Influence of powder/liquid mixing ratio on the performance of a restorative glass-ionomer dental

cement
Biomaterials

Volume 24, Issue 23, October 2003, Pages 41734179
Garry J.P. Fleming, ,
Ahmed A. Farooq,
Jake E. Barralet

Show more

DOI: 10.1016/S0142-9612(03)00301-6
Get rights and content

Abstract

The influence of powder/liquid mixing regime on the performance of a hand-mixed restorative glass-
ionomer cement (GIC) was evaluated in terms of compressive strength, working characteristics and the
porosity distribution.

Mean compressive fracture strengths, standard deviations and associated Weibull moduli (m) were
determined from series of 20 cylindrical specimens (6 mm height, 4 mm diameter) prepared by hand-
mixing the relative proportions of the powder and liquid constituents. Working characteristics were
assessed using an oscillating rheometer whilst scanning electron microscopy and image analysis were
used to investigate the influence of the mixing regime on pore distribution.

For a constant volume of liquid (1 ml) the mean compressive strength decreased from 102.123.1 MPa
for 7.4 g of powder, to 93.822.9, 82.618.5 and 55.717.2 MPa for 6.66, 5.94 and 3.7 g of powder,
respectively. A concomitant increase in both the working and setting times was also observed.

GICs manipulated to a powder/liquid mixing consistency below the manufacturers recommend ratio,
for a constant volume of liquid, resulted in reduced porosity levels in the cement mass and extended
working and setting times. Unfortunately, a reduction in the concentration of reinforcing glass particles
in the set material below that specified by the manufacturers decreases the cements load bearing
capacity so that they fail at lower compressive stress levels in the posterior region of the mouth.
Keywords

Glass-ionomer dental cement;
Mixing ratio;
Weibull statistics;
Compressive strength

1. Introduction

Glass-ionomer cements (GICs) are utilised in a variety of different applications from luting cements to
posterior filling materials [1]. Since the introduction of GICs in the early 1970s [2] the cements have
generally been supplied as powder and liquid constituents with the relative proportions being
determined by the technical experience of the operator. The problems identified with hand-mixing GICs
in clinical practice were manifested as powder/liquid variations resulting from operator variability have
been reported previously in the literature [3] and [4]. The authors identified that the GIC investigated
was mixed to a wide range of powder/liquid ratios and the range of mixing ratios did not include the
manufacturers recommended ratio for luting purposes *3+. In an attempt to eliminate operator induced
variability on mixing, GICs have been supplied in encapsulated form where the optimum powder and
liquid proportions are pre-determined and supplied as capsules. A recent study to evaluate the effect of
the mixing method on the compressive strength and porosity of encapsulated GICs reported variations
between materials in mixing method and properties and the overall effect was dependent on the
powder/liquid ratio, the initial viscosity of the cementitious mass and the mixing technique [5].
However, encapsulated GIC systems are more expensive compared with more traditional hand-mixed
cements and consequently, hand-mixed GICs are extensively used in clinical practice.

Variations in hand-mixed cement consistencies utilised by practitioners for posterior filling purposes are
expected as a result of proportioning the powder by eye or with the aid of scoops where the volume of
powder dispensed is dependent upon the method of filling the scoop [1], [4], [5] and [6]. Variations in
liquid volume are dependent upon the positioning of the liquid bottle when held to disperse a drop of
liquid, such that the liquid volume will vary depending on the inclusion of air bubbles [1], [6] and [7].
External factors, namely the temperature and humidity of the mixing environment, and the mixing
conditions (time and manipulation technique) may also introduce operator induced variability in the
cement consistency produced. As a result GICs produced for posterior filling purposes may not have the
same physical, chemical or mechanical properties as cements manipulated according to the
manufacturers recommended mixing conditions. Billington et al. *3+ reported that the compressive or
diametral tensile strength of the luting GIC investigated failed to meet the requirements of the British
Standard [8] when the mean GIC powder/liquid mixing ratio identified in clinical practice was used.
Consequently, the aim of the current study was to investigate the influence of mixing regime for a
restorative hand-mixed GIC in terms of compressive strength, porosity levels within the cementitious
mass and the working characteristics.
2. Materials and methods
2.1. Cement manipulation

The restorative posterior filling GIC used in the current study to investigate the influence of
powder/liquid ratio hand-mixing on cement performance was ChemFil Superior (Dentsply DeTrey,
Konstanz, Germany). The range of powder/liquid mixing ratios identified previously by Billington et al.
[3] was representative of variation in mixing ratio of GICs for luting and filling purposes utilised less
powder than that recommended by the manufacturers. In the current study, Group G100 consisted of a
GIC specimens prepared to the manufacturers recommended mixing ratio (7.40/1.0 g/ml) and further
groups prepared with 90% (G90), 80% (G80) and 50% (G50) of the manufacturers recommended
powder content (6.66, 5.92 and 3.7 g, respectively) for a constant 1 ml volume of liquid. The variability
of the compressive strength of 20 nominally identical cylindrical specimens (height 6.00.1 mm,
diameter 4.00.1 mm) was assessed using an Instron Tensile Testing apparatus (Instron, High
Wycombe, England) in accordance with the procedure outlined previously [7]. The samples were
prepared by a single operator utilising a Teflon split-mould, capable of holding a maximum of five
samples, that was coated prior to filling with a PTFE dry film lubricant (RS Components, Northants,
England) to facilitate the removal of hardened cement specimens. The mould was placed on the base of
an aluminium jig that had been covered with an acetate strip to prevent the setting cement from
adhering to the jig.

The GIC under investigation was prepared on one end of a glass slab and manipulated in accordance
with the manufacturers instructions. The cement was then applied to one side of the mould and packed
to excess with a non-corrodable spatula within 60 s from the end of mixing. The samples were covered
with an acetate strip, isolated from the surrounding atmosphere with a second glass slab and sealed
with a G-clamp. Not more than 120 s after the completion of mixing, the entire assembly was
transferred to a water-bath maintained at 371C in an attempt to simulate oral conditions more
closely. One hour after the completion of mixing, the specimens were removed from the mould, the
ends ground using water as a lubricant on P800 carborundum (Silicon Carbide) abrasive paper (Struers,
Glasgow, Scotland) to remove excess cement. Specimens with non-uniform ends, residual surface
defects or visually apparent pores were discarded and the remaining specimens returned to the water-
bath and stored wet for a further 23 h at which time the test specimens were compressive loaded (24 h
after the completion of mixing).
2.2. Compressive loading

The diameters of the specimens were measured with a micrometer screw gauge (Moore and Wright,
Sheffield, England) accurate to 10 m. The flat ends of the specimens were covered with a wet piece of
filter paper to ensure the specimens were tested wet and a compressive load applied, with a
crosshead speed of 1 mm/min, to the long axis of the specimens. The maximum load to failure was
recorded and the procedure repeated so that a minimum of 20 nominally identical standard cylindrical
specimens had been fractured for each powder/liquid mixing ratio investigated. The compressive
strength, P (MPa), of each individual cylindrical specimen was calculated by dividing the fracture force
(F) by the area of the specimen where D was the specimen diameter:
equation(1)

2.3. Scanning electron microscopy

Cylindrical specimens (Section 2.1) were mounted horizontally in a cold-setting epoxy resin (Epoxy HQ,
Struers, Glasgow, Scotland) and ground flat through a series of silicon carbide abrasive papers (P220,
P500, P800 and P1200 for 30 s followed by P2400 and P4000 for 60 s) on a Dap-7 Pedemin polishing
machine (Struers, Copenhagen, Denmark) at a force of 10 N. Following grinding, polishing was
completed using a DP Dac cloth with a series of diamond paste (6, 3 and 1 m, respectively) at a force of
10 N for 60 s each. The mounted specimens were gold coated for 60 s (Denton Vacuum, NJ, USA), placed
onto aluminium stubs and a strip of silver dag was painted from the specimen to contact with the stub.
The silver dag facilitated the removal of the excess electrons from the specimen surface during Scanning
electron microscopy (SEM) (Jeol JSM 5300LV, Jeol Ltd., Akishima Tokyo, Japan) to prevent zones of
charge. The scanning electron microscope used in backscatter electron mode, which allowed for the
assessment of particles present within the samples at a voltage of 20 kV.
2.4. Image analysis

Image analysis was used to determine porosity within three standard cylindrical cement specimens
embedded in a cold-setting resin and polished in accordance with the procedure outlined in Section 2.3.
Cylinders were serially sectioned transversally to maximise the number of fields in the analysis in
accordance with a procedure outlined previously [9] and [10] based on a trial series of specimens and
the data interpreted according to stereological methods [11] and [12]. Two random non-overlapping
locations from the surface of each specimen were utilised and the procedure repeated at every 0.4 mm
depth to prevent counting pores twice. Consequently, a total of 16 fields per cylinder were analysed
with Optimas, version 4.0 (Bioscan Inc., Washington, USA) which automatically isolated the pores
identified on a transmitted light microscope (Leitz, West Germany) fitted with a Cohu High Performance
4910 Series CCD Camera (San Diego, USA) at 4 magnification. The values, based on image intensity,
returned the largest diameter in calibrated units and frequency tables of the largest diameters were
constructed to approximate the pore distribution. Image analysis used in the current study detected
pores greater than 4 m diameter and the equivalent diameter results (Table 2) provide an indication
of the pore distribution across the nominally identical standard cylindrical GIC specimens. It was also
possible to compare porosity levels across the range of GIC consistencies investigated as a result of the
differences in initial viscosities of the hand-mixed posterior filling cements on mixing.
2.5. Working characteristics

The working and setting times of the GIC at the powder/liquid mixing ratios under investigation were
determined using an oscillating rheometer [13] where the oscillation amplitude was measured and
plotted against time using a chart recorder. The time for the amplitude to reduce to 100% of its value
was taken as the setting time of the cement which was performed at 371C and working time was
taken as the time for the amplitude to reduce to 95% which was performed at 231C [14].
2.6. Statistical analysis

Multiple comparisons of group means were made using a one-way analysis of variance (ANOVA) and a
Tukey's multiple range test was employed at a significance level of P<0.05. The compressive fracture
strength data was ranked in ascending order and a Weibull analysis [15] performed on the resultant
data. The basic form of the Weibull distribution is
equation(2)
where 0 and m are all constants. m is known as the Weibull modulus characterising the brittleness of
a material. A higher value of m indicates a close grouping of the flexure stress data whilst a lower value
represents a larger scatter in the flexure stress data *16+. 0 is the normalising constant. Confidence
limits for the groups were calculated and differences were considered to be significant when the
confidence intervals did not overlap [17]. The stress at the 10% failure probability was calculated for
each consistency investigated.

3. Results
3.1. Compressive strength

The compressive fracture strength data for the hand-mixed GIC resulted in significant differences
between the mean compressive strength of groups G100-G50 prepared to different powder/liquid
posterior filling consistencies (Table 1), when the one-way ANOVA and Tukey test comparisons were
analysed (P<0.05). Decreasing the powder content to 80% and 50% of the manufacturers recommended
posterior filling consistency G80 and G50, respectively, resulted in a significant decrease in the mean
compressive fracture strength data and an associated reduction in the 95% compressive failure stress (
Table 1). Reducing the powder content to 50% the manufacturers recommended posterior filling
consistency increased the associated m although the differences were not considered to be significant
as the 95% confidence interval of m significantly overlapped with the other powder/liquid posterior
filling consistencies investigated.

Table 1.

Compressive fracture strength data for series of 20 cylindrical specimens prepared to varying powder
contents for a constant volume of liquid (figures in brackets represent standard deviations)
Property G100 G90 G80 G50
Compressive strength range (MPa) 65.7134.7 54.8144.4 48.1122.5 35.176.3
Mean compressive strength (MPa) 102.1 (23.1)a 93.7 (22.9)a,b 82.6 (18.5)b 55.7 (17.2)c
Weibull modulus (m) 4.5 (0.5) 4.4 (0.5) 4.5 (0.5) 5.3 (0.5)
Confidence intervals (95%) 4.04.7 4.04.8 4.14.9 4.65.9
R2 value 0.94 0.97 0.97 0.94
95% failure stress (MPa) 71.2 63.5 52.5 42.6

Note: Different superscripts represent significant differences at the 95% level when the one-way
ANOVA and Tukey's multiple range test were employed.

Table options

3.2. Scanning electron microscopy

SEM examination of the polished hand-mixed restorative specimens to 200 revealed no discernible
change in the morphology of particles of the hand-mixed GIC restorative (Fig. 1ac). However,
examination by SEM did provide an indication of the degree of porosity, filler size and distribution in the
hand-mixed cements (Fig. 1ac). Unfortunately, a two-dimensional cross section of a GIC cylinder does
not provide any quantifiable information on the distribution of pores within the specimen and as a
result image analysis studies were carried out.

Full-size image (55 K)
Fig. 1.

Scanning electron micrographs of the GIC specimens prepared using powder contents of (a) 7.40, (b)
5.92 and (c) 3.20 g for a constant volume of liquid (1 ml). The micrographs illustrate a two-dimensional
cross section highlighting the increased number of reinforcing powder particles and concomitant
increase in porosity at the higher mixing ratios.
Figure options

3.3. Image analysis

Hand-mixing using 100% and 90% the manufacturer's recommended powder content for a restorative
filling material (with 1 ml of liquid), G100 and G90, respectively, increased the number of large pores
(>26 m) compared with using 80% and 50% the manufacturers recommended GIC powder content
(G80 and G50, respectively). As a result the maximum pore diameter was increased with increasing
powder content for a constant volume of liquid.
3.4. Working characteristics

In general, hand-mixing the cement to powder liquid mixing ratios below that recommended by the
manufacturers increased the working and setting times of the GIC under investigation compared with
hand-mixing at the recommended posterior filling consistency (Table 3). No significant differences in
working times were discernable when the one-way ANOVA was employed at the 95% significance level.
However, significant reduction in setting times were noted when using 80% and 50% the manufacturers
recommended GIC powder content (G80 and G50, respectively) compared with G100, when the one-
way ANOVA was employed at the 95% significance level.
4. Discussion

Billington et al. [3] identified that variances between operators can influence mechanical properties,
namely compressive and diametral tensile strength, and that hand-mixed GICs are regularly manipulated
to consistencies below that recommended by the manufacturer in clinical practice. As a result, the
current study reviewed the effect of mixing a GIC to a range of consistencies by a single operator. In line
with previous observations *3+ and *4+ the mixing ratios used included the manufacturers
recommended mixing ratio and further ratios with 90%, 80% and 50% powder content for the
manufacturers recommended posterior filling consistency manipulated with a constant volume of
liquid, similar to those commonly used in clinical practice. The compressive testing regime based upon
the specifications outlined in BS 9917 for water-based cements [18] was employed as the method allows
sensitive changes in the mechanical properties of brittle materials to be distinguished through changes
in composition and levels of porosity [5]. The use of diametral tensile strength as used previously [3] was
not employed due to difficulties in specimen loading to achieve an intimate point contact [19].

The microstructure of GICs is formed as a result of the acidbase reaction between the proton donating
acidic liquid and proton accepting basic powder resulting in filler glass particles distributed within a salt-
like hydrogel [20]. The compressive fracture strength of the cement arises from the reinforcing glass
filler particles, which resists compressive forces on loading rather than the weak matrix [21]. Variation in
the restorative filling consistencies used in the study had a causative effect on the compressive strength
data (Table 1) where the mean compressive fracture strength decreased with decreasing powder
content for a constant volume of liquid. The reduction in the volume of reinforcing glass particles which
acted as a strengthening filler reduced the GICs ability to resist the compressive forces induced on
loading and was manifested as failure at lower compressive loads [4], [21], [22], [23] and [24]. This was
demonstrated in the survival probability distributions for the compressive fracture strength data for the
posterior filling consistencies investigated where the 95% failure stress level was identified as being
significantly decreased with powder contents below that recommended by the manufacturer for a
posterior filling material.

SEM has been employed previously as a tool to provide an indication of surface morphology, filler size
and distribution and porosity [24], [25], [26], [27] and [28]. Fig. 2 identifies the particulate filled
composite structure with reinforcing glass particles within an extensively porous cement matrix. As the
powder content was decreased a reduction in the volume of reinforcing glass particles was evident and
a concomitant decrease in porosity level with reduced powder content (for a constant volume of liquid)
was also evident (Fig. 1ac). Unfortunately, a two-dimensional cross section through a GIC cylinder does
not provide any quantifiable information on the distribution of pores within the specimen since random
sectioning through a porous solid produces a distribution of sectioned diameters [11] and [12]. The
sectioned diameter varies from the lower bound limit (when the pore is sectioned at the smallest
possible diameter) to the true diameter (when the pore is sectioned at the largest diameter). However,
porosity levels can in theory be assessed using the size distribution of sectioned pore areas so that the
area-equivalent diameter (the diameter of a disc with an area equal to the section area of a pore) can
be determined by assuming the pore shape is spherical and hence its section to be circular [9] and [10].
In accordance with the serial sectioning technique outlined in Section 2.3 and previous studies [11] and
*12+ the integrated frequency histogram of the area-equivalent diameters was constructed to provide a
lower bound for the pore distributions within the specimens (Table 2).

Full-size image (4 K)
Fig. 2.

Probability of survival versus compressive fracture stress for ChemFil Superior specimens mixed to
varying restorative filling consistencies.
Figure options

Table 2.

Lower bound porosity distributions within cylindrical specimens prepared for ChemFil Superior when
the serial sectioning technique was employed (figures in brackets represent standard deviations)
Pore diameter (m) G100 G90 G80 G50
46 60.0 (7.5) 49.7 (19.4) 45.0 (15.7) 41.2 (15.7)
610 57.3 (7.0) 54.3 (15.0) 46.0 (4.6) 44.3 (2.1)
1014 21.0 (9.0) 28.0 (4.4) 20.0 (7) 12.0 (2.6)
1418 14.0 (3.5) 14.0 (7.8) 13.0 (2.6) 5.3 (3.2)
1822 6.0 (1.7) 8.7 (5.0) 5.3 (2.1) 5.7 (1.5)
2226 5.3 (0.6) 5.0 (1.0) 2.3 (1.5) 3.7 (1.2)
2630 3.0 (1.0) 4.3 (0.6) 1.7 (1.2) 2.0 (1.7)
3034 3.0 (1.0) 3.3 (1.5) 1.0 (1.0) 0.7 (0.6)
3438 1.7 (1.2) 3.3 (0.6) 0.3 (0.6) 1.3 (1.2)
3842 2.0 (2.6) 2.0 (0.0) 0.7 (0.6) 0.7 (1.2)
>42 6.7(6.4) 10.0 (2.0) 1.3 (1.5) 1.7 (0.6)
Maximum pore size 157.7 m 117.0 m 86.1 m 50.6 m
Table options

It is likely that porosity levels identified in Table 2 resulted from a combination of two major sources.
Firstly, porosity introduced on manipulating the powder and liquid constituents, dictated by the
spatulation force used to mix the powder and liquid. Increasing the powder content for a constant
volume of liquid would be expected to increase the spatulation force required to adequately mix the
cementitious mass and therefore result in increased porosity with increasing powder content [29] as
illustrated in Table 2. Secondly, larger pores introduced on filling the split-mould where it was possible
that air bubbles may become incorporated or entrapped into the cement mass on filling resulting in
internal porosity during filling. Previous studies on a zinc phosphate cement identified that increasing
the amount of powder, mixed with a constant volume of liquid (under similar mixing conditions), fluid
mixes flowed more easily into the mould than more viscous mixes [29] which resulted in cylinders which
were consolidated better with less internal porosity. A similar trend is highlighted for the GIC
investigated whereby large voids in excess of 42 m diameter were formed on consolidating hand-mixed
cements prepared at higher mixing ratios. The largest pore diameters recorded for cements prepared
using 3.70, 5.94, 6.66 and 7.4 g of powder (for 1 ml of liquid) were 50.6, 86.1, 117.0 and 157.7 m,
respectively, which confirmed that larger voids were formed on consolidating cements prepared to
higher powder/liquid mixing ratios. The survival probability against the compressive fracture strength
plot for the powder/liquid consistencies investigated highlight that the distribution of fracture strengths
for group G50 are narrower. This is indicative of a higher Weibull modulus; however, the reduced
porosity level in group G50 compared with G100, G90 and G80 is responsible for the increased m.
4.1. Working characteristics

The oscillating rheometer [13] was used to measure the setting characteristics of the GIC mixed to the
manufacturers recommended consistency and to consistencies below that specified by the
manufacturer. The oscillating rheometer has caused controversy among the dental profession [30], [31]
and [32] as the sensitivity of the device is determined by the spring tension that reduces motion of the
lower platen. If the chosen springs were weak the initial changes during setting were emphasised;
conversely if the springs were too strong the final stages of setting would be more discernible. During
the current study a compromise between the two spring tensions was selected in accordance with the
recommendations of the designers [13] in an attempt to minimise problems associated with the
accuracy of the apparatus resulting from spring tension. The use of hand-mixed cements prepared to
consistencies below that recommend by the manufacturers extends the working time available to the
operator (Table 3). The increased fluidity of the mass for cements prepared below the manufacturer's
recommended posterior filling consistency (compared with the recommended consistency) allows
increased time for the operator for more precise placement. Unfortunately, extending the working and
setting times to 20.2 and 3.50.01 min, respectively, for a powder/liquid ratio of 3.70 g/ml from
1.50.02 and 2.50.03 min, respectively, at the manufacturers recommended restorative filling
consistency would increase both the likelihood of moisture damage to the surface of the hardened
cement to which GICs are especially susceptible [33].

Table 3.

Working characteristics for the series of three cement specimens prepared to varying powder
contents for a constant volume of liquid (figures in brackets represent standard deviations)
Working time (min) Setting time (min)
G100 1.5 (0.02)a 2.5 (0.03)a
G90 1.5 (0.7)a 2.8 (0.3)a,b
G80 1.8 (0.1)a 3.1 (0.06)b
G50 2.0 (0.2)a 3.5 (0.01)b

Note: Different superscripts represent significant differences at the 95% level when the one-way
ANOVA and Tukey's multiple range test were employed.

Table options

Utilising a cement manipulated to the manufacturers recommended powder/liquid mixing ratio (7.4/1.0
g/ml) to place a posterior restoration reduced the working time (compared with cements manipulated
to lower powder contents for a constant volume of liquid) so that the cement may possibly have set
before the restoration had been completely placed. Ideally, a GIC restorative filling material should
allow the operator sufficient time for placement of the restoration, limit the effect of moisture damage
to the surface of the hardening cement whilst maintaining adequate mechanical properties.
Unfortunately, it is likely that GICs manipulated in practice at mixing ratios below that specified by the
manufacturers do not have adequate strength in terms of the masticatory forces likely to be
encountered in service.
5. Conclusion

Powder/liquid mixing ratio influences the mechanical properties and therefore possibly the resultant
clinical performance of the high strength posterior GIC investigated. Reducing the powder content for a
constant volume of liquid reduces the porosity levels in the cement mass and extends the working and
setting times. Unfortunately, decreasing the powder content reduces the concentration of reinforcing
glass particles in the set material and is manifested as a reduction in the load bearing capacity of the GIC
in service.