POLY(CARBORANE-CO-SILOXANES) AND METHODS FOR THEIR PRODUCTION

The present application relates to methods of producing poly(carborane-co-siloxane) polymers, and to polymers produced by such methods.

BACKGROUND

Poly(carborane-co-siloxane)s are attractive polymers for use in a wide range of applications due to their excellent thermal stability, chemical resistance, high boron content, and relatively low glass transition temperatures.

The thermal stability of such materials means that they are well suited for use in high temperature environments and applications which require long service lifetimes. For example, such polymers have proven commercially useful as stationary phases for high temperature gas chromatography columns, and are sought after materials for gaskets, seals and electronic shielding in the aerospace industry. The high boron content of such materials means that they are also of interest in applications where neutron shielding is required (for example, shielding materials for neutron generators and electronics) and in neutron capture therapy of cancer, since the ¹⁰ B isotope (~20% natural abundance) has a high cross section for thermal neutron capture.

Poly(carborane-dimethylsiloxane)s were developed and commercialised by the Olin

Chemical Company in the 1970s under the tradename Dexsil®. Typically, Dexsil® polymers comprise dialkyl- or diarylsiloxane repeat units linked by m-dicarba-c/oso-dodecaborane fragments, together with a small fraction of a vinyl functionality to promote cross-linking. The simplest Dexsil® polymers are essentially co-polymers of dimethylsiloxane and m-dicarba- c/oso-dodecaborane, as illustrated below (N.B., unmarked vertices of the carborane cage indicate B-H g

Dexsil® 100

Dexsil® 200

Dexsil® 300

Dexsil® 400 As can be seen, the first digit in the identification number (i.e., 1 , 2, 3, 4) indicates the number of Si-0 units in the repeat unit. The last digit can be used to indicate different substituents on the Si backbone. The number of Si-O units in the repeat unit and the degree of cross-linking influences the elastomeric and thermal properties of the material.

Despite the large range of applications for Dexsil® polymers and other poly(carborane-cp- siloxane)s, their uptake has been hindered by the lack of suitable methods for their production.

Poly(carborane-co-siloxane)s have conventionally been prepared via an FeC -mediated condensation between dichloro- and dimethoxy- terminated siloxane monomers

[Knollmueller]. For example, bis(methoxydimethylsilyl)-m-carborane self-condenses in the presence of FeCI ₃ to produce Dexsil® 100, whilst addition of dichlorosilanes produces higher order Dexsil® polymers.

However, there are several significant disadvantages associated with this route:

Firstly, the chlorosilane starting materials are moisture-sensitive and form oligomers and volatile cyclic siloxanes, hence reaction conditions must be carefully controlled in order to maintain reagent purity.

Secondly, under the reaction conditions methoxy- and chlorosilane groups undergo exchange and bis(methoxydimethylsilyl)-m-carborane undergoes self-condensation. As a result, the material produced in the reaction is not a true co-polymer, and the material's properties are difficult to predict or model since the polymer structure is poorly controlled.

Thirdly, the reaction must be carried out at high temperatures leading to increased expense, control issues and safety concerns. In addition, to achieve efficient polymerisation the temperature of the reaction must be altered over the course of the reaction, suggesting a complex stepwise mechanism. For example, polymerisation at 116 °C has been shown to slow at around 50% conversion of the monomer, and the addition of more catalyst and utilisation of higher temperatures are required to drive the reaction further. This indicates that it is difficult to control the reaction through control of temperature - a fact which is particularly relevant for reproducibility between different reaction scales.

Fourthly, there is unavoidable cross-linking between polymer chains during the reaction, possibly due to the reaction proceeding via reactive radical intermediates. As well as preventing control over the properties of the final product, this cross-linking also means that iron-containing impurities (including FeC catalyst) is trapped within the cross-linked network of the final product. These impurities result in a coloured (black or brown) product, and have the potential to compromise the performance of the materials over time (e.g., due to further reactions in the final product).

Fifthly, the use of FeCI ₃ as a catalyst prevents the introduction of solvent resistant moieties, such as trifluoropropyl groups, or cross-linkable moieties, such as vinyl groups. Sixthly, gaseous products are evolved from the reaction, including chloromethane - a toxic and flammable greenhouse gas [Dietrich].

Given the serious shortfalls with the FeCI ₃ catalysed route, various alternative reaction schemes have been proposed.

Hedaya and co-workers have proposed production of poly(carborane-co-siloxane)s by reaction of bis(hydroxyorganosilyl)carboranes with either a silyl diamine, a ureido-silane or a silyl dicarbamate [Hedaya (a)], [Hedaya (b)], [Stewart]. However, there are problems associated with this reaction. For example, it has been found that reactions involving bis(N,N'-dimethylamino)dimethylsilane and bis(N,N'-dimethylcarbamato)dimethylsilane lead to low molecular weight products, because the dimethylamine byproduct in these reactions facilely splits the Si-C carborane bond in bis(hydroxyorganosilyl)carborane [Hedaya (a)].

In addition, reaction routes starting from ureido-silane are problematic because the preferred ureido-silane starting material is air- and water-sensitive which requires air-sensitive stepwise synthesis starting from hazardous isocyanate [Zhang].

More recently, Zhang and colleagues have examined the production of poly(carborane-co- siloxane)s through reaction of cyclotrisilazanes with bis(hydroxydimethyl)dicarba-c/oso- dodecaborane [Zhang]. They found that, in the absence of a catalyst, the reaction produced <70% yields of low molecular weight polymer, containing residual silazane bonds, which were prone to react with atmospheric moisture to release NH3. However, addition of catalytic (NH SO2 improved the reaction, leading to higher yields of higher molecular weight polymer. The scheme is limited to the production of poly(carborane-co-organotrisiloxane) polymers, however, so does not provide a general scheme for accessing a series of poly(carborane-co-siloxane)s.

Thus, there remains a need to develop general syntheses for the production of

poly(carborane-co-siloxane)s.

SUMMARY OF THE INVENTION

In a first aspect, the present invention provides methods of producing

poly(carborane-co-siloxane)s by reacting a siloxane compound having at least two Si-OR° bonds with an organohydrosiloxane compound having at least two Si-H bonds in the presence of a Lewis acid catalyst containing boron, wherein the siloxane and/or

organohydrosiloxane comprise a divalent carboranyl group.

More specifically, the present invention provides a method of producing a poly(carborane-co- siloxane) comprising a polymerisation step in which:

(a) a siloxane compound having at least two Si-OR° bonds is reacted with

(b) an organohydrosiloxane compound having at least two Si-H bonds in the presence of,

(c) a Lewis acid catalyst containing boron, (a) and/or (b) comprise a divalent carboranyl group;

-R° is -H or -R';

-R' is -R ⁰⁰ , -R ⁰⁰² , or -L°-R ⁰⁰² ;

-R ⁰⁰¹ is an optionally substituted -CMO alkyl, -Cwo haloalkyl -C2-io alkenyl, or -C2-10 alkynyl group;

-R ⁰⁰² is an optionally substituted -C3-8 cycloalkyl, -Ce-io aryl, or -C _6- io haloaryl group; and

-L°- is C1-4 alkylene.

This method has a number of significant advantages over known processes:

• the method provides a general route to a wide range of poly(carborane-co-siloxane)s, for example, poly(carborane-co-organotrisiloxane)s, poly(carborane-co- organotetrasiloxane)s, poly(carborane-co-organopentasiloxane)s, etc., unlike some prior art processes (such as that proposed by Zhang and co-workers) which are limited to poly(carborane-co-organotrisiloxane)s.

• the polymerisation uses reagents which are able to tolerate moisture, unlike the

chlorosilanes used in the conventional FeC -catalysed reaction, and the silazanes used in other prior art processes.

• the polymerisation can be performed at ambient (room) temperature, avoiding the need for expensive and potentially dangerous heating systems.

• the polymerisation produces a well-controlled material structure, and thus materials' properties are well-defined - the reaction should result in a true co-polymer of alternating monomers. This means that the reaction can be used to reliably and reproducibly produce poly(carborane-co-siloxane)s having desired characteristics.

• the polymerisation scheme of the present invention allows the avoidance of

unintended cross-linking side reactions, since there are no radical intermediates.

• the Lewis acid containing boron preferentially catalyses the reaction between Si-OR° and Si-H bonds, meaning that other functional groups (e.g., crosslinkable groups) can be left unaffected during the polymerisation step, and reacted after the polymerisation step.

• the method can be used to produce products over a range of molecular weights.

• the catalyst is easily removed post-polymerisation.

• the product is free from residual metal and other unwanted high atomic number

elements.

• the product is clear/opaque and colourless.

In a second aspect, the present invention provides a poly(carborane-co-siloxane) obtainable by the method of the first aspect. As discussed above, the method of the present invention allows the production of clear/opaque and colourless products having a controlled molecular weight and degree of crosslinking, unlike the conventional FeCI ₃ catalysed products.

In a third aspect, the present invention provides use of a Lewis acid catalyst containing boron during a polymerisation process to produce a poly(carborane-co-siloxane).

DETAILED DESCRIPTION

Catalyst

Preferably, the Lewis acid catalyst has the formula (R ^A ) _a B(R ^B )b, wherein: each -R ^A is independently -H, -OH, or halo; each -R ^B is independently -Ci-ioalkyl, -C2-10 alkenyl, -C3-12 cycloalkyl, or -C6-ioaryl, or two -R ^B radicals bond to one another so as to form, with the boron atom to which they are bonded, a 5- to 14-membered ring; with said ring being able to be saturated, unsaturated, bridged and/or aromatic, and to comprise one or more heteroatoms chosen from oxygen, nitrogen and boron atoms; wherein each -R ^B is optionally substituted with one or more (e.g., 1 , 2, 3, 4 or 5) electron-withdrawing groups; a is 0, 1 or 2; b is 1 , 2 or 3; and and a+b is 3.

Preferably -R ^B is -C6-io aryl (preferably phenyl) optionally substituted with one or more electron-withdrawing groups.

Suitable electron-withdrawing groups for substituting -R ^B are, for example, halo, -CF ₃ , -N0 ₂ , -CN, -OCF3, -SF ₅ , or -OSO2CF3. Preferably, each -R ^B is substituted with at least one electron-withdrawing group, more preferably at least one halogen atom, most preferably at least one fluorine atom.

For example, each -R ^B may be a -Ce-io aryl (preferably phenyl) substituted with 3, 4 or 5 halo atoms (preferably fluorine atoms).

The catalyst may be, for example, B(R ^B )3. Most preferably, the catalyst is B(C ₆ F ₅ )3.

Advantageously, this catalyst is very effective at catalysing the coupling of components (a) and (b), and is thermally robust, oxidation-resistant and water-tolerant. B(C ₆ F ₅ ) ₃ is believed to activate the silicon atom of a silane to attack by the oxygen of an alkoxysilane, as shown in Scheme 1 [Chojnowski].

RO-Si-0

I SiR'3

The proposed reaction mechanism proceeds via initial formation of borohydride and oxonium intermediates; the resulting transition state can conceivably collapse by hydride transfer via three pathways to either regenerate silane and alkoxysilane (a reversible exchange process, pathways a and b in Scheme 1) or expulsion of an alkane with concomitant generation of the siloxane bond (a redox process, pathway c, Scheme 1). The latter process is irreversible under the reaction conditions and is driven by the high entropy associated with the release of the gaseous alkane.

Component (a)

Component (a) is a siloxane molecule, preferably containing a divalent carboranyl group, which contains at least two Si-OR° bonds per molecule (preferably at least two Si-OH bonds).

Preferably, component (a) is an optionally substituted bis(organosilyl)carborane having at least two Si-OR° bonds (most preferably at least two Si-OH bonds).

More preferably, component (a) is an optionally substituted bis(organosilyl)carborane having at least one Si-OR° bond per organosilyl group (i.e., there is an Si-OR° bond either side of the carboranyl group). Advantageously, having at least one Si-OR° bond per organosilyl group allows the polymer to grow so that the carboranyl moiety appears in the backbone of the polymer (as opposed to in a side-chain group). This can be used as a general route to producing Dexsil® type polymers with various advantages over the routes discussed above.

For example, component (a) may be an optionally substituted

bis(alkoxydialkylsilyl)carborane, an optionally substituted bis(alkoxydiarylsilyl)carborane, an optionally substituted bis(alkoxy(alkyl)(aryl)silyl)carborane, an optionally substituted bis(hydroxydialkylsilyl)carborane, an optionally substituted bis(hydroxydiarylsilyl)carborane, or an optionally substituted bis(hydroxy(alkyl)(aryl)silyl)carborane.

Most preferably, component (a) is an optionally substituted bis(organosilyl)carborane having at least one Si-OH bond per organosilyl group.

Suitable carborane starting materials are commercially available, for example from KatChem Ltd. The carbon sites of such carboranes can be functionalised with organosilyl groups using methods known in the art, such as those described in [Gomez] and

[Gonzalez-Campo]. For example, carboranes may be functionalised by deprotonating with ⁿ BuLi and reacting with a chlorosilane under dry, inert conditions in THF or diethyl ether solvent at low temperature (e.g., -70°C).

Component (a) may be, for example, a compound having the following formula:

wherein:

-X- is a carboranyl group or -0-; each -R ¹ and -R ^1* is independently -OR ⁰ or -R ^s ; each -R ² and -R ^2* , when present, is independently -OR ⁰ or -R ^s ; each -R ^s is independently -R ^c or -N=C=0; each -R ^c is -R ^CC1 , -R ^CC2 , or -L ^C -R ^CC2 ; each -R ^CC1 is an optionally substituted -C1-10 alkyl, -CMO haloalkyl, -C2-io alkenyl, or -C2-io alkynyl group; each -R ^CC2 is an optionally substituted -C3-8 cycloalkyl, -C3-8 heterocycloalkyl, -Ce-10 aryl, -Ce-io haloaryl or -Cs-io heteroaryl group; each -L ^c - is Ci-4 alkylene; p is≥ 0; q is≥ 0; and there are at least two Si-OR ⁰ bonds per molecule.

Preferably, component (a) is a compound having the following formula:

wherein X, R°, R\ R , R ² , R ² ", p and q are as defined above.

In such embodiments, all R ¹ , R ^1* , R ² , and R ^2* groups may be -R ^s i.e., there may be only two Si-OR° bonds per molecule. Advantageously, this allows the production of linear polymers, which can be optionally crosslinked through the inclusion of suitable crosslinkable groups in the polymer chain.

In such embodiments, component (a) preferably has the following formula:

wherein X, R°, R\ R ^1* , R ² , R ^2* , p and q are as defined above.

The upper limit for the integers p and q may be, for example, 100, 50, 20, 15, 10, 8, 6, 5, 4, 3 or 2.

The lower limit for the integers p and q may be, for example, 0, 1 , 2, 3 or 4. For example, the integers p arid q may be 0-20, 0-10, 0-5 or 0-3.

Preferably, p = q. When p = q = 0, the component (a) may be a compound having the following formula:

H O— Si— X Si— O H

Preferably, -X- is a carboranyl group, preferably a closo carborane such as a dicarba-c/oso- dodecaboranyl group, i.e., a group having the formula -CB10H10C-. For example, -X- may be 1 ,12-dicarba-c/oso-dodecaboranyl (p-dicarba-c/oso-dodecaboranyl, often referred to as "p- carboranyl") or, most preferably, 1 ,7-dicarba-c/oso-dodecaboranyl (m-dicarba-c/oso- dodecaboranyl, often referred to as "m-carboranyl"). Dicarba-c/oso-decaborane groups are particularly useful due to their high boron content, chemical stability, and their ability to confer radiation and thermal stabilities to the poly(carborane-co-siloxane).

In particularly preferred embodiments, component (a) is a bis(hydroxy(dialkyl)silyl)carborane (in particular, bis(hydroxy(dimethyl)silyl)carborane)), in contrast to the FeCI ₃ catalysed process which starts from a bis(methoxy(dialkyl)silyl)carborane. Surprisingly, such starting materials result in quick polymerisation at high yields despite their relatively low solubility in the organic solvents typically used in coupling reactions. In addition, such reagents do not contain the acidic-impurities typically present in bis(alkoxy(dialkyl)silyl)carboranes, which can compromise the stability of the polymer product.

Component (b)

Component (b) may be, for example, a compound having the following formula:

wherein: each -R ³ and -R ^3* is independently H or -R ^s ; each -R ⁴ and -R ^4* , when present, is independently H or -R ^s ; each -R ⁵ and -R ^5* , when present, is independently H or -R ^s ; each -R ^s is independently -R ^c or -N=C=0; -R ^c is -R ^CC , -R ^CC2 , or -L ^C -R ^CC2 ;

-R ^CC1 is an optionally substituted -Ci-ioalkyl, -CMO haloalkyl, -C _2- io alkenyl, or -C2-10 alkynyl group;

-RCC2 j _{S an} optionally substituted -C ₃ -8 cycloalkyl, -C3-8 heterocycloalkyl, -C _6- io aryl, -C _6- io haloaryl or -Cs-io heteroaryl, group;

-L ^c - is C1-4 alkylene; and either:

(i) -Y- is an oxygen atom, r is≥ 0, s is 0, t is 0, u is≥ 0; or

(ii) -Y- is a carboranyl group, r is≥ 0, s is 1 , t is 1 , u is≥ 0; wherein there are at least two Si-H bonds per molecule.

Preferably at least two of the groups -R ³ , -R ^3* , -R ⁴ and -R ^4* are -H.

More preferably, at least one -R ³ or -R ⁴ is -H and at least one -R ^3* and -R ^4* is -H.

The upper limit for the integers r and u may be, for example, 100, 50, 20, 15, 10, 9, 8, 7, 6, 5, 4, 3 or 2.

The lower limit for the integers r and u may be, for example, 0, 1 , 2, 3 or 4. For example, the integers r and u may be 0-20, 0-10, 0-5 or 0-3. Preferably, r = u.

Optionally, component (b) does not include a divalent carboranyl group. In such

embodiments, component (b) may have the following formula:

wherein -R ³ , -R ^3* , -R ⁴ and r are as defined above.

Preferably, at least one -R ³ and one -R ^3* are -H, i.e., component (b) is a compound having the following formula:

When -Y- is a carboranyl group, it is preferred that the group is a closo carborane such as a dicarba-c/oso-dodecaboranyl group having the formula -CB10H10C-. For example, -Y- may be 1 ,12-dicarba-c/oso-dodecaboranyl (p-dicarba-c/oso-dodecaboranyl, often referred to as "p-carboranyl") or, most preferably, 1 ,7-dicarba-c/oso-dodecaboranyl (m-dicarba-c/oso- dodecaboranyl, often referred to as "m-carboranyl").

Suitably, in embodiments in which -Y- is a carboranyl group, at least one -R ³ or -R ⁴ group is -H and at least one R ^3* or R ^4* group is H, i.e., there at least two Si-H bonds which are not adjacent to the carboranyl group. The inventors have found that the carboranyl moiety can deactivate adjacent Si-H bonds, and thus ensuring that there at least two Si-H bonds which are not adjacent to the carboranyl group promotes polymerisation.

Relative amounts of components

The ratio of component (a) to component (b) may be, for example 0.5:1 to 1.5:1 , more preferably 0.8:1 to 1.2:1 , most preferably around 1.0:1.0.

It is possible to include a slight excess of component (a) compared to component (b) (for example, 1.0:1.0 to 1.1 :1.0) to favour incorporation of -Si-OR° groups in the final polymer, which can undergo further functionalisation.

Catalyst removal

Suitably, the catalyst (c) is removed after the polymerisation step. For example, the catalyst may be bound by a sequestering/chelating agent, such as AI2O3, to form a complex which is then separated from the reaction mixture (e.g., by filtration), in accordance with literature reports [Thompson]. Advantageously, the removal of catalyst (c) from the

poly(carborane-co-siloxane) product is relatively easy compared to removal of FeCI ₃ catalyst from poly(carborane-co-siloxane)s produced by the conventional manufacturing method, since the present method does not result in uncontrolled crosslinking.

Crosslinkinq

Optionally, the method of producing a poly(carborane-co-siloxane) involves

crosslinking/vulcanising the poly(carborane-co-siloxane) to create a crosslinked network, (e.g. an elastomeric material). To achieve such crosslinking, the reacting components may contain crosslinkable moieties, for crosslinking poly(carborane-co-siloxane) chains.

For example, component (a) may be substituted with at least one crosslinkable moiety in addition to the two Si-OR° bonds (i.e., component (a) may be a tri-functional, tetra-functional etc. component). Optionally, all component (a) molecules are substituted with at least one crosslinkable moiety in addition to the two Si-OR° bonds. Alternatively, only a fraction of component (a) molecules have a crosslinkable moiety in addition to the two Si-OR° bonds. For example, the percentage of component (a) molecules having a crosslinkable moiety may be 1 mol% or less, 2 mol% or less, 3 mol% or less, 5 mol% or less, 8 mol% or less, 10 mol% or less, 15 mol% or less, 20 mol% or less, 30 mol% or less, 40 mol% or less, 50 mol% or less, 75 mol% or less, or 90 mol% or less (the mol% being based on the total number of component (a) molecules in the polymerisation step).

Additionally or alternatively, component (b) may be substituted with at least one

crosslinkable moiety in addition to the two Si-H bonds. Optionally, all component (b) molecules are substituted with at least one crosslinkable moiety in addition to the two Si-H bonds. Alternatively, only a fraction of component (b) molecules have a crosslinkable moiety in addition to the two Si-H bonds. For example, the percentage of component (b) molecules having a crosslinkable moiety may be 1 mol% or less, 2 mol% or less, 3 mol% or less, 5 mol% or less, 8 mol% or less, 10 mol% or less, 15 mol% or less, 20 mol% or less, 30 mol% or less, 40 mol% or less, 50 mol% or less, 75 mol% or less, or 90 mol% or less (the mol% being based on the total number of component (b) molecules in the polymerisation step).

Additionally or alternatively, the polymerisation step may include a further component (d) including at least one crosslinkable moiety for crosslinking the poly(carborane-co-siloxane) chains. For example, component (d) may be a silane or siloxane having at least one crosslinkable moiety and at least two bonds selected from Si-OR° and Si-H, preferably at least two Si-OR° bonds or at least two Si-H bonds (where R° is as defined above).

Component (d) may be present at 1 mol% or less, 2 mol% or less, 3 mol% or less, 5 mol% or less, 8 mol% or less, 10 mol% or less, 15 mol% or less, 20 mol% or less, 30 mol% or less, 40 mol% or less, or 50 mol% or less (the mol % being based on the total amount of components (a), (b) and (d)).

When component (a) or (d) contain three or more bonds selected from Si-OR° and Si-H, these bonds themselves may serve as the crosslinkable moiety. For example, when component (a) or (d) contains three or more Si-OR° bonds, one of the Si-OR° groups may serve as the crosslinkable moiety. Similarly, when component (b) or (d) contains three or more Si-H bonds, one of the Si-H bonds may serve as a crosslinkable moiety. In such instances, crosslinking between polymer chains occurs during the polymerisation step, and is catalysed by the Lewis acid catalyst.

Preferably, the optional crosslinkable moiety in components (a), (b) and/or (d) are selected from -C2-10 alkenyl or -C2-10 alkynyl groups, more preferably -C2-10 alkenyl groups, most preferably vinyl groups. Advantageously, alkenyl and alkynyl groups are relatively unreactive during the polymerisation step, meaning that the groups remain in the

polymerisation product.

For example, component (d) may be dialkoxy(organo)vinyl silane or dihydroxy(organo)vinyl silanes, such as dialkoxy(alkyl)vinyl silanes, dihydroxy(alkyl)vinyl silanes, dialkoxy(aryl)vinyl silanes or dihydroxy(aryl)vinyl silanes. For example, component (d) may be

dimethoxy(methyl)vinyl silane or diethoxy(methyl)vinyl silane. Preferably, component (d) is a dimethoxy(organo)vinyl silane, most preferably dimethoxy(methyl)vinyl silane. Advantageously, using dimethoxy(methyl)vinyl silane as component (d) provides an excellent means of introducing vinyl groups into the polymer backbone, since coupling reactions involving the compound are fast and high yielding.

Preferably, the crosslinkable moieties do not react during the polymerisation step. In such instances, the method of producing a poly(carborane-co-siloxane) may involve the step of crosslinking the poly(carborane-co-siloxane) chains after the polymerisation step, i.e, the method involves at least two steps: (A) a polymerisation step as defined above followed by (B) a crosslinking step. Advantageously, having separate reaction and crosslinking steps means that the conditions of each of the steps can be tailored individually, which allows excellent control over the degree of crosslinking of the final product. This is in contrast to the conventional FeCb catalysed route to poly(carborane-co-siloxane)s where uncontrollable crosslinking takes place during the polymerisation step. In addition, having a separate reaction allows the linear polymer to be accumulated and stored before crosslinking, and makes it simpler to form shaped components (since polymerisation and crosslinking into a shaped component occur in separate steps).

During the crosslinking step the poly(carborane-co-siloxane) molecules may be crosslinked directly with one another via the crosslinkable moieties. Additionally or alternatively, the poly(carborane-co-siloxane) molecules may be crosslinked by a crosslinking component (e) which reacts with the crosslinkable moieties.

The crosslinking step (B) may involve crosslinking/vulcanising the poly(carborane-co- siloxane) product of step (A) using a crosslinking agent. Suitable crosslinking agents include, for example, peroxide-based crosslinking agents ("peroxide crosslinkers"). In such embodiments, the peroxide crosslinker decomposes to produce radicals which subsequently react with the poly(carborane-co-siloxane) to create polymer radicals. These polymer radicals then react with other polymer molecules to create crosslinked polymer chains. Suitable peroxide crosslinking agents include, for example, diacyl peroxides, peroxy esters, diaralkylperoxides, alkyl-arakyl peroxides and di-alkyl peroxides. The poly(styrene) equivalent weight average molecular weight (M _w ) of the poly(carborane-co-siloxane) before crosslinking may be, for example, 10 kDa or more, 20 kDa or more, 30 kDa or more, 40 kDa or more or 50 kDa or more.

In embodiments in which the crosslinkable moiety is a -C2-10 alkenyl group, the crosslinking step (B) may involve reacting the poly(carborane-co-siloxane) with an alkenyl-specific peroxide crosslinker. Suitable alkenyl-specific peroxide crosslinkers include, for example, diaralkyl peroxides, alkyl-arakyl peroxides and dialkyl peroxides. More specifically, the alkenyl-specific peroxide crosslinker may be, for example, dicumyl peroxide,

di(fert-butylperoxyisopropyl benzene), 2,5-di(fert-butylperoxy)-2,5-dimethylhexane, or di-ie t-butyl peroxide. Such reagents are particularly effective when the crosslinkable moiety is a vinyl group.

In embodiments in which the crosslinkable moiety is a -C _2- io alkenyl group, the crosslinking step (B) may involve reacting the poly(carborane-co-siloxane) product of step (A) with component (e): a siloxane crosslinker having at least two Si-H bonds, in the presence of a hydrosilation catalyst.

The reaction may occur by homogenous hydrosilation, non-aqueous heterogeneous hydrosilation, aqueous heterogenous hydrosilation or photochemical hydrosilation. Suitable hydrosilation catalysts include, but are not limited to, Karstedt's catalyst, Speier's catalyst (HzPtC ), chloroplatinic acid, Pt(COD) ₂ , Pt(PPh ₃ ) ₄ , PtCI ₂ (PPh ₃ )2, [Rh(COD)CI] ₂ ,

PtCI ₂ (PhCN) ₂ , PtCI ₂ (diop); PtCI ₂ (dppb), RhCI(PPh ₃ ) ₃ , Cp*Rh(C ₂ H ₄ )(SiR ₃ )H, RhCh-3H ₂ 0, Rh(PPh) ₃ CI, [Cp*Rh] ₂ CI ₄ , [Cp*Rh ₂ ]CI ₃ H, {[Cp ^* Rh] ₂ (OH) ₃ }*, Me ₂ SiCp* ₂ Th(n-Bu) ₂ , Pt(acac) ₂₎ and Fe(CO) ₃ , or combinations of the above.

Additionally or alternatively, the crosslinkable moieties may be introduced to the

poly(carborane-co-siloxane) after the polymerisation step and before the crosslinking step. In such instances; the method or producing a poly(carborane-co-siloxane) may involve (A) a polymerisation step as defined above followed by, (Α') a functionalisation step, in which crosslinkable moieties are added to the product of step (A), followed by (B) a crosslinking step. Advantageously, introducing the crosslinkable moiety after the polymerisation reaction limits or prevents crosslinking during the polymerisation step, which allows excellent control over the degree of crosslinking of the final product.

In such embodiments, the reacting components may contain a leaving group to facilitate introduction of a crosslinkable moiety after the polymerisation step.

For example, component (a) may be substituted with at least one -CMO haloalkyl group, which can subsequently undergo reaction to replace a halogen atom with a crosslinkable moiety (e.g., conversion to a methacrylate or acrylate group via reactions analogous to those described in [Wakita] , [Altmann], [Ervithayasuporn], and [Li]). For example, at least one of -R , -R ¹ , -R ² or -R ^2* may be a -Ci-io haloalkyl group, such as chloropropyl. Additionally or alternatively, component (b), or a separate component may be substituted with at least one -Ci-io haloalkyl group.

Carboranyl group

The term "carboranyl", as used herein pertains to a group consisting of a cluster of formula -CBgQhC- wherein g is an integer of from 3 to 16, h is an integer of from 3 to 16 and Q is -H, a -Ci-6 alkyl or a halogen group. Suitably, h = g + g*, where g* = 0 to 4, e.g., g* = 0 to 2. Optionally, h - g.

For example, the carboranyl group may be 1 ,2-dicarba-c/oso-dodecaboranyl (1 ,2-C ₂ BioHio, o-carborane), 1 ,7-dicarba-c/oso-dodecaboranyl (1 ,7-C ₂ BioHio, m-carborane), 1 ,12-dicarba- c/oso-dodecaboranyl (1 ,12-C ₂ BioHio, p-carborane), 1 ,10-dicarbadecaboranyl (1 ,10-C ₂ B ₈ H ₈ ); 1 ,6-dicarbadecaboranyl (1 ,6-C ₂ B ₈ H ₈ ); 2,4-dicarbaheptaboranyl (2,4-C ₂ B ₅ H5); 1 ,6- dicarbahexaboranyl (1 ,6-C ₂ B ₄ H ₄ ); 9-alkyl-1 ,7-dicarba-c/oso-dodecaboranyl

(9-alkyl-1 ,7-C ₂ BioH ₉ ); 9,10-dialkyl-1 ,7-dicarba-c/oso-dodecaboranyl

(9,10-dialkyl-1 ,7-C ₂ Bi ₀ H ₈ ); 2-alkyl-1 ,10-dicarbadecaboranyl (2-alkyl-1 ,10-C ₂ B ₈ H ₇ );

8-alkyl-1 ,6-dicarbadecaboranyl (8-alkyl-1 ,6-C ₂ B ₈ H ₇ ); decachloro-1 ,7-dicarba-c/oso- dodecaboranyl (1 ,7-C ₂ BioClio); octachloro-1 ,10-dicarbadecaboranyl (1 ,10-C ₂ B ₈ CI ₈ ); decafluoro-1 ,7-dicarba-c/oso-dodecaboranyl (1 ,7-C2Bi ₀ Fi ₀ ); or octafluoro-1 ,10- dicarbadecaboranyl (1 ,10-C2B ₈ F ₈ ).

The divalent carboranyl group is preferably a closo carborane group, most preferably a dicarbadodecaborane group, i.e., a group having the formula -CB10H10C-.

Dicarbadodecaborane groups are particularly useful due to their high boron content, chemical stability, and their ability to confer radiation and thermal stabilities to the poly(carborane-co-siloxane).

Dicarbadodecaborane exists in three structural isomers, designated as ortho-, meta- and para- (abbreviated by 0-, m- and p- prefixes hereinafter), with the prefix describing the relative positions of the carbon atoms in the cage. These structures are shown in the figure below, where the unmarked vertices indicate B-H groups.

Ort/?o-dicarbadodecaborane /nefa-dicarbadodecaborane para-dicarbadodecaborane (1 ,2-dicarbadodecaborane) (1 ,7-dicarbadodecaborane) (1 ,12-dicarbadodecaborane)

Preferably the divalent carboranyl group (e.g., X and Y as defined above) is

1 ,7-dicarbadodecaborane (m-dicarbadodecaborane) or 1 ,12-dicarbadodecaborane

(p-dicarbadodecaborane).

Isotopes

The constituent atoms of the starting materials used in the present invention may be in any isotopic form. For example, H may be in any isotopic form, including ¹ H, ² H (D), and ³ H (T); C may be in any isotopic form, including ¹² C, ¹³ C, and ⁴ C; O may be in any isotopic form, including ¹⁶ 0 and ¹⁸ 0; Si may be in any isotopic form, including ²⁸ Si, ²⁹ Si, and ³⁰ Si; and the like.

The relative amounts of ¹⁰ B and ¹¹ B boron isotopes in the carboranyl groups of the starting materials may reflect the naturally occurring relative abundance of these isotopes (~20% ¹⁰ B and -80% ¹¹ B). However, in certain applications it may be advantageous for the carboranyl groups to be enriched in ¹⁰ B or ¹¹ B.

For example, the amount of ⁰ B atoms as a percentage of the total boron atoms of the starting materials used in the present invention may be greater than 20%, greater than 30%, greater than 40%, greater than 50%, greater than, 75%, greater than 90% or greater than 95%. Advantageously, increasing the ¹⁰ B content of the starting materials allows the production of poly(carborane-co-siloxane)s with improved ability to capture neutrons, without negatively impacting other properties of the material. This makes the materials particularly useful for neutron shielding and neutron capture applications.

Alternatively, the amount of ¹¹ B atoms as a percentage of the total boron atoms of the starting materials used in the present invention may be greater than 80%, greater than 90% or greater than 95%. Increasing the ¹ B content of the starting materials allows the production of poly(carborane-co-siloxane)s which capture fewer neutrons, without negatively impacting other properties of the material.

Solvent

The polymerisation step may be carried out in a hydrocarbon solvent, e.g., pentane, hexane, xylene, benzene, or toluene. Preferably the polymerisation step is carried out in hexane or xylene, most preferably toluene.

Further aspects

In a second aspect, the present invention provides a poly(carborane-co-siloxane) obtainable by a method of the first aspect.

Preferably, the poly(carborane-co-siloxane) is elastomeric.

The lower limit for the poly(styrene) equivalent weight average molecular weight (M _w ) of the poly(carborane-co-siloxane) may be, for example, 1 kDa, 5 kDa, 10 kDa, 20 kDa, 30 MDa or 50 MDa. The upper limit for the poly(styrene) equivalent M _w of the

poly(carborane-co-siloxane) may be, for example, 60 kDa, 100 kDa, 200 kDa, 500 kDa, 1 MDa or 10 MDa. For example, the poly(styrene) equivalent M _w of the

poly(carborane-co-siloxane) may be 20 kDa or more, 20 kDa to 100 kDa, or 20 kDa to 60 kDa.

The poly(styrene) equivalent M _w may be determined by gel permeation chromatography following the procedures outlined in ISO 11344 (entitled "Rubber, raw synthetic - determination of the molecular-mass distribution of solution polymers by gel permeation chromatography"). For example, M _w may be determined according to the following protocol: gel permeation chromatography (GPC) is performed on Viscotek TDA or Viscotek 250 instruments at an elution flow rate of 1 mlJmin using two Polymer Labs PLgel columns (Mixed D) in series with THF solvent and heated at 35°C. Known solution concentrations (~ 2 mg mL ) of the poly(carborane-co-siloxane) are prepared in THF and allowed to stand until full dissolution of the samples prior to analysis. Data is analysed using OmniSEC software. Molecular weight data is given as poly(styrene) equivalents against low molecular weight narrow poly(styrene) standards used as calibrants

As noted above, the conventional FeC catalysed method for producing

poly(carborane-co-siloxane)s results in crosslinked polymers which contain residual iron- containing impurities trapped within the crosslinked polymer network. Furthermore, removal of these impurities is complicated by the fact that the crosslinked polymers are generally insoluble in solvents such as toluene. Even for those polymers which do dissolve, removal of the impurities is still problematic. Thus, the polymers produced in the conventional method have a brown/black discolouration arising from the iron-containing impurities, which reduces the range of applications which the polymers can be used for. For example, the polymers cannot be used in applications which require clear, transparent or translucent materials, such as coatings for windows or fabrics. Furthermore, iron (probably in the form of iron oxide) is known to accelerate ageing of polymeric matrices, so the presence of iron- containing impurities is undesirable for this reason as well. The FeCI ₃ catalyst itself is corrosive and reacts with moisture to form HCI which is hazardous and detrimental to siloxane-based polymers.

Advantageously, the methods of the present invention provide a general route to polymers which do not contain any iron-containing impurities. Therefore, suitably, the

poly(carborane-cosiloxane) is substantially free of iron, in particular, free of FeCI ₃ .

Preferably, the poly(carborane-co-siloxane) is substantially free of residual catalyst.

By "substantially free" we mean that the total amount of the relevant impurity or impurities is 0.5 wt.% or less, 0.1 wt.% or less, 0.05 wt.% or less, 0.01 wt.% or less (the weight percent being relative to the total weight of poly(carborane-co-siloxane)). Preferably, there is no detectable quantity of the relevant impurity or impurities. The amount of iron in the sample can be determined using inductively coupled plasma atomic emission spectroscopy (ICP- AES). For example, the polymer being tested may be digested (heated in acid) and the quantity of its constituent elements measured using ICP-AES against an appropriate standard.

Preferably, the poly(ca formula (I):

wherein:

-X- is a carboranyl group; each -R ⁶ is independently -R ^s each -R ^s is independently -R ^c or -N=C=0; each -R ^c is -R ^CC1 , -R ^CC2 , or -L ^C -R ^CC2 ; each -R ^CC1 is an optionally substituted -CMO alkyl, -Ci-io haloalkyl, -C2-io alkenyl, or -C2-10 alkynyl group? each -R ^CC2 is an optionally substituted -C3-8 cycloalkyl, -C3-8 heterocycloalkyl, -Cs-10 aryl, -C _6- io haloaryl or -C _5- io heteroaryl, group; each -L ^c - is alkylene; each -LA is a crosslinking moiety, crosslinking recurring units of formula (I); and i is > 3.

Such compounds encompass poly(carborane-co-organotetrasiloxane)s, poly(carborane-co- organopentasiloxane)s, poly(carborane-coorganohexasiloxane)s etc. and higher siloxane copolymers, and derivatives of such compounds. These polymers are analogous to conventional Dexsil® polymers, but unlike Dexsil® polymers the method of the present invention allows the polymer to contain precisely repeating units of Formula (I), that is, polymers where "i" is the same for all or substantially all recurring units in the polymer.

By "substantially all recurring units in the polymer" we mean, for example, at least 80%, at least 85%, at least 90%, at least 95%, at least 97%, at least 98%, or at least 99% of the units in the polymer have the same value for "i" (as a percentage of the total number of units of formula (I) in the polymer). The relative amount of different recurring units can be determined by, for example, NMR spectroscopy, such as ¹ H and ²⁹ Si-NMR spectroscopy. An example of a suitable method for analysing the structure of poly(carborane-co-siloxane)s by NMR is provided in [Kahlig].

The upper limit for the integer i may be, for example, 100, 50, 20, 15, 10, 9, 8, 7, 6, 5, 4, or 3. The lower limit for the integer i may be, for example, 4 or 5. For example, the integer i may be 3-20, 3-10, or 3-5.

Suitably, all -R ⁶ groups may be -R ^s , in which case the polymer is a linear polymer with no crosslinking between polymer chains. Advantageously, the present invention provides a general route to a range of linear poly(carborane-co-siloxane)s with precisely repeating recurring units, which are not accessible using conventional synthesis methods.

In instances where a component (d) is included, the polymer will not necessarily contain precisely repeating recurring units. Therefore, in instances where the polymer contains crosslinked recurring units (i.e., -LA groups are present), the method of the present invention allows the production of a polymer network where all or substantially all recurring units which do not contain -LA groups have the same value for "i".

At least some (i.e., >0%), optionally all, of the -R ⁶ groups may be -CLIO alkenyl groups. The upper limit for the amount of -R ⁶ groups which are -C1-10 alkenyl groups, as a percentage of the total number of -R ⁶ groups may be, for example, 90%, 80%, 70%, 60%, 50%, 40%, 30%, 25%, 20%, 15%, 10%, 8%, or 5%. The lower limit may be, for example, 0%, 0.1 %, 0.5%, 1 %, 2%, 3%, 4%, 5%, 10%, 15% or 20%. For example, example, the total amount of -R ⁶ groups which are -CMO alkenyl groups may be 0-50%, or 5-25%. Preferred combinations

In preferred embodiments, the method of producing a poly(carborane-co-siloxane) comprises a polymerisation step in which:

(a) a siloxane compound having the following formula:

is reacted with organohydrosiloxane compound of the following formula:

,5*

R-— Si- -0— Si 0— Si- Si— 0- Si— O- -Si— R

5*

R" R"

in the presence of

(c) a Lewis acid catalyst having the formula (R ^A ) _a B(R ^B ) _b , wherein X, Y, R\ R ^1* , R ² , R ^2* , R ³ , R ³ \ R ⁴ , R ^4* , R ⁵ , R ^5* , R ^A , R ^B , p, q, r, s, t, u, a and b are as defined above.

In such embodiments, it is preferred that component (a) has the following formula:

R'

HO— SH-0— Si Si— CM— Si— O H q

wherein X is a carboranyl group, most preferably -C2B10H10-, and R ¹ , R ^1* , R ² , R ^2* , p and q are as defined above, and component (b) has the following formula:

In addition, it is preferred that the Lewis acid catalyst is B(R ^B >3, where each -R ^B is a Ce-ιο aryl (preferably phenyl) substituted with 3, 4 or 5 halo atoms. Most preferably, the catalyst is B(C ₆ F ₅ ) ₃ .

Advantageously, such methods provide a fast, reliable general polymerisation reaction for producing high-molecular weight poly(carborane-co-siloxane) polymers. In contrast to the conventional FeCI ₃ -catalysed reaction route, the above preferred embodiments start from hydroxyl-terminated siloxanes instead of methoxy-terminated siloxanes. The present inventors have found that hydroxyl-terminated siloxanes are excellent starting materials for producing poly(carborane-co-siloxane)s, resulting in high yields over a short time period, which is surprising since hydroxyl-terminated siloxanes have low solubility in the

hydrocarbon solvents typically used in the reactions (lower solubility than methoxy- terminated siloxanes). in addition, hydroxyl-terminated siloxanes are more readily purified to remove acidic- impurities, which result from their method of production, than methoxy-terminated siloxanes. Thus, commercially available hydroxyl-terminated siloxanes typically do not contain residual acid, in contrast to commerciall methoxy-terminated siloxanes which typically contain residual HCI (in this regard, the present inventor found that commercially available methoxy- terminated siloxanes used in this work evolved acidic vapour over time). Thus, it is possible to make polymers from hydroxyl-terminated siloxanes which have low or no acid content, which improves the stability of the polymers, since the presence of acid can degrade polymer performance over time.

Further options and preferences

The invention may have any one or more of the following, generally combinable, optional or preferable features.

=B£ ^■ ■ ^'

Preferably, -R° is -R ⁰⁰¹ : -R° ⁰¹

-R°° ¹ may be optionally substituted with one or more groups -R ^M1 .

.R001 _ma y _{De A} .Q^ alkyl, -C _2- 6 alkenyl, or -C2-6 alkynyl group, said group being optionally substituted with one or more groups -R ^M1 .

-R°° ¹ may be a -C1- ₁ 0 alkyl, said group being optionally substituted with one or more groups

-RM1. -R ⁰⁰ may be a -Ci-e alkyl, said group being optionally substituted with one or more groups -R ^M .

Preferably, -R° ⁰¹ is methyl, ethyl, /-propyl, n-propyl, /-butyl, s-butyl, n-butyl, or f-butyl.

More preferably, -R°° ¹ is methyl, ethyl, /-propyl, n-propyl, /-butyl, s-butyl, or n-butyl.

More preferably still, -R ⁰⁰ is methyl, ethyl, /-propyl, or n-propyl.

More preferably still, -R° ⁰¹ is methyl or ethyl.

Most preferably, -R ⁰⁰¹ is methyl.

-R°° ²

.R00 ² _may b _e optionally substituted with one or more groups -R ^M2 .

Preferably, -R°° ² is a phenyl group optionally substituted with one or more groups -R ^M2 .

Preferably, -L°-R° ⁰² , if present, is -CH ₂ -phenyl optionally substituted with one or more groups -R ^M2 .

-R ^M1 is halo, -OH, -OR", -CF ₃ , -OCF ₃ , -C(0)OH, or -C(0)OR". Preferably, -R ^M1 is halo, -OH, -OR", -CF ₃ , or -OCF ₃ . More preferably, -R ¹ is halo or -CF ₃ .

-R ^M2

-R ^M2 is C1-4 alkyl, -R ^M1 , or -L ^M -R ^M1 .

Optionally, -R ^M2 is methyl, ethyl, /-propyl, or n-propyl, -R ^M1 , or -L ^M -R ^M1 . Optionally ,-R ^M2 is methyl, ethyl, halo or CF ₃ . Optionally,-R ^M2 is methyl.

_|_M

-L - is C1-4 alkylene, for example, -CH ₂ -, -C ₂ H ₄ - or -C ₃ H ₆ -. M -

Preferably, -R" is a -Ci-e alkyl group.

For example, -R" may be methyl, ethyl, /-propyl, n-propyl, /-butyl, s-butyl, n-butyl, or f-butyl. More preferably, -R" is methyl, ethyl, /-propyl, or n-propyl. More preferably still, -R ^" is methyl or ethyl. Most preferably, -R" is methyl.

Preferably, each -R ^s is independently -R ^c .

Preferably, each -R ^CC1 is a -Ci-i ₀ alkyl, -C2-io alkenyl, or -C _2- io alkynyl group, said group being optionally substituted with one or more groups -R ^M3 .

More preferably, each -R ^CC1 is a -Ci-6 alkyl, -C2-6 alkenyl, or -C2-6 alkynyl group, said group being optionally substituted with one or more groups -R ^M3 .

Preferably, each -R ^CC2 is a -C^e cycloalkyi, or -Ce-io aryl group, said groups being optionally substituted with one or more groups R ^M4 .

More preferably, each -R ^CC2 is a -C _6- io aryl group (e.g., phenyl), said groups being optionally substituted with one or more groups R ⁴ .

Preferably, -L ^C -R ^CC2 , if present, is -CH ₂ -phenyl optionally substituted with one or more groups -R ⁴ .

-L ^c -

-L ^c - may be, for example, -CH2-, -C2H4-, or -C3H6-. -R ^M3

-R ^M3 is halo, -OH, -OR", -CF ₃ , -OCF3, -C(0)OH, -C(0)OR", -N ₃ or -CN. Preferably, -R" ³ is halo, -OH, -OR", -CF ₃ , -OCF3 or -N ₃ . More preferably, -R ^M3 is halo or -CF ₃ .

-R ^M4 is Ci-4 alkyl, -R ^M3 , or— L ^M -R ^M3 .

Optionally, -R ^M4 is methyl, ethyl, /-propyl, or n-propyl, -R ^M3 , or -L ^M -R ^M3 . Optionally, -R ^M4 is methyl, ethyl, halo or CF ₃ . Optionally, -R ⁴ is methyl. Definitions

The term "C _x-y alkyl", as used herein pertains to a monovalent moiety obtained by removing a hydrogen atom from a carbon atom of an unsaturated hydrocarbon compound having from x to y carbon atoms. Unless specified otherwise, the term covers both linear and branched hydrocarbon chains.

The term "C _x-y alkylene", as used herein, pertains to a bidentate moiety obtained by removing two hydrogen atoms, either both from the same carbon atom, or one from each of two different carbon atoms, or a saturated hydrocarbon compound having from x to y carbon atoms. Unless specified otherwise, the term covers both linear and branched hydrocarbon chains.

The term "C _x-y cycloalkyl" as used herein pertains to an alkyl group which is also a cyclyl group; that is, a monovalent moiety obtained by removing a hydrogen atom from an alicyclic ring atom of a cyclic hydrocarbon (carbocyclic) compound, which moiety has from x to y carbon atoms, including from x to y ring atoms.

The term "C _x-y heterocyclyl" as used herein pertains to a monovalent moiety obtained by removing a hydrogen atom from a ring atom of a heterocylic compound, which moiety has from x to y ring atoms.

The term "C _x-y aryl" as used herein pertains to a monovalent moiety obtained by removing a hydrogen atom from an aromatic ring atom of an aromatic compound, which moiety has from x to y ring atoms.

The term "C _x-y heteroaryl" as used herein pertains to a monovalent moiety obtained by removing a hydrogen atom from an aromatic ring atom of an heteroaromatic compound, which moiety has from x to y ring atoms.

For the avoidance of doubt, the index "C _x-y " in terms such as "Cs-ioheteroaryl",

"C3-7heterocyclyl", and the like, refers to the number of ring atoms, which may be carbon atoms or heteroatoms (e.g., N, O, S). For example, pyridyl is an example of a Ceheteroaryl group, and piperidino is an example of a Ceheterocyclyl group.

The term "heteroaryl" refers to a group that is attached to the rest of the molecule by an atom that is part of an aromatic ring, wherein the aromatic ring is part of an aromatic ring system, and the aromatic ring system has one or more heteroatoms (e.g., N, O, S). For example, pyridyl is an example of a Ceheteroaryl group, and quinolyl is an example of a Cioheteroaryl group.

The term "heterocyclyl" refers to a group that is attached to the rest of the molecule by a ring atom that is not part of an aromatic ring {i.e., the ring is partially or fully saturated), and the ring contains one or more heteroatoms (e.g., N, O, S). For example, piperidino is an example of a Ceheterocyclyl group. BRIEF DESCRIPTION OF THE FIGURES

Figure 1 shows gel permeation chromatography (GPC) data obtained under differential viscosity detection for two poly(m-carborane-co-1 ,1 ,3,3,5,5,7,7-octamethyltetrasiloxane)s produced according to the method of the present invention (Samples B300-1 and B300-2);

Figure 2 shows GPC data obtained under refractive index detection of vinyl-pendent poly(carborane-co-siloxane)s prepared using Me(vinyl)Si(OEt) ₂ according to the present invention;

Figure 3 shows GPC data obtained under refractive index detection of vinyl-pendent poly(carborane-co-siloxane)s prepared using Me(vinyl)Si(OMe) ₂ according to the present invention;

Figure 4 shows dynamic mechanical analysis plots (only the fifth compression cycle, for clarity) for elastomeric pads B400C-1 and B500C-1

Figure 5 shows different N.M.R. shifts for H, C and Si atoms in different environments in conventional Dexsil®-type poly(carborane-co-siloxane)s;

Figure 6 shows GPC data obtained under refractive index detection for Dexsil® 200 polymers produced according to the conventional FeC -catalysed method;

Figure 7 shows differential scanning calorimetry data for Dexsil® 200 prepared using two heating stages (Sample F200-1) and three heating stages (Sample F200-3);

EXAMPLES

The present invention will now be discussed with reference to the following non-limiting examples.

Materials

1 ,7-Bis(dimethylsilane)-m-carborane was synthesised and purified by flash column chromatography before use. All other carboranyl precursors and Dexsil-200 were procured from KatChem. Poly(siloxane)s including DMS-S35 were supplied by Gelest/ABCR chemical companies. B(CeF ₅ )3 was supplied by the Aldrich Chemical Company. Solvents were standard laboratory grade. Alt reagents were used without further purification.

Characterisation

Infra-red spectra were recorded on a Nicolet Avatar 360 FT-IR E.S.P spectrophotometer using the SMART Golden Gate accessory for attenuated total reflectance (ATR) for liquids and solid samples. Data were collected and analysed using OMNIC software.

¹ H-, ³ C-, ¹ B- and ²⁹ Si-NMR spectra were recorded using a Varian VNMRS-400

spectrometer operating under 9.4 T magnetic field (400 MHz ¹ H frequency) with a dual broadband multinuclear probe. Spectra were recorded at 298 K as dilute solutions in deuterated solvent as specified. Data were analysed using Varian VNMRJ version 2.3a software. Data are expressed as chemical shifts (δ) in parts per million (p _; p.m.) on the unified scale for all nuclei relative to the ¹ H resonance of tetramethylsilane in dilute solution in chloroform [Harris(a), Harris(b)]. ¹ H-NMR were not routinely run boron-decoupled.

1 ³ C-NMR were acquired using an Attached Proton Test (APT) program where specified. 1B-NMR were routinely run proton-decoupled. ²⁹ Si-NMR spectra were acquired using an Insensitive Nuclei Enhanced by Polarisation Transfer (INEPT) pulse sequence where specified. The multiplicity of each signal is designated by the following abbreviations: s, singlet; b, broad; m, multiplet.

Differential scanning calorimetry (DSC) was performed on a Perkin Elmer Diamond

DSC instrument following the standard procedure and the routine: i) sample (in aluminium pan) cooled to -160 °C and held isothermally at -160 °C for 1 min.;

ii) heating from -160 °C to +100 °C at 100 °C min ^"1 ;

iii) sample held isothermally at +100 °C for 0.5 min.;

iv) cooling from +100 °C to -160 °C at 100 °C min- ¹ ;

v) sample held isothermally at -160 C for 1 min.;

vi) repeat from (ii) four times.

The first run was discarded and repeat runs showed identical traces in each case, to ensure results were not artefacts from unknown thermal history. Data were collected and analysed using Perkin Elmer Pyris software.

Dynamic mechanical analysis (DMA) was performed using a TA Instruments Q800 DMA equipped with the compression clamp arrangement using the procedure: i) equilibrate temperature at 35 °C, preload of 0.05 N;

ii) Ramp compression force from 0.05 N to 18.0 N at 3 N min ^-1 ;

iii) Ramp compression force from 18.0 N to 0.05 N at 3 N min ^-1 ;

iv) Repeat step (ii) five times.

Data were collected on all runs, with data presented herein from the fifth compression cycle after repeated conditioning cycles in order to remove Mullins' effects and achieve an equilibrium stress-strain curve [Harwood, Mullins]. Poisson's ratio for the material has not been measured separately, and a default value of 0.5, as generally accepted for elastomeric materials, has been used in the calculation of the mechanical response. Data were collected and analysed using TA Instruments Universal Analysis software. Compressive Young's moduli were calculated from gradient of the stress-strain profile at the end of the

compression part of the cycle.

Gel permeation chromatography (GPC) was performed on Viscotek TDA or Viscotek 250 instruments using two Polymer Labs PLgel columns (preferably Mixed D) in series with THF solvent and heated at 35 °C. Known solution concentrations (~ 2 mg mL ¹ ) were prepared in THF and allowed to stand until full dissolution of the samples prior to analysis. Data was analysed using OmniSEC software. Molecular weight data is given as poly(styrene) equivalents against low molecular weight narrow poly(styrene) standards used as calibrants Centrifugal mixing was performed in a Siemens Speedmixer™ DAC 150 FVZ-K at

3,540 r.p.m. for the times specified herein.

1 ,7-Bis(methoxydimethylsilyl)-m-dicarbadodecaborane 1 (1.08 g, 3.37 mmol, 1.0 eq.) and 1 ,1 ,3,3-tetramethyldisiloxane 2 (0.45 g, 0.60 ml_, 3.37 mmol, 1.0 eq.) were dissolved in hexane (2 mL) and stirred at room temperature in air, without any special precaution to exclude moisture. Tris(pentafluorophenyl)borane (20 μΙ_ of a solution of catalyst (10.5 mg) in CH2CI2 (500 pL)) was added in one portion and the reaction mixture exhibited gradual gas evolution which increased in vigour before a very vigorous gas expulsion and exotherm after 5 minutes following catalyst addition; the reaction mixture continued to evolve gas for some time.

After complete reaction (2.5 hours) Al ₂ 0 ₃ (Brockmann, neutral, 500 mg) was added to the reaction mixture to sequester catalyst before filtration through filter paper under gravity to yield 300 (1 ,50 g, quant.) as a clear, colourless/translucent viscous oil. The product is referred to herein as "Sample Β300- .

Product characteristics: u _max /cm- ¹ (ATR) 2961 (u(C-H)), 2592 (υ(Β-Η)), 1258, 1031 , 856, 791 ; δ _Η (400 MHz, CDCU) 3.2 - 1 :5 (10H, bm, 10 * H), 0.19 (12H, 2 χ Si(CH ₃ ) ₂ ), 0.08 (12H, 2 * Si(CAV ₃ ) ₂ ); 5c (400 MHz, CDCI3) 68.4 (C), 1.0 (2 * Si(CH ₃ ) ₂ ) 0.5 (2 * Si(CH ₃ ) ₂ ); δ _Β (400 MHz, CDCI ₃ ) ; -2.58 (2B, bs), -7.63 (2B, bs), -9.74 (4B, bs), -13.78 (2B, bs); 6 _Si (400 MHz, INEPT, CDCI ₃ ) -0.93 (2 x S/(CH ₃ )2), -20.49 (2 S/(CH ₃ ) ₂ ).

1 ,7-Bis(hydroxydimethylsilyl)-m-dicarbadodecaborane 3 (447 mg, 1.53 mmol, 1.0 eq.) and 1 ,1 ,3,3-tetramethyldisiloxane 2 (0.21 g, 0.27 mL, 1.53 mmol, 1.0 eq.) were taken up in hexane (3 mL) to afford a reaction mixture with a suspension of undissolved carborane component. The reaction mixture was stirred at room temperature in air, without any special precaution to exclude moisture. Tris(pentafluorophenyl)borane (20 pL of a solution of catalyst (10.5 mg) in CH2CI2 (500 pL)) was added in one portion and the reaction mixture immediately and vigorously evolved gas. Over five minutes the reaction mixture became a clear colourless solution after complete dissolution of the carborane reagent.

After 2.5 hours Al ₂ 0 ₃ (Brockmann, neutral, 250 mg) was added to the slightly translucent reaction mixture to sequester catalyst before filtration through filter paper under gravity to yield 300 (0.63 g, 97%) as a clear, colourless/translucent viscous oil. The product is referred to herein as "Sample B300-2".

Product characteristics: Dmax/crrr ¹ (ATR) 2961 (u(C-H)), 2592 (υ(Β-Η)), 1258, 1031 , 856, 790; δ _Η (400 MHz, CDCI ₃ ) 3.2 - 1.5 (10H, bm, 10 BH), 0.19 (12H, 2 ^χ Si(CH ₃ ) ₂ ), 0.08 (12H, 2 x Si(CH ₃ ) ₂ ); δ ₀ (400 MHz, CDCI3) 68.4 (C), 1.0 (2 Si(CH ₃ ) ₂ ) 0.5 (2 Si(CH ₃ ) ₂ ); δ _Β (400 MHz, CDCI ₃ ) -2.75 (2B, bs), -7.68 (2B, bs), -9.76 (4B, bs), -13.88 (2B, bs); 5 _Si (400 MHz, INEPT CDCI ₃ ) -0.93 (2 x S/(CH ₃ ) ₂ ), -20.49 (2 S/ ^' (CH ₃ ) ₂ ); T _g (by DSC at scan rate of 100 °C min ¹ ) at onset = -69.3°C, at half Cp = -66.8°C (Literature value for Dexsil® 300 = -68°C [Heying]); ACp (by DSC) = 0.18 J g ^_1 K~ ¹ .

Example 3 - Polv(fn-dicarbadodecaborane-co-1.1.3.3.5.5.7.7-octamethyltet rasiloxane) (300) Poly(m-dicarbadodecaborane-co-1 ,1 ,3,3,5,5,7,7-octamethyltetrasiloxane) was produced in a similar fashion to Example 2, but the reaction was conducted in toluene.

1 ,7-Bis(hydroxydimethylsiiyl)-meia-dicarbadodecaborane 3 (5.01 g, 17.1 mmol, 1.0 eq.) and 1 ,1 ,3,3-tetramethyldisiloxane 2 (2.30 g, 3.03 ml_, 17.1 mmol, 1.0 eq.) were taken up in toluene (38.9 mL) to afford a reaction mixture with a suspension of undissolved carborane component (0.44 M concentration of disilane). The reaction mixture was stirred at room temperature, without any special precaution to exclude moisture or air.

Tris(pentafluorophenyl)borane (50 μΙ_ of a solution of catalyst (30 mg) in toluene (300 μΐ.) was added in one portion. After an induction period the reaction mixture evolved gas, with gas evolution becoming more vigorous and subsiding to afford a clearer colourless solution after complete dissolution of the carborane reagent.

After 20 minutes Al ₂ 0 ₃ (Brockmann, neutral, ~ 0.5 g) was added and the reaction mixture stirred at room temperature for 20 minutes before filtration through filter paper under gravity and removal of the solvent by evaporation under reduced pressure to afford the product polymer 300 as a clear/opaque colourless viscous oil (6.66 g, 91 %). The product is referred to herein as "Sample B300-4".

The experiment was repeated, but with the initial concentration of disilane adjusted to be 0.175 M or 0.70 M, to give products referred to herein as "Sample B300-3" and "Sample B300-5" respectively.

Example 4 - Polv(An-dicarbadodecaborane-co-1.1.3.3.5.5.7.7.9.9-decamethy lpentasiloxane (400)

The same general procedure as Example 3 was followed, except that 1 equivalent of 1 ,1 ,3,3,5,5-hexamethyltrisilane S was used instead of 1 ,1 ,3,3-tetramethyldisiioxane 2. The reaction was conducted at three different initial concentrations (0.175 M, 0.44M and 0.70 M) to give products referred to herein as "Sample B400-1", "Sample B400-2 ", and "Sample B400-3".

The same general procedure as Example 3 was followed, except that 1 equivalent of 1 ,1 ,3,3,5,5-hexamethyltrisilane 6 was used instead of 1 ,1 ,3,3-tetramethyldisiloxane 2. The reaction was conducted at three different initial concentrations (0.175 M, 0.44M and 0.70 M) to give products referred to herein as "Sample B500-1", "Sample B500-2 ", and "Sample B500-3".

Example 6 - Polv(m-dicarbadodecaborane-co-1.1.3.3.5.5-hexamethyltrisilox ane) (200)

H

1 ,7-Bis(hydroxyl(dimethyl)silyl)-meia-dicarbadodecaborane 3 (0.490 g, 1.71 mmol, 1 eq.) was crushed to a fine powder and transferred to a dry round bottom flask with stirring flea. The reagent was taken up in THF (10 mL, dry) to afford an opaque solution/suspension. Chlorodimethylsilane 7 (0.57 mL, 5.12 mmol, 3 eq., excess) was added in one portion to the stirring reaction mixture at room temperature, and the reaction followed by thin layer chromatography.

After three hours no starting material 3 was evident and the reaction was quenched by addition of water (10 ml_). The reaction mixture was partitioned by addition of brine (10 mL) and the aqueous layer was extracted with CHCI ₃ (3 x 10 mL). The combined organic layers were dried (MgS0 ₄ ) and the resulting clear colourless solution was concentrated by evaporation under reduced pressure to afford a clear colourless oil (534 mg) that contained the required product and the half-reacted intermediate. The crude oil was purified by flash column chromatography over silica (hexane:EtOAc 9:1 ) affording 8 (0.289 g, 41 %) as a clear colourless oil: Rf = 0.68 (hexane:EtOAc 9:1); δ _Η (400 MHz, CDCI ₃ ) 4.71 (2H, sept, J 2.8, 2 χ S\H), 0.21 (12H, d, J 2.8, 2 Si(CH ₃ ) ₂ H), 0.20 (12H, s, 2 Si(CH ₃ ) ₂ ); δο (400 MHz, CDCI ₃ ) 68.5 (C) 0.6 ((CH ₃ ) ₂ ), 0.3 ((CH ₃ ) ₂ ); δ _Β (400 MHz, CDCI ₃ ) -2.79 (2B), -7.69 (2B), -9.80 (4B), -13.93 (2B); 5 _Si (400 MHz, CDCI ₃ ) 0.89 (2 χ S/(CH ₃ ) ₂ ), -4.23 (2 S;(CH ₃ ) ₂ H).

Further elution afforded the half-reacted intermediate 1-(1 ,1 ,3,3-tetramethyldisiloxanyl)-7- (hydroxyl(dimethyl)silyl)-mefa-dicarbadodecaborane 9: ■ ^>

9

Product characteristics of 9: Rf = 0.21 (hexane:EtOAc 9:1); δ _Η (400 MHz, CDCI ₃ ) 4.70 (1 H, sept, J 2.8, SiH), 0.27 (6H, s, Si(CH ₃ ) ₂ OH), 0.21 (6H, d, J 2.8, Si(CH ₃ ) ₂ H), 0.20 (6H, s, Si(CH ₃ ) ₂ ); 6c (400 MHz, CDCI ₃ ) 68.8 (C), 68:1 (C), 0.6 ((CH ₃ ) ₂ ), 0.3 ((CH ₃ ) ₂ ), -0.4

(Si(CH ₃ ) ₂ OH); δ _Β (400 MHz, CDCI ₃ ) -2.74 (2B), -7.55 (2B), -9.76 (4B), -14.03 (2B); 6 _Si (400 MHz, INEPT, CDCIs) 9.60 (S;(CH ₃ ) ₂ OH), 0.88 (S/(CH ₃ ) ₂ ), -4.07 (S/(CH ₃ ) ₂ H).

B(C ₆ F ₅ ) ₃ (5 pL of a toluene solution of B(C ₆ F ₅ ) ₃ (30 mg) in toluene (300 μΙ_), i.e. 0.5 mg catalyst) was added in one portion to a solution/suspension of 1 ,7-bis(1 ,1 ,3,3- tetramethyldisiloxanyl)-mefa-dicarbadodecaborane 8 (136.7 mg, 0.33 mmol, 1 eq.) and 1 ,7- bis(hydroxyl(dimethyl)silyl)-mefa-dicarbadodecaborane 3 (97.8 mg, 0.33 mmol, 1 eq.) in toluene (0.76 mL, to afford a 0.44 M concentration of each monomer.

After a short period (~ 2 minutes) of steady gas evolution, vigorous gas evolution occurred, which subsided to afford a clearer and colourless solution. The reaction mixture was stirred at room temperature for a total of 20 minutes. Then Al ₂ 0 ₃ (Brockmann neutral, ~ 0.1 g) was added to the reaction mixture with further stirring for 10 minutes at room temperature. The reaction mixture was then filtered (with toluene washings) and the filtrate concentrated by evaporation under reduced pressure to afford 200 as a colourless/opaque waxy film/brittle solid. The product is referred to herein as "Sample B200-1".

Product characteristics of B200-1 : (2.23 mg, 91 % c.y.): δ _Η (400MHz, CDCI ₃ ) 3.2 - 1.5 (10H, bm, 10 x BH), 0.22 (12H, 2 χ Si(CH ₃ ) ₂ ), 0.11 (6H,OSi(CH ₃ ) ₂ 0); 5 _C (400 MHz, CDCI ₃ ) 68.3 (C), 1.0 (OSi(CH ₃ )20), 0.5 ((mefa-carborane)Si(CH ₃ ) ₂ 0); δ _Β (400 MHz, CDCI ₃ ) -2.82 (2B), - 7.72 (2B), -9.78 (4B), -13.69 (2B); 6 _Si (400 MHz, CDCI ₃ ) -0.7 ((mefa-carborane)S/(CH ₃ ) ₂ 0), - 19.0 (OS/ ^' (CH ₃ ) ₂ O). Example 7

Poly(carborane-co-siloxane)s having pendent vinyl groups were produced according to the scheme above, following a B(C ₆ F ₅ )3-catalysed polymerisation procedure analogous to Example 3.

hydrosilation with 1 ,1 ,3,3-tetramethyldisiloxane using Karstedt's catalyst (in vinyl terminated PDMS), with a viny silane stoichiometry of 1 :1. The reagent siloxanes were mixed with the catalyst and moulded under pressure, and afforded an opaque colourless elastomeric material after curing within two hours at room temperature.

Example 9

1 ,7-Bis(hydroxydimethylsilyl)-m-dicarbadodecaborane 3 (2.020 g, 6.91 mmol, 1.1 eq.) and 1 ,1 ,3,3,5,5-hexamethyltrisiloxane 5 (1.59 mL, 6.28 mmol, 1.0 eq.) were taken up in hexane (20 mL) to afford a reaction mixture with a suspension of undissolved carborahe component. The reaction mixture was stirred at room temperature in air, without any special precaution to exclude moisture. B(CeF ₅ )3 (10 pL of a solution of catalyst (22.0 mg) in toluene (200 pL)) was added in one portion, and the reaction mixture immediately and vigorously evolyed gas. Over five minutes the reaction mixture became a clear colourless solution after complete solution _. of the carborane reagent. After 2 hours AI2O3 (Brockmann, neutral, 0.5 g) was added to the reaction solution/suspension to sequester catalyst before filtration through filter paper under gravity to yield 400 (3.105 g, 91 %) as a clear, colourless/translucent viscous oil.

Product characteristics: i½ax/cm- ¹ (ATR) 2961 (u(C-H)), 2592 (υ(Β-Η)), 1258, 1026, 856, 790; δ _Η (400 MHz, CDC ) 3.2 - 1.5 (10H, bm, 10 x BH), 0.20 - 0.20 (12H, m, 2 Si(CH ₃ ) ₂ ), 0.11 - 0.08 (18H, m, 3 x Si(CH ₃ ) ₂ ); _C (400 MHz, APT, CDCI ₃ ) 68.4 (C), 1.0 (Si(CH ₃ ) ₂ ), 0.5 (2 ^χ Si(CH ₃ ) ₂ ); δ _Β (400 MHz, CDCI ₃ ) -2.97 (2B, bs), -7.73 (2B, bs), -9.79 (4B, bs), -13.70 (2B, bs); 5si (400 MHz, INEPT CDCI ₃ ) -1.00 (2 ^χ S/(CH ₃ ) ₂ ), -20.41 (2 x S/(CH ₃ ) ₂ ),-21.87 (1 x S/(CH ₃ ) ₂ ).

The poly(dicarbadodecaborane-co-decamethylpentasiloxane) 400 (2.43 g, 0.44 mmol, 1 .0 eq.) and a hydroxyl-terminated polydimethylsiloxane ("DMS-S35", M _w 49,000; 0.07%(OH), viscosity 5,000 cSt, 5.66 g, 0.12 mmol, 3.8 eq.) were centrifugally mixed for one minute to afford a colourless opaque mixture. Tetraethoxysilane (85 μΙ_, 0.51 mmol, 1.15 eq.) was added and the reaction mixture was centrifugally mixed for 30 seconds. Sn(ll)oct ₂ (0.1 ml_, oct = 2-ethyl hexanoate) catalyst was added and the reaction mixture was immediately mixed centrifugally for 15 seconds before immediately casting the material into a Perspex circular mould (diameter = 70 mm, depth = 2 mm) and curing under pressure (6 T), to afford a colourless opaque elastomeric material. Once the material had cured sufficiently, the pad was removed and post-cured at 75 °C for 10 h., to afford a colourless opaque elastomeric material having a compressive Young's modulus (DMA) of -1.17 MPa and a boron content of 6.5 wt% (calculated from the molecular formulae of the components and the reaction stoichiometry). This material is referred to herein as "Sample B400C-1".

Example 10

Poly(m-dicarbadodecaborane-co-1 ,1 ,3,3,5,5,7,7,9,9,11 ,11-dodecamethylhexasiloxane) 500 was produced under analogous conditions to Example 9, using 1 ,7- Bis(hydroxydimethylsilyl)-m-dicarbadodecaborane 3 (2.023 g, 6.92 mmol, 1 .1 eq.) and 1 ,1 ,3,3,5,5,7,7-octamethyltetrasiloxane 5 (2.06 ml_. 6.29 mmol, 1 .0 eq) as starting materials.

Product characteristics: u _max /cm- ¹ (ATR) 2961 (u(C-H)), 2592 (υ(Β-Η)), 1258, 1023, 856, 789; δ _Η (400 MHz, CDCI ₃ ) 3.2 - 1.5 (10H, bm, 10 ^χ BH), 0.20 (6H, s, 1 ^χ Si(Ctf ₃ ) ₂ ), 0.18 (1 H, s, 1 x Si(CH ₃ ) ₂ ), 0.12 - 0.08 (24H, m, 4 x Si(CH ₃ ) ₂ ); 6 _C (400 MHz, APT, CDCI ₃ ) 68.4 (C), 1.0 (Si(CH ₃ ) ₂ ), 0.5 (2 Si(CH ₃ ) ₂ ); δ _Β (400 MHz, CDCI ₃ ) -2.85 (2B, bs), -7.90 (2B, bs), -9.79 (4B, bs), -13.58 (2B, bs); 6 _Si (400 MHz, INEPT CDCI ₃ ) -0.79 (1 x S/(CH ₃ ) ₂ ), -1.03 (1 x S/(CH ₃ ) ₂ ), - 20.45 (2 x S/(CH ₃ ) ₂ ), -21.84 (1 S/(CH ₃ ) ₂ ), -22.24 (1 x S/(CH ₃ ) ₂ ).

The poly(dicarbadodecaborane-co-dodecamethylhexasiloxane) 500 (2.99 g, 0.48 mmol, 1.0 ' eq) was subsequently reacted with DMS-S35 (6.06g, 0.124 mmol, 3.8 eq) and

tetraethoxysilane_(90 μΙ_, 0.55 mmol, 1.15 eq.) in the same way as described for Example 9 to produce an elastomeric pad having a Young's modulus (DMA) of -0.97 MPa and a boron content of 6.2 wt% (calculated from the molecular formulae of the components and the reaction stoichiometry). This material is referred to herein as "Sample B500C-1 ". Example 11

A poly(dicarbadodecaborane-co-decamethylpentasiloxane) incorporating 3% vinyl groups and having an M _p of 27,700 Da was produced following the procedure outlined in Example 7. The resulting polymer was blended with the peroxide crosslinking reagent Trigonox 101 (40- 50% 2,5-di(tert-butylperoxy)-2,5-dimethylhexane in silicone oil, Akzo Nobel) or Perkadox BC (40-50 wt% dicumyl peroxide in silicone oil, Akzo Nobel) using a Torrex mixer at a ratio of 100 parts polymer to 1-5 parts peroxide, before allowing the mixture to degas in a vacuum chamber at ambient temperature. The polymer and peroxide mixture was then transferred by pipette to a mould, and heated to 160°C at ambient pressure for a minimum period of one hour to produce a moulded product.

1 ,7-Bis(methoxydimethylsilyl)-/77-dicarbadodecaborane 1 was stirred at room temperature in the presence of an excess of pentamethyldisiloxane 10. B(C ₆ F ₅ )3 was added as a dilute solution of B(CeF ₅ )3 in toluene. After addition of the catalyst there was a short induction period, followed by rapid methane gas evolution (for ~ 5 minutes) and a large exotherm that was sufficient to boil off low molecular weight volatiles.

The reaction mixture was stirred over neutral alumina to sequester catalyst after complete reaction. Filtration to remove the catalyst and alumina was followed by concentration of the filtrate under reduced pressure to give the required 1 ,7-bis(1 ,1 ,3,3,5,5- heptamethyltrisiloxanyl)-/D-dicarbadodecaborane 11 in quantitative yield. Failure to remove the catalyst in a control study resulted in changes to the mass spectrum of the material over time, presumably due to cross-linking and degradation reactions taking place in the continued presence of the catalyst. Product that had been purified with alumina to sequester catalyst showed no change after storage under ambient conditions for several months.

The ²⁹ Si-INEPT-NMR spectrum of 11 in CDCI ₃ shows three peaks at -1.21 , -20.07 and +7.55 p.p.m. which are assigned to the (carborane)-S Me2-, 0-S/Me2-0 and terminal 0-S/ e3 groups respectively.

Reference Example 2

1 ,7-Bis(hydroxydimethylsilyl)-m-dicarbadodecaborane 3 was stirred in hexane in the presence of B(C ₆ F ₅ )3, and was unreactive toward the catalyst under the reaction conditions.

Reference Example 3

1 ,7- Bis(methoxydimethylsilyl)-m-dicarbadodecaborane 1 was stirred in hexane in the presence of B(CeFs)3 and afforded recovered starting material 1 and hydrolysed product 3 (with analytical data matching those of bis(hydroxydimethylsilyl)-m-dicarbadodecaborane 3). Reference Example 4

1 ,1 ,3,3-tetramethyldisiloxane 2 was stirred in the presence of B(C6F ₅ )3 overnight to yield a mixture of products, with the consumption of starting material. The ²⁹ Si-INEPT-NMR spectrum of 1 ,1 ,3,3-tetramethyldisiloxane 2 in CDC is a single peak at -4.55 p. p.m., and after reaction the crude mixture's spectrum has a peak at -8.33 and a collection of peaks in the range -19.08 - -22.52 p.p.m.

Reference Example 5

1 ,7-Bis(dimethylsilyl)-m-dicarbadodecaborane 12 did not react in the presence of B(C ₆ Fs)3 - only starting material was recovered (with the chemical shift in the ²⁹ Si-INEPT-NMR spectrum of the silicon atom in CDCI ₃ remaining unchanged at -3.1 p.p.m.).

Reference Example 6

Me Me Me

H— S Ii— CB ₁₀ H ₁₀ C— S Ii— H + EtO -S Ii— OEt

Me Me

t ^'

13

1 ,7-Bis(dimethylsilyl)-m-dicarbadodecaborane 12 was treated with B(C6F ₅ )3 in the presence of diethoxy(methyl)vinylsilane 13, with only 12 recovered in quantitative yield after evaporation of solvent and the volatile reaction components.

Reference Example 7

Me Me Me Me

H S Ii— CB ₁₀ H ₁₀ C— S Ii— H + MeO— Si— CB ₁₀ H ₁₀ C— Si— OMe

Me Me Me Me

12 1

1 ,7-Bis(dimethylsilyl)-m-dicarbadodecaborane 12 was treated with B(C6Fs)3 in the presence of 1 ,7-Bis(methoxydimethylsilyl)-m-dicarbadodecaborane 1. The starting materials 10 and 1 were recovered, along with compound 3 resulting from hydrolysis of the starting reagents.

Comparative Example 1

A Dexsil® 200 sample was prepared following literature precedent [Heying]. All glassware was oven-dried (140 °C) and assembled under a dry nitrogen gas atmosphere. Syntheses were performed using a two-neck round bottomed flask equipped with stirrer bar and condenser, and heated using a silicone oil bath. The chlorosilane reagent(s) was transferred to the reaction. vessel via a syringe. Reactions were maintained under an inert (nitrogen) atmosphere employing a nitrogen flow across the top of the condenser and exiting through an oil bubbler. The completion of the reaction at each stage could be checked by closing the inert gas flow to the apparatus and checking for gas evolution via the bubbler.

1 ,7-Bis(methoxy(dimethyl)silyl)-/w-dicarbadodecaborane 1 (1.225 g, 3.82 mmol, 1.0 eq.), dichlorodimethylsilane 14 (460 pL, 3.82 mmol, 1.0 eq.) and FeCI ₃ (6.0 mg, 38.2 mol, ~1 mol% of 1 ,7-bis(methoxy(dimethyl)silyl)-A77-dicarbadodecaborane) were heated to 125°C over 45 minutes and maintained at this temperature for a further 50 minutes, during which time the reaction mixture became dark yellow with slow gas evolution observed. Once gas evolution had ceased the reaction mixture was allowed to cool to room temperature to afford a dark yellow liquid.

A second portion of FeCI ₃ (5.0 mg, ~1 mol% of

1 ,7-bis(methoxy(dimethyl)silyI)-m-dicarbadodecaborane starting material) was added to. the reaction mixture. The reaction mixture was heated to 170°C over 30 minutes and maintained at this temperature for a further 60 minutes, during which time the reaction mixture became dark brown with slow gas evolution observed. Once gas evolution had ceased the reaction mixture was allowed to cool to room temperature to afford a dark brown waxy solid (crude recovered yield of 1.16 g, 87%).

Product characteristics: δ _Η (400 MHz, CDCI ₃ ) 3.50 (bs, (mefa-carborane)Si(CH ₃ ) ₂ OCH ₃ ), 3.2 - 1.4 (bm, BH), 0.52 (bs, impurity), 0.30 - 0.16 (appbm, (meia-carborane)Si(CH ₃ ) ₂ OMe and (mefa-carborane)Si(CH3)20), 0.16 - 0.04 (appbs, OSi(CH ₃ ) ₂ 0,); δ _ε (400 MHz, APT, CDCI ₃ ) 68.4 (C), 2.5 (impurity), 1.2 - 1.0 (OSi(CH ₃ ) ₂ 0), 0.7 - 0.5 (bm, (meia-carborane)Si(CH ₃ ) ₂ 0), -2.3 (b, (mefa-carborane)Si(CH ₃ ) ₂ OCH ₃ ); δ _Β (400 MHz, CDCI ₃ ) -2.60 (2B), -7.58 (2B), -9.74 (4B), -13.87 (2B); 6 _Si (400 MHz, INEPT, CDCI ₃ ) 24.1 (impurity), 10.0 (S/(CH ₃ ) ₂ OCH ₃ ), 2.1 - 1.8 (m, (mefa-carborane)S/(CH ₃ ) ₂ -0, Dexsil® 100), -0.7 - -0.7 (m,

(mefa-carborane)S/(CH ₃ ) ₂ , Dexsil® 200), -19.0 - -19.0 (m, OS/(CH ₃ ) ₂ 0, Dexsil® 200).

The crude product was acidic and was soluble in chloroform, tetrahydrofuran and diethyl ether; sparingly soluble in acetone (cloudy/opaque suspension) and insoluble in /so-propyl alcohol, methanol and water.

The crude product was purified by washing a diethyl ether solution (10 mL) with water (3 ^ 5 ml_). The acidic aqueous layer was discarded and the organic solution concentrated by ¹ evaporation under reduced pressure. The washed crude product was then extracted sequentially with water, water/acetone (10%v/v) and water to afford a waxy brown solid. The product is referred to herein as "Sample F200-1".

Product characteristics: Dmax/cnr ¹ (ATR) 2962, 2592, 2557, 1258, 1051 , 855, 793, 666; δ _Η (400 MHz, CDCIs) 3.48 (s, Si(CH ₃ ) ₂ OCH ₃ ), 3.2 - 1.4 (m, BH), 0.26 (s, {meta- carborane)Si(CH ₃ ) ₂ OH), 0.23 - 0.23 (m, (mefa-carborane)Si(CW ₃ ) ₂ OCH ₃ and (meta- carborane)Si(CH ₃ ) ₂ 0, Dexsil® 100), 0.21 - 0.20 (m, (me.a-carborane)Si(CH ₃ ) ₂ 0, Dexsil® 200), 0.10 - 0.09 (m, OSi(CH ₃ ) ₂ 0, Dexsil® 200); δ ₀ (400 MHz, APT, CDCI ₃ ) 68.3 (C), 1.0 (0Si(CH ₃ ) ₂ 0), 0.6 ((mefa-carborane)Si(CH ₃ ) ₂ 0), 0.5 ((meia-carborane)Si(CH ₃ ) ₂ 0), -0.4 (mefa-carborane)Si(CH ₃ ) ₂ OH), -2.3 (mefa-carborane)Si(CH ₃ ) ₂ OCH ₃ ); δ _Β (400 MHz, CDCI ₃ ) - 2.60 (2B), -7.58 (2B), -9.74 (4B), -13.87 (2B); δ _& (400 MHz, INEPT, CDCfe) 10.0 {(meta- carborane)S/(CH3)20CH ₃ ), 9.5 ((mefa-carborane)S/(CH ₃ ) ₂ OH), 1.9 ((carb0rane)S/(CH ₃ ) ₂ 0, Dexsil® 100), -0.7 ((carborane)S/ ^' (CH ₃ ) ₂ 0, Dexsil® 200), -19.0 ((carborane)S/(CH ₃ ) ₂ 0, Dexsil® 200); DSC (10 °C min- ¹ ) glass transition temperature (at half AC _P ) = -39.2 °C, melting temperature = 26.6 °C, AH _f = 0.449 J g- ¹ .

Comparative Example 2

A further Dexsil® 200 sample was prepared in a similar fashion to Comparative Example 1.

1 ,7-Bis(methoxy(dimethyl)silyl)-m-dicarbadodecaborane 1 (0.945 g, 2.95 mmol, 1.0 eq.), dichlorodimethylsilane 14 (355 μΙ_, 2.95 mmol, 1.0 eq.) and FeCI ₃ (5.3 mg, 32.6 μηηοΙ, ~1 mol% of 1 ,7-bis(methoxy(dimethyl)silyl)-m-dicarbadodecaborane) were heated to 120°C over 10 minutes and maintained at this temperature for a further 40 minutes, during which time the reaction mixture became yellow with slow gas evolution observed (one bubble every 2-3 seconds). Once gas evolution had ceased the reaction mixture was allowed to cool to room temperature to afford a dark yellow liquid.

A second portion of FeCI ₃ (7.0 mg, ~1 mol% of 1 ,7-bis(methoxy(dimethyl)silyl)-m- dicarbadodecaborane starting material) was added to the reaction mixture. The reaction mixture was heated to 180°C over 30 minutes and maintained at this temperature for a further 85 minutes, during which time the reaction mixture became dark yellow/brown with slow gas evolution observed (one bubble every 2 seconds). Once gas evolution had ceased the reaction mixture was allowed to cool to room temperature to afford a dark brown waxy solid. .

A third portion of FeCI ₃ (9.5 mg, ~1 mol% of 1 ,7-bis(methoxy(dimethyl)silyl)-m- dicarbadodecaborane starting material) was added to the reaction mixture. The reaction mixture was heated to 180°C over 25 minutes arid maintained at this temperature for a further 80 minutes, during which time the reaction mixture became dark brown with slow gas evolution observed. Once gas evolution had ceased the reaction mixture was allowed to cool to room temperature to afford a dark brown waxy solid (crude recovered yield of 0.78 g, 76%). The crude, dark brown waxy product was dissolved in Et20 (20 mL) and the cloudy brown solution was washed with distilled water (3 x 10 mL). Brine was added to separate the organic and aqueous layers and the organic layer was dried (MgS0 ₄ ) and concentrated by evaporation to afford the product as a waxy brown solid. The product is referred to herein as "Sample F200-2".

Product characteristics: u _raax /cm- ¹ (ATR) 2961 , 2591 , 1258, 1046, 856, 792, 666; δ _Η (400 MHz, CDCI ₃ ) 3.48 (s, Si-OCH ₃ ), 3.2 - 1.4 (m, BH), 0.23 (bs, (mefa-carborane)Si(CH ₃ ) ₂ OMe and (meta-carborane)-Si(CH ₃ ) ₂ -0, Dexsil® 100), 0.21 - 0.20 (m, (meta- carborane)Si(CH ₃ ) ₂ 0, Dexsil® 200), 0.10 - 0.09 (m, OSi(CH ₃ ) ₂ 0, Dexsil® 200); δ ₀ (400 MHz, APT, CDCI ₃ ) 68.3 (2 ^χ C), 0.9 (0-Si(CH ₃ ) ₂ -0), 0.5 ((carborane)Si(CH ₃ ) ₂ 0, Dexsil® 100), 0.5 ((carborane)Si(CH ₃ ) ₂ 0, Dexsil® 200), -2.3 ((me.a-carborane)Si(CH ₃ ) ₂ OMe); δ _Β (400 MHz, CDCI ₃ ) -2.55 (2B), -7.60 (2B), -9.74 (4B), -13.95 (2B); 6 _Si (400 MHz, INEPT, CDCIs) 10.0 ((mefa-carborane)S/(CH ₃ ) ₂ OMe), 1.9 ((meia-carborane)S/(CH ₃ ) ₂ 0, Dexsil® 100), 0.8 ((meia-carborane)S/(CH ₃ ) ₂ 0, Dexsil® 200), -19.0 ((carborane)S/(CH ₃ ) ₂ 0, Dexsil® 200).

Comparative Example 3

A sample of the crude product from Comparative Example 2 was cut into thin pieces and further purified by sequential washing with acetone, 10% acetone/water and acetone, leaving the material to swell and extract with gentle agitation [Heying]. The material did not become visibly less coloured. The product is referred to herein as "Sample F200-3".

The data was comparable to that of Comparative Example 1 : DSC (10 °C min ^-1 ) glass transition temperature (at half C _p ) = -34.1 °C, melting temperature = 26.7°C,

Comparative Example 4

A sample of Dexsil® 200 was purchased from KatChem Ltd. The sample was a brown opaque waxy solid. The product is referred to herein as "Sample F200-4".

Product characteristics: (found: B, 30.50; requires B, 31.01); δ _Η (400 MHz, CDCI ₃ ) 3.20 - 1.50 (H, bm, BH), 0.26 (s, (/ne/a-carborane)Si(CH ₃ ) ₂ OH), 0.23 (s,

(mefa-carborane)Si(CH ₃ ) ₂ 0, Dexsil® 100), 0.21 (bs, (mefa-carborane)Si(CH ₃ ) ₂ 0, Dexsil® 200 and Dexsil® 300), 0.1 - 0.1 (m, OSi(CH ₃ ) ₂ 0, Dexsil® 200 and Dexsil® 300); 6 _C (400 MHz, APT, CDCIa) 68.3 - 68.3 (m, C), 1.0 - 1.0 (OSi(CH ₃ ) ₂ 0), 0.5 - 0.5 (m, (meta- carborane)Si(CH ₃ ) ₂ 0), -0.4 ((meia-carborane)Si(CH ₃ ) ₂ OH); δ _Β (400 MHz, CDCI ₃ ) -2.75 (2B, bs), -7.61 (2B, bs), -9.81 (4B, bs), -14.05 (2B, bs); 6 _Si (400 MHz, INEPT, CDCI ₃ ) 9.5

((meia-carborane)S/(CH ₃ ) ₂ OH), 1.9-1.8 (m, (mefa-carborane)S/ ^' (CH ₃ ) ₂ 0, Dexsil® 100), -0.7 - -0.9 (m, (mefa-carborane)S/(CH ₃ ) ₂ 0, Dexsil® 200 and Dexsil® 300), -18.9 - -19.0 (m, OS/(CH ₃ ) ₂ 0, Dexsil® 200), -20.4 - -20.5 (m, OS/(CH ₃ ) ₂ 0, Dexsil® 300); DSC (100°C min.-1) glass transition temperature at onset = -41.8°C and at half Cp = -39.1 °C,

ACp = 0.16 J g ^" K ^~1 :

Comparative Example 5

1 ,7-Bis(methoxy(dimethyl)silyl)-m-dicarbadodecaborane 1 (1.048 g, 3.30 mmol, 1.0 eq.), dichlorodimethylsilane 15 (355 μΙ_, 2.95 mmol, 0.9 eq.), dichloro(methyl)vinylsilane 16 (46.5 mg, 330 μιηοΙ, 0.33 mmol, 0.1 eq.) and FeCI ₃ (5.4 mg, 33.0 μηηοΙ, ~1 mol% of

bis(meth0xy(dimethyl)silyl)-m-dicarbadodecaborane) were heated to 120°C over 10 minuntes and maintained at this temperature for a further 30 minutes, during which time the reaction mixture became yellow with slow gas evolution observed (one bubble every 2 seconds). Once gas evolution had ceased the reaction mixture was allowed to cool to room temperature to afford a yellow liquid.

A second portion of FeCh (4.5 mg, ~1 mol% of bis(methoxy(dimethyl)silyl)-m- dicarbadodecaborane starting material) was added to the reaction mixture. The reaction mixture was heated to 180°C (over 15 minutes) and maintained at this temperature for a further 55 minutes, during which time the reaction mixture became dark yellow/brown with slow gas evolution observed (one bubble every 2 seconds). Once gas evolution had ceased the reaction mixture was allowed to cool to room temperature to afford a dark brown tacky waxy solid (recovered yield of 0.76 g) that was soluble in THF, E.2O and CHCI3 and insoluble in /so-propyl alcohol and water.

The crude product was dissolved in Et20 (20 mL) and washed with water (3 * 10 ml_). Brine was added to separate the layers and the organic layer was dried (MgS0 ₄ ) and

concentrated to afford the product as a tacky waxy brown solid. The product is referred to herein as "Sample F200A-1".

Results and Discussion

Control Experiments

To study the viability of the use of boron-based Lewis acid catalysts in preparing

poly(carborane-cosiloxane) polymers, various coupling reactions were carried out in the presence of B(C6Fs)3-

Firstly, 1 ,7-bis(methoxydimethylsilyl)- 77-dicarbadodecaborane 1 was stirred at room temperature in the presence of an excess of pentamethyldisiloxane 10 and B(C ₆ F ₅ )3 as a catalyst (Reference Example 1 ). This reaction produced 1 ,7- bis(1 ,1 ,3,3,5,5-heptamethyltrisiloxanyl)-m-dicarbadodecaborane 11 in quantitative yield, confirming that B(CeF ₅ )3 is an effective catalyst for coupling reactions involving

alkoxy(carboranyl)silanes at room temperature with a high yield. There was no evidence of an improved reaction when the synthesis of 11 was repeated with rigorous. exclusion of moisture, with the intended coupling being rapid and high yielding.

To test whether the various starting materials used in the reaction undergo homo-coupling in the presence of a boron-based catalyst, various control experiments were conducted.

1 ,7-Bis(hydroxydimethylsilyl)-m-dicarbadodecaborane 3 was stirred in hexane in the presence of B(CeF ₅ )3, and was unreactive toward the catalyst under the reaction conditions (Reference Example 2). Bis(methoxydimethylsilyl)-m-dicarbadodecaborane 1 yielded some hydrolysed product (with analytical data matching those of bis(hydroxydimethylsilyl)-m- dicarbadodecaborane 3) under the same reaction conditions, presumably via the catalyst behaving as a traditional Lewis acid and mediating hydrolysis by trace water present in the reaction (Reference Example 3). It should be noted that the hydrolysis of

methoxydimethylsilyl- to hydroxydimethylsilyl- does not affect the general reactivity of the functional group towards the proposed borohydride intermediate, or the repeat unit of the final co-polymer that would be prepared from either monomer.

1 ,1 ,3,3-tetramethyldisiloxane 2 stirred in the presence of B(C ₆ F ₅ )3 overnight yielded a mixture of products (Reference Example 4), with the consumption of starting material. The ²⁹ Si-INEPT-NMR spectrum of 1 ,1 ,3,3-tetramethyldisiloxane 2 in CDC is a single peak at - 4.55 p. p.m., and after reaction the crude mixture's spectrum has a peak at -8.33 and a collection of peaks in the range -19.08 - -22.52 p.p.m. This observation is consistent with the silane groups, as for pentamethyldisiloxane 10, reacting with B(CeF5)3 to form a borohydride intermediate which may then be nucleophilically-attacked by trace moisture in the system or by the siloxane bond within disiloxane 2. Since pentamethyldisiloxane has been shown to react cleanly (see Reference Example 1) it is likely that the former pathway predominates, and that the oxonium anion formed in situ undergoes nucleophiiic attack by water and subsequent reaction of the transition state. Since this initially forms the silanol 1 ,1 ,3,3- pentamethyldisiloxan-1-ol, which can go on to couple with further activated silane groups, the product distribution is expected to contain an equilibrium mixture of cyclic, oligomeric and polymeric poly(siloxane) products, which is consistent with the experimental observations.

1 ,7-Bis(dimethylsilyl)-m-dicarbadodecaborane 12 did not react in the presence of B(C6F ₅ )3 (Reference Example 5) which is expected since there is no source of siloxane to co-ordinate with any borohydride intermediate that may form - only starting material was recovered from this control reaction (with the chemical shift in the ²⁹ Si-INEPT-NMR spectrum of the silicon atom of 12 in CDCI3 remaining unchanged at -3.1 p.p.m.).

No evidence of gas evolution was observed upon addition of the catalyst in Reference Examples 2 to 5, and any hydrolysis or coupling is believed to occur at a far slower rate relative to the proposed B(C6F ₅ )3-catalysed borohydride formation and redox process between suitable coupling partners.

The lack of reaction of 1 ,7-bis(dimethylsilyl)-m-dicarbadodecaborane 12 in the presence of B(C ₆ F ₅ )3 could be due to the postulated borohydride intermediate failing to form, or the low reactivity of the intermediate which does not react in the presence of trace moisture.

1 ,7-Bis(dimethylsilyl)-m-dicarbadodecaborane 12 was treated with B(CeFs)3 in the presence of the alkoxysilane-containing reagent diethoxy(methyl)vinylsilane 13, with only 12 recovered in quantitative yield after evaporation of solvent and the volatile reaction components (Reference Example 6). Diethoxy(methyl)vinylsilane 13 was selected as a potential coupling partner since it is considered a standard alkoxysilane and is expected to couple with the proposed borohydride intermediate should it form in situ. Attempted catalytic coupling between 1 and 12 in an attempt to prepare the carborane-rich poly(carborane-co- tetramethyldisiloxane) linear polymeric structure also failed, resulting in the return of starting material and hydrolysed product (Reference Example 7).

Both of these results are in agreement with 1 ,7-bis(dimethylsilyl)-m-dicarbadodecaborane 12 being unreactive towards B(C ₆ F ₅ )3 under the reaction conditions, possibly due to the electron-withdrawing properties of the adjacent carborane cage deactivating the silane group and not favouring formation of the proposed borohydride intermediate. Thus 12 is not a suitable bis(silane)-component for the preparation of poly(carborane-co-siloxane)s via this methodology.

The control reactions discussed here are relevant to the production of poly(carborane-co- siloxane) polymers since for a controlled synthesis of a co-polymer from alternating monomer units each monomer should not homo-couple. The individual bis(alkoxysilyl)- and bis(hydrosilyl)- monomers have been shown to not appreciably homo-couple, and the reaction mechanism dictates that a hydrosilane group couples with an alkoxysilane or hydroxysilane. The coupling reaction is inherently controlled by the chemistry, and selectively yields regular poly(carborane-co-siloxane)s.

Preparation of polv(carborane-co-siloxanes) via boron-based Lewis acid-catalvsed coupling The control reactions above showed that 1 ,7-bis(methoxydimethylsilyl)-/n- dicarbadodecaborane 1 couples with monosilane 10 in the presence of B(CeF ₅ )3 (Reference Example 1). To prepare polymers from this methodology both 1 ,7-bis(methoxydimethylsilyl)- m-dicarbadodecaborane 1 and 1 ,7-bis(hydroxydimethylsilyl)-m-dicarbadodecaborane 3 were coupled under B(CeFs)3 catalysis with 1 ,1 ,3,3-tetramethyldisiloxane 2 to form linear poly(dicarbadodecaborane-co-octamethyltetrasiloxane) polymers, referred to herein as Samples Ε330Ό-1 and B300-2 (described respectively in Examples 1 and 2).

From the control reactions and by consideration of the reaction mechanism (see above) the carborane monomers do not undergo homo-coupling under the reaction conditions. The silane groups are expected to react to yield the borohydride intermediate which activates the silicon atom to nucleophilic attack and it has been shown that poly(siloxane) oligomers and cyclics may form in the presence of trace amounts of moisture. When there is a large excess of a competing alkoxysilyl or hydroxysilyl- nucleophile it is expected that the intended coupling will be the most favourable hence yielding co-polymer - indeed, experiments performed under anhydrous conditions did not show any improvement in the yield or purity of the product polymer. From these observations, the polymer would be expected to be highly ordered and to contain the repeat unit as depicted in Examples 1 and 2 above.

The peak molecular weights of B300-1 and B300-2 were determined by gel permeation chromatography using Varian PLgel mixed-B columns in toluene (Table 1). The polymers under investigation are almost /so-refractive with toluene, and therefore in this solvent the polymers give a weak refractive response, but suitable retention time data could be obtained under differential viscosity detection (Figure 1). The differential pressure (DP)

chromatogram for Example 2 shows a small shoulder at lower retention volume (near 14 mL, and not visible under Rl detection) which suggests a small amount of high molecular weight material in this sample.

Table 1 The NMR data for B300-1 and B300-2 are identical, within experimental error, confirming the identical repeat unit in each polymer. This is as expected from the synthesis as both the hydroxydimethylsilyl- and methoxydimethylsilyl- groups are expected to react in the same manner. No end groups are visible in the NMR spectra for B300-1 and B300-2, unlike for the Dexsil®-200 material of Comparative Example 4, supporting the conclusion that the B300-1 and B300-2 are of higher molecular weight than the Dexsil® 200 of Comparative Example 4, or possibly that the polymers have self-condensed to form large (i.e. macro) cycles.

Infra-red spectra for B300-1 and B300-2 are in agreement with literature data [Mohadger] and DFT modelling.

Furthermore, the polymer product of the B(C6Fs)3-catalysed reaction, regardless of the choice of carborane-containing precursor, is visibly cleaner than the commercially available - Dexsil®-200 of Comparative Example 4.

Synthesis of polymers 300, 400 and 500 under lower initial monomer concentrations in toluene was carried out as described in Examples 3, 4 and 5, and the molecular weight of the resultant samples measured by gel permeation chromatography (Table 2). These syntheses were performed at increased scale of starting material (~ 4-5 g

1 ,7-bis(hydroxyl(dimethyl)silyl-/nefa-dicarbadodecaborane) and a 1 :1 ratio of the carborane to disilane monomers.

All polymers so produced were non-waxy and fluid with excellent flow properties for mould filling, and are therefore well suited to preparing shaped elastomeric materials via a suitable cross-linking chemical reaction. the molecular weight data show that the polymerisation reactions give a major peak at lower retention volume, with a quantity of product at greater retention volume (around 16 mL on the setup used) corresponding to low molecular Weight.

It should be noted that the refractive index of THF (1.407) [Lide] is almost identical to poly(dimethylsiloxane)s (1.408) [Gelest] and therefore as the carborane content of the polymer decreases (and the polymer structure more closely resembles pure

poly(dimethylsiloxane) the polymer's refractive index converges on that of the solvent used. Therefore, for the same polymer concentration the intensity of the measured refractive index decreases in the order 300 > 400 > 500. This means that the relative intensities of the chromatograms 300, 400 and 500 are not readily comparable, even though the sample concentrations were similar.

The samples produced from reactions with the lowest initial monomer concentration

(0.175 M) showed a significant quantity of low molecular weight product (at a retention volume of -16 mL). The respective ¹ H and ²⁹ Si-NMR data show only peaks corresponding to a single component, meaning that the low and higher molecular weight peaks are from material that has the same structural composition but that must differ in molecular weight.

Preparation of Dolv(m-dicarbadodecaborane-co-1.1.3.3.5.5-hexamethyltrisilox ane) via boron- based Lewis acid-catalvsed coupling

Dexsil® 200 prepared via the conventional FeCI ₃ -mediated route has been shown to produce non-regular poly(carborane-co-siloxane), comprising of meia-carborane units separated by predominantly two siloxane groups, but with a proportion of 'Dexsil® 100' and 'Dexsil® 300' repeat units in the material. Hence the polymer is best described as comprising of different types of main chain placements.

The non-uniformity of the polymer produced via the FeCb-catalysed procedure arises from the exchange of Si-OMe and Si-CI groups between the reagents during the reaction in the presence of the catalyst, before monomer condensation and the irreversible formation of chloromethane [Dietrich]. Any moisture in the reaction will also hydrolyse Si-CI groups to Si-OH which can self-condense under the forcing (i.e. high concentration and temperature) conditions.

As shown in Reference Example 7, attempting to react 1 ,7-bis(methoxy(dimethyl)silyl)-/neia- dicarbadodecaborane 12 and 1 ,7-bis(dimethylsilyl)-meia-dicarbadodecaborane 1 in the presence of B(C ₆ F ₅ )3 affords only starting materials and some hydrolysis product 3, rather than the carborane-rich poly(dicarbadodecaborane-co-tetramethyldisiloxane) polymer. This may be rationalised by the electron-withdrawing electronic properties of the carborane in 12 deactivating the silane group and not favouring formation of the proposed borohydride intermediate (see Scheme 1 above). Thus, an alternative route to poly(dicarbadodecaborane-co-hexamethyltrisiloxane) was devised: 1 ,7-bis(1 ,1 ,3,3-tetramethyldisiloxanyl)-mefa-dicarbadodecaborane 8 was prepared by the condensation of 1 ,7-bis(hydroxyl(dimethyl)silyl)-mefa-dicarbadodecaborane 3 with an excess of chlorodimethylsilane 7 (see Example 6). The crude product was purified by flash column chromatography to afford the pure required product; although no recovered 3 was observed in the crude product a significant quantity of the half-reacted intermediate 9 was observed, and a portion was isolated and characterised to confirm its structure. This unoptimised route to 1 ,7-bis(1 ,1 ,3,3- tetramethyldisiloxanylj-mefa-dicarbadodecaborane affords a relatively low yield, likely due to the acidic workup of the reaction and the instability of the required product to silica in the purification step. However, sufficient co-monomer was prepared to enable the catalysed coupling with 1,7-bis(hydroxyl(dimethyl)silyl)-mefa- dicarbadodecaborane 3 to be undertaken to produce polymer 200.

The NMR spectra of product B200-1 is in agreement with the poly(dicarbadodecaborane-co- tetramethyldisiloxane) structure but also contain lower intensity peaks in the ²⁹ Si-NMR (and some low intensity unassigned peaks in the ¹ H- and ³ C-NMR spectra). These peaks are not consistent with any other connectivity in the polymer backbone (i.e.

poly(dicafbadodecaborane-co-tetramethyldisiloxane) ('Dexsil® 100'-like) or

poly(dicarbadodecaborane-co-octamethyltetrasiloxane) ('Dexsil® 300'-like) and higher structures) and are likely to be due to some cyclic arrangements that will give slightly different chemical shifts.

As discussed below, Dexsil® 200 produced via the conventional FeCh-mediated route has been shown to have an irregular structure. In contrast, the equivalent material prepared from the B(C ₆ Fs)3-catalysed condensation is regular and far cleaner, with NMR spectra only showing peaks consistent with a regular 'Dexsil® 200'-like structure.

Gel permeation chromatography of B200-1 (Table 3) showed that the polymer is of comparable molecular weight to that measured for a related poly(dicarbadodecaborane-co- octamethyltetrasiloxane) polymer (sample B300-6, produced in an analogous fashion to Example 3 above) produced via B(CeF ₅ )3 at the same initial monomer concentration of 0.44 M (see also Table 2).

Table 3

Poly(styrene) equivalent molecular weight

Preparation of polv(carborane-co-siloxane)s with pendent vinyl groups The incorporation of pendent functional groups along the polymer backbone at locations for subsequent cross-linking reactions offers an attractive approach to producing cross-linked polymers.

B(C ₆ F ₅ )3 has been reported to catalyse hydrosilation of alkenes with R3S1H, utilising 5 mol% catalyst in dry CH2CI2 [Rubin]. However, B(C6F ₅ )3-catalysed siloxane formation is unperturbed in the presence of alkenylsilanes when low boron catalyst loadings are employed.

Therefore, vinylsilane groups are considered to be non-reactive under the general reaction conditions employed for the dehydrocarbonative condensation of alkoxysilanes and hydrosilanes, and hence should be readily incorporated into a polymer backbone using appropriate monomers. Vinyl groups are useful to incorporate pendent to the polymer backbone since they are an attractive reactive handle for cross-linking, for example through hydrosilation or peroxide-mediated cross-linking.

A series of vinyl-pendent poly(carborane-co-siloxane)s were prepared under B(C6F ₅ )3 catalysis, as described in Example 7, using diethoxy(methyl)vinylsilane (R = Et) as the vinylsilane source (Table 4). The ratio disilane:silanol/alkoxysilane was maintained at 1 :1 for the vinyl-pendent poly(carborane-co-siloxahe)s, (i.e. α + γ = β in the scheme shown in Example 7). Prepared polymers were analysed by gel permeation chromatography (see Figure 2 and Table 4).

Synthesised polymers showed unreacted vinyl groups by ¹ H, ³ C and ²⁹ Si-NMR

spectroscopy, confirming that the groups had not reacted under the reaction conditions. The resonances for the vinyl groups showed slight broadening compared to the starting material, consistent with their incorporation within the polymer.

Table 4

*Poly(styrene) equivalent

Using Me(vinyl)Si(OEt) ₂ in the 300 polymer series synthesis appears to decrease the molecular weight of the product, with greater Me(vinyl)Si(OEt)2 further reducing the measured molecular weight relative to the control sample (B300-7, which has 0% vinyl content). The NMR spectra of B300A-3 shows unreacted ethoxysiiane groups which is evidence that the ethoxysiiane groups of the Me(vinyl)Si(OEt) ₂ monomer are less reactive than silanol groups of 1 ,7-bis(hydroxyl(dimethyl)silyl)- r7efa-dicarbadodecaborane. This results in preferential coupling of the carborane-containing monomer with the disilane, leaving a reservoir of unreacted Me(vinyl)Si(OEt)2 and/or oligomer chains terminated with - SiMe(vinyl)OEt.

The polymerisation reaction proceeds via a step-growth mechanism and any deviation from the optimum 1 :1 ratio of reactive groups will greatly reduce the maximum molecular weight that is achieved.

Due to the difference in reactivity between the silanol and ethoxysiiane groups increasing the Me(vinyl)Si(OEt)2 content of the reaction mixture effectively reduces the reaction

stoichiometry between the silane and alkoxysilane/silanol from the optimum 1 :1 , and this contributes to the lower molecular weight achieved in the polymerisations using

Me(vinyl)Si(OEt) ₂ .

The proposed borohydride intermediate (in Scheme 1 ) formed by the reaction of silane and B(C ₆ F ₅ )3 catalyst preferentially reacts with and consumes silanol, and as the free silanol concentration in solution decreases there is competition between coupling with ethoxysiiane monomer in solution and intramolecular cyclisation, with concomitant polymerisation termination. If cyclisation and termination is favoured on reactivity grounds then this will afford low molecular weight cycles and relatively low molecular weight oligomers and a reservoir of un/partially reacted Me(vinyl)Si(OEt)2.

For the vinyl-pendent 400 and 500 variants a significant amount of low molecular weight product was observed in the gel permeation chromatograms, as was found in the case for the pure 400 and 500 products. Analysis of these polymers also showed unreacted ethoxysiiane groups.

Me(vinyl)Si(OMe) ₂ was considered as an alternative monomer with which to introduce vinylsilane into the polymer backbone: the dimethoxysilane functionality is less hindered than the corresponding diethoxysilane and was predicted to react at a greater rate with the borohydride intermediate and to tend towards complete reaction [Thompson].

The closer the reaction rates of silane with Me(vinyl)Si(OMe) ₂ and silanol with silane under B(C ₆ F ₅ )3 catalysis the closer the effective silanol/alkoxysilane monomer concentration in solution will be, and hence the greater the molecular weight achieved via the step-growth polymerisation and the more complete the polymerisation reaction should be.

Vinyl-pendent poly(carborane-co-siloxane)s were prepared from Me(vinyl)Si(OMe)2 under B(C ₆ F ₅ )3 catalysis and analysed by gel permeation chromatography (Figure 3 and Table 5).

The gel permeation data for the syntheses shows that employing Me(vinyl)Si(OMe)2 over Me(vinyl)Si(OEt) ₂ favours higher molecular weight (Figure 3); NMR spectra of the polymer products derived from Me(vinyl)Si(OMe)2 also showed no residual methoxysilane groups. These two observations are consistent with the greater reaction rate of Me(vinyl)Si(OMe) ₂ over Me(vinyl)Si(OEt)2, and complete reaction as intended.

The molecular weight data for the vinyl-pendent polymers prepared using Me(vinyl)Si(OMe) ₂ are close to the pure poly(carborane-co-siloxane) B300-7 control (Figure 3 and Table 4).

The experiments reported above have shown that vinyl-pendent poly(carborane-co- siloxane)s in the 300 series can be prepared with comparable molecular weight to the non- vinyl containing polymer using Me(vinyl)Si(OMe)2 as a co-monomer.

Preparation of elastomeric material from polv(carborane-co-siloxane)s

Hvdrosilation crosslinkina

To produce a cross-linked material, the sample B300A-2 was crosslinked using 1 ,1 ,3,3- tetramethyldisiloxane, as described in Example 8. The cross-linked material was an elastomeric pure carborane-siloxane material, with over 24%wt boron content. This elastomeric material contains the boron-rich carborane moiety chemically bound within the polymer backbone, homogeneously distributed throughout the material.

Sn-catalvsed crosslinkina

As an alternative route to elastomeric poly(carborane-co-silpxane), elastomeric pads were formed by crosslinking poly(carborane-co-siloxane)s in combination with a

poly(dimethylsiloxane) and tetra(ethoxysilane) in the presence of Sn(ll)(oct)2 following the method described in Examples 9 and 10. Dynamic mechanical analysis (DMA) testing on the pads yielded the stress-strain data shown in Figure 4 from which the compressive Young's modulus was calculated to be -1.17 MPa for Sample B400C-1 and -0.97 MPa for sample B500C-1. The reduction in the compressive Young's modulus upon reducing the carborane content (and hence the %wt boron, i.e. Sample B500C-1 with respect to Sample B400C-1) is as expected, since it is known from the Dexsil series that reducing the carborane content produces a more flexible material with a reduced glass transition temperature [Korshak]. Carborane present in the backbone generally increases the stiffness and/or the viscosity of the bulk material. The components of Samples B400C-1 and B500C-1 yielded an opaque mixture before curing, suggesting that the presence of carborane in the backbone prevents close packing of the polymer chains at the molecular level. Samples B400C-1 and B500C-1 remained opaque after catalysed crosslinking curing. The DMA traces (Figure 4) only show a single compression/relaxation for each material, which is to be expected for a homogeneous material. Hence the opaque nature of the carborane-containing pads is not likely to be due to phase separation of the material's components post curing, but could originate from aggregation or microcrystalline domain formation.

Peroxide crosslinking

As a further alternative route,_elastomeric pads were formed from vinyl-containing poly(carborane-co-siloxane) using peroxide-based crosslinking agents, as described in Example 1 1. Advantageously, this approach allowed the poly(carborane-co-siloxane) polymer chains to bond directly to one another.

Preparation of elastomeric materials via FeCIa catalysed coupling

Dexsil®-200 polymers prepared from 1 ,7-bis(methoxy(dimethyl)silyl)-mefa- dicarbadodecaborane and a dichlorosilane were prepared following literature precedent [Heying]. The procedure is multistep and requires heating the reagents in the presence of FeCI ₃ at 120°C before addition of further FeCI ₃ catalyst and increasing the reaction temperature to 180°C. The catalyst is deactivated during the reaction and hence further portions must be added at intervals to enable further reaction [Dietrich]. The reactions were performed solvent-free under anhydrous conditions using purified dichlorosilanes. Heating was performed using a silicone oil bath that enabled controlled and uniform heating of the reaction flask. Reactions were performed under a positive inert dry nitrogen atmosphere, venting via a bubbler; stages were deemed to have reached completion when no gas evolution through the bubbler was observed when the gas flow to the reaction flask and condenser was briefly closed.

Samples of Dexsil®-200 were prepared from 1 ,7-bis(methoxy(dimethyl)silyl)-m- dicarbadodecaborane 1 and dichlorodimethylsilane 14 with additional portions of FeCI ₃ added before each heating stage (Comparative Examples 1 and 2). After the first heating stage (to ~ 120 °C) the reaction mixture was a clear yellow fluid oil, and NMR analysis showed that a large proportion of Si(CH ₃ )20Me groups remained unreacted and peaks corresponding to reaction intermediates (and quenched chlorosilanes to silanol groups) were observed. A subsequent heating stage(s) (at ~ 180 °C) with additional FeCI ₃ resulted in the reaction mixture darkening and thickening, and reaction of the Si(CH ₃ )20Me groups and the formation of oligomers and polymer as inferred from NMR spectroscopy.

Two-stage (Comparative Example. 1 ) and three-stage (Comparative Example 2) heating regimes were followed to afford waxy brown solid (after cooling to room temperature) products referred to as F200-1 and F200-2 respectively. The products are soluble in common organic solvents (CHCI ₃ , THF, Et ₂ 0) and insoluble in /so-prppyl alcohol, acetone and water, in accordance with the published observations. The solubility in common organic solvents implies the formation of linear (or only lightly cross-linked) polymer. Since the reaction is performed solvent-free and above the boiling point of dichlorodimethyisilane (b.p. = 70 °C at s.t.p.) hot all of the dichlorodimethylsilane is in intimate contact with the catalyst and the carborane precursor or the forming oligomers and polymer, and hence the isolated yield is less than quantitative. Unreacted chlorosilane hydrolysed readily upon exposure of the product to atmospheric moisture, resulting in an acidic crude product. NMR spectroscopy of F200-1 in CDCI ₃ gave broad spectra, likely to be due to chloroform-insoluble impurities in the product (e.g. ionic salts) and residual iron- containing compounds. Crude polymer of F200-2 was dissolved in diethyl ether, and the organic solution washed with water to purify the product, which was recovered by removal of the organic solvent under reduced pressure to afford a sample referred to as F200-3. Peaks in the NMR spectra of the crude product (0.52, 2.5 and 24.1 p.p.m in the ¹ H-NMR, ¹³ C- and ²⁹ Si-NMR spectra respectively) were effectively removed via this washing process.

The ¹ H-NMR peak at the chemical shift of 0.26 p.p.m. corresponds to the methylsilane protons in a (mefa-carborane)-Si(CH ₃ )20H end group. The corresponding resonance of the hydroxyl group is too broad to be visible in the ¹ H-NMR spectrum itself (as is the case for pure 1 ,7-bis(hydroxyl(dimethylsilyl)-m-carborane, 3, most likely do to rapid proton exchange), and the ¹³ C- and ²⁹ Si-NMR resonances at -0.4 and 10.0 p.p.m. respectively are in agreement with this assignment. Hence the purified polymer of Comparative Example 1 contains significant amounts of both methoxysilane and hydroxysilane end groups, which correspond to a low polymer molecular weight.

The ³ C- and ²⁹ Si-NMR spectra of the crude product before purification do not show evidence for the (/T7efa-carborane)Si(CH ₃ ) ₂ OI-l end groups, which shows that the hydroxyl- terminated end groups form after the reaction during work up. The reaction was performed under anhydrous conditions until workup and purification, which involved an aqueous wash under acidic conditions. Unreacted chlorosilane groups will hydrolyse to silanols, and the acidic conditions may also hydrolyse some of the remaining methoxysilane end groups.

The B-H groups of the mefa-carborane cage give a broad multiplet in the chemical shift range 3.2-1.4 p.p.m. (in CDCI ₃ ) in the boron-coupled H-NMR spectrum.

The ¹¹ B-NMR of F200-1 is similar to that for the starting material,

1 ,7-bis(methoxy(dimethyl)silyl)-/n-dicarbadodecaborane, with four apparent broad singlets in a 2:2:4:2 ratio as expected for the four environments of boron in the mefa-carborane cage. ¹⁰ B and ¹¹ B are quadrupolar nuclei occurring in 19.6% and 80.4% natural abundances with intrinsic nuclear spin of 5/2 and 3/2 respectively. Due to the multiple couplings between adjacent boron nuclei and the many permutations of boron isotopes adjacent to each cage ¹¹ B nucleus the ¹¹ B-NMR spectrum of the cage appears as broad peaks in the ratio of the different boron environments.

The NMR spectra of F200-1 show the presence of Dexsil® 200 and Dexsil® 100 fragments. Integral analysis of the ¹ H-NMR spectrum enables an estimate of the empirical structure to be calculated, considering that there was an equal number of (m-carborane)Si(CH ₃ )20Me and (/77-carborane)Si(CH3)20H end groups within the sample analysed (Table 6). The majority of the F200-1 product contains Dexsil® 200 fragments, with about 25% Dexsil® 100 content by fragment. No Dexsil® 300 fragments were observed in product F200-1.

F200-2 was prepared in the same manner as F200-1 with addition of further FeCI ₃ (~ 1mol% based on 1 ,7-bis(methoxy(dimethyl)silyl)-mefa-dicarbadodecaborane starting material) before a third heating stage (see Comparative Example 2).

The ¹ H-NMR of F200-2 in CDCI ₃ shows only a trace quantity of (m-carborane)Si(CH ₃ ) ₂ OCH ₃ end groups present (-Si(CH ₃ )20CH ₃ , with a ¹ H chemical shift of 3.48 p.p.m.), with corresponding resonances in the ¹³ C- and ²⁹ Si-NMR spectra (at chemical shifts of -2.3 and 10.1 p.p.m. for -Si(CH ₃ )20CH ₃ and -S/(CH ₃ ) ₂ OCH ₃ respectively).

There is no evidence of hydroxysilane-terminated (i.e. (m-carborane)Si(CH ₃ )20H) end groups by H, ¹³ C and ²⁹ Si-NMR spectroscopy. This, in conjunction with the reduced methoxysilane content, shows that the polymerisation has progressed further to completion for F200-2 than for F200-1 , and hence F200-2 has a higher molecular weight with fewer unreacted methoxy groups from the starting material. Aqueous work up of F200-2 did not cause hydrolysis of residual chlorosilane to form terminal silanol groups, as was observed in F200-1.

Integral analysis of the ¹ H-NMR spectrum of F200-2 confirms the greater molecular weight, assuming that each polymer chain is terminated by (/nefa-carborane)Si(CH ₃ )20Me. The polymer is predominantly comprised of Dexsil® 200 fragments, with about a third Dexsil® 100 content by fragment (Table 6).

The commercially-supplied sample of 'Dexsil® 200', referred to herein as F200-4

(Comparative Example 4) does not contain (m-carborane)Si(CH ₃ )20Me end groups but is terminated by (/77-carborane)Si(CH ₃ ) ₂ OH silanol groups, by reference to the ¹ H, ¹³ C and ²⁹ Si- NMR spectra of silanol- and methoxysilane-terminated m-carborane (Figure 5). F200-4 also contains a substantial proportion of the Dexsil® 300 fragment that is not present in purified F200-1 or F200-2; integral analysis of the ¹ H-NMR spectrum allows an empirical structure of the F200-4 material to be calculated (Table 6).

Table 6

The commercially-supplied F200-4 has a higher molecular weight than the material produced from the two stage process, but a lower molecular weight than the material from the three stage process. This is in agreement with integral analysis of the ¹ H-NMR spectra and the calculated end group content (Table 6), and confirms that the polymer molecular weight can be increased by heating the reaction mixture for longer with additional catalyst. Three heating stages also appears to narrow the polydispersity index (PDI, i.e. M _w /M _n ) from 2 to 1.3 for the main peak observed in the GPC data.

Sample F200-3, which was obtained by subjecting sample F200-2 to an additional purification step (extraction sequentially with acetone, 10% acetone/water and acetone, Table 2 entry 4) gave an identical result to the unextracted material; this demonstrates that the washing process did not affect the material's molecular weight distribution. This is important since it confirms that there is not a soluble low molecular weight fraction that is readily solvent-extracted, which would change the nature of the polymer distribution.

Differential scanning calorimetry of F200-1 and F200-3 shows clear glass transition temperatures (Figure 7). The glass transition temperature of F200-1 is similar to that of F200-4, whereas the glass transition temperature of F200-3 is a significant 5 °C higher (Figure 7 and Table 8). The data for F200-1 above the melting temperature shows some peaks that are likely to be due to solvent and other volatiles incorporated within the product being driven out of the material. From the above analysis and discussion F200-1 is known not to be the product of a complete reaction, and therefore a small proportion of low molecular weight content is expected.

Table 8

An increase in the glass transition temperature is evidence that there is less backbone flexibility within the polymer, and this is consistent with the polymer compositions as determined by integral analysis of the ¹ H-NMR spectra. F200-1 contains predominantly the Dexsil® 200 repeat unit, with about 25% of the carborane content being the carborane-richer (and hence less flexible) Dexsil® 100 fragment. F200-3 contains about 30% of the carborane in Dexsil® 100 fragments, and it is this increase in the proportion of the more rigid Dexsil® 100 content that results in the increased T _g . Although glass transition temperatures reported under different conditions may not be directly comparable, the lower glass transition temperature of F200-4 can be explained by the Dexsil® 300 content which acts to increase the backbone flexibility.

Synthesis of Dexsil® polymer incorporating vinylsilane (10 mol% of carborane content) was performed using a two stage heating process (Comparative Example 5). NMR spectroscopy of the product, referred to as F200A-1 , shows that all of the vinyl groups have reacted to form a saturated -CH ₂ CH ₂ - group, as expected under the reaction conditions, resulting in light cross-linking between the relatively short polymer chains. The waxy brown material produced is soluble in common organic solvents but was tacky which is consistent with light cross-linking. This shows that, in contrast to reactions carried out according to the present invention, the FeCU-catalysed reaction is unable to produce products with pendent vinyl groups for subsequent reaction.

In summary, a range of poly(carborane-co-siloxane)s have been produced according to the method of the present invention, and compared to polymers produced according to the conventional FeCb method. Characterisation of the products shows that the method of the present invention allows good control over the molecular weight of the polymer product, and greater flexibility over the functional groups which can be incorporated in the starting materials. Furthermore, the method of the present invention is more selective than the FeCI ₃ catalysed route over the repeat unit of the polymer. In addition, the reactions can be carried out under mild conditions without the need to carefully control ambient conditions (air and moisture content, and temperature).

REFERENCES

Altmann, S.; Pfeiffer, J.; Monatshefte fur Chemie, 2003, 134, 1081-1092.

Chojnowski, J. et al.; Organometallics 2005, 24, 6077-6084.

Dietrich, H. J. er a/.; T. L. Makromolekulare Chemie 1974, 1 5,425-440.

Ervithayasuporn V.; Chimjarn, S.; Inorg. Chem. 2013, 52, 13108-13112.

Gelest Inc., Reactive Silicones: Forging New Polymer Links. Supplement to Gelest Catalogue Silicon, Germanium & Tin Compounds, Metal Alkoxides and Metal Diketonates; 2004.

Gomez, F.A., et al J. Am. Chem. Soc, 1991 , 113, 5915-5917 Gonzalez-Campo, A., et al., Macromolecules, 2007, 40, 5644-5652

(a) Harris, R. K ef a/.; Pure Appl. Chem. 2001 , 73, 1795-1818;

(b) Harris, R. K et a/.; Pure Appl. Chem. 2008, 80, 59-84.

Harwood, J. A. C; Payne, A. R.; J. Appl. Polym. Sci. 1966, 10, 315-324;

(a) Hedaya, E. er a/.; US Patent No. 4,145,504, published 20 March 1979.

(b) Hedaya E. et al.; J Polym Sci Polym Chem Ed 1977;15(9):2229-2238. Heying, T. L. et al.; US Patent No. 3,388,091 , published 11 June 1968. Kahlig, H.; Mayer-Helm, B.X. J. Chrom. A, 2006, 1131, 235-241. Knollmueller, R. N. ef a/.; J. Polym. Sci. PartA-1 1971 , 9, 1071-1088. Korshak, V.V. et al.; J. Polym. Sci. Part C - Polym. Lett., 1979, 17, 115-120. Li, L. et al.;J. Colloid and interface Sci. 2013, 406, 30-36.

Lide, D. R., Ed. Solvents for Liquid Chromatography. In Handbook of Chemistry and Physics, 75th Edition; CRC Press; Boca Raton, Florida; 1994; p. 8-64. Reprinted from Bruno, T. J., Svoronos, P. D. N., CRC Handbook of Basic Tables for Chemical Analysis; CRC Press; Boca Raton, Florida; 1989; p. 89. Mohadger, Y. et al.; J. App. Poly. Sci. 1973, 14, 2635-2652. Mullins, L. Rubber Chem. Technol. 1969, 42, 339-362.

Patel, M., Swain, A. C. Polymers Incorporating Icosahedral C\oso-Dicarbaborane Units. In Macromolecules Containing Metal and Metal-Like Elements, Vol. 8, Boron-Containing Polymers; Abd-EI-Aziz, A. S., Carraher (Jr), C. E.; Pittman (Jr), C. U., Zeldin, M., Eds.; Wiley Interscience; New Jersey; 2007; pp. 77-102.

Rubin, M. et al.; J. Org. Chem. 2002, 67, 1936-1940.

Stewart D.D., et al.; Macromolecules 1979; 12(3): 373-377.

Thompson, D. B.; Brook, M. A. J. Am. Chem. Soc. 2008, 130, 32-33.

Wakita, K.; Shirahata, A.; US Patent Application US 2002/0115878, published 22 August 2002.

Zhang, X. et al.; Polymer 2011 , 52, 4777-4784.