Development of a New Poly(Dimethylsiloxane) MaxillofacialProsthetic Material

Tariq Aziz,1 Mark Waters,1 Robert Jagger2

1 Departments of Dental Health and Biological Sciences, University of Wales College of Medicine,Heath Park, Cardiff, United Kingdom

2 Adult Dental Health, University of Wales College of Medicine, Heath Park, Cardiff, United Kingdom

Received 3 December 2001; revised 13 August 2002; accepted 20 August 2002

Abstract: Maxillofacial prosthetic materials are used to rehabilitate patients suffering fromacquired or congenital facial disfigurement. Silicone rubber materials [poly(dimethylsiloxane)rubbers] are the materials of choice. However, current formulations have inadequate me-chanical properties, in particular, tear strength. The purpose of this study was to formulate anew poly(dimethylsiloxane) material with improved mechanical properties in comparison tocommercially available materials. The formulations developed were based on condensationchemistry containing a hydroxy–end-blocked poly(dimethylsiloxane), hydrophobic silica filler,silane cross-linking agents, and a tin catalyst. The following parameters were varied in orderto formulate improved materials: (a) bimodal polymeric mix (i.e., ratio of high- and low-molecular-weight polymers); (b) silica filler type and quantity. By varying these parameters aformulation was developed with a base polymer mix of 80%:20% high- to low-molecular-weight polymers into which 20% w/w filler was added. This formulation had a tear strengthof 22.23 � 1.6 m N/mm compared to the commercial materials having tear strengths in therange 4.53–17.63 N/mm. © 2003 Wiley Periodicals, Inc. J Biomed Mater Res Part B: Appl Biomater 65B:252–261, 2003

Keywords: maxillofacial prosthesis; poly(dimethylsiloxane); tear strength; hardness; molec-ular weight between the cross links

INTRODUCTION

Maxillofacial prostheses are used to rehabilitate patients suf-fering acquired or congenital facial disfigurement. Poly(si-loxane) rubber materials were introduced into maxillofacialtechnology in the 1960s1 and today are the materials ofchoice. Two main types of systems are used within maxillo-facial prosthetics. Poly(dimethylsiloxanes) (PDMS) termi-nated with hyroxyl groups and their subsequent polyconden-sation is the basis of room temperature vulcanizing (RTV)systems in which tin compounds are used as catalysts. An-other RTV or heat-cured system is based on unsaturated(vinyl) terminated poly(siloxanes) that undergo polyadditionreactions with the aid of a platinum catalyst. There are anumber of problems with the current materials, notably lowtear strength, poor color stability, and poor wettability, whichmay lead to abrasion of tissues that surround the prosthe-sis.2–4 It is also important to have a material with adequatetensile properties and appropriate hardness. The tear strength

needs to be adequate, so that very thin margins of the pros-thesis can be produced to blend in with surrounding tissues.It is also important that the material is not excessively hard,and ideally the material should be as skin-like as possible.Finally, the basic strength and flexibility of the material needsto be adequate so that the material can be removed, washed,and handled without any chance of damage. This is mostimportant during the processing of prostheses, particularlywhen materials are removed from molds. Thus there is a needfor improved materials with enhanced properties.

The mechanical properties of a silicone elastomer aredependent on three main factors. The first of these is themolecular-weight distribution, which affects the mechanicalproperties of the elastomer. The blending of long and shortchains of the same polymer gives a broader, bimodal molec-ular-weight distribution, and a network prepared from such ablend is known as a bimodal network.5 The practical signif-icance of such networks is to achieve elastomers possessing acombination of good mechanical properties such as tearstrength, tensile behavior, and resilience.6 A bimodal formu-lation has a local high cross-link density between the rela-tively long chain polymers, blended with a small proportionof short chains prior to vulcanization.5,7,8

Correspondence to: Mark Waters, Department of Dental Health and Biologicalsciences, University of Wales College of Medicine, Heath Park, Cardiff CF14 4XY, UK

© 2003 Wiley Periodicals, Inc.

252

A second factor affecting mechanical properties is theincorporation of a hydrophobic surface-treated silica filler(with dimethyl silyl or trimethyl silyl groups) with a lowparticle size and therefore a high surface area. Facial pros-theses are usually cleaned in water, and therefore the incor-poration of hydrophobic silica filler into the base polymer isa good choice for a medical-grade material. Surface-treatedfillers are also better at dispersing into the silicone polymerand have a reduced base viscosity compared to the baseviscosity of non-surface-treated silica filler system. Underdeformation, these surface-treated fillers help to increase thestrength of the elastomer by allowing the polymer chains touncoil and slide past neighboring filler particles increasingthe crystallization between neighboring PDMS chains. Clin-ically acceptable mechanical properties are only achieved atthe correct filler concentration.9

The third main factor is the degree of cross-linking be-tween polymer chains. An elastomer with a very high cross-link density produces an inelastic brittle material whereas avery low cross-link density would produce a very weakmaterial with very low tear and tensile strength. The correctconcentration of cross-linker is essential for optimum tensileand tear strength of a material.9

Within this study the above factors were used as a basis fordesigning a new maxillofacial material. Hence formulationswere made using varying ratios of high- and low-molecular-weight silicone polymers, a range of surface-treated silicafiller, and an optimal cross-link density.

The main objectives of this study, therefore, were todevelop a new improved maxillofacial material with en-hanced mechanical properties with the use of the rationaleexplained above. In addition, the effect of varying the keyparameter explained above on hardness, tensile, and tearproperties will be assessed and explained in terms of molec-ular weight between cross links.

MATERIALS AND METHODS

Materials Used in Condensation System

The materials used in the development of the condensationcuring system were supplied by ABCR Ltd. and are listed inTable I. The commercial materials used as a comparison inthe study were Cosmesil (HC) and Cosmesil (standard) (Prin-cipality Medical Ltd., Newport, UK), both of which consist ofhydroxy-end-blocked silicone polymer, silica filler, tin cata-lyst, and silane cross-linking agents. In addition, A-2186(Factor II, Lakeside, AZ), Premium Facial and Body Elas-tomer (Prestige Dental, Bradford, UK), and MED-4920 (Nu-sil Technology, Carpinteria, CA) were tested, which consistof vinyl–end-blocked silicone polymers, silica, hydride func-tional cross-linker, and platinum catalyst.

Selection of Ideal Silica Filler

The physical properties of the six different silica fillers usedin this study are shown in Table II. The grades used differed

TABLE I. Materials Used in the Development of the Condensation Curing System

Materials Description Manufacturer

PDMS base polymer C50 Silanol terminated poly (dimethylsiloxane) F.W. � 65,000 Bayer plc, Bayer House,Newbury, RG14 1JA,UK

PDMS base polymer DMS-S21 Silanol terminated poly (dimethylsiloxane), 0.8–0.9% �OH,V: 90–120 m2/s, F.W. � 4200 (Structure same as C50)

ABCR Ltd.,Manchester, M15 6JJ,UK

Catalyst SND3260 Di-nButyldilauryltin C32H64O4Sn ABCR Ltd.Cross-linkers A-172 (a) A-174 (b) (a) 50% vinyl(trimethoxyethoxy)silane � (b) 50%

gammamethacryloxypropyltrimethoxysilaneUnion Carbide UK Ltd.

RickmansworthWD31RB UK.

Silica filler* R104, R106, R202, R972, R974, R812s Degusa Ltd, Cheshire,UK

Note: F.W. � formula weight, V � viscosity.*All silica fillers are described in detail in Table II.

253NEW POLY(DIMETHYLSILOXANE) MAXILLOFACIAL PROSTHETIC MATERIAL

in specific surface area and in the organic groups situated onthe surface of the silica.10 The silica fillers used in thisinvestigation were hydrophobic, and therefore prevent waterabsorption into the cured PDMS elastomer.

Preparation of Polymeric Base

The required concentration of silica filler was mixed with500 g of C50 PDMS base polymer in a planetary mixer for2 h. The cross-linking reaction was achieved by the additionof the tin catalyst (0. 2g per 10 g of base polymer) and crosslinker (0.36 g per 10 g of base polymer) and placed underpressure in plaster dental molds for 24 h at room temperature.A flow diagram of the condensation curing reaction from thestarting materials is shown in Figure 1. Tear test specimenswere prepared for each of the different silica fillers (Table II).

Selection of an Ideal Filler Concentration

The condensation base polymer was prepared, as describedpreviously, with R812s silica filler at six different filler con-centrations, 0, 5, 10, 15, 20, and 25 wt.% in C50 basepolymer. Tear test specimens were prepared for each differ-ent formulation.

Selection of an Ideal Cross-Linker Concentration

The condensation base polymer was prepared at filler con-centration of 20 wt.% R812s in C50 base polymer. Formu-lations were prepared at the following cross-linker concen-trations: 0.12, 0.20, 0.28, 0.36, 0.54, and 0.72 g (per 10 g ofbase polymer) and formulations were cured with the use of atin catalyst (0.2 g per 10 g of base polymer). Tear testspecimens were prepared for each different formulation.

Determination of Molecular Weight between the CrossLinks for Condensation Curing System

Unfilled formulations were prepared at the following cross-linker concentrations: 0.08, 0.12, 0.20, 0.28, 0.36, 0.54, and

0.72 g (per 10 g of C50). All formulations were cured withthe use of tin catalyst (0.2 g per 10 g of base polymer).

The unfilled rubber samples previously prepared for tearanalysis were used for solvent swell studies. Five rubberswell specimens (10 mm � 20 mm � 2 mm) were preparedfor each material.

TABLE II. Physico-Chemical Characteristics of Hydrophobic Fumed Aerosil Silica Fillers (Supplied by Degusa Ltd., Cheshire, UK)

Properties

Silica Fillers

R104 R106 R202 R972 R974 R812s

BETa surface area (m2/g) 150 � 25 250 � 30 100 � 20 110 � 20 170 � 20 220 � 25Mean particle size (nm) 12 7 14 16 12 7SiO2 �99.8 �99.8 �99.8 �99.8 �99.8 �99.8Moisture content (%) �0.5 �0.5 �0.5 �0.5 �0.5 �0.5Transparency in PDMS (%) 27.1 37.0 19.5 19.5 27.0 28.1Hydrophobic treatmentb D4c D4c Silicone oild DDSe DDSe HMDSf

N.B. Taken from Reference 10.aBrunauer, Emmett, and TellerbHydrophobic treatment.cD4 � octamethylcyclotetrasiloxanedSilicone oil � dihydroxypolysiloxaneeDDS � dimethyldichlorosilanefHMDS � hexamethyldisilazane

Figure 1. Flow diagram of the condensation curing reaction from thestarting materials.

254 AZI, WATERS, AND JAGGER

Samples were placed in a desiccator containing phospho-rus pentoxide and calcium chloride until they achieved aconstant weight to an accuracy of 0.0001 g. The specimenswere then placed in glass screw-topped jars containing cy-clohexane (50 ml) and maintained in a water bath at 25 °C.The specimens were reweighed after 5 days; the sampleswere blotted to remove excess hexane and reweighed again toan accuracy of 0.0001 g. The five specimens were air driedand reweighed after 5 days, again to an accuracy of 0.0001 g.

The average weight readings were used to calculate themolecular weight between the cross links, Mc, with the use ofthe following equations:11

swelling ratio, Q �Ws�2

Wd�1; (1)

volume fraction of dried network, � �1

1�Q�2

�1

; (2)

molecular weight between cross links,

Mc/g mol�1 �� V1�2�

13

� � ��2 � ln (1 � �); (3)

where�1 � density of solvent,� 2 � density of dried network � 0.97 g/cm3,8

V1 � molar volume of solvent �108.17 cm3/mol 8

� � Flory interaction parameter � 0.42 8

W0 � mass of rubberWs � Mass of swollen gel at equilibriumWd � Mass of dried network

Addition of Low-Molecular-WeightPoly(Dimethylsiloxane)

The planetary mixer was used to prepare a master batch ofC50 (high-molecular-weight polymer) with 20 wt.% R812sfiller and a master batch of DMS-S21 (low-molecular-weightpolymer) with 20 wt.% R812s. Formulations were preparedby mixing C50 with DMS-S21 (0–50 wt.%) base polymers.The cross-linking reaction was achieved by the addition of tincatalyst (0.4 g per 10 g of base polymer) and cross- linker(0.72 g per 10 g of base polymer). The tear strength, tensileproperties, and hardness were investigated at increasingDMS-S21 (bimodal) concentration.

Determination of Molecular Weight between the CrossLinks for Bimodal Condensation Curing System

Unfilled bimodal formulations were prepared at the followingDMS-S21 concentrations 0, 10, 20, 30, 40 and 50 wt.%. Allformulations were cured by adding tin catalyst (0.4 g per 10 gof C50/ DMS-S21 bimodal base polymer) and cross linker(0.72 g per 10 g of C50/ DMS-S21 bimodal base polymer).

Solvent swell specimens were prepared for each differentformulation.

Testing of Formulations

Hardness Test. Hardness specimens (45 � 45 � 4mm)were made with the use of a conventional dental flaskingtechnique. Five hardness specimens were prepared for eachmaterial.

The hardness test used in this study was based on mea-surements of the indentation of a rigid ball into the testspecimen under specified conditions and conducted accordingto ISO R.48 (1968) (E);12 ASTM D-1415 (1983)13 with theuse of a Wallace Dead Load Hardness Tester (H.W. Wallaceand Co. Ltd., Croydon, UK) with results being read directlyin International Rubber Hardness Degrees (IRHD).

For each specimen, 10 hardness readings were made at 10different positions on the surface of the specimen.

Tensile Test. Tensile specimens (100 � 80 mm and 2mm) were made with the use of a conventional dental flaskingtechnique. Ten dumbbell-shaped specimens (Figure 2) ofrubber material [in accordance with BS 903 (1979) Type 2test pieces]14 were cut from strips of processed material.

The specimens were tested according to BS 903 Part A2(1979) Determination of Tensile Stress Strain Properties.14

The test was carried out on Lloyd Instruments LR10K tensilemachine (Lloyd Instruments Ltd., Fareham, Hampshire, UK)at rate of 500 mm/min.

With the use of the following equations, the softwareautomatically calculated tensile strength and tensile failurestrain:

tensile strength (N m�2)

� (ultimate load/initial gauge section area)

Figure 2. Dimensions of dumbbell tensile test specimen.

255NEW POLY(DIMETHYLSILOXANE) MAXILLOFACIAL PROSTHETIC MATERIAL

tensile failure strain (%)

� 100 � (elongation at break/original gauge length)

Tear Test. Test specimens (100 � 80 mm and 2 mm)were made with the use of a conventional dental flaskingtechnique. Ten tear specimens were cut from the processedmaterial with final dimensions of 50 � 10 mm � 2 mm, witha 4-mm cut placed from one edge. This test specimen is amodification of the test specimen described by ASTM D624:Standard Test Method for Rubber Property—Tear Strength.15

Testing was carried out with the use of the Lloyd Instru-ments LR 10K testing machine fitted with a 1-kN load cell.Specimens were tested at a constant cross-head speed of 20mm/min at a gauge length of 25 mm. On failure of thespecimen the computer software automatically calculated thetear strength by using the following equation:

Ts � F/t,

where

Ts � tear strength (N/mm),

F � maximum load (N),

t � median thickness of specimen (mm).

Statistical Analysis

For each property determined for each of the materials tested,the mean (X) and standard deviations (SD) were calculated. Aone-way analysis of variance, with the Bonferroni methodbeing utilized to produce corrected p values, was used todetermine the statistical nature of the difference in the pa-

rameter mean values for the pairs of materials being consid-ered.

RESULTS

The tear strength obtained from a selection of commerciallyavailable surface-treated silica fillers in C50-based formula-tions are shown graphically in Figure 3. The tear strengthobtained from the silica filler R812s (7.41 � 1.88 N/mm) wassignificantly greater than all the other silica fillers tested (p �0.001) and so became the filler of choice. There was nosignificant difference (p � 0.05) in the tear strength of theremaining silica fillers R104 (3.01 � 0.44 N/mm), R106(4.38 � 0.49 N/mm), R202 (3.01 � 0.31 N/mm), R972(3.41 � 0.19 N/mm), and R974 (3.97 � 0.36 N/mm).

The tear strength of C50 formulations containing an in-creasing concentration of silica filler R812s is shown inFigure 4. There was no significant difference (p � 0.05)between the tear strengths in formulations at 0 wt.% (0.660 �0.09 N/mm), 5 wt.% (1.08 � 0.12 N/mm) and 10 wt.%(1.94 � 0.09 N/mm) R812s silica filler concentration. Therewas no statistical difference (p � 0.05) in the tear strengths informulations as the silica filler concentration was increasedfrom 10 wt.% to 15 wt.% (3.19 � 0.15 N/mm) R812s silicafiller. There was a significant increase (p � 0.001) in tearstrength as the silica filler concentration was increased from15 to 20 wt % (14.48 � 1.38 N/mm) R812s silica filler. Asexpected there was also a significant increase in tear strengthas the filler concentration was increased from 20 to 25 wt.%(17.36 � 3.03 N/mm) R812s silica filler (p � 0.01); however,formulations at this filler loading proved very difficult to mix.For subsequent experiments the filler concentration used was20% by weight.

The tear strength at increasing cross-linker concentrationin formulations (at constant tin catalyst concentration) at a

Figure 3. Comparison of the tear strength of elastomers containingvarious surface treated silica fillers (16.7 wt.% and 0.72 g cross linkerper 10 g base) in C50 base polymer formulations.

Figure 4. Effect of increasing R812s filler concentration in C50 basepolymer formulations on the tear strength (0.72 g cross-linker per 10 gbase).

256 AZI, WATERS, AND JAGGER

silica filler concentration of 20 wt.% in C50 base polymer isshown in Table III. There was no significant difference (p �0.05) in the tear strength of formulations containing 0.12%and 0.20% of cross linker. As the cross-linker concentrationwas increased to 0.28% (16.62 � 3.37 N/mm), there was asignificant increase in the tear strength resulting in the for-mulation with the greatest tear strength (p � 0.01). There wasno significant difference (p � 0.05) in the tear strengthobtained from formulations containing 0.36, 0.54, and 0.72%of cross-linker (p � 0.05).

The molecular weight between the cross links, Mc, inunfilled formulations cured at increasing cross-linker concen-tration (at constant tin catalyst concentration) in C50 basepolymer is shown Figure 5. Solvent swell measurements onunfilled PDMS (swelled in cyclohexane) were used to esti-mate the cross-link density by measuring the average molec-ular weight between the cross links, Mc (g/mol). A decreasein the molecular weight between cross links is a measure of

increasing cross-link density. The results show that there wasa significant decrease in the molecular weight between thecross links, Mc, as the cross-linker concentration was in-creased in all the formulations (p � 0.001). The molecularweight between the cross links, Mc, decreased significantly(p � 0.001) from 0.08 % of cross linker (25,127 g/mol) to0.2% of cross linker (21717 g/mol). The level of increase inthe cross-link density from 0.54% (10057 g/mol) of crosslinker to 0.72% (10,057 g/mol) of cross linker was approxi-mately 1400 g/mol (p � 0.01). Although statistically therewas a significant increase in the cross-link density in consec-utive formulations, the level of increase gradually decreasesacross the group mainly due to unavailable cross-linking sites(silanol terminated groups) at high cross-linker concentra-tions.

The tear strength in filled (20 wt.% R812s) formulationscured at increasingly low molecular weight PDMS (DMS-S21) concentration in C50 base polymer is shown Figure 6.There was no statistical significance in the increase in tearstrength from 0 wt.% DMS-S21 (16.46 � 2.37 N/mm) to 10wt.% (19.50 � 2.88 N/mm, p � 0.05). There was a signifi-cant increase in the tear strength from 0 wt.% DMS-S21 to 20wt.% DMS-S21 (22.23 � 1.65 N/mm, p � 0.001). Thegreatest tear strength was achieved at 20 wt.% DMS-S21concentration. There was a significant decrease in the tearstrength from 20 wt.% DMS-S21 to 30 wt.% (15.35 �3.74N/mm) DMS-S21 (p � 0.0001). The tear strength decreased

TABLE III. Effect of Increasing Cross-Linker Concentration onthe Tear Strength of Filled Samples (20 wt.% R812s Filler inC50 Base Polymer Formulations)

Cross-Linker (%) Tear Strength (N/mm)

0.12 11.82 � 1.640.20 12.15 � 1.820.28 16.62 � 3.370.36 15.55 � 1.730.54 14.62 � 2.220.72 16.13 � 2.93

Figure 5. Effect of increasing cross-linker concentration on the mo-lecular weight between cross links (20 wt.% R812s in C50 basepolymer formulations) on the tear strength.

Figure 6. Effect of increasing DMS-S21 concentration (20 wt.%R812s in C50 base polymer formulations and 0.36-g cross linker per10 g base) on the tear strength.

257NEW POLY(DIMETHYLSILOXANE) MAXILLOFACIAL PROSTHETIC MATERIAL

significantly in formulations cured at 40 wt.% (2.95 � 0.24N/mm) and 50 wt.% (2.67 � 0.24 N/mm) DMS-S21 concen-trations when compared to all the other formulations tested inthis group (p � 0.001).

The tensile strength and elongation at break in filled (20wt.% R812s) formulations cured at increasingly low molec-ular weight PDMS (DMS-S21) concentration in C50 basepolymer is shown in Table IV. There was a significant in-crease in the tensile strength in the bimodal formulationscured at 20 wt.% (4.016 MPa � 0.419) and 30 wt.% (4.137MPa � 0.460) DMS-S21 concentration when compared withthe tensile strength at 0 wt.% (2.936 MPa � 0.429) DMS-S21concentration. There was no significant difference (p � 0.05)in the tensile strength of formulations cured at 0, 10, 40, and50 wt.% DMS-S21 concentration.

The greatest elongation at break was in the nonbimodalformulation (475.9 % � 73.4, p � 0.001), and there was asignificant decrease (p � 0.001) in the percentage elongationas the concentration of DMS-S21 increased from 0 to 50wt.% DMS-S21 (155.3% � 22.4). There was no significantdifference (p � 0.05) in the percentage elongation between10 wt.% (397.9% � 44.4) and 20 wt.% (349.3% � 42.8)DMS-S21 concentration.

The hardness in filled (20 wt.% R812s) formulations curedat increasingly low molecular weight PDMS (0 to 50 wt.%DMS-S21) concentration in C50 base polymer is shownFigure 7. There was a significant increase (p � 0.001) in thehardness as the concentration of DMS-S21 was increasedfrom the nonbimodal formulation at 0 wt.% DMS-S21(50.416 � 0.309 IRHD) and the hardest formulation was at50 wt.% DMS-S21 (68.15 � 0.078 IRHD) concentration.There was no significant difference (p � 0.05) in the hardnessof formulations at 40 wt.% (67.968 � 0.413 IRHD) and 50wt.% DMS-S21. The hardness at 20 wt.% DMS-S21(57.36 � 0.37 IRHD) concentration, the formulation with thegreatest tear strength will be subsequently referred to as C1(i.e., base polymer: C50/DMS-S21 80%/20%; Filler: R812S20%).

The molecular weight between the cross links, Mc, inunfilled formulations cured at (constant cross-linker and cat-alyst concentration) increasingly low molecular weightPDMS (0 to 50 wt.% DMS-S21) concentration in C50 basepolymer is shown in Figure 8. Solvent swell measurements

on unfilled PDMS were used to estimate the cross-link den-sity in bimodal formulations. The results show that there wasa significant decrease (p � 0.001) in the molecular weightbetween the cross links, Mc, as the concentration of low-

TABLE IV. Effect of Increasing DMS21 Concentration onTensile Strength and Tensile Failure Strain of Samples (20 wt.%R812s in C50 Base Polymer Formulations and 0.36 g Cross-Linker per 10 g Base)

DMS-S21Conc. (wt.%)

TensileStrength (MPa)

Tensile FailureStrain (%)

0 2.936 � 0.429 475.9 � 73.410 3.474 � 0.305 397.9 � 44.420 4.016 � 0.419 349.3 � 42.830 4.137 � 0.460 196.6 � 23.040 3.189 � 0.458 122.1 � 18.050 3.608 � 0.476 155.3 � 22.4

Figure 7. Effect of increasing DMS21 concentration (20 wt.% R812sin C50 base polymer formulations) on hardness.

Figure 8. Effect of increasing DMS21 concentration on distance be-tween cross links (20 wt.% R812s in C50 base polymer formulations;0.72 g cross-linker/ 0.4 g catalyst per 10 g base ).

258 AZI, WATERS, AND JAGGER

molecular-weight PDMS was increased from 0 to 50 wt.%DMS-S21. This increase in cross-link density in all the bi-modal formulations occurs at a fixed cross-linker concentra-tion (0.72 g per 10 g of bimodal base polymer). The averagemolecular weight between the cross links, Mc, in the nonbi-modal formulation 0 wt.% DMS-S21 (10057 g/mol � 131)decreased significantly at 20 wt.% DMS-S21 (7050 g/mol �88), the bimodal formulations with the greatest tear strengthin the filled formulations. The molecular weight between thecross links may be used as a guide to the cross-link density inthe equivalent filled formulations.

The tear strength for the bimodal condensation curingsystem (C1) and the commercial materials is shown in Figure9. The tear strength of the bimodal condensation system, C1(22.23 � 1.65 N/mm) was significantly greater than tearstrength of Cosmesil HC (15.55 � 1.73 N/mm, p � 0.001),Cosmesil St (4.87 � 0.25 N/mm, p � 0.001), Factor II(17.63 � 2.21 N/mm, p � 0.01), Prestige (4.53 � 1.1 N/mm,p � 0.001) and Nusil (14.45 � 4.91 N/mm, p � 0.001). Thehardness of the bimodal condensation curing system (C1) andthe commercial materials is shown in Figure 10. The hardnessof the bimodal condensation system, C1 (57.36 � 0.37IRHD) was significantly greater (p � 0.001) than the hard-ness of all the commercial materials (Cosmesil HC (44.47 �0.36 IRHD), Cosmesil St (44.99 � 0.91 IRHD), Factor II(16.26 � 2.53 IRHD), Prestige (16.21 � 0.43 IRHD) andNusil (24.33 � 0.76 IRHD).

DISCUSSION

The aim of this work was to develop an improved siliconerubber material for use as a maxillofacial prosthetic material.The main focus was to improve the tear strength, which is amajor deficiency of currently used materials. To achieve this,formulations were developed with varying type and quantityof silica filler, varying cross-link densities, and varying ratiosof high- and low-molecular-weight base polymers.

In the first instance a selection of hydrophobic silica fillersvarying in particle size, BET surface area, and type of surfacetreatment were incorporated (Table II) into PDMS (C50) at aconcentration of 16.7 wt.%. It was noted that the viscosity ofthe resulting polymer/filler base polymers for R104, R106,R202, R972, and R974 was extremely high due to strongpolymer/filler interactions; this phenomenon is known as the“thixotropic effect.” Silica particles with a high specific sur-face area or a high silanol density produce systems with avery high base viscosity. This results in difficult processingcharacteristics experienced when adding catalyst and crosslinker prior to cross linking.10 There was no significant dif-ference between the tear strengths of the silicone rubbersproduced from these fillers. The C50/R812s silica filler com-bination produced the lowest base viscosity and the greatesttear strength in comparison to other silica fillers tested andwas selected as the best hydrophobic filler in the developmentof the condensation system.

The tear strength and the viscosity of the PDMS basepolymer increased as the filler concentration was increased(at constant cross-linker and catalyst concentration). Thisreinforcing effect was due to the interaction between thesilanol groups on the silica surface and the siloxane chains ofthe polymer.9 These physical interactions are essentially re-sponsible for the rheological behavior of the base polymerprior to cross linking and for the mechanical strength of thecross-linked silicone rubber. The thixotropic properties wereat their greatest when the reinforcing effect achieved anoptimum value at filler concentration of 20% by weight, togive an acceptable base viscosity and a tear strength (14.48 �1.38 N/mm). A further increase in filler concentration to 25wt.% lead to filler overloading producing the highest baseviscosity and tear strength.

The tear strength of specimens cured at increasing cross-linker amounts was used to determine the optimum cross-linker concentration. There was a significant increase in thetear strength from 0.12% to 0.28% of cross-linker concentra-tion per 10 g of base polymer. There was no significantdifference in the tear strength at cross-linker concentrationsabove 0.28%, but the standard deviation in specimens at this

Figure 10. The hardness of C1 in comparison to the commercialmaterials.

Figure 9. The tear strength of C1 in comparison to the commercialmaterials.

259NEW POLY(DIMETHYLSILOXANE) MAXILLOFACIAL PROSTHETIC MATERIAL

concentration was greater than the standard deviation in spec-imens cured at a cross-linker concentration of 0.36%. Thecross-linking reaction of the PDMS chains is responsible forall the important elastic properties of the cross-linked net-work. A low cross-linker concentration produces a very lowdegree of cross linking, and therefore poor mechanical andtear properties. As the cross-linker concentration is increased,optimum mechanical properties such as tensile strength, tearstrength, and elongation at break are achieved. A very highcross-linker concentration produces an inelastic hard materialwith poor mechanical and physical properties.

The effect of increasing the cross linker concentration onthe molecular weight between the cross links can be clearlyobserved in solvent swell studies. Solvent swell studies onunfilled PDMS (swelled in cyclohexane at room temperature)were used to estimate the cross-link density by measuring theaverage molecular weight between the cross links, Mc (g/mol). A decrease in the molecular weight between the crosslinks is a measure of increasing cross-link density. The effectof increasing the cross-linker concentration was inverselyproportional to the molecular weight between the cross links,Mc, and hence the cross-link density increases. The cross-linkdensity increased significantly from 0.08% of cross-linker(25127 g/mol) to 0.72% (10,057 g/mol) of cross-linker. Therewas a significant increase in the cross-link density in consec-utive formulations; this level of increase gradually decreasedacross the group mainly due to unavailable cross-linking sites(silanol-terminated groups) at high cross-linker concentra-tions. The optimum cross-linker concentration can only bedetermined experimentally by measuring the tear strength inspecimens at increasing cross-linker concentration. The opti-mum tear strength occurred at a cross-linker concentration of0.36%, and the cross-link density in solvent swell specimenscured at this cross-linker concentration was 14,985 g/mol.

The addition of a lowmolecular-weight polymer (DMS-S21) into the high-molecular-weight polymer, C50, producesa broader bimodal molecular weight distribution. Increases intear and tensile strengths of the cross-linked network occuronly at low concentrations of DMS-S21. The extra cross-links from the low-molecular-weight polymer tighten the

cross-linked network while at the same time retaining theflexibility of the long PDMS (C50) polymer chains.7 Thelow-molecular-weight polymer chains are considered theweakest points in the cross-linked structure due to theirlimited elasticity when compared to the long C50 polymerchains. Under an applied tensile or tearing force the low-molecular-weight polymer chains rupture, but the long C50cross-linked network remains intact and is able to extendfully, allowing stress-induced crystallization to take place asin a nonbimodal system. The optimum bimodal concentrationwas achieved at 20 wt.% DMS-S21 (22.23 � 1.65 N/mm) inthis condensation system.5,7

At higher DMS-S21 concentration the cross-linked net-work tightens to such a point that the flexibility is reducedand extremely brittle specimens with poor tensile propertiesare produced. The tear resistance and the elongation at breakin the condensation system decreased significantly in formu-lations cured above 30 wt.% DMS-S21. Similarly, the hard-ness of the condensation system increased from 50 IRHD at0 wt.% DMS-S21, to 68 IRHD at 50 wt.% DMS-S21in thebimodal system due to the increased cross-link density andreduced flexibility of the formulations.5,7

Solvent swell measurements on unfilled PDMS were usedto estimate the cross-link density in bimodal formulations. Itmust be noted that the equations used in this analysis [Eqs.(1)–(3)] are only intended for a nonbimodal system thatcontains polymer chains of equal length, end linked with atetrafunctional cross linker. The results from this analysisshowed that there was a significant decrease in the molecularweight between the cross links, Mc, as the concentration oflow-molecular-weight polymer was increased. This increasein cross-link density in all the bimodal formulations occurredat a fixed cross-linker concentration (0.72 g per 10 g ofbimodal base polymer). This study showed that the DMS-S21polymer chains reduced the average molecular weight be-tween the cross links from 10057 g/mol in the nonbimodalformulation to 7050 g/mol at 20 wt.% DMS-S21 in theformulation with the highest tear strength (C1).

To summarize, a formulation (C1) has been developedwith a superior tear strength in comparison to commerciallyavailable materials (Figure 9). This has been achieved bycarefully tailoring the filler type and quantity, cross-link

Figure 11. The tensile strength of C1 in comparison to the commer-cial materials.

Figure 12. The tensile failure strain of C1 in comparison to thecommercial materials.

260 AZI, WATERS, AND JAGGER

density and ratio of high- and low-molecular-weight basepolymers. The tear strength was considered the most impor-tant property of a maxillofacial material to improve andtherefore was prioritized in this study. It is presumed that thecommercial materials therefore have cross-link densities thatare not as optimized for high tear strength as C1. A problemseen by producing this high tear strength is that the hardnessof the new material is relatively high in comparison to com-mercial materials (Figure 10). This is a consequence of thehigher filler loading and high cross-link density, which isneeded to improve the tear strength. The tensile strength andtensile failure strain are also affected by the optimization oftear properties (Figures 11 and 12); however, these propertiesare not as vital for a successful prosthesis, and the measuredvalues would be more than adequate for clinical use. Futurework will look to address the problem of high hardness whilemaintaining high tear strength by further manipulation of theparameters of bimodal networks, cross-link densities, andfiller loading.

REFERENCES

1. Chalian VA, Drane JB, Standish SM, editors, Maxillofacialprosthetics. Multidisciplinary practice. Baltimore: Williams andWilkins; 1972.

2. Aziz T. Development of a poly(dimethylsiloxane) maxillofacialprosthetic material. Ph.D. thesis, University of Wales College ofMedicine ; 2001.

3. Waters MGJ, Jagger RG, Polyzois GL. Wettability of siliconerubber maxillofacial prosthetic materials. J Prosthet Dent 1999;81;439–443.

4. Wolfaardt JF, Chandler HD, Smith BA. Mechanical propertiesof a new facial prosthetic material. J Prosthet Dent 1985;53:228–234.

5. Shah GB, Winter RW. Effect of bimodality on tear properties ofsilicone networks. Macromol Chem Phys 1996;197:2201–2208.

6. Bhowmick AK, Stephens HL, editors. Handbook of elastomers;new developments and technology. Plastic Engineering Series.New York: Marcel Dekker;1988.

7. Mark JE. Elastomeric networks with a bimodal chain-lengthdistribution. Acc Chem Res 1994;27;271–278.

8. Shah GB. Aspects of the polycondensation of hydrofunctionalpolydimethylsiloxanes. Ph.D. thesis, University of Wales Col-lege of Medicine, 1990.

9. Raymond P. Some observations on silicone materials and theirapplication in the maxillofacial field. Proc Inst Maxillofac Tech1977;177–195.

10. Degussa-Silicas. Product technical bulletin. 1992;15;12.11. Winter RW. The Cosmesil facial prosthesis system. Proc Inst

Maxillofac Tech 1983;381–386.12. ISO. Vulcanised rubbers—Determination of hardness. ISO 48;

1979.13. ASTM. D1415-81, Standard test method for rubber property—

International hardness. Philadelphia: ASTM; 1983. pp 317–321.14. British Standards Institute. BS-903 Part A3, Determination of

tensile stress strain properties; 1971.15. ASTM. D624-81, Rubber property—Tear resistance. Philadel-

phia: ASTM; 1983.

261NEW POLY(DIMETHYLSILOXANE) MAXILLOFACIAL PROSTHETIC MATERIAL