NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

Institute of Medicine (US) Committee on the Safety of Silicone Breast Implants; Bondurant S, Ernster V, Herdman R, editors. Safety of Silicone Breast Implants. Washington (DC): National Academies Press (US); 1999.

Cover of Safety of Silicone Breast Implants

Safety of Silicone Breast Implants.

Show details

2Silicone Chemistry

Silica, Silicon, and Silicone

Silica is the most common substance on earth. It is a constituent of most rocks. Beach sand is almost pure crystalline silica, as is quartz, which in its purest form is a clear or rosy-colored gemstone, found in geodes, or, if less pure, may be found as amethyst, agate, flint, or "petrified wood." The molecular formula of silica is SiO 2, silicon dioxide. Silicon dioxide is a three-dimensional network of silicon (Si) atoms linked by oxygen (O) atoms in a 2:1 ratio; each silicon atom is linked to four oxygen atoms, and each oxygen to two silicon atoms.

A crystalline substance is one whose atoms form a regular pattern over large distances. This regularity is usually measured by the diffraction of x-rays. The constructive and destructive interference of x-ray waveforms causes the x-ray beam to be redirected into a reproducible pattern that can be detected by photographic film. The characteristics of this pattern allow calculation of the precise atomic spacing. This regularity is also a key to the hardness and strength of most crystalline substances.

Although human exposure to crystalline silica is extensive and generally to no ill effect, tissue (especially lung) exposure to particulate silica or silica dust has well defined toxic, inflammatory outcomes (American Thoracic Society, 1997). Silica is also found in less toxic, amorphous forms (Warheit et al., 1995). Amorphous materials, in which the atoms are not found in regular arrays even though the atomic ratios are the same, do not give crystalline patterns. Amorphous forms of silica include a vitreous form like the glass used in higher-wattage halogen lamps, (e.g., automobile headlamps). Window and bottle glasses are diluted, low-melting forms of silica. Sodium and calcium oxides are used as diluents in sodalime window and bottle glass. Silica aerogel, silica smoke, fumed silica, and precipitated silica are names for amorphous silica powders that are important constituents of medical rubber-like goods, including breast implants. In fact, amorphous silica is used in almost all silicone elastomers, as well as in special-purpose isoprene (natural rubber) elastomers. Less often used is diatomaceous earth, the silica skeletal residue of diatoms, which are microscopic sea creatures (Heaney et al., 1.994; Iler, 1979a,b Iler, 1981). Some forms of amorphous silica have been approved by the Food and Drug Administration (FDA) (D. Benz, FDA, personal communication, 1998) for use in pharmaceuticals and food; they are widely distributed in foods and foodstuff manufacture (Villota and Hawkes, 1986).

Silicon is a semimetallic element, located just below carbon in the periodic table, that is not found in nature in its elemental form. It is perhaps best known as the shiny semiconducting metalloid used to make computer chips. Silicon can be made by heating silica with carbon (coke or charcoal), typically in an electric arc furnace. At high enough temperatures, the elements silicon, carbon, and oxygen can exchange places, and the driving force of the reaction is the loss of gaseous carbon dioxide (CO2) leaving silicon and any excess carbon behind. The impure reaction product can be used to make silicones (see below) but requires extensive purification before it can be fabricated into computer chips. Although silicon is in a high-energy, unstable state with respect to its oxide, silica does not form spontaneously in air on its surface, as rust does on iron or aluminum oxide on aluminum. Like gold or platinum, silicon retains its shiny metallic appearance and electrical properties. However, silicon can burn in air to give a thick, white smoke of amorphous silica (LeVier et al., 1993).

Silicone refers to a large family of organic silicon polymer products with a main chain of alternating silicon and oxygen atoms. Typically, each silicon in the chain carries two methyl groups (CH3-, which can also be written as Me-, where C = carbon and H = hydrogen), and the material is called poly(dimethylsiloxane) (PDMS):

Image img00000.jpg

The tilde (˜) at the chain end implies that the sequence is repeated, typically with hundreds to thousands of silicon-oxygen links. As the number of units increases from two to hundreds or thousands, the compounds formed have very different properties: for example, hexamethyldisiloxane has a viscosity (0.65 centistoke [cS]) less than water (1.0 cS) and is absorbed from the gastrointestinal tract, whereas compounds with 3,000 units are relatively inert biologically and are solids having viscosities of millions of centistokes (see Chapter 4 and Table 2-1).

TABLE 2-1. Viscosity of Silicone Compounds.


Viscosity of Silicone Compounds.

Chemistry of Silicones

Silicone is made by the reaction of dimethyldichlorosilane (Me2SiCl 2) with water to give PDMS and hydrochloric acid (HCl):

Image img00001.jpg

The Me2Si(OH)2 first formed polymerizes spontaneously to silicone hydrolysate HOMe2SiO(Me2SiO)nMe2SiOH, where n can vary from 0 to 50, depending on reaction conditions. Cyclic compounds are also formed in greater or lesser amounts depending on conditions.


Me2SiCl2 is made from impure silicon in the "direct process." Silicon powder is heated in a stirred or fluidized bed with gaseous methyl chloride (which, in turn, is made from wood alcohol and hydrochloric acid) to produce a mixture of chlorosilanes.

Image img00002.jpg

The mixture can be separated into its components by distillation of the liquid chlorosilanes. This provides sufficient purity for many applications. Me2SiCl2 is hydrolyzed to PDMS, which is known as crude hydrolysate at this stage (see above). Me3SiCl is hydrolyzed to Me3SiOSiMe3, and MeSiCl3 hydrolyzes to MeSiO3/2, called "T-gel," a cross-linked hard material that seems closer to sand than to silicone rubber. The 3/2 in the formula refers to the three oxygens each shared with two silicon atoms. The ratio, of 1.5 oxygen atoms per silicon atom, implies that three bonds are formed to other silicon atoms, since silicon is always tetravalent. Thus, a three-dimensional network, similar to the silica network, results, but the material is usually softer because of the valences occupied by methyl groups.

SiCl4 hydrolyzes to a hydrated silica also known as silicic acid (a polymerized form of Si(OH)4), which forms the drying agent silica gel when oven dried. It is a glassy, amorphous structure that has shrunk with the loss of water to give a porous, spongy structure. SiCl4 is also hydrolyzed as a gaseous stream with steam. A high temperature flame of this tetrachlorosilane reacting with steam gives a white amorphous silica smoke or aerogel, which is gathered to provide the very finely divided filler used in silicone rubber.

Functionality and Nomenclature

Another shorthand polymer nomenclature is useful as well. Me3SiOSiMe 3 is also known as M2 in a popular shorthand that denotes Me3SiO1/2 as M, a monofunctional polymer component or "chainstopper" because it regulates polymer chain length. The 1/2 subscript indicates that only one-half of the oxygen belongs in the group, since it is shared by two silicons. In practical terms, the M (or monofunctionality) means that silicon binds to only one oxygen, and the form of the whole molecule can be deduced from this and the following information. The D in D4 or in MD2M refers to the difunctional dimethylsiloxane unit Me2SiO, the polymer building block that can add to itself to form enormously long chains known as high polymers. PDMS is made up almost entirely of D units. Since D means that silicon binds to two oxygens (i.e., is difunctional), the silicon at one end of the polymer chain must bind to the oxygen at the other end to form a circular (cyclic) molecule (as in D4) unless the last silicon has three methyl groups attached and is, therefore, a chainstopper, giving a chain such as MDnM (where n can be any number). The T in T-gel refers to the trifunctional MeSiO3/2 which, when present as an impurity, leads to branching of polymer chains. However, its overall effect on molecular weight can be controlled to some extent by the addition of more M units, since chainstopping one of the oxygen-silicon units makes MT which is equivalent to D. The so-called Q units are quadrifunctional and result from hydrolysis of silicon tetrachloride (SiCl4). Q can be modified with M units since M2Q is equivalent to D because two of the oxygen-silicon units are chainstopped. The various mixtures of M, D, T, and Q control molecular weight, branching, and molecular shape and are used to formulate various types of silicone resins (varnishes, fiberglass bonding solids, pressure-sensitive adhesives, and even the release paper for protecting adhesive tapes). The symbol L is also sometimes used to denote D units in linear polymers, with Dn reserved for cyclics (see below). Thus, L6 would be a linear hexamer (the same as MD4M) with no indication of the chemical groups at the ends of the chain. The formation of cyclic, branched, or linear compounds and the substitution of other groups for methyl will change the physical and biological properties of silicone molecules, often to a very great extent (LeVier et al., 1993).

High Polymers

Pure D polymer units, without M, T, or Q, give the highest molecular weight unbranched polymers and thus are most useful for elastomer manufacture. Crude hydrolysate was once used for elastomers by polymerization with acid.

Image img00003.jpg

The catalyst used was usually sulfuric acid, removed from the high molecular weight PDMS by washing with water. Alternatively, acid ion-exchange resin or acid clay was used. The resin or clay could be removed by filtration in low-viscosity products or by neutralization. The limitation of polymers made from hydrolysate was the variable concentration of M, T, and Q groups, leading to uncertain properties since varying amounts of these groups caused more or less cross-linking, branching and chain length which affected molecular weight and physical properties of the final silicone product.

Methyl Tetramer (D4)

The solution to this problem was the development of cyclic silicones, usually tetramers, as intermediates that could be highly purified. Tetramer is distilled from a kettle containing hydrolysate and a basic catalyst such as potassium hydroxide. The base forms the potassium silanolate ion via the removal of water. The silanolate ion is a powerful "rearrangement" catalyst, causing siloxane molecular bonds to exchange parts with each other. Some cyclics, mostly tetramer, are formed in a random fashion and can be distilled off. Removal of cyclics upsets the equilibrium, so more cyclics are generated, and the process continues until all of the polymer is converted to volatile cyclics. The tetramer can then be fraction-ally distilled to a very high degree of purity, eliminating the interfering M, T, and Q units. One of the most sensitive measures of purity is the molecular weight of a polymer made under controlled conditions. Industrial-grade tetramer can be polymerized to a molecular weight greater than several million Daltons. For even higher purity, a technique called zone refining has been used to form polymers of more than 40,000,000 Daltons (Martellock, 1966).

The D4 tetramer is an eight-membered ring of alternating silicon and oxygen atoms with eight methyl groups attached, two per silicon (octamethylcyclotetrasiloxane):

Image img00004.jpg

The polymerization of this cyclic tetramer is somewhat different from the condensation reaction that leads to hydrolysate and higher polymers by splitting off water molecules from pairs of silanol groups. The tetramer D4 has no silanol groups, but polymerizes by a process known as ring-opening polymerization. It is essentially the same process that led to generation of tetramer, except the equilibrium now favors high polymer, yielding about 86% high polymer with 14% mixed cyclics (Brown and Slusarczuk, 1965; Carmichael and Winger, 1965).

MD2M, a short chain of four silicon atoms alternating with three oxygen atoms and fully substituted with methyl groups, is the chainstopper of choice in industry because it matches the reactivity of D4. The two are heated together in the presence of a trace of basic or acidic catalyst to yield a linear polymer whose molecular weight, which controls the viscosity, is regulated by the small amounts of MD2M. This polymerization is called equilibration because it leads to an equilibrium mixture of linear and cyclic silicones, about 14% cyclics. The latter are mostly removed by steam or vacuum distillation after the catalyst is removed. The molecular weight of crude hydrolysate is controlled by the chainstopping effect of dimethylsilanol (-SiMe 2OH) groups, but these are unstable since further polymerization is possible under some conditions. The silanol (-SiOH) groups can condense, losing water and increasing chain length. Conversely, siloxane (-SiOSi-) groups can, in theory, revert in the presence of water to give pairs of silanol groups, thereby decreasing chain length. Both reactions can occur under mild acidic or basic conditions. However, the reversion reaction is inhibited by the lack of solubility of water in PDMS, i.e., the physical inaccessibility of water to siloxane, so for practical purposes it does not present a problem (Martellock, 1966).

Uses of Silicones

PDMS polymers are useful for hundreds of applications (Silicones Environmental Health and Safely Council, 1994). Low molecular weight oils (fluids) are used as (1) lubricating oils (Slipicone), (2) lubricants for syringe needles and barrels, (3) substances to improve the "hand" of fabrics and give them water repellency, (4) skin cream modifiers in cosmetics, (5) antifoam agents in the food and chemical industry, and (6) cures for stomach gas (Simethicone in Gas-X). Higher molecular weight, more viscous oils are widely used as high-temperature hydraulic and brake fluids. A list of silicone containing medical devices includes hydrocephalus shunts, foldable intraocular lenses, soft tissue implants for congenital and cancer reconstructive surgery, cardiac pacing and defibrillation devices, implantable infusion pumps, elastomeric toe and finger joints, incontinence and impotence devices, infusion ports, larynx implants, tissue expanders, and many shunts and catheters (Compton, 1997).

Silicone oil (fluid) has a very low surface energy, which causes it to spread on higher-energy surfaces and to make these surfaces water repellent. Silicone resins, highly cross-linked with T or Q units to give the required hardness, also have this effect and are used to coat plastic eyeglass lenses, or glass bottles with a very thin film to increase scratch and break resistance and aid in emptying aqueous contents. Similar resins are used as release agents in commercial bakeware. Silicone rubber films are used to coat rough-service light bulbs to increase break resistance and to capture shattered fragments. There are actually thousands of silicone products that impinge upon all aspects of modern life (Silicones Environmental Health and Safety Council, 1994).

Silicone Elastomers

Image img00005.jpg

Three broad categories of silicone are used in implant manufacture: (1) platinum-cured (gel or LSR [liquid silicone rubber]); (2) RTV (room-temperature vulcanized) rubber; and (3) gum-based peroxide-cured (heat-cured) rubber. All require a final oven bake to attain optimal purity, stability, and physical properties (Lynch, 1978). The simplest of these gels is the platinum-cured gel that expanded most early implants. The original embodiment was developed for a plastic surgeon, Thomas D. Cronin, by Dow Corning. It consisted of a slightly vinyl substituted PDMS fluid that was cross-linked with a hydrogen-containing PDMS fluid in a platinum-catalyzed reaction. This means that a few silicon atoms in the chain had a vinyl (Vi) instead of a methyl substituent. The vinyl group was susceptible to bonding with a receptive hydride group on a neighboring siloxane polymer chain, creating a cross-link. This curing reaction is known as hydrosilation. It forms a very lightly cross-linked unfilled elastomer that gives the desired softness and compliance. In chemical terms, each silicon hydride group (the ''hydrogen-stopped PDMS" below) adds across a vinyl double bond (CH2 = CH-in the MeVi tetramer below), thereby converting the vinyl group to an ethylene (-CH2-CH2-) bridge linking two polymer molecules together. Since a few of the polymer molecules had more than two reactive groups per molecule, the reaction results in a cross-linked system with no new soluble or leachable components.

Silicon Addition Cure Chemistry

The Cronin gel probably had a vinyl level less than 0.1 mole %, which was reacted with insufficient silicon hydride so that the hydrogen would be completely used up and not cause problems with hydrogen gas evolution at a later time. Therefore, the final gel had an even lower level of excess vinyl, and the cross-link density was controlled by the concentration of hydride groups added. See Table 2-2 for the composition of several versions of the Dow Corning gel.

TABLE 2-2. Dow Corning Silicone Implant Gel, Platinum Cure.


Dow Corning Silicone Implant Gel, Platinum Cure.

Certain silicone fluids, chainstopped with alkyltriacetoxysilane and thickened with amorphous silica filler, cure to RTV silicone rubbers when exposed to moist air. The triacetoxy groups are hydrolyzed by the moisture and produce T groups at the chain ends, which react with each other to form cross-links and thus a rubbery structure. The hydrolysis also produces acetic acid, which catalyzes the cross-linking and accounts for the vinegary smell of this material. A tin soap such as stannous oleate or octoate is usually used as a catalyst to speed the cure of water-cured, condensation-type silicone adhesives. The filler used to confer strength is amorphous silica. It is usually treated to control its surface reactivity because the latter might impinge on the shelf life of the product. RTVs are used as adhesives and sealants in assembling the implants. Table 2-3 gives details of the composition of Dow Corning RTV.

TABLE 2-3. Silicone Adhesive to Seal Injections Sites, RTV Cure.


Silicone Adhesive to Seal Injections Sites, RTV Cure.

The filler treatment involves silica aerogel filler. This filler, whose manufacture is outlined, is amorphous silica with a surface high in silanol groups. In a nonpolar environment such as silicone gum, these silanols tend to bond between filler particles, causing aggregation. Although the primary particle size is ca. 5-7 nm (LeVier et al., 1993), this aggregation can form long chains and/or agglomerates of filler particles. The agglomerated filler acts as a second cross-linking mechanism, leading to a large increase in the stiffness of an uncured gum or the hardness of a cured elastomer, clearly an unwanted outcome (Boonstra et al., 1975). Two methods have been developed to control this—process aids and filler treatment.

Process aids are typically very low molecular weight silanol-terminated silicone fluids that are added to the gum or filler mixture. The silanols in the process aid interact with those on the filler surface to give a stable coating and thus prevent filler agglomeration. This is also called in situ filler treatment. The coating molecules are held in place by a phenomenon known as hydrogen bonding, a moderately strong association of two oxygen atoms (of the silanol OH groups) held together by a shared hydrogen atom. Filler treatments similarly passivate the filler surface silanols. They may also rely on hydrogen bonding, (i.e., treatment of filler with D4 vapor in a kettle). However, it is preferable and more reliable to bond M groups directly to silanol groups on the filler surface. Hexamethyldisilazane is a reagent that can bring this about. The resulting treated fillers can then be mixed more easily into elastomer compounds and yield stable products. Dow Corning has disclosed the use of hexamethyldisilazane as an in situ filler treatment where it is added to the polymer along with the untreated filler. About 60% of the silanol groups react, but the resulting coverage is complete by an "umbrella effect" because the remaining silanols are not free to interact.

High-molecular-weight, viscous silicone gums filled with amorphous silica (Cab-O-Sil or Aerosil) are mixed with process aids and with peroxides such as 2,4-dichlorobenzoyl peroxide and cured under heat and pressure in metal molds. The molds help to protect the elastomer from cure inhibition by oxygen in the air, as well as to control thickness and shape (Lynch, 1978). Typically, all three types of silicone elastomers are heated or "postcured" in circulating air ovens when necessary to improve purity and stability. All three types of elastomeric materials have been used to produce silicone breast implants, joint implants, surgical drains, pace-maker covers, indwelling catheters, and the like (Batich and DePalma, 1992; Batich et al., 1996; Clarson and Semlyen, 1993; Kennan and Lane, personal communication, 1998).

Silicone Breast Implants

The chemistry of silicone elastomer and gel in a silicone breast implant described here is based substantially on disclosures by Dow Corning Corporation and is therefore only an example of one version of this technology. Details of implants of other manufacturers may be substantially similar but are proprietary, although similar processes have been described by NuSil chemists (Compton, 1997). Additional details of Dow Corning and other implants are discussed in Chapter 3 of this report (Lane and Bums, 1998; T. H. Lane and J. J. Kennan, personal communications, 1998; T. H. Lane et al., personal communication, 1998). The earliest Dow Corning shells were made of high molecular weight polymer (gum) filled with amorphous silica and a process aid designed to passivate the filler surface. They were mixed with 2,4-dichlorobenzoyl peroxide and cured in a heated mold. Two shell faces, front and back, were produced and then glued together to make a seamed shell. This heat-cured elastomer system was used only in the early Dow Corning implant shells designed for Dr. Cronin and in patches for sealing them; Table 2-4 lists the details. The uncured gel mixture was injected though a hole and cured in place with added heat. The hole was sealed with a patch or sealant, and the final product was thoroughly oven baked to remove all traces of volatile reactants and reaction products, especially 2,4-dichlorobenzoic acid, which may be toxic (a residue of 20 parts per million remained; T. H. Lane et al., personal communication, 1998). The gel used was the previously described platinum curing material, designed to be as soft as possible, with barely enough strength to maintain its shape (see Table 2-2). This gel was responsible for the tactile feel of the implant, roughly approximating the feel of human adipose tissue. The shell helped the gel to maintain its shape.

TABLE 2-4. Cronin Seamed Shell Elastomer, Peroxide Cure.


Cronin Seamed Shell Elastomer, Peroxide Cure.

Both the shell and the gel were modified for a softer feel in the 1970s (see Chapter 3). The shell could be made thinner and of more uniform thickness by using platinum-cured liquid silicone rubber (LSR) and a dip-coating process similar to the process used to form surgical gloves. In the dip-coating process, the shell thickness is controlled by the viscosity and draining rate of the LSR mixture and therefore by its dilution with solvent (Table 2-5). The shell was formed in one piece using a breast-shaped mandrel. The LSR was diluted to 15% solids with trichloroethylene (TCE) solvent, dipped and drained, dried, heat-set and peeled off. The large hole on the back left by the mandrel shaft was covered with an elastomer patch material (Table 2-6). This elastomer also had an amorphous silica filler whose silanol surface reactivity had been masked with M groups, probably by reaction with hexamethyldisilazane, Me3SiNHSiMe3, the reactive nitrogen analogue of M2. The liquid gel precursor was also modified by the addition of 80% by volume of low molecular weight (1,000 cS, 25,000 Daltons) nonreactive PDMS fluid to swell the gel and make it even softer when cured (Table 2-2). This effect is analogous to the swelling of a sponge in water. Added low molecular weight silicone fluid is not chemically attached to breast implant gel. The network gel, 20% of the material mass, is a single giant molecule that is swollen by low molecular weight fluid, which compromises 80% of the mass, that can move or be extracted like water in a sponge.

TABLE 2-5. Implant Sealing Patch Rubber, Platinum Cure.


Implant Sealing Patch Rubber, Platinum Cure.

TABLE 2-6. Seamless Dip-Coated Shell Elastomer, Platinum Cure.


Seamless Dip-Coated Shell Elastomer, Platinum Cure.

Barrier-Layer Implants

The movement or diffusion of silicone gel fluid was addressed by Dow Corning with a fluorosilicone elastomer barrier-layer shell. Other manufacturers have used different technologies. A copolymer of trifluoropropylmethylsiloxane (F3PrMeSiO) and methylvinylsiloxane was mixed with a solvent, a silica aerogel filler (in situ treated with a process aid), an alkynol cure inhibitor, MeHSiO, and platinum complex catalyst (see Table 2-5). This was dip coated at 2.5% solids to form a very thin interior layer. A second layer approximating the earlier single shell layer was then added for strength. The shell was finished as before. The function and performance of this and other barrier layers are discussed in Chapter 3.

The toxicology of silicone and silicone breast implants is also discussed there and in Chapter 4, and the biological and physical properties of some other silicone compounds are noted. As in any large chemical family, some organic silicone compounds are very toxic, some have biological activity, and some are relatively inert and do not have significant biological activity (LeVier et al., 1995).

Copyright © 2000, National Academy of Sciences.
Bookshelf ID: NBK44788


Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...