This invention is concerned with superplastic forming and in particular with an improved method of superplastic forming sandwich structures from a plurality of aluminium alloy sheets.

The formation of sheet metal into useful configurations is usually achieved by forces applied to the material by hard tooling, such as in press forming. By contrast, in superplastic forming processes, the required deformation of the workpiece is accomplished by application of gas pressure. A die cavity delimits expansion of the workpiece, but intimate contact between the die and the deforming sheet does not occur until the desired configuration has been achieved in the vicinity of the contact zone.

Many aluminium alloys have the potential to undergo superplastic forming and consequently there is a desire amongst aerospace manufacturers to produce aluminium alloy structures using combined diffusion bonding/superplastic forming processes (DB/SPF) similar to those currently used for titanium components.

However, the propensity of titanium alloys to undergo diffusion bonding is superior to that of the majority of structural aluminium alloys and in many cases strengths of titanium diffusion bonds are comparable to that of the bulk metal. By contrast, the bond interface in joints between aluminium alloys may be much weaker than the parent material. This is especially true at the elevated temperatures required for superplastic forming. As a result of this, when a stress is applied normal to the diffusion bond, the joint is susceptible to peeling under a relatively low peel force.

Low peel strength precludes or limits the use of diffusion bonds in multiple sheet structures with the result that manufacturers are actively seeking alternatives to superplastic forming of aluminium structures lest peel f; _ure should occur before stresses in the sheet assembly reach t- levels required to effect superplastic deformation.

In our earlier patent application GB 224l 1^ we disclose a method for diffusion bonding aluminium-lithium alloys which gives rise

to materials having improved properties compared with previously-known methods. The resultant materials are capable of withstanding high shear stresses, but they have insufficient peel strengths to undergo superplastic forming except in the case of very thin sheet sections cf 1mm thickness or less.

This is in marked contrast to the high peel strengths observed in diffusion bonded titanium alloy joints and it is therefore an object of the present invention to effect an increase in the peel resistance of diffusion bonded aluminium alloys to the extent that such materials are rendered suitable for processing by superplastic forming.

The invention is a method of superplastically forming a sandwich structure from a plurality of aluminium alloy sheets, the method comprising:

assembling a stack of aluminium alloy sheets in face-to-face relationship, said stack comprising a pair of outer sheets and at least one core sheet;

diffusion bonding the sheets in bonding zones according to a predetermined pattern to create a workpiece having internal cavities;

heating the workpiece to the superplastic forming temperature of the alloy sheets, and

applying gas pressure to the cavities to effect superplastic deformation of the workpiece,

characterised in that reinforcing material is provided in the stack prior to the diffusion bonding step in the zones where diffusion bonding is to be carried out, said reinforcing material thereby serving to prevent the tensile stresses exerted during subsequent superplastic forming from causing peel fracture of the diffusion bonds such that superplastic strain occurs selectively in those regions of the workpiece between the diffusion bonded zones.

The reinforcing material may take the form of separate pieces such as strips or patches which are bonded in a predetermined pattern on

the parent sheets from which the workpiece is to be formed. Conveniently, these separate pieces are diffusion bonded to the parent sheets in the same diffusion bonding operation which joins the sheets together. Alternatively, the reinforcing material may be integrally formed with the parent sheet material. This is achieved' by machining or chemically milling thicker sheets in selected regions to create a predetermined pattern of stepped areas, the raised portions of which are used in diffusion bonding to adjacent sheets. The optimum thickness and stiffness of the reinforcing material will depend on the superplastic forming conditions used and on the properties of the workpiece. For example, a thicker workpiece will require a stiffer or thicker reinforcement for achieving the same degree of peel resistance. In order to minimise weight increases resulting from the inclusion of reinforcing material, it is preferable that it should be formed of relatively stiff material whenever possible. Thus, in the cases where the reinforcing material is bonded to plain sheets, as distinct from the cases where it is integrally formed from thicker sheet stock, the reinforcements are advantageously formed from a stiff material such as a metal matrix composite, or of another alloy which is stiffer than the parent sheets, or possibly even a ceramic material. It is particularly advantageous for the stiffness of the reinforcing material to be relatively higher than the stiffness of the parent alloy sheets at the superplastic forming temperature of the latter, i.e. typically in the temperature range 15"530 ^β C.

Depending on the material from which the reinforcements are made, it may sometimes be necessary to use a liquid phase diffusion bonding technique to achieve effective adhesion of the reinforcing material to the alloy sheets.

Further measures aimed at minimising weight include contouring the reinforcements, for example b*~ providing them with grooves normal or parallel to the edge of the diffusion bond. The invention will now be described by way of example with reference to the drawings, in which:

Figure 1 shows a peel test specimen without reinforcing material before and after testing; Figure 2 shows corresponding views of a test specimen provided with reinforcing material; Figure 3 is a graph comparing the peel performance of test specimens configured as shown in Figures 1 and 2; ^• Figure 4 is a cross-sectional view of an assembly of alloy sheets with separate reinforcing material prior to diffusion bonding; Figure 5 is a cross-sectional view of an assembly of alloy sheets having integral reinforcing material prior to diffusion bonding; Figure 6 shows the sheet assembly of Figure 4 after diffusion bonding; Figure 7 shows the diffusion bonded sheet assembly of Figure 6 during superplastic forming; Figure 8 shows completion of the superplastic forming operation, and Figure 9 shows a finished truss sandwich.

Referring now to Figures 1 and 2, Figure la shows a 90° T peel test piece 10 prior to peel testing. This test piece comprised two strips 11, 12 of 8090 aluminium-lithium alloy (Designation of the Aluminum Association of America) 1.6mm in thickness bonded together along part of their lengths at 13, with their free ends bent at 90° tc form the arms 14, 15 of the "T". The bond 13 was formed by solid state diffusion bonding according to the method disclosed in our earlier patent application GB 2241 9l4 A, using a pressure of 0.75 ^Mp a for 4 hours at 560°C.

Elevated temperature peel resistance of the test piece 10 was assessed by applying a tensile load across the arms 14, 1 to effect deformation. This assessment was carried out at 530°C under a strain rate of 3 x lO^s ^"1 between the loading points, these being the conditions under which 8090 aluminium-lithium alloy exhibits superplastic behaviour.

Figure lb shows test piece 10 after deformation. The bond 13 peeled open under the test conditions by progressive incremental peel fracture and bending of the arms in the diffusion bonded zone 13. without any apparent plastic deformation of the arms 14 and 15•

Figure 2a shows a modified 1.6mm sheet test piece 20 machined frcTr.

4mm 8090 aluminium-lithium alloy sheet stock. The full 4mm thickness was retained at the extremities of the strips 21, 22 to prevent failure of the test piece at the loading pin holes, but otherwise the arms 24, 25 were reduced to 1.6mm thickness to facilitate direct comparison with the Figure 1 example, these arms being diffusion bonded to each other at 23- Test piece 20 was further modified by having strips 26, 27 of reinforcing material solid state diffusion bonded to the strips 21, 22 in the region of diffusion bond 23. The reinforcing strips 26, 27 were composed of 5 ^m " ^tn i k metal matrix composite consisting of 8090 aluminium-lithium alloy containing 20% by weight of silicon carbide particles of average diameter 3μπι.

Figure 2b shows test piece 20 after it was subjected to the same temperature conditions and tensile stresses as test piece 10 above. About 50 superplastic elongation was obtained in the arms 24, 2 without any sign of peel at the bond line 23. Figure 3 shows the peel strength v. cross-head displacement curves for the test pieces 10 and 20 at 530°C. The solid curve represents the behaviour of the unmodified test piece 10 and shows a load peak at 6.7N_-m ^-1 followed by a plateau region. The peak marks the onset of peel in the diffusion bond 13 and the plateau indicates that peel fracture propagates at more or less constant load.

The performance of modified test piece 20 is represented by the broken line in Figure 3- Here a load peak at a strength of about 7.6Nmm ^"1 is reached after relatively little displacement, this peak indicating the onset of superplastic extension of the arms 24, 25- The curve appears to show that this superplastic deformation takes place under decreasing load, but in fact the superplastic flow stress is nominally constant. The downward slope of the curve is caused by the reduction in both width and gauge of the arms 24, 25 as extension takes place.

This improvement in peel strength is also reflected in room temperature performance: After solution heat treatment at 530°C, T peel test pieces similar to test pieces 10 an: 20 above were air cooled and aged, then subjected to tensile loading as before. The load peak obtained for the modified test piece with reinforcements was approximately five times greater than that for the unmodified test piece, as indicated below:

Further tests using specimens having 4mm thick aluminium alloy sheets stiffened with 5 ^mm thicknesses of metal matrix composite as before produced a peak peel strength of 17.4Nmm ^_1 prior to the onset of peel fracture. Whilst the above examples show the efficacy of this technique for solid state diffusion bonded specimens, tests have shown that it is equally applicable to those situations where liquid phase diffusion bonding is employed. The peak and plateau values for peel strength of unstiffened specimens constructed from 1.6mm thick aluminium alloy sheets which had been diffusion bonded using a lOμ interlayer of copper by a liquid phase technique such as that described in our earlier patent application GB 224l 914 A, were 4.4Nmm ^"1 and 1.9Nmm ^-1 , respectively. These values are too low to sustain a superplastic forming operation without peel fracture. By contrast, liquid phase diffusion bonded specimens which had been reinforced in accordance with the invention were able to undergo superplastic forming quite satisfactorily.

The improvements in peel strength for stiffened liquid phase diffusion bonded articles are such that they may be superplastically formed without fear of peel fracture. At room temperature, peel strengths for stiffened liquid phase diffusion bonded specimens are comparable to their solid state diffusion bonded counterparts. The behaviour of the test pieces modified with reinforcing material indicates the potential for suitably adapted workpieces to undergo superplastic forming without the risk of joint peel compromising the strength of finished components.

Figures 4 to 8 show the method of the present invention applied tc a truss core sandwich of aluminium alloy sheets. This type of structure has been commonly formed from titanium alloys by superplastic forming/diffusion bonding techniques, but the peel

stresses encountered during the superplastic deformation stage have hitherto precluded the manufacture of such articles from diffusion bonded aluminium alloys.

In Figure 4 there is shown a three-layer arrangement of aluminium- lithium alloy sheets 30, 31 and 3 with intermediate strips 33 of reinforcing material placed between the outer skin sheets 30. 32 and the core sheet 31• The strips 33 are formed of a stiff material, such as a metal matrix composite of aluminium-lithium alloy reinforced with particulate silicon carbide, or of another stiffer alloy, or of a ceramic material. Barrier coatings 34 are applied where necessary to prevent diffusion bonding in undesired areas.

Figure 5 shows an alternative arrangement in which the outer skin sheets 40, 42 overlie a contoured core sheet 4l having integral reinforcements 43. The contoured core sheet 4l is made from a thicker sheet which is selectively machined or chemically milled to leave the desired pattern of stepped portions which form the reinforcements 43- Barrier coatings 44 are used as before in the Figure 4 example.

To effect diffusion bonding between the respective layers, the components are pressed together as shown in Figure 6 under appropriate temperature and pressure conditions. For 8090 aluminium-lithium alloy this typically involves initial pressing at 120MPa as the temperature of the bonding pieces is raised to 250°C, followed by a spell at 7MPa pressure whilst the temperature climbs to 500°C and finally pressing at 0.75MPa for a period of at least 1 hour at a temperature between 500 and 5δ0 ^β C. Although viewed in cross-section, the workpiece is shown here without hatched lines in order that the diffusion bonded areas depicted by dotted lines 35 can be more clearly seen.

The dies 36, 37 m y be contoured as shown to produce a profiled workpiece having kinked faces over the diffusion bonded regions which are to be superplastically formed. This profiling is an optional measure which reduces the gas pressure required to initiate deformation when the workpiece is subsequently presented to a further die for superplastic forming.

The workpiece is heated to a temperature of 515~530°C suitable for superplastic forming and then gas pressure is applied to the cavities P _j , P ₂ and P- in Figure 7 to form the workpiece into shaped die 38- An upper die member (not shown) maintains the sheet 30 in flat form

whilst the sheets 31.an (to a lesser extent) 32 undergo deformation. The increased resistance to peel due to reinforcing strips 33 concentrates the tensile stresses in core sheet 31 whilst inhibiting the effect of tensile peel stress components acting normal to the bond interfaces. This effectively means that superplastic strain is experienced in the core sheet without causing peel fracture at the bonded joints.

Figure 8 shows the truss core sandwich at the end of the superplastic forming operation, with outer layer 3 pressed firmly into the corners of shaped die 38. Figure 9 shows the finished component after removal from the die.

It should be noted here that the 90° T peel tests discussed above in relation to Figures 1 to 3 illustrate the worst possible case. In general, structures made by superplastic forming techniques will be similar to the truss core sandwich shown in Figure 3 , where the core is inclined to the bond plane at an angle 0 which is less than 90 ^c . This means that the tensile force which must be exerted by the core sheet to initiate peel in the truss sandwich is greater than that required in the 90° case:

Peel force = Tensile force x sin 0

This is true for both room temperature and elevated temperature situations. Various modifications may be made to reduce the weight of finished products. As indicated above, the reinforcing material may be formed with grooves extending normal or parallel to the edge of the diffusion bonded areas. Alternatively, or possibly in combination with such profiling, selected reinforcing strips 33 in Figure 4 can be omitted where their adhesion to the outer sheets 30 and 3 is prevented by barrier coatings 3^- In the Figure 5 embodiment, this selective omission would be effected by removing those stepped portions 3 which have overlying barrier coatings 44.

The production of titanium alloy sheet structures by superplastic forming/diffusion bonding has led to substantial reductions in manufacturing costs, not only because of the decrease in the number cf operations and tooling jigs required, but also because it enables

several components to be made simultaneously. Significant weight reductions are also achievable through the avoidance of rivets or similar fasteners. This has the added benefit of avoiding the stress concentrations which are associated with such fasteners, and also of minimising corrosion effects.

The present invention enables these advantages to be extended to diffusion bonded materials which were hitherto susceptible to peel fracture during superplastic forming, most especially to diffusion bonded aluminium alloy structures. Examples of the types of components which may be formed using this method of DB/SPF processing include inspection covers and doors for pressure cabins. These are structures which are exposed to considerable stresses in use, so their reliability in service must be beyond question. This reliability is assured by the provision of reinforcing material in accordance with the invention as defined by the following claims.