ANALYSIS OF COLLOID AND TRACER BREAKTHROUGH CURVES

We consider the dispersion and elution of colloids and dissolved nonso rbing tracers within saturated heterogeneous porous media. Since flow path geometry in natural systems is often ill-characterized macroscopi c (mean) flow rates and dispersion tensors are utilized in order to ac count for the sub-model scale microscopic fluctuations in media struct ure (and the consequent hydrodynamic profile). Even for tracer migrati on and dispersal this issue is far from settled.Here we consider how c olloid and tracer migration phenomena can be treated consistently. The oretical calculations for model flow geometries yield two quantitative predictions for the transport of free (not yet captured) colloids wit h reference to a non-sorbing dissolved tracer within the same medium: the average migration velocity of the free colloids is higher than tha t of the tracer; and that the ratio of the equivalent hydrodynamic dis persion rates of colloids and tracer is dependent only upon properties of the colloids and the porous medium, it is independent of pathlengt hs and fluid flux, once length scales are large enough. The first of t hese is well known, since even in simple flow paths free colloids must stay more centre stream. The second, if validated suggests how solute and colloid dispersion may be dealt with consistently in macroscopic migration models. This is crucial since dispersion is usually ill-char acterized and unaddressed by the experimental literature. In this pape r we present evidence based upon an existing Drigg field injection tes t for the validity of these predictions. We show that starting from ex perimental data the fitted dispersion rates of both colloids and non-s orbing tracers increase with the measured elution rates (obeying sligh tly different rules for tracers and colloids); and that the ratio of c olloid and nonsorbing tracer elution rates, and the ratio of colloid a nd nonsorbing tracer dispersion rates may be dependent upon properties of the colloids and the medium (not the flow regime). It is important to realize that even for unretarded species, an earlier peak in the b reakthrough curve does not necessarily correspond to a faster mean elu tion rate, or vice versa. But rather that a colloid may elute faster b ut disperse less than an equivalent tracer. Hence its peak may be reta rded compared to that of the tracer, even assuming no retardation. Hen ce one must consider a combination of mean elution rate and mean dispe rsion rate, and not rely on ''peak times'' to corroborate chromatograp hic effects. The importance of this lies in the fact that these proces ses are not independent and yet upscale differently. Thus realistic es timates of effective colloid dispersion rates should be upscaled in a way consistent with that adopted for tracers within the same system.