Page 71 - Copper and Bronze in Art: Corrosion, Colorants, Getty Museum Conservation, By David Scott

brasses; and nickel sulfate on nickel-silver alloys in urban atmospheres. In an interesting Ger­
man study Riederer (1972 a,b) examined the corrosive deterioration of more than two hundred
bronze statues erected before the end of the nineteenth century. The study found that leaded
bronzes were more rapidly attacked, while the common statuary alloys of copper-tin-zinc were
more resistant.
Holm and Mattsson (1982) found that brasses with high amounts of zinc (24-40%) were
susceptible to dezincification; the most severe attack was found in the urban atmosphere, the
least severe in the rural atmosphere. For brasses with less than 15% zinc and for arsenical
alpha brasses, no significant dezincification was found. Deeper loss of zinc was noticed in the
alpha+beta brasses, mainly in the beta phase, as would be expected, compared with those
brasses that were in either the alpha range or the beta range of composition. The depth of attack
in the alpha brasses did not show any significant decrease with decreasing zinc content, but the
type of corrosion attack gradually changed from selective to general corrosion of the surface
with zinc contents below 15%. For the majority of copper and copper alloys, the decrease in ulti­
mate tensile strength was found to be less than 5%, and in elongation, less than 10%. Around the
time when Holm and Mattsson began their study in Sweden, Costas (1982) began an atmo­
spheric study at thirteen different test sites in the United States for periods lasting from fifteen
to twenty years. Although this study does not appear to amplify the conclusions reached by
Holm and Mattsson, it found that the rate of corrosion of the copper alloys at four locations after
these time periods varied from a maximum of 2.3 μιη per year to a low of 0.22 μιη per year. The
corrosion rate was the highest at the industrial sites, and the green sulfate patina developed
there only. Rural and urban sites developed cuprite patinas only after twenty years; and the
basic copper chloride, paratacamite, was detected in patinas that developed in marine locations.
These studies, while useful, present only a simple, linear picture of corrosion development com­
pared with the more complex development of the corrosion crusts that are found on bronze stat­
ues exposed for much longer periods of time, as the data from Vernon's work in the 1930s
illustrates (see TABLE 1.4).
A useful study of thickness profiles of patina development over time was initiated by Franey
and Davis (i987), who examined the thickness of copper alloy patinas sampled in the New York
City metropolitan area. The samples were taken from copper alloys that had been exposed for
periods from one to almost one hundred years, during the period of 1886-1983. The data show
that most of these patinas achieved a thickness of approximately 12 μιη in about fifty years, and
probably not much more than that in one hundred years. This gives a corrosion rate for urban
areas of 0.12-0.24 μιη per year, substantially less than the initial corrosion rates determined
experimentally over shorter periods of time and, incidentally, far lower than the rate observed
by Tracy for the Christ Church roof in Philadelphia. Deducing corrosion rates from patina
thicknesses developed during outdoor exposure may be misleading, however, since the mass
loss from the corrosion events is not known, especially over long periods of exposure.

C H A P T E R O N E
54