Environ. Sci. Technol. 2001, 35, 4327-4333
Fixation of Radionuclides in Soil and Minerals by Heating BRIAN P. SPALDING† Environmental Sciences Division, Oak Ridge National Laboratory, Oak Ridge, Tennessee 37831-6036
Heating of fine sand-sized common mineral powders (quartz, feldspar, or calcite) or a soil (from the Department of Energy’s Hanford site) up to 1000 °C, in contact with sorbed radioisotopes (85Sr, 57Co, 134Cs, or U), markedly increased each isotope’s immobilization. A sequential extraction procedure was applied after heating the materials to assess the changes in each isotope’s functional form among water-soluble, cation-exchangeable, acid-soluble, and residual phases. The overall immobilization effects were consistent with rapid high temperature ionic diffusion from the initially contaminated surfaces into the mineral matrices; subsequent diffusion out of mineral particles at ambient temperature, as measured by the sequential extraction behavior, would be such a slow process that the radionuclides may be considered sequestered from further potential environmental mobilization. In the Hanford soil, the effect was found to follow an Arrhenius-type relationship with treatment temperature up to 1000 °C for 57Co, 85Sr, and U, and immobilization was independent of previous thermal treatment of the materials. Although 134Cs exhibited its largest immobilization in the Hanford soil after heating to 1000 °C, the large immobilization of 134Cs at all temperature and even in unheated Hanford soil made it difficult to observe a strong temperature dependence. A general and promising technique for environmental remediation of contaminated soil by high-temperature heating without melting can be extrapolated directly from the empirical leaching information.
Introduction Radioactively contaminated soil remains a difficult and expensive problem for the U.S. Department of Energy (DOE) to address in an environmentally comprehensive way with present technology. The slow and continued leaching by groundwater of major fission and activation products and transuranic isotopes as well as associated nonradioactive hazardous species, through surrounding contaminated soil into aquifers and/or surface water, often requires corrective action or, at least, careful, continuous, and thus expensive environmental monitoring. Possible remedial actions at former radioactive waste burial sites fall into several categories including (1) containment of present contamination with engineered covers and other barriers, (2) treatment of retrieved soil and waste to remove radioactive contaminants, leaving “clean” soil and residuals while generating secondary waste of hopefully much smaller volume, or (3) treatment of waste and soil, preferably in situ, to decrease the leachability or environmental mobility of contaminating radionuclides. † Corresponding author phone: (865)574-7265; fax: (865)576-8543; e-mail:
[email protected].
10.1021/es010608n CCC: $20.00 Published on Web 10/05/2001
2001 American Chemical Society
However, it is a rare scenario where soil is contaminated by only one or a few radionuclides or hazardous chemical species; thus, many promising soil and waste extraction or treatment techniques are found difficult to apply to more than one radionuclide. Thermal treatment of soil has been intermittently evaluated as a contaminant extraction (via volatilization), destruction, and/or stabilization technology (1, 2). But, unless residual material is vitrified, significant concerns usually remain about potential secondary wastes and disposal of residual contaminants. Investigations into the effects of thermal treatment of soil and mineral materials on subsequent leachability of contaminants have usually been incidental to a primary objective such as thermal destruction or incineration of hazardous organic contaminants (3-6), control of volatile emissions during incineration (7-13), or vitrification of the material to prepare a durable lowleachability waste form (14, 15). Occasionally, thermally altered soil has been found to exhibit a noticeable decrease in contaminant extractability (often based on a single extract or extraction method) which effect is either left unexplained or hypothesized to result from unique reactivity of that contaminant with component chemical phases in the soil or geologic material (11, 16, 17). One of the difficulties of extrapolating these findings into a general mechanistic interpretation has been the lack of sufficient evidence with a variety of radionuclides or contaminants over a broad range of temperature imposed using several disposal site soils or their common compositional minerals. Additionally, the use of a rigorous sequential extraction procedure, which allows inferences of the contaminant chemical forms in soil and their long-term environmental performance, would greatly aid in this evaluation. In a recent investigation of the effects of thermal treatment of contaminated hardened cement paste on the extractive behavior of major radionuclides (90Sr, 137Cs, 60Co, and U) at DOE facilities (18), a 15-step sequential extraction procedure was employed to characterize changes in extractability behavior of these radionuclides after heating between 200 and 1350 °C (the vitrification temperature of the cement paste); this sequential extraction protocol was designed for separation of contaminants from various earthen matrices in addition to cement paste; this procedure involved five repeated water extractions (for water-soluble or environmentally mobile phases), followed by five extractions with 0.1 N CaCl2 (for cation-exchangeable or environmentally labile phases), followed by five extractions with 0.2 N HCl (for acid-soluble or environmentally weakly sequestered phases). Although the use of multiple extractants to characterize soil phases of differing environmental availability has recently attained more common use and acceptance (19-22), the use of multiple steps for each extractant offers a less frequently employed but powerful technique to obtain evidence that the inferred phases have been exhaustively removed and separated before subsequent harsher extractants are imposed to remove more recalcitrant phases. Although thermal treatments of cement paste were generally found to decrease the extractability of these radionuclides from aggregate-free contaminated cement paste (18), the next hypothesized step to evaluate possible thermal treatments of concrete was to perform similar leaching tests with common and, presumably, inert mineral aggregates (i.e., quartz and feldspar sands and limestone gravels) which compose the remaining 80% of the weight of contaminated concrete. Given the generally observed weak sorption of radionuclides by such inert geologic aggregate materials at VOL. 35, NO. 21, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
4327
ambient temperature, it was presumed that these minerals would also exhibit comparatively weak reactivity with radionuclides during and after heating. Thus, this investigation was begun with the strong bias that it would be, at best, a time-consuming but necessary process to verify that common concrete aggregate materials remained inert during heating while simultaneous marked effects on contaminant extractability were occurring in the surrounding and more reactive cement paste. Fortunately, this bias has proven to be completely unsupported by the present observations that an alternate interpretation of radionuclide-mineral interactions can be supported and appears to have a direct and beneficial application for thermal remediation of most contaminated soils.
Experimental Section Four mineral materials (a quartz sand, a feldspar sand, a calcitic limestone, and one soil from the DOE Hanford reservation (see Supporting Information, Table SI-1, for elemental compositions and descriptions)) were investigated after spiking with one of four radionuclides (85Sr, 134Cs, 57Co, and depleted U); leachability tests of these spiked materials were performed after thermal treatments up to 1200 °C using the detailed 15-step sequential extraction protocol (18). Details of the sequential extraction methods are presented in the Supporting Information. All mineral materials were ground prior to spiking in an agate ball-mill using the isolated -100/+200 mesh (75-150 µm equivalent diameter) fraction so that comparison among these materials would be based on equivalent surface areas per unit weight, and the findings could be compared with the previous observations employing pulverized cement paste of similar size. The general experimental approach was to assay the gamma activity of a single isotope (23) spiked into a 1-g mineral powder sample before and after imposing a final temperature in a furnace open to the atmosphere (to detect any volatilization of the radionuclide during heating), and the distribution of that activity throughout the subsequent 15 extracts and the residual material, using corrections for both radioactive-decay and differing counting geometry for the various sample matrices, yielded a normalized distribution for each isotope. Because the experimental variable matrix [4 materials × 4 radioisotopes × 2 (and often 3) replicates × n treatment temperatures/ regimes] became rather unwieldy, initial investigations were performed only at room (untreated) temperature-, 500 °C-, and 1000 °C-treatments imposed for 16-20 h (overnight). Standard solutions of carrier-free radioisotopes [85Sr(II), 134Cs(I), and 57Co(II)] were obtained commercially at nominal 1 mCi/mL in 0.1 M HCl. Aliquots of these radioisotopic solutions were diluted 2000-fold with 0.01 N CaCl2, as a neutral dilute salt carrier solution, for use in spiking the mineral powders. Only one isotope was spiked into a given sample using a nominal 0.025 µCi of activity dissolved in 100 or 200 µL. For uranium, a standard solution (Spex Certiprep Inc.) of 9977 µg of depleted U(VI)/g in 5% HNO3 was employed directly using 700 µL spikes for a nominal activity of 0.005 µCi total uranium. The radionuclide spike was added to a weighed amount of mineral powder (0.999 ( 0.001 g) contained in a 2-mL-capacity 13-mm diameter × 25-mm high cylindrical refractory (99.8% alumina) gastight crucible (Coors Ceramics Co.). The spiking volume was placed into a central depression within the mineral powder such that wetted powder did not directly contact the walls of the crucible except in the case of the larger 700 µL U spike; this larger spiking volume was required for uranium to attain the needed activity for detection. After air-drying overnight, the spiked powder was mixed within the crucible by stirring thoroughly with a stainless steel spatula and the crucible was reweighed. Prior to imposing a thermal treatment, spiked powders were generally aged for 20-40 days while contained 4328
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 21, 2001
in 20-mL-capacity polypropylene scintillation vials after capping each crucible with a small piece of parafilm and a cork to prevent further drying or spilling during automatic sample changes during gamma spectroscopy. All crucibles and extracts were assayed via gamma spectroscopy using a NaI detector (23); further details of the assay methods are given in the section on gamma activity assay techniques in the Supporting Information. After completing a sequential extraction sequence, each crucible and its residual solids (including the filter used to prepare the extracts) were assayed again via gamma spectroscopy using geometrically similar isotopic standards and correcting for decay of the shorterlived isotopes. Final quantitative uranium assays were completed on extracts and residual solids after allowing at least seven half-lives (i.e., >168 days) for activities of the 234Th and 234Pa daughters to reestablish secular equilibria. The sum of decay- and counting geometry-corrected activities in the 15 extracts and residual material were compared on a percent basis with the sample activity prior to extraction to calculate the radioisotopic recovery for each sample after processing into these 16 fractions. After completing initial gamma assays of the spiked samples, the crucibles and their contents were subjected to a variety of heating treatments. Within each suite of samples spiked with a given radionuclide but subjected to different final temperatures, one group of triplicate samples was subjected to air-drying to constant weight (
