Hot Rolling and Explosive Lamination of Copper-Nickel on Copper Core Material
By Roger W. Burdette, special to CoinWeek …..
When the Treasury Department faced the necessity of switching from a .900 fine silver alloy to a base metal composition for dimes, quarters, and half dollars, they commissioned a report from Battelle Memorial Institute to examine suitable coinage alloys.[1] Among the most promising materials was a laminated (or clad) structure consisting of a core of pure copper bonded to outer layers of approximately 75% copper, 25% nickel alloy. A version of this was eventually adopted by the United States as a substitute for the 90% silver, 10% copper alloy in 10-cent and 25-cent coins, and eventually 50-cent and both large and small one-dollar coins. Cupronickel clad was not a commonly used material at the time and presented unusual problems in fabrication.
To understand the difficulties faced in producing a clad core or laminated coinage material, we must review the conditions necessary to manufacture coins from metal. Using silver copper alloy as an example, first, the pure metals are melted in a crucible. The goal is a uniform mixture, or alloy, of silver and copper. The mixture is poured into iron molds to form specially shaped bars called ingots. These are taken to pairs of large horizontal steel rollers and passed between them multiple times. Each pass is made with the rolls slightly closer together. After five or six passes, the metal hardens and has to be heat-treated (annealed) to restore its softness. This continues until the desired thickness is obtained.
Once the silver copper alloy is at the planned thickness, the metal strips are taken to a machine that punches circular blanks of the correct coin diameter. The blanks then go to an upsetting machine that forms a slightly raised rim on the blank and ensures that it is perfectly smooth and uniform. Last, these planchets go to a coining press where obverse, edge, and reverse dies impress designs into the surfaces.
Under this basic process, the metal has to be capable of certain physical and mechanical conditions. It must be uniform in composition, and it must be pliable enough to allow rolling into thin strips. It must be treatable to prevent it from becoming too hard or brittle, and it must take a faithful impression from the dies without cracking or splitting.
Olin Mathieson Chemical Corporation Laminates
Laminated coinage material had to meet similar mechanical criteria, and its manufacture was more difficult than commonly used coinage alloys.
Normal cold rolling (room temperature) of clad and core sheets could be used, but the bond was often weak, failing when the material was bent or heated.
Hot rolling didn’t help much and increased other problems, such as thermal expansion.[2]
When pure copper is used for the core, there is a tendency for cladding to form a thin layer of brittle compounds where the copper and cladding meet. This brittle layer is easily shattered when the laminate is flexed, bent, or struck by coinage dies.
A further problem with copper core materials is a tendency for copper to oxidize, especially at the temperatures necessary for hot rolling. This produces a layer that inhibits bonding with the cupronickel outer layers. The unintentional presence of iron oxide or aluminum oxide can result in inconsistent bonding of layers.
All of these limit the production of uniformly bonded cupronickel-copper core laminates suitable for coinage. The challenges were both in materials and in their composite production for suitable coinage use.[3]
Considerable industrial experimentation occurred during the late 1950s. By 1962, Joseph Winter, a young research engineer at the Olin Brass Material Research Laboratories in New Haven, Connecticut, developed a method for producing high-quality copper-nickel, copper core laminate for coinage.[4] Winter’s method is surprisingly simple yet resulted in effective clad material that could be produced in large coils for later mint use.[5] Similar methodology was used by Texas Instruments, Union Carbide, Engelhard Industries, and others who supplied clad coinage material to the United States Mint.[6]
The process begins with a copper core plate that is at least 12.7 mm thick and upper and lower copper-nickel plates less than 6.35 mm thick. The core plate is heated to between 150ºC and 540ºC, but the copper-nickel cladding plates are not heated. The three are passed between rolls capable of reducing the thickness by 40% to 80% during a single pass. The two cladding plates contact the rolls first and the copper core is pressed between them at the initial “bite” of the rolls. In addition to heating only the core, an important part of the process is that the top and bottom cladding enter the rolls at an angle of between 10 and 22 degrees from horizontal, while the heated copper core is horizontal. Rolling proceeds at a rate of at least 30.5 meters per minute. The output strip can be of any convenient length and spooled on a take-up reel as if it were a normal binary alloy.
Winter explains in his patent why his method produces good bonding:
On the front side of the rolls (entering side) the cladding and the rolls are travelling at different linear speeds; whereas, at the exit side they are going at the same speed due to the reduction in thickness of the composite. The difference in travelling speeds between the cladding and the rolls, coupled with the pre-contact between the cladding and the rolls, generates a shearing strain between the cladding and the rolls. This introduces shearing at the bite of the rolls to the core-clad interface. This shearing strain at the core-clad interface results in turbulent flow of metal at the interface which causes more intimate bonding by increasing the interfacial linear surface of the composite by 20% or more.[7]
The angled entrance for cladding into the rolls is critical. If the angle is too shallow, or if the cladding and core come into contact before the rolls bite, results will be poor, and the material will delaminate locally if annealed or severely flexed.[8]
Manufacturing Clad Coinage with Explosive Bonding
In 1960, Belgian patent number 599,918 to George R. Cowan, John J. Douglass, and Arnold H. Holtzman described the formation of a solid metallurgical bond between two or more layers of metals by explosively driving the layers together. It was registered in the United States by the same inventors as assignors to E.I. du Pont de Nemours and Company, titled “Explosive Bonding”[9], patent 3,137,937.
The inventors described several potential defects in hot-rolled laminated metals and claimed that their explosive bonding process:
…[Provides] a means whereby virtually any two or more metals or metal alloys may be bonded together a form a composite multilayered system. A still further object of the invention is the provision of clad metal systems wherein the cladding layer is nonpervious, uniform, and adherent, by a method that is convenient and economical. A still further object of this invention is to provide unique clad assemblies from metals which heretofore could not be bonded effectively together.[10]
Their approach involved layering a material to be cladded (the “core”), a cladding metal, and an explosive. As might be anticipated, the explosive had to be carefully selected so as not to destroy the metals or factory.
As described in this 1960 patent, DuPont’s process involved placing a piece of cladding metal above and slightly separated from the metal to be clad. The separation can be as little as 0.025 mm up to about 10 mm, provided air was excluded if the separation was greater than about 1 mm. A layer of explosive was placed on the outer surface of the cladding metal. The explosive was selected to have a velocity of detonation less than 120% of the speed of sound in the medium of the metals being joined. If the velocity was too high, the cladding metal would be sheared or shredded before it could make solid contact with the material to be clad. A shockwave created by the explosive causes brief plasticization of the facing surfaces of both metals and also forces the two together, forming a solidly bonded interface.[11] This method is especially useful for materials that cannot be welded or easily bonded, including metals to ceramics.
In the conceptual illustration above, the material to be clad (or the “core”) is at the bottom. Spaced slightly above is the cladding metal, and above that is a layer of explosive. As the explosive detonates (left to right), the cladding is forced onto the core, with the plastic (plasma) surfaces of both metals forming a complete metallurgical bond. The explosion must move linearly from the point of detonation to exhaustion.
In earlier work, DuPont engineers demonstrated explosive bonding, which they called Detaclad, only on small pieces of metal, generally not larger than about 700 mm x 700 mm. By 1964, they had succeeded in making Detaclad in long strips suitable for coiling. We do not have DuPont’s detailed manufacturing process, but we have tantalizing hints contained in newspaper articles and a bound volume prepared by Richard A. Brown of Revere Copper and Brass, Inc. titled A Pictorial Summary of the Conversion of Cupro-Nickel Coinage Metal for U.S. Mint containing 21 color photos documenting the Detaclad process.[12]
In practice, DuPont likely combined processes from several patents to produce sufficient laminate for the production of blanks by the Mint. The process below conforms to DuPont’s patents and basic materials engineering principles. Using available information as a guide, the author has created a possible scenario for DuPont’s production. A stack of metal plates several centimeters thick and approximately one meter wide was placed on a supporting plywood support. The top and bottom plates were 75% copper and 25% nickel cladding, and the center plate was pure copper. The metals were of appropriate thickness so that after bonding and rolling, coin blanks of correct thickness and metal proportions could be cut.
The cupronickel plates had been lightly rolled with regularly spaced small surface dimples as shown in the conceptual illustration. These served to provide uniform separation between copper core and cladding. These were described in DuPont patent US #3140539 and are also suggested by photomicrographs made by Bell Labs in November 1965. A layer of masking tape, polystyrene foam, or other inert material covered the outside surface of the cladding to protect it from direct effects of the explosive layer.[13]
The preferred explosive was probably a flexible composition consisting of 20% very fine pentaerythritol tetranitrate (PETN), 70% red lead, and a binder of 10% of a 50/50 mixture of butyl rubber and a thermoplastic terpene resin adhesive [a mixture of polymers of β-pinene of formula (C 16 H 6 )n], commercially available as “Piccolyte” S-10 (manufactured by Pennsylvania Industrial Chemical Corporation, now Hercules Inc.). This material was readily rolled into sheets so that a single sheet could extend the full length of the production trough. This explosive detonates at a velocity of about 4,100 meters per second and has a uniform velocity over distances of several hundred meters.[14] The explosive extends slightly beyond the cladding to allow for convenient attachment of an initiator, and to ensure that the detonation front will have achieved maximum velocity when it is adjacent to the edge of the cladding assembly. Once bonding was complete, the resulting ingot was cleaned and passed through heavy rolls to the final thickness.
Evidently, DuPont created its laminated coinage material by bonding both outside layers simultaneously with dual layers of explosive.[15]
To demonstrate Detaclad’s utility for coinage, DuPont had 500 quarter-size tokens struck by August C. Frank, Co. for distribution to Treasury and Mint officers, and others. Quantities of Detaclad were supplied to the U.S Mint[16] and were tested at both the Philadelphia and Denver mints.[17]
In addition to quarter-size coinage samples, DuPont sent explosive bonded material (or cut blanks) to Inco, or possibly directly to vending equipment manufacturers such as National Rejector Co. Two of these are shown below.
Small sample plates of coinage composition and souvenir ingots of other compositions were used for internal company meetings and promotion.
The range of promotional items indicates that DuPont was interested in much more than the possible U.S. Mint use of explosive bonding. Ultimately, cost and manufacturing limitations caused the company to back away from this technology except for small-scale work. The Detaclad subsidiary, originally organized at Pompton Lakes, New Jersey, in 1965, was sold to Dynamic Materials Corp. in July 1996 for about $5 million.
* * *
Notes
[1] Battell Memorial Institute. “A Study of Alloys Suitable for use as United States Coinage to U. S. Department of the Treasury, Bureau of the Mint”. February 12, 1965.
[2] See: Shapiro, Eugene and Joseph Winter. “Stress Relaxation in Spring Materials”. Olin Brass Material Research Laboratories, New Haven, CT. 1981.
[3] Winter, Joseph, assignor to Olin Mathieson Chemical Corporation. “Process for Obtaining a Clad Article with a Copper Base Alloy Core,” Continuation in part of application Ser. No. 229,262, Oct. 2, 1962; application May 7, 1965, Ser. No. 454,182. Patent number 3,381,364. p.1.
[4] Ibid. Other companies were working along similar lines of research, and the U.S. Mint’s first orders for cupronickel clad material went to the Metals and Controls Division of Texas Instruments, Inc. in December 1964.
[5] Kula, Eric, and Volker Weiss, eds. “Residual Stress and Stress Relaxation”, 28th Sagamore Army Materials Research Conference Proceedings. Springer Science. 1981.
[6] See: United States General Accounting Office (GAO). “Briefing Report to the Honorable John F. Kerry, U.S. Senate: Procurement of Clad Metal for Coins”. Government Printing Office, Washington D.C., May 17, 1991, for a summary and overview of clad coinage material procurement limitations.
[7] Winter, p. 2.
[8] See also U.S. Pat. Nos. 3,496,621, 3,496,625, 3,381,366, and 3,753,669 for variations on this process.
[9] George R. Cowan, John J. Douglass, and Arnold H. Holtzman, assignors to E.I. du Pont de Nemours and Company. “Explosive Bonding”, Patent 3,137,937 filed October 26, 1960. Cowan was a physicist, Douglas a mechanical engineer, and Holtzman a metallurgist.
[10] Cowan et al., p. 2.
[11] Ibid.
[12] Heritage Auctions Long Beach U.S. Coin Signature Auction, June 5-9, 2013, Lot 4934. This material dates from 1964, about a year before DuPont developed a continuous clad bonding process. Revere Copper and Brass worked with DuPont to develop a manufacturing capability for the explosive bonding process.
[13] Separation could also have been created by scattering fine copper or cupronickel particles across the exposed surfaces. The dimple approach is much more uniform and reliable, especially with large metal plates.
[14] Other explosives could be used depending on the materials being bonded.
[15] “U.S. Could Use DuPont’s Process”, The Morning News. Wilmington, DE. June 4, 1965. p.38.
[16] Ibid.
[17] Gibbs, William T. “Daughter Preserves Father’s Record of 1964 Coinage Testing”, Coin World. June 4, 2013.
* * *