Physical Property Evaluation of Four Composite Materials Open Access

Manufacturers of dental materials have attempted to distinguish their products by adjusting composite formulations to find a market niche. In an effort to identify which composition will have adequate clinical performance and meet minimum standards, physical property tests are used to evaluate the materials.

Polymerization shrinkage of a dental composite is of significant clinical importance because shrinkage leads to mechanical stresses on enamel and dentin at the adhesive interface.¹  The degree of shrinkage of a composite is related to the molecular weight of the monomer, the amount of monomer in the composite, and the degree of polymer conversion. The polymerization shrinkage of low-molecular-weight monomers is more pronounced than that of high-molecular-weight monomers.¹  Shrinkage can lead to marginal leakage when the restoration does not adhere to the tooth structure or when the stress generated at the restoration-tooth interface causes the composite to debond from the tooth structure.²  Marginal leakage and cohesive failures have been generated by stresses of as little as 2% of volume shrinkage of typical composites.³ 

There are two general methods to test polymerization shrinkage: non–volume dilatometry and volume dilatometry.⁴  Non–volume dilatometric methods include the use of a linometer^5,6  and video imaging of a sample as it undergoes polymerization.⁷  A mercury dilatometer⁸  and a water-filled dilatometer²  are common methods of volumetric measurements for shrinkage. A significant consideration in the use of a mercury or water dilatometer is that the fluids are very sensitive to ambient temperature. Thermal control is mandatory; otherwise, the expansion or contraction of the fluids will affect the results.⁸  In addition, application of light to polymerize an experimental resin is difficult because of the opacity of mercury.

Strength (tensile, shear, compressive, or flexural) values are often relied on as indicators of structural performance for brittle dental materials, including composites.⁹  The three-point flexural strength test has been used to measure mechanical properties of a composite material and has been selected by the International Standards Organization (ISO) for screening resin-based filling materials.^9,10  Flexural strength is defined as the failure stress of a material as measured in bending.¹¹  Composite restorations are subjected to flexural stresses, especially in stress-bearing cavities (Class I, Class II, and Class IV).¹²  Flexural strength has been shown to be a more discriminating test and more sensitive to subtle changes in a materials substructure than compressive strength.¹³  Currently, continuous loading machines are employed in flexural strength testing, which can be conducted using either three-point or four-point loading.^14,15 

The satisfactory clinical performance of a restorative material is partially determined by its adaptability to prepared tooth structure and its resistance to degradation in the oral environment.¹⁶  Degradation is the result of a combination of effects including abrasive wear, masticating forces, and damage by oral fluids. The hardness of a composite is a quality of the material that resists the degradation forces and improves the function of the restoration. The use of hardness tests has become prevalent because of their relatively simple testing method and the reliability of the results.¹⁷  Another advantage that has been cited is the correlation between hardness values and degree of conversion.¹⁸  In general, Knoop hardness is the most commonly indicated method for evaluation of polymeric materials, such as resin composites, because it minimizes the effect of elastic recovery. However, other authors have used the Vickers hardness test as an indicator of the degree of resin polymerization.^19,20  The Vickers indentation involves a larger surface area and is more representative of the surface of multiphase materials.

This study was designed to compare flexural strength, polymerization shrinkage, and surface hardness of four composites, selected on the basis of filler content and resin composition: Filtek LS (3M-ESPE), Aelite LS (Bisco), Kalore (GC America), and Empress Direct (Ivoclar). In addition, a comparison of the Knoop and Vickers hardness test was made to determine if there was a correlation in the precision between the two. The null hypotheses tested were that there were no differences in each of the physical properties among the four composites and that there is not a significant direct correlation between Knoop and Vickers hardness tests.

This study was divided into three parts to evaluate the following physical properties: polymerization shrinkage, surface hardness using Knoop and Vickers and flexural strength. For each part, 10 samples of each composite were made in shade A2 as listed in Table 1.

Measurements for linear polymerization shrinkage were performed with a linometer as described by Herrero and others⁶  and are shown as a schematic representation in Figure 1. The setup consisted of a noncontact displacement transducer (Kaman Instrumentation Corp, Colorado Springs, CO) with the sensor placed in a vertically oriented quartz tube. The target and samples are positioned at the required offset distance of 0.13 mm and within the 1-mm range below the sensor. The Kaman linometer was calibrated at the beginning of each measurement period with a fixed sensor, and the aluminum target was placed at 0, 500, and 1000 μm distances, such that each unit = 1 μm on a standard micrometer gauge.

The samples of disk-shaped uncured composite were approximately 5 mm in diameter and between 1.5- and 2.0-mm thick. All the procedures were performed at room temperature. A cylinder-shaped specimen of each composite resin was extruded from the syringe and preflattened on the mixing pad to the desired thickness. The samples were transferred to a glass slide that was coated with a separating medium (Al-Cote, Caulk, Dentsply, Milford, DE) to allow the composite to shrink free of surface adhesion.

The top of the sample was covered by a flat aluminum target (10-mm diameter × 1-mm thick) placed parallel to the glass slide and coated with a lubricant. The entire assembly was mounted in a vertical position to use gravity to maintain the aluminum target position on the top surface of the sample (Figure 1). Once in position, the composite specimens were polymerized for 40 seconds using a Smartlite iQ2 (Model No. 200, Dentsply). Linear shrinkage data were recorded as ΔL after 40 seconds of light exposure. The percentage of shrinkage was calculated using the formula proposed by de Gee and others⁵  and used by Herrero and others⁶:

in which ΔL is the recorded displacement in micrometers and L is the thickness of the specimen in micrometers after polymerization. The Lin% is converted to a volumetric value using the following formula^5,6  in which the last term is negligible:

A metal mold, lubricated with Al-Cote, was used to fabricate composite samples that were 14 mm in diameter and 2 mm in thickness. Under low-light conditions, composite was placed in the mold, and the excess was removed using a wax spatula. A lubricated glass slide was placed on top of the sample, and finger pressure was applied to achieve a smooth surface and good adaptation of the composite in the mold. Each sample was light cured for 40 seconds through the glass slide using the Smartlight iQ2. The samples were removed, and the bottom or the side of the sample that was against the base was labeled using a permanent marker.

The light-cured side of each sample was polished for 5 seconds in an elliptical pattern with 600-grit wet silicon carbide paper (MAGER Scientific, Dexter, MI) to remove the resin-rich surface. A rotary grinder-polisher (Micro Star 2000, Concord, Ontario, Canada) was used at a rotation speed of 102 rpm to polish the samples. Each sample was first polished with a 9-μm polycrystalline diamond suspension on a polishing cloth for four minutes, followed by a 1-μm diamond suspension for two minutes on a finer polishing cloth (PC-532 Pan, MAGER Scientific). Each sample was rinsed with tap water and kept in an envelope in a dark drawer for 24 hours at room temperature prior to testing. Following the storage period, a Tukon 2100B (Wilson Instrument, Norwood, MA) was prepared and calibrated for the microhardness testing.

For Knoop hardness, a 200-g load was applied to the sample surface with a dwell time of 15 seconds. After removal of the load, the microscopic indentation obtained from this procedure was evaluated for length and/or width as a function of penetration. For Knoop hardness, the diamond is rhombic shaped, and measurements are performed on the longest diagonal of the impression.

Three indentations were made on each sample, and the filar readings were measured using 20× magnification (Figure 2A). The three readings were averaged to produce a single hardness value for each sample. Knoop hardness is the ratio of the load applied to the area of the indentation calculated from the following formula¹⁵:

where L = the applied load in kilograms, I = the length of the long diagonal of indentation (mm), and Cp = a constant related to the area of the indentation (0.07028). Means and standard deviations were then calculated.

Vickers indentations were made with a 500-g load, applied for 15 seconds. The diamond indenter is pyramidal, and hence a square-shaped indentation is obtained. Measurements were made on both diagonals, and mean values were obtained. The first measurement was made by positioning the filar lines at the tips of the horizontal indentation. For the second measurement, the eyepiece assembly was rotated 90° clockwise, and the filar lines were placed on the tips of vertical axis of the Vickers indentation (Figure 2B). Three readings were made, and they were averaged to represent a Vickers hardness value for each sample. The Vickers hardness number is computed from the following equation¹⁵:

where P = test force in kg and A = surface area of indent in mm². Means and standard deviations were then calculated.

Samples of each composite were prepared in a split aluminum mold 25 ± 2 mm × 2 ± 0.1 mm × 2 ± 0.1 mm in accordance with ISO specification No. 4049/2000 and ANSI/ADA specification No. 27. The base and mold were lubricated with a thin layer of Al-Cote. Under low-light conditions, sufficient composite was applied to fill the mold, and the excess composite was removed using a wax spatula. A thin coat of Al-Cote was applied to a clean glass slide, and the slide was placed on top of the sample. Finger pressure was applied to achieve a smooth surface and good adaptation of the composite. Each sample was light cured for 40 seconds through the glass slide using the Smartlight iQ2. Since the length of the rectangular bars for the three-point flexure test exceeded the diameter of the curing-light tip, three overlapping curing times were employed until the entire length of the samples was covered. During all sample preparation, light intensity (620 mW/cm²) was checked periodically with the Cure Rite radiometer (Efos Inc, Mississauga, Ontario, Canada). After polymerization, the specimens were finished using 600-grit wet silicon carbide paper and stored in distilled water at room temperature for 24 hours prior to testing.

The specimens were mounted on a jig, and the load was applied with an Instron Universal Testing Machine (Instron 5565, Instron Corp, Norwood, MA) at a crosshead speed of 0.75 ± 0.25 mm/min until the sample fractured. The maximum load exerted on the samples was recorded, and the flexural strength at failure was calculated by the following formula:

where P = maximum load in Newtons, S = distance between supports (20 mm) accurate to ±0.01 mm, and B and H are the width and thickness of the sample (2 mm) measured immediately prior to testing.

Multiple comparisons of material group means for each property were made using a one-way analysis of variance (ANOVA). A Tukey's multiple comparison test was then used to determine significant differences between the means at 95% confidence (p<0.05). A Pearson correlation coefficient was calculated to test for correlation between the Knoop vs Vickers hardness tests (p<0.05).

Four commercially available resin composites were evaluated for polymerization shrinkage, flexural strength, and surface hardness (n=10). Products were selected based on resin composition and filler particle size. In addition, a comparison of Knoop and Vickers hardness tests was made to determine if there was a correlation between the two tests.

Table 2 illustrates the range, means, and standard deviations for volumetric shrinkage. Filtek LS had a significantly lower shrinkage value than the other materials, based on a one-way ANOVA test (p=0.001) and a Tukey multiple comparison test (p<0.05). Aelite LS, Kalore, and Empress Direct have statistically similar but higher shrinkage values.

The distribution of volumetric shrinkage values for the 10 samples within each material showed that Empress Direct exhibited the highest variation in values among samples (range=2.12) compared with the other material groups. On the other hand, Filtek LS showed the most consistent values (range=0.45).

Table 3 illustrates the range, means, and standard deviations for Knoop hardness. Aelite LS exhibited a significantly higher mean value compared with the other materials. Filtek LS was significantly lower than Aelite LS; Kalore was significantly lower than Filtek LS, and Empress Direct exhibited the lowest value. Statistical analysis demonstrated significant differences between all the materials (p<0.05).

The distribution of Knoop hardness values for the 10 samples within each material showed that Aelite LS had greater variation in values (range=31.63) than the other materials. Filtek LS, Empress Direct, and Kalore all had similar variation (ranges=5.10-5.84).

Table 4 illustrates the range, means, and standard deviations for Vickers hardness. Aelite LS exhibited a significantly higher mean value than the other materials. Filtek LS was significantly lower than Aelite LS; Kalore was significantly lower than Filtek LS, and Empress Direct exhibited the lowest value. Statistical analysis showed significant differences between all the materials (p<0.05).

The distribution of Vickers hardness values for the 10 samples within each material again showed that Aelite LS had more variation in values (range=19.24) than the other materials. Filtek LS had less variation than Aelite LS (range=7.03), and Kalore had less variation than Filtek LS (range=5.47). Empress Direct had the least amount of variation (range=3.20).

Figure 3 illustrates the comparison between the Knoop and Vickers hardness data for all materials. The Pearson correlation coefficient of the Knoop vs Vickers hardness tests showed that there was a very strong correlation between the two tests (r=0.991), with Vickers hardness values always significantly greater (independent t-tests, p<0.01).

Table 5 illustrates the range, means, and standard deviations for the flexural strength test. Aelite LS exhibited the highest value compared with other materials but was not significantly different from Filtek LS. Kalore and Empress Direct were similar but were significantly lower (p<0.05).

The distribution of flexural strength values for each material showed that Aelite LS had the greatest variation (range=55.01), while the other materials showed similar but lower variation (ranges=25.28-30.00).

This study tested the following hypotheses:

1. Ho₁: 

   There is no significant difference in polymerization shrinkage of different composite systems.

2. Ho₂: 

   There is no significant difference in flexural strength among different composite systems.

3. Ho₃: 

   There is no significant difference in surface hardness among different composite systems.

4. Ho₄: 

   There is no significant difference in the precision between Knoop and Vickers hardness tests.

There was a statistically significant difference among the materials tested for each physical property. Filtek LS had less shrinkage than the other composites. Aelite LS showed greater hardness for both Knoop and Vickers tests. Filtek LS had greater hardness than Kalore and Empress Direct. Aelite LS and Filtek LS had greater flexural strength than Kalore and Empress Direct. Therefore, the first three null hypotheses were rejected, and the fourth hypothesis was accepted. Composite shade A2 was selected for this study since it has a greater potential for full depth of cure at 2-mm thickness and has been used in prior research protocols.²¹ 

There are two general approaches for determining material shrinkage: volume dilatometry or non–volume dilatometric methods.⁴  Dental composites polymerize from a viscous to a rigid material, and during the polymer conversion, the resin matrix produces a gel. It is at this gel point that the material can no longer provide viscous flow to keep up with the curing contraction. Therefore, the results of shrinkage determinations are dependent on the flow of the material. Conclusions regarding shrinkage values where the displacement transducer requires activation can monitor only the postgel part of the curing contraction because that is when the material is sufficiently strong to exert forces.²²  In general, volumetric curing contraction determinations account for both the pregel and postgel curing contraction.²  These are termed free shrinkage measurements. Non–volume dilatometric experiments are usually designed with either a contacting or noncontacting transducer. The dimensional change during contraction is considered “hindered” and is generally thought to account for only the postgel contraction.²²  For this reason, the various linear shrinkage determinations are not standardized and are hard to compare.

The dental literature reports many devices in which volumetric polymerization contraction has been measured. Many experiments are designed with either a mercury dilatometer^8,23  or a water dilatometer.²  Dilatometry is arduous and protracted. It is also subject to data scattering when used for low-viscosity resins. Mercury dilatometry requires the handling of mercury (a toxic material), and as with water dilatometry, strict adherence to temperature control of the liquid medium is required. Maintaining a constant temperature environment for the water dilatometer is critical. To illustrate this point, the thermal expansion coefficient of water at 20°C is 20.7 × 10⁻⁴. For a dilatometer that contains 25 mL of water, a 0.1°C increase in water temperature will cause an increase of 2.67 mm in meniscus reading. A typical meniscus drop associated with the shrinkage in the work by Lai and Johnson²  was 20 mm for the samples. Therefore, a 0.1°C temperature fluctuation could introduce 13.3% error. Considering the limitations of mercury and water dilatometry, these methods were not chosen for this experiment.

A more recent experimental design is the use of video imaging to measure volumetric shrinkage.⁷  Video imaging has been shown to produce volumetric shrinkage measurements of composites that are comparable to those obtained with mercury dilatometry.⁷  In addition, it is free of the hazards of handling toxic materials such as mercury. However, the instrumentation is expensive and not readily available.

Non–volumetric dilatometric experiments measuring polymerization shrinkage can be determined linearly.²⁴  In 1991, Watts and Cash⁴  used a method to determine volumetric shrinkage indirectly by calculating the postgel linear displacement of a deflecting disk placed on composite. This technique measured the shrinkage between a glass microscope slide and a flexible glass diaphragm using a linear variable differential transformer displacement transducer. Grajower and Guelmann²⁵  determined the dimensional change of glass polyalkenoate cements by means of a linear displacement transducer. Feilzer and others²⁶  presented a modified technique for measuring the wall-to-wall shrinkage of composites. They used a linometer to determine the linear curing contraction from the start of the setting reaction. The linear contraction was determined by means of a noncontact displacement transducer. A thin layer of grease was needed to maintain contact between the two walls of the device and to ensure free shrinkage of the composite, thus determining the total (pregel and postgel) shrinkage. The current project methodology was patterned after the work of Herrero and others⁶  employing a KμDA (Kaman Instrumentation Corp) contactless displacement instrument to measure the linear shrinkage of composite polymerized with an LED curing light.

One challenge in conducting the experiment was to reduce the amount of ambient light during the testing process. During each measurement, the laboratory lights were turned off, and the testing assembly was located in a remote area of the room to minimize the effect from air drafts. The composite was pliable and did not appear to set prior to the start of light exposure. Future shrinkage experiments should be done in total darkness or at least in a filtered environment that eliminates the ambient light that would cause the material to precure before polymer activation.

Another challenge in this experiment was the preparation of uniform uncured composite samples. To make the prepolymerization depth of 2 mm for each sample, approximately 2 mm of composite was expressed from the compule and placed on a glass slide in the center of a 2-mm-thick × 10-mm-wide Teflon ring. A second glass slide was lightly coated with Alcote and placed on the top of the composite. Finger pressure was applied to the top slide so that it eventually rested on the top of the ring. However, the composite tended to adhere to the top glass slide, and some manipulation was required to release it from the sample. Thus, there was opportunity to develop voids and nonuniform sample thicknesses.

In the present study, Filtek LS had significantly less shrinkage than the other materials. Filtek LS was the only composite with the ring opening organic matrix, silorane. Aelite LS used ethoxylated Bis-GMA, Kalore employed urethane dimethacrylate and a proprietary matrix formula DX-511, and Empress Direct was composed of dimethacrylates. According to the manufacturer, silorane chemistry does not contain methacrylates, and the ring-opening monomers allow for lower polymerization shrinkage.

In this study, Aelite LS showed greater values for both Knoop and Vickers hardness, likely due to the high inorganic filler content of that material. Filtek LS had greater hardness than Kalore and Empress Direct, possibly due to the polymerization characteristic of the silorane matrix. Significant differences in hardness among the materials tested are related to the resin formulations, polymerization kinetics, and the type and loading of the filler particles.

For polymeric materials, an elastic recovery is present after removal of the load, and it could affect both measurements in the Vickers indentation. In a Knoop indentation, the elastic recovery would affect the shorter diagonal more than the longer diagonal. Consequently, since the Knoop measurement is based on the longest diagonal, the occurrence of elastic recovery will have less of an impact on the Knoop measurements than the Vickers measurements.¹⁷ 

When the results for the Knoop and Vickers hardness tests are compared, there was a strong correlation between results obtained using the two tests. The Vickers values were slightly higher than Knoop hardness measurements, and this could be due to the elastic recovery. The depth of the indentation is also related to the homogeneity of the surface being tested. One area of a sample may have higher filler content or larger particle sizes than another similar area and thus produce greater variation. This variation is most notable in Aelite LS, where the filler content is higher and the range of hardness values is much greater, especially for the Knoop test. Another factor is that the Knoop value is determined from a linear measurement, while the Vickers value is determined from an area measurement. Therefore, it is not possible to say which values are the most accurate. It would appear that either test could be used to make material comparisons, but the Vickers test may be more relevant. As a result, the fourth null hypothesis can be accepted.

An alternative measurement involves the use of depth-sensing devices. These offer advantages because of their high resolution and ability to obtain information regarding the elastic properties of the tested material.²⁷  In this technique, the depth of the indentation is measured instead of a surface dimension. The depth-sensing device reports depth alterations under load and unload cycles. Therefore, the hardness and elastic modulus may be automatically reported. This is different from the visual measurement performed in a conventional hardness test. Lower loads are used, which produce much smaller indentations than Knoop and Vickers hardness tests. The Ultra Micro-Indentation System has been used to measure the hardness and elastic modulus of dental structures and is considered a reliable and reproducible method with relative operator ease in measuring.²⁸ 

Composite restorations are subjected to flexural stresses, especially in stress-bearing cavities (Classes I, II, and IV).¹²  The flexural three-point bending test has been used to measure mechanical properties of a composite material and for predicting clinical performance of restorative materials (ISO specification 4049).¹⁰  It is stated that ISO specifications for height and width are acceptable because the dimensions of the specimens permit effective polymerization.⁹ 

In this experiment, the sample bars (25 × 2 × 2 mm) were prepared in accordance with the ISO 4049 specification. The length of the bar-shaped specimen exceeded the exit window diameter of the curing-light tip, so an overlapping light-curing exposure was required as curing progressed down the length of the bar. Recent studies have questioned the effect of the overlapping method on the flexural properties of dental composites, whereby the overlapped regions of bar-shaped specimens are subject to an increased light energy density, resulting in specimens that are cured inhomogeneously.⁹  It was also observed that the ISO 4049 standard is not relative to clinical placement of composite restorations since the geometry of the large bar-shaped specimen is disparate to that of actual direct fillings.

A possible solution is the use of biaxial flexure strength tests. This process has been recognized for the evaluation of dental ceramics²⁹  and has advantages over uniaxial, diametral tensile, and compressive strength tests.³⁰  One study showed that biaxial flexure strength tests provided a more reliable testing method than did three-point flexure tests.³¹  The shape for a biaxial test disk (12-mm diameter × 2-mm thick) allows for a complete reproducible single-exposure cure without the need for overlapping curing light exposures.

In the present study, the flexural strength of Aelite LS and Filtek LS were not significantly different, but both were significantly greater than Kalore and Empress Direct, which were not significantly different. Increasing the load of reinforcing filler particles has enhanced the mechanical properties of composites. Aelite LS has a higher filler content than the other three materials, and this could account for the greater amount of flexural strength and surface hardness. However, Filtek LS had the lowest filler content and yet high flexural strength. This could be due to the use of silorane as a matrix material, which may improve resin resistance to flexural stress.

Within the limits of this in vitro study, the following conclusions were drawn:

1. 1

   There were statistically significant differences in polymerization shrinkage, surface hardness, and flexural strength among composite materials. Therefore, the first three null hypotheses of this study were rejected.

2. 2

   For polymerization shrinkage, Filtek LS had significantly lower shrinkage than Aelite LS, Kalore, or Empress Direct, which were not significantly different.

3. 3

   For surface hardness, all materials were significantly different for both Knoop and Vickers tests. Aelite LS > Filtek LS > Kalore > Empress Direct for both tests.

4. 4

   For flexural strength, Aelite LS and Filtek LS had significantly greater flexural strengths than Kalore and Empress Direct. Aelite LS = Filtek LS > Kalore = Empress Direct.

5. 5

   The Vickers hardness values were significantly higher than Knoop hardness measurements for each material. There was a strong correlation (r=0.99) between the two hardness tests. The fourth null hypothesis was accepted.

The United States Navy for graduate student support.

Delta Dental Fund for partial support of this research.

3M ESPE, Bisco, Kerr, GC America, and Ivoclar Vivadent for providing composite materials.

The authors of this manuscript certify that they have no proprietary, financial, or other personal interest of any nature or kind in any product, service, and/or company that is presented in this article.