Identification of industrial stony residues in urban soils made of filling materials : a step toward prediction of trace element content ?

Industrial stony residues were widely used for centuries as filling material in urban areas of the Walloon region (Belgium). Their composition and properties widely differ according to the industrial process which generated them. It is assumed, therefore, that the identification of such materials in the field can contribute to (1) better assess potential risks at an early stage when investigating such soil and (2) to devise remediation operations that are more appropriate. However, identification criteria for industrial residues are still poorly documented and might be affected by regional context. This study aims at establishing identification criteria for industrial stony residues of the Walloon region and linking the nature and amounts of such residues to the content of trace metal in a large population of soil samples from this area. On this purpose, we selected 75 reference samples of industrial stony residues in order to include a wide range of morphology and texture basing on hand specimen visual description. They were analyzed by ICP either after alkaline fusion (Li-borate) for major elements (Si, Ca, Fe, Al, Mn, Mg, K, Na, S and P) or after tri-acid dissolution for trace metals (Cu, Pb, Zn, Ni, Cr, As, Cd, Co). Associated industrial processes were identified (i) according to historical information on the sampling site, (ii) by comparing chemical composition to bibliographic data and (iii) by using morphologic and texture diagnostic criteria from literature, when available. Measured chemical compositions were compared to the composition of samples from literature using graphical and statistical methods. The samples were linked to 16 different industrial processes, mainly related to steel industry and coal power plants. On the basis of the 75 reference samples, we derived criteria for field hand specimen identification. They are based on both morphology and texture features (colour, shape, fracture, luster, vacuolar texture …) and simple tests commonly used in mineralogy (HCl reaction, streak colour, hardness, density …). We used these criteria to identify and quantify stony materials in 380 soils samples collected in 115 filling soil profiles in urban areas from Wallonia. Stony materials were also classified into 6 major classes: stones naturally occurring in soils, mining residues, demolition rubbles, industrial waste, household waste and unidentified materials. Industrial stony residues were found in 18% of the surface layer and 25% of deeper layers (down to a 1.3 m depth). Industrial stony residues account on average for 7% and 12% (volumic percentages) of stony materials in surface and deeper horizons respectively. They mainly consist of coal combustion residues, blast furnace slags, steel slags and coke fragments. Results for all the stony materials were compared to the chemical data measured on the fine earth fraction (i.e. the granulometric fraction under 2 mm) of corresponding soil samples in order to detect the classes which are associated with a high pollutant content. Such information make it possible to rapidly assess on the field any potential risk for human health or and the environment and to devise cleaning or conservation measures appropriate to the type of material and to its industrial origin