by Lyndsie Selwyn, Senior Conservation Scientist, Conservation Processes and Materials Research
Have you ever wondered why bronze statues turn green?
Before I came to CCI, someone told me that copper carbonates were responsible and I saw no reason to doubt it. After all, Ottawa is not an industrial town and is not close to an ocean, so the green compounds were unlikely to contain sulphur or chlorine. And, as I later learned, archaeological bronzes are typically covered with the copper carbonates malachite (green) and azurite (blue). But within days of starting work at CCI in 1987, I became involved in a project to study the material on outdoor bronze statues on Parliament Hill, and to my amazement I soon discovered that the green was indeed caused by those unlikely copper sulphates and chlorides.
Sixteen outdoor monuments containing more than 50 individual cast bronze statues surround the Parliament Buildings in Ottawa. Some of these monuments have been outside for more than 100 years. The oldest monument, that of Sir Georges-Étienne Cartier, was unveiled in 1885 and the most recent one as of 1987 was John Diefenbaker, unveiled in 1986. By 1987, many of these had disfiguring black areas and uneven streaking, and the older ones had turned green.
As part of this conservation program, a project was initiated between PWGSC and CCI to collect and analyse representative surface samples from the statues before their conservation.
Samples were collected from the outer surface of the bronze statues and 168 were analysed by powder X-ray diffraction to determine the crystalline components. Twenty-four of these samples were further analysed using Fourier transform infrared spectroscopy to characterize any organic materials present.
We identified 94 crystalline compounds and 9 organic materials. Table 1 lists a few of these by name and chemical formula, together with the number of occurrences of a particular compound in the 168 surface samples. The table has been divided into ‘surface corrosion' (corrosion products most likely produced when the bronze alloy reacted with the local environment) and ‘surface contamination' (extraneous materials most likely produced from wind-blown debris, casting residues, or applied coatings).
Changes to the surface of bronze depend on the local environment—temperature and humidity fluctuations, wind-blown material, and even bird droppings (Figure 1)—but the main corrosion products forming after prolonged outdoor exposure are copper compounds because bronze contains mostly copper with small amounts of tin, zinc, and lead. When copper first corrodes it becomes covered with a layer of copper (I) compounds and we detected cuprite, a common copper (I) compound, in many of the samples. After a while, this layer becomes covered with another layer containing copper (II) compounds. We detected many copper (II) compounds, mainly green copper sulphate hydroxides (brochantite and antlerite) and green copper chloride hydroxides (atacamite and paratacamite). These are the compounds that give weathered bronzes their green appearance. But why did these compounds contain the unexpected sulphur and chlorine, and not the expected carbonate?
It is high levels of sulphur and chlorine around the statues that encourage the formation of copper sulphates and chlorides rather than copper carbonates. In Ottawa, the sulphur comes from acid precipitation caused by pollutant gases (such as sulphur dioxide and nitrous oxides) that react with water to form sulphuric acid and nitric acid. The sulphur dioxide originates from industries in the Ohio River Valley in the United States and Sudbury in Canada, and from 1979 to 1994 the average annual pH in Ottawa was between 4.2 and 4.4. Although chloride-containing compounds are usually associated with marine environments, in Ottawa they originate from the salt used to melt road ice. More than half of the surface samples we tested contained some chloride. Only one sample, however, contained nantokite, the copper (I) chloride CuCl often associated with ‘bronze disease'. Nantokite has been identified on archaeological bronzes, usually next to the metal surface, but it has rarely been identified on outdoor bronzes.
Many crystalline materials identified in the analysis are not associated with bronze corrosion, and were most likely surface contaminants. The most important of these was quartz, found in 75% of the surface samples. The quartz could be residual material from the casting process, but could also be from wind-blown sand, suggesting the statues are being slowly sand-blasted by their environment. Another find was gypsum, a calcium sulphate. Gypsum is often used during casting to fill the inner core of the mould. The foundry usually removes this core material, but it is impossible to remove it completely. Because gypsum is slightly soluble in water, some of it will dissolve in the moisture that inevitably condenses inside hollow statues. Wherever this water flows out of cracks, holes, or casting flaws in the bronze, it leaves a residue, and over time a deposit can build up. These deposits are usually not the expected white colour of gypsum, but instead are gray or green because they are ontaminated with impurities, usually copper (II) compounds.
We detected several organic materials. There were waxes (mainly beeswax and paraffin wax), a drying oil which in a few cases was identified as linseed oil, and a silicone oil called poly(dimethyl siloxane) which is often added to furniture waxes. These had probably been applied as protective coatings in the past. We also identified both a calcium and a copper oxalate, formed by reactions with oxalic acid which is secreted by microorganisms such as lichen and, in urban areas, is abundant in fog and rain. Finally, we identified uric acid and its hydrate, which are most likely from bird excrement.
Bronze corrosion requires contact with water, and if salts are dissolved in the water, corrosion is faster than if no salts are present. The materials identified in our surface samples possess a range of solubilities: some are soluble or slightly soluble in pure water; others are essentially insoluble in water but become more soluble as acidity increases. The presence of these water-soluble salts makes it important to carry out proper surface cleaning to remove them during the initial stages of a conservation treatment. Furthermore, the application and maintenance of a protective coating on bronze is important because it limits contact between the bronze and water, thereby slowing corrosion.
The results from this study document an impressive range of compounds on the surface of outdoor bronzes. Although such results rarely alter the conservation treatment, they are important in confirming the need for a treatment that includes both surface cleaning and the application of a protective coating.
I would like to give credit to the people who played an important role in this project: the scientists hired on contract (David Downham, Marilyn Laver, Jacques Poitras) to collect and analyse the samples, Nancy Binnie of CCI who kept track of the data, and the staff (past and present) of CCI's Analytical Research Laboratory, whose help, analytical expertise, and advice were invaluable in interpreting and presenting the results for publication. Thanks also to the members of the Curatorial Committee and the staff of Craig Johnson Restorations for cooperation and advice.
