Reaction mass of Potassium aluminium tetrafluoride and Dipotassium aluminium pentafluoride monohydrate

Adsorption / desorption

Description of key information

As cryolite dissociates in water and the risks are assumed to be determined by fluoride, it is more appropriate to assess the adsorption behaviour of fluoride. In a column leaching test, Koc values varying from 603 - 6502 were found for fluoride in 4 different soils, tested at cryolite concentrations varying from 5 - 50 ppm (Dykeman, 1985). As the percentage of fluoride in the molar weight in both substances is practically identical, namely 54% for cryolite and 53% for aluminium potassium fluoride, similar Koc values can be expected for aluminium potassium fluoride. The geometric mean of 1498 (log Koc 3.18) will be used in the assessment.

Additional information

As no information is available on adsorption/desorption of aluminium potassium fluoride, data from its structural analogue cryolite is taken into account. In a column leaching study using four different soils with cryolite at an equivalent application rate of 16 lb/acre, fluoride (the only species monitored) showed little mobility. A fluoride ion specific electrode was used for quantitation. Background fluoride concentrations from the control soils, which varied from about 2 to 14 ppm, were subtracted from the treated soils. Most fluoride remained within the top 24 cm of the columns. Some extraneous leaching did occur to a maximum depth interval of 36-42 cm, but was probably an artefact of method limitations and/or natural soil variation. No fluoride was detected in the leachate of the 42 cm columns. For comparison, sodium fluoride, which was run through equivalent soil columns at equivalent fluoride concentration, showed virtually the same leaching profile for fluoride as cryolite (Dykeman, 1985). Koc values are given in the Table below.

 

Table: Measured Kd values and calculated approximate Koc value based on the assumption that % organic carbon ~ % organic matter / 1.724

Soil	% organic matter (measured)	% organic carbon (calculated)	Measured Kd range	Calculated Koc range
Lakeland sand	0.4	0.23	1.4 – 6.6	603-2,844
Ramona Sandy Loam	0.9	0.52	8.1 – 15.1	1,552 -2,892
Aguila Clay Loam	1.4	0.81	7.9 -10.7	973 – 1.318
Calhoun Silt Loam	1.4	0.81	19.3 – 52.8	2,377 – 6,502

The data show two striking features. First, there are large soil-dependent differences in sorption that were mentioned before, with "apparent" simple Kd values ranging from 1 to 53 in standard units (calculated range of apparent Koc values is from approximately 600 to 6500, see Table). Second, there is a pronounced, regular concentration-dependent spread of Kd values within three of the four individual soils. For these three soils the higher the concentration, the lower the Kd. For example, in sand with 5 ppm of cryolite, the Kd is 6.6, but at 50 ppm of cryolite, the Kd is only 1.4.

These features prompted the U.S. EPA to conduct a Freundlich analysis using the reported data. Results yielded exponents (1/n values) of approximately 1/2 for three soils (exponents of 0.56, 0.49, and 0.69 for sand, sandy loam, and silt loam soils, respectively) (U.S. EPA, 1996). The exponent of about 1/2, the seeming approach to "saturation" of fluoride, and the apparent lack of correlation with organic matter in these soils suggested that the mineral precipitation with a divalent cation is responsible for the observed behaviour.

As calcium is usually a dominant exchangeable cation in soils, and also forms insoluble calcium fluoride, the U.S. EPA tested the precipitation hypothesis using the adsorption data, the solubility product constant for calcium fluoride (the mineral fluorite), and the assumptions that approximately half of the fluorine in cryolite is available as fluoride and that exchangeable calcium ion in many soils usually accounts for about 0.1 to 0.2 of the maximum CEC (individual exchangeable cations were not reported). Calculations using the various measured Kd's and water to soil ratios showed fluoride concentrations consistent with those predicted.

Unlike the other three soils, the fourth soil (Aguila clay loam) had uniform sorption coefficients for all four of the tested concentrations. Kd values averaged approximately 8.9 ± 1.3 in standard units, the pH is 8.0, and its CEC is given as 43.6 meq/100 g. These high values are typical of a calcareous soil and require special interpretation. With a large reserve of calcium, small changes in its equilibrium concentration due to precipitation with fluoride are offset, and the soil is far from being saturated with fluoride. Additional calcium ion available from equilibrium with abundant solid carbonate opposes any shifts in dissolved calcium concentration. Thus, the observed sorption behaviour is again explainable if calcium fluoride precipitation occurs. As the percentage of fluoride in the molar weight in both substances is practically identical, namely 54% for cryolite and 53% for aluminium potassium fluoride, similarKoc values can be expected for aluminium potassium fluoride.