2,4,6-trimethyl-2,4,6-tris(3,3,3-trifluoropropyl)cyclotrisiloxane

Hydrolysis

Link to relevant study record(s)

Reference

Endpoint:

hydrolysis

Type of information:

experimental study

Adequacy of study:

key study

Study period:

2013-03-18 to 2013-05-02

Reliability:

2 (reliable with restrictions)

Rationale for reliability incl. deficiencies:

guideline study with acceptable restrictions

Qualifier:

according to guideline

Guideline:

OECD Guideline 111 (Hydrolysis as a Function of pH)

Deviations:

yes

Remarks:

20% co-solvent was used to achieve a higher test and reference substance concentration. In addition, only one single temperature was used and changes in in only parent concentration was monitored

GLP compliance:

yes

Radiolabelling:

no

Analytical monitoring:

yes

Buffers:

- pH: 5
- Type and final molarity of buffer: 0.005 M
- Composition of buffer: dilute acetic acid and dilute lithium hydroxide solution

- pH: 7
- Type and final molarity of buffer: 0.005 M
- Composition of buffer: imidazole and dilute hydrochloric acid

- pH: 9
- Type and final molarity of buffer: 0.005 M
- Composition of buffer: dilute boric acid and dilute lithium hydroxide solution

Details on test conditions:

The buffer solutions were sparged with an inert gas for a minimum of 5 min to exclude oxygen and carbon dioxide and sterilised using Nalgene sterile filtration unit with a 0.2 µm cellulose nitrate (CN) membrane. Acetonitrile was mixed with the buffers to be 20% by volume of the final test solution.

The pH 9 reaction samples for F-D3 and D3 were aged at ambient temperature of the laboratory; as rate of hydrolysis was expected from method development to be quite fast (half-life <15 min), it was not practicable to age them in an incubator. The pH 5 and pH 7 reaction samples for F-D3 and D3 were aged in a dark incubator at an average temperature of 25°C. Incubator temperature was recorded once per day of sampling. Temperature measurements were made using a thermometer that had been standardised against a NIST traceable thermometer. A digital clock was used to record experiment initiation and sampling times used to calculate sample age (i.e. reaction time). Configuration conditions were ~2500 rpm for ~10 min.

Duration:

46 d

pH:

7

Temp.:

25

Initial conc. measured:

390 - 430 other: ppb

Positive controls:

no

Negative controls:

no

Preliminary study:

Preliminary study was not performed

Transformation products:

yes

No.:

#1

No.:

#2

No.:

#3

Details on hydrolysis and appearance of transformation product(s):

The initial product of hydrolysis of F-D3 is 1,5-dihydroxy-1,3,5-trimethyl-1,3,5-tris(3,3,3-trifluoropropyl)trisiloxane (the trimerdiol), which is formed by ring-opening of the parent substance. This then undergo further hydrolysis with the final hydrolysis product being 3,3,3-trifluoropropylmethysilanediol (see Figure 1 attached).

% Recovery:

98.8 - 107.6

pH:

5

Temp.:

25 °C

Duration:

0 - 2 min

% Recovery:

91.5 - 97.7

pH:

7

Temp.:

25 °C

Duration:

< 1 min

% Recovery:

58.1 - 69

pH:

9

Temp.:

25 °C

Duration:

< 1 min

Key result

pH:

5

Temp.:

25 °C

Hydrolysis rate constant:

0.031 h-1

DT50:

> 7.5 d

Type:

(pseudo-)first order (= half-life)

Remarks on result:

other: Hydrolysis half-life of F-D3 in 20% acetonitrile/80% aqueous buffer (submission substance)

Key result

pH:

7

Temp.:

25 °C

Hydrolysis rate constant:

0.23 h-1

DT50:

6 d

Type:

(pseudo-)first order (= half-life)

Remarks on result:

other: Hydrolysis half-life of F-D3 in 20% acetonitrile/80% aqueous buffer (submission substance)

Key result

pH:

9

Temp.:

25 °C

Hydrolysis rate constant:

0.342 min-1

DT50:

11 min

Type:

(pseudo-)first order (= half-life)

Remarks on result:

other: Hydrolysis half-life of F-D3 in 20% acetonitrile/80% aqueous buffer (submission substance)

Key result

pH:

5

Temp.:

25 °C

Hydrolysis rate constant:

0.31 h-1

DT50:

2 h

Type:

(pseudo-)first order (= half-life)

Remarks on result:

other: Hydrolysis half-life of D3 in 20% acetonitrile/80% aqueous buffer (reference substance)

Key result

pH:

7

Temp.:

25 °C

Hydrolysis rate constant:

0.139 h-1

DT50:

5.4 h

Type:

(pseudo-)first order (= half-life)

Remarks on result:

other: Hydrolysis half-life of D3 in 20% acetonitrile/80% aqueous buffer (reference substance)

Key result

pH:

9

Temp.:

25 °C

Hydrolysis rate constant:

0.563 min-1

DT50:

1.1 min

Type:

(pseudo-)first order (= half-life)

Remarks on result:

other: Hydrolysis half-life of D3 in 20% acetonitrile/80% aqueous buffer (reference substance)

Key result

pH:

5

Temp.:

25 °C

DT50:

2 min

Type:

(pseudo-)first order (= half-life)

Remarks on result:

other: Hydrolysis half-life of D3 in 100% aqueous buffer (reference substance)

Key result

pH:

7

Temp.:

25 °C

DT50:

23 min

Type:

(pseudo-)first order (= half-life)

Remarks on result:

other: Hydrolysis half-life of D3 in 100% aqueous buffer (reference substance)

Key result

pH:

9

Temp.:

25 °C

DT50:

0.44 min

Type:

(pseudo-)first order (= half-life)

Remarks on result:

other: Hydrolysis half-life of D3 in 100% aqueous buffer (reference substance)

F-D3:

Hydrolysis of F-D3 did not proceed directly to completion under conditions of the experiments. Log_etransformed plots for the fraction of the test substance remaining against time were non-linear as indicated by r²= 0.44 and r²= 0.78 and for pH 5 and pH 7 hydrolysis experiments respectively. The kinetic range for the data points varied from 49 seconds to 6 days for pH 5, 48 seconds to 7.2 days for pH 7 and 6 seconds to 2 hours for pH 9.

After initial rapid decline for pH 7, the kinetic profile was characterised by attainment of a plateau at about 50% remaining. The kinetic profile for pH 9 was characterised by a steep drop (~40%) between the first measured time point at 6 seconds and the next time point at 38 seconds, followed by a more gradual decline to 9% parent remaining at 2 hours, which suggested rapid ring opening of the test substance initially until sufficient 1,3,5-tris[(3,3,3-trifluoropropyl)methyl]trisiloxane-1,5-diol was built up to observe reformation of the ring by intramolecular condensation. The kinetic profile for pH 5 was characterised by an initial decline, not quite as rapid as for pH 7 and 9, followed by attainment of a plateau t about 70% remaining. Perhaps the gradual decline observed for pH 9, rather than a plateau as observed for pH 5 and 7, is due to differences in the relative rates of hydrolysis for the parent substance and first intermediate. The plateau observed for pH 5 and 7 suggested that the test substance was in equilibrium with the inferred trisiloxanediol hydrolysis intermediate and that further changes in parent substance were controlled by slow degradation of this intermediate.

 

The test substance data (on a fraction parent remaining vs time basis) were fit into a reversible first order kinetic model using non-linear regression to test the hypothesis of equilibrium between hydrolysis ring opening F-D3 (Kf) and intramolecular condensation of 1,3,5-tris[(3,3,3-trifluoropropyl)methyl]trisiloxane-1,5-diol (Kr), with subsequent slow hydrolysis of 1,3,5-tris[(3,3,3-trifluoropropyl)methyl]trisiloxane-1,5-diol (k2). All data points were included in the non-linear regression analysis of the test substance. Table 1 contained values for the hydrolysis rate constants kf, kr, k2 half-lives as well as kinetic range and number of time points included in the regression for each pH experiment. Data from 7-11 sampling intervals (with replications) were included in each analysis. The experimental half-lives were 6 days at pH 7, 11 min at pH 9 at nominal 25°C. The half-life for pH 5 at nominal 25°C was >7.5 days, due to the plateau observed, 50% F-D3 was not reached during duration of the experiment.

A small fraction of F-D4 was observed in chromatograms for pH 5, 7 and 9 extracts. The kinetic profiles for pH 5 and 7show that there was no detectable response for F-D4 initially, followed by an increase up to 5 and 8% respectively of F-D3 applied. The kinetic profile for pH 9 shows that F-D4 fraction varied from <1% to 7% of the F-D3 applied. The higher initial concentration of F-D3 tested during method development, first at 7 ppm and later at 1.5 ppm resulted in larger fraction of F-D4 observed, up to 30% and 12% respectively. By decreasing the initial test substance concentration to ~0.4 ppm for the definitive study, the fraction of F-D3 was reduced while still being able to quantify F-D3. No hydrolysis products or condensation products of the test substance was observed in the GC-FID chromatogram due to poor extraction efficiency of these polar degradates and/or co-elution with extraction solvent.

Table 1: summary of kinetic results for test substance

pH buffer	pH with 20% ACN	Kinetic range (%)	No of points	Parameter, kf, kr, k2	SE, k	r^2	Half-life
4.991	5.310	108-67	14	Kf = 0.0311/h Kr = 0.0761/h K2 = 1.000/h	-	-	>7.5 d
7.077	6.810	97 – 48	22	Kf = 0.2298/h Kr = 0.5620/h K2 = 0.0085/h	-	-	6 d
8.958	9.526	69-9	18	Kf = 0.3422/min Kr = 0.4379/min K2 = 0.0320/min	-	-	11 min

 

D3:

Examination of the reference substance plots depicted the expected first-order behaviour spanning 1.3 – 2.0 half-lives for the pH 5, 7 and 9 hydrolysis data. The kinetic range for data points varied from 44 seconds to 4.9 h for pH 5, 56 seconds to 9.5 h for pH 7, and 8 seconds to 2.5 minutes for pH 9. The pH 7 and 9 plots appeared to reach a plateau at around 10% D3 remaining, while there are insufficient data to determine whether this was also the case for pH 5. All data points were included in the linear regression analysis for the reference substance except the later time points with <20% parent remaining. Table 2 contains value of the hydrolysis rate constants, k, half-lives, kinetic data range and number of time point included in the regression for each pH experiment. The results for the linear regression analysis yielded r2 value of 0.994, 0.974 and 0.982 for pH 5, 7 and 9 respectively. The experimental half-lives were 2 hours, 5.4 hours and 1.1 min at pH 5,7 and 9 respectively and nominal 25°C. Given the position of the plateau at low percent remaining, and the bias concern; the reference substance data were not fitted using non-linear regression. No hydrolysis products or condensation products of the test substance was observed in the GC-FID chromatogram due to poor extraction efficiency of these polar degradates and/or co-elution with extraction solvent.

Table 2: summary of kinetic results for reference substance

pH buffer	pH with 20% ACN	Kinetic range (%)	No of points	Slope,k	SE, k	r^2	Half-life
4.991	5.310	101-22	18	0.3103/h	0.0077/h	0.9903	2 h
7.077	6.810	115-25	14	0.1387/h	0.0067/h	0.9729	5.4 h
8.958	9.526	93-25	8	0.5634/min	0.337/min	0.9790	1.1 min

 

Time zero recoveries:

The time zero recoveries for both modified time zero samples and those treated the same way as the sample after spiking are presented in Table 3. By modifying the time zero recoveries for the test and reference substance experiments at pH 9, the recoveries were significantly improved to the range 94-95% and 87-88% respectively. In comparison, the non-modified time zero samples for the test and reference substances at pH 9 were 58-69% and 59-60% respectively due to the fast hydrolysis rate. There was no significant benefit to modify the procedure for the time zero samples for the test and reference substance at pH 5 and 7 because the average recovery for non-modified time zero samples were within acceptable range of 90 -110%.

Table 3: summary of time zero recoveries

pH buffer	pH with 20% ACN	Test substance (%)	Test substance (modified time zero) (%)	Reference substance (%)	Reference substance (modified time zero) (%)
4.991	5.310	107.6 (49 sec) 98.8 (70.9 sec)	-	101.1 (43.9 sec) 94.7 (47.9 sec)	-
7.077	6.810	91.5 (47.9 sec) 97.7 (50 sec)	-	108.9 (59 sec) 99.1 (56.2 sec)	-
8.958	9.526	69 (38 sec) 58.1 (43 sec)	94.4 (6 sec) 95.1 (9 sec)	59.4 (41 sec) 59.7 (42 sec)	86.5 (8 sec) 88.3 (9 sec)

 

LC-MS analysis

LC-MS analysis was conducted to confirm that methyl(3,3,3-trifluoropropyl)silanediol and DMSD concentrations were consistent with those expected based on degradation of the spiked F-D3 and D3 respectively.

F-D3

Table 4 shows the initial concentration of test substance in each sample, expected and measured concentrations of the final hydrolysis product methyl(3,3,3-trifluoropropyl)silanediol , and the percent of expected final hydrolysis product (i.e. recoveries). The recoveries of methyl(3,3,3-trifluoropropyl)silanediol relative to initial test substance concentration ranged from 105-127% for the six aged samples (aged 27-45 days). These results indicated that material balance was conserved in the hydrolysis of the test substance. The methyl(3,3,3-trifluoropropyl)silanediol standards prepared from the dimethoxy –methyl(3,3,3-trifluoropropyl)silane were used to construct a calibration curve. The calibration graph resulted in r²of 0.9925 indicating good linearity for the LC-MS response to the standard concentrations. The slope and y-intercept derived from the calibration graph were used to quantitate the six kinetic samples.

Table 4: Summary of LC-MS Analysis for test substance

Sample reference*	Nominal pH	Initial test substance concentration µg/g	Expected methyl(3,3,3-trifluoropropyl)silanediol concentration µg/g	Reaction time**, days	% remaining TFP by GC-MS	Measured***methyl(3,3,3-trifluoropropyl)silanediol concentration µg/g	% of target
TFP5 11	5	0.3919	0.4371	14.87	0	0.52	118
TFP5 12	5	0.3943	0.4397	14.87	0.8933	0.52	116
TFP7 8	7	0.3890	0.4339	22.96	0	0.46	105
TFP7 16	7	0.3887	0.4335	22.96	0	0.53	121
TFP9 21	9	0.4303	0.4799	5.74	0	0.61	127
TFP9 22	9	0.4312	0.4809	5.75	0	0.52	108

*TFP = FD-3

**Reaction time (spiked to LC-MS analysis) was an additional 22 days after reaction time shown in the table

*** Results with 5 calibration standard covering a large range with an r²= 0.9909

D3:

Table 5 shows initial concentration of D3 in each sample, expected concentration of the final hydrolysis product, DMSD, sample aging time (reaction time) and percent D3 remaining by GC-FID. The LC-MS analysis revealed that only the undiluted 560 ppm DMSD stock could be detected under same LC-MS conditions used for methyl(3,3,3-trifluoropropyl)silanediol. DMSD was not detected in 10x and 100x dilutios of DMSD standard in pH 7, 80/20 water/ACN buffer. Kinetic samples were not analysed because calibration graph could not be constructed for DMSD in the appropriate concentration range. Reported hydrolysis for other cyclic methyl siloxanes such as D4 and D5 indicate the final hydrolysis product is DMSD and that losses due to volatilisation is only significant for conditions associated with longer half-lives (days to weeks).

The rate of disappearance of the reference substance was faster than that of the test substance. For the reference substance, the apparent half-life was slower than that reported for 100% aqueous buffer for the same pH and temperature (pH 5 = 2.0, pH 7 = 23 and pH 9 = 0.44 min), which suggested that the addition of the ACN co-solvent had a decelerating effect. However, the half-life for D3 was comparable to that reported for 100% aqueous buffer at similar pH and temperature in that the slowest hydrolysis occurred at pH 7 and fastest hydrolysis at pH 9.

Table 5: Summary of LC-MS Analysis for reference substance

Sample reference	Nominal pH	Initial D3 concentration µg/g	Expected DMSD concentration µg/g	Reaction time*, days	% remaining D3 by GC-MS	Measured**DMSD concentration µg/g	% of target
D35 23	5	0.4312	0.5360	12.07	4.165	N/A	N/A
D35 24	5	0.4350	0.5407	12.07	2.953	N/A	N/A
D37 23	7	0.4296	0.5340	1.04	0.0019	N/A	N/A
D37 24	7	0.4354	0.5412	1.04	-1.143	N/A	N/A
D39 21	9	0.4831	0.6005	1.05	2.597	N/A	N/A
D39 22	9	0.4864	0.6046	1.05	0.756	N/A	N/A

*Reaction time (spiked to LC-MS analysis) was an additional 7-18 days after reaction time shown in the table

**Samples were not analysed, could not detect DMSD in standard dilutions to generate a calibration graph, therefore quantitative analysis could not be performed for DMSD

Analysis of hydrolysis kinetic data:

Test substance/test medium	Half-lives
	pH 5	pH 7	pH 9
Half-life of F-D3 in 20% ACN/80% aqueous buffer	>7.5 days	6 days	11 min
Half-life of D3 in 20% ACN/80% aqueous buffer	2 h	5.4 h	1.1 min
Half-life of D3 in 100% aqueous buffer*	2 min	23 min	0.44 min

*Sun, Ying Hydrolysis of Hexamethylcyclotrisiloxane (D3, CAS No: 541-05-9), HES Study number: 9760-102, Dow Corning internal Report No: 2005-I0000-54960

 

Conclusions:

Hydrolysis half-lives >7.5 d at pH 5, 6 days at pH 7 and 11 min at pH 9 and 25°C were determined for the substance using a relevant test method. The result is considered reliable.

Description of key information

Hydrolysis half-life: >7.5 days at pH 5, 6 days at pH 7 and 11 min at pH 9 and 25°C (Modified OECD 111)

Key value for chemical safety assessment

Half-life for hydrolysis:

6 d

at the temperature of:

25 °C

Additional information

Hydrolysis half-times of >7.5 days at pH 5, 6 days at pH 7 and 11 minutes at pH 9 and 25°C were determined for the submission substance (F-D3) in a modified study conducted in accordance with OECD 111. Extra co-solvent (20% acetonitrile) was added because the substance has a low solubility in water (1.3E-06 mg/l at 20°C, predicted). A higher concentration of co-solvent was used to enhance solubility and minimise losses by sorption and volatilisation. The hydrolysis half-lives of a reference substance, hexamethylcyclotrisiloxane (D3, CAS 541-05-9) for which a standard (100% aqueous buffer) OECD 111 study is available were also measured. The half-lives for D3 (reference substance) under the modified OECD 111 were 2 h at pH 5, 5.4 h at pH 7 and 1.1 minutes at pH 9 and 25°C. Under standard OECD 111 study conditions, the half-lives of D3 were 2 minutes at pH 5, 23 minutes at pH 7 and 0.44 minutes at pH 9 and 25°C. The results for the half-lives of D3 under standard conditions suggest that the half-lives for the submission substance, F-D3 under standard conditions could be shorter than those measured in the non-standard study. However, the low solubility and high adsorption capacity of F-D3 are expected to limit the hydrolysis observed in practise. The measured half-times are used for assessment purposes.

 

The half-times reported are for the hydrolysis (ring-opening) of the parent substance. The initial product of the hydrolysis of F-D3 is 1,5-dihydroxy-1,3,5-trimethyl-1,3,5-tris(3,3,3-trifluoropropyl)trisiloxane (the trimerdiol), which is formed by ring-opening of the parent substance. F-D3 did not show irreversible first-order behaviour. Instead, the kinetic profiles showed initial decline followed by attainment of a plateau. This was shown to be consistent with reversible hydrolytic ring-opening of the cyclic trimer followed by slower hydrolysis of the trimerdiol. The final hydrolysis product was demonstrated to be methyl(3,3,3-trifluoropropyl)silanediol.

Stoichiometric amounts of this final hydrolysis product were observed by LC-MS after 15 days at pH 5, 23 days at pH 7 and 6 days at pH 9, thus confirming hydrolysis of F-D3 rather than other possible loss mechanisms (such as volatilisation).

 

The attached figure shows the proposed hydrolysis reaction mechanism (water omitted) where A is F-D3 (parent), B is 1,3,5-tris[(3,3,3-trifluroproyl)methyl]trisiloxane-1,5-diol (the trimerdiol), C is 1,3-tris[(3,3,3-trifluroproyl)methyl]disiloxane-1,3-diol (the dimerdiol) and D and E are methyl(3,3,3-trifluoropropyl)silanediol (the silanol hydrolysis product). Based on numerous studies of hydrolytic degradation of cyclic and linear permethylsiloxanes in water, soil and sediment under environmentally relevant conditions of pH and temperature, siloxanes degrade by hydrolytic cleavage of the Si-O bond (Xu 1999, Lehmann et al 1994, Durham 2005, Xu 2009). However, the Si-C bond is not susceptible to abiotic hydrolysis. Therefore, the mechanism presented in Figure 4.1 is the only plausible one to explain the observed kinetic profiles from the study; it is consistent with the established chemistries of the silicon moieties. The observation that the rate of disappearance of the parent substance was much greater at the beginning and then slowed as the reaction proceeded, suggested that the hydrolytic ring opening was sufficiently faster than the hydrolysis of the intermediate trimerdiol such that the effect of the ring-chain hydrolysis-condensation equilibrium was observed as attainment of a (pseudo)plateau in the kinetic profiles. Although direct evidence of the formation of the proposed hydrolysis intermediates was not obtained, stoichiometrically quantitative formation of the final silanol product supported this scheme.

 

Figure 4.1 Proposed hydrolysis reaction mechanism

Formation of the cyclic tetramer, F-D4 (CAS# 429-67-4, EC 207-060-9, 2,4,6,8-tetramethyl-2,4,6,8-tetrakis(3,3,3-trifluoropropyl)cyclotetrasiloxane) was also demonstrated during the study. Up to 5% at pH 5, 8% at pH 7 and 7% at pH 9 F-D4 was observed by GC-FID. F-D4 can be formed by condensation reactions of the intermediate and final silanol hydrolysis products. Higher concentrations of the cyclic tetramer were observed during method development when using higher initial test substance concentration.

 

References

Xu S, Fate of Cyclic Methylsiloxanes in Soils. 1. The Degradation Pathway.Environ. Sci. Technol.1999, 33, 603-608.

Lehmann RG, Varaprath S, Frye CL, Degradation of Silicon Polymers in soil.Environ. Toxicol. Chem. 1994, 13, 1061-1064.

Durham J, Hydrolysis of Octamethylcyclotetrasiloxane (D4). Final report to the Silicones Environmental, Health and Safety Council (SEHSC) on HES Study No. 10000; Dow Corning Corporation: Auburn, MI, 2005.

Xu S, Anaerobic Transformation of Octamethylcyclotetrasiloxane (14C-D4) in Aquatic Sediment Systems. Final report to the Centre Européen des Silicones (CES) on HES Study No. 11101; Dow Corning Corporation: Auburn, MI, 2009.