Trilithium hexafluoroaluminate

Adsorption / desorption

Link to relevant study record(s)

Reference

Reference 1

Endpoint:

adsorption / desorption, other

Remarks:

measurement on aquifer material

Type of information:

not specified

Adequacy of study:

supporting study

Study period:

not applicable

Reliability:

2 (reliable with restrictions)

Rationale for reliability incl. deficiencies:

data from handbook or collection of data

Remarks:

Data reported in HSDB peer-reviewed by the Scientific Review Panel (SRP).

Principles of method if other than guideline:

HSDB is a toxicology data file on the National Library of Medicine's (NML) Toxicology Data Network (TOXNET(R)). It focuses on the toxicology of potentially hazardous chemicals. It is enhanced with information on human exposure, industrial hygiene, emergency handling procedures, environmental fate, regulatory requirements, and related areas. All data are referenced and derived from a core set of books, government documents, technical reports and selected primary journal literature. HSDB is peer-reviewed by the Scientific Review Panel (SRP), a commitee of experts in the major subject areas within the data bank's scope. HSDB is organized into over 5000 individual chemical records.

GLP compliance:

no

Adsorption and desorption constants:

The adsorption of lithium was measured on aquifer material; Freundlich coefficients ranging from 4.5 to 5.5.
(reference: Vereecken H et al; pp. 627-48 in Environ Fate Xenobiot, Proc Symp Pestic Chem, 10th, 1996. Del Re AAM, ed., Goliardica Pavese: Italy (1996))

Lithium has been found to sorb slightly to humic soils with a Kp of 4.6 at pH 5
(reference : (Oman C, Rosqvist H; Wat Res 33: 2247-54 (1999))

These data indicate that lithium compounds are not expected to adsorb strongly to soils and sediments(SRC).

Validity criteria fulfilled:

not applicable

Conclusions:

lithium compounds are not expected to adsorb strongly to soils and sediments

Reference 2

Endpoint:

adsorption / desorption, other

Remarks:

Column leaching

Type of information:

experimental study

Adequacy of study:

key study

Study period:

1985

Reliability:

2 (reliable with restrictions)

Rationale for reliability incl. deficiencies:

test procedure in accordance with national standard methods with acceptable restrictions

Remarks:

Although the study was not carried out according to a guideline and GLP, it is well described and the result is considered acceptable.

Qualifier:

according to guideline

Guideline:

other: EPA Guideline Subdivision N 163-1

GLP compliance:

no

Type of method:

other: Column leaching

Media:

soil

Test temperature:

22 ± 2ºC

Analytical monitoring:

yes

Key result

Type:

Koc

Value:

>= 603 - <= 6 502 L/kg

Key result

Type:

log Koc

Value:

>= 2.8 - <= 3.8 L/kg

Background fluoride levels for the soils prior to use in columns were: Lakeland sand, 3.4 parts per million (ppm); Ramona sandy loam, 3.0 ppm; Aguila clay loam, 11.7 ppm; and Calhoun silt loam, 15.0 ppm.

Results of fluoride analysis of control (background), cryolite-spiked, and sodium fluoride-spiked columns for each soil type are summarized in the Table below. All values are corrected for analytical method efficiency; spiked column values are net concentrations above the background level for each segment.

Despite the constant 5 cm head of water maintained during each test, water flow for all soils except Lakeland decreased during the course of each column run, presumably as fine soil particles migrated downward sufficiently to fill interstices between the larger particles. Lakeland sand columns ran in 40 minutes, Calhoun silt loam in about 14 hours, Aguila clay loam in 36 to 48 hours, and Ramona sandy loam in 60 to 96 hours. Fluoride concentration in the eluted water was very low for all soils - 0.2 ppm for Aguila clay loam and 0.1 ppm for the other soils. These concentrations were identical in control and in spiked columns (i.e., the spiked column fluoride levels above background were all zero).

In general, fluoride concentrations in the top ("cover") segment in the control columns were less than in lower segments - considerably so in the Aguila and Ramona soils and, to a lesser degree, in Calhoun and Lakeland soils, indicating the leaching of naturally occurring fluoride from this layer. Excluding this layer, the maximum variation of background fluoride concentration was 9% of the highest concentration for Lakeland soil, 15% for Ramona, 11% for Calhoun, and 27% for Aguila soil (14% for Aguila if a single low value of 7.7 ppm is excluded). This variation is to be expected in soils which are, by their nature, heterogeneous.

Table - Parts per million (ppm) fluoride in each segment control and spiked soil columns and in recovered water. All values are corrected for analytical method efficiency. Spiked column values are net concentrations above background levels.

	Aguila Clay Loam			Calhoun Silt Loam			Ramona Sandy Loam			Lakeland Sand
Layer	control	cryolite	NaF	control	cryolite	NaF	control	cryolite	NaF	control	cryolite	NaF
Cover	4.4	0.0	0.0	11.2	0.4	1.2	1.6	0.4	0.8	3.5	0.0	0.0
Spiked	10.4	0.1	3.3	12.5	19.1	16.2	2.9	2.0	2.0	4.2	2.9	0.3
0-6 cm	9.9	6.3	4.5	12.8	2.1	2.9	3.1	6.6	6.4	3.9	2.5	0.9
6-12 cm	9.2	0.0	0.0	12.6	0.0	0.9	3.1	2.7	2.2	4.0	1.4	1.6
12-18 cm	9.0	0.1	0.4	13.9	0.0	0.0	3.3	0.0	0.0	4.3	0.5	0.8
18-24 cm	7.7	1.9	0.0	14.0	0.0	0.0	3.4	0.0	0.1	4.1	0.4	0.7
24-30 cm	9.7	0.0	0.0	12.7	0.4	0.0	3.2	0.0	0.0	4.2	0.0	1.1
30-36 cm	10.5	0.0	0.0	13.6	0.0	0.4	3.2	0.0	0.0	4.3	0.0	0.7
36-42 cm	9.4	0.0	0.0	13.4	0.4	0.6	3.4	0.3	0.0	4.0	0.0	0.0
water	0.2	0.0	0.0	0.1	0.0	0.0	0.1	0.0	0.0	0.1	0.0	0.0

The greatest movement of fluoride in cryolite-spiked columns was in Lakeland sand, where fluoride concentrations of 0.5 and 0.4 ppm were detected at the 18 to 24 cm level. This soil also exhibited the greatest movement of all the sodium fluoride-spiked columns, with detectable fluoride concentrations down to the 30 to 36 cm segment (0.7 ppm).

An isolated 1.9 ppm fluoride concentration at 18 to 24 cm in the clay loam cryolite column was probably caused by an anomalously high concentration of background fluoride in this segment, as the two segments above it contained only 0.0 and 0.1 ppm fluoride.

With the exception of this isolated high value and three other isolated low values of 0.4 ppm at 24 to 30 cm and 36 to 42 cm in the silt loam and 0.3 ppm at 36 to 42 cm in the sandy loam (also probably caused by high background anomalies), none of the cryolite-spiked soil columns except Lakeland sand exhibited any detectable migration of added fluoride below the 6 to

12 cm segments. Fluoride concentration in Aguila clay loam and Ramona sandy loam was greatest in the 0 to 6 cm segment (6.3 and 6.6 ppm, respectively). There was virtually no movement of added fluoride in Calhoun silt loam where the highest concentration was in the spiked layer.

With the exception of Lakeland sand, sodium fluoride-spiked columns displayed fluoride migration profiles similar to those of cryolite in the matching cryolite columns. As with the cryolite columns, there were low level isolated fluoride concentrations probably caused by background fluoride anomalies. With the exception of these isolated values, spiked fluoride in the sodium fluoride columns was not detectable below the 6 to 12 cm segments in these three soils.

Sorption coefficients (Kd) determined for each of the soils are presented in Table below. The highest sorption coefficients were obtained for Calhoun silt loam, which is in agreement with the low fluoride mobility displayed in the leaching columns using this soil. Coefficients for Aguila clay loam and Ramona sandy loam were considerably lower; these soils also displayed more fluoride mobility in the leaching columns. The lowest sorption coefficients were obtained with Lakeland sand, which also displayed.the greatest movement of fluoride in both cryolite- and sodium fluoride-spiked soil columns.

Table - Sorption coefficients (Kd) of cryolite in four test soils. Each value is the mean of three replicates.

Soil	Water: Soil	Initial ppm Cryolite in water	Replicate Kd values			Mean Kd	Std Dev
Lakelandsand	5:1	5.0	6.9	5.9	6.9	6.6	0.6
		10.0	3.5	4.4	4.3	4.1	0.5
		20.0	3.3	3.4	3.5	3.4	0.1
		50.0	1.6	1.3	1.3	1.4	0.2
Ramona Sandy Loam	15:1	5.0	16.2	16.2	12.8	15.1	2.0
		10.0	14.4	14.4	12.3	13.7	1.2
		20.0	10.0	12.0	12.3	11.4	1.3
		50.0	8.1	8.0	8.1	8.1	0.1
Aguila Clay Loam	15:1	5.0	10.9	7.7	7.1	8.6	2.0
		10.0	10.0	10.4	11.8	10.7	0.9
		20.0	7.7	8.3	9.0	8.3	0.7
		50.0	7.1	8.2	8.3	7.9	0.7
Calhoun Silt Loam	50:1	5.0	54.2	50.0	54.2	52.8	2.4
		10.0	39.3	44.3	48.0	43.9	4.4
		20.0	34.0	37.7	31.3	34.3	3.2
		50.0	18.3	19.8	19.8	19.3	0.9

The most evident result of these tests is that there was extremely limited movement of cryolite or sodium fluoride in any of the soils with the exception of sodium fluoride in Lakeland sand. In Calhoun silt loam, there was virtually no movement of either chemical away from the spiked layer. In addition, fluoride movement in the columns was nearly identical for any given soil for both the cryolite- and sodium fluoride-spikedcolumns; again, with the exception of the sand.

There are many factors which can affect the movement of a chemical through a soil. Factors characteristic of a soil type include soil particle size, the amount and type of clay present, soil pH, cation exchange capacity, and organic matter content. The rate of water movement through the soil will influence both the depth of chemical movement and the dispersion or tailing of an eluted chemical in a column of soil. This includes both the speed of water movement and the rate of application of the water.

In evaluating these factors, it is important to consider that nearly all available data in the literature, as well as EPA Guidelines dealing with leaching and sorption/desorption are predicated on the fate of organic chemicals in soils. Sorption of organic compounds in soil occurs at the soil particle surface, primarily through physical binding involving van der Waals forces which do not necessarily involve alteration of structures. Cryolite is an inorganic chemical with the potential to react readily with constituent soil anions and cations

to form more stable and less soluble complexes.

It might be supposed that sodium fluoride with a solubility in water of 0.4%, which is ten times more than that of cryolite at 0.04%, would show much greater mobility in soil than was observed. This supports a conclusion that fluoride ion in solution tends to be complexed into less soluble forms. For example, fluoride ion would likely compete for calcium ions to form highly insoluble CaF₂ (0.0015% aqueous solubility).

With regard to other soil properties which may influence the degree of leaching of a compound, there was no apparent correlation of fluoride movement with soil pH in either cryolite or sodium fluoride columns. While cation exchange capacity (CEC) was smallest in the sand where fluoride movement was greatest, fluoride movement in the other three soils does not appear to display any relationship to CEC.

Soil organic matter content is often the primary soil property responsible for sorption and leaching retardation of organic compounds but a marked influence cannot be detected in the case of inorganic cryolite. As noted, sand (0.4% organic matter) displayed the greatest movement of fluoride. However, movement in sandy loam and clay loam (0.9% and 1.4% organic matter, respectively) was very similar and greater than in the clay loam (also 1.4%).

Water flow rate through the columns cannot be isolated as a significant factor in fluoride movement in these tests. Relative flow rates may be expressed as: sand >> silt loam > clay loam > sandy loam, while fluoride movement was generally: sand > sandy loam > clay loam > silt loam. Peak broadening, which is expected to be greatest for organic compounds at lowest flow rates, was not evident in this study.

So-called sorption coefficients (Kd) were found to be accurate predicators of fluoride mobility in these soils. Kd values for sand ranged from 1.4 to 6.6, depending on the amount of cryolite initially in the water. There was some overlap of clay loam (Kd = 7.9 to 10.7) and sandy loam (Kd = 8.1 to 15.1) values, which is reflected by the similarity of leaching profiles in these soils. Coefficients for silt loam (Kd = 19.3 to 52.8) indicate a soil of high sorbtivity, which was borne out by test results in which little fluoride movement occurred away from the spiked layer.

It is evident that these laboratory tests support observations of the behavior of cryolite in field use. Even with multiple seasonal applications over a period of years, significant leaching of cryolite into ground water has not been observed, as evidenced by analyses of soil run-off water and well water. Mobility of cryolite in the four representative soils tested is very low. In sand there is somewhat more movement than in the other soil types. Mobility/sorption of cryolite is influenced primarily by soil mineral composition and soil particle size. Mobility is suppressed by means of chemical complexation which occurs primarily with aluminum, iron, and calcium minerals to form highly stable insoluble fluorides.Smaller soil particle size increases reactive surface area per soil volume. Other soil characteristics and water flow rate do not have an appreciable influence on cryolite mobility.

The data show two striking features. First, there are large soil-dependent differences in sorption that were mentioned before, with "apparent" simple Kd's ranging from 1 to 53 in standard units (calculated range of apparent Koc's is from approximately 600 to 6500, see Table 4). 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 Lakeland 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)^a. 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 behavior.

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 registrant 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's 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 calcarious 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 behavior is again explainable if calcium fluoride precipitation occurs.

In the report Kd values are presented, whereas Koc values are needed. Koc values can be calculated based on the organic carbon content. Although the organic carbon contents are not reported, organic matter contents are available. According to Schüürmann et al.,(2007)^b, the organic carbon content can be calculated by the following equation, if no suitable experimental data are available:

% organic carbon = % organic matter/ 1.724

In the Table below, calculated Koc ranges are presented:

Table - Measured Kd values and calculated approximate Koc value based on the assumption that % organic carbon ~ % organic matter / 1.724 (Schüürmann et al.,2007)^b

Soil	% organic matter (measured)	% organic carbon (calculated)	Measured Kd range	Calculated Koc range
Lakeland sand	0.4	0.23	1.4 – 6.6	603 – 2844
Ramona Sandy Loam	0.9	0.52	8.1 – 15.1	1552 – 2892
Aguila Clay Loam	1.4	0.81	7.9 – 10.7	973 – 1318
Calhoun Silt Loam	1.4	0.81	19.3 – 52.8	2377 - 6502

^aU.S. EPA (1996), Reregistration Eligibility Decision (RED) Cryolite

^bSchüürmann, G., Ebert, R.U., Nendza, M, Dearden, J.C., Paschke, A and Kühne, R. (2007), Chapter 9, Predicting fate-related physicochemical properties, Risk Assessment of Chemicals: An Introduction, Second edition (Eds.Vanand Vermeire, T.G.), Springer, The Netherlands

Description of key information

As Lithium cryolite dissociates in water and as the risks are assumed to be determined by fluoride, a read across approach with a data available on Sodium cryolite is performed. For the cryolite moiety, a Koc of 1498 (log Koc 3.18) will be used in the assessment.

For Lithium moiety, low adsorption is expected as reported in HSDB.

Key value for chemical safety assessment

Koc at 20 °C:

1 498

Additional information

No data on the adsorption/desorption potential of Lithium cryolite is available. A reliable key study on Sodium cryolite is available, and is used in a read-across approach (See the endpoint study summary "Environmental fate and pathways" for the read-across justification). This study is described here below:

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 artifact 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, 1985b). 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 – 2844
Ramona Sandy Loam	0.9	0.52	8.1 – 15.1	1552 – 2892
Aguila Clay Loam	1.4	0.81	7.9 – 10.7	973 – 1318
Calhoun Silt Loam	1.4	0.81	19.3 – 52.8	2377 - 6502

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 Lakeland 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 behavior.

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 behavior is again explainable if calcium fluoride precipitation occurs.

The geometric mean of 1,498 (log Koc 3.18) will be used in the assessment.

As Lithium cryolite dissociates in water, the risks are assumed to be determined by fluoride. As the Kd is the ratio of the concentration of the substance in two different matrices, the value expressed in Fluoride, Sodium cryolite or Lithium cryolite will be the same. Therefore, the calculated koc value for Sodium cryolite is directly applicable to Lithium cryolite.

[LogKoc: 3.18]