BEVERAGE CAN BODY SHEET FROM CONTINUOUSLY CAST ALUMINUM ALLOY

Abstract

A continuous casting method and apparatus are described that generate alpha-phase Al(Fe,Mn)Si intermetallic particles on the surface of an aluminum alloy strip to facilitate the manufacture of beverage cans. The alpha phase particle size is generated so that subsequent rolling will ensure the particles will provide galling-free operation in the ironing stages of the can body making process. The control of alpha-particle size upon the surface layers of the strip ensures galling-free operation.

Description

FIELD OF THE INVENTION

The present invention relates to production of beverage can body stock using continuously cast aluminum feedstock which improves the economy, efficiency, recycle content and reduces emissions of the manufacturing process.

PRIOR ART

Beverage cans present the largest market sector for aluminum by volume. The cans are produced from wide sheet stock by a process in which the sheet is first blanked into a circular configuration and cupped in a single operation. The sidewalls are then drawn and ironed by passing the cup through a series of dies of progressively smaller bores. The dies thus “iron” out and lengthen the sidewall to produce a can body thinner in its walls than its bottom. The resulting cylindrical can body provides a shape yielding high strength using minimum metal.

Aluminum can body stock is currently produced by large and sophisticated machinery in a few dedicated rolling plants. Such massive high technology machinery represents a significant cost factor both from a capital investment and an operating cost perspective. Many batch processes are employed in a sequence of separate steps. In the typical case, a large ingot of 10 to 20 tons and approximately 650 mm thick is cast and cooled to ambient temperature. The ingot is then stored for inventory management. Prior to further processing, the ingot's surfaces are machined (“scalped”, in the vernacular) to remove defects such as segregation, pits, folds, liquation and handling damage. It is then heated to a required homogenization temperature for several hours to ensure that the components of the alloy are uniformly distributed through the metallurgical structure and to promote the formation of dispersoids of Al₆FeMn within the bulk. It is then cooled to a lower temperature suitable for hot rolling. The ingot is subjected to breakdown hot rolling in a number of passes using reversing or non-reversing hot rolling mill stands. After breakdown hot rolling, the slab is typically transferred to a tandem mill for hot rolling to finish hot line gauge, which is nominally 3 mm, after which the sheet stock is coiled. The hot rolling speed and operating parameters are usually set up so that the coil temperature is sufficient to cause a “self-anneal” during the air cooling of the coil. The coil may alternatively be annealed in a furnace in a batch step. The coiled and annealed sheet stock is then further reduced to final can stock gauge (about 0.2 mm) by cold rolling using unwinders, single and/or tandem rolling mills, and rewinders. The coil is then typically stabilize-annealed, slit to final width, oiled and packaged for shipment. The above-described process is known as the conventional or ingot method for making can body stock.

The many batch processes and the associated material handling operations in the ingot method are labor and energy intensive and demand significant processing times. It is typical for the entire process, from molten metal to finished can stock, to take three weeks to complete. Further, scrap is generated in most of the foregoing steps, in the form of scalping chips, end crops, edge trim, scrapped ingots and scrapped coils. Total loss ranges from 25% to 40% and represents a significant cost of the manufacturing process.

Continuous casting (cc) of aluminum strip or slab can eliminate a substantial number of steps and time from the production path and reduce losses, energy consumption and emissions. In one method, described in U.S. Pat. No. 4,238,248 assigned to Swiss Aluminum Ltd., for example, the casting is performed by continuous slab casting (12-25 mm thick) followed by hot rolling, coiling and cooling. Thereafter, the coil is annealed and cold rolled to finish gauge. In a second method, as described in several US patents such as U.S. Pat. Nos. 5,470,405, 5,514,228, 5,655,593, 5,769,972, 5,772,799, the strip is cast thin enough (about 2.5 mm) for processing to final gauge by in-line rolling. This method includes an anneal (typically a rapid or “flash” anneal) before the final cold rolling steps.

These previous attempts to make can body sheet from cc feedstock have not reached successful commercial application. Two general issues have appeared with sheet processed from cc feedstock. One is the excessive earing exhibited in the ironing operation when using continuous cast sheet. This issue, however, can be resolved by adjustment of annealing conditions, the amount of subsequent cold rolling and alloy composition. The second issue, referred to as “galling” or “scoring”, is observed with continuous cast sheet in the ironing steps of the can making process, and this issue has proven to be more problematic. Debris collects and builds up on the ironing dies and leads to unsightly scratches and other flaws on the outer surface of the can. Additionally, this galling can cause failures and jamming of the body maker equipment, shutting down production and resulting in significant scrap losses. Galling has been attributed to the finer size of the intermetallic (IM) particles in cc feedstock compared to those present in the conventional ingot path, which capitalizes on the large intermetallic dispersoid particles formed during the homogenization process. The finer intermetallic particles of continuous cast products are not able to act as “scrubbing agents”, a function performed by the larger particles present in the conventional material. The galling effect is observed relatively early with cc feedstock whereas coils from conventional ingot path can run through many hundreds of thousands of cans before the dies need to be re-dressed or changed. This galling issue in continuous cast can body stock remains unresolved.

Two families of solutions have been proposed and studied in detail for making larger particles in the surface layers of continuous cast feedstock to provide galling resistance. One method involves the homogenization of the metal at high temperatures to grow the intermetallic particles by diffusion. In this method, described by Hitchler and Klein in U.S. Pat. No. 4,111,721 (assigned to American Can Company), treatment is carried out on the as-cast sheet, though it may alternatively be carried out on partially rolled sheet providing there are enough subsequent cold rolling passes to reach the required strength level in the finished product. Temperatures greater than 480° C. and lengthy thermal durations of 4 to 24 hours are required to be performed on coiled product. A thermal treatment at a temperature of 538° C. for 10 hours is recommended in U.S. Pat. No. 4,111,721, with temperatures around 620° C. for 16 to 24 hours considered even more advantageous. Tests on 3004 alloy showed that the thermal treatment caused the Al6FeMn intermetallic particles to grow from a size range of 0.1-1 μm present in the as-cast material to the 2 to 3 μm range needed for gall-free drawing and ironing operations. The inventors were not specific about the method of continuous casting but gave the as-cast gauge of the strip to be between 5 to 25 mm. This range would cover any one of the three cc methods in use in the aluminum industry: conventional roll casting (5-12 mm), block cast slab (19 mm) and twin-belt cast slab (18-25 mm). In a variation of this method, U.S. Pat. No. 4,872,921 describes an oxidation procedure for introducing magnesium (Mg) and aluminum (Al) oxides to the surface of the sheet. For this purpose, the strip, in as-cast state or after the first step of rolling, is first chemically etched and then oxidized in a furnace at 350° to 400° C. for 1 to 2 hours, during which time crystalline oxides of Al and Mg grow on the surface. These particles are 2-15 μm in diameter and less than 5 μm thick. Oxidation conditions are selected so as to cover 10 to 25% of the surface with such oxide particles. After the sheet is rolled to final gauge, the particles perform the desired scrubbing action during ironing steps of the can making process.

In a second approach, a thin liquation layer is formed on the surface of the as-cast strip to serve the same purpose as the coarser intermetallic particles. Such layers are known to be rich in solute and in intermetallic particles. In U.S. Pat. No. 6,120,621, awarded to Kim and Fitzsimmons, for example, such layers were produced in a twin-belt caster by the use of “rough” steel belts. Satisfactory results were obtained when the surface texture was in the form of sharp peaks, imparted to the belts by shot blasting. A belt surface roughness in 4 to 15 μm range, measured as Ra, was preferred. This belt texture promoted the formation of Al₆(Fe, Mn) particles in liquated areas at the surface. The average size of the Al₆(Fe, Mn) particles in the surface layer was 3.5 μm, substantially coarser than the 1.7 μm observed in the interior of the strip. Significantly, in the absence of this inventive method, it was noted that the IM particles would be of the alpha-phase Al(Fe, Mn)Si type with an average size of 1.5 μm all over the strip, which is too fine for effective galling resistance during ironing. The performance of the inventive method was demonstrated in can making tests. It was shown that at least 50,000 cans could be made without scoring of the can surface owing to the presence of Al₆(Fe,Mn) particles.

The benefits of such a thin liquated layer were known from earlier work on strips from a block caster (also known as Caster II), described by Mcauliffe in Reference 1. It was argued that a 10 μm thick liquation layer was present on strip of alloy 5107 made in Caster II and the rolled sheet was galling-free. McAuliffe considered this feature to be an integral part of the block caster operation. It is unclear however, whether this was a natural result of solidification or the caster was being operated in a specific manner to obtain the said thin liquation layer which is normally considered undesirable for aluminum sheet.

Another method of generating a suitable size IM particle distribution on the sheet surface in the as-cast strip was recently described in U.S. Pat. No. 10,633,724 (assigned to Arconic, and known as Micromill™). Galling-free sheet was obtained by selection of the alloy composition in the hyper-eutectic range. The alloys described contained higher Fe (0.6-2%), Mn (0.8-2.2%) and Mg (up to 3%) than is normally used in the can making industries. This method does not rely on the liquation effect described in the earlier work, and does produce the required Al₆(Fe,Mn) particles. Presence of large Al₆(Fe, Mn) IM particles at the surface is an integral part of the microstructure development in this method. Unfortunately, this method requires high melt temperatures (which are potentially unsafe), high alloy content, and often promotes detrimental inclusions in the sheet outside a very narrow casting operating window.

There is thus a need to provide a continuous, preferably in-line, process for producing aluminum alloy can body stock. Preferably the alloy should be a lean conventional alloy such as AA3004 or AA3104 or AA5017, that can be processed in industrial can making equipment. To be acceptable to standard can making manufacturing, the sheet product must exhibit earing of less than 3% and be free of galling during the ironing stage of the process. It is accordingly an object of the present invention to provide a process for producing aluminum alloy can body stock which can be carried out in a continuous method without the need to employ separate batch operations. It is a more specific object of the invention to provide a method for commercially producing an aluminum alloy can body stock in a continuous process which can be operated economically and provide a product having equivalent or better metallurgical properties needed for can making than the current methods. These and other objects and advantages of the invention are shown below.

SUMMARY OF THE INVENTION

The present invention provides a method and apparatus for generating alpha-phase Al(Fe,Mn)Si IM particles of suitable size and density in the surface layers of a continuously cast strip. The fundamental principle is to control the size of alpha-phase IM particles on the surface of the cast strip so as to produce sufficiently large alpha-phase particles to provide for galling-free operation during can making. By selectively controlling the solidification rate and subsequent thermal path of the sheet surface, this distribution in IM particles can be achieved without the use of hypereutectic compositions, thus permitting the use of conventional beverage can alloy feedstock. The IM particles remain in the surface layers and break up to smaller sizes during rolling to provide the required galling-free performance in can making operations. Those skilled in the art will understand and appreciate that the present invention has no impact on earing performance.

The invention requires the use of textured casting surfaces or molds. The molds are comprised of belts or rolls or blocks in the caster, depending on the type of caster in use. The invention is applicable to all three types of casters. For the sake of easy reference, the casting surfaces will be referred to as “rolls” or “mold” hereinafter.

To generate IM particles of suitable size on the surface of the strip, the texture on the mold surface needs to be so designed as to promote solidification in certain regions of the mold without physical contact with the molten metal. Solidification will occur at slower rates in those areas and will produce larger IM particles. In other regions, direct physical contact with the casting surface occurs and the well-known fine IM particles are produced as a result of the higher solidification rates. The net result is thus two families of IM particles on the strip surface; one area with larger size IM particles formed in no-contact areas and another area of fine IM particles in the contact areas. By proper selection of the chemical composition of the alloy, and the type and area percentage of contact allowed, the amount and size of surface IM particles is controlled to achieve the required no-galling performance in the ironing operations of the can making process. Another advantage of this procedure is that the fine structure of the IM particles within the interior of the strip is still maintained. The inventive process is more tolerant of impurities such as Fe and Si, typically found in UBC (used beverage can) based feedstock than conventional methods (typically only about 60%), permitting a greater use of recycled content than allowed in conventional can body stock alloy formulations, significantly enhancing the economic benefit of the inventive method.

Control of the size and distribution of the fine alpha-phase IM particles on the surface of the continuously cast strip can be exerted through rapid heating of the strip surface. The surface may be rapidly heated to a temperature close to the solidus temperature of the alloy in a continuous heating process during which a thin layer of large alpha-phase IM particles is developed. As the strip surface is pre-populated with alpha phase IM particles over a range of sizes, the growth of these alpha phase particles is rapid and can be performed continuously. Therefore, the long cycle times at high temperatures and batch processing described by Hitchler and Klein are thus avoided, thereby permitting in-line continuous processing.

BRIEF DESCRIPTION OF FIGURES

FIG. 1 shows sub-micron particles of alpha-Al(Fe,Mn)Si on the surface of 3003 strip made in an experimental high-speed twin roll caster (from reference 2).

FIG. 2 shows large dendritic particles on the surface of AA5050 alloy strip produced in a commercial Hunter-Douglas block caster (from reference 3).

FIG. 3 is a schematic representation of rapidly heating the surface of the as-cast strip

FIG. 4 shows the suspension of molten metal over a groove in the mold.

FIG. 5 shows the depth of molten metal penetration as a function of groove width on the mold for three metallostatic head heights.

FIG. 6 is a plot showing the useful ranges of groove width and groove area percentage.

FIG. 7 shows the depth of molten metal penetration as a function of dimple diameter on the mold for three metallostatic head heights.

FIG. 8 is a plot showing the range of dimple diameters and dimple area percentage suitable for generating large intermetallic particles on the surface.

FIG. 9 shows dendrites of alpha phase Al₁₂(Fe,Mn)Si on the surface of AA3304 alloy cast on a substrate textured with protrusions of about 300 μm projected diameter.

DETAILED DESCRIPTION OF THE INVENTION

Example 1

Particles on the surface of a continuously cast aluminum alloy strip may be seen in scanning electron microscope (SEM) micrographs. The micrograph in FIG. 1 is typical for the process and shows sub-micron particles which are too small to be effective as scrubbing agents during ironing (image from reference 2). FIG. 2 shows the surface of an AA5050 strip produced by the Hunter-Douglas (block caster) cc method (see reference 3). The strip was 25 mm thick and 230 mm wide. It shows a dense network of IM particles on the surface. The IM particles vary in size, the largest being about 35 μm. These are therefore much coarser than the average 1.7 μm IM normally found on the surface of cc aluminum strip cast by other commercial methods. Energy Dispersive X-Ray (EDX) examination shows these dendritic IM particles to be quaternary Al—Fe—Mn—Si phases, alpha phase Al(Fe,Mn)Si. They contain 2-8% Si, 1-3% Mn and 10-20% Fe by weight. IM particles in the main alloying elements Fe, Si and Mn near the surface reach many times the bulk content in the alloy. Depth profiles of the sample show that this enrichment near the surface is confined to a depth of no more than 1 μm (reference 3). The large IM particles are thus present only on the surface of the as-cast strip. Some or all of these dendritic IM particles would break up during the subsequent rolling operations and be present in the final gauge sheet in the median size range of 2-7 μm that is preferred for imparting galling resistance. It is noted that these particles are not a product of liquation or exudation which would run tens of micron deep. Rather, they are formed from molten aluminum that was suspended between the high points on the mold surface. Without physical contact this resulted in slower solidification than that in contact areas. Therefore, it is possible to control the size of surface IM particles by proper selection of the mold surface texture and the bulk chemistry of the alloy.

TABLE 1
Chemical Composition of the Aluminum Alloy AA5050 (reference 3)
Element	Si	Fe	Cu	Mn	Mg	Cr	Zn	Ti	Pb	Be	Na
AA5050 (wt %)	0.28	0.8	0.12	0.33	1.20	0.05	0.07	0.03	0.01	0.0006	0.003

Example 2

It is also possible to modify the size of intermetallic particles on the surface of cc strip after casting. This can be done by heating the surface layers close to the solidus temperature of the alloy to grow the smaller alpha-phase particles by diffusion into larger particles at the surface. Consider a strip or slab that has been continuously cast and optionally rolled. The bulk temperature of said strip or slab could be close to the melting point if it has recently been cast. More specifically, the bulk strip or slab temperature could be within 50° C. of the solidus temperature. Alternatively, the bulk strip or slab temperature may be close to room temperature if it has been rolled and allowed to cool. By heating the surfaces of the strip sufficiently rapidly, it is possible to raise the surface temperatures of the strip close to the solidus temperature, while the average bulk temperature of the strip or slab remains below the solidus temperature. It is important that the bulk temperature remains below the solidus, so that the strip or slab does not fall apart due to bulk melting. Once the heating medium is removed, the strip or slab equilibrates to a new average temperature that is higher than the initial average temperature, but lower than the solidus temperature. During this heating, beneficially large intermetallic alpha phase Al(Fe, Mn)Si particles form, and are dispersed over the surface of the strip or slab. Subsequent rolling operations, especially hot rolling, preserve the beneficial particles in the proper size range at the surface, and they are thus available to scrub the dies during the ironing step of beverage can production.

Heating of the surface close to the melting temperature, while the average temperature of the strip or slab remains below melting, requires a high heat transfer method. A method and apparatus for rapidly heating the surface of the strip or slab is infra-red radiation (IR). FIG. 3 illustrates the concept for the case of a strip running through a high heat transfer infrared radiation heater.

In the figure, an infra-red radiation source [2] is placed proximate to the as-cast strip or slab surface [1]. Only the top half of the strip or slab is shown, but it is understood that the same method and apparatus can be simultaneously applied on the bottom half of the strip. The strip is travelling out of the caster at speed v, with average bulk temperature Tc, thickness t and the length of the infrared radiation source is L. The infrared radiation source has a temperature Tr. After equilibration, the average strip or slab temperature is Tf. The time, τ, spent heating up in the radiation zone is:

$\begin{array}{cc}\tau +L/v& \mathrm{Eqn}\left(1\right)\end{array}$

Heat is transferred to the strip by radiation. A simple approximation of the heat transferred to the strip may be obtained using the equation for two parallel plates:

$\begin{array}{cc}Q={\sigma }_{\mathrm{SB}}\frac{{\mathrm{Tr}}^4-{\mathrm{Tc}}^4}{\frac{1}{ϵr}+\frac{1}{ϵc}-1}& \mathrm{Eqn}\left(2\right)\end{array}$

• • where Q=heat transfer [W/m²], σ_SB=5.67×10⁻⁸ W/m² K⁴ (the Stefan-Boltzmann constant), ϵr and ϵc are the emissivities of the radiation source and the as-cast strip or slab, respectively. Energy imparted to the strip or slab, E, is approximately:

$\begin{array}{cc}E=Q*\tau *L*W& \mathrm{Eqn}\left(3\right)\end{array}$

• • where W is the width of the irradiated strip or slab. The mass of strip or slab in the irradiated area is:

$\begin{array}{cc}M=\rho *L*W*t/2& \mathrm{Eqn}\left(4\right)\end{array}$

• • where ρ is the material density of the strip or slab. The final average temperature of the strip or slab can be approximated as:

$\begin{array}{cc}\mathrm{Tf}=\mathrm{Tc}+\frac{E}{\mathrm{Cp}*M}& \mathrm{Eqn}\left(5\right)\end{array}$

• • where Cp is the specific heat capacity of the strip or slab.

For the present invention, Tf must always be below the solidus temperature of the strip or slab. To achieve this, either the irradiation length L or the speed v may be controlled. It is noteworthy that irradiation temperature Tr and thickness h may also be controlled to regulate the final temperature, but from a practical perspective, these two variables are difficult to alter and are therefore normally held constant.

An important feature of surface heating and the growth of large intermetallic alpha phase Al(Fe,Mn)Si particles is an abrupt change of appearance of the strip surfaces. In particular, the strip surfaces change from a shiny appearance to a dull or matte appearance. This indicates that the growth of the IM has occurred. Temperatures close to the solidus facilitate the rapid growth of beneficial large intermetallic alpha phase Al(Fe,Mn)Si particles. The dull or matte surface appearance is an important indicator of the success of the process. It causes the surface emissivity to jump from around 0.1 for the rolled shiny surface to a level of up to 0.9 for the dull surface. This is visible to the naked eye and can be measured by a pyrometer. Under those conditions, the radiation heat flux at the surface increases by about the same ratio. This may be understood from Eqn. 2 by taking ϵr=1 for the ideal radiation source and noting that the heat flux then becomes proportional to strip surface emissivity ϵc. This factor enables the strip to reach around the solidus temperature at the surface and generate the large alpha-phase particles without breaking the strip as the bulk of the strip is still at a lower temperature.

This concept is demonstrated by the following example. An infrared radiation (IR) source was used to heat aluminum alloy strip from room temperature. The alloy was in the 3xxx family, more specifically the alloy was close to AA3105 composition. The strip had been continuously cast at high speed and rolled. Initially, the strip speed was set high to ensure that the final average temperature was well below the solidus temperature of the strip. The strip speed was then reduced gradually until the surface appearance of the strip became dull, at which point the speed was held constant and the product was made at that operating point. Table 2 below shows the operating conditions:

TABLE 2
Operating Conditions of Infrared Heating
Aluminum Strip to Achieve Matte Surface
	MATERIAL	Symbol	Units	Value
	ALLOY			AA 3105
	Heat capacity	Cp	J/kg K	1004
	Density	ρ	kg/m3	2700
	Thickness*	t	m	0.00083
	Width*	W	m	0.165
	Strip emissivity	εc		0.1
	IR radiator emissivity	εr		0.95
	IR radiator temperature*	Tr	C	1000
	IR radiator length*	L	m	7.6
	Initial speed*		m/s	0.254
	Final speed*	v	m/s	0.167
	Initial strip temperature*	Tc	C	13
	Final strip temperature*	Tf	C	510
	Solidus temperature		C	633
	Initial surface appearance			Shiny
	Final surface appearance			Matte
	Dwell time in furnace	τ	s	45
	*measurements

Using the equations 1-5 above, the calculated final strip temperature is 519° C. This is reasonably close to the measured final strip temperature of 510° C., the difference being attributable to simplifying approximations used in the equations, and measurement errors. The measured surface temperature of 510° C. was thus below the equilibrium solidus temperature of 633° C., but sufficiently high to promote the changes in alpha-phase size and morphology.

The strip produced under these operating conditions was cold rolled to finish gauge and beverage cans were made from it. The drawn and ironed cans were free of galling after the production of more than 100,000 cans.

Example 3

Consider the casting of molten aluminum on a mold surface textured with straight grooves arranged in the casting direction of width w in the form of rectangular or hemi-cylindrical or wedge-shaped channels separated by land areas between them. This is illustrated in FIG. 3 for a wedge-shaped groove. Molten metal makes contact with the land area and penetrates into the grooves by a distance h. The surface of the solidified strip therefore shows protrusions over the grooves and relatively flat areas over the lands. Under these conditions, solidification of the metal suspending into the grooves is momentarily delayed compared to the parts solidifying at faster rates in contact with the lands. This delay allows the molten metal in suspension to take up solidification shrinkage of the newly-formed solid. The molten metal in suspension does this by pulling itself up over the grooves. As a result, the protrusion of the metal in the lands from the surface of the solidified strip is smaller than the penetration of the molten metal into the grooves.

Solidification on a mold with a groove pattern promotes the formation of large IM particles over the grooves where solidification is delayed and takes place at a substantially slower rate. The effect is restricted to the surface of the strip because as solidification progresses, differences between the groove area and land area are reduced and eventually become diffuse.

On the surface of the strip, two families of IM particles are observed. One family is in the land areas that are in contact with the mold and experience high solidification rates. This family has round IM, less than 1.5 μm or finer, as is well known from experience with cc. An example of such fine IM particles is shown in FIG. 1. Such fine particles do not provide the scrubbing effect needed during the ironing stage of can-making. The second family forms over the grooves in dendritic form and in sizes up to 15 μm, such as those illustrated in FIG. 2. These large IM fracture during subsequent rolling and reach an average size of about 3 μm required for the scrubbing action and thus avoidance of galling during the ironing stage of can manufacture.

Estimates of the penetration depth are used to design the width range and depth for groove patterns. Consider the ideal case of a column of molten metal suspending into a groove of width w, FIG. 4. The depth of the groove and the angle of the wedge are designed so as to prevent molten metal making contact with the walls and the root of the groove, respectively. Under these conditions, the radius of curvature, r, of the suspension can be calculated from the pressure of the molten metal and its surface tension using equation 6 below. If the pressure inside the molten metal is entirely due to metallostatic head, then for a cylindrical suspension

$\begin{array}{cc}\Delta p=\rho gH=\sigma /r& \mathrm{Eqn}\left(6\right)\end{array}$

• • where Δp=head pressure, ρ=density of molten aluminum, g=gravitational acceleration (9.8 m s⁻²), H=metallostatic head, σ=surface tension (a property of the molten aluminum alloy) and r is the radius of the cylindrical bubble over the groove. The depth of penetration h is calculated from the geometry of the bubble using:

$\begin{array}{cc}h=r\left(1-\cos \left(\theta \right)\right)& \mathrm{Eqn}\left(7\right)\end{array}$

where w=2 r sin(θ)

• • and θ is one half of the subtended angle as shown in FIG. 3. This value of h applies to the suspension of the metal in molten state. It is recognized that the actual penetration (observed as protrusion on the strip surface) will be somewhat less than this due to the pull up during solidification caused by shrinkage.

Calculations for molten aluminum with a nominal density of 2300 kg/m³ and surface tension σ=0.86 J/m² are shown in Table 3. It is clear that the higher the head level H and the wider the grooves, the deeper is the penetration h of the molten metal. An upper limit of h=30 μm may typically be placed on the penetration depth h since deeper penetration results in a very rough surface that is aesthetically undesirable even after rolling to final gauge.

TABLE 3
Effect of Groove Width on Penetration of Molten Metal
head (mm)	25.4		38		51
ΔP (Pa)	573		859		1145
r (mm)	1.50	r (mm)	1.00	r (mm)	0.75
groove
width		h		h		h
mm	sin(θ)	μm	sin(θ)	μm	sin(θ)	μm
0.10	0.033	1	0.050	1	0.067	2
0.20	0.067	3	0.100	5	0.133	7
0.30	0.100	8	0.150	11	0.200	15
0.40	0.133	13	0.200	20	0.266	27
0.50	0.166	21	0.250	32	0.333	43
0.60	0.200	30	0.300	46	0.399	63
0.70	0.233	41	0.350	63	0.466	87
0.80	0.266	54	0.399	83	0.533	115
0.90	0.300	69	0.449	107	0.599	150
1.00	0.333	86	0.499	134	0.666	191

FIG. 5 illustrates the relationship between groove width w and depth of penetration h, for various head heights H. Head heights H between 25 mm and 50 mm are considered practical by those skilled in the art of horizontal continuous casting. It is noted that the maximum groove width w depends on the operating metallostatic head and varies from about 0.4 to 0.6 mm. It is also clear that narrower grooves are preferred if the metal protrusion into the grooves is to be kept lower.

The depth of the grooves needed for efficient functioning may be understood as follows. Shallow grooves will not function because the depth of the grooves must be substantially greater than the calculated suspension depth to prevent contact with the walls. There also needs to be enough opening left in the grooves to allow the escape of gases (typically air) to prevent surface defects in the strip. Referring to Table 3, if the grooves were cylindrical, the molten metal would fill the grooves to about 30 μm depth at a metal head of 50 mm. On the basis of this observation, for triangular grooves as shown in FIG. 3 a minimum groove depth of 40 μm is recommended. However, for practical reasons an even deeper pattern needs to be applied since the casting surface is subject to wear and the groove depth is thus reduced during use.

Acceptable useful groove width and area percentage are shown in FIG. 6. A freshly textured mold surface would have the largest recommended groove width (about 0.5 mm) and the highest groove area percentage (75%). As the mold surface wears, the grooves become smaller and groove area would also be reduced until the texture loses functionality.

One other parameter for this family of textures is the depth of the grooves. A minimum 40 μm depth is required for proper functionality, as discussed above. Grooves can be created in a deterministic fashion using laser ablation, photoelectric etching, cutting, knurling or embossing. The upper limit for groove depth is also dictated by practical considerations such as the action of the grooves as potential stress risers and their influence on strength of the mold. A groove depth of up to 0.40 mm is considered suitable for a newly textured surface so that when the grooves stop functioning at ^˜40 μm depth, the land percentage would still be within the preferred range.

Example 4

Another form of surface roughness suitable for suspending the molten metal is depressions or dimples on the surface of the casting mold. The metal suspends into the dimples from the rims, equivalent to lands of the groove pattern. Such patterns can be created by many methods such as shot peening, grit blasting, electro-discharge texturing, cutting, laser ablation, photoelectric etching or embossing. The suitable size of such dimples can be calculated assuming the indentation of the dimple is a circle of radius R. The molten metal is modelled as a bubble of radius r above the dimple. This radius r and the depth of penetration h are calculated from Eqn. (8) and (9), respectively.

$\begin{array}{cc}\Delta p=\rho gh=2\sigma /r& \mathrm{Eqn}.\left(8\right)\end{array}$ $\begin{array}{cc}h=r\left(1-\cos \left(\theta \right)\right)& \mathrm{Eqn}.\left(9\right)\end{array}$

• • where θ represents the half-angle subtended over the dimple as in FIG. 3. The results of these calculations are shown in Table 4 for several dimple diameters (d=2R) and metallostatic head. Dimple diameters up to 1 mm can be used while the depth of penetration is still kept below 30 μm for the nominal metallostatic head of 38 mm. This calculation shows that the diameter of the dimples can be larger than the width of the grooves described in example 3 and still provide for suspension of the molten metal in the desired range. In practice, however, such large dimples are not preferred because the molten metal solidifying over the dimple is likely to develop surface cracks along a diameter as those areas are the last to solidify. This is due to the constraint from the edges of the dimples where the metal solidifies earlier in contact with the mold. Therefore, dimple dimensions similar to the width of the grooves described above (less than 0.7 mm) will be more useful. Dimple diameters applied to the casting mold surface should therefore be less than this maximum.

FIG. 5 shows the predicted depth of suspension on a dimple pattern for operation at three levels of metal head. It is seen that for any given dimple opening, the higher the metal head, the deeper is the predicted penetration. Dimple diameters up to 0.7 mm are seen to be suitable for keeping the suspension depth to a maximum of 20 μm. This lower value of suspension is needed for the dimpled surfaces to avoid lubricant trapping in the depressions of the cast strip surface during subsequent rolling steps. For operation at 50 mm metallostatic head, a value of 0.2 mm is recommended for the lower end of the dimple diameter for effective suspension of the molten metal and the desired delay in solidification over dimples.

Surface coverage of dimples, measured as area percentage, can vary over a wide range. With freshly textured mold surfaces, the high end of the dimpled area coverage and dimple size are recommended. This is because of the wear of the mold surface during casting as a result of which both the dimple diameters and the dimple area percentage will get smaller. With high coverage and wide dimples, the textures will last longer before they lose their functionality. It is recognized that processes such as shot peening create a range of dimple sizes, but not a single size. Aluminum allows a large dimple diameter up to 0.7 mm. Dimple coverage of 64% is recommended. These recommendations are shown in FIG. 8. A freshly textured mold surface would have the largest recommended dimple diameter (0.7 mm) and the highest coverage (64%). As the mold surface wears, the dimples would become smaller and coverage would also be reduced until the texture loses functionality. One other parameter for this family of textures is the depth of the dimples. For this 10-100 μm is recommended. Similar dimples can be created in a deterministic fashion using electro-discharge texturing, laser ablation, photoelectric etching, cutting, knurling or embossing.

In Table 4, the depth of penetration of molten aluminum into a mold surface with dimples (H=metallostatic head, h=penetration, d=dimple diameter, r=radius of the molten metal bubble above the dimple, θ=subtended angle; FIG. 3 may be used to illustrate this for dimples as well as grooves) are calculated.

TABLE 4
Effect of Dimple Diameter on Penetration of Molten Metal
	H(mm)	25.4		38		51
	r (mm)	3.00	r (mm)	2.00	r (mm)	1.50
dimple d		h		h		h
mm	sin(θ)	μm	sin(θ)	μm	sin(θ)	μm
0.20	0.033	1.66	0.050	2.50	0.067	3.33
0.30	0.050	3.75	0.075	5.62	0.100	7.51
0.40	0.067	6.66	0.100	10.01	0.133	13.37
0.50	0.083	10.42	0.125	15.66	0.166	20.95
0.60	0.100	15.02	0.150	22.60	0.200	30.26
0.70	0.117	20.46	0.175	30.82	0.233	41.34
0.80	0.133	26.75	0.200	40.35	0.266	54.24
0.90	0.150	33.89	0.225	51.21	0.300	68.99
1.00	0.166	41.90	0.250	63.41	0.333	85.66

Example 5

Another surface texture particularly advantageous is discrete protrusions on the mold surface. Such protrusions can be in the form of hemispheres or cylindrical rods with a hemispherical cap. The dendritic IM particles of alpha phase Al(Fe,Mn)Si family grown on such a surface is shown in FIG. 9 for alloy AA3304 with individual IM particles reaching up to 35 μm size. Such large IM particles break down in rolling operations to reach a size suitable for the scrubbing action needed during the ironing steps. It is considered that similar size and height ranges would apply to such protrusions as those described in Example 4 above for depressions. It is noted, as above, that these large IM particles are confined to the surface of the strip while IM size inside the strip thickness remains fine.

Claims

1. An aluminum alloy sheet product, suitable for the production of cans, made by continuous casting followed by rolling, annealing and cold rolling to final gauge wherein intermetallic particles of alpha phase Al(Fe, Mn)Si of diameters between 0.2 to 35 μm are present on at least one surface of the cast strip such that during the ironing steps of the can making process the product is substantially free of galling defects.

2. A method of making the product of claim 1 by continuous casting on molds with surface textures such as grooves or dimples or protrusions so as to produce intermetallic particles of alpha phase Al(Fe, Mn)Si on at least one surface of the cast strip.

3. The method of claim 2 wherein the casting mold surfaces have dimples or protrusions of diameter between 0.3 to 0.7 mm and depth or height between 0.01 to 0.2 mm with an area coverage in the range of 5% to 64%.

4. The method of claim 2 where the mold surface grooves or dimples or protrusions are produced by shot peening, laser shot blasting, grit blasting, electro-discharge texturing, laser ablation, photoelectric etching, cutting, knurling, roll forming or embossing.

5. The method of claim 2 wherein the casting mold is characterized by longitudinal grooves of widths between 0.04 to 0.50 mm and depths between 0.040 to 0.40 mm and groove area coverage between 2% to 75%.

6. The rolled sheet product of claim 1 in which the intermetallic particles of alpha phase Al(Fe,Mn)Si are in the diameter range of 0.2 to 15 μm with a median diameter between 2 and 7 um on at least one surface of the sheet.

7. The cast strip product of claim 1 in which the surface protrusions due to mold grooves, or mold dimples/protrusions, is 20 μm or less, or 30 μm or less, respectively.

8. A method of producing an aluminum alloy strip having intermetallic particles of alpha phase Al(Fe, Mn)Si by rapidly heating the strip surface to a temperature near the solidus temperature so as to coarsen the alpha phase Al(Fe, Mn)Si to a median diameter greater than 2 μm.

9. The method of claim 8 where at least one surface develops a matte finish indicating the presence of large alpha phase Al(Fe,Mn)Si particles.

10. The method of claim 8 wherein the surface reaches a temperature at which the surface develops a matte finish with a maximum average strip temperature 10° C. or more below the solidus temperature.

11. The method of claim 8 wherein the surface develops a matte finish with maximum average strip temperature 20° C. or more below the solidus temperature.

12. An aluminum alloy strip according to claim 1 that is rolled to finish gauge for making beverage cans that are substantially free of galling.