aluminum can

Makers of beer and soft-drinkcontainers in the U.S. produce300 million aluminum bever-

age cans a day, 100 billion of them ev-ery year. The industryÕs output, theequivalent of one can per American perday, outstrips even the production ofnails and paper clips. If asked whetherthe beverage can requires any morespecial care in its manufacture than dothose other homey objects, most of uswould probably answer negatively. Infact, manufacturers of aluminum cansexercise the same attention and preci-sion as do makers of the metal in anaircraft wing. The engineers who pressthe design of cans toward perfectionapply the same analytical methodsused for space vehicles.

As a result of these eÝorts, todayÕscan weighs about 0.48 ounce, downfrom about 0.66 ounce in the 1960s,when such containers were Þrst con-structed. The standard American alu-minum can, which holds 12 ounces ofliquid, is not only light in weight andrugged but is also about the sameheight and diameter as the traditionaldrinking tumbler. Such a can, whosewall surfaces are thinner than two pag-es from this magazine, withstands morethan 90 pounds of pressure per square

inchÑthree times the pressure in anautomobile tire.

Yet the can industry is not standingpat on its achievement. Strong econom-ic incentives motivate it toward furtherimprovements. Engineers are seekingways to maintain the canÕs performancewhile continuing to trim the amount ofmaterial needed. Reducing the canÕsmass by 1 percent will save approxi-mately $20 million a year in aluminum(and make still easier and even lessmeaningful the macho gesture of crush-ing an empty can with a bare hand).

Aside from the savings it yields, themodern manufacturing process impartsa highly reßective surface to the canÕsexterior, which acts as a superb base fordecorative printing. This attribute addsto the enthusiasm for the aluminum canamong those who market beverages.Indeed, that industry consumes abouta Þfth of all aluminum used in the U.S.Consequently, beverage cans haveemerged as the single most importantmarket for aluminum. Until 1985, mostcans held beer, but now two thirds ofthem store nonalcoholic drinks.

The aluminum beverage can is adirect descendant of the steelcan. The Þrst of these vessels ap-

peared in 1935, marketed by KreugerBrewing Company, then in Richmond,Va. Similar to food cans, this early bev-erage container comprised three piecesof steel : a rolled and seamed cylinderand two end pieces. Some steel canseven had conical tops that were sealedby bottle caps. During World War II, thegovernment shipped great quantitiesof beer in steel cans to servicemen over-seas. After the war, much of the produc-tion reverted to bottles. But veteransretained a fondness for canned beer, somanufacturers did not completely aban-don the technology even though thethree-piece cans were more expensiveto produce than the bottles.

The Þrst aluminum beverage canwent on the market in 1958. Developedby Adolph Coors Company in Golden,

Colo., and introduced to the public bythe Hawaiian brewery Primo, it wasmade from two pieces of aluminum. Toproduce such cans, Coors employed aso-called impact-extrusion process. Themethod begins with a circular slug thathas a diameter equal to that of the can.A punch driven into the slug forces ma-terial to ßow backward around it, form-ing the can. The process thus made theside walls and the bottom from onepiece. The top was added after Þlling.

This early technique proved inade-quate for mass manufacturing. Produc-tion was slow, and tooling problemsplagued the process. Moreover, the re-sulting product could hold only sevenounces and was not eÛcient structural-ly : the base could not be made thinnerthan 0.03 inch, which was much thick-er than it needed to be to withstandthe internal forces.

Nevertheless, the popularity of theproduct encouraged Coors and othercompanies to look for a better way tomake the cans. A few years later Rey-nolds Metals pioneered the contempor-ary method of production, fabricatingthe Þrst commercial 12-ounce alumi-num can in 1963. Coors, in conjunctionwith Kaiser Aluminum & Chemical Cor-poration, soon followed. But pressurefrom large can companies, which alsopurchased steel from Kaiser for three-piece cans, is said to have obliged Kai-ser to withdraw temporarily from alu-minum-can development. Apparently,these steel-can makers feared the com-petition of a new breed of container.Hamms Brewery in St. Paul, Minn., be-gan to sell beer in 12-ounce aluminumcans in 1964. By 1967 Coca-Cola andPepsiCo were using these cans.

Today aluminum has virtually dis-placed steel in all beverage containers.The production of steel three-piece cans,which are now rarely made, reached itspeak of 30 billion cans in 1973. Thenumber of two-piece steel cans toppedout at 10 billion in the late 1970s. Thisdesign now accounts for less than 1 per-cent of the cans in the U.S. market (they

48 SCIENTIFIC AMERICAN September 1994

The Aluminum Beverage CanProduced by the hundreds of millions every day, the modern can—

robust enough to support the weight of an average adult—is a tribute to precision design and engineering

by William F. Hosford and John L. Duncan

WILLIAM F. HOSFORD and JOHN L.DUNCAN have been active in researchon sheet-metal forming for more than30 years and act as consultants to alu-minum producers. Hosford is professorof materials science and engineering atthe University of Michigan. He receivedhis doctorate in metallurgical engineer-ing from the Massachusetts Institute ofTechnology and has written books onmetal forming and the plasticity of ma-terials. Duncan, who received his Ph.D.in mechanical engineering from the Uni-versity of Manchester in England, is pro-fessor of mechanical engineering at theUniversity of Auckland in New Zealand.Like Hosford, Duncan has written a text-book on the forming of sheet metal.

Copyright 1994 Scientific American, Inc.

SCIENTIFIC AMERICAN September 1994 49

ANATOMY OF MODERN BEVERAGE CAN reveals the dimen-sions that design and engineering must achieve on a dailybasis. The goal of can makers is to reduce the amount of alu-

minum needed without sacriÞcing structural integrity. A cannow weighs about 0.48 ounce; the industry hopes to reducethat weight by about 20 percent.

SCORED OPENINGThe lid is scored so that themetal piece pushes in easilywithout detaching.

NECKThe body of the can isnarrowed here to accom-modate the smaller lid.

BODYThis aluminum alloy typ-ically incorporates by weight1 percent magnesium, 1 per-cent manganese, 0.4 per-cent iron, 0.2 percent siliconand 0.15 percent copper. Itis ironed to dimensions with-in 0.0001 inch and is madethicker at the bottom foradded integrity. It withstandsan internal pressure of 90pounds per square inch andcan support 250 pounds.

RIVETUsed to secure the tab tothe can, this integral piece ofthe lid is made by stretchingthe center of the lid upwardslightly. It is then drawn toform a rivet.

TABThis separate piece of met-al is held in place by the in-tegral rivet.

FLANGEAfter the top of thecan is trimmed, it isbent and seamedto secure the lid af-ter filling.

BASEThe bottom of the can as-sumes a dome shape in orderto resist the internal pressure.

LABELThe ironing process that thinsthe body of the can producesa highly reflective surfacesuitable for decoration. Themirrorlike finish may be oneof the main reasons mar-keters of beverages adoptedthe aluminum can.

LIDThe lid may make up 25 per-cent of the total weight. Itconsists of an alloy that con-tains less manganese butmore magnesium than thebody does, making it stronger.To save on the mass, manu-facturers make the diameterof the lid smaller than that ofthe body.

0.012”

0.006”

0.003”

0.005”


are, however, more popular in Europe).The process that Reynolds initiated

is known as two-piece drawing and wallironing. Aluminum producers beginwith a molten alloy, composed mostlyof aluminum but also containing smallamounts of magnesium, manganese,iron, silicon and copper. The alloy iscast into ingots. Rolling mills then ßat-ten the alloy into sheets.

The Þrst step in can making is cut-ting circular blanks, 5.5 inches in diam-eter. Obviously, cutting circles from asheet produces scrap. The theoreticalloss for close-packed circles is 9 per-cent; in practice, the loss amounts to12 to 14 percent. To reduce this waste,sheets are made wide enough to incor-porate 14 cups laid out in two stag-gered rows. Each blank is drawn into a3.5-inch-diameter cup.

The next three forming operationsfor the can body are done in one con-tinuous punch stroke by a second ma-chineÑin about one Þfth of a second.First, the cup is redrawn to a Þnal in-side diameter of about 2.6 inches, whichincreases the height from 1.3 to 2.25inches. Then, a sequence of three iron-ing operations thins and stretches thewalls, so that the body reaches a heightof about Þve inches. In the last step, thepunch presses the base of the can bodyagainst a metal dome, giving the bottomof the can its inward bulge. This curvebehaves like the arch of a bridge in thatit helps to prevent the bottom frombulging out under pressure. For addedintegrity, the base of the can and thebottom of the side walls are made thick-er than any other part of the can body.

Because the alloy does not have thesame properties in all directions, thecan body emerges from the forming op-

erations with walls whose top edges arewavy, or Òeared.Ó To ensure a ßat top,machinery must trim about a quarterinch from the top. After trimming, thecup goes through a number of high-speed operations, including washing,printing and lacquering. Finally, thecan is automatically checked for cracksand pinholes. Typically, about one canin 50,000 is defective.

Ironing is perhaps the most criticaloperation in making the body of thecan. The precisely dimensioned punchholds and pushes the cup through twoor three carbide ironing rings. To thinand elongate the can, the punch mustmove faster than the metal does in theironing zone. The clearance betweenthe punch and each ring is less thanthe thickness of the metal. The frictiongenerated at the punch surface assistsin pushing the metal through the iron-ing rings. To increase this friction, thepunch may be slightly roughened witha criss-cross scratch pattern (which can

be seen, impressed on the inside of acan). On the exterior of the can theshearing of the surface against the iron-ing rings yields the desired mirror Þnish.

The side walls can be thinned with-out loss of integrity because, structur-ally, the can is a Òpressure vessel.Ó Thatis, it relies for part of its strength on theinternal force exerted by carbon diox-ide in beer and soft drinks or by the ni-trogen that is now infused into suchuncarbonated liquids as fruit juice. In-deed, most beers are pasteurized in thecan, a process that exerts nearly 90pounds per square inch on the materi-al. Carbonated beverages in hot weath-er may also build up a similar pressure.

Filling introduces a diÝerent kind ofstress on the can. During this stage, thecan (without its lid) is pressed tight-ly against a seat in a Þlling machine. Itmust not buckle, either during Þlling andsealing or when Þlled cans are stackedone on another. Hence, can makers spec-ify a minimum Òcolumn strengthÓ of


STEPS IN CAN MANUFACTURE begin with an aluminum alloysheet. Blanks 5.5 inches in diameter are cut from the sheet; apunch draws the circle to form a 3.5-inch-diameter cup. A

second machine then redraws the blank, irons the walls andgives the base its domeÑall in approximately one Þfth of asecond. These procedures give the can wall its Þnal dimen-

DRAWING AND IRONING constitute the modern method of beverage can manufac-ture. The initial draw transforms the blank into a small cup (1 ). The cup is trans-

��

1 2

SLEEVE

ALUMINUMBLANK

PUNCH

��

BLANK AND DRAWFROM SHEET

REDRAW IRON AND DOME TRIMEARS

CLEAN


about 250 pounds for an empty canbody. Thin-walled structures do not eas-ily meet such a requirement. The slight-est eccentricity of the loadÑeven a dentin the can wallÑcauses a catastrophiccollapse. This crushing can be demon-strated by standing (carefully) on an up-right, empty can. Manufacturers avoidfailures by using machines that holdthe cans precisely.

The second piece of the can, the lid,must be stiÝer than the body. That isbecause its ßat geometry is inherentlyless robust than a curved shape (dams,for instance, bow inward, presenting aconvex surface to the waters they re-strain). Can makers strengthen the lidby constructing it from an alloy thathas less manganese and more magne-sium than that of the body. They alsomake the lid thicker than the walls. In-deed, the lid constitutes about onefourth the total weight of the can. Tosave on the mass, can makers decreasethe diameter of the lid so that it is

smaller than the diameter of the cylin-der. Then they Òneck downÓ the top partof the cylindrical wall, from 2.6 to 2.1inches, to accommodate the lid. An in-genious integral rivet connects the tabto the lid. The lid is scored so that thecan opens easily, but the piece of metalthat is pushed in remains connected.

In addition to clever design, makingbillions of cans a year demands reliableproduction machinery. It has been saidthat in order to prove himself, an ap-prentice Swiss watchmaker was not re-quired to make a watch but rather tomake the tools to do so. That sentimentapplies to can manufacturing. As oneproduction manager remarked, ÒIf atthe end of a bad day, you are a half mil-lion cans short, someone is sure to no-tice.Ó A contemporary set of ironing diescan produce 250,000 cans before theyrequire regrinding. That quantity isequivalent to more than 20 miles of alu-minum stretched to tolerances of 0.0001inch. Die rings are replaced as soon as

their dimensions fall out of speciÞca-tion, which occurs sometimes more thanonce a day.

Much of the success behind theconsistent and precise produc-tion lies in the strong yet form-

able alloy sheet. The metallurgical prop-erties responsible for the performanceof modern can sheet have been propri-etary and therefore not well known.Only within the past decade has thatsituation changed. Through the eÝortsof Harish D. Merchant of Gould Elec-tronics in Eastlake, Ohio, James G. Mor-ris of the University of Kentucky andothers, scientiÞc papers on the metal-lurgy of can sheet have become morewidely published.

We now know that three basic factorsincrease the strength of aluminum. Wehave already mentioned one of them:manganese and magnesium dissolvedinto the material. These atoms displacesome of the aluminum ones in the sub-stance. Because they are slightly diÝer-ent in size, the manganese and magne-sium atoms distort the crystal lattice.The distortions resist deformation, thusadding strength to the sheet.

The second contribution comes fromthe presence of so-called intermetallicparticles. Such particles, which formduring the processing of the sheet, con-sist of a combination of diÝerent met-als in the alloy (mostly iron and man-ganese). They tend to be harder thanthe alloy itself, thus supplying strength.

Perhaps the most important contri-bution to sheet strength, however, isthe work hardening that occurs whenthe sheets are cold-rolled (ßattened atroom temperature). During this shap-ing, dislocations, or imperfections, inthe lattice materialize. As the metal de-forms, the dislocations move about andincrease in number. Eventually they be-come entangled with one another, mak-ing further deformation more diÛcult.


sions. After the ÒearsÓ at the top of the walls are trimmed, the can is cleaned, deco-rated and then ÒneckedÓ to accommodate the smaller lid. The top is ßanged to se-cure the lid. Once Þlled and seamed shut, the can is ready for sale.

ferred to a second punch, which redraws the can; the sleeveholds the can in place to prevent wrinkling (2 ). The punch

pushes the can past ironing rings, which thin the walls (3 ). Fi-nally, the bottom is shaped against a metal dome (4 ).

3 4CAN WALL

IRONINGRING

��

DECORATE NECK FLANGE FILL ANDSEAM


Unfortunately, this work hardeningdramatically reduces the ability of thematerial to stretch. Tensile tests indi-cate that the elongation capacity dropsfrom 30 percent to about 2 or 3 percent.Conventional wisdom had it that sheetscan be formed only if the material hasa high tensile elongation. Certainly inthe automotive industry, body parts areformed from fully annealed sheets thatcan elongate more than 40 percent. Thisphilosophy guided the early attemptsto make two-piece aluminum cans. Re-searchers concentrated on annealed orpartially work-hardened sheets, whichsacriÞced strength for ductility.

The understanding of formability re-ceived a major boost from studies inthe 1960s by Stuart P. Keeler and Wal-ter A. Backofen of the MassachusettsInstitute of Technology and ZdzislawMarciniak of the Technical University inWarsaw, among others. Looking at thebehavior of various sheet metals, theyconsidered more than just the behaviorunder tension applied in one direction(as is done in the tension test). Theyalso looked at what happens when ten-sion is applied simultaneously in two

directions. They showed that a smallwindow of strains exists that permitsforming without structural failure. Al-though work hardening greatly reducesthe size of this window, a small slitnonetheless remains openÑenough topermit the doming of the base anddrawing and redrawing of the side walls.

The crucial advance that made thealuminum can economical, however,came from Linton D. Bylund of Rey-nolds. He realized that cans could bemade from a fully work-hardened sheetusing a carefully designed process thatspeciÞed the placement of the ironingrings, the shape of the punch and dies,and many other parameters. The strong,fully work-hardened sheet made it pos-sible to use sheet that was thinner, sav-ing enough weight to make the canseconomically competitive.

Nowhere is the technique of formingwork-hardened sheet more apparentthan it is in the cleverly designed rivetthat holds the tab on the can lid. Therivet is an integral piece of the lid. Tomake it, the center of the lid must bestretched by bulging it upward a bit.This ÒextraÓ material is drawn to form

a rivet and then ßattened to secure thetab (which is a separate piece of metal).

Besides making the can sheetstronger, manufacturers alsosought to reduce the amount of

aluminum needed by controlling thewaviness, or earing, which as we haveseen takes place at the top of the canafter ironing. The eÝect derives fromthe crystallographic texture of the alu-minum sheet, that is, the orientation ofits crystal structure. Hence, earing is in-evitable to some extent. Hans-JoachimBunge of the Technical University inClausthal, Germany, and Ryong-JoonRoe of Du Pont and others have devel-oped x-ray diÝraction techniques to de-scribe qualitatively the textures thatcause earing. Laboratory techniciansprepare specimens by grinding awaylayers of the sheet to expose materialat diÝerent depths. X-ray diÝractioncoupled with elegant analytical tech-niques automatically produces three-dimensional diagrams that reveal thepreferred orientation of crystals as afunction of depth in the sheet.

Such diagnostic approaches have en-abled aluminum companies to producesheet that yields much smaller ears.Metallurgists balance the two predomi-nant crystallographic textures that existin the aluminum. One kind of texturearises during annealing of the alloy af-ter the alloy is hot-rolled from ingots. Itcauses four ears to appear every 90 de-grees (at 0, 90, 180 and 270 degrees)around the circumference of the can.The second kind of texture results fromcold-rolling the sheet, which producesan ear at 45, 135, 225 and 315 degrees.Proper control of annealing and rollingcan lead to a combination of the twotextures such that ears caused by oneÞll the valleys caused by the other. Theresult is eight very low ears. The maxi-mum height of an ear is often less than1 percent of the height of the cup.

Consistent processing of metal andcareful design have now made each partof the can about as strong as any other.It is not unusual to Þnd cans in whichthe opening on the lid fractures, andthe bottom dome and lid bulge at near-ly the same pressure, within the rangeof 100 to 115 pounds per square inch.

Despite the success of current designand manufacture, can makers are stillsearching for reÞnements. Much of theinvestigation focuses on ways to usealuminum more eÛciently, because themetal represents half the cost of thecan. One possibility for saving wouldbe to cast the molten alloy into thinslabs rather than into thick ingots, as iscurrently done. A typical ingot may be30 inches thick, which is rolled down


ANNUAL BEVERAGE CAN PRODUCTION in the U.S. has increased by several billionover the past few years. The two-piece aluminum can overwhelmingly dominatesthe market; steel cans constitute less than 1 percent. Three-piece steel cans, whichare now rarely made, reached their peak production in the mid-1970s.

NU

MB

ER

OF

CA

NS

PR

OD

UC

ED

IN T

HE

U.S

. (B

ILLI

ON

S)

100

80

60

40

20

01965 1970 1975 1980 1985 1990

TOTAL BEVERAGECANS

TWO-PIECEALUMINUM

THREE-PIECESTEEL

TWO-PIECE STEEL

THREE-PIECECAN

TWO-PIECECAN


by a factor of 2,500 to 0.011 or 0.012inch. So much rolling requires expensivecapital equipmentÑfurnaces and rollingmillsÑand consumes a lot of energy.

It is possible to cast aluminum contin-uously into slabs that are an inch thickor less. These thin slabs would requiremuch less rolling to reach the desiredÞnal sheet thickness. Continuous cast-ing is used for some soft aluminum al-loysÑfor example, aluminum foil ismade from material cast to a thicknessof 0.1 inch.

Unfortunately, production of satisfac-tory can stock from thin slabs thwartsthe metallurgists. The faster coolingand decreased rolling inherent in con-tinuous casting do not yield the desiredmetallurgical structure. Two main prob-lems arise. First, crystallographic texturecannot be properly controlled to pre-vent large ears. Second, the faster cool-ing rate produces severe diÛculties inironing the can walls.

These ironing problems develop be-cause of the nature of the intermetallicparticles that form when the molten al-loy solidiÞes. Intermetallic particles thatdevelop during solidiÞcation are muchlarger than those that originate duringprocessing (which as we have seen im-part strength to the sheet). Because oftheir size, they play a key role in iron-ing. During this procedure, aluminumtends to adhere to the ironing rings.Ordinarily, the intermetallic particles,which are about Þve microns in size,act like very Þne sandpaper and polishthe ironing rings. The faster coolingrates of continuous casting, however,produce intermetallic particles that aremuch smaller (about one micron). Atthis size, the particles are not very ef-fective in removing aluminum thatsticks to the ironing rings. As a result,aluminum builds up on the rings andeventually causes unsightly scoring onthe can walls. The problem of achiev-ing thin slabs with the desired interme-tallic particles may yet be solved, per-haps by altering the composition of thealloy or by shifting the rate of solidiÞ-cation from the materialÕs molten state.

The control of casting epitomizesa recurrent feature of the wholecan story: one behavior is care-

fully traded oÝ against another, fromthe control of earing and ironability toeconomical sheet production, from canweight to structural integrity. Yet onecost element eludes an easy balance:the energy needed to make cans. Mostof this outlay lies in the aluminum it-self. Taking into account ineÛcienciesin electricity distribution and smelting,industry experts estimate that 2.3 mega-joules of energy is needed to produce

the aluminum in one can. This value isequal to about the amount of energyexpended to keep a 100-watt bulb litfor six hours, or about 1.7 percent ofthe energy of a gallon of gasoline. Al-though small, it represents the majorexpenditure of a can.

One way to reduce this expense isthrough recycling, which can save upto 95 percent of the energy cost. Indeed,more than 63 percent of aluminum cansare now returned for remelting. Recy-cling also has an important part withinthe aluminum mill. For every ton of canbodies made, a ton of scrap metal isproduced. This scrap is remelted andthus injected back into the manufactur-ing cycle. Developing simpler ways ofproducing can sheet and Þnding strong-er materials that can lead to lighter cansshould save more money and energy.

Meeting these goals presents a greatchallenge. Existing cans already use ahighly strengthened, well-controlledsheet. Their shape is Þnely engineeredfor structural strength and minimumweight. And with little tool wear, theproduction machinery in a single plantis capable of making many millions of cans a day with few defects. The re-

wards of even small improvements,however, are quite substantial. The de-mand for aluminum beverage cans con-tinues to grow everywhere in the world;their production increases by severalbillion every year. The success of thecan is an industrial lesson about whatcan be achieved when scientiÞc and en-gineering skills are combined with hu-man perseverance.


EASY-OPENING LIDS were introduced on three-piece steel cans in 1961. The originalcaption reads: ÒHousewives of ancient Greece and the space age compare contain-ers for the kitchen at the press debut of the new canning innovation by the Can-TopMachinery Corp., Bala-Cynwyd, Pa.Ó

FURTHER READING

A GOLDEN RESOURCE. Harold Sohn andKaren Kreig Clark. Ball Corporation,1987.

FROM MONOPOLY TO COMPETITION: THETRANSFORMATIONS OF ALCOA, 1888Ð 1986. George David Smith. CambridgeUniversity Press, 1988.

THE MECHANICS OF SHEET METAL FORM-ING. Z. Marciniak and J. L. Duncan. Ed-ward Arnold, 1992.

ALUMINUM ALLOYS FOR PACKAGING. Ed-ited by J. G. Morris, H. D. Merchant, E. J.Westerman and P. L. Morris. Minerals,Metals and Materials Society, Warren-dale, Pa., 1993.

METAL FORMING: MECHANICS AND MET-ALLURGY. William F. Hosford and Rob-ert M. Caddell. Prentice Hall, 1993.


aluminum can

Documents