Registration Dossier

Data platform availability banner - registered substances factsheets

Please be aware that this old REACH registration data factsheet is no longer maintained; it remains frozen as of 19th May 2023.

The new ECHA CHEM database has been released by ECHA, and it now contains all REACH registration data. There are more details on the transition of ECHA's published data to ECHA CHEM here.

Diss Factsheets

Administrative data

Link to relevant study record(s)

Description of key information

Key value for chemical safety assessment

Bioaccumulation potential:
low bioaccumulation potential

Additional information

The registered substance is a monoconstituent substance, 4,4,13,13-Tetraethoxy-3,14-dioxa-8,9-dithia-4,13-disilahexadecane (“S2”; CAS No. 56706-10-6, EC No. 260-350-7). There are no in vivo data on the toxicokinetics of the registered substance.

The following summary has therefore been prepared based on validated predictions of the physicochemical properties of the substance itself and related sulfidosilane test substances, using these data in algorithms that are the basis of many computer-based physiologically based pharmacokinetic or toxicokinetic (PBTK) prediction models. The main input variable for the majority of these algorithms is log Kow so by using this and, where appropriate, other known or predicted physicochemical properties of the substance or its hydrolysis products, reasonable predictions or statements may be made about their potential absorption, distribution, metabolism and excretion (ADME) properties.

S2 is a liquid of low volatility, with a measured vapour pressure of 9 Pa at 20°C. The substance hydrolyses at a moderate rate in contact with water (predicted half-life approximately 40-80 hours at pH 7), generating (3-{[3-(trihydroxysilyl)propyl]disulfanyl}propyl)silanetriol and ethanol hydrolysis products.

Human exposure can occur via the inhalation or dermal routes. Relevant exposure would be to the parent substance as significant hydrolysis in contact with skin or lung tissue is not expected to occur. However, based on available data for the hydrolysis of triethoxysilanes at acidic pH, if the substance is ingested, rapid hydrolysis may be expected.

The toxicokinetics of ethanol have been reviewed in other major reviews and are not considered further here.



Significant oral exposure is not expected for the parent substance S2.

However, oral exposure to humans via the environment may be relevant for the silanol hydrolysis product, (3-{[3-(trihydroxysilyl)propyl]disulfanyl}propyl)silanetriol. When oral exposure takes place it can be assumed, except for the most extreme of insoluble substances, that uptake through intestinal walls into the blood occurs. Uptake from intestines can be assumed to be possible for all substances that have appreciable solubility in water or lipid. Other mechanisms by which substances can be absorbed in the gastrointestinal tract include the passage of small water-soluble molecules (molecular weight up to around 200 g mol-1) through aqueous pores or carriage of such molecules across membranes with the bulk passage of water (Renwick, 1993).

In the event of exposure to the parent substance via the oral route, rapid hydrolysis to ethanol and (3-{[3-(trihydroxysilyl)propyl]disulfanyl}propyl)silanetriol is expected, therefore systemic exposure to the parent substance is not expected. Based on the high water solubility of the (3-{[3-(trihydroxysilyl)propyl]disulfanyl}propyl)silanetriol, uptake into the systemic circulation is considered likely. The key repeated oral toxicity study conducted with the parent substance and with a closely related substance (polysuphides containing 5-10% S1, 60-70% S2 and 15-20% S3) shows evidence of absorption of substance-related material following ingestion as adverse systemic effects were observed. Therefore, it is considered that should oral exposure take place it is reasonable to assume that resulting systemic exposure will occur also. 


The fat solubility and therefore potential dermal penetration of a substance can be estimated by using the water solubility and log Kow values. Substances with log Kow values between 1 and 4 favour dermal absorption (values between 2 and 3 are optimal) particularly if water solubility is high.

With a predicted log Kow of 5.2 and measured water solubility of <1 mg/l absorption of the parent substance across the skin is highly unlikely to occur. There may be penetration of the stratum corneum (limited by molecular weight), but there will then be extremely limited transfer from the stratum corneum to the epidermis. For the hydrolysis product (3-{[3-(trihydroxysilyl)propyl]disulfanyl}propyl)silanetriol, although the predicted water solubility is high, the predicted log Kow of -3.0 is far from the ideal range so penetration would be limited.



There is a QSPR to estimate the blood:air partition coefficient for human subjects as published by Meulenberg and Vijverberg (2000). The resulting algorithm uses the dimensionless Henry coefficient and the octanol:air partition coefficient (Koct:air) as input variables.

Using these values for the parent substance S2 results in a blood:air partition coefficient of approximately 0.5:1 meaning that if lung exposure occurred there would be little or no uptake into the systemic circulation. The high water solubility of the silanol hydrolysis product, results in an extremely high blood:air partition coefficient of approximately 3.6E+12, assuming a predicted water solubility of 1.0E+06 mg/l. Therefore once hydrolysis has occurred, as it would be expected to in the lungs, then significant uptake would be expected into the systemic circulation. However, the high water solubility of the hydrolysis product may lead to some of it being retained in the mucus of the lungs so once hydrolysis has occurred, absorption is likely to slow down.


For blood:tissue partitioning a QSPR algorithm has been developed by De Jongh et al. (1997) in which the distribution of compounds between blood and human body tissues as a function of water and lipid content of tissues and the n-octanol:water partition coefficient (Kow) is described.

Using this value for the parent substance predicts that, should systemic exposure occur, distribution would primarily be into fat, with potential slight distribution into liver, muscle, brain and kidney.

For the hydrolysis product, (3-{[3-(trihydroxysilyl)propyl]disulfanyl}propyl)silanetriol, distribution into the main body compartments would be minimal with tissue:blood partition coefficients of less than 1 for all major tissues (approaching zero for fat).

Table 1: tissue:blood partition coefficients


Log Kow







Parent substance S2

















Rapid hydrolysis to ethanol and the corresponding silanol is expected in the stomach following oral ingestion. Animals in the key repeated oral toxicity study were shown to have liver hypertrophy, which could be an indication of hepatic metabolism following ingestion of S2. In genetic toxicity tests in vitro, there were no observable differences in effects with or without metabolic activation.

Information on the metabolism of disulfides compared to tri- and tetrasulfides, based on non-silicon substances, indicates that there is also no apparent difference in metabolic pathways.

The labile nature of the S-S bond results in a variety of metabolic routes for detoxification. The disulfide bond may be reduced to give the corresponding thiol in a reversible reactionin vivo. Therefore, the metabolic transformations of thiols are also applicable to disulfides. Thiol-disulfide exchange reactions are common in vivo and result from nucleophilic substitution by sulfur. Thiol-disulfide exchange reactions with endogenous cellular thiols (reduced glutathione, GuSH) or disulfides (oxidised glutathione, GuSSGu) will produce mixed disulfides that may also undergo reduction. Also, thiol groups on proteins (surface cysteine residues) or other nucleophilic groups may be involved and in many cases such thiolate substitution will affect the biological function of the protein. In blood plasma, the main protein-containing thiol is Cys34 of serum albumin, which constitutes ~80% of the free thiols in blood and which reacts selectively with reactive oxygen and nitrogen species (Carballa et al. 2003). Under normal conditions, disulfide exchange reactions control the cellular concentrations of endogenous thiols (reduced and oxidised glutathione and maintenance of an adequate GuSH/GuSSGu ratio is essential for cell survival and function (Cotgreave et al., 1989; Brigelius, 1985; Sies et al., 1987). 

Further metabolism of any free thiols produced will generate thiosulfinates, thiosulfones and eventually polar sulfates, which are generally excreted. Oxidation of thiols is catalysed by cytochrome P450 and flavin mono-oxygenases, generally in the liver. Thiols may also be methylated via S-adenosylmethionine (SAM)-dependent thiol methylation to yield a thio-ether, which is usually oxidised to the sulfoxide (major) and sulfone (minor) polar metabolites, which are then excreted. The above general metabolites are those typically seen in the metabolism of diallyldisulfide (DAD) by rat and human hepatocytes or perfused liver (ex vivo) by Germaine et al.(2003), who detected the following metabolites (shown with relevant pathways):

·        DADH2C=CH-CH2-S(=O)-S-CH2-CH=CH2(diallylthiosulfinate, allicin)

·        Diallyltiosulfinate + GuSHGuS- S-CH2-CH=CH2(non enzymatic)

·        DAD + SAMCH3-S-CH2-CH=CH2CH3-S(=O)-CH2-CH=CH2H3-S(=O)2-CH2-CH=CH2

·        DAD + GuSHGu-S-CH2-CH=CH2(direct nucleophilic attack)

·        DADH2C=CH-CH2-SH (allyl mercaptan) 

The rate of metabolism of DAD by liver was very rapid, with a t½ of approximately 6 min. It can thus be deduced that disulfides are rapidly metabolised in mammalian systems by a combination of thiol-disulfide exchange reactions and oxidation steps. The metabolic mobility of thiols and disulfides can also be seen in experiments where disulfides and thioethers are detected in vivo when a thiol is administered. For example, when rats were exposed to methanethiol via inhalation, both dimethylsulfide and dimethyldisulfide were detected in the expired air (Blomet al., 1990).

No studies on the metabolic pathways of tri- or polysulfides could be found. However, we may infer that thiol-trisulfide exchange reactions would be even more facile than with disulfides as the inter-sulfur bond energy is weaker (70, 45 and 36 kcal mol-1 for di-, tri and tetrasulfides, respectively) (Pickeringet al., 1967). The same paper also showed that both dimethyltrisulfide and dimethyltetrasulfide were thermally unstable, forming the disulfide, polysulfides and elemental sulfur when heated to 80oC. A free radical mechanism was proposed. The scission of the sulfur-sulfur bonds of trisulfides by nucleophiles occurs readily. This has been attributed to the large polarisable sulfur atom, which may accommodate the negative charge of a thiol or a hydrodisulfide ion thus making these ions good leaving groups in such reactions. Hence, for trisulfides where X-is the nucleophile and the centre sulfur is the electrophile:

R-S-S-S-R + X-R-S-S-X + RS-RSSR + SX (Ash, 1973).

However, there is evidence that nucleophilic attack on a terminal, rather than the central sulfur of an organic trisulfide also occurs, possibly preferentially, with thiol nucleophiles. This evidence was afforded by the reaction of diethyltrisulfide with ethane thiol in the presence of piperidine (a non-nucleophilic nitrogenous base), which produced two moles of diethyl disulfide per mole of diethyl trisulfide plus hydrogen sulfide (Evans and Saville, 1962)viz: Et-S-S-S-Et + 2Et-SH2Et-S-S-Et + H2S, whereas attack on the central sulfur would lead to no net change in products.

In the case of glutathione as the nucleophile, this may be represented as:

Et-S-S-S-Et + 2GuSH2GuS-S-Et + H2S

Consequently, the trisulfides as well as disulfides would be expected to participate in thiol-exchange reactions and in the case of trisulfides to generate a mixed disulfide and hydrogen sulfide. Tetrasulfides produce mainly disulfides under similar reaction conditions (Steudel, 2002).

Clearly, if this mechanism held true in vivo, H2S would be generated from trisulfides but not tetrasulfides or disulfides and due to the high acute toxicity of hydrogen sulfide, this may have a toxicological consequence.  Hydrogen sulfide is an acute respiratory poison that reacts irreversibly with the haem of cytochrome oxidase, a toxic mechanism shared by cyanide, azide and carbon monoxide. The available data for H2S are mainly for inhalation/respiratory effects which are not relevant for the sulfidosilanes.  Studies on the reproductive effects of sub-lethal chronic exposure to hydrogen sulfide have been reported in both animal studies and human exposure reports.  These reports are somewhat equivocal and contradictory; however, the mammalian reproductive effects of chronic exposure to hydrogen sulfide appear to be slight. In particular, a GLP study on rats reported by Dormanet al. (2000) no significant treatment related effects on reproductive performance or offspring were observed. Hydrogen sulfide is not classified for reproductive toxicity in Annex VI to Regulation EC (No) 1272/2008. The in vitro evidence that the hydrogen sulfide might be produced in vivo from the trisulfide would not affect the validity of read-across from Polysulfides to S2, as the former substance contains a higher percentage of the trisulfide.  Consequently, the mixture (CAS 211519-85-6) would be more likely to manifest highertoxicity.

Furthermore, a European Food Standards Agency Opinion (EFSA, 2005) reports the following:

“Disulfides [FL-no: 13.113, 13.144, or 13.178] can be reduced to give the free thiols, or can be oxidised to give thiosulfinates or thiosulfones.For the one trisulfide candidate substance [FL-no:13.146], no information on biotransformation was available, but it may be expected that this substance is metabolised via similar routes as disulfides.”

“No information about the metabolism of trisulfides could be found. However, as these substances may be considered as a special case of disulfides, it may be assumed that some of the reduction and oxidation reactions described above may also apply to them.”

In order to assess whether read-across of toxicological data is valid from information on a mixture of di-, tri and tetrasulfidosilanes (CAS 211519-85-6) to S2 (CAS 56706-10-6) and S2/S3 (EC No. 915-748-1), data on the comparative toxicology between disulfides and tri- (or higher) sulfides was sought. Although there are many toxicological data in the public domain on the toxicology of disulfides, no useful information on the toxicity of higher homologues could be found. However, a regulatory precedent exists for using read-across from furfuryldisulfides to a furfuryltrisulfide used as a food flavouring in a European Food Safety Agency (EFSA, 2005) report on the basis of inferred similarity of in vivo bio-transformations via thiol-disulfide exchange (EFSA, 2005). Organic linear disulfides are rapidly metabolised in vivo via facile and rapid thiol-disulfide exchange reactions (often involving glutathione), and oxidation steps, forming mixed disulfides and oxidised sulfur species. There is mechanistic and energetic evidence that with higher homologues (tri- and tetrasulfides) the thiol-disulfide exchange reactions are even more energetically favourable, although in the case of trisulfides there is mechanistic evidence that hydrogen sulfide may be an additional metabolite. In all cases, the disulfide is a major product of breakdown product of tri- and tetrasulfides (Steudel, 2002).

In view of the inferred similar biotransformation routes of tri- and tetrasulfides and their disulfide homologues, in addition to the lack of acute toxicity for either substance and the similar toxicological profile seen in repeated-dose oral toxicity studies, read-across of reproductive and developmental toxicity data from a mixture of di-, tri, and tetrasulfidosilanes (CAS 211519-85-6) to S2 (CAS 56706-10-6) and to S2/S3 (EC No. 915-748-1) is considered to be valid and no disproportionate effects would be expected from treatment with the disulfidosilane alone.

As discussed above, there are a number of routes for the metabolic transformation of molecules containing disulfide bridges and these are likely to be applicable to all the substances in the sulfidosilane group. The enzymes involved in these pathways are not substrate specific, as indicated by the range of reactions that have been characterised a few of which are described above (see for example Hodgson, E. and Levi, PE, 1988).

The impurity S1 is also likely to be metabolised by commonly occurring metabolic pathways. The metabolism of sulfides of general structure R-S-R has been investigated (Hodgson, E. and Levi, PE, 1988). Sulfides are initially oxidised in the presence of NADPH, and then undergo conjugation with glutathione or similar endogenous compounds. The final products of metabolism are therefore unlikely to be distinguishable from those of the other constituents and impurities.

In conclusion, the available metabolic pathways produce the same product for all the substances, with the possible exception of the S1 impurity in the non-registered substance low purity S2.


A determinant of the extent of urinary excretion is the soluble fraction in blood. QPSRs as developed by De Jongh et al. (1997) using log Kow as an input parameter, calculate the solubility in blood based on lipid fractions in the blood assuming that human blood contains 0.7% lipids.

Using this algorithm, the soluble fraction of the parent substance S2 in blood is <0.001%, therefore, should systemic exposure occur it would not be eliminated via the urine. However, the corresponding figure for the silanol hydrolysis product is >99.0% meaning that it is likely to be effectively eliminated via the kidneys in urine.  


Renwick A. G. (1993) Data-derived safety factors for the evaluation of food additives and environmental contaminants. Fd. Addit. Contam.10: 275-305.

Meulenberg, C.J. and H.P. Vijverberg, Empirical relations predicting human and rat tissue:air partition coefficients of volatile organic compounds. Toxicol Appl Pharmacol, 2000. 165(3): p. 206-16.

DeJongh, J., H.J. Verhaar, and J.L. Hermens, A quantitative property-property relationship (QPPR) approach to estimate in vitro tissue-blood partition coefficients of organic chemicals in rats and humans. Arch Toxicol, 1997.72(1): p. 17-25.