Releases and fluxes of copper

Corrosion of metals forms part of the natural cycle in which the metal is striving through spontaneous chemical and electrochemical processes to reach back to its most stable condition, i.e. as a mineral, from which it was originally refined. When copper sheet is used as a roofing material, it undergoes gradual degradation, induced by wet and dry deposition of environmental pollutants, such as SO2, NOx, O3, HCl and NaCl. A prerequisite for a corrosion process to occur is the presence of water, e.g. humidity, dew, fog or precipitation. He (2002) has made a survey of the most common corrosion products, formed on a copper roof in an urban environment:

The product formed first (within minutes or hours) is cuprite, Cu2O, which may transform into the intermediary amorphous copper sulphate and then - after months or years - into any of the two (greenish) copper sulphate hydroxides, posnjakite and brochantite (usually known as "patina"). In chloride-containing environments such as road traffic or marine atmospheres, the copper chlorides nantokite (CuCl) and atacamite (Cu2Cl(OH)3) may form. Typical corrosion rates, expressed in ^m/y, have been given as ~ 0.5 in rural environment, ~ 1.0 in marine, 1-2 in urban, and < 2.5 in industrial environment (Leygraf, 1995).

When both corrosion rate and runoff rate on fresh copper sheet are expressed in g/m2, y , the relationship between the two, when measured over 48 weeks of exposure to the urban atmosphere in Stockholm (annual precipitation 575 mm, average atmospheric concentration of SO2 = 3 ^g/m3 and of NO2 = 50 ^g/m3), is as follows: the annual corrosion rate was 6.7 g Cu/m2, y and the runoff rate 1.3 g Cu/m2, y (He, 2002). Exposure of a naturally (in situ, in the urban environment) patinated copper sheet of an age of 100 years, resulted in a somewhat higher annual runoff rate: about 2.0 g Cu/m2, y. In laboratory follow-up experiments it was shown that the "first flush" (rain on a dry surface) contributed to the total runoff with a higher percentage (44%) from the 100-year old copper plate than from the fresh copper plate (18%).

Additional measurements (Karlen et al., 2002; Odnevall Wallinder and Leygraf, 2001) have confirmed that runoff rates of copper from the urban testing site in Stockholm varied between 1.0 and 1.7 g Cu/m2, y for fresh copper (<5 years old) and between 1.3 and 2.0 g Cu/m2, y for green, naturally patinated copper (>30 years old). The variation within each category of panel age was mainly attributed to differences in annual precipitation quantity. The reason for the discrepancy in runoff rates between fresh and old copper plates (primarily during the first flush) was due to the morphology and other surface characteristics of the corrosion products within the patina. For example, the adsorption capacity of water was higher in the porous structure of green-patinated plates. Hence, corrosion and dissolution can take place even during relatively dry periods, since water is trapped within the patina. Therefore, the first flush effect is more pronounced on green-patinated copper compared to brown-patinated copper (Karlén, 2001; He, 2002).

During extended exposures for more than four years of both fresh copper sheet and naturally patinated, 130 years old copper sheet in the Stockholm atmosphere, further confirmation was obtained that the annual runoff rates were 1.1 - 1.6 g Cu/m2, y and 1.6 - 2.0 g Cu/m2, y, respectively (Odnevall Wallinder et al., 2002a).

In order to assess the influence of the amount of precipitation and of the air quality on the runoff rate, a comparative study was set up with simultaneous copper sheet exposures in Singapore and Stockholm. The total annual rainfall is 5-8 times higher in Singapore compared to Stockholm and also the rain frequency and duration of humid periods are much higher. Moreover, the atmospheric concentration of SO2 is 7-8 times higher (about 20 ^g/m3) in Singapore compared to Stockholm. After about one year of exposure, it was found that runoff rates from fresh copper sheet in Singapore were 5.6-5.7 g Cu/m2, y. At the same time, the copper runoff from pre-patinated copper sheets was estimated, and it was found that although the Singaporean runoff rate from brown pre-patinated sheet was almost identical with that from fresh copper sheet, the runoff rate from green pre-patinated sheet was significantly higher, 8.4-8.8 g Cu/m2, y (Odnevall Wallinder et al., 2002a).

Previous calculations carried out in 1998 (L & L - Cu) came to the conclusion that the total amount of copper primarily washed off from copper roofs and copper-faced façades in Stockholm would be in the range 1.1 - 1.3 t/y. Now, with the more refined measurements of the runoff from copper roofs of different ages and with a more sophisticated assessment of the total surface of copper coated buildings in Stockholm, based on digital air photo processing (Ekstrand et al., 2001), it turns out that the new and refined estimates end up with a somewhat smaller figure: on average, 1.0 t of copper leaves the copper coated roofs every year.

It should, however, be clearly recognized that the indicated quantity of copper is what is released from the roofs in the city. In order to assess the further fate of this quantity of copper, it is important to undertake a thorough speciation of the metal, in particular determine the reactivity, mobility and bioavailability of the released metal. This has been the subject of a licentiate thesis, prepared by one member of the KTH team (Karlén, 2001), and of one of the included papers, specifically dealing with copper (Karlén et al., 2002).

Samples were taken at the edge of the roof, equipped with either a fresh copper panel or with a naturally patinated, 30 years old, copper sheet, which were exposed to the Stockholm atmosphere for 3 years. The methods used for speciation of copper were:

• measurements of the cupric ion concentration using a ion selective electrode, ISE (ORION Model 9629);

• computer modeling, using the water-ligand model MINTEQA2 (Allison and Brown, 1991), with input data of pH, alkalinity and concentrations of sulphate and chloride in the runoff water;

• use of the cell-based biosensor BIOMET (a gene-modified strain of Alcaligenes eutrophus), specifically developed for detection of bioavailable copper (Corbisier, 1997);

• application of the OECD "Algal Growth Inhibition Test" to detemine the copper concentration in the runoff causing 50% inhibition of algal growth, and then relate this concentration to the total concentration of copper in the runoff water.

Karlen et al. (2002) confirmed the previous, repeatedly found, results that annual runoff rates varied between 1.0 and 1.5 g Cu/m2 for naturally patinated copper sheets of varying age. The concentration of total copper in the runoff water samples ranged from 0.9 to 9.7 mg/l. Both computer modeling and direct measurements by means of ISE and the biosensor revealed that nearly all the copper in the runoff water, sampled directly at the edge of the roof was bioavailable. Some 60-100% of the released copper was present as free hydrated cupric ion, Cu(H2O)62+, which is considered the most bioavailable copper species. Other copper species in the runoff water, such as Cu(OH)- and Cu2(OH)22+, are also bioavailable.

Testing of toxic effects on the alga Raphidocelis subcapitata, during standard exposure for 72 hours, demonstrated that the toxicity of the runoff water leaving the roof was high: the EC50 value ranged from 0.29 to 0.67 vol-% (150-350 "Toxic Units", TU = 100/ EC50). When the 72-h EC50 values for the runoff water were expressed in ^g of total Cu per litre, it was clearly demonstrated that the range (6-24 ^g Cu/l) was very close to the 72-h EC50 reference value for CuSO4 (18-27 ^g Cu/l), mainly representing free cupric ion. This means that the same chemical species, or species with the same degree of bioavailability, would have been present in the runoff water.

After having provided a satisfactory description of the copper species occurring in the runoff water immediately after leaving the roof, the next step is to evaluate what transformations of these copper species may take place, when the runoff water comes into contact with various sorbing materials and solid surfaces on its way to the STP or to the final receiving body of water. It is assumed that - during further transport in roof gutters, downpipes, street gutters and sewers - copper undergoes changes in chemical speciation, which may result in immobilisation of some of the copper, thus reducing the amount transported further on, and possibly reducing the bioavailability of the remaining amount.

These aspects have been studied by Bertling and co-workers, first in a laboratory study, where the capacity of limestone to retain and immobilize copper was investigated (Bertling et al., 2002a). A series of follow-up field and laboratory investigations were then conducted, where runoff water, containing copper, was passed through reactors containing various sorbing surfaces, such as limestone, soil or concrete, that mimic the normal situation prevailing where runoff water encounters solid materials during environmental entry (Bertling et al., 2002b).

In the laboratory study, which was conducted according to a fractional factorial design, a great number of parameters that may influence the copper immobilization by limestone were tested. The most significant parameters were the specific surface area of limestone, the copper concentration in the runoff water and the amount of limestone used in the reactor (Bertling et al., 2002a).

In the follow-up experiments, the runoff water was allowed to interact for short periods of time with limestone, concrete or soil (natural European soils and an OECD standard soil).

The retention capacity of copper appeared to be highest in soil (9699.8%), but also quite substantial in concrete (18-95%) and limestone (547%). The interaction between the copper-containing runoff water and the interfaces in these materials also caused a significant reduction of the fraction of free, hydrated cupric ion, thereby drastically decreasing the concentration of bioavailable metal (Bertling et al., 2002b).

Thus, roof runoff water containing copper, where the dominant fraction consists of free cupric ions when the runoff is leaving the roof, will rapidly become impoverished in both free cupric ions and total copper, due to the high reactivity of the metal species when interacting with materials that are common in the catchment of the runoff water. So far, it has not been possible to quantify the amount of copper being retained in the close surroundings of copper roofs, e.g. in Stockholm. Nor has it been possible to evaluate the impact of copper roofing and facing materials on the range of concentrations of cupric ions - and other bioavailable copper species - in receiving waters where protection of natural biota has a high priority.

For the time being, it is only possible to make an educated guess -very similar to the speculation made in (L & L - Cu) in 1998 - that of the 1.2 t/y of total copper being released from copper roofs and façades in Stockholm, at least half of the released amount is retained in solid materials close to the source. Thus, about 0.6-0.7 t/y of total copper would be physically transported away from the close surroundings of copper-clad buildings, and of this amount, some 50-60% would still be in a bioavailable form. During further transport in the storm-water sewers - or combined sewers - (often made of concrete), an additional fraction of unknown magnitude of the released copper will be immobilised and retained before the runoff water arrives at the STP.

0 0

Post a comment