#316 — The Patina

Seeds: Pilling-Bedworth ratio (13851), Statue of Liberty copper patina (13852), timber charring rate (13853), yakisugi technique (13854), iron oxide triple-layer failure (13855), passivation thesis (13856). 6 source nodes across metallurgy, architectural history, fire engineering, and corrosion science.

On October 28, 1886, the Statue of Liberty was unveiled in New York Harbor. Its copper skin — three hundred sheets, each 2.4 millimeters thick, hammered over wooden forms by the repoussé technique — was the color of a new penny: shiny, reddish-gold, reflective in afternoon light. Frédéric Auguste Bartholdi expected it to darken. He did not expect it to turn green.

The transformation proceeded in three chemical stages. First, copper reacted with atmospheric oxygen to form cuprite — copper(I) oxide, Cu₂O — producing a reddish-brown darkening visible within the first years. Then the cuprite further oxidized to tenorite — copper(II) oxide, CuO — turning the surface black. Finally, the tenorite reacted with moisture, sulfur dioxide from coal-burning industry, and carbon dioxide to produce brochantite — Cu₄SO₄(OH)₆ — the basic copper sulfate responsible for the characteristic green. By approximately 1906, twenty years after installation, the transformation was essentially complete.

In 1906, Congress was alarmed enough by the discoloration to appropriate sixty-two thousand dollars for restoration. The Army Corps of Engineers studied the patina and reversed the intuition: the green layer was not decay. It was protection. The cuprite base layer bonded directly to the copper surface — dense, adherent, chemically stable — and the brochantite above it sealed the cuprite from further atmospheric attack. Removing the patina would expose fresh copper to the cycle again. The corrosion product was the corrosion barrier.

When the statue was comprehensively restored between 1984 and 1986, the exterior copper and its patina were again left untouched. The interior told a different story. Gustave Eiffel's iron armature — approximately 1,800 puddled-iron saddle bars connecting the copper skin to the structural framework — had corroded catastrophically. Shellac-soaked asbestos cloth, originally insulating the two metals, had disintegrated over a century, allowing salt air and moisture to create a galvanic cell. Two-thirds of the bars were badly damaged. They were replaced with 316L stainless steel, selected after four and a half years of accelerated corrosion testing simulating a further century of exposure.

After more than a hundred years in one of the harshest atmospheric environments on earth — industrial pollution, salt spray, acid rain, freeze-thaw cycling — the copper had lost less than 0.127 millimeters. Five percent of its original thickness. The iron, carrying less structural load and receiving less direct exposure, had lost two-thirds of its integrity. Two metals. One environment. The copper corroded and survived. The iron corroded and was consumed.

In 1923, N.B. Pilling and R.E. Bedworth published "The Oxidation of Metals at High Temperatures" in the Journal of the Institute of Metals. Their question had practical urgency — which metals could survive high-temperature service — and their answer was geometric. When a metal oxidizes, the oxide occupies a different volume than the metal it consumed. Pilling and Bedworth calculated this ratio for a range of metals and found that it predicted, almost perfectly, which oxides were protective.

The Pilling-Bedworth ratio is the volume of the metal oxide produced divided by the volume of the metal consumed. Aluminum oxide — Al₂O₃ — has a ratio of 1.28. The oxide is twenty-eight percent larger than the metal it replaced. This produces a gentle compressive stress that keeps the film dense, coherent, and bonded to the surface beneath. Copper's cuprite — the base layer that protects the Statue of Liberty — has a ratio of 1.64. Nickel, 1.66. Titanium, 1.77.

Below a ratio of one, the oxide is smaller than the metal it consumed. The film cannot cover the surface. Cracks and pores expose fresh metal to continued attack. Sodium, at 0.54, and potassium, at 0.47, corrode so rapidly in air that they are stored under oil or argon. Their oxides are too small to seal the wound they came from.

Between one and two, the oxide fits. The compressive stress is strong enough to keep the film coherent but not so strong that it buckles. This is the narrow corridor — a factor-of-two geometric window — in which the product of damage happens to have the right size to prevent further damage.

Above two, the oxide is too large. The compressive stresses exceed the mechanical strength of the film. It buckles, cracks, spalls off, and exposes fresh metal. The process repeats. The damage produces a product that cannot stay in place, and so the damage never ends.

The corridor is narrow. The chemistry is the same in every case — metal atoms lose electrons to oxygen. What determines the outcome is the geometry of the result.

In 1967, E.L. Schaffer of the USDA Forest Products Laboratory published "Charring rate of selected woods — transverse to grain," establishing a number that would become the foundation of timber fire engineering. Under standard fire exposure, softwood chars at approximately 0.65 millimeters per minute. The rate is remarkably consistent across species and conditions — consistent enough to be codified in Eurocode 5 and the International Building Code as a structural design parameter.

The char layer that forms is a ruin. The wood's cellulose has decomposed. Its volatile compounds have been driven off as combustion gases. What remains is porous carbon — a skeleton of the original cellular structure, drained of everything that could burn. And this ruin is an insulator. The thermal conductivity of the char layer is roughly one-third that of intact wood. The temperature gradient across a burning timber member is so steep that at thirty-five to forty millimeters behind the fire surface, the wood is at ambient temperature. Full strength. Full stiffness.

This produces a paradox that building codes have formalized. An unprotected steel beam exposed to a standard fire reaches 550 degrees Celsius — the temperature at which it retains only sixty percent of its yield strength — in fifteen to twenty minutes. The steel does not burn. It softens. It has no mechanism to produce a barrier between itself and the heat. A heavy timber beam sacrifices its outer surface to create the insulation that protects its structural core. The combustible material outperforms the non-combustible material because the combustible material passivates and the non-combustible material does not.

The principle was understood before the engineering. In eighteenth-century Japan, during the Edo period, a technique called yakisugi — charred cedar — involved deliberately burning the surface of cryptomeria boards with stacked bonfires. The pre-charred surface resisted water, rot, insects, and further fire. The method was considered low-class, a merchant's expedient for warehouse stock. It is deliberate pre-passivation: inflicting the damage before the environment can, so that the protective product is already in place when the threat arrives. Eurocode 5 formalizes the same logic as the effective charring depth — charring rate multiplied by time, plus a seven-millimeter zero-strength layer — and permits the remaining cross-section to be designed at full capacity.

The counter-case is iron.

In the 1830s, Michael Faraday discovered that iron could be passivated. He placed iron in dilute nitric acid and observed vigorous dissolution — the metal dissolving, hydrogen evolving from the surface. He then placed iron in concentrated nitric acid, expecting faster attack. The metal sat inert. When he scratched the surface, a brief burst of bubbles appeared and then ceased. The concentrated acid, being a powerful oxidizer, had forced a thin magnetite film onto the surface. The stronger acid produced a stronger defense.

In the atmosphere, iron cannot sustain this defense. The reason is that iron forms not one oxide but three, layered from the metal surface outward: wüstite (FeO, Pilling-Bedworth ratio 1.77), magnetite (Fe₃O₄, ratio 2.10), and hematite (Fe₂O₃, ratio 2.14). The innermost layer — the one in direct contact with the metal, the one that should serve as the primary barrier — is thermodynamically unstable below 570 degrees Celsius. At ambient temperature, wüstite decomposes into magnetite and metallic iron through a eutectoid reaction, producing a porous, incoherent mixture at the most critical location. Furthermore, wüstite is non-stoichiometric: its actual composition is Fe₀.₈₃O to Fe₀.₉₅O, with up to seventeen percent of its iron lattice sites vacant. These vacancies form a fast diffusion pathway for iron ions moving outward toward the oxygen. The oxide that should block further oxidation is a highway accelerating it.

The three oxide layers have different crystal structures — cubic, inverse spinel, rhombohedral — with thermal expansion coefficients that diverge by up to thirty percent near their respective magnetic transition temperatures. Any temperature change creates shear stress at the interfaces between layers. The outer hematite buckles and spalls. The middle magnetite cracks. Fresh iron is re-exposed. New oxide forms, buckles, spalls. The cycle continues until the piece is consumed.

Faraday could passivate iron by forcing it, electrochemically, into a state where the protective film was stable. Friedrich Flade, in 1911, identified the exact potential at which this passivation breaks down — the boundary between two stable states of the same metal. In concentrated nitric acid, iron sits above the Flade potential. In the atmosphere, it sits below. The chemistry is identical. The outcome is opposite.

The same process — oxidation — protects copper and aluminum and destroys iron, and the difference resides entirely in the product. The chemistry of the damage is the same: metal atoms lose electrons to oxygen atoms. What differs is the geometry, the phase stability, and the adhesion of what the damage leaves behind. Aluminum's oxide is twenty-eight percent larger than the metal it consumed, producing a gentle compressive stress that keeps the film sealed. Iron's oxides are seventy-seven to a hundred and fourteen percent larger, and layered across three incompatible crystal structures, and the innermost one is unstable at the temperature where it is needed.

The timber case extends the principle beyond chemistry. Char insulates because it has already been consumed — it contains nothing further for the fire to take. The sacrifice is complete at the surface, and the completeness is what produces the barrier. Yakisugi is the engineering conclusion: if the product of damage is protective, inflict the damage first.

The principle is not that damage heals itself. Damage does not heal. The oxide is not the original metal. The char is not the original wood. The patina is not the original copper. What forms is a different substance — a ruin of the original, transformed by the process that destroyed it — that happens, by geometry or thermodynamics or the physics of heat transfer, to prevent the process from continuing. Stability is not the absence of damage. It is the condition in which the damage has produced something that fits.