The alloy 3003 is recorded in the May 1987 Registration

Record of the Aluminum Association Inc as having the composition in wt %:-

Si up to 0.6 Fe up to 0.7 Cu 0.05 - 0.20

Mn 1.0 - 1.5 Zn up to 0.1 Although AA3003 is well known as an extrusion alloy, it is much more widely used for rolling to sheet form. This invention is concerned with a particular alloy composition, within or close to the edges of the AA3003 specification, which is particularly well adapted for extrusion. While all AA3003 alloys can be extruded, the alloys with which this invention is concerned can be extruded faster, at lower extrusion pressures and with improved extrudate surface finish. In one aspect the invention provides a method of making an extrusion billet, which method comprises providing a billet of composition in wt %

Si 0.15 - 0.6 preferably 0.20 - 0.30 Fe up to 0.5 preferably 0.10 - 0.35 Cu up to 0.3 preferably 0.05 - 0.20

Mn 0.8 to 1.2 preferably 0.9 - 1.2 Zn up to 0.1

Others up to 0.05 each, 0.15 total, Al balance heating the billet at a homogenisation temperature of at least 590°C,

and cooling the homogenised billet to 300 - 400°C at a cooling rate of 10 - 250°C per hour.

By virtue of the alloy composition and homogenising conditions, the starting extrusion billet contains dispersoids of composition Al (Fe Mn) Si (provided the alloy contains Fe). These generally have a predominantly cubic α-crystal structure and are approximately equiaxed i.e. generally spherical. By dispersoids we mean intermetallic particles formed by a solid-state reaction for example during homogenising. The mean size of dispersoids in extrusion billets of this invention is typically in the range of 0.25 - 0.5 μm.

In another aspect the invention provides an extrusion billet of composition in wt %

Si 0.15 - 0.6 preferably 0.20 - 0.30 Fe up to 0.5 preferably 0.10 - 0.35 Cu up to 0.3 preferably 0.05 - 0.20

Mn 0.8 to 1.2 preferably 0.9 - 1.2 Zn up to 0.1

Others up to 0.05 each, 0.15 total, Al balance. Preferably the microstructure exhibits dispersoids containing Mn at a dispersoid inter-particle spacing of at least about 1 μm in the bulk of grain interiors.

For the purposes of this specification, dispersoid inter-particle spacing is measured by the following technique. Samples taken from the half radius positions of transverse slices of the homogenised billet are examined by optical microscopy at a magnification of 500x. To reveal the dispersoids, the samples are prepared by mechanical polishing and then etching in a solution of 0.5 vol.%HF in water at room temperature for 30 seconds. Dispersoid spacing is then measured by the standard linear intercept technique. It is recognised that this technique does not reveal all

the dispersoids and may exaggerate the size of the dispersoids. Nevertheless, it is a quick means of characterising an important feature of the billet microstructure.

Preferably the extrusion billet is of composition in wt.% Si 0.20 - 0.30

Fe 0.15 - 0.35

Cu up to 0.3

Mn 0.9 - 1.2

Zn up to 0.1 Others up to 0.05 each, 0.15 total,

Al balance.

Preferably the extrusion billet is cylindrical, in contrast to billets for rolling which are generally of rectangular cross-section, although there is no reason in principle why billets of square or rectangular or other cross section should not be used.

These extrusion billets extrude faster and are easier to extrude, permit the use of lower extrusion pressures and/or colder extrudate exit temperatures, and produce extruded sections having much better surface finish than currently supplied billets of AA3003. These factors combine to allow faster extrusion speeds, normally limited by surface finish, and thus significant improvements in productivity.

The Si content of the alloy is specified as 0.15 to 0.6 preferably 0.20 - 0.30. This is rather higher than is usual in AA3003 alloys. As demonstrated in the examples below, a Si level of at least 0.15 or 0.20 is helpful in reducing extrusion pressure. But there is some equivocal evidence that Si contents above 0.3 may be less advantageous, or even disadvantageous, in this respect.

The Fe level is specified as being up to 0.5 preferably 0.10 or 0.15 to 0.35 wt %. This is below the level in commonly used AA3003 alloys. A low Fe level contributes to reduced extrusion pressure and

improved extrudate surface finish. There is no critical lower limit. But the price of Al alloys progressively increases as the Fe content is decreased, especially for remelt billet, and it is not economic to use metal with a very low Fe content. In the above phrase "extrusion alloy" the word "extrusion" indicates that the alloy is suitable for extrusion on a commercial scale in an economically viable manner. The composition and treatment herein described allow improved extrusion performance in terms of both speed and surface finish. The parameter that most strongly influences the extrusion behaviour is the inter-particle spacing of the dispersoids. This in turn depends on the amount of Mn precipitated and the way in which precipitation occurs. During the homogenisation soak, Mn dispersoids undergo a process of Ostwald ripening in which the finer precipitates dissolve and the coarser ones grow larger. This increases the inter-particle spacing ofthe dispersoids. The effect is greater as the temperature is raised and the homogenising time increased. A rather large inter-particle spacing is required. If the dispersoids spacing is too small, the flow stress of the billet (as measured by hardness or high temperature flow stress) is high and the resulting extrudability is too low. This is the situation that occurs when the homogenising temperature is below about 590°C.

On the other hand, the inter-particle spacing resulting from Ostwald ripening greatly affects the re-precipitation of Mn during cooling. Mn most readily re-precipitates on existing dispersoids. If the dispersoids are close enough, the Mn coming out of solution on cooling will precipitate on existing dispersoids coarsening them but not greatly affecting the inter- particle spacing. If the spacing between the dispersoids is too great, some Mn will be unable to diffuse to the closest dispersoid and will form new precipitates at other sites in the matrix. This new precipitation is fine, closely spaced and greatly increases the room temperature hardness and

the hot flow stress of the billet. Extrudability is reduced.

Arising out of this, the billet is held at a homogenisation temperature of at least 590°C, preferably 600 - 630°C for a time to coarsen dispersoids containing Mn and to increase the dispersoids inter-particle spacing to at least about 1μm in the bulk of grain interiors. Then the homogenised billet is cooled to 300 - 400°C at a cooling rate of 10 -250°C per hour, preferably 25 - 200°C per hour, chosen to cause Mn to precipitate on to the existing dispersoids with substantially no precipitation at the new sites. The maximum permissible rate of cooling is thus related to the dispersoids inter-particle spacing, being lower at greater inter- particle spacings. There results a homogenised extrusion billet containing coarse dispersoids at an inter-particle spacing of at least about 1μm in the bulk of grain interiors, with no or substantially no intervening fine Mn- containing dispersoids. In general, the coarse dispersoids are large enough to be revealed by etching and examination in an optical microscope, while fine dispersoids, should any be present, would be detectable only by other methods.

The resulting homogenised extrusion billet has a low Vickers Hardness, preferably a hardness less than 31.5 VPN measured at room temperature. Room temperature hardness is a good guide to extrusion flow stress in this alloy, because the same strengthening mechanisms exist both as regard to hardness and extrusion pressure, i.e. predominantly inter-particle spacing. Extrudability generally declines when the hardness is greater than 31.5 VPN. The billets herein described are suitable for extrusion by virtue of their microstructure. Although this invention is based on results not theories, the following theoretical comments may be of interest.

Major features of the microstructure are the Mn and Cu in solid solution and the spacing of Mn bearing dispersoids. These are influenced by composition and homogenising conditions, heating rate, soak

temperature, soak time and cooling rate afterwards. The effect of reducing the Fe content is mainly to lower the volume fraction of coarse cell boundary particles (FeMn)AI ₆ or cubic α AI(Fe Mn)Si, but also to increase the Mn solid solubility. The coarse Fe bearing particles are not expected to be of sufficient numbers to affect flow stress. Nevertheless a lower Fe content has been found to reduce flow stress which reduces the extrusion pressure. In contrast, more Mn in solid solution, measured by a decrease in electrical conductivity, may be expected to increase the extrusion pressure and hence counterbalance the effect of coarse particles. This is addressed by tightening of the Si content which requires a minimum level, as shown above, to aid extrudability. Fewer coarse particles has an additional benefit of substantially improving the surface finish of extruded sections. Possibly the reduction in coarse particles in a cell boundary network may aid the break up ofthe billet microstructure required for easy flow in the extrusion container. In any event the combination of these two features, lower pressure and better surface allows the alloy to be extruded at speeds substantially higher than current AA3003.

Extrusion pressure is principally determined by three microstructural features:- i) Solid solution content (Mn and Cu) ii) Dispersoid size and spacing of αAI(Fe Mn)Si, iii) Volume fraction and morphology of coarse Fe containing particles. Laboratory tests using programmable furnaces, conductivity measurements, optical metallography and hot torsion testing have highlighted the homogenising conditions of importance and the microstructural features being controlled. Differences exist between the standard 3003 alloy and the preferred low Fe variant, particularly at the higher soak temperatures that are preferable. Whilst the degree of coarse particle spheroidisation is the same, considerably more dispersoids are

present in the new alloy making the cooling rate after homogenisation vitally important. The interparticle spacing of the dispersoids is thus the principal feature that alters extrusion pressure as a function of homogenising. This feature is controlled by particle volume fraction and size, both of which can change during homogenisation. Semi-quantitative analysis of optical micrographs indicates that ingots which contain spacings of greaterthan 1 μm in the bulk of the grain interiors (this excludes the dispersoid free region at cell boundaries) are preferred. In contrast, the main feature discriminating between 3003 and the preferred low Fe variant is the coarse Fe particle volume fraction which is much lower in the new variant. The final microstructural feature, Mn in solid solution, varies with alloy composition and homogenisation but the preferred range, in conjunction with the dispersoid spacing, is 0.4 to 0.7 wt % Mn (as measured by electrical conductivity 35-44% IACS). High temperature homogenisation coarsens the Mn bearing dispersoids thus reducing flow stress. This may be more critical at low Fe levels, perhaps because there are then fewer Fe containing particles to absorb Mn during homogenisation, resulting in more Mn dispersoids in the alloy. Si encourages the precipitation of Mn from solid solution thus reducing the solid solution strengthening. However this has to be balanced against any extra dispersion strengthening which may result from the additional Mn bearing particles.

Homogenisation of these extrusion billets is generally necessary, preferably at a temperature of at least 590°C. Preferred homogenisation conditions involve a billet heating rate of 10 - 1000°C per hour, preferably 25 - 500°C per hour, a hold temperature of 600 - 630°C, a hold time of 1 - 24 hours, preferably 2 - 12 hours.

Cooling from the soak temperature down to about 300 - 400°C is preferably effected quite slowly, e.g. at 10 - 250°C per hour

preferably 25 - 200°C/hr e.g. 50 - 150°C /hr, in order to establish the required metallurgical structure. The cooling billet can be held at 400 - 500°C for 1 - 24 hours, preferably 2 - 10 hours, to achieve an equivalent effect. The cooling rate from about 300 - 400°C to ambient temperature has no material effect on the alloy microstructure and may be effected fast to reduce furnace dwell time.

By contrast, cast AA3003 sheet alloy would normally be homogenised prior to rolling to form sheet. A long homogenising treatment is used with the objective of producing a coarse Mn containing dispersion distribution which allows the rolled sheet to readily recrystallise to form equiaxed grain about 20-40 μm in diameter. Typically the treatment involves holding at a high temperature around 600°C for about 24 hours followed by holding at about 500°C for about 12 hours prior to hot rolling. By contrast, the relatively short homogenising treatment in the present invention results in a much finer Mn containing dispersion. Extrudability cannot be predicted or inferred from the rolling characteristics of an alloy. The improved extrusion properties of homogenised billets according to this invention are believed surprising in the light of the known properties of AA3003 sheet alloys. These alloys are hot extruded, preferably at higher extrusion speed and/or lower extrusion pressure, to yield extruded sections which preferably have improved surface finish. Preferred extrusion conditions involve a billet temperature of 400 or 420 to 510° and a container temperature of 400-460°C. As a result of the rather high homogenisation temperatures employed, the extruded section is generally fully recrystallised.

Reference is directed to the accompanying drawings in which:-

Figure 1 is a graph of peak extrusion pressure against Si content. Figure 2 is a graph of peak extrusion pressure against Fe

content. Both graphs show the effect of ingot composition on extrudability.

Figure 3 comprises a pair of bar charts showing extrusion pressure against ingot composition and homogenisation temperature.

Figure 4 is a graph showing the effect of homogenisation temperature on Mn solid solution content.

Figure 5 is a graph showing the influence of cooling rate on billet microstructure.

Figures 6A and 6B are bar charts showing the effect of homogenisation conditions on hot flow stress of (a) a standard alloy and (b) a low Fe variant.

Figure 7 is a bar chart showing extrusion loads as a function of alloy and homogenisation.

Figure 8 is a graph showing effect of alloying on extrudate tearing using flat bar. The following examples illustrate the invention.

EXAMPLE 1

Alloys having the composition shown in Table 1 were DC cast and homogenised at 610°C for four hours. Heating rate to homogenising temperature was 100°/hr, average cooling rate to room temperature 2257hr.

Extrusion of the homogenised billet was carried out on an experimental press using the following conditions: billet temperature 450°C relevant ratio 26:1 ram speed 20 mm/sec extrudate air cooled.

The extrusion pressure was measured throughout the extrusion cycle and the data used to calculate the peak and the steady state extrusion pressure for each alloy. Peak and steady state pressures

are dependent on the compositions. Peak pressure is reduced by adding Si up to about 0.15% and remains constant thereafter within the range examined. By contrast, adding Fe above 0.5% increases the extrusion pressure. The conductivity of the as cast and the homogenised billets was measured. A low conductivity implies high Mn and/or Cu in solution. The results are set out in Table 1. Homogenising substantially reduces the Mn in solid solution. With reference to Table 1 :

A: Low Mn giving low Mn in solution and hence low extrusion pressure.

B: mid range alloy.

C: High Mn and Cu in solid solution, but low Fe giving a low volume fraction of coarse Fe particles, hence low extrusion pressure.

D: Low Mn + Cu in solution, but high Fe giving high volume fraction of coarse Fe particles, hence high extrusion pressure.

E: Low Si therefore low dispersoid density i.e. high Mn in solid solution which increases the extrusion pressure.

F: High Si therefore high dispersoid density i.e. low Mn in solid solution which decreases the extrusion pressure. Some of this data is used in Figures 1 and 2 to show the effect of Si content and Fe content respectively on peak extrusion pressure.

Table 1

Our Composition wt % Extrusions Press As Cast Homogenised Designation Si Fe Cu Mn Steady Peak Conductivity % Mn in Conductivity %Mn in State Pressure solution ^1,2 solution ^{1 2}

MPa MPa IACS% IACS%

A 0.16 0.4 0.09 0.78 194 228 34.6 .73 41.9 .49

B 0.19 0.39 0.12 0.98 191 234 31.2 .84 40.1 .55

C 0.22 0.2 0.1 1 196 234 30.2 .87 40.5 .54

D 0.2 0.6 0.13 0.98 199 248 31.2 .84 41.9 .49

E 0.09 0.4 0.1 1 205 250 31.1 .84 39.2 .58

0.29 0.39 0.09 0.97 189 232 31.5 .83 44 .42

^F

¹ Calculated from: %Mn = 1.86-0.0326IACS%

This equation was derived for binary alloys but believed to be reasonably accurate for the present case.

^! All Cu assumed to be in solution after homogenisation and not to contribute significantly to conductivity.

EXAMPLE 2

Alloys having the composition shown in Table 2 were homogenised using two conditions as set out in Table 3.

Table 2: Chemical Compositions of Billets

Alloy Si Fe Cu Mn Zn Tl

G Standard 0.23 0.58 0.085 1.07 0.015 0.011

H Low Fe medium Si 0.22 0.20 0.10 1.00 0.018 0.015

I Low Fe high Si 0.29 0.20 0.08 1.05 0.016 0.010

Table 3: Homogenising Conditions

Heating Rate Hold Cooling Rate

Low Temperature 100°C/hr 585°C 2 hours 350°C/hr

High Temperature 100°C/hr 610°C 4 hours 225°C/hr

Extrusion ofthe homogenised billet was carried out on an experimental press using the following conditions: Billet temperature 450°C Extrusion ratio 72:1 Ram speed 10 mm/sec Extrudate air cooled. The high temperature homogenising treatment results in a substantial reduction in the extrusion pressure of all of the alloys. The effect is greater in the alloy having the preferred composition suggesting

that the combined effects of homogenisation and composition act together.

The peak and steady state extrusion pressures were measured as before and are presented in Figure 3. The data confirms the effect of composition on extrusion pressure. Low Fe (H and I) results in a reduction in the extrusion pressure. After the high temperature homogenisation, however, the higher Si level (I) appears to increase slightly the extrusion pressure, suggesting that there is an upper limit to the Si content for optimum extrudability.

EXAMPLE 3

The alloys listed in Table 4 were DC cast to 18 cm diameter billets. All billets were homogenised by heating to temperature at a rate of 50°C/hour and holding for 4 hours at between 610 and 630°C followed by cooling to room temperature at an average rate of 150°C/hr.

Table 4: Chemical Compositions

Alloy Variant Si Fe Cu Mn Ti

Low Fe J 0.22 0.21 0.082 1.03 0.013

Standard K 0.23 0.59 0.081 1.04 0.017

Extrusion was carried out on a commercial press using the following conditions:

Billet temperature 470 - 482°C Container temperature 420 - 430°C. Two shapes were chosen for the trial both produced on a single hole die: 1. Round tube (1.5 cm diameter) coiled at the press. Extrusion speed was limited by the surface quality produced.

2. Flat section cut to length at the press.

The extrusion pressure during the press cycle and the cycle time for each billet was recorded. The extrudates were assessed in terms of surface finish (visual and profilometry).

Typical extrudate speeds for the round tube were 72 - 81 m/min. It was noted during the trial that the faster speeds were observed with the low Fe variant J and this is corroborated by shorter cycle times for this alloy.

A 2% reduction in extrusion pressure was obtained when extruding round tube in the low Fe alloy J compared with the standard alloy K. A 4% reduction was achieved with the flat section.

Two batches of billet having compositions close to K and J and produced on a full plant scale were homogenised as shown in the Table 5 below. After cooling to room temperature, the hardness of the billets was measured using a Vickers Hardness Testing Machine with a 10- kg load.

Table 5

Hardness VPN ₁₀

Condition

Alloy K Alloy J

As Cast 39.2 38.5

0hr/630°C 33.6 33.6

2hr/630°C 31.6 31.4

4hr/625°C 28.9 29.1

After homogenising, the alloys were cooled to 300 - 400°C at 75°C/hr and then fast cooled. The Mn-containing dispersoid inter-particle spacing of billets homogenised for 2 hours or 4 hours was greaterthan 1μm.

EXAMPLE 4

Extrusion billets, designated low Fe (J) and standard (K) in Example 3, were homogenised as in Example 3 and extruded in a commercial press using the following conditions: Billet temperature 465 - 500°C

Container temperature 438°C

Extrusion ratio 82:1

Extrusion speed 67 - 102 m/min

Extrudate water quenched Extrudate: 20mm OD round tube produced on a 6 hole die.

Under these conditions, the extrusion pressures required to extrude both alloys are approximately the same. However, the extrusion speed of the low Fe alloy J could be increased to 102 m/min whilst retaining an excellent surface finish. By comparison, the maximum speed at which the standard alloy K could be extruded, whilst maintaining a good surface finish, was 67 m/min.

EXAMPLE 5

A 27 mm OD tube was extruded from the same alloys as in Example 4 using the following conditions:

Billet temperature 465 - 500°C

Container temperature 423°C

Extrusion ratio 70:1

Die: three hole Extrusion speed 85 m/min for both alloys (this is the maximum speed for the press with this die).

The standard alloy (K) gave very poor surface finish under these conditions especially on the bore of the tube. The low Fe alloy (J) gave an extremely good finish both on the exterior and on the bore of the tube. Surface finish, especially on the interior of the tube, is particularly

important in this application because the tube is intended to be cold drawn to a smaller diameter. Surface defects interfere with the drawing process and result in fracture ofthe tube during drawing.

EXAMPLE 6

Three alloys were DC cast for laboratory evaluation to the following compositions:

Alloy ID Si t % Fe wt % Cu wt % Mn wt %

L Low Fe 0.23 0.20 0.086 1.05

M Standard 0.23 0.55 0.076 1.02

N Low Fe + Cu 0.20 0.19 0.15 1.02

The extrusion billets were homogenised using a variety of conditions and the suitability for high speed extrusion assessed by several means; microstructural evaluation, 'traditional' hot torsion testing (see Mechanical Metallurgy by G E Dieter, publ. McGraw-Hill 1976, p378) and direct extrusion. The choice of homogenisation practice influences the three microstructural parameters discussed previously; spheroidisation of constituent particles, spacing of dispersoids and Mn solid solution content. This influence also depends on the alloy choice, i.e. standard composition M versus the new alloys L, N. Figures 4 and 5 show the effect of homogenisation soak time and cooling rate after soaking on the Mn solid solution content. Figure 4 shows Mn in solution after cooling to room temperature at 100°C/hr. A difference in behaviour between alloys can be seen at high temperatures with the low Fe variants L and N continuing to dissolve dispersoids and return Mn to solution as the temperature is raised.

The standard alloy M exhibits a maximum solute content due to the partitioning of Mn to the high number of constituent Fe particles.

The amount of Mn removal during cooling depends on the cooling rate and this can be seen in Figure 5 where the low temperature homogenisation was at 580°C and the high temperature homogenisation was at 625°C. The low Fe alloy L is more sensitive to cooling rate and, with the slowest rates, it responds by removing more Mn. Thus by a combination of high soak temperature to spheroidise constituent particles and coarsen dispersoids, followed by slow cooling to remove Mn from solid solution, the desirable billet microstructure can be produced, with Mn- containing dispersoids present at a dispersoid inter-particle spacing of greater than 1 μm in the bulk of grain interiors. The benefit of the correct microstructure can be seen in extrudability experiments.

Hot torsion testing at 450°C demonstrates one aspect of the desirable microstructure achieved through processing and composition in reducing the flow stress of the alloy/billet. Figure 6 shows the effect of homogenisation conditions on torsion data and also allows a comparison between alloys. The softest microstructure is achieved by the alloy as part of the invention, i.e. low Fe figure 6b, in combination with a coarse dispersoid distribution achieved by high temperature homogenisation and slow cooling.

The critical dispersoid spacing required depends on the subsequent cooling rate. Slow cooling at 100°C/hr reduces the flow stress for all homogenising treatments. Under these conditions the hot flow stress falls with increasing homogenising temperature Figure 6b. The critical dispersoid spacing is close enough to allow substantially all the Mn to precipitate on existing dispersoids. However, if the cooling rate is increased to 350°C/hr a closer critical spacing is necessary to allow all the Mn to precipitate on dispersoids. Homogenising at 610°C provides the appropriate spacing. By contrast, homogenising at 625°C takes a greater

amount of Mn into solution which, combined with the greater dispersoid spacing, encourages Mn precipitation at new sites with a consequent increase in flow stress. Cooling at 600°C/hr results in a high flow stress at all compositions and homogenising treatments. This implies that the dispersoids are precipitating on new sites whatever the initial dispersoid spacing.

This work enabled a small confirmation trial to be conducted using extrusion of the 3 alloys L, M and N which were homogenised using two practices a) and b) derived from the above results. These homogenising practices were: a) hold at 620°C ± 10°C for 4 hours followed by cooling at 150°C/hr and b) hold at 620°C ±10°C for 4 hours followed by cooling at 75°C/hr to 330°C and then rapidly cooling to ambient. Figure7 shows the average extrusion loads recorded for several billets from the alloys noted above. The data demonstrate that the lowest peak load is achieved by the new alloy and modified practice which has developed the desirable microstructure discussed above. Figure 8 demonstrates that the benefits are not simply seen in extrusion loads but also in surface finish. In this experiment the alloy/homogenisation conditions were tested over a range of extrusion temperatures and speeds to develop a limit diagram. The advantage ofthe new alloy over existing billet is clearly demonstrated. Both figures also indicate that the control of Cu content is also important with the higher Cu in the low Fe variant significantly affecting extrudability.