This invention relates to aluminium alloys containing lithium which are particularly suitable for aerospace construction and have been found to have improved damage tolerance in regard to their fatigue crack deviation behaviour.

In International patent application PCT/GB 90/01851 published under No. ¥0 91/08319, there is described and claimed a method of producing sheet or strip material of lithium-containing aluminium alloy having improved cold rolling characteristics preferably with improved damage tolerance. In that specification, fatigue crack growth in such alloys is discussed and fatigue crack deviation is mentioned. This is an important phenomenon in the consideration of an alloy's damage tolerant behaviour, especially in pressurised fuselage constructions, and is a particular consideration for lithium-containing alloys where the lithium content is relatively high, such as in that known commercially as "8090". At the present time no clear understanding exists of the causes of fatigue crack deviation on a macroscopic scale, and it is recognised that it is not possible to extrapolate an alloy's large scale cracking behaviour from, for example, its susceptibility to very small scale crack branching.

As used herein "fatigue crack deviation" refers to the change in mean direction of a propagating fatigue crack over a distance of several and sometimes tens of millimetres as compared with the type of microscopic crack branching which occurs across grains with a mean length of only a few tenths of a millimetre.

In the aforementioned PCT specification, a production method is described which has the purpose of producing sheet or strip material having improved cold rolling

characteristics. Specifically, there is claimed a method of producing sheet or strip material of improved cold rolling characteristics preferably with improved damage tolerance which comprises the steps of:-

(a) providing, in a condition suitable for hot rolling, a billet of an alloy of the composition in weight percent:-

lithium 1.9 to 2.6 magnesium 0.4 to 1.4 copper 1.0 to 2.2 manganese 0 to 0.9 zirconium 0 to 0.25 at least one other grain-controlling element 0 to 0.5 nickel 0 to 0.5 zinc 0 to 0.5 aluminium balance (except for incidental impurities)

wherein the other grain-controlling elements are selected from hafnium, niobium, scandium, cerium, chromium, titanium and vanadium, and wherein at least one of (i) manganese, (ii) zirconium and (iii) one of the said other grain controlling elements is present,

Cb) hot rolling the billet to produce an intermediate shape suitable for annealing,

(c) annealing the said intermediate shape at a temperature sufficiently high for the intermediate shape to be softened sufficiently to be subsequently rolled, and high enough for essentially no « precipitate to be formed, but not so high as to form any significant amount of C phase, and for a time sufficient to precipitate any soluble constituents therein to an extent sufficient to decrease significantly the extent of work hardening needed in step (d),

(d) cold rolling the annealed intermediate shape to an extent sufficient to cause an essentially fully recrystallised grain structure to be formed therein during step (e) and to produce a sheet or strip of the desired thickness, and

(e) rapidly heating and rapidly cooling the cold rolled sheet or strip material to produce an essentially fully recrystallised grain structure therein.

Where a production route based on an initial casting method is desired the described processing steps are as follows:-

1. The alloy is cast, preferably by the direct chill method, and then heated at a controlled rate to a temperature sufficient to relieve internal stresses caused by the cooling from melt of the molten alloy.

For the described preferred alloys, this is generally between 300 and 500°C, preferably between 300 and 400°C. During this heating, some precipitation of at least some of the constituents held in super¬ saturated solid solution may occur.

2. Either with intermediate cooling or following directly on from the heating step 1 , the stress- relieved billet is heated at a controlled rate such that the low melting point phases are substantially all dissolved without melting, and the billet homogenised by holding it at a temperature and for a time sufficient to dissolve substantially all of the soluble phases. The billet may then be cooled to room temperature and scalped.

3. The homogenised billet is then reheated generally to between 535 and 545°C and hot rolled, optionally with re-heating at intermediate stages, and optionally with hot widening, i.e. cross-rolling at elevated

-it-

temperature, to produce an intermediate shape suitable for annealing. If desired, the hot rolled metal may be heated to about 450°C in order to allow alteration of the distribution of the second phase particles to occur.

4. The hot rolled material is then annealed in order to precipitate any soluble constituents therein in order to reduce the extent of work hardening during cold rolling. For the described preferred alloys this is generally performed at between about 270°C and 350°C, preferably between about 270° and 325°C, and more preferably about 300°C, depending on the precise composition of the alloy used. The annealing temperature should be sufficiently high for the intermediate shape to be softened sufficiently to be subsequently rolled, and high enough for essentially no S ' precipitate to be formed, but not so high as to form any significant amount of C phase.

5. The annealed material is then cold rolled to its final thickness, optionally with inter-annealing usually between 270 and 350°C, such that sufficient cold work is imparted to the sheet or strip to cause a fine re-crystallised grain structure to be formed during solution treatment.

6. The cold-rolled sheet or strip is then rapidly heated to a suitable heat-treatment temperature, preferably in a salt bath, and rapidly cooled, preferably by water quench, in order to produce a solution-treated, fully recrystallised grain structure therein. It should be noted that this heat treatment can be done in two steps, the first step at a lower temperature of from about 450°C to below about 530°C in order to bring about recrystallisation and then a second step at about 530°C followed by water quench to solution

treat the sheet or strip. The heating step can be carried out. using a continuous heat treatment furnace, an air-recirculating furnace or by induction heating, but a salt bath is preferred.

7. Optionally recrystallisation can be performed again starting again from step 4 or from step 5.

8. The quenched sheet or strip is then if desired stretched and/or planished and then under aged, for example at about 150°C for 24 hours, to produce the finished product. Natural ageing may be possible for certain alloys depending on the particular combination of toughness and strength that is desired.

It will be noted that in accordance with conventional wisdom in order to improve the resulting alloy's damage tolerance underaging is described, particularly at about 150°C for 24 hours.

It has now been surprisingly discovered that if instead of under ageing the alloy it is aged near to, at, or over its peak aged condition followed by a relatively high temperature short time reversion treatment its susceptibility to fatigue crack deviation is significantly reduced.

In accordance with the present invention there is provided a method of producing a damage tolerant sheet of lithium- containing aluminium alloy having reduced susceptibility to fatigue crack deviation comprising providing a sheet of the said alloy with an unaged structure, ageing the sheet at a temperature and for a time near to, at, or over the peak aged condition for that alloy based upon its proof stress properties sufficient to increase precipitation of at least one planar slip blocking phase, heating the sheet

to an elevated temperature higher than its ageing temperature and holding the sheet at that temperature for a time sufficient to facilitate significant dissolution of the phase precipitates therein without significant dissolution of any planar slip blocking phase, and then cooling the sheet.

The heating of an aged alloy to a temperature higher than its ageing temperature for a relatively short time is generally referred to as a reversion treatment. Although there is discussed in an article by A.K. Vasudevan et al in Treatise on Material Science and Technology Volume 31, 15 at pages 445 to 462 and entitled "Fracture and Fatigue Characteristics in Aluminum Alloys" the effect of such reversion treatments on microscopic fatigue crack branching, there is no mention therein of the susceptibility of a lithium-containing alloy to macroscopic fatigue crack deviation, or recognition of the importance of the ageing treatment to which the alloy is subjected prior to the reversion treatment. Furthermore the authors were concerned only with fatigue behaviour at relatively low A K values, e.g. 4 to 10 MPav/πT, i.e. in the near threshold regime.

In EP-0273837 a number of thermal treatment methods are described and claimed which involve a reversion treatment of the type essential to the present invention. That specification is, however, concerned mainly with the improvement of an alloy's corrosion resistance and its greater isotropy of mechanical properties, and there is no recognition in that specification of the effect of such reversion treatments on an alloy's fatigue properties, and specifically its susceptibility to fatigue crack deviation.

Although not wishing to be bound by theory, it is believed that fatigue crack deviation is a consequence of planar

slip and crystallographic preferred orientation. The latter is considered to be significant, particularly that known as GOSS texture, but current sheet production routes inevitably produce such texture. Planar slip, on the other hand, is known to be encouraged by the presence of $ precipitate and is discouraged by the presence of precipitates such as S' phase and/or other phases in the structure. It has been found that standard ageing, i.e. under ageing, treatments result in the formation of high levels of & ' precipitate before an adequate amount of the at least one planar slip blocking phase, usually S', has been formed. The high levels of & formed by such conventional ageing are undesirable because, in addition to encouraging planar slip, this phase increases the strength and decreases the fracture toughness of the resulting alloy. In other words, when a standard ageing treatment is used, by the time there is adequate S' or similar phase to suppress planar slip, there is already so much present that the alloy's strength is too high and its toughness has fallen to an unacceptably low level. The reversion treatment of the present invention is therefore designed to produce a controlled amount of & while retaining the S' or similar phase in the structure. Two such similar phases are the T-, phase and the 0' phase.

In contrast to conventional under ageing, the present invention utilizes at least near peak under ageing, which generally means ageing to a proof strength of at least about 85% of peak aged proof strength. The optimum ageing treatment will depend upon the composition of the particular alloy used. The present invention is particularly applicable to control the fatigue behaviour of alloys at relatively high values of Δ K above the threshold region, e.g. between 10 and 50 MPa N/ΠT, especially between 15 and 35 MPa m.

The present invention can be applied to any lithium- containing aluminium alloy which exhibits a £ ' phase and

at least one planar slip blocking phase such as S' or T^ Examples of such alloys are those containing lithium 1.7- 2.8%, copper 1.0-3.0% and magnesium 0-1.9%, in addition to other components, such as the alloys designated as AA's 8090, 8091, 2090 and 2091.

The reversion temperature is from 200 to the £ ' solvus temperature which is about 300°C, the optimum reversion conditions depending on the composition of the particular alloy used and on its ageing treatment. Preferably the reversion temperature is from 210 to 270°C and for 8090 more preferably from 220 to 250°C and most preferably about 230°c. The holding time is generally from 10 seconds to 8 minutes preferably about 4 minutes. Preferably the sheet is aged at between about 170 and 190°C for at least 50 hours at the lower temperature, or at least 12 hours at the higher temperature prior to the reversion treatment of the present invention, and usually the sheet is stretched prior to ageing.

Preferably the alloy has a recrystallised structure, and the sheet is rapidly heated to its reversion temperature and is rapidly cooled thereafter. Generally the reversion treatment is carried out after cooling from the first ageing temperature to some intermediate temperature e.g. most conveniently, although not necessarily, room temperature, optionally holding for some time at the intermediate temperature. Alternatively it could be done without cooling directly following the first ageing treatment.

Although the alloy's susceptibility to fatigue crack deviation is significantly reduced by the described reversion treatment, it has been found that for some alloys a better balance of its macroscopic properties, including tensile strength, can be obtained by subjecting the treated alloy to a further ageing step at an elevated

te perature, but not so high as to allow significant precipitation to occur at the grain boundaries. Preferably this further and final ageing step is carried out at or below the first ageing temperature. Such secondary ageing treatment can improve the long term stability of the mechanical properties of the alloy.

Although the method of the present invention can be applied to any sheet of a lithium-containing aluminium alloy, such as described in EP-B-0088511 , EP-B-0124286, EP-A-0157711 and EP-A-0210112, it is preferred that the initial sheet be produced by a process as described in the aforementioned PCT specification. It will be appreciated that the method of the present invention is well suited to the use of a continuous heat treatment furnace, although batch treatment can be used for example in an air furnace or in salt baths.

Embodiments of the present invention will now be described by way of example with reference to the following Examples.

Example 1

A 1.6 mm sheet was prepared in the T3 condition from a cast alloy number E140, the composition of which is set out below in Table 1. It is known that this material in its standard damage tolerant temper, viz. aged for 24 hours at 150°C, is susceptible to fatigue crack deviation.

Table 1 Composition (wt%) Cast No. Li Cu Mg Zr Fe Si Ti Na

E140 2.34 1.24 0.98 0.06 0.07 0.04 0.017 0.0006

Tensile data was obtained at 170°c and 210°c and two peak aged conditions selected (82h at 170°C and 4h at 210°C).

The peak aged materials were then given a reversion treatment consisting of 1 min. at 270°C in a salt bath followed by a cold water quench. The water quench was used in order to minimize both matrix and grain boundary precipitation which might occur during a slower cool. It was noted that no distortion occurred during the water quench.

Tensile data was then obtained on the reverted materials and is shown in Table 2 below where it is compared with data for the standard damage tolerant temper and the specification minimum values. In the Table "AC" stands for "air cooled" and " Q" stands for "water quenched". The proof strength levels obtained in the reverted materials were acceptable in both starting conditions, but the tensile strength levels were low. The elongation values obtained were acceptable for the 170°C aged sample but rather low for the 210°C aged sample. For this reason only the 170°C aged material was selected for subsequent fatigue testing.

Five fatigue crack deviation tests were performed on the 170°C aged and reverted material, using 75 mm wide centre cracked panels in the T-L orientation. Three tests were performed at an P. ratio of 0.385 and two tests at an R ratio of 0.1, covering the range of A and K__ _v conditions where crack deviation occurs in susceptible material ( __„ = 28 MPa ^m ^" ). The fatigue cracks ran straight in all five samples with no evidence of any deviation.

In order to assess the relative toughness values of the different materials, notched tensile tests were performed and the NTS/YS ratio calculated, as described in ASTM Standard E338-81. Double edge notched tensile samples were prepared from 100 x 25 mm blanks. Each sample contained two 60° notches, 4 mm deep with a root radius of

0.013 mm. The samples were taken from the broken halves of the fatigue test panels together with standard tensile specimen blanks, all in the transverse direction. The tensile and NTS/YS values obtained are shown in Table 3 below. The tensile results are comparable with the tensile results shown in Table 2, except for marginally higher strength levels, and lower ductility values which resulted from the specimens failing outside the gauge length. The proof strength of the reverted material is slightly lower than the standard aged material and the tensile strength is significantly lower. The NTS/YS values are about 8% higher for the reverted material and show that the toughness of the reverted material is higher than the standard aged material.

Examination of the fractured notched specimens in the region of the notch showed a significantly greater amount of plasticity in the reverted material than in the standard aged material.

The microstructures of the materials were examined in the TEM and the results are summarized below.

Material aged 82 hours at 170°C

The matrix of this material contained uniform distributions of both fine $' and coarse S' precipitate. The grain boundaries showed narrow precipitate free zones (pfz's) and some precipitation of both I phase and phase.

Material aged 82 hours at 170°C + 1 minute at 270°C and Water Quenched

The matrix of this material showed a very fine uniform precipitation of $> ' resulting from the reversion treatment, plus some coarser undissolved & , presumably

resulting from incomplete dissolution at 270°C The material therefore contained a bimodal distribution of S . The distribution of coarse S' precipitate was essentially unchanged from the distribution observed prior to reversion. A small amount of another phase, however, was observed which has not yet been identified but is probably the T _* phase. The grain boundaries showed essentially the same amount of precipitation as observed prior to reversion but the pfz's were much wider.

Material aged 4 hours at 210°C + 1 minute at 270°C and Water Quenched

The material aged 4h at 210°C was not examined in detail but the reverted material showed the following features. The matrix showed a bimodal distribution of S ' , as described above for the 82h 170°C + 1 min 270°C material, and a uniform distribution of S' also similar to that observed in the 170°C aged material. There was, however, a greater amount of the unidentified phase present. The grain boundaries showed more precipitation than the 170°C aged material and wider pfz's.

Table 2

Ageing Treatment Dire- 0.2%PS TS El. NTS NTS/YS ction (MPa) (MPa) (%) (MPa)

24h 150°C AC

82h 170°C AC

+ 1 min. 270°C WQ

* Broke outside gauge length

The fracture toughness of the reverted material was then measured by generating R curves, as described in ASTM Standard E561-86, using 760 mm wide centre cracked panels in both L-T and T-L orientations. The R curves obtained for the two orientations were almost identical, showing that the material was highly isotropic, much more so than for the standard age. The reverted material also showed more stable crack growth before failure, a = 140 mm compared with typically A a = 70 - 100 mm for standard aged material. The R curves and hence toughness were

-Ill-

equivalent to, or slightly better than, the established data for similar standard aged material. The actual Kc numbers obtained were invalid due to net section yielding. The fatigue crack growth rate was then measured using a 160 mm wide centre cracked panel, in the T-L orientation, with an R value (load ratio) of 0.385. The fatigue crack ran straight; this confirmed the absence of crack deviation found in the 75 mm wide panels previously tested. The fatigue crack growth rate was higher at low K and lower at high A compared with similar material given the standard age. The higher growth rate at low was attributed to less planar slip. The lower growth rate at high A K was attributed to the higher toughness.

Example 2

In order to demonstrate the effect of peak ageing as opposed to under ageing on reversion treatment, samples of the E140 material of Example 1 were subjected to either peak ageing (PA) for 82 hours at 170°C or near peak under ageing (UA) for 32 hours at 170°C, before being subjected to reversion treatment in a salt bath at either 270°C or 250°C for times between one and eight minutes, followed by water quenching. The transverse tensile and NTS/YS properties were then measured.

The double edged notched tensile samples had a gauge width of 12.7 mm and contained two 60° notches 2.12 mm deep with a root radius of 0.018 mm.

Secondary ageing was then performed at 150°C for between one half and two hours and the samples' properties were then re-measured. The results are set out in Tables 4 (reversion temperature 270°C) and 5 (reversion temperature 250°C) below, and it will be noted that the secondary ageing times were selected so as to give a target proof strength of 300 MPa for the evaluation of final tensile properties and NTS/YS ratio.

Table 4

Ageing

PA: PA: PA:

UA UA UA

Ageing

PA: PA: PA:

UA UA UA

Example 3

In this Example the effect of reversion on fatigue crack deviation using a different material than that used previously (E140) was examined. The material selected was from heat treatment lot L0318 which has the following composition and characteristics:

Composition (wt%)

Lot Li Cu Mg Zr Fe Si Ti Na No.

L0318 2.34 1.19 0.77 0.06 0.05 0.04 0.026 0.0007

The material was in the damage tolerant temper (24h/150°C) and the sample had a thickness of 3.2 mm. Previous work had shown that in this condition it was susceptible to fatigue crack deviation.

The material was additionally aged for 82 h at 170°C (as used for E140) and then reverted for 4 min. at 270°C in a salt bath followed by a cold water quench. The 4 min. time was selected as a more commercially feasible time for this treatment than the 1 min. used previously. Tensile properties were measured on the as-reverted material and the following data obtained:

P.S.

Two crack deviation tests were then performed on the reverted material using 75 mm centre cracked panels in the

T-L orientation, at an R value of 0.385. The cracks ran straight on both samples. This confirmed the previous finding on E140 sheet that in the as-reverted condition the material did not show crack deviation.

Example 4

In this Example the effect of reversion temperature on fatigue crack deviation was examined for the cast alloy No. E140 of Examples 1 and 2. A total of eight panels were tested, each of which was aged and then subjected to a reversion treatment in either a salt bath or an air furnace. For most of the panels ageing at 170°C for either 32 or 82 hours was carried out, but for Panel 8 ageing for 24 hours at 150°C was used followed by a reversion treatment of 5 minutes at 190°C for the sake of comparison to show the effect of under ageing and too low a reversion temperature. Specimens from the panels were then tested.

For two of the panels, numbers 1 and 5, a secondary ;geing treatment was carried out to the reverted material, and again their susceptibility to fatigue crack deviation was tested.

The fatigue crack deviation test used in this and the other Examples was a standard centre cracked panel test in which there exists the possibility of two fatigue crack deviations per sample being formed, one from either end of the centre crack. The results are presented in Table 6, and it will be noted that the reversion treatment of the present invention is effective in suppressing fatigue crack deviation where the reversion temperature was at least 230°C. Reversion at 210°C and 190°C resulted in some deviation, particularly for the lower temperature.

Panels 1 and 5 were subjected to a secondary ageing treatment with half of the specimens of Panel 5 being subjected to the same secondary ageing treatment as for Panel 1, i.e. 1 hour at 150°C, whilst the other half of the specimens were subjected to secondary ageing at 70 C for 672 hours. The results of the centre cracked panel tests for these secondary aged specimens are set out in

Table 6 and show that for the particular heat treatments used secondary ageing of the reverted material resulted in crack deviation starting to return.

Table 6

PANELS AGEING REV.TIME/ FRACTION SEC. AGE FRACTION _h/ o _C TEMP. DEVIATING TIME/TEMP. DEVIATING

(min/°C) (CRACKS/ (CRACKS/ NO. OF (h/ ^u C) NO. OF SPECIMENS) SPECIMENS)

REVERSION IN SALT

1/150 1/5

1/150 2/5 672/70 2/5

7 32/170 4/230 0/5 8* 24/150 5/190 2/4

*Mot of the present invention. Example 5

Correlating all of the data obtained from the mechanical testing of the various specimens, Table 7 presents the effect of reversion temperature on the tensile properties of the alloy tested.

32/170 4/230 312 407 8.3 1.30

On the basis of a minimum specification for this alloy of PS 280 MPa, TS 410 MPa and El. 9%, it can be seen that reversion should be carried out at 230°C. For those reversion conditions, the effect of secondary ageing on the tensile properties is presented in Table 8.

Table 8

AGEING REV. PROOF UTS El. MTS/YS SECONDARY AGE PROOF UTS El. NTS/YS _fh/ o _p TIME/TEMP. STRESS (MPa) (%) TIME/TEMP. STRESS (MPa) (%) ' (min/°C) ^(MPa) (h/°C) ^(MPa)

32/170 4/230 308 411 10.3 1.31 1/150 341 444 7-7 1.19

672/70 328 452 8.1

For this alloy of cast No. E140, Table 9 sets out its mechanical properties under different ageing conditions and presents its transverse tensile proof stress as a percentage of that achieved by peak ageing the alloy.

Table 9

Aged for 32 hours at 170°C:

Age Trans 0.2% Percentage of ( /°C) ^{PS MPa) peak} °' ² ^ ^ps

Peak Age 82/170 392 100

Near peak Under Age 32/170 343 87.5 Conventional Under Age 24/150 318 81.1