Registration Dossier

Ecotoxicological information

Long-term toxicity to fish

Currently viewing:

Administrative data

Link to relevant study record(s)

Description of key information

Given the hydrolytic instability of LiPF6, long-term toxicity to aquatic organisms is best defined in terms of the toxicity of its F-, Li+ and PO4(3-) hydrolysis products.  Review of the known toxicities of these leads to a clear conclusion that the fluoride released from LiPF6 is the “toxic marker” for long-term toxicity to fish: from the lowest reported fluoride long-term fish NOEC of 2.3 mg F-/l a long-term NOEC value of 3.1 mg LiPF6/l is calculated (based on the expected complete F- release).

Key value for chemical safety assessment

EC10, LC10 or NOEC for freshwater fish:
3.1 mg/L

Additional information


In the aquatic environment, HF will principally be present in the form of fluoride ion. For this reason, test data obtained using soluble inorganic fluorides can be used to evaluate HF toxicity and LC50 or NOEC values expressed in terms of mg F-/l are appropriate for assessment of HF toxicity to aquatic organisms (HF: EU Risk Assessment Report, 2001).



A number of long-term toxicity tests with freshwater fish have been reviewed (WHO EHC 227, 2002): in a 20-day semistatic test using soft water (CaCO3 <3 mg/l), long-term LC50 values of 2.7 – 4.7 mg/l were recorded, but other workers calculated that 5.1 and 7.5 mg F-/l represent safe concentrations for rainbow and brown trout respectively. A long-term NOEC for rainbow trout (21-day semistatic test LC5 value) of 4 mg F-/l in very soft water (CaCO3 12 mg/l) is cited in the HF: EU Risk Assessment Report (2001), and a test of mullet (marine fish) exposed to fluoride at 5.9 or 5.5 mg/l for 68 or 113 days had no effect on survival, although growth of smaller fish was reduced (WHO EHC 227, 2002).  In a detailed study of public domain information of fluoride toxicity to various fish species, Fleiss (2011) showed that among LC50 values from studies longer than 4 days, rainbow trout gave both the lowest and highest values: the lowest reported LC50 value was 2.3 mg/l. The lower 95thpercentile value calculated from all longer term LC50s considered was 2.62 mgF-/l.



After three separate experiments on rainbow trout, looking at egg integrity, fertilisation, embryogenesis, hatching and then sac-fry and juvenile survival, Emery, Klopfer and Skalski (1981) reported “lowest rejected concentration tested” (LOEC) values of 3.3 mg Li+/l following LiF exposure and 0.6 mg Li+/l after Li2CO3 exposure. 


Kzos, Beauchamp and Stewart (2003) measured growth inhibition of fathead minnow larvae exposed to lithium for 7 days in low sodium (ca. 2.8 mg/l) water or natural stream water (ca. 17 mg Na/l) and found lithium toxicity to be inversely related to sodium level: with added sodium at 40 mg/l, exposure to Li+ at 4 mg/l caused no evident toxicity while but with 10 mg Na/l the same Li+ concentration proved 100% lethal. In the low sodium water an IC50 value for growth of 0.57 mg/l was reported. These workers concluded that in natural waters the presence of sodium is sufficient to prevent lithium toxicity, noting that high lithium tolerance limits for three species ofriver fish were reported by workers who used reconstituted river water (49 mg Na/l) in their tests.


A fish early life stage study using the fathead minnow (conducted according to US EPA and ASTM guidelines and under GLP) and lithium chloride is available (Long, Brown and Woodburn, 1998). Effects on embryos and larvae were assessed over a 26-day period (including 22-23 days post hatching) in treated river water (hardness 56-70 mg/l as CaCO3). EC50 and NOEC values calculated in terms of Li+ ion concentration were 1.0 and 0.2 mg/l respectively.



A study of phosphorus supplementation in rainbow trout over 53 days found no evidence that addition of Na2HPO4 into low-phosphorus basal diet at levels giving up to 10.96 g P/kg dry weight (equivalent to 33.6 g PO4/kg or 50.2 g Na2HPO4/kg) adversely affected growth or survival; indeed, weight gain was increased with phosphate supplementation up to a level of approximately 5 g P/kg in diet (equivalent to 8g PO4/kg) (Rodehutscord, 1996). 


Key value for assessment

The NOEC value reported for lithium in a fish early life stage test (0.2 mg/l) corresponds to an LiPF6 concentration of 4.4 mg/l. Given the range of long-term fish NOEC and LC50 values reported for fluoride (variable mainly due to variations in water hardness), it is reasonable to consider the lowest long-term LC50 value of 2.3 mg F-/l (below the long-term fish NOEC cited in the HF: EU Risk Assessment Report, 2001 and below the level occurring naturally in surface waters of areas where fluoride-containing rock formations exist) as an indicative threshold value for fluoride toxicity: this corresponds to an LiPF6 concentration of 3.1 mg/l. Due to its low toxicity, the contribution of phosphate to LiPF6 toxicity is considered insignificant. It is therefore concluded that the fluoride released from LiPF6 in water is the “toxic marker” for long-term toxicity to fish and a long-term NOEC value of 3.1 mg/l for LiPF6 is taken for use in assessment.