5 COPPER CORROSION

Artifacts are often made of copper containing several percent of alloying elements. Corrosion of such alloys is complex due to the large number of corrosion products that can form in the biosphere. The following discussion is limited to copper objects that corrode in water sufficiently oxidizing and near neutrality to permit the formation of cuprite in the absence of SRB. This limitation excludes reducing environments in which copper does not corrode and acid environments in which copper dissolves. These restrictions eliminate consideration of corrosion in peat bogs, some hot springs, and ocean bottoms off the continental shelf.

The first product of copper corrosion, cuprite, forms epitaxially as a product of the direct reaction of copper with dissolved O2 or with water molecules (North and Pryor 1970). Cuprite has a high electrical conductivity and permits transport of copper ions (North and Pryor 1970) through the cuprite layers, allowing the copper ions to dissolve in water and reprecipitate. If the water chemistry approximates that of seawater, copper ions reprecipitate as botallackite (Cu2(OH)3Cl) (Pollard et al. 1989), which can convert in minutes or hours to either paratacamite or atacamite (alternate crystal structures of botallackite), depending on local water chemistry. Paratacamite is the more common form (Kato and Pickering 1984). If the water contains little Cl− but is in equilibrium with the atmosphere with regard to carbonate species, malachite will form. A direct solid-state transformation of cuprite to malachite is possible, while a transformation to a hydroxychloride mineral would require dissolution and precipitation (fig. 4). Malachite and hydroxychloride minerals normally precipitate from solution. If both Cl− and carbonate species are lacking, tenorite (CuO) is produced. The hydroxychlorides, atacamite and paratacamite, form in inhomogeneous layered structures. Malachite can form Liesegang structures layered with cuprite (Fisher and Lasaga 1981; Scott 1985). Malachite also forms botryoidal (Gettens 1969) or hairlike structures (Scott 1990b). The reasons for these differing morphologies are not known.

Corrosion of copper-nickel alloys in saline environments is much less rapid than corrosion of pure copper. Swartzendruber et al. (1973) proposed that this corrosion resistance was due to changes in the crystal chemistry of the base metal. This theory is no longer accepted. North and Pryor (1970) maintained that the improved corrosion resistance of copper-nickel alloys was due to changes in the mass transport characteristics of the oxide as a result of the addition of nickel. A comprehensive review of recent experimental work (Hack et al. 1986) indicates that, under hydrodynamic conditions that prevent accumulation of porous outer layers of hydroxychloride or hydroxycarbonate corrosion products, the critical effect of the alloying elements is to alter the cathodic oxygen reduction.

The mineralogy of the corrosion products observed during long-term corrosion of copper alloys is different from that commonly found on pure copper. Nickel compounds are observed in an adherent corrosion product layer on copper-nickel alloys. Copper-tin alloys tend to form layered structures with the characteristics of Liesegang structures (Scott 1985). Zinc tends to be selectively removed from copper-zinc alloys, producing a spongy copper material (Walker 1977), although the incorporation of zinc in corrosion products to produce rosasite ((Cu,Zn)2CO3(OH)2) is known (Gettens 1963b). Dealloying of the copper from tin bronzes has been reviewed by Geilmann (1956). Partitioning of alloying elements between remaining metal ions in the corrosion product and the electrolyte has received little attention (Zolotarev et al. 1987). Review of the corrosion processes of bronzes and brasses might lead to reconsideration of their original alloy chemistries.

If conditions at a copper surface permit precipitation of nantokite under the cuprite layer, the alloy becomes vulnerable to bronze disease (Scott 1990a) or type 1 pitting corrosion (Lucey 1967), depending on mass transport conditions. A biofilm containing acid-producing bacteria could help produce the requisite conditions by increasing acidity in anodic areas and reducing copper ion transport from the surface, resulting in a higher local copper concentration (Mohr and McNeil 1992). While nantokite-based corrosion can also occur by nonbiological paths, the nonisometric morphology frequently observed on precipitated cuprite crystals suggests that microbiological poisoning of growth planes is a factor. The presence of alloying elements does not protect against the formation of nantokite and the consequent bronze disease corrosion (Mond and Cuboni 1893; Scott 1990a).

In addition to nantokite, melanothallite (Cu2OCl2), eriochalcite (CuCl2�H2O), and other copper chlorides and hydroxychlorides may produce bronze disease-type reactions. The failure to identify bronze disease precursors other than nantokite may be due to their rarity as corrosion products, to kinetic or thermodynamic factors, or to a lack of adequate experimental study.

Copper alloys (except possibly arsenical bronzes) are also subject to sulfide-induced corrosion by SRB within a biofilm. Under these circumstances chalcocite forms easily, covellite less easily (fig. 3)(Syrett 1977, 1981; Kato et al. 1984). The corrosion layer may contain other sulfides buried under other corrosion products (Mor and Beccaria 1975; North and MacLeod 1986). Nonetheless, the poor adherence and mechanical properties of the sulfides make these layers nonprotective. The presence of iron ions in solution can lead to the formation of a layer of chalcopyrite (CuFeS2) (Daubr�e 1862; de Gouvernain 1875; Duncan and Ganiaris 1987). The circumstances under which chalcocite and covellite form and their relation to other sulfide corrosion products such as djurleite (Cu1.96S), and digenite (Cu5S9) have been addressed by McNeil et al. (1991).

Under some conditions copper phosphate corrosion products can form. Libethenite (Cu2PO4OH) and sampleite ((Na,K) Ca Cu5 PO4Cl�5H2O) have been identified by Geilmann (1956) and Fabrizi et al. (1989), respectively. Some research on the consequences of deliberately phosphating copper surfaces has been reported (Reiber 1989), but basic data on copper phosphate minerals are lacking.