the effect of particle size distribution

7/27/2019 The Effect of Particle Size Distribution

1/6

The effect of particle size distribution

on an experimental glass-ionomer cement

Leon H. Prentice, Martin J. Tyas*, Michael F. Burrow

School of Dental Science, University of Melbourne, Parkville, Vic. 3010, Australia

Received 26 February 2004; received in revised form 8 June 2004; accepted 29 July 2004

KEYWORDSPolyalkenoate

cement;Compressive strength;Working time;Setting time;Reactive fluoride

glass;Particle size;Viscosity;Glass-ionomer

Summary Objectives. The role of particle size and size distribution of glasspowders in glass-ionomer cements (GICs) has been largely overlooked, being limitedto demonstrations of the classical inverse sizestrength relationship. This studyinvestigated variation in properties of an experimental glass-ionomer cement when acombination of large (Powder A) and small (Powder B) particles was used.

Methods. Large- (mean size 9.60 mm) and small-particle (3.34 mm) glass powderswere blended in various proportions and mixed with powdered polyacrylic acid tomake a range of glass-ionomer powders. These powders were mixed with a glass-ionomer liquid (SDI Ltd, Australia) at powder to liquid ratios of 2:1, 2.5:1, and 3:1,and the resultant cements evaluated for working time, setting time, clinicalhandling, and compressive strength. Results were analysed by ANOVA.

Results. An increased proportion of smaller particles corresponded to higher

strengths, and an increased proportion of larger particles with a decrease in viscosityof the unset cement. When 2030% by weight of small particles was used, the pastedemonstrated a peak in cohesion and working time, with a viscosity similar tocommercial restorative GICs.

Significance. Optimisation of particle sizing and distribution may thus lead to glass-ionomer cements with improved clinical handling characteristics and greaterstrength, which may increase the longevity of the restoration.Q 2004 Academy of Dental Materials. Published by Elsevier Ltd. All rights reserved.

Introduction

Glass-ionomer cements (GICs), more properly

referred to as glass polyalkenoate cements, areformed from powdered glass and aqueous poly-acids. The glass is generally calcium or strontiumfluoroaluminosilicate; strontium provides the radio-pacity and fluorine, in the form of fluoride ions,

enters the matrix phase, from where it can bereleased to promote remineralisation of caries-affected tooth structures, and confer antimicrobial

properties to the final cement[14]. The polyacidsreact with the glass in an acidbase reaction,leaching calcium/strontium and aluminium ions toform the set restorative cement, which is a mixtureof unreacted glass in a calcium/strontium andaluminium polycarboxylate matrix.

Extensive work has been undertaken examiningthe interfacial reaction between the glass andthe polyacid solution [59]. However, little of

Dental Materials (2005) 21, 505510

www.intl.elsevierhealth.com/journals/dema

0109-5641/$ - see front matter Q 2004 Academy of Dental Materials. Published by Elsevier Ltd. All rights reserved.doi:10.1016/j.dental.2004.07.016

* Corresponding author. Tel.: C61 3 9341 0231; fax: C61 39341 0437.

E-mail address:[email protected] (M.J. Tyas).
http://www.intl.elsevierhealth.com/journals/demahttp://www.intl.elsevierhealth.com/journals/dema


2/6

the conventional understanding of solid or liquidslurries and composite materials, such as the role offiller size in dental resin composites, or the flowcharacteristics of slurries in mineral processing, hasbeen applied to GICs, even though variation inparticle size distribution has been identified as amajor route to improved mechanical properties

[10]. There has been minimal investigation into theeffects of particle size, and in particular particlesize distribution, of the glass in conventional GICs.

The early work of Kent and Wilson [11]has beencontinued recently by Brune and Smith [12], whofound mean particle sizing (based on sieve tech-niques) had little effect on compressive strength. Acomparison with composites was used to explain anincrease in abrasion resistance in association with adecrease in glass particle sizes[13]. It is commonlyknown that GICs have larger mean particle sizesthan other restorative materials [14], which is

recognised as a contributing factor to the relativeweakness of the material[10,14]. The lower limitsof particle size in GICs are set by the viscosity of theresultant cement and by the necessity of significantglass content, to ensure that strength, ease of useand setting time remain suitable for a restorativematerial.

It is hypothesised that variation in particle sizedistribution has no significant effect on the hand-ling, working time, setting time, and compressivestrength of an experimental GIC (null hypothesis);this investigation aims to explore this hypothesis.

Materials and methods

Preparation of glass-ionomer powderand liquid

A strontium fluoroaluminosilicate glass was milledusing a laboratory ball mill, wet-sieved through a25-mm mesh, and the suspension allowed to settlefor 10 min. The supernatant was removed, and thesettled glass filtered, dried, and re-sieved through a

150-mm mesh to remove agglomerates, resulting ina uniform glass powder (powder A). Powder A hada mean particle diameter of 9.60mm (Table 1),measured with a laser-diffraction particle sizeanalyser (Mastersizer S, Malvern Instruments, Mal-vern, UK). The supernatant was filtered and dried toobtain a glass powder (powder B) with a meandiameter of 3.34 mm (Table 1). Particle sizedistribution and powder surface areas (sphericalmodel) for both powders are given inTable 1. Bothglass powders were surface-treated in 5% acetic

acid for 2 h, filtered, dried, and re-sieved through a150-mm sieve prior to use.

A proprietary developmental glass-ionomer liquid(SDI Ltd, Bayswater, Australia) was used as thestandard liquid for all tests, consisting of polyacrylicacid (2535%), L(C)-tartaric acid (515%), water,and accelerators.

Clinical handling

A table of clinical ranking scores was developed forthe clinical handling assessment of the pastes(Table 2). Unset material was evaluated immedi-ately after mixing and during the setting reaction inorder to estimate the clinical suitability of thematerial. A score of 0 was given for inability toextrude material from the capsule, while a score of5 corresponded to pastes with characteristicssimilar to commercial GICs.

Sample preparation

Glasses A and B were combined in various pro-portions as shown in Table 3, and each resultantpowder was mixed with a powdered polyacrylic acid(Plex 4779L, Rohm GmbH, Darmstadt, Germany) tomake the final range of powders.

The powders and liquids were measured intounassembled capsules (Riva SC, SDI Ltd) at threedifferent powder: liquid ratios; 2:1, 2.5:1 and 3:1.Each capsule was assembled, activated, and mixedin an amalgamator (Ultramatw 2, SDI Ltd) for 10 s.

All testing was carried out at (20G1) 8C.

Working time and initial setting time

Working time was determined as the time when thematerial could no longer cohesively string to aheight of 10 mm when lifted with a spatula. Initialsetting time was determined in a manner analogousto ISO9917E [15] as the point at which a 1-mmamalgam plugger no longer caused a permanentindentation in the material.

Table 1 Particle size distributions (mean diameters)of powders A and B.

D0.1(mm) D0.5(mm) D0.9(mm) Surfacearea(m2 gK1)

Powder A 3.59 9.60 20.86 0.30

Powder B 1.87 3.34 6.40 0.70Subscripts refer to proportions, by volume, which fall belowthe specified diameter.

L.H. Prentice et al.506


3/6

Compressive strength

Cylindrical specimens of height (6.0G0.1) mm anddiameter (4.0G0.1) mm were made and tested forcompressive strength according to ISO9917 [15].The specimens were stored at O80% humidity at37 8C for 1 h, removed from the moulds and

immersed in water at 37 8C for a further 23 h, andloaded to fracture on a universal testing machine(Instron 5568, Instron Ltd, Milton Keynes, UK) at acrosshead speed of 0.5 mm/min. Strength datawere analysed using one-way ANOVA (p!0.05)and Tukey post-hoc analysis.

Results

Working time, setting time, and clinical

handling

Fig. 1 shows the variation in working and settingtime for pastes formed at the powder:liquid ratio of2:1. The initial setting time decreased as theproportion of small particles (powder B) increased,while the working time showed a peak around 10%powder B. At this point also the paste achieved itshighest score for handling (Table 3). The decreasein initial setting time followed an s-shaped curve,dropping more steeply after approximately 10%powder B before flattening out.

The graph of the powder:liquid ratio of 2.5:1

(Fig. 2) showed a peak in working time similar to thegraph of 2:1. However, the maximum in the workingtime occurred at a proportion of powder B ofapproximately 1520%. The data for handling(Table 3) also followed a similar pattern, withincreased scores for the paste up to around 20%powder B. At this point the paste became moreviscous and exhibited reduced working and initialsetting times.

Table 2 Ranking scores for clinical handling charac-teristics.

Paste rank-

ing score

Detailed description

0 Working time was too brief to allowextrusion from capsule. Under high

force, some separation of liquid fromglass particles was observed1 Coarse, non-cohesive, clay-like paste.

Manipulation was not possiblematerial did not form a cohesivepaste. Difficult to extrude, settingtime was very fast

2 Very poor cohesion. Paste had a shortworking time, but was still soft undermanipulation. Very weak, clay-likematerial prior to completion of settingtime

3 Paste was soft and glossy. Somecohesion was evident, but working

time was reduced by the coarseparticles in the material

4 Low-viscosity paste, extrusion waseasy. Working time was extended bycohesive properties, but some coarse-ness was evident. Cement resistsfragmentation prior to setting time,and final cement was strong

5 Low-viscosity paste. Extrusion waseasy. Unset material was glossy andcohesive. Working time was extendedby cohesive properties as viscosityincreases, and cement maintainedhigh cohesion prior to setting time

Table 3 Ranking scores for a paste at variouspowder:liquid ratios.

% Powder B Powder:liquid ratio

2:1 2.5:1 3:1

Paste ranking score0 2 1 02.5 2 1 05 2 2 110 3 2 112.5 3 3 215 3 4 320 4 4 425 4 3 430 4 3 550 3 3 560 N/T N/T 4100 2 1 0

N/T, Not tested. Figure 1 Working time and initial setting time (min-

utes) of powder blends at a powder:liquid ratio of 2:1.

Particle size effects in GICs 507


4/6

For pastes at the powder:liquid ratio of 3:1, nodetermination of working time, initial setting time,or strength could be made at either extremes ofpowders A or B content. Pastes made solely frompowder A had paste properties similar to dry clay,

and even up to 5% powder B did not form cohesivepastes (Table 3). Pastes from 10 to 50% powder Bformed workable pastes, with a maximum in theworking time occurring for 30% powder B content(Fig. 3). The paste with 30% powder B displayedexcellent paste properties, and had a working timeand initial setting time corresponding to conven-tional regular-set glass-ionomer restoratives.

Increasing the proportion of small particles(powder B) also increased the viscosity of thepaste, but this was noticeable only above a criticalvalue, which depended on the powder:liquid ratio

and was lower at higher ratios. At very low powder Bconcentrationsbelow 15%and at powder:liquidratios of 2:1 or 2.5:1, the increase in viscosity dueto the greater surface area of small (powder B)particles (Table 1) was small (Table 2). At 2:1, themarked increase in viscosity due to the highsurface area of powder B occurred at around 60%powder B. At 2.5:1, the viscosity-increasing proper-ties of powder B were demonstrated at lower

concentrations, being evident at around 20% pow-der B (Table 3). At 3:1, the viscosity of a materialcontaining only powder A was indeterminate, and apoint of transition to properties characteristic ofmaterials with small particles was unobservable.

Compressive strength

The 24-h compressive strength of the 3:1 materialsshowed a linear increase (RZ0.976) with powder Bcontent (Fig. 4, Table 4). Although, no workablepaste could be formed at 100% powder B, theextrapolation of the fit indicated a theoreticalstrength at this point of 180 MPa, which approachesother GICs.

Discussion

A theoretical modelling of the flow and viscosity ofglass-ionomer pastes is extremely difficult, given thecomplexity of the particleparticle and particleliquid interactions, polymeric characteristics, andgelation reaction. Even two-phase dense solidliquidsystems do not compare well with theoretical

Figure 2 Working time and initial setting time (min-utes) of powder blends at a powder:liquid ratio of 2.5:1.

Figure 3 Working time and initial setting time (min-utes) of powder blends at a powder:liquid ratio of 3:1.

Figure 4 24-h compressive strength (MPa) as a functionof percentage powder B incorporated into the powders.

Table 4 24-h compressive strength (MPa) as afunction of percentage powder B incorporated intothe powders.

% Powder B Strength (MPa)10 58.0 (4.5)a,b

12.5 55.9 (4.6)a

15 61.2 (6.3)a,b

20 65.4 (6.9)b

25 77.9 (2.3)c

30 79.8 (4.6)c

40 100.3 (2.1)d

50 103.0 (2.4)d

60 116.7 (8.3)e

Superscripts denote significant differences (p!0.05).



5/6

calculations[16]. Although, conventional glass-iono-mer cements fall into the two broad categories ofconventional Type II cements and highly-viscouscements[17], these are not defined quantitatively.

At each powder:liquid ratio, the improvement incohesion and the corresponding increase in workingtime were clear, though the locations at which the

peaks occurred varied for the different ratios. Theintermediate powders, containing various amounts ofboth powder A (coarse) and powder B (fine), demon-strated handling characteristics that were improvedover either extreme. Powder A suffered from a lackof cohesion and, especially at higher powder to liquidratios (Table 3), demonstrated a shear phase-separ-ation (the separation of the liquid from the solidphase under high loads) typical of clays. The pastescontaining over 50% powder B demonstrated workingtimes that were too short (less than 30 s) to beclinically suitable. It was not possible to mix a paste ateither extremely high or low powder B content at a

powder:liquid ratio of 3:1 (Fig. 3,Table 3).As the solids content of the mixed pastes

increased, the optimal concentration of powder Balso increases (Figs. 13), from approximately 10%at 2:1, to 1520% at 2.5:1, to around 30% at 3:1. Thisis most l ikely due to the particle packinginteractions and the progressing polycarboxylatecross-linking reaction, which form two competingmechanisms. At low powder:liquid ratios, thegreater amount of polymer solution per surfacearea of glass will prolong the working time, as agreater amount of ions must be leached from the

glass before gelation of the polymer matrix occurs.The acidbase reaction therefore, dominates thesetting characteristics, and the greater surfacearea of the smaller particles contributes to amore rapid reactionand hence a reduced workingtimeat higher percentage powder B.

At higher powder:liquid ratios, particle packingbenefits achieved by the addition of small particles(powder B) are more clearly seen, and the increasedcohesion has a greater effect on the paste viscositythan the reaction kinetics. Although, the increasedsurface area would hasten the reaction anddecrease the working time of the pasteespeciallygiven the smaller relative volume of liquid phase inthe pastepowder B has a greater effect over thehandling of the paste than over the rate of reaction.Also, as the powder:liquid ratio is increased, so toodoes the necessity of the cohesive influence of thefiner particles.

The increase in strength corresponded closely toa decrease in mean particle size and increase insurface area (Fig. 4). This is consistent withbasic theory of reinforced composite-like struc-tures, where the reinforcement comes from

semi-spherical particles bound to the matrix. Theparticles increase the fracture energy of thecomposite materialthe energy required to propa-gate a crackthereby increasing the strength andtoughness of the set cement. The improvement instrength is especially noticeable when the particlesare intimately bound to the matrix, as is the case for

GICs. The strengthening effect has been previouslydiscussed and analysed [12,14,18,19]. A recentstudy by Mitsuhashi et al. [20] found reducedfracture strength with reductions in particle sizes,but these results are for failures in tension of resin-modified glass-ionomer cements, where the mech-anism of failure is significantly different.

Overall, while the strength of the materials fallsbelow the standards required of restorative glass-ionomer cements [15], it is believed that themechanism of particle size is only partially con-tributory to strength. Other factorsreactivity,glass formulation, surface treatment, etc.have a

much more pronounced effect on strength, and theresults obtained in these experiments can be easilytransplanted. Further work into the particle sizedistributions of glass-ionomer cements, combinedwith continuing activity into the other fieldsmentioned above, will result in an optimal restora-tive, more able to reliably meet the growingchallenges as a versatile material.

Conclusion

Modern GICs provide for the clinician both chal-lenges and opportunities. The requirements of astrong material, combined with optimal cohesionand good working and set times, means thatexperimental investigation into the fundamentalsof particle interaction remains a necessary researcharea. GICs composed of large particles (around10 mm) formed a clay-like, non-cohesive paste,while those composed of finer particles (around3.4 mm) were strong but too fast-setting and viscousfor clinical usage. Mixtures of fine and coarseparticles resulted in a low-viscosity, cohesivepaste with properties similar to commercial GICs.

Further, investigations into the optimisation ofparticle size distributions in glass-ionomer cementswill further enhance the properties and clinicalsuitability of this class of restorative.

Acknowledgements

The authors wish to acknowledge the support of SDILtd (Melbourne, Australia) in the supply ofmaterials and use of experimental equipment.

Particle size effects in GICs 509


6/6

References

[1] Demoor RJG, Verbeeck RMH, Demaeyer EAP. Fluoriderelease profiles of restorative glass ionomer formulations.Dent Mater1996;12(2):8895.

[2] Cattani-Lorente MA, Dupuis V, Moya F, Payan J, Meyer J-M.Comparative study of the physical properties of a polyacid-modified composite resin and a resin-modified glass

ionomer cement. Dent Mater1999;15(1):2132.[3] Levallois B, Fovet Y, Lapeyre L, Gal JY. In vitro fluoride

release from restorative materials in water versus artificialsaliva medium (SAGF). Dent Mater1998;14(6):4417.

[4] Vermeersch G, Leloup G, Vreven J. Fluoride release fromglass-ionomer cements, compomers and resin composites.

J Oral Rehabil2001;28(1):2632.[5] Crisp S, Pringuer MA, Wardleworth D, Wilson AD. Reactions

in glass ionomer cements: II. An infrared spectroscopicstudy.J Dent Res 1974;53(6):14149.

[6] De Maeyer EAP, Verbeeck RMH, Vercruysse CWJ. Stoichi-ometry of the leaching process of fluoride-containingaluminosilicate glass-ionomer glasses. J Dent Res 1999;78(7):13128.

[7] Griffin SG, Hill RG. Influence of glass composition on the

properties of glass polyalkenoate cements. Part I: influenceof aluminium to silicon ratio. Biomaterials 1999;20(17):157986.

[8] Griffin SG, Hill RG. Influence of glass composition on theproperties of glass polyalkenoate cements. Part II: influenceof phosphate content. Biomaterials2000;21(4):399403.

[9] Nicholson JW. Chemistry of glass-ionomer cementsareview. Biomaterials1998;19(6):48594.

[10] Guggenberger R, May R, Stefan KP. New trends in glass-ionomer chemistry. Biomaterials1998;19(6):47983.

[11] Kent BE, Wilson AD. Dental silicate cements. XV. Effect ofparticle size of the powder.J Dent Res1971;50(6):161620.

[12] Brune D, Smith D. Microstructure and strength properties ofsilicate and glass ionomer cements. Acta Odontol Scand1982;40(6):38996.

[13] Wang WP, Dibenedetto AT, Goldberg AJ. Abrasive weartesting of dental restorative materials. Wear1998;219(2):

2139.[14] Gladys S, Vanmeerbeek B, Braem M, Lambrechts P,

Vanherle G. Comparative physico-mechanical characteriz-ation of new hybrid restorative materials with conventionalglass-ionomer and resin composite restorative materials.

J Dent Res 1997;76(4):88394.[15] ISO. ISO 9917 Dental water-based cements. Geneva:

International Organization for Standardization; 1991.[16] Soo SL. Particulates and continuummultiphase fluid

dynamics. New York: Hemisphere; 1989.[17] Davidson CL, Mjor IA, editors. Advances in glass-ionomer

cements. Berlin: Quintessence; 1999.[18] Krishnan VK, Yamuna V. Evaluation of a light cured dental

composite containing a radiopaque glass filler. Curr SciIndia1997;72(3):1925.

[19] Jandt KD, Al-Jasser AMO, Al-Ateeq K, Vowles RW, Allen GC.Mechanical properties and radiopacity of experimentalglasssilicametal hybrid composites. Dent Mater 2002;18(6):42935.

[20] Mitsuhashi A, Hanaoka K, Teranaka T. Fracture toughness ofresin-modified glass ionomer restorative materials: effectof powder/liquid ratio and powder particle size reductionon fracture toughness.Dent Mater2003;19(8):74757.


the effect of particle size distribution

Documents