral Processing"
Again, it has been demonstrated that the use of this route is, in principle, feasible. The Supral route requires a high level of super- saturation with zirconium which, in turn, requires abnormally high casting temperatures and casting equipment sepcially designed to minimise the residence time of the molten alloy in the casting head. This was not easy for the Supral alloys and, given the very high reactivity of molten aluminium-lithium based alloys, presents daunting difficulties for aluminium-lithium. Nevertheless, Figure 1 shows that good superplastic ductilities can be achieved over a fairly wide temperature range although the optimum superplastic behaviour is achieved at tenlperatures too low for simultaneous solution treatment. A further disadvantage of the low forming temperature is that the flow stresses involved are considerably higher than those for sheet with a higher forming temperature.
Substantially larger superimposed hydrostatic stresses would, therefore, be
required to suppress cavitation and much greater demands would be made on forming machinery. The low forming temperature does have the advantage of allowing use of relatively cheap mould materials for forming.
3 . Normal zirconium-dynamic recrystallisation "Modified Supral processing"
As indicated above some superplastic behaviour can be obtained from many standard aluminium-lithium alloy sheets, particularly if the starting condition of the sheet is "as rolled". This has the major advantages of using entirely standard alloy compositions - where the compositions have been developed because of the service properties conferred - and of being superplastically formed in the solution treatment temperature range. It has t.he disadvantages of giving inconsistent behaviour, frequently including high flow stresses, a high cavitational tendency and poor surface finish. The susceptibility to cavitation is greatly increased if the sheet contains any coarser recrystall- ised grains and this is illustrated in Figure 2 where gross cavities have formed in the immediate vicinity of the coarser grains.
Optimisation of the Modified Supral route by Laboratory Processing
In view of the strong arguments in favour of the third of the above routes, a series of laboratory based experiments was conducted with the objective of retaining the advantages of the route while producing consistent good behaviour in terms of isotropy, flow stress and cavitation. Basically
t h r e e r o u t e s w e r e selected in which, starting from entirely standard 8090 and
8091 ingot, metal was processed by a variety of thermal/mechanical treatments to sheet gauge. Details of the processing remain proprietorial and, in the remainder of this paper, they are simply designated A, B and C. However, Table 1 indicates the optimum superplastic strains obtained in longitudinal and
transverse directions in both alloys.
Table 1 : Optimum superplastic ductilities for experimentally processed 8090 and 8091 sheet
In the progression from route A to route C the average superplastic ductility has, generally, increased while the anisotropy has decreased. The general level of superplasticity in the 8091 was higher than that in 8090.
Perhaps the most important trend, however, was in cavitational behaviour as Manufacturing Route
A
B
C
Test Direction
L T L T L T
Optimum Elongation (%) 8090
350 2 00 265 34 0 3 3 5 4 80
8091 -
395
425
63 0
570
710
54 0
C3-242 JOURNAL DE PHYSIQUE
illustrated in Figure 3. The route A material showed a far higher rate of development of cavities in specimens strained in the transverse direction while in the longitudinal direction the rate of cavity development was lower, but, nevertheless considerably faster than that for the route C material tested in the longitudinal direction. Route C material had roughly comparable cavitational tendencies in both directions of testing.
Confirmation of. Route C with Factory Production
Having developed the thermal/mechanical processing route on the laboratory scale, a number of ingots in both 8090 and 8091 compositions was factory processed to confirm that scaling up would be satisfactory. The 8091 was processed to gauges between 1.5mm and 4mm while the 8090 was produced at a
gauge of 2mm.
Figure 4 shows the uniaxial elongations in the 8091 sheet, but, for the sake of clarity, longitudinal and transverse ductilities have been averaged.
In the great majority of cases the longitudinal and transverse test results differed by less than 75%. The majority of the tests employed an initial strain rate of 1 x lo-* per second but a limited number of tests was also performed with an initial strain rate of 1 x loq3 per second, giving approximately a doubling in ductility. All the tests were performed in air at atmospheric pressure and there seems every reason to suppose that substantially higher strains would have been achieved had hydrostatic stress been applied.
This has, for example, been demonstrated by ~hosh(~) using 8091 manufactured by, essentially, the same route. Figure 5 shows the ductilities for 8090 sheet tested under the same conditions. The superplastic strain capability of the material was good enough to allow the manufacture of reasonably difficult components, without application of back pressure, and components made from both 8090 and 8091 are illustrated in Figure 6.
A series of top hats was produced to give bi-axially formed material upon which to measure mechanical properties and, in particular, to assess the quench sensitivity of the material. The maximum thickness strain for most of the top hats was restricted to -100% in order to minimise strength reduction effects from cavitation. The top hats were formed at 5 3 0 ~ ~ and were either still air cooled (SAC), cooled by a cold air blast (AQ) or cold water quenched (CWQ).
Samples from the base of an SAC top hat were also re-solution treated and cold water quenched. Figure 7, for the 8091 illustrates that there is no significant difference between the strengths developed in components quenched off the press and those that are re-solution treated. Properties deteriorate significantly as the cooling rate after forming decreases and the deterioration is appreciably greater in the thicker gauges of sheet. In contrast, 2mm 8090 top hats showed no consistent reduction in properties between cold water quenching and air blast and only slight deterioration for still air cooling (Figure 8).
Large superplastic strains will, inevitably, lead to significant
cavitation and serious loss of properties. This effect is illustrated in
Figure 9 for 2mm 8090 sheet. Satisfactory manufacture of structural components
will certainly require superimposed hydrostatic stress to suppress the
formation of cavities and Table illustrates the results for cones formed to
failure at a temperature of 5 2 5 O ~ while subjected to a back pressure of 2.1Mpa.
Table 2 : Cone formation with superimposed hydrostatic stress
While the Route C material and standard unrecrystallised 8090 sheet achieved almost identical depths to failure it should be noted that the distribution of metal in the Route C material was considerably superior to that in the standard sheet although the standard sheet took more than twice as long to form. At failure cavitation can be detected in all the cones but Figure 10, comparing the grain structure and cavitation near the pole for standard and specially processed sheet demonstrates that there is a far greater cavitational susceptibility in the standard sheet than in the specially processed material.
Material
2mm 8091 "Route C"
3mm 8090 "Route C"
3mm unrecrystallised 8090
By optimiation of the thermal mechanical processing 8090 and 8091 sheet can be produced that combines good isotropic superplastic capabilities with the ability to be simultaneously formed and solution treated. The surface quality of biaxially formed components is superior to that of the majority of components superplastically formed from non-lithium containing aluminium alloys.
The 8090 sheet shows very little quench sensitivity and in gauges up to about 3mm, gives virtually full ~.trength properties after air quenching off the forming machine. The 8091 sheet exhibits greater quench sensitivity and air quenching results in a detectable reduction in properties comparied with cold water quenching particularly in gauges thicker than about 2mm.
Depth to failure (mm>
9 7 9 9 9 9
The final properties in the superplastically formed component are more isotropic than the properties in a conventionally formed component.
Sheet made by a conventional manufacturing route may, at best, exhibit reasonable superplastic strain capability. However, compared with spf route sheet, the flow stress is higher, cavitational tendency greater and anisotropy and surface quality worse.
Time to failure (mins)
23 2 7 68
Acknowledgements
Average Thickness at pole
(mm)
0.5 0.65 0.4
The contribution of British Aerospace, Warton Division in performing the
cone tests is acknowledged.
JOURNAL DE PHYSIQUE
References
1. R. Grimes Sheet Metal Industries, 59, (1982), 885 2. R. Grimes and W.S. Miller, pp 153-167 in
Aluminium-Lithium Alloys I1 T.H. Sanders and E.A. Starke (Eds), AIME, New York, NY, (1983)
3. R. Grimes, C. Baker, M.J. Stowell and B.M. Watts, Aluminium, 51, (1975), 720
4. J. Wadsworth, I.G. Palmer, D.D. Crooks and R.E. Lewis pp 111-135 in
~luminium-~ithium Alloys 11 T.H. Sanders and E.A. ~ t a k k e (Eds), AIME, N e w ~ o r k , NY, (1983)
5. R.J. Lederich and S.M.L. Sastry, ibid, pp 137-151
6. A.K. Ghosh, Westec 87 to be published by ASM
Figure 1 Superplastic ductilities of high zirconium sheet produced b y the "Supral" route
Figure 2:
Cross cavitation in the vicinity of coarse re- crystallised grain
etched
micro-anodised and
photographed under
polerised light
C3-246 JOURNAL DE PHYSIQUE
Route A Route C
100% strain
300% strain
Figure 3 Development of cavities with uniaxial strain (atmospheric pressure)
under optimum straining conditions for 8090 sheet processed by
Route A and Route C
Figure 4 Superplastic Figure 5 Superplastic ductilities (average of L and T) ductilities for 2mm 8090 sheet f o r various gauges o f 8091
sheet. Initial strain r a t e as indicated.
1.5mm.
A 2.3mm.
V kOmm.
Figure 6 Components superplastically formed without back pressure a t Superform ~ d l e t a l s . F r o ~ ~ t component i s 8091, other two is 8090 sheet
700
- 5 0 0 -
p
Z
e
C 4
w
Z5 300
Y
r - - - - - I x 1 0 - L C - 1
IX10-2 sec-1
C
- -A
/
P--C 8 0 0
. .
# / /
o*
/$j 6 0 0
-
Z9
C 4
