Sources and Colloid Transport of 234U-238U in the Baltic Sea and Kalix River Watershed

D. Porcelli Lunatic Asylum, Div. Geological and Planetary Sci., California Institute of Technology,

Pasadena, CA 91125, USA

P. S. Andersson Swedish Museum of Natural History, Laboratory for Isotope Geology, Box 50007,

104 05 Stockholm, Sweden

G. J. Wasserburg Lunatic Asylum, Div. Geological and Planetary Sci., California Institute of Technology,

Pasadena, CA 91125, USA

J. Ingri Div. Applied Geology, Luleå University of Technology, 971 87 Luleå, Sweden

M. Baskaran Dept. Oceanography, Texas A&M University at Galveston, 5007 Avenue U, Galveston, TX 77551, USA

The role of colloids in the transport of U in the Baltic Sea and the Kalix River, which drains a mire-rich region of northern Sweden, was investigated. Andersson et al. (1995) argued that colloids were important for the transport of Th into the Baltic Sea. These authors also found the 234U/238U ratio in the Kalix River to be unusually high compared to other rivers, and this was attributed to preferential transport of 234U from U-rich mires (groundwater-fed peatlands). The present TIMS study evaluates the sources of U isotopes in the river basin and the importance of U transport by colloids in the watershed and during estuarine mixing in the Baltic Sea. Andersson et al. (1996) examine the role of particles (>0.45µm) in transporting U in this system.

Ultrafiltration techniques were used to separate U associated with colloids > ~10,000 daltons (including inorganic particles, humic acids and other large organic molecules) from 'dissolved' U (U in solution or associated with smaller colloids) using Amicon hollow fiber membranes. Samples were first filtered using a 0.45µm filter. For each sample, the ultrafiltrate and the fraction that was excluded from the ultrafilter (the 'colloid concentrate') were collected, as well as a final 10% HCl rinse intended to remove U adsorbed onto
the system walls or in the ultrafilter. This U is likely to be associated with the colloidal fraction of the waters, although it is possible that 'dissolved' U was scavenged onto the ultrafilter. Measurements of the ultrafiltrate, colloid concentrate, and acid rinse of several samples (brackish water and organic-rich river water) found 94-97% recovery of the total sample U, demonstrating that U can be quantitatively recovered in these experiments. Substanital differences in the 234U/238U ratio between different fractions of the same sample were not found.

In the Kalix River, U concentrations in 0.45µm-filtered waters were found to increase downstream, while the fraction of this U associated with >10k dalton colloids also increased. At the head of the Kalix River, above the mire region, 0.45mm-filtered water had 55ng/kg U and d234U=1010. 60% of this U is associated with colloids >10k daltons, while the remaining 40% (<10k dalton fraction) was in solution or associated with smaller colloids. Near the mouth of the Kalix River, where the waters are rich in organic matter derived from the mire region, 0.45µm-filtered water had 184ng/kg U and d234U=900, and ~80% of this U was associated with >10k dalton colloids. These results clearly demonstrate the importance of colloids as U carriers throughout the river.

Within the Gulf of Bothnia, waters with low salinity (~3.3 ”) have 356ng/kg U and d234U=250, and ~50% of the U is associated with >10k dalton colloids, indicating that colloids are also important for U in these brackish waters. It was found earlier (Andersson et al., 1995) that U in waters in the northern Gulf of Bothnia could be explained by conservative mixing of seawater and a river component with ~40ng/kg U. Although this is much lower than the total U concentration in Kalix River water, the present data show that 40ng/kg U (~20% of the U in 0.45µm-filtered water) in the Kalix River mouth is in the <10k dalton 'dissolved' fraction. We suggest that it is this 'dissolved' U that behaves conservatively, while riverine U associated with >10k dalton colloids is removed during estuarine mixing. The association of U with colloids therefore may be an important parameter in determining U estuarine behavior. However, a significant fraction of U in the Gulf of Bothnia is also associated with colloids, indicating a continuing association of U with colloids in brackish waters. Within the Baltic Sea proper and at higher salinities (>6 ”), ~10% of the U is associated with >10k dalton colloids, a smaller but still significant fraction of the total U.

Within the Kalix River basin, peat samples from mires have U concentrations in ashed residues substantially higher than in typical crustal rocks, and dry weight concentrations of ~105 times that of associated mire waters. This indicates that U is highly concentrated in the organic matrix under the reducing conditions in the mires. The high 234U/238U in the Kalix River was earlier earlier (Andersson et al., 1995) attributed to mire water inputs. Although mire waters are enriched in 234U relative to the peats, these waters have low U concentrations (<20ng/kg) and 234U/238U ratios comparable to the river. These waters are therefore not the primary source of the excess 234U in the river. The most likely source of excess 234U in this area is groundwater. This is division contribution No. 5626 (916).


Andersson, P.S. et al., Earth Planet. Sci. Lett. 130, 217-234 (1995).

Andersson, P.S. et al., J. Conf. Abs. 1(1) 16 (1996).