INTRODUCTION

Despite the emergence of novel, high-performance glass ionomer cements1,2), resin composites are still the most widely used materials for dental restoration3,4). Among them, a possible choice of the dentist is between composites intended for use in direct restorations or in indirect restorations. The direct composite restorations are made in one chair session at relatively low costs and are thus preferred by many clinicians5). Nevertheless, the direct composites are generally not indicated for large restorations because of the limited penetration depth of light during photo-curing (~2 mm)6,7) and due to the residual stress and possible edge gaps resulting from the composite shrinkage8). However, while the indirect composite inlays/onlays appear more resistant to wear thanks to a higher degree of conversion originating from the thermal-assisted polymerization in the oven9), their use is more time consuming and expensive. Additionally, some recent literature is concerned with the possible differences in the performance of these two classes of composites with respect to their plaque accumulation properties10-12), which is known to be critical for the possible occurrence of recurring caries.

The aim of the present study was to assess the viscoelastic response of different composites selected by the dentist (GD) according to his knowledge of the market. The importance of the elastic i.e. storage modulus of restorative composites is well-known, as it should be high enough to withstand masticatory forces especially for posterior restorations, and close to the replaced native tissue13). The clinical implications of

loss tangent, describing the viscous response of the material, are more controversial, as will be pointed out in the DISCUSSION. The study was performed with particular attention on the storage modulus, to analyze its possible difference between the two groups of direct and indirect composites. We also addressed the distribution of modulus within the imaged areas, to understand if major discontinuities exist at the filler-matrix interface. A novel atomic force microscope (AFM) method was used, which is promising for: (i) amount of information provided (i.e. the imaginary part of the generalized complex modulus in addition to the real part13)); (ii) reliable information (i.e. no need to know probe properties to be fed into physical models); (iii) speed of acquisition at given spatial resolution (i.e. it is a real-time scanning technique different from AFM force spectroscopy). The AFM measurements are supported by scanning electron microscope (SEM) imaging and its associated energy-dispersive spectroscopy (EDS) technique. The hypotheses tested were that: (i) the locally different properties at the filler-matrix edges may be observed with the used technique; (ii) there exists a difference in viscoelastic response between direct and indirect composites.

MATERIALS AND METHODS

Selected compositesA total of eight commercially available resin composites were used in the present study. Among them five were direct composites: 1) Venus Diamond from Heraeus Kulzer (abbreviated VD), 2) Adonis from Sweden and Martina, 3) Optifil from IDS Dental, 4) Enamel Plus HRi from Micerium (abbreviated EPH), and 5) Clearfil

Local viscoelastic response of direct and indirect dental restorative composites measured by AFMLaura GRATTAROLA1, Giacomo DERCHI2, Alberto DIASPRO3, Carla GAMBARO1 and Marco SALERNO3

1 Department of Mechanical Engineering, University of Genova, via all’Opera Pia 15, 16145 Genova, Italy2 Istituto Stomatologico Toscano, strada statale 1 via Aurelia, 55041 Lido di Camaiore, Pisa, Italy3 Department of Nanophysics, Istituto Italiano di Tecnologia, via Morego 30, 16163 Genova, ItalyCorresponding author, Marco SALERNO; E-mail: marco.salerno@iit.it

We investigated the viscoelastic response of direct and indirect dental restorative composites by the novel technique of AM-FM atomic force microscopy. We selected four composites for direct restorations (Adonis, Optifil, EPH, CME) and three composites for indirect restorations (Gradia, Estenia, Signum). Scanning electron microscopy with micro-analysis was also used to support the results. The mean storage modulus of all composites was in the range of 10.2–15.2 GPa. EPH was the stiffest (p<0.05 vs. all other composites but Adonis and Estenia), while no significant difference was observed between direct and indirect group (p≥0.05). For the loss tangent, Gradia had the highest value (~0.3), different (p<0.05) from Optifil (~0.01) and EPH (~0.04) despite the large coefficient of variation (24%), and the direct composites showed higher loss tangent (p<0.01) than the indirect composites. All composites exhibited minor contrast at the edge of fillers, showing that these are pre-polymerized, as confirmed by EDS.

Keywords: Dental restorative composites, Direct and indirect restorations, Atomic force microscope, Viscoelastic modulus, SEM/EDS

Color figures can be viewed in the online issue, which is avail-able at J-STAGE.Received Feb 16, 2017: Accepted Jun 14, 2017doi:10.4012/dmj.2017-048 JOI JST.JSTAGE/dmj/2017-048

Dental Materials Journal 2018; 37(3): 365–373

Table 1 Composite materials investigated in the present work (MA=methacrylate)

Composite Manufacturer Class Filler size Filler typeLoading (wt%)

Resin base

(1)Venus Diamond (VD)

Heraeus Kulzer (Hanau, Germany)

Direct (calibration)

Nano/micro-hybrid [12] (5 nm-20 µm) [13]

Ba Al F glass (prepolymerized)

80–82 [12]

TCD-DI-HEA, UDMA [10]

(2) GradiaGC (Tokyo, Japan)

Indirect

Micro-hybrid,(100 nm-17 µm [10], 0.85 µm [13])

SiO2, fumed SiO2, Sr and lanthanoid F, Al F silicate (pre-polymerized) [14]

—MA [10]; UDMA [14]

(3) EsteniaKuraray (Tokyo, Japan)

IndirectNano-hybrid [15](2 nm-2 µm)

treated Al2O392

[16]UDMA [10]; UTMA [15]

(4) SignumHeraeus Kulzer

IndirectMicro-hybrid(~0.6 µm)

SiO274

[15]

Bis-GMA, UDMA, TEGMA

(5) OptifilIDS dental(Savona, Italy)

DirectNano-hybrid(50 nm-1.5 µm)

Ba glass80

[13]

Di- and poli-functional MA

(6)Enamel Plus HRi(EPH)

Micerium(Avegno, Italy)

DirectNano-hybrid(40 nm-1 µm)

nanozirconia, high refractive index glass

81 [17]

Bis-GMA, UDMA, Butandiol-MA

(7)

Clearfil Majesty Esthetic(CME)

Kuraray DirectMicro-hybrid(370 nm-1.5 µm)

Al2O3, glass ceramic (prepolymerized)

92 [18]

4- and 6-functional urethane MA

(8) AdonisSweden and Martina(Padova, Italy)

DirectNano-hybrid(50 nm-0.9 µm)

Ba glass, pyrogenated Si

80Di- and poli-functional MA

Indirect and direct composites are (2)–(8), respectively. VD is (1), even though direct, because is used here as a reference.

Majesty Esthetic from Kuraray (abbreviated CME); and three were indirect composites: 6) Gradia from GC, 7) Estenia from Kuraray, and 8) Signum from Heraeus Kulzer, (for details see Table 1). VD was not measured in this work but just used as a reference.

Specimens preparationThe composite pastes were poured into round pits of a steel mold (~6 mm diameter and ~2 mm height) and then covered with a Mylar strip to minimize the formation of the oxygen-inhibited layer. The direct composites were then polymerized for 40 s through the strip with a chair-side light-curing unit (Valo, Ultradent, South Jordan, UT, USA) at a standard irradiance of ~1,000 mW/cm2. The indirect composites were polymerized in the laboratory oven (Labolight LV-III GC, estimated irradiance of ~1.89 mW/cm2 14)) for 2 min. All the specimens were then removed from the mold, checked for visible surface defects, and polished according to

the manufacturer’s recommendations. Finally, all the specimens were gently cleaned using ethanol (70%) and applicator brush tips (3M ESPE, St. Paul, MN, USA), followed by thorough rinsing in distilled water. For each composite, three specimens were prepared for the measurements.

In order to access the inner bulk material, we ground the surface of each specimen by means of a Dimple Grinder (Gatan, Pleasanton, CA, USA). This device used a rotating copper wheel (diameter ~15 mm and thickness ~2 mm) wet with a polishing slurry containing a ~1:1 mixture of 0.5 and 0.25 µm diameter abrasive diamond particles. After typically 20 min rotation at intermediate speed under a load of 50 g, a depth of 80–100 µm was ground, within a spherical cap with upper diameter of ~2 mm (small regions in Figs. 1a, b among the blue pen marks) and a flat bottom circle with ~0.5 mm diameter (visible in the magnified field of Fig. 1c). A surface cleaning was then carried out as previously described,

366 Dent Mater J 2018; 37(3): 365–373

Fig. 1 a and b: typical aspect of the specimens, when prepared for AM-FM and SEM/EDS measurements, respectively. c: Top-view of the AFM camera during the measurement of a specimen.

The bottom flat circle of the evenly ground area is visible, with the surrounding concentric circles along the slopes towards the upper original level. The five different scan regions within this area are identified by white squares.

at which point the specimens were ready for the AM-FM measurements, after mounting with carbon double-sided sticky pads onto optical glass slides (Fig. 1a).

AM-FM characterizationThe used AFM mode is called AM-FM15) after the simultaneous bimodal cantilever oscillation and feedback in both amplitude modulation (AM) and frequency modulation (FM) regime. AM-FM mode is based on a single-pass, during which the cantilever is dithered at two different resonance frequencies, namely the fundamental one in air, f1,air, same as in standard Tapping mode, and a higher harmonic, in our case the second mode for AC160TS, which is at f2,air~6.2 f1,air

16).The amplitude of oscillation at the lower frequency

is used in an AM feedback loop to track the surface and to obtain the loss tangent tanδ17). The phase of oscillation at the higher resonance frequency enters an FM feedback loop, which adjusts the frequency to keep the phase at 90° (i.e. the oscillation on resonance). The shift in frequency required to this goal, ∆f2,air, depends on the local stiffness of the sample surface.

By AM-FM it is possible to measure the viscoelastic properties of the material, as represented by the generalized complex modulus E=E’+iE”. More precisely, one can measure both the real part of E, namely the storage modulus E’, and the ratio of the imaginary part of E to its real part, known as the loss tangent tanδ=E”/E’. It can be demonstrated that the local storage modulus of the sample at any point on its surface is E’=C ∆f2,air

15,18,19). The technique requires calibration of the coefficient of proportionality C by measurement of a known sample with given elastic modulus E’. To this goal, we used the composite VD (see Table 1) that was formerly investigated extensively in our group, with both

traditional AFM modes (force-volume spectroscopy20)) and nano-indentation21). The value of modulus assumed valid for VD is 11.5 GPa.

The AM-FM imaging of the composites was performed on different regions of each specimen within the ground area (five regions for one specimen of each type, see Fig. 1c, two regions for the other two specimens), for a total of N=9 measurements. Each region had a scan size of 50×50 µm2, large enough for several micro-fillers to possibly appear within the image. The AFM instrument used was an MFP-3D (Asylum Research, Goleta, CA, USA), with an AC160TS probe (Olympus, Tokyo, Japan). The probe has a standard tip (nominal apex diameter ~18 nm, length ~11 µm) at the terminal corner of an arrow-shaped cantilever, with nominal spring constant and resonant frequency of ~40 N/m and ~300 kHz, respectively.

The reference (calibration) sample as close as possible to the unknown sample to be measured is recommended by the company that developed the method (Asylum Research), and for the tooth we used metallic Mg, which we scored at 45 GPa as both due to literature data22) and independent macroscopic measurement in tensile mode on our same specimen (data not shown).

SEM and EDSThe specimens to be used for this analysis were mounted onto aluminium stubs (see Fig. 1b) and underwent overcoating of an electron drain-layer of carbon with nominal ~50 nm thickness, by means of a carbon coater K950X (Emitech, Ashford, UK). The SEM used was a JSM 6490AL (JEOL, Tokyo, Japan). We worked at an acceleration voltage of 15 kV, with an aperture of 2, at a working distance of 10 cm. The specimens were imaged with both secondary electrons and backscattered

367Dent Mater J 2018; 37(3): 365–373

Fig. 2 Representative data on one of the five analyzed regions for the indirect composites. a, d, g: height images, b, e, h: storage modulus images, c, f, i: histograms of levels for

the modulus images. a–c: Gradia, d–f: Estenia, g–i: Signum.

electron signals, seeking for the type of image that provided best contrast among the existing features associated with different physical-chemical phases. We also carried out elemental mapping by means of EDS.

Statistical analysisFrom each AFM image a total of 256×256 pixel is available, which gives rise to its own distribution of values. From this distribution, assumed to be Gaussian, a single representative mean value was extracted, along with its standard deviation representing the distribution spread.

For both the storage modulus E’ and the loss tangent tanδ, the distributions of mean values on the different images have been analyzed for statistical differences by means of one-way analysis of variance (ANOVA), with a post hoc test (Tukey) for multiple pair comparisons. A level of significance of 0.05 was considered to assess the significance of the apparent difference among distributions. Where significant difference occurred, a higher significance level of 0.01 was also probed, to check for so-called strong difference (p<0.01).

RESULTS

In Fig. 2 representative images of the indirect composites are reported. Both the topography (panels a, d, g) and the map of storage modulus (panels b, e, h) are shown. The maps of loss tangent were in all cases featureless and are not presented in the main text. According to a standard convention, in all AFM images the light regions correspond to higher values of the plotted quantity.

From the topography, we see that in Gradia the dominant type of surface morphology consists of grooves, probably due to the grinding. In Estenia instead an isotropic roughness prevails, with fine grain features. Only in Signum large grains appear, mainly rectangular in shape, which could possibly be associated with micro-fillers dispersed in the matrix.

For the E’ values, the surfaces appear to be homogeneous. On the right of the images of modulus, the respective distributions of values (256×256 pixels in the image) are shown. Clearly, each distribution of modulus is mono-modal (i.e. represents a single population) and symmetrical. The distributions have been the fitted with Gaussian profiles, and the values of mean±one standard deviation for each area presented in Fig. 2 are 12.4±1.7, 11.2±2.3 and 10.6±1.8 for Gradia, Estenia and Signum,

368 Dent Mater J 2018; 37(3): 365–373

Fig. 3 Representative data on one of the five analyzed regions for the direct composites. a, d, g, j: height images, b, e, h, k: storage modulus images, c, f, i, l: histograms of levels

for the modulus images. a–c: Optifil, d–f: EPH, g–i: CME, j–l: Adonis.

respectively.In Fig. 3 the same as Fig. 2 but for direct composites

is shown. One can see that for two composites, namely EPH (panel d) and Adonis (panel j), no fillers appear in the topography. Consistently, also in the images of modulus (panels e, k) no internal contrast emerges in the images: again the distributions (panels f, l) are uniform, mono-modal, and roughly Gaussian, and a single E’ value can be extracted from each image. In particular, EPH is stiffest of all the seven composites investigated, showing a modulus of ~15 GPa. However, the other two direct composites, Optifil (panels a–c) and CME (panels g–i) are more interesting. In the topography, trapezoidal areas appear that present locally smoother surface (see the red contours identifying the respective areas in Figs. 3a, g). These areas can be associated with

the presence of micro-fillers. In the images of modulus the micro-fillers show a local mean value slightly different from the outside areas. This difference is hardly visible for Optifil, as shown by the histograms of the different areas in Fig. 3c. The red histogram holds for the total image, while the blue and green histograms describe the two separate masked portions of the image split at the red boundary, representing micro-filler and outer region, respectively. The blue histogram has only slightly lower mean than the green one in this case. For CME the distinction of the histograms (Fig. 3i) is a bit clearer, and one can conclude that the closed (‘micro-filler’) area exhibits higher modulus (~14% higher) than the surrounding area. However, the difference is still low as compared to the distribution width, so is not statistically significant.

369Dent Mater J 2018; 37(3): 365–373

Fig. 4 Representative SEM images (secondary electrons) and EDS maps of selected composites.

a–c: Gradia, d–f: Adonis, g–i: CME, j–l: Optifil. a, d, g, j: morphological SEM images; b, e, h, k: compositional EDS maps of Si, c, f, i, l: compositional EDS maps of the main radio-opaque element (Sr in c, Ba in f, i, l).

In Fig. 4 some representative results of the SEM/EDS analysis are presented. On the rows different composites are listed: one indirect composite is shown in the top row (panels a–c), namely Gradia, and three indirect composites appear in the subsequent rows, namely Adonis, CME and Optifil, from top to bottom. On to the columns, the left-most one reports the SEM morphology and the other two show elemental mapping made by EDS. The middle column, in green, shows the localization of Si within the imaged area, whereas the right column, in red, shows the localization of the radio-active element (Sr in Gradia and Ba in Adonis, CME and Optifil).

For Gradia the fillers are hardly visible (Fig. 4a), and this corresponds to even distributions of both Si (Fig. 4b) and Al (Fig. 4c). This situation is representative also for the other indirect composites and seems in

agreement with the missing mechanical contrast of the AM-FM images in Fig. 2. However, for the representative direct composites shown here (Adonis, CME, and Optifil) fillers with a typical diameter of ~µm (micro-fillers) appear clearly in the morphological SEM images (Figs. 4d, g, j, respectively). Consistently, the respective EDS maps exhibit a locally different composition in the micro-fillers areas. For all the three materials the micro-fillers present an excess of Si (green maps in panels e, h, k), and a low level of Al (red maps in panels f, i, l). Nevertheless, no mechanical contrast between micro-fillers and external regions was observed for Adonis (see Fig. 3k), and very low contrast for CME (Fig. 3h) and Optifil (Fig. 3b).

The results of all the AM-FM measurements on the composites (N=9) are presented in graphical form in Fig. 5, where the dark-gray bars represent the indirect

370 Dent Mater J 2018; 37(3): 365–373

Fig. 5 Compendium of the viscoelastic measurements on all the composites. a, b: storage modulus E’, c, d: loss tangent tanδ. a, c: individual composite values,

b, d: values grouped according to the two classes of direct and indirect composites. Bar heights are the means, error bars are ±1 σ (standard deviation), n=9 for each composite.

Fig. 6 Compendium of the elemental compositions of the composites, after EDS.

Bars heights are the mass percentages of the given element.

composites and the light-gray bars the direct ones. For E’ in Fig. 5a, the white bar represents of the reference (calibration) composite, VD.

For the modulus, the stiffer material (Fig. 5a) clearly appears to be EPH (15.2±2.1 GPa), while all the others are very close together (in the 10.2–11.6 GPa range). Indeed, the ANOVA disclosed a statistically significant difference (at the 95% confidence level) only between EPH and all the other investigated composites but Estenia (11.5±1.0 GPa) and Adonis (11.6±1.9 GPa). In particular, the difference was confirmed at the highest considered level (99% confidence level) for three of four cases (namely Gradia, Signum and CME).

The comparison between the two groups of direct and indirect composites according to the storage modulus (Fig. 5b) showed a slightly higher mean value for the direct ones (12.1±2.6 GPa vs. 10.8±1.9 GPa) yet without statistically significant difference (p>0.05).

For the loss tangent (Fig. 5c), statistically significant differences (p<0.05) were observed here only between Gradia (0.30±0.17) and Optifil (0.08±0.06) and EPH (0.11±0.06). However, despite the generally large spreads within the sets of single composite values, in Fig. 5d a statistically significant difference in loss tangent appeared (p<0.05) between the indirect composites (0.27±0.15) than the direct ones (0.10±0.13).

In Fig. 6 we made a similar bar plot also for the elemental compositions obtained by EDS. The main elements found for all the composites are represented in Fig. 6 by bars of different colors. Only C has not been plotted, which is due not only to the resin but also to the coating used during specimen preparation. One

371Dent Mater J 2018; 37(3): 365–373

can see that the dominant element in all materials is O, and the other element that is always present is Si. Additionally, Al is present in all composites but Signum, and a radiopaque element is always present (also but in Signum) among Sr (in Gradia and EPH), Ba (in Adonis, CME and Optifil) or La (in Estenia).

DISCUSSION

Resin-based dental restorative composites are complex hybrid materials consisting of an organic resin matrix and inorganic filler particles, coated with a bonding agent joining the two phases at their interface. Some discontinuity at this interface is expected in terms of physical-chemical properties. Nevertheless, in most images of modulus for both indirect (Fig. 2) and direct (Fig. 3) composites, no or little contrast was observed between the outer region and the micro-fillers, where these emerged as visible (Figs. 3a, g). The complementary SEM/EDS data provided additional information, useful for interpretation of this lack of mechanical contrast. It should be noted that, while we used is a traditional hot-filament SEM, the resolution was high enough for our purposes (observation of micro-fillers). Apart from O and C (not plotted), Si and Al were the most abundant elements identified in the EDS spectra of Fig. 6, and were thus plotted in the EDS maps of Fig. 4. The large amount of Si appearing in Fig. 4 (‘green’ column) at the micro-fillers is a hint that they are made in all cases of glass (silica) as the main material. Another difference of the micro-fillers with the outer region is a deficiency of Al (data not shown). The same is for the radiopaque element, namely Ba for Adonis, CME, and Optifil, and Sr for Gradia (‘red’ column). Since the latter elements can not be associated with the organic matrix, they have to originate from much smaller fillers (nano-fillers), not resolved in the images yet uniformly present outside the micro-fillers.

These nano-fillers account for the lack of strong contrast in modulus between the micro-fillers and the surrounding regions. Obviously, the stiffer Al associated to alumina nano-fillers compensated the stiffness due to silica inside the micro-fillers. Still, if the micro-fillers were of bare silica, a value of modulus of ~70 GPa would be expected. Hence, the homogeneous modulus across the micro-fillers and the outer nano-filler region may only be due to the micro-fillers being pre-polymerized, thus also containing a certain amount of resin same as in the outer nano-fillers region. Indeed, the EDS maps of C (not shown) where uniform everywhere, and this could not be ascribed solely to the C overcoating (drain layer), due to the deep penetration of the energetic primary electrons (~1 µm). For this reason we should not call the region external to the micro-fillers in Figs. 3 and 4 the ‘matrix’, because it does not contain only comparatively soft organics such as the resin and its additives (photosensitizers, inhibitors, opacifiers, etc.), but also Al and Ba (or Sr) nano-fillers. In turn, the pre-polymerized micro-fillers contain resin as well. The broad use of pre-polymerized fillers is already known

in dental composites and was previously observed for the reference material, VD20). For the composites in our list (Table 1) only two are explicitly declared by the manufacturers to hold this property (Gradia and CME), however it is plausible that all the current composites are similarly designed.

The results in Fig. 5b are in contrast to those of Mesquita et al.23), who observed significantly higher modulus for two direct composites (Diamond Lite and Grandio) versus two indirect ones (Artglass and Vita Zeta LC). However the different conclusion could be due to either the different composites selected or the different technique used (macroscopic dynamic mechanical analysis).

In conclusion, the null hypothesis (i) that the microscale stiffness contrast can be observed with the used AM-FM technique is confirmed, given the minor differences observed for cases in Figs. 3d, j, together with the general explanation of the weak contrast after the use of pre-polymerized micro-fillers together with surrounding nano-fillers.

For the null hypothesis (ii), it is confirmed that a difference exists in viscoelastic properties between the direct and indirect composites selected, but this holds in the loss tangent rather than the storage modulus of the materials. The measured loss tangent represents the dissipation in the system due to the residual sample viscosity, and appeared higher for the indirect composites, which is a somewhat unexpected result. It is generally assumed that the photo-curing of direct composites is less effective than the combined thermal and light curing carried out in the oven for the indirect composites, and as a consequence conversion is ongoing for the direct composites during the first days following the photo-cure24). The clinical implications of loss tangent values are not clear to date, and mainly the subject of speculation only. Whereas some authors consider high loss tangent as an indicator of negative performance of the composites25), others consider that, for similar values of storage modulus, composites with higher loss tangent provide alternative means of relief of applied stress, thus probably enhancing the material strength26). We do not pretend here to make any conclusion and simply provide the experimental observations for future interpretation. The clinical implications of the present results have to be investigated in future work, possibly in conditions closer to practice.

Some limitations apply to the observed result. The loss tangent values of the individual composites (Fig. 5c) are affected by much higher relative errors than for the modulus, particularly for Gradia (almost 50%), Signum (more than 50%), and especially CME (more than 100%). It should be pointed out that the viscous dissipation at the tip-sample interface may be affected not only by the sample properties but also by the environmental conditions (humidity of ambient air) as well as by surface contaminants. As a consequence, it is known that the loss tangent values measured by AM-FM can be considerably overestimated, of a factor up to one order of magnitude17,27). Nevertheless, the relative ranking of the

372 Dent Mater J 2018; 37(3): 365–373

materials in the same conditions (i.e. same systematic error) should not be affected.

The EDS spectra and the elemental color maps derived thereof (e.g. Fig. 4) were not only useful for qualitative identification of the chemical elements present, but could be processed for quantitative evaluation of the relative amounts, as shown in the weight percentages of Fig. 6. One interesting observation is about the presence of O, which mostly comes from the filler oxides. In agreement with this, the stiffest material is that one with highest O content, namely EPH, which confirms the reinforcing effect of the inorganic fillers.

CONCLUSION

In this study different direct and indirect resin composites have been tested, by means of a novel AFM technique called AM-FM, measuring the complex elastic modulus of materials. The weak contrast of elastic modulus at the surface of micro-fillers was ascribed to the use of prepolymerized fillers, along with interstitial nanoscale fillers. For the chemical composition we identified both mechanical reinforcement and radio-opacity agents in both classes of composites, without significant differences. The storage modulus was equivalent for both classes of composites but a difference appeared in their viscous response, showing higher loss tangent for the indirect composites.

REFERENCES

1) Khoroushi M, Keshani F. A review of glass-ionomers: From conventional glass-ionomer to bioactive glass-ionomer. Dent Res J 2013; 10: 411-420.

2) Lohbauer U. Dental glass ionomer cements as permanent filling materials? Properties, limitations and future trends. Materials 2010; 3: 76-96.

3) Chan KHS, Mai Y, Kim H, Tong KCT, Ng D, Hsiao JCM. Review: Resin composite filling. Materials 2010; 3: 1228-1243.

4) Cramer NB, Stansbury JW, Bowman CN. Recent advances and developments in composite dental restorative materials. J Dent Res 2011; 90: 402-416.

5) Tyas MJ, Anusavice KJ, Frencken JE, Mount GJ. Minimal intervention dentistry —a review. Int Dent J 2000; 50: 1-12.

6) Moore BK, Platt JA, Borges G, Chu TM, Katsilieri I. Depth of cure of dental resin composites : ISO 4049 depth and microhardness of types of materials and shades. Oper Dent 2008; 33: 408-412.

7) Monte Alto RV, Guimaraes JGA, Poskus LT, da Silva EM. Depth of cure of dental composites submitted to different light-curing modes. J Appl Oral Sci 2006; 14: 71-76.

8) Dietschi D, Scampa U, Campanile G, Holz J. Marginal adaptation and seal of direct and indirect Class II composite resin restorations: an in vitro evaluation. Quintessence Int 1995; 26: 127-138.

9) Tezvergil-Mutluay A, Lassila LVJ, Vallittu PK. Degree of conversion of dual-cure luting resins light-polymerized through various materials. Acta Odontol Scand 2007; 65: 201-205.

10) Bernardo M, Luis H, Martin MD, Leroux BG, Rue T, Leitão J, DeRouen TA. Survival and reasons for failure of amalgam versus composite posterior restorations placed in a randomized clinical trial. J Am Dent Assoc 2007; 138: 775-783.

11) De Fúcio SBP, Puppin-Rontani RM, De Carvalho FG, De Mattos-Graner RO, Correr-Sobrinho L, Garcia-Godoy F. Analyses of biofilms accumulated on dental restorative materials. Am J Dent 2009; 22: 131-136.

12) Derchi G, Vano M, Barone A, Covani U, Diaspro A, Salerno M. Bacterial adhesion on direct and indirect dental restorative composites: an in vitro study on a real biofilm. J Prosthet Dent 2017; 117: 669-676.

13) Darvell BW. Materials Science for Dentistry. Ninth Edit. Woodhead Publishing; 2009.

14) Hirata M, Koizumi H, Tanoue N, Ogino T, Murakami M. Influence of laboratory light sources on the wear characteristics of indirect composites. Dent Mater J 2011; 30: 127-135.

15) Ebeling D, Solares SD. Bimodal atomic force microscopy driving the higher eigenmode in frequency-modulation mode: Implementation, advantages, disadvantages and comparison to the open-loop case. Beilstein J Nanotechnol 2013; 4: 198-207.

16) Butt HJ, Jaschke M. Calculation of thermal noise in atomic force microscopy. Nanotechnology 1995; 6: 1-7.

17) Proksch R, Kocun M, Hurley D, Viani M, Labuda A, Meinhold W, Bemis J. Practical loss tangent imaging with amplitude-modulated atomic force microscopy. J Appl Phys 2016; 134901: 1-11.

18) Grattarola L. Misura con metodo AMFM del modulo di Young di compositi dentali. MSc Thesis. University of Genova; 2016. Available from: www.tesionline.it/default/tesi.asp?idt=50776

19) Garcia R, Proksch R. Nanomechanical mapping of soft matter by bimodal force microscopy. Eur Polym J 2013; 49: 1897-906.

20) Salerno M, Patra N, Diaspro A. Atomic force microscopy nanoindentation of a dental restorative midifill composite. Dent Mater 2012; 28: 197-203.

21) Salerno M, Derchi G, Thorat S, Ceseracciu L, Ruffilli R, Barone AC. Surface morphology and mechanical properties of new-generation flowable resin composites for dental restoration. Dent Mater 2011; 27: 1221-1228.

22) The Engineering ToolBox. Modulus of Elasticity or Young’s Modulus - and Tensile Modulus for common Materials. [cited 2016 May 31]. Available from: http://www.engineeringtoolbox.com/young-modulus-d_417.html

23) Mesquita R V, Axmann D, Geis-Gerstorfer J. Dynamic visco-elastic properties of dental composite resins. Dent Mater 2006; 22: 258-267.

24) Steinhaus J, Frentzen M, Rosentritt M, Möginger B. Dielectric analysis of short-term and long-term curing of novel photo-curing dental filling materials. Macromol Symp 2010; 296: 622-625.

25) Papadogiannis Y, Helvatjoglu-Antoniades M, Lakes RS. Dynamic mechanical analysis of viscoelastic functions in packable composite resins measured by torsional resonance. J Biomed Mater Res - Part B Appl Biomater 2004; 71: 327-335.

26) Tanimoto Y, Nishiwaki T, Nemoto K. Dynamic viscoelastic behavior of dental composites measured by Split Hopkinson pressure bar. Dent Mater J 2006; 25: 234-240.

27) Proksch R, Yablon DG. Loss tangent imaging: Theory and simulations of repulsive-mode tapping atomic force microscopy. Appl Phys Lett 2012; 100: 2012-2014.

373Dent Mater J 2018; 37(3): 365–373