Isolation of colloidal monomethyl mercury in natural waters using cross-flow ultrafiltration techniques

A series of experiments was conducted to evaluate the appropriateness of cross-flow ultrafiltration (CFUF) techniques for the determination of the phase speciation of monomethyl mercury (MeHg) in natural waters. Spiral-wound cartridge (Amicon S1Y1) and Miniplate (Amicon) were evaluated for their nominal molecular weight cut-offs of 1 and 10 kDa, respectively. The ultrafiltration behavior of standard macromolecules showed that the permeation of high molecular weight (HMW) organic macromolecules was not significant when a concentration factor (CF)>15 was used. The retention of low molecular weight (LMW) molecules was significant, especially at a low CF<5, suggesting that the use of a high CF (15) will minimize the retention of LMW molecules. Sorptive losses of MeHg in the solution phase to the 1 kDa membrane were negligible, but MeHg bound to HMW macromolecules was still retained (20%), even with a preconditioned membrane. The mass balance recovery of MeHg during ultrafiltration averaged 101±15% (n=7) and 105±14% (n=5) for the 1 and 10 kDa membranes, respectively. Sample storage over 24 h caused significant coagulation (47%) of the <10 kDa MeHg into the 10 kDa–0.45 μm colloidal or the particulate MeHg pool. The 1 kDa–0.45 μm colloidal MeHg in Galveston Bay and the Trinity River water samples accounted for 40–48% of the filter-passing MeHg, although the most abundant fraction (52–60%) of MeHg was the truly dissolved fraction (<1 kDa). The partition coefficient between the colloidal (1 kDa–0.45 μm) and truly dissolved MeHg (average log K_C=5.2) was higher than the partition coefficient based on particle/filter-passing (average log K_D=4.6) or particle/truly dissolved MeHg (average log K_P=4.8), suggesting that MeHg has stronger affinity for natural colloids than macroparticulate materials (>0.45 μm).