National Park Service Restoring The Statue of Liberty By Sharon Ofenstein National Park Service Technical Bulletin CRM Bulletin
Volume 9: No. 2 Cultural Resources Management • A National Park Service
Technical Bulletin
April 1986
Restoring the Statue of Liberty
Sharon Ofenstein
Inspiration and perspiration—each has played a major role in the current restoration of the Statue of Liberty. Only the use of innovative, high-tech processes and materials made the project possible. But the application of these methods and materials has required the most diligent, persistent—even tedious—workmanship.
Consider, for example, the effort to replace nearly all of the statue's approximately1,800 armature bars. (A few original bars were left in place, in the sole of the sandal on theright foot.) These 2- by 5/8-inch, ribbon-like iron straps form a grid conforming exactly tothe inner surface of the statue’s copper "skin," which measures 11,000 square feet. Theskin is attached to the straps by means of U-shaped copper "saddles," which are pieces of metal bent over the bars and riveted to the skin.
This system was commonly used for sculptures of this size at the time the statue was created, in the 1880s. However, the method whereby the grid of armature bars was supported was unusual. It seems to have been part of the ingenious internal structure designed for the statue by the brilliant French engineer, Gustave Eiffel. Rather than being hung from the interior structural members, the armature bars were affixed to the outer ends
of iron "flat bars" that projected upward, like struts, from the interior structure. Both the armature bars and the flat bars consisted originally of wrought iron, which was fairly supple. They allowed the skin to move in response to wind and thermal stresses.
System Fails
Nearly a century later, however, this system was failing. The statue had been maintained since its erection in 1886, but its large size and waterbound location tended to make care somewhat episodic. Nevertheless, the statue received over the years a series of repairs, as well as improvements to the visitor experience. An unusually thorough inspection in 1980, however, revealed that the iron of the statue's armature bars was
reacting galvanically with the copper of the saddles and skin. Water was entering the nterior of the statue through a variety of openings—in particular, at rivet holes, through the joints of the copper sheets, and around the deteriorated glazing of the flame of the torch. This water, made saline by the marine environment of New York Harbor, was acting as an electrolyte. It caused the iron of the armature bars, which was sacrificial to the copper of he saddles and skin, to corrode--everywhere that it was in contact with the copper. The rust ate into the armature bars, reducing their strength. It also caused many saddles to become detached from the skin. This occurred because the growth of rust was generating more volume under the saddles than had been occupied by the original iron bars at those points. The slow but powerful expansion of the rust under the saddles would pull the latter's rivets right through the skin. This left holes that admitted more water, hastening the process of deterioration.
The 19th-century craftsmen who manufactured the statue knew that such a galvanic eaction could develop. Traces of asbestos felt impregnated with shellac were found in the vicinity of the saddles; they appear to have been part of an attempt to isolate electrolytically the iron from the copper. The materials chosen, unfortunately, were inadequate for the task. Asbestos could not prevent the movement of moisture through it. The shellac had this capability, but it was short-lived. By 1911, water penetration was severe enough to prompt the application of a bituminous paint (coal tar) to the interior elements. A number of the armature bars were replaced, due to corrosion, in 1937-1938. By 1980, however, fully two-thirds of the approximately 1,800 armature bars were badly corroded, while all were affected to some degree. The iron flat bars supporting the armature bars were affected less, but some had buckled. It was clear that the existing destructive galvanic couple had to be
eliminated. This would require the replacement of virtually all of the iron armature bars and flat bars. This work would constitute the most extensive aspect of the centennial restoration campaign.
New Materials
Advanced technology proved invaluable in selecting suitable replacement materials. The metal for the new armature bars and flat bars would have to be malleable, to replicate the complex forms involved, and galvanically passive with the copper, to prevent corrosion. At the same time, it needed to be as much like the original wrought iron as possible to comply with the Secretary of the Interior's Standards for Rehabilitation and because of the indeterminate nature of the statue's structural-stress system. The convoluted shapes of the armature bars in particular made precise stress analysis difficult; it was not known to what degree the new bars could deviate safely from the old ones. Since wrought ron had worked well mechanically for almost a century, it was thought best to find a new material that would match its modulus of elasticity.
Responsibility for the choice of materials rested with the National Park Service's North Atlantic Historic Preservation Center (NAHPC). Discussions with metallurgists revealed that copper alloys would be slightly more passive with respect to the copper saddles and skin, but that iron alloys would behave more like the wrought iron. Extensive tests by a number of interested parties were conducted under the supervision of the
NAHPC. The copper alloys tested were stronger than the original wrought iron. In order to achieve the same modulus of elasticity, however, bars made of the copper alloys would have had to be thicker, and thus 30% heavier, than the original bars. Eventually, two types of stainless steel were selected. The flat bars, and the bolts of the armature bars' splice plates, would be made of Ferralium 225, a ferritic and austenitic (nonmagnetic) alloy. This duplex stainless steel exhibited minimal reaction with the copper; it offered thermal expansion and elasticity similar to that of the wrought iron, but was stronger. Ferralium would have been too difficult to shape into the complex forms of the armature bars, however. The metal selected for them was the extra-low-carbon, 316L stainless steel. It also exhibited minimal reactivity with the copper, and had a modulus of elasticity similar to the wrought iron. In addition, it had the same specific gravity as the wrought iron, and was stronger, but still easier to work with than the Ferralium. (The low carbon content of 316L makes it more makes it more ductile, as well as able to retain its resistance to corrosion through shaping treatments requiring high heat.)
As a further precaution, the armature bars and flat bars were to be sandblasted and
cleaned with nitric acid before installation. This would remove from the surfaces of the bars
bits of iron imbedded in them by the rolling mills—bits that would rust if not so treated.
Finally, Teflon tape was to be used to isolate electrolytically the armature bars from the
copper. The tape, which is backed with pressure-sensitive silicone adhesive, would be
applied to the side of the armature bars facing the copper, and to the inner surfaces of the
copper saddles. The Teflon has an indefinitely long lifespan, and a high resistance to the
transfer of ions necessary for galvanic corrosion to occur. It also is the industrial polymer
with the lowest coefficient of friction: it will hold up even during shearing motions. The
silicone adhesive is particularly resistant to oxidative degradation, which tends to be
associated with copper and copper oxide.
The advanced technology employed in the selection of replacement materials was of little use in the actual manufacture and installation of the bars. Traditional craft techniques were needed— albeit on an enormous scale that required unusual patience and persistence. Only four armature bars and their flat bars could be removed from any given area at one time, and only four such areas—far apart from each other—could be treated simultaneously. Thus, only 16 armature bars and their flat bars could be removed at a time.
The most complex armature bars, which required the use of a coal forge for the shaping of their replacements, were sent off Liberty Island to the workshop of Nab/Fiebiger—A Joint Venture. (These bars comprised approximately 200 of the 1,800 armature bars.) Less complicated bars were cold-worked by Nab/Fiebiger employees in the "restoration pavilion" that was set up on the island. Tools used included a 120-ton hydraulic press, an acetylene torch for heating difficult areas, and a variety of hand tools. The original bars had
to be stripped of any paint on them, measured, and then carefully duplicated in new, annealed metal. A bar could require edge bending, flat bending, twisting, or any combination of these. The workmen started with the most complex part of each bar, and then proceeded to form it out to both ends. Continual precise measurements and comparisons were required to replicate each piece. In either case—whether a bar was orged or cold-worked—it was removed, and its replica was made and installed within 36 hours.
Saddles and Skin
Closely related to the replacement of the armature bars was the replacement of the approximately 3,000 copper saddles, and the repair of specific sections of the copper skin. (Several original saddles were left in place on the original armature bars retained in the right foot.) The failure of the saddles' attachments to the skin, as explained previously, was caused by the expansion of the rusting iron armature bars. New saddles and rivets were
manufactured of toughpitch copper. They were reinstalled in the same locations as the old saddles, using the same rivet holes. On the whole, the copper skin is still in excellent condition. The deterioration of specific sections was caused by several factors. Comparative measurements were made by the NAHPC, using ultrasonic calipers, of the average thickness of the skin in many different areas. The original average thickness— 3/32nds of an inch—has eroded only 4/1000ths of an inch in almost a century, despite the harsh environment. However, certain original conditions and later alterations caused particular areas of the copper skin to deteriorate to the point where they had to be replaced. For example, some sections of the skin are much thinner than other sections, due to the heavy amount of hammering they underwent during the fabrication process. The metal of the torch's filigreed handle was only one millimeter thick when the statue was new. A hundred years later, erosion of this element had reached a critical point. Other, less highly worked areas of the copper skin also had deteriorated. Some of these were the result of the absence, blockage, or incorrect placement of "weep holes." These holes would allow water that did get into the statue to drain out. Wherever water collected and could not escape, metal-thinning corrosion occurred. This was the case with the tip of the nose, with three curls of the hair, and with the finial at the bottom of the torch's handle. The deterioration of the torch's flame, however, was caused by alterations made to it over the years, which inadvertently turned it into the statue's single worst source f water penetration. The flame originally was fabricated from solid sheets of copper. Immediately preceding the official dedication of the statue on October 28, 1886, however, two rows of holes were cut in the lower half of the flame, to permit its illumination from the inside via electrically powered arc lamps. An entire band of copper at the level of the upper row of holes was removed in 1892, and replaced with glass panels. Almost 25 years later, the entire surface of the flame was perforated with panes, creating a mosaic appearance. This form remained unchanged into the 1980s, but the inevitable failure of its elaborate glazing system allowed water to enter freely. The decision was made during the current restoration campaign to replace the flame, not repair it and lose its historic fabric. Also, repairing the mosaic-like glazing system would always be a liability, in a particularly inaccessible location. It was decided to remove the original, altered flame from the statue, and to display it in the lobby of the museum in the statue's pedestal base. A new flame would be made of gilded, solid-sheet copper formed to match the original design.
Working with Copper
Before any copper repairs were begun, however—even before any armature bars were replaced—the interior surface of the copper skin had to be stripped of the 1911 bituminous paint and a number of coats of other paint. Thermal, chemical, and mechanical means of emoval were tested, to see which could remove the deposits without harming the copper. The low-dust abrasive, aluminum oxide was to be used to strip the paint from the iron interior structure. This method would have been too harsh for use on the copper. The NAHPC proposed the use of liquid nitrogen for cryogenic removal. Application of this product did achieve embrittlement and shedding of the paint layers. (Similar experiments with liquid nitrogen both prior and subsequent to this work failed to obtain the same success.) However, a dark-colored copper compound, caused by the interaction of the
copper skin and the coal tar over time, remained. This had to be removed by blasting with bicarbonate of soda.
An essential component of the statue, both aesthetically and physically, is the green patina that has developed on the outer surface of the skin over the years. The new copper elements and the repair patches lacked this coloration. The flame of the torch was to be gilded, so that the new element was not a problem. The other repairs, however, would have been conspicuous. Thus, Les Metalliers Champenois treated the metal of the new torch with copper chloride, while the workmen of Nab/Fiebiger used copper sulfate on their repairs to achieve a greenish color quickly. This artificially accelerated patina will be supplanted gradually from behind by the patina that will develop naturally. The heads of all rivets were patinated with copper sulfate as part of their manufacture. The installation process frequently caused this layer to be knocked off. However, rivets that have lost their accelerated patina in this way still display a tendency to patinate at a rate faster than normal. Thus, the repair work soon will blend in with the statue's overall patina.
The condition of this overall patina has been a subject of intense concern during the current restoration. The presence of a stable patina is desirable, although it constitutes the corrosion of the outermost part of the statue, because it protects the rest of the metal below it. If the patina washes or blows off, new corrosion will take place on the exposed metal. Thus, the protection and enhancement of the existing patina has been a goal of the NPS
investigators. This patina is nonuniform in color and thickness, and the NAHPC undertook extensive studies to ascertain its condition. Closest to the copper skin is cuprite (cuprous oxide). Above this is a blackish layer composed mostly of brochantite (copper sulfate tribasic). This layer can be seen in places on the statue where the topmost, greenish-blue layer is absent. The greenish-blue layer also contains mostly brochantite, but with a
crystalline structure different from that of the blackish brochantite. The layer, in addition, contains antlerite (copper sulfate dibasic) in varying concentrations. The brochantite is a fairly stable product, and provides good protection for the copper statue. The antlerite is more porous, more soluble, and less protective.
The chief concern on the part of the NPS is the likelihood that the brochantite is being converted by acid precipitation (sulfuric acid) to antlerite, which tends to wash away. Examination of photographs taken in the 1950s shows a visible loss of the greenish color in some areas since that time. This is especially true on the north side, where weathering is most severe. Review of a study done in the 1960s suggests that the ratio of antlerite to
brochantite has increased significantly since then. Coatings and chemical corrosion nhibitors are commonly used to protect outdoor sculpture. Their use on the statue was considered but rejected, due to their short lifespan, coupled with the large size and general inaccessibility of the statue itself. The National Park Service will continue to monitor and analyze the condition of the patina to determine exactly what is taking place. Indeed, the NPS will continue to monitor and study many aspects of the Statue of Liberty long after the centennial restoration is completed. This launching of a long-term program of material and environmental investigation is perhaps one of the. most important egacies of the current repair work. As might be expected, new technology and traditional practices will be combined in the undertaking. This information will help the NPS make nformed, timely decisions about the care of the statue in the next 100 years and beyond. The author is a technical publications writer/editor for the North Atlantic Historic Preservation Center. The center is part of the Division of Planning and Resource Preservation, North Atlantic Region (NAR). It contains laboratories and analytical equipment, and is staffed by historic preservation conservators and exhibit specialists who provide technical support to the parks primarily within the North Atlantic Region. Ofenstein was aided in the preparation of this article by Blaine Cliver, Chief of Historic Preservation or the NAR; John Robbins, Historical Architect for the NAHPC; Ed McManus, Objects Conservator for the NAR; and Carole Perrault, Historic Preservation Conservator for the NAHPC.