Poly-Vinylphosphonic Acid and Metal Oxide or Cermet or Glass Ionomer Cement

This invention relates to poly(vinyl phosphonic acid), in particular to a method of synthesising it. In addition, the invention relates to a glass ionomer or metal oxide or cermet cement made using poly(vinyl phosphonic acid). This cement may find application for specialised surgical purposes, such as dental fillings and cements and as splint bandage materials.

Glass-ionomer cements were developed in the early 1970's to meet requirements for a more acid-resistant filling material to replace dental silicate cement, formerly in widespread use. Glass-ionomer cements set by an acid-base reaction between an acid-decomposable glass powder, which acts as a base, and a concentrated solution (typically 50%) of poly(acrylic acid) ##STR1## or its copolymers with methacrylic or itaconic acids. On mixing the glass and the acid, cations (Ca.sup.2+ and Al.sup.3+), released from the glass, cross-link the polyacid chains. A metal polysalt is formed which acts as a binding matrix for the partially reacted glass particles. Thus the glass-ionomer cement is a composite material, consisting of a glassy filler embedded in an insoluble metal polyacrylate matrix.

These cements are unusual in being adherent to tooth dentine and enamel, providing excellent seals when used as filling materials. Thus, they prevent ingress of fluid and debris into the tooth at the restorative/cavity wall margin, and so prevent secondary tooth decay. They are however susceptible, before they are fully set, to contamination by moisture. They are also not sufficiently translucent to match tooth enamel optically. Poly(vinyl phosphonic acid) ##STR2## is a stronger acid than poly(acrylic acid), and hence offers the possibility of producing stronger cements, with the adhesion characteristic of existing cements. However, it has not proved easy to synthesise, nor can it be simply used as a direct replacement for poly(acrylic acid).

According to the present invention, a cement composition comprises an intimately blended mixture of a water-containing liquid, a cation-catalysed cross-linkable polymeric acid containing on average one phosphonic acid group per one to three backbone carbon atoms, such as poly(vinyl phosphonic acid) PVPA optionally containing poly(acrylic acid), and a cation-leachable surgically acceptable glass powder containing Si and Al in the mole proportions 0.6-0.2:1 if previously heat-treated, otherwise exceeding 1.6:1, in the proportions (1-x)g phosphonic acid e.g. PVPA: 1-5 g glass: x g liquid, where x is from 0.3-0.7. The solvent is preferably aqueous, i.e. water, and preferably dissolves the acid, the glass being kept separate until use. However, the acid may be dried and mixed with the glass: the invention extends to this mixture, which is preferably packed in a sealed capsule (PVPA being hygroscopic) and which is made into cement by adding the solvent. In a modification, the glass may be replaced by a metal oxide such as MgO, preferably heat-treated, preferably of dense powder morphology, or by a cermet. (A cermet is a metallised glass powder, typically formed by fusing (by sintering) a metal powder (e.g. silver or tin) onto the surface of a glass powder, and may offer lower friction when burnished and greater wear resistance than glass in these cements.)

The invention extends to a pack comprising two pastes which when mixed form a cement composition as set forth above; the first paste may be the acid(s) plus water, and the second paste may be the glass powder suspended in aqueous thickening agent e.g. methylcellulose. If the two pastes have been formulated to appropriate concentrations, one could in use squeeze out equal lengths of paste from two tubes, or scoop out equal numbers of spoonfuls from two tubs, as an easy way to ensure that the mixture is of the correct composition.

The composition may further contain acid-neutralising agents such as zinc oxide or aluminum phosphate or both, in a total amount of up to 10% by mass based on the acid solution, and/or complexing agents such as phosphonic-acid-based materials.

The glass powder preferably consists of particles substantially all of which are smaller than 100 microns, preferably smaller than 60 microns. Preferably the glass powder has been heat-treated at at least 400.degree. C. (preferably 450.degree.-600.degree. C.) for at least 40 minutes (preferably at least 55 minutes). The Si:Al range of 0.6-2:1 yields an opaque product, which may be acceptable in appropriate cases. The glass may be washed in a dilute acid such as acetic acid, preferably in a y % concentration for z hours where yz<4, preferably .ltoreq.2. In place of glass, MgO deactivated at at least 900.degree. C. may be used.

The invention also extends to a pack comprising the components of the cement composition set forth above, so packed that when unpacked and mixed they form the cement.

Poly(vinyl phosphonic acid), PVPA, can be prepared by a free radical polymerisation in an inert solvent, using for example vinyl phosphonic acid monomer. The polymer is a light brown very hygroscopic solid, which dissolves readily in water to give solutions of up to 80-85% by weight. The solutions show good long term stability with no evidence of gelling. A 50% solution typically has a viscosity of 0.8-0.9 cP.

Poly(vinyl phosphonic acid) is a dibasic acid with pKa values (acidic strength) of 2 and 8 respectively, compared with the pK.sub.a of 5-6 for poly(acrylyc acid). It is expected therefore that the aforementioned cross-linking cations released from glass powder (Ca.sup.2+ and Al.sup.3+) would form stronger associations between the polyacid chains, leading to stronger cements.

In practice, however, cement formation with the glass types used in current glass-ionomer cements is difficult when PVPA is used in place of PAA. For example, a glass which has a setting time of 3-4 minutes with PAA solution will set in probably less than 45 seconds with PVPA solutions. It thus proved necessary to modify the cement system to allow a reasonable working and setting time for the cement paste, by:

i) using glasses with a different formulation which allows adequate working time, or by deactivating reactive glasses;

ii) partially neutralising the PVPA solution, so reducing its reactivity, and

iii) use of additives, such as complexing agents, in the liquid; these are believed to tie up cations as they are released from the glass and delay the setting of the cement.

Using a cement according to the invention as above, cements with setting time of about 5 minutes can be obtained, which is comparable with existing glass-ionomer cements (5-20 minutes).

1. Adhesion

Adhesion of a dental restorative material to the tooth substrate is of prime importance. The dental silicate cement (now no longer used since inadequately acid-resistant), typical of phosphoric acid based cements, did not adhere to the tooth enamel and as a consequence fluid and debris could leak into the tooth at the edge of the filling, possibly resulting in secondary tooth decay. An adhesive filling material will (in contrast) effectively seal this edge. In addition, with non-adherent restoratives, the filling needs to be mechanically keyed into the tooth by undercutting, with the loss of excessive sound tooth material. This is not necessary with adhesive materials.

2. Biocompatibility and toxicity

Glass-ionomer cements are known to be compatible with the oral tissues and do not irritate the dental pulp as phosphoric acid based cements do. This improvement is attributable to the very low diffusion of bulky polymer molecules through the dental tubinals leading to the pulp. Cements based on PVPA are expected to be equally bland, and like PAA cements they also show very small temperature rises on setting.

Current glass-ionomer cements show resistance to acid attack and to staining and PVPA-based cements are expected to show similar properties.

The toxicity of the PVPA starting material vinyl phosphonyl dichloride is known to be high, but this is converted 100% to vinyl phosphonic acid monomer and polymer. The toxicity of the VPA monomer is not known, but its complete removal from the polymer solution prior to cement formation is readily possible as described later. As a matter of interest, the monomer has been found to form insoluble cements with certain metal oxides, and would not therefore be expected to be leachable even if present in the initial polyacid solution.

3. Translucency

If a restorative material is to be successfully used as an anterior filling material, then its translucency should match that of the surrounding tooth material, to give a "live" appearance. Dentine has an opacity of 0.51-0.93 (C.sub.0.70 value, test disc thickness 1 mm, two discs compared for strength of reflected light when one sits on a white (reflectivity 0.70) and the other on a black background), while enamel has an opacity of 0.21-0.67. The opacity of glass-ionomer restorative material is affected by the refractive indices of the glass particles and of the gel matrix; a close match of the refractive indices results in a low-opacity cement.

The British Standard for glass-ionomer cements quotes limits for the opacity of cements as 0.35-0.90 (C.sub.0.70 value), and typical values are 0.65-0.75 for filling cements, above the range for enamel. Dental silicate cements had a much lower opacity (B.S. opacity value 0.35-0.55, typically 0.45-0.55), but these are no longer in use. Measurement of the opacity of a cement prepared from PVPA and a modified dental silicate glass gave a C.sub.0.70 value of 0.40-0.46 compared with 0.5-0.6 for a cement prepared from the same glass and PAA of a similar molecular weight as the PVPA. Such low opacity indicates that aesthetic anterior fillings should be possible using PVPA cements.

The invention will now be described by way of example.

The three accompanying drawings, which are described in more detail at the end, illustrate the setting properties of cements (both within and outwith the invention) with reference to their composition as plotted on ternary composition diagrams, in which the counts are of mole per cent of Ca atoms, mole per cent of Si atoms and mole per cent of Al atoms unless otherwise stated. The same letters refer to the same glasses.

This part describes the preparation of a polymer of vinyl phosphonic acid from vinyl phosphonyl dichloride (VPDC). VPDC is dissolved in an equal volume of 1,1,1-trichloroethane and to this solution is added the initiator, azo-bis-isobutyronitrile, at a level of 3% by weight of the monomer. The mixture is heated under nitrogen, with stirring, in a flask equipped with a reflux condenser to a temperature of 70.degree. C. for two hours. Heating is by means of a water bath, surrounding the flask, to maintain adequate temperature control. If the entire polymerisation (18 hours) were carried out at 70.degree. C., the yield would be 50-60%. Instead, after two hours at 70.degree. C., the temperature is lowered to 40.degree.-45.degree. C. and maintained at this temperature for a further 16-18 hours, whereby the yield is increased to 85-90%.

A viscous, orange-brown solution results, which is a solution of the crude polymer of vinyl phosphonyl dichloride. This is then hydrolysed by pouring slowly into a large volume of chilled water with stirring, removing the bulk of the HCl gas as it is formed by means of a vacuum. The hydrolysed product is then concentrated, and the excess organic solvent removed by vacuum distillation or by means of a rotary evaporator.

Analysis of the product typically shows it to contain 85-90% of the desired PVPA-poly(vinyl phosphonic acid), a solid, together with some 10-15% residual monomer.

This part describes the purification of the product prepared as described in Part 1. The PVPA must first be isolated by removing the water introduced by the hydrolysis, using a rotary evaporator or similar. The resulting solid is then dissolved in an equal or lesser weight of ethanol (other alcohols can be used), with gentle warming as necessary. The solution is then poured slowly into a large excess of a non-solvent, such as ethyl acetate, with stirring. A white precipitate forms, which is isolated from the solvent mixture, and contains some 5% of residual monomer. Repeating the dissolution in ethanol and the precipitation with ethyl acetate leaves no residual monomer detectable by P-31 NMR spectroscopy.

Pure PVPA is a hygroscopic, off-white solid. It is readily soluble in water up to concentrations greater than 80% by weight when prepared as described. The viscosity of an infinitely dilute solution (intrinsic viscosity), obtained by the extrapolation of the measured viscosities of solutions of concentrations 1, 2, 3 g/100 ml (sodium salt of PVPA in 1M NaCl) is 0.08 ml/g, suggesting a polymer of low molecular weight, probably in the range 3000-5000.

This part describes the cement-forming properties of aqueous solutions of poly(vinyl phosphonic acid) with various metal oxide powders, i.e. not with glass as such. In all cases a 50% by weight aqueous solution of poly(vinyl phosphonic acid), prepared as described in Parts 1 and 2, is used as the cement-forming liquid. The liquid and a powdered metal oxide are mixed together on a glass block with a metal spatula at a g powder: ml liquid ratio which gives a homogenous paste, and which is from 1:1 to 5:1 depending on the reactivity and bulk density of the metal oxide powder. Many metal oxides can be reacted with poly(vinyl phosphonic acid) in this way to yield fast setting, water-stable cements at room temperature.

Metal oxides which are suitable for forming water-stable cements are:

Cu(I), Cu(II), Co(II), Co(III), Sn(II), Bi(III), Pb(II), Pb(II,IV), Hg(II), Cd(II), Y(III), La(III), Mo(VI).

In addition, the following oxides reacted very vigorously with poly(vinyl phosphonic acid), giving cement-like products in less than 15 seconds: Zn(II), Ca(II), Mg(II). In an attempt to deactivate (reduce the reactivity of) these oxides, they were heat treated at 1000.degree. C. for at least 24 hours. Only in the case of magnesium oxide could a practical cement be mixed in this way. The morphology of the MgO was also found to affect setting. "Light" MgO, a low density form, reacted very rapidly despite the heat treatment, whereas "heavy" MgO, a much denser powder, gave a suitable setting time.

Typical physical properties of some metal oxide-PVPA cements are given in the Table.

The MgO setting time at 37.degree. C. (as in the mouth) was under 200 seconds. That product (after 24 hours at 100% relative humidity) adhered not only to tooth enamel (5.0 MPa) and dentine (3.4 MPa) but also to base metals useful in dental surgery such as stainless steel and cobalt-chrome alloys.

This part illustrates the properties of a typical cement prepared from an aqueous solution of poly(vinyl phosphonic acid) and an ion-leachable aluminosilicate glass. The polyacid solution used is 50% m/m. The glass is prepared by mixing together 437 parts by weight silica, 230 parts by weight alumina, 129 parts by weight calcium fluoride, 175 parts by weight cryolite and 29 parts by weight aluminium phosphate and heating to 1300.degree. C. for 75 minutes. The melt is cooled rapidly by pouring into water. The resulting glass is ground and sieved, and the fraction of particle size less than 45 microns used to prepare the cement. Before use, the glass powder is deactivated by heating in a furnace at 450.degree. C. for 90 minutes.

The setting properties of the poly(vinyl phosphonic acid) cement of Part 4 are compared with commercial glass ionomer cements containing poly(acrylic acid) or similar. Setting times were determined using a 453 g Gillmore needle. Working times and setting rate were determined from Wilson rheograms.

The PVPA coment could be made with glass acetic-acid washed for 1 or 2 hours, acid concentration being 1% or 2%. All combinations extended the working time and slowed the setting rate, which could be useful, but 2 hours/2% did so rather strongly.

The mechanical properties of the poly(vinyl phosphonic acid) cement were compared with commercial glass ionomer cements. The cements were mixed at the powder/liquid ratios given in Part 4A. Compressive strengths were determined on cylindrical specimens (12 mm length, 6 mm diameter), after storage for 24 hours.

Flexural strength was determined using specimens of dimensions 25.times.3.times.3 mm.

This Part illustrates the translucency of the poly(vinyl phosphonic acid) cement. The cement paste is placed in a brass mould, such as will give a cement disc of diameter 2.0 cm and thickness 0.1 cm. The mould is closed with steel plates and the cement allowed to set. The cement disc is removed from the mould after one hour, and subsequently stored in water for a further 24 hours. The opacity of this cement disc, when measured on a Hunterlab D25A-9 Tristimulus Colorimeter is superior to that of a similar cement prepared using poly(acrylic acid) as the cement forming liquid, and compares favourably with the opacity of a similarly prepared commercial dental silicate cements (DSC) and glass ionomer cements (GI). The C.sub.0.70 opacities obtained for the cements are given in the Table.

This Part illustrates the rapid development of water stability of cements prepared from poly(vinyl phosphonic acid), and compares this feature with similar cements prepared from poly(acrylic acid) and with commercial glass ionomer cements. Thus, a specimen is prepared by placing the cement paste (powder/liquid ratio 3 g: 1 ml) in a brass mould such as to give set cement disc of diameter 2 cm and thickness 0.2 cm. A length of unwaxed dental floss is placed in the cement paste, by which to suspend the set cement disc. After being allowed to set for 7 minutes or 1 hour, as required, the disc is removed from the mould, and immediately suspended in a tared weighing bottle containing a fixed volume of water. After 24 hours in water, the disc is removed, and the water evaporated, leaving the residue leached from the cement disc. The water stability of the cement is thus determined from the difference in weight of the bottle and the weight of the cement disc.

The water stabilities of cements thus determined are given in the Table.

This example describes the cement forming properties of a range of glass powders based on that described in Part 4, and prepared by altering the pre-fired composition of the glass. Thus the effect of altering the amount of aluminium phosphate (5A-5D), the Si/Al ratio (5E-5H), and the amount of fluoride (calcium fluoride) (5J-5M) in the glass is described. The glass compositions are listed below as parts by weight.

The setting times of cements prepared from these glass powders are given, measured at 23.degree.-24.degree. C. and ambient humidity. Deactivation of the glass powders is achieved by heating at 450.degree. C. for 90 minutes.

This Part describes the preparation of a cement from an ion-leachable glass and an aqueous solution of PVPA. A calcium fluoroaluminosilicate glass is prepared by mixing together 67 parts by weight of silica, 100 parts by weight of alumina and 100 parts by weight of calcium fluoride, and heating to a temperature of 1300 degC for 90 mins. The melt is rapidly cooled by pouring into cold water, and the resulting glass-ceramic crushed and sieved so as to collect the fraction of particle size less than 45 microns. The cement is prepared using a 50% by weight solution of poly(vinyl phosphonic acid) at a p/1 ratio of 3:1. The resulting cement-forming reaction is very vigorous, and a hard, dense cement-like mass results in less than 10 seconds.

The glass powder can be deactivated by heating to 600.degree. C. for at least 1 hour. When mixed with the poly(vinyl phosphonic acid) solution, this deactivated glass powder gives a setting time (at 23-24 degC) of 105 seconds. The compressive and flexural strengths of this cement are given below and are compared with the compressive strengths of similar cements prepared using poly(acrylic acid) as the cement forming liquid, and the same (unheated) glass powder.

This Part illustrates the effect of altering the Si/Al ratio on the cement properties of the glass type described in Part 6. Thus a series of glass powders was prepared as described, with Si/Al mole ratios from 2.5:1 to 0.23:1, by altering the amount of silica in the pre-fired mixture. This is the ratio by number of the silicon atoms to the aluminium atoms present. Cements were mixed at 2.5 g powder: 1 ml liquid and setting times measured at 23.degree.-24.degree. C. The Table also shows the effect of heating to 600.degree. C. on the reactivity of the glasses.

Glasses 7G, 7H and 7I after heating were partly crystallised, so that the "real" (amorphous) glass apart from the crystals would have had a composition deviating from the overall nominal composition. That actual composition was not determined. It may account for the reversal in the trend of setting times.

Similarly the Si/Al ratio has an effect on the physical properties of the set cements, as shown in the Table below. The glass powders for making the last three cements were heat-treated at 600.degree. C. for 6 hours.

This Part illustrates the effect of altering the fluoride content of the type of glass described in Part 4. Thus cements were prepared as described, using glass powders of constant Si/Al mole ratio (viz. 1.48:1), but with varying amounts of fluoride. The effect on the setting time of such cements (mixed at 2.5 g powder: 1 ml liquid) is given in the Table.

Fluoride acts as a fluxing agent, making the glass easier to work (as may be imagined--in glass 8A, fluorine atoms are 40.6% by number of the whole, all other species of atom added together being only 59.4%). Fluoride makes the glass more susceptible to acid attack, and it moderates the setting reaction.

Increasing the Ca:Al ratio generally decreases the setting time if no other factors are varied.

This Part describes the effect of the incorporation of aluminium phosphate, zinc oxide and mixtures thereof in the aqueous poly(vinylphosphonic acid) solution, on the working time and the setting rate of the glass described in Example 4. A Wilson rheometer is used to determine these data on cements (3 g powder: 1 ml liquid) mixed at 23.degree.-24.degree. C. The glass powder was deactivated by heat-treating at 450.degree. C. for 90 minutes before use.

This Part describes the effect of adding chelating compounds on the working properties of the cements described above. The additives were incorporated into the polymer solution so as to give a final solution containing 45% m/m PVPA and 5% m/m (unless otherwise stated) of the additive. A number of additives were tried: all were phosphonic acids manufactured by Monsanto Co. under the trade name "Dequest". The Dequests tried and found to be useful were:

Dequest 2000 aminotris(methyl phosphonic acid)

Dequest 2010 hydroxyethyl-diphosphonic acid

Dequest 2060 diethylenetriamine-penta(methylenephosphonic acid)

The modified PVPA solution was mixed with two types of glass (Parts 4 and 6 above, 3 g powder: 1 ml liquid. The setting time was determined with a 453 g Gillmore needle at 23.degree.-24.degree. C. and ambient humidity (55%). The effect of Dequests on the setting of cements prepared from untreated and heat treated glasses was studied, and the results shown below. (The effect of Dequests on the compressive and flexural strengths of the cements of Part 7 was studied and found to be insignificant).

An estimate of the working time (manipulability) of each of the cements was made by examination of the rheograms obtained using a Wilson rheometer for each of the cements.

An ideal dental filling material will have a 37.degree. C. setting time of 3-6 minutes from starting the mixing or 2-5 minutes from completing the mixing (itself done at 23.degree. C.), following a relatively leisurely working time.

This part describes the cement formed between an aqueous solution of poly(vinyl phosphonic acid) and a tin cermet. The tin cermet was prepared by mixing glass powder (as described in Part 6, but ground to 18 .mu.m) with tin powder (Goodfellow Metals, sieved to <45 .mu.m) in equal volumes (1:4 by weight). The mixture was compressed to form a 2.5 cm diameter disc under a pressure of 20 tonnes. The discs were then sintered in a vacuum furnace at 210.degree. C. for 30 minutes before being ground to a powder and passed through a 15 .mu.m sieve.

A cement was prepared by mixing a 50% by mass solution of PVPA with the cermet material prepared as described above, in the ratio 5 g powder: 1 ml liquid. The cement was water stable on setting. It had the following properties:

Turning to the drawings, FIG. 1 shows the cement setting properties of SiO.sub.2 --Al.sub.2 O.sub.3 --CaF.sub.2 powdered glass used with PVPA, the glass having not been heat-treated. Foregoing Examples 7A-7H are Glasses A-H respectively in this FIG.; 7I is J. The other Examples and the other Glasses do not correspond. The most suitable compositions describe an arc across the diagram.

FIG. 2 shows a selection of the glasses examined in FIG. 1, with the difference that they have been heat treated (600.degree. C. for 90 minutes), the glasses containing equal CaF.sub.2 and Al.sub.2 O.sub.3 plus varying amounts of SiO.sub.2.

FIG. 3 compares some of the foregoing compositions and some with CaO instead of CaF.sub.2, and 2.5 g glass used per ml PVPA.

Cements prepared by mixing two pastes, as opposed to a powder and liquid, have the advantage of ease of measuring and give an easier mix. Solutions of PVPA at concentrations of 80% m/m (by mass) can be prepared, these being viscous liquids. The possibility of using such concentrated solutions as par of a two paste glass ionomer type cement was investigated, the other paste being a glass powder suspended in water and some dispersing aid. Materials.

80% m/m PVPA solution was used as one component of the cement. The glass paste was prepared by mixing G5 (i.e. the glass of Part 4 above) with sufficient water such that when this paste was mixed with the PVPA solution, a 50% m/m solution of PVPA in water resulted. In addition, various dispersing aids were added to the glass paste to prevent drying out. Sufficient glass to give an approximately 3:1, or more usually 5:1, p/l ratio with the 50% PVPA solution was used. A density of 1.0 g cm.sup.-3 was assumed for a 50% PVPA solution.

The dispersing aids used in this study were "Cellosolve" (i.e. ethoxyethanol), Tween 80 (a non-ionic surfactant) and Dequest 2010 (explained in Part 10). The composition of the glass pastes are given in Table 3.

The G5 glass was used both unheated and heat treated (450.degree. C. 90 mins) where marked with asterisk (G5*).

Firstly, a cement was prepared from 80% PVPA and glass powder. A maximum p/l ratio of 1:1 was possible, giving a stiff but workable paste (Paste 0). This cement had a very long setting time, and remained soft and pliable at >1 hr. The poor setting characteristics of this cement confirmed that there is a minimum quantity of water required in the cement paste for the reaction to occur. Generally cements are prepared from 50-60% m/m PVPA solutions.

The setting behaviour of the cements prepared from the glass pastes of Table 1 are shown in Table 5. The compressive strengths of the cements are shown in Table 2. All of the cements when mixed as two pastes gave a homogenous paste within 10 seconds, except paste no. 6. In this case the unheated glass reacted with D2010 in the glass paste, forming a friable solid. When mixed with the PVPA solution the resulting cement set very rapidly (<20 seconds). With the heat deactivated glass (e.g. in Pastes 4 and 5) there was no reaction noticed between preparing the glass paste and forming the cement from it, generally about 5-10 minutes, although storage of such a paste may have shown similar problems to that observed for Paste 6. By adding the D2010 to the PVPA solution however, (Paste 7), such problems were overcome. This cement (Paste 7) used non-deactivated G5, and consequently had a working time similar to that obtained for a standard powder-liquid mix. However, the p/l ratio of this cement was almost double that which could be achieved with a normal mix.

In general, mixing cements from two pastes gave a homogeneous mix of higher p/l ratio than could be achieved by mixing a powder and liquid. The working times of the cement pastes were also increased, especially when using cellosolve in the glass paste. However, the disadvantage with all of the cements was their low strength when compared with the normal powder-liquid cements, Table 2. Although the incorporation of D2010 did increase the strength of the cements, it was not able to fully compensate for the weakening effect of the other additives (i.e. cellosolve or Tween 80). Typically, a cement prepared from 50% PVPA, containing 5% D2010, and G5*, at a p/l ratio of 3:1 has a compressive strength of 85-90 MPa.

The two paste cements may therefore be limited in their applications to temporary restoratives, for example. Further work is required to identify other dispersing aid, which might have less effect on the strength of the resulting cements.

All cements were prepared by mixing the glass pastes with 0.15 g of 80% m/m PVPA solution.

Compressive strengths were measured in cements 24 hours old, stored at 37.degree. C., 100% RH.