QUANTITATIVE MODELING OF CORROSION PROCESSES ON COPPER ARTIFACTS, AND
ASSESSMENT OF THE RELATIVE AGE OF BREAKAGE OF COPPER ARTIFACTS
Jeffery A. Colwell, Ph.D. , Research Leader
Applied Metallurgy Department
Battelle Columbus Laboratories, Columbus, OH
with Commentary by
Christopher Carr, Professor
Department of Anthropology
Arizona State University
Tempe, AZ 85287-2402
Statement of Research Goals and Research Questions
This component of the project had two major goals. First was to evaluate theoretically and quantitatively whether the varieties of copper corrosion products found on individual headplates, breastplates, celts, earspools, and other copper items of the Ohio Hopewell were likely the result of fully natural, random, in situ corrosion or, instead, the product of some kind of intentional, cultural, artistic activity (e.g.., painting, patination). The second aim was to assess whether the broken edges of some copper breastplates were relatively modern or ancient in age, in order to determine whether the artifacts may have been decommissioned by breaking them into culturally prescribed forms found elsewhere in Hopewell art. This report is divided into two sections that address these two issues.
Corrosion on Copper Artifacts
The general problem of the natural or cultural origin of the copper corrosion minerals on the artifacts was approached with the following specific questions in mind:
(1) Under what soil conditions (pH, free energy, temperature, dissolved ionic species) are corrosion products of various mineral species formed?
(2) Are any of the kinds of copper corrosion products on the artifacts different enough in the conditions under which they form that they would not likely have developed together naturally, even though they might in actuality be found together on the copper artifacts?
(3) What is the length of time required to form the various kinds of copper corrosion products found on the artifacts, assuming common soil conditions in the Scioto valley of Ohio? Could copper corrosion minerals be culturally induced to form (e.g., as in patination) in a short amount of time that could represent a single artistic venture (a few days to months), or a staged venture within a charnel burial house (months to years)?
(4) Are there scenarios that can be envisioned whereby simple, naturally occurring substances (i.e., natural salts and acids) readily available to the Hopewell might have been used to intentionally create corrosion products to make artistic imagery, perhaps with the aid of heat, water, and/or some other accelerating method?
Corrosion Products on Copper Buried in Soils
The corrosion of copper in soil is a complex interplay among many variables. Corrosion is an electrochemical process that requires an electrolyte, i.e., water, to be present in at least minimal amounts to provide some conductivity within the soil. The corrosivity of a soil is a function of its water content, but also the local mineral content and the dissociation of those compounds into cationic and anionic species. Given the variability of the water content of a Midwestern soil throughout the year as a function of season, the corrosion rate of buried artifacts is a process that is not continuous and in fact varies over time.
The development and stability of corrosion products on copper thus depends on many local environmental variables: e.g., pH, the composition of dissolved ionic species, temperature, and time-of-wetness, to name just a few. Notwithstanding the expected variation over time in the macro-environment that is present over an entire copper artifact in relation to general soil composition and soil water conditions, there may also be variation in the micro-environment across the surface of the artifact. This variation may result from materials (e.g., textiles, feathers, hide, paints) that were intentionally fixed to the artifacts, or that were fortuitously in contact with them, or that were in the native soil but different significantly in chemistry from the surrounding soil. Such materials may hold water differently than the surrounding soil, allow dissolution of ionic species that would contribute to local chemical or electrochemical processes, or may themselves participate in reactions on the surface of the copper.
Corrosion behavior is comprised of two factors: (1) the compounds, if any, that form and (2) the rate of that formation. The rate of formation of those compounds on the sample of artifacts of concern would be very hard to predict, given probable variations in environmental conditions over long exposure time. However, the kinds of copper compounds that might be expected to develop can be evaluated using thermodynamics.
Thermodynamics can be used to show what compound, metal, or cation of a metal, is stable under what defined set of conditions. One method of visualizing this aspect of the corrosion behavior of metals was pioneered by Pourbaix. The method uses graphical representations, which are called Pourbaix Diagrams today. For our purposes, the copper-water diagram at 25 C° is of most importance and is shown in Figure 1. The sloped dotted lines in Figure 1 represent the zone where water is stable, which bounds the potential for this case. At potentials lower than the bottom dotted line , water decomposes, generating hydrogen. Above the upper dotted line, water decomposes to generate oxygen. As can be seen, the diagram maps the range of pH and free energy (as potential, E) where copper metal and copper corrosion products are stable.
Diagrams such as these are constructed purely from Gibbs free energy data and represent the expected compounds at equilibrium under those conditions. Thus, the formation of specific compounds will occur only under the specific mapped conditions of pH, potential, and temperature. There are computerized techniques today that can be used to extend the Pourbaix format to multiple components, but only if data from the additional compounds that might form are available.
An important distinction to be made from the diagrams is that only one compound can be stable in one zone. Two compounds can coexist only along lines between the stable phase fields, which is not very likely if they form as corrosion products, given the requirement that the conditions remain constant at that very specific pH and/or electrical potential. For both adjacent compounds to be in equilibrium, excess quantity of each must be present to allow for buffering the aqueous phase and fixing the pH, for example. This could occur practically only by the addition of both compounds intentionally, because it is likely that soil conditions would not be so highly variable over small distances, and even small variations would be averaged as the soil components dissolved in water on the surface of the artifact during corrosion.
A suite of Pourbaix diagrams was constructed for the addition of multiple species to water in order to determine whether specific compounds might have formed on the artifacts as a result of corrosion. These are provided in Figures 2 through 9. The Cu-H2O diagram baseline case shows that for near-neutral pH’s of soils (between 7 and 7.5 at the Hopewell site, for instance), copper metal, or either one of the two oxides of copper can be stable, depending on the oxidizing power of the water phase. However, the potential can be assumed to be higher than the Cu/Cu2O dividing line because bright, uncorroded copper metal (Cu) is not found on the surfaces of artifacts within Ohio Hopewell archaeological sites. Cu2O is found, instead. Furthermore, oxides are in general very stable thermodynamically, and more so than other compounds. Thus, when adding additional components to calculate new diagrams, it is not surprising to see the oxides of copper still to be the dominant stable phase in the range of pH of archaeological interest.
One exception to the formation of copper oxides is when carbon, which forms carbonates, is added to the chemical system(Figure 10). In this case, malachite (CuCO3)(OH)2) is the thermodynamically favored phase in addition to Cu2O. Indeed, malachite and cuprite are the two most common corrosion products found on the surfaces of Hopewell copper artifacts (see Part I). Note, however, that these phases are predicted from thermodynamics to coexist at only a very specific potential at a very specific pH, which is statistically unlikely.1
The importance of these types of diagrams is very significant to the interpretation of the compounds found and identified on the Hopewellian plates. As stated previously, each zone on the diagram shows a zone of thermodynamic stability of a single compound. In the case of a corrosion reaction, for example, copper metal would form either a single oxide or sulfide or carbonate of copper, depending upon the potential and pH. Consequently, if one or more compounds are found in the same location on a copper plate, it is not likely that they resulted from naturally occurring and random corrosion reactions. Other formation processes, such as painting with multiple copper corrosion pigments, or artistic patination, would be suggested. If corrosion had indeed caused the formation and distribution of those different minerals, then it would indicate radically different environmental conditions in very close proximity on the surface of the plate. This situation is highly improbable in nature.
One might postulate that another plausible hypothesis for natural occurrence of adjacent compounds on one plate: that environmental conditions changed over time and allowed different equilibrium phases to form in close proximity. However, had this occurred, with for example an initially stable oxide condition then altered to a carbonate- stable condition, then the oxide would become unstable and would be converted to carbonate at equilibrium, thus leading to a single phase after some period of time.
In summary, by using thermodynamic arguments, it has been shown that random corrosion processes is not a reasonable explanation for the observed diversity of copper compounds on the Ohio Hopewellian artifacts. A reasonable explanation of the diversity is that specific minerals were added at specific locations to generate the artistic imagery, i.e., paints. Another possibility is that local micro-environments on a plate were intentionally changed to create the imagery using readily available chemicals, or materials that had the same effect, i.e., artistic patination.
Independent experiments have been conducted by Christopher Carr and David Pimentel at Arizona State University on bare copper to test the hypothesis that “forced corrosion” (i.e., patination) could have been used to alter the copper plates and form a diversity of specific corrosion compounds, thus creating the imagery. The results of that investigation answer to the positive, and will be discussed elsewhere.
In order to determine whether the hypothesis of intentional prehistoric breakage is true, or whether alternatively the breaks were recent and fortuitously matched known artistic patterns, twenty copper artifacts that had broken edges and that appeared to represent culturally prescribed forms were examined. The sample included sixteen breastplates and four headplates (Table 3.1). The surfaces, edges, and internal fractures of the plates were examined using low power microscopy (about 10-20X).
Twenty artifacts were examined and, for the most part, the fracture edges were judged to be corroded. Some fresh, i.e., modern, fractures were seen (indicated by bright uncorroded copper), but primarily in small locations or corners where pieces of the object had undoubtedly been broken during the archival of the objects. There were also several cracks observed that have not fractured through, but were clearly “old”. The old and new cracks and fractures have been noted in the figures on file at Arizona State University with the PI.
The corrosion products identified by this visual examination of the edges of the artifacts were cuprite and malachite. In most cases these were mixed along the edges, but generally in zones where either one or the other was predominant. Referring back to Figure 10, it can be seen that these two compounds would be the ones expected under long term exposure to near neutral pH water if carbonates were the dominant soil species.2
The fact that only malachite and cuprite were seen on the fracture edges indicates the very strong likelihood that the edges corroded in the soil according to the expected thermodynamics, and that the edges were fabricated to specific artistic shapes prior to burial of the artifacts.
2In addition, from the Pourbaix diagram in Figure 10, one would not expect pure copper metal to ever co-exist with malachite under natural conditions In other words, malachite does not form directly on the surface of copper, because the diagram shows that there are no conditions where copper and malachite are in equilibrium. Instead, as copper is converted to malachite, there will always be at least a thin layer of Cu2O between the uncorroded copper and malachite. There was one plate in particular, B060, where an edge fracture could be seen that confirmed this. An oblique view of the edge showed malachite on the surfaces above and below the copper plate, but with cuprite between the copper plate and malachite. Other examples of this stratigraphic relationship are given by Nicoll in Part I.