TYPE Original Research PUBLISHED 20 October 2022 DOI 10.3389/fevo.2022.975283 Accuracy of dental microwear OPEN ACCESS impressions by physical EDITED BY Ferran Estebaranz-Sánchez, Milá y Fontanals Institution properties of silicone materials for Research in Humanities (CSIC), Spain Ryohei Sawaura1,2,3*† , Yuri Kimura4,5*† and Mugino O. Kubo6*† REVIEWED BY 1 Maria Ana Correia, Okinawa Prefectural Museum and Art Museum, Naha, Japan, 2 Department of Anthropology, University of São Paulo, Brazil National Museum of Nature and Science, Tsukuba, Japan, 3 Division of Oral and Craniofacial Lisa Schunk, Anatomy, Graduate School of Dentistry, Tohoku University, Sendai, Japan, 4 Department of Geology University of Wrocław, Poland and Paleontology, National Museum of Nature and Science, Tsukuba, Japan, 5 Institut Català de Paleontologia Miquel Crusafont, Barcelona, Spain, 6 Department of Natural Environmental Studies, *CORRESPONDENCE Graduate School of Frontier Sciences, The University of Tokyo, Kashiwa, Japan Ryohei Sawaura sawaura@gmail.com Yuri Kimura ykimura.research@gmail.com Mugino O. Kubo Dental microwear analysis is an oft-used paleodietary estimation method, and mugino@k.u-tokyo.ac.jp † These the impression molds or resin casts are often analyzed rather than the original authors have contributed equally to this work tooth surfaces. A choice of silicone products for dental impressions is crucial SPECIALTY SECTION because the quality of microwear data is affected by the impression accuracy This article was submitted to of the molds. For this reason, microwear researchers have heavily depended Paleoecology, on a few commercial products such as “President” (Coltène/Whaledent AG, a section of the journal Frontiers in Ecology and Evolution Switzerland) to avoid analytical errors caused using different silicone materials. RECEIVED 22June 2022 Considering that the production business might be terminated, however, ACCEPTED 16 September 2022 heavy reliance on specific products could be a potential weakness in the PUBLISHED 20 October 2022 field. In this study, we aimed at identifying specific indexes of physical CITATION properties of silicone materials with satisfactory accuracy. For this purpose, Sawaura R, Kimura Y and Kubo MO (2022) Accuracy of dental microwear we measured dynamic viscoelasticity and shrinkage rates of various silicone impressions by physical properties compounds, including the standard impression material President and other of silicone materials. Front. Ecol. Evol. 10:975283. eight affordable dental silicones. We scanned both original tooth surface and doi: 10.3389/fevo.2022.975283 dental impression molds with a confocal laser microscope and conducted COPYRIGHT dental microwear texture analysis (DMTA) to quantitatively compare the © 2022 Sawaura, Kimura and Kubo. This is an open-access article scanned surfaces. The results showed relationships between the material distributed under the terms of the properties of silicones and impression accuracy, indicating that the materials Creative Commons Attribution License that cured slowly and began to shrink relatively early in the hardening process (CC BY). The use, distribution or reproduction in other forums is were less accurate. Some of these dental impression molds showed blurred permitted, provided the original surfaces, implying that molds were peeled off from the tooth surface at the author(s) and the copyright owner(s) are credited and that the original microscopic level, as the shrinkage speed might exceed the curing speed. The publication in this journal is cited, in following indices provided in the product information were found to be helpful accordance with accepted academic practice. No use, distribution or in the search for substitutes: (1) medium viscosity, (2) short curing time after reproduction is permitted which does mixing (5–6 min), and (3) delayed change in shrinkage. not comply with these terms. KEYWORDS dental microwear texture (DMT), accuracy of silicone molds, material properties, dynamic viscoelasticity, shrinkage rate Frontiers in Ecology and Evolution 01 frontiersin.org Sawaura et al. 10.3389/fevo.2022.975283 Introduction In this study, we aim at identifying the physical properties of silicones to find appropriate silicones, which mold the Microscopic use-wear marks on tooth surfaces (microwear) tooth surface with satisfactory accuracy, by examining the are important as a morphological indicator for estimating reproducibility of DMT among various silicone compounds, the diets of extinct and extant animals and have long been including President and other affordable dental silicones. investigated using scanning electron microscopes or light In addition to comparing the general properties of the microscopes (reviewed in Ungar, 2015). In the early stages of tested silicones provided by product companies, we dental microwear research, optical microscopy and SEM were measured changes in viscosity and shrinkage over time, the main methods of observation. Even when the same area was which are important properties in the process of making analyzed, the observation settings and inter- or intra-observer impression molds. errors were often problematic (Grine et al., 2002; Galbany et al., 2005; Purnell et al., 2006). So, dental microwear was difficult to evaluate quantitatively. However, the recent progress Materials in the evaluation of microwear by optical profilometers has generated a new field of microwear research (Ungar et al., 2003; Tooth specimen Scott et al., 2005; Schulz et al., 2010). Currently, an increasing number of studies reported characteristics of microscopic 3D We used an isolated right mandibular second molar of a topography, that is dental microwear texture (DMT), of various Pleistocene Japanese macaque (Macaca fuscata) (lower right types of vertebrates (Krueger et al., 2008, 2017; Caporale and of Figure 1). The analyzed specimen (NMNS PV 6166-7) Ungar, 2016; DeSantis and Patterson, 2017; Aiba et al., 2019; was originally reported by Hasegawa et al. (1968) and is now Kubo and Fujita, 2021). stored in the Department of Geology and Paleontology at the In the studies of 3D dental microwear texture analysis National Museum of Nature and Science, Japan. Following (DMTA) on anthropological and paleontological materials, the standard of primate dental microwear studies (Kay and impression molds and resin casts are analyzed rather than Hiiemae, 1974), facet 9 of Phase II was observed for the analysis. the original tooth surfaces. Two major reasons are considered. Three arbitrary sites on facet 9 (f9-1, f9-2, and f9-3) were First, large specimens, such as teeth in the jawbone, cannot be selected based on the accessibility to the focal points, which placed properly on a microscope stage. Second, cultural and had characteristic landmarks (e.g., pits, scratches, and cracks) paleontological resources may be protected under laws and (Figure 1). regulations, and thus casts and molds are the best alternatives for further analyses in laboratories. Therefore, the accuracy of dental silicones is crucial in DMTA. In previous studies, Silicone impression materials microwear researchers have utilized a few commercial products of silicone compounds such as President (Coltène/Whaledent We used the following nine dental silicone compounds AG, Switzerland) to avoid analytical errors caused by using including the standard silicone impression material (President different materials. However, such an effort of analytical quality Regular Body) in microwear research (Table 1): President control (i.e., relying on certain products) is a fundamental Regular Body, Affinis regular body, Affinis light body weakness in the field, considering that the production business (Coltène/Whaledent AG, Switzerland), Dr. Silicon regular might be terminated for miscellaneous reasons as in the (BSA Sakurai, Japan), Fusion II wash, EXAHIFLEX regular case of President. (GC, Japan), JM Silicone (Nissin Dental Products INC., Japan), A few studies have tested the President silicone in the IMPRINSIS regular (Tokuyama Dental Corporation, Japan), method of DMTA. Goodall et al. (2015) tested some substitutes and SILDE FIT regular (Shofu INC., Japan). Dr. Silicon is a local for President and reported that silicone compounds with product mainly in East Asia, whereas others are commercially medium viscosity were more accurate than low viscosity available worldwide. However, President Regular Body was no (Speedex, MM913, and Accutrans) and high viscosity (Microset longer available in Japan and was replaced by the subsequent 101 RF and MM240TV). They detected a few differences product line called Affinis. Therefore, we imported it from the of statistical significance between the original specimen and United Kingdom in 2017. Around that time, the “President” casts made with President. On the other hand, Mihlbachler series became a new product line, “PRESIDENT The Original.” et al. (2019) used President to examine differences between Since this material was not available (and still is not in Japan) at the original specimen and casts in higher magnification the time we initiated this study, we did not include it but used scans (150×) and reported many significant differences. Their the older “President.” contradictory conclusions indicate that different choices of All products are cartridge-type in dentistry quality, and molding and casting materials, the surface textures analyzed, their material properties such as shrinkage rate and cure and scanning magnifications may have affected the results. time are provided by the companies. High-viscosity materials Frontiers in Ecology and Evolution 02 frontiersin.org Sawaura et al. 10.3389/fevo.2022.975283 FIGURE 1 Comparison of dental impressions among nine impression materials. 2D images were obtained by a confocal laser microscope equipped with a 20× objective lens. The size of each image is 540 µm × 725 µm. An occlusal view of the right lower second molar of the Japanese macaque used in this study is shown in the right lower photo and schema. The black rectangle on the tooth photo indicates the area where we obtained silicone impression molds. The white rectangles in the 20× tooth surface image obtained from the original tooth indicate the locations (f9-1, 2, and 3) where higher magnification images and 3D texture data were acquired. Wear state and location of facet 9 are shown in the schema. were excluded in this study because Goodall et al. (2015) made sure to apply the silicones well before their expiration reported poor accuracy with high-viscosity materials. For date to avoid the impact of quality deterioration on the mixing, we used a cartridge dispenser and mixing tips. We results. Frontiers in Ecology and Evolution 03 frontiersin.org Sawaura et al. 10.3389/fevo.2022.975283 Viscoelasticity Methods JIS T6513:2005 is a Japanese Industrial Standard based on the international standard ISO 4823:2000 Dentistry—Elastomeric impression materials and JIS and ISO have the same standard for pot life and final shrinkage (at the time this test was conducted). at 500 s (log kPa/s)e 47.4 40.6 35.8 37.9 12.7 19.9 13.4 22.5 7.4 Measurement of physical properties of silicones TABLE 1 Silicone materials tested in this study and their physical properties shown in their product material safety data sheet (MSDS) and obtained through the measurements in this study. To monitor the time-dependent elastic behaviors of the Viscosity in product silicone materials after mixing, we measured the dynamic Medium Medium Medium Medium MSDS Low Low Low Low Low viscoelasticity and shrinkage rate of the silicone materials. Dynamic viscoelasticity was measured by a rheometer (HAAKE Mars III, Thermofisher Scientific) at the Tokyo Metropolitan Industrial Technology Research Institute. For studyd every measurement, about 3 ml of silicone material was tested Final shrinkage (%) This 0.65 0.35 0.63 0.54 0.38 0.67 0.74 0.82 0 and monitored for 10 min. Nagrath et al. (2014) reported that humidity at curing affects impression accuracy. In this study, dynamic viscoelasticity measurements were performed under average room temperature and humidity conditions. Product MSDSa 0–1.5* 0–1.5* 50.2* 0.1 0.1 0.1 0.2 0.1 0.2 Time-sequential change of shrinkage was monitored under a room temperature of 25◦ C and ambient humidity by a Custron device (AcroEdge, Tokyo, Japan) for 10 min, starting 90 s after silicone materials were drawn on a discal pit (diameter = 10 mm Hardnessc a Measured under JIS T6513:2005. b Defined as duration to reach constant viscosity in Figure 2A. c End hardness in Shore A. d See Figure 2B. e See Figure 2A. and depth = 1 mm) in a glass plate. 40.5 NA NA NA NA NA NA 55 57 Evaluation of impression accuracy studyb Preparation for microscope scanning This >600 >600 >600 >600 >600 325 430 415 500 Cure time (s) Each silicone was applied to the targeted area of the tooth surface and then pressed against its surface with a small ball of silicone putty (EXAFINE PUTTY TYPE, GC) to assure that details are imprinted on the silicone mold. This procedure Product MSDSa 180* 120* 180* was done as an imitation of the common one when taking 120 190 180 180 180 240 dental impressions in clinical situations: (1) a dentist applies an impression material using a cartridge dispenser on the teeth of a patient, and (2) the patient bites a clinic tray Pot life filled with putty-type silicon. The mold was taken off from 120* 135* 135* (s)a 120 120 120 138 120 120 the specimen in 5–6.5 min. Before applying the next silicone material, we cleaned the specimen surface with 2% NaOCl- Asterisks (*) indicate that JIS is not mentioned in the product information. soaked cotton sheets followed by rubbing with acetone-soaked Regular body Regular body industrial cotton swabs. After taking off from the specimen, Light body Regular & denture Regular Regular Regular Regular Type Wash we trimmed the silicon mold to an appropriate size using a knife and attached it to a sliding glass using superglue for laser Product microscope observation. EXAHIFLEX IMPLINSIS JM Silicone SILDE FIT Dr. Silicon Confocal laser microscope scanning President Fusion II Name Affinis The original specimen and silicone molds were scanned with a confocal laser microscope (VK-9700, Keyence, Osaka, Japan) housed in the Department of Natural Environmental Studies, Graduate School of Frontier Sciences, the University Coltène/Whaledent AG of Tokyo. We used lower magnification lenses (20× and 50×) to identify the scan locations (i.e., f9-1, f9-2, and f9-3). Then Company BSA sakurai we used a 100× long-distance lens (N.A. = 0.95) to scan the Tokuyama focal sites and generated 3D surface texture data by lateral Nissin Shofu GC (x, y) sampling with an interval of 0.138 µm and a vertical Frontiers in Ecology and Evolution 04 frontiersin.org Sawaura et al. 10.3389/fevo.2022.975283 resolution (z) of 0.001 µm. For each scan, 3D surface data of TABLE 2 The number of repeated scans for each sample and scan location. an area of 140 µm × 105 µm was obtained. Moreover, grayscale photographs of the tooth surfaces were procured for graphical Sample Facet Facet Facet Total comparison. We scanned both the original tooth surface and 9-1 9-2 9-3 the silicone molds with dental impressions by the following Original tooth 29 27 27 83 procedures. We first placed the original specimen on the stage 1. Affinis light body 12 12 12 36 of the laser microscope and scanned the focal sites. To replicate 2. Affinis regular body 3 3 3 9 the process, we took the specimen out of the stage and then 3. Dr. Silicon regular type 4 4 4 12 replaced it on the stage to scan the same targeted area again. 4. Fusion II wash type 7 7 7 21 The relocation was done manually by finding common features 5. EXAHIFLEX regular type 3 3 3 9 (see Supplementary Figures 1–3), so this may have influenced 6. JM Silicone regular type 3 3 3 9 the results to some extent. The scanning was repeated at least 7. President jet regular body 4 4 4 12 three times for each site. We scanned the silicone molds in the 8. IMPRINSIS regular type 3 3 3 9 same way, but for every scanning, we made new impression 9. SILDE FIT regular & 7 7 7 21 molds to take the intrasample errors of impression molds denture type into account. The original tooth was scanned repeatedly on Total 75 73 73 221 different days to check intrapersonal variation in the choice of the targeted area. The accessibility to the focal scan points varied among the samples because it was difficult to find the silicone molds, Steel’s multiple-comparison test was performed same scan points in the molding samples with less precise with the original surface scan as a control treatment. The surface replication. Therefore, the number of repeated scans number of significant differences was counted for each of the differed among the molding materials (Table 2). Note that the nine impression materials at each scan location (f9-1, f9-2, and scanning of molds was done within 1–2 weeks after molding, f9-3). To visualize the variation of surface roughness among so there is little need to consider the effects of dimensional original tooth surfaces and the molds, we conducted principal changes as reported in Rodriguez and Bartlett (2011) on the component analysis (PCA) with Varimax rotation for the 32 acquired data. DMT parameters (30 ISO and 2 SSFA parameters), which are presented as scatter plots along the principal component (PC) Qualitative evaluation of 2D image axes. All statistical analyses were conducted using JMP Pro For a simple evaluation of the accuracy and precision of 16.0.0 (SAS Institute Inc.). the silicone impression molds, we first checked air bubble contamination and surface smoothing on 2D images obtained by the 100 × lens. Results Quantitative evaluation of 3D surface Dynamic viscoelasticity microtopography We analyzed the 3D data obtained by the confocal Figure 2A shows the dynamic viscoelasticity (log kPa/s) laser microscope using Mountains Map Imaging Topography of silicone materials measured by a rheometer. The analyzed (version 9.0.9653, Digital Surf, Besançon, France). The 3D data silicone materials are categorized into two types based on were preprocessed to remove measurement noises and outliers viscoelastic behaviors during hardening. by the following procedure described in Aiba et al. (2019), which included (1) mirroring (only for molds), (2) leveling Rapid completion type (i.e., steep changes in (least square plane by subtraction), (3) application of S-filter viscoelasticity) (a robust Gaussian filter with a cutoff value of 0.8 µm, as Four products of silicone compounds are included, which defined in ISO 25178), (4) form removal function (polynomial are President, Affinis (regular), Dr. Silicon, and Affinis (light) of increasing power = 2) to remove large-scale curvature of in this order. Of these, Dr. Silicon, Affinis (regular), and enamel bands (F-operation in ISO 25178), (5) outlier removal Affinis (light) have low initial viscoelasticity. The hardening with a slope >80◦ , (6) application of a threshold removing of President was completed 6 min after mixing of main the upper and lower 0.1% of the data, and (7) filling of the and cure agents. The final viscosity of President is the non-measured points. After data cleaning, 30 parameters of highest of all. The hardening of Affinis (regular), Dr. ISO 25178-2 and two (fractal complexity Asfc and length-scale Silicon, and Affinis (light) has completed about 7 min anisotropy epLsar) from scale-sensitive fractal analysis (SSFA) after mixing. Noticeably, these four products with rapid were calculated (Table 3). completion of hardening have high impression accuracy by For non-parametric pair-wise tests between the parameter qualitative evaluation of two-dimensional microscopic images values from the original tooth surface and those from the (Supplementary Figures 1–3). Frontiers in Ecology and Evolution 05 frontiersin.org Sawaura et al. 10.3389/fevo.2022.975283 TABLE 3 Names and descriptions of dental microwear texture parameters used in this study. Parameter Description Unit Standard Sq Standard deviation of the height distribution µm ISO 25178-2 Ssk Skewness of the height distribution No unit ISO 25178-2 Sku Kurtosis of the height distribution No unit ISO 25178-2 Sp Maximum peak height, height between the highest peak and the mean plane µm ISO 25178-2 Sv Maximum pit height, depth between the mean plane and the deepest pit µm ISO 25178-2 Sz Maximum height, sum of the maximum peak height and the maximum pit µm ISO 25178-2 height (Sp + Sv) Sa Arithmetic mean height µm ISO 25178-2 Smr Areal material ratio, ratio of the area of the material at a specified height c % ISO 25178-2 (c = 1µm under the highest peak) Smc Inverse areal material ratio, height at which a given areal material ratio (p = 10%) µm ISO 25178-2 Sdc(Previously Sxp) Peak extreme height, difference in height between the p and q material ratio µm ISO 25178-2 (p = 50%, q = 97.5%) Sal Autocorrelation length (s = 0.2) µm ISO 25178-2 Sdq Root mean square gradient No unit ISO 25178-2 Sdr Developed interfacial area ratio % ISO 25178-2 Vm Material volume at a given material ratio (p = 10%) µm3 /µm2 ISO 25178-2 Vv Void volume at a given material ratio (p = 10%) µm3 /µm2 ISO 25178-2 Vmc Material volume of the core at a given material ratio (p = 10%, q = 80%) µm3 /µm2 ISO 25178-2 Vvc Void volume of the core (p = 10%, q = 80%) µm3 /µm2 ISO 25178-2 Vvv Void volume of the dale at a given material ratio (q = 80%) µm3 /µm2 ISO 25178-2 Spd Density of peaks 1/µm2 ISO 25178-2 Spc Arithmetic mean peak curvature 1/µm ISO 25178-2 S10z Ten-point height µm ISO 25178-2 S5p Five-point peak height µm ISO 25178-2 S5v Five-point pit height µm ISO 25178-2 Sda Closed dale area µm2 ISO 25178-2 Sha Closed hill area µm2 ISO 25178-2 Sdv Closed dale volume µm3 ISO 25178-2 Shv Closed hill volume µm3 ISO 25178-2 Sk Distance between the highest and lowest level of the core surface µm ISO 25178-2 Spk Average height of the protruding peaks above the core surface µm ISO 25178-2 Svk Average height of the protruding dales below the core surface µm ISO 25178-2 epLsar Exact proportion of length scale anisotropy of relief No unit Scale sensitive fractal analysis (Ungar et al., 2003) Asfc Area-scale fractal complexity No unit Scale sensitive fractal analysis (Ungar et al., 2003) Slow completion type (i.e., gradual changes in they are largely categorized into two types with a threshold of viscoelasticity) 0.2% in 200 seconds (s) as follows. The remaining five products are categorized in this type: SILDE FIT, JM Silicone, IMPRINSIS, EXAHIFLEX, and Fusion Delayed change in shrinkage II had higher viscosity in this order. The hardening of these The shrinkage rates exceed 0.2% after 200 s, showing more products did not complete perfectly even 10 min after mixing. gradual curves of shrinkage compared to the fast shrinkage type. This includes President (two-step slope with an initial shrinkage starting in 180 s after mixing and reaching 0.74% in 480 s), Shrinkage rate Affinis regular (one-step slope with an initial shrinkage starting in 200 s after mixing and reaching 0.35% in 330 s), and Dr. Figure 2B shows the shrinkage rate changes of silicone Silicon (one-step gradual slope, with an initial shrinkage starting materials during cure time. More complicated patterns are in 280 s and reaching at 0.58% after 650 s). Final shrinkage observed for shrinkage than for dynamic viscosity. However, rates fall between 0.35 and 0.65%. Irregularly, no shrinkage was Frontiers in Ecology and Evolution 06 frontiersin.org Sawaura et al. 10.3389/fevo.2022.975283 FIGURE 2 Dynamic viscoelastic properties and shrinkage rates of silicone materials. (A) Chronological change of viscoelasticity after mixing. (B) Chronological change of shrinkage rates after mixing. The dashed and gray areas in the figure show the timing of silicon application, pressing, and removal. observed for Affinis light, which could be an unknown analytical lower magnification images (Figure 1), it was obvious in error since its final shrinkage rate provided by the company was the higher magnification images that several impression comparable to that of Affinis regular (Table 1). Alternatively, molds (EXAHIFLEX, JM Silicone, and IMPRINSIS) had poor this result on Affinis light could be batch-dependent, but due reproducibility of original DMT, that is, their microscopic to the budget and time available for this study, we were not able images showed blurred surfaces and contamination of bubbles to repeat the measurements. (Supplementary Figures 1–3). In all fields of f9-1, f9-2, and f9-3, blurring was observed in EXAHIFLEX, JM Silicone, and Fast change in shrinkage IMPRINSIS. Molds made by SILDEFIT roughly reproduced The shrinkage rates exceed 0.2% within 200 s, showing overall microwear features but had many traces of small bubbles increased shrinkage rates immediately after mixing. This group contamination on the surfaces in all target views. Among other includes JM Silicone (1st slope, reaching a shrinkage rate of impression materials with relatively good accuracy, President 0.5% in 200 s, followed by the 2nd gradual slope), SILDE and Dr. Silicon had the best accuracy with fine microwear FIT (1st slope, reaching 0.6% in 200 s, followed by the 2nd features (e.g., very fine scratches). gradual slope), Fusion II (one-step steep slope, reaching 0.55% in 250 s), IMPRINSIS (one-step steep slope, reaching 0.74% in 280 s), and EXAHIFLEX (1st slope, staying at 0.2% until Impression accuracy evaluated by ∼240 s and reaching at 0.38% after 420 s, followed by the 2nd dental microwear texture parameters gradual slope between 300 and 400 s). Final shrinkage rates range from 0.38 to 0.82%. The difference between the original surface and molds was further evaluated by statistical comparisons of 32 DMT parameters. The raw values of the DMT parameters of Impression accuracy judged by 2D all scans (N = 221) used in this study are listed in images Supplementary Table 1. Summary statistics (sample size, mean, and S.D.) of the DMT parameters for scan targets (i.e., the Though it was confirmed that there were differences original tooth and nine impression materials) are provided in in impression accuracy of silicone materials even in the Supplementary Table 2. Results of Steel’s non-parametric test Frontiers in Ecology and Evolution 07 frontiersin.org Sawaura et al. 10.3389/fevo.2022.975283 comparing the original tooth surface and dental molds are of ISO 25178-2, indicating that PC1 could be interpreted as presented in Supplementary Table 3. overall surface roughness, in other words, the size of microwear Figure 3 shows the number of significant differences in features. On the other hand, PC2 had larger positive values for Steel’s non-parametric test presented for each silicone material Sdq, Sdr, Asfc, and Spd, which were all related to the fineness and scan location. Thus, lower values on the vertical axis are of the surface. In addition, the parameters related to surface better, capturing morphological details of the original surface. segmentation (Sda, Sha, Svd, and Shv) showed large negative President Jet Regular body and Dr. Silicon regular type are values. Therefore, the larger PC2 values indicated that the better (= the smallest number of significant differences) than surface was characterized by finer features and segmented into other products. The accuracy of Affinis light and SILDE FIT smaller hills and dales. The scatter plots of PCs are shown follows President and Dr. Silicon, and the difference is small. in Figure 4. We found a non-negligible variation in repeated To present the precision of dental molds, the deviation of the scans of the original tooth surface, which indicated uncontrolled parameter value of the mold from the average parameter value errors associated with placing the tooth on the microscope stage, of the original tooth surface was obtained for each parameter tilting the tooth, and searching and scanning the focal sites: all (Supplementary Figure 4). In Supplementary Figure 4, the of them cannot be perfectly repeated and therefore this leads to ranges of deviations (mold—original tooth surface) are wider data variation. Most of the scan data from the dental molds lay for some impression materials (IMPRINSIS and JM Silicone), within the 90% confidence ellipses of the original tooth scans, whereas others show equivalent ranges. Affinis light shows though some departed even from the ellipses. Notably, the data narrower ranges since it has a larger number of molding and of molds made by IMPRINSIS regular type, the material with scanning. In most cases, the ranges include zero, suggesting the least replicability (Supplementary Figures 1–3), are placed that there was no outstanding impression material and that beyond the ellipses for all three scan locations. all show decent levels of precision. However, as seen in the 100 × 2D images of Supplementary Figures 1–3, SILDE FIT molds contain many bubbles. And the Affinis light molds are slightly blurred. Therefore, we evaluated them not as accurate as Discussion President and Dr. Silicon. IMPRINSIS regular type is the worst among all (Figure 3), which corresponds to visual observations Dynamic viscoelasticity as a good of 2D images (Supplementary Figures 1–3). indicator to find suitable silicones for PCA of the 32 DMT parameters clarified that the first and dental microwear texture analysis the second principal components explained 52.6 and 19.1% of the observed variation. The factor loadings of PC1 (Table 4) Based on 2D observations and statistical results of the showed larger positive values of height and volume parameters DMT parameters, we clearly showed dynamic viscoelasticity FIGURE 3 The number of significant differences in 32 DMT parameters between the original tooth surface and silicone molds. Pair-wise tests were conducted for each scan location on facet 9 (f9-1, f9-2, and f9-3). Frontiers in Ecology and Evolution 08 frontiersin.org Sawaura et al. 10.3389/fevo.2022.975283 TABLE 4 The factor loadings obtained from a PCA using the silicones provided by the product companies (Table 1), the 32 DMT parameters. range of cure time of the slow completion type does not differ PC1 PC2 from that of the rapid completion type. This finding indicates that the most conclusive and quantitative way to find suitable Variance explained (%) 52.6 19.1 silicones for DMTA is to measure dynamic viscoelasticity by a Sq 0.9860 −0.0872 rheometer and identify the product with similar curves to those Ssk −0.3811 0.4047 of the rapid completion type. Sku 0.1615 −0.2427 Sp 0.8890 0.1019 Sv 0.9415 −0.1742 Criteria for silicones with the high Sz 0.9840 −0.0776 impression accuracy Sa 0.9813 −0.0780 Smr −0.6484 −0.1563 Although dynamic viscoelasticity is the best indicator, Smc 0.9729 −0.0390 we suggest the following criteria for silicones as reliable Sdc 0.9684 −0.1372 as President. Sal 0.5220 −0.3105 Sdq 0.5495 0.7938 Rapid completion of curing in 5–6 min after Sdr 0.4881 0.7770 mixing Vm 0.8096 0.0160 The silicones with rapid completion showed better Vv 0.9764 −0.0371 impression accuracy. It may be related to the timing of Vmc 0.9735 −0.0938 removal of the impression molds from the specimen, Vvc 0.9713 −0.0203 which was set as 5–6.5 min after mixing. If the pressure Vvv 0.9336 −0.1865 with putty-type silicone had been continued until the Spd 0.0902 0.8415 completion of hardening, the slow completion type may Spc 0.5694 0.5266 have presented a better impression than the resultant S10z 0.9674 0.1615 impression of this study. Nevertheless, we suggest that the S5p 0.8451 0.3353 5–6 min completion of curing is a reasonable time for S5v 0.9435 0.0370 DMTA research, considering the procedures for making Sda 0.1007 −0.7624 many impression molds of tooth specimens in a limited Sha 0.1177 −0.7742 working time (e.g., visiting museums). It must be noted Sdv 0.0167 −0.5166 that the data of cure time provided by companies were not Shv 0.1496 −0.6440 concordant with dynamic viscoelasticity measured in this Sk 0.7664 0.4959 study; the provided cure time data of the slow completion Spk 0.5859 0.4824 type do not differ from those of the rapid completion Svk 0.6901 0.4659 type (Table 1). This is partly because the cure time in epLsar −0.2506 0.3188 product material safety data sheet (MSDS) is idealized Asfc 0.4542 0.7987 for the application of in vivo human dentition in the oral cavity, whose temperature is much higher than the ambient temperature condition of 25◦ C for measuring the dynamic is the best indicator to find appropriate dental impression viscoelasticity in this study. Also, there should be variability materials, which can mold accurately the tooth surface as due to production batches. The divergence of acquired President. In our study, dynamic viscoelasticity categorized the viscoelasticity and shrinkage data in the present study from analyzed silicones into two types based on the hardening speed. product MSDS should be further investigated in future Despite a wide range of initial values, the rapid completion type collaboration with professionals of dental materials and product commonly shows steep viscoelastic curves. The silicones of this manufacturers. type include President, Affinis regular, Affinis light, and Dr. Silicone. They show better reproducibility of microwear features Delayed change in shrinkage, not exceeding with less blurring and less air bubble contamination than the 0.2% in 200 s after mixing slow completion type (Figures 1, 3, 4). Especially, Dr. Silicone The analyzed silicones show more complicated patterns presents better reproducibility of fine microwear features than for the time-sequential shrinkage rate than for dynamic others. In contrast, the slow completion type, showing gradual viscoelasticity. In our study, the silicones that do not exceed the viscoelastic curves, did not complete hardening in 600 s from 0.2% shrinkage in 200 s after mixing showed better impression the start. Nevertheless, according to the physical properties of accuracy. These products were also categorized as the rapid Frontiers in Ecology and Evolution 09 frontiersin.org Sawaura et al. 10.3389/fevo.2022.975283 FIGURE 4 Scatter plots of the PC scores obtained from 32 DMT parameters, presented respectively for each scan location on facet 9 (f9-1, f9-2, and f9-3). Variation of repeated scanning of the original tooth surface was presented by the ellipses, within which 90% of the data lie. Symbols for the dental impression materials are shown in right lower corner. completion type, which is also associated with better impression Medium viscosity accuracy. In contrast, poor impression accuracy was observed Among the four products categorized as the type with for the products that began shrinking immediately after mixing. rapid completion of curing through the measurements Delayed change in shrinkage with better impression of dynamic viscoelasticity, President and Affinis (regular) accuracy is related to the timing of applying the silicone to the are medium viscosity, whereas Affinis (light) is low specimen (approximately 70 s after mixing) and pressing it to viscosity in the ISO 4823 classification. Although the the specimen with another putty-type silicone (approximately viscosity classification is unknown for Dr. Silicone, 180–200 s after mixing) (Figure 2B). The delayed start based on handling these materials, the product is of shrinkage seemed to provide enough time for silicone assumed to be medium viscosity. These medium viscosity compounds to extend over the microscopic details. On the products show more accurate impressions of microwear other hand, the materials that cured slowly and began to shrink features, which is concordant with the previous finding relatively early in the hardening process were less accurate (Goodall et al., 2015). Affinis light body was overall and their molds showed blurred surfaces (JM Silicone and good in parameter values, but from our experience, IMPRINSIS). This implies that the molds were peeled off from this material should contain a larger amount of oil the tooth surface at the microscopic level before capturing than medium viscosity materials and the oil seems to details due to the shrinkage, as the shrinkage speed might exceed be associated with the slightly blurred surfaces of the the curing speed. molds. Frontiers in Ecology and Evolution 10 frontiersin.org Sawaura et al. 10.3389/fevo.2022.975283 Air bubble contaminations Ethics statement Adhesion of air bubbles was remarkable in the materials Ethical review and approval were not required for this not only with relatively high viscosity immediately after the study as this study used a fossil museum specimen and the mixing but also with slow completion of curing (EXAHIFLEX, investigation of the fossil was officially approved by the curator. JM Silicone, IMPRINSIS, and SILDE FIT). Exceptionally, very few bubbles were observed in President molds despite the high initial viscosity. There may be some material characteristics Author contributions or components that make it difficult for bubbles to be mixed. Anyway, the factors that contribute to bubble contamination YK conceived the experiments. RS and YK conducted the and deaeration would need to be considered with other physical experiments. MOK and RS analyzed the results. All authors properties of materials and molding procedures. reviewed the manuscript. As pointed above, the materials in combination of the rapid completion type and delayed change in shrinkage result in higher impression accuracy than other materials. This was supported not only by qualitative evaluation based on Funding comparisons of 2D images but also by quantitative evaluation based on statistical analysis using 32 DMT parameters. This work was funded by JSPS grants No. 18K12567 (to RS), Nos. 18K13650 and 21K15176 (to YK), and Nos. 16K18615 and 22H00027 (to MOK). Conclusion For seeking dental impression materials suitable for Acknowledgments DMTA research, this study quantitatively evaluates the We thank Tai Kubo (The University Museum, The physical properties of silicone impression materials (changes University of Tokyo), Eisuke Yamada (Yamanashi Prefectural of viscoelasticity and shrinkage over time) and the accuracy Museum), and Konoka Aiba (Japan Wildlife Research Center) of dental impression molds by DMTA. Our results, show that for supporting our study. dynamic viscoelasticity is strongly related to the accuracy of silicone impression materials. On the timing of the impression molding procedure, it is important for the impression accuracy that the curing of silicone occurs immediately after mixing Conflict of interest and completes by the time the mold is removed from the specimen, while the shrinkage starts after application to the The authors declare that the research was conducted in the specimen. In addition, it is suggested that the physical property absence of any commercial or financial relationships that could information supplied by the product companies is not practical be construed as a potential conflict of interest. in searching for impression materials suitable for DMTA, even if it is measured according to ISO standards. Further collecting and examining the data on physical properties of other dental Publisher’s note impression materials, which are globally or locally purchasable, is to be of use for standardizing the methodology of DMTA. All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the Data availability statement reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed The original contributions presented in this study are or endorsed by the publisher. included in the article/Supplementary material, further inquiries can be directed to the corresponding authors. In addition, we uploaded original raw scan data saved Supplementary material in “.vk3” format, all the analytical files in “.mnt” format, and processed surface 3D data in “.sur” format to the The Supplementary Material for this article can be online repository Zenodo. DOI of the uploaded files is found online at: https://www.frontiersin.org/articles/10.3389/ https://doi.org/10.5281/zenodo.7052345. fevo.2022.975283/full#supplementary-material Frontiers in Ecology and Evolution 11 frontiersin.org Sawaura et al. 10.3389/fevo.2022.975283 References Aiba, K., Miura, S., and Kubo, M. O. (2019). Dental microwear texture analysis variability in Neandertal behavior. J. Hum. Evol. 105, 13–23. doi: 10.1016/j.jhevol. in two ruminants, Japanese serow (Capricornis crispus) and sika deer (Cervus 2017.01.004 nippon), from Central Japan. Mamm. Study 44, 183–192. doi: 10.3106/ms2018- Kubo, M. O., and Fujita, M. (2021). Diets of Pleistocene insular dwarf 0081 deer revealed by dental microwear texture analysis. Palaeogeogr. Palaeoclimatol. Caporale, S. S., and Ungar, P. S. (2016). Rodent incisor microwear as a proxy for Palaeoecol. 562:110098. doi: 10.1016/j.palaeo.2020.110098 ecological reconstruction. Palaeogeogr. Palaeoclimatol. Palaeoecol. 446, 225–233. Mihlbachler, M. C., Foy, M., and Beatty, B. L. (2019). Surface replication, doi: 10.1016/j.palaeo.2016.01.013 fidelity and data loss in traditional dental microwear and dental microwear texture DeSantis, L. R. G., and Patterson, B. D. (2017). Dietary behaviour of man- analysis. Sci. Rep. 9:1595. doi: 10.1038/s41598-018-37682-5 eating lions as revealed by dental microwear textures. Sci. Rep. 7:904. doi: 10.1038/ Nagrath, R., Lahori, M., and Agrawal, M. (2014). A comparative evaluation of s41598-017-00948-5 dimensional accuracy and surface detail reproduction of four hydrophilic Vinyl Galbany, J., Martínez, L. M., López-Amor, H. M., Espurz, V., Horaldo, O., polysiloxane impression materials tested under dry, moist, and wet conditions- Romero, A., et al. (2005). Error rates in buccal-dental microwear quantification an in vitro study. J. Indian Prosthodont. Soc. 14, 59–66. doi: 10.1007/s13191-014- using scanning electron microscopy. Scanning 27, 23–29. doi: 10.1002/sca. 0365-z 4950270105 Purnell, M. A., Hart, P. J. B., Baines, D. C., and Bell, M. A. (2006). Quantitative Goodall, R., Darras, L., and Purnell, M. (2015). Accuracy and precision of silicon analysis of dental microwear in threespine stickleback: a new approach to analysis based impression media for quantitative areal texture analysis. Sci. Rep. 5:10800. of trophic ecology in aquatic vertebrates. J. Ani. Ecol. 75, 967–977. doi: 10.1111/j. doi: 10.1038/srep10800 1365-2656.2006.01116.x Grine, F. E., Ungar, P. S., and Teaford, M. F. (2002). Error rates in dental Rodriguez, J. M., and Bartlett, D. W. (2011). The dimensional stability of microwear quantification using scanning electron microscopy. Scanning 24, 144– impression materials and its effect on in vitro tooth wear studies. Dental Mater. 153. doi: 10.1002/sca.4950240307 27, 253–258. doi: 10.1016/j.dental.2010.10.010 Hasegawa, Y., Yamauti, H., and Okafuji, G. (1968). A fossil assemblage of Schulz, E., Calandra, I., and Kaiser, T. M. (2010). Applying tribology to teeth of Macaca and Homo from Ojikado-cave of Hiraodai karst plateau, northern Kyushu, hoofed mammals. Scanning 32, 162–182. doi: 10.1002/sca.20181 Japan. Trans. Proc. Paleontol. Soc. Japan New Ser. 69, 218–229. Scott, R. S., Ungar, P. S., Bergstrom, T. S., Brown, C. A., Grine, F. E., Teaford, Kay, R. F., and Hiiemae, K. M. (1974). Jaw movement and tooth use in M. F., et al. (2005). Dental microwear texture analysis shows within-species diet recent and fossil primates. Am. J. Phys. Anthropol. 40, 227–256. doi: 10.1002/ajpa. variability in fossil hominins. Nature 436, 693–695. doi: 10.1038/nature03822 1330400210 Ungar, P. S. (2015). Mammalian dental function and wear: a review. Biosurf. Krueger, K. L., Scott, J. R., Kay, R. F., and Ungar, P. S. (2008). Technical note: Biotribol. 1, 25–41. doi: 10.1016/j.bsbt.2014.12.001 dental microwear textures of ‘Phase I’ and ‘Phase II’ facets. Am. J. Phys. Anthropol. Ungar, P. S., Brown, C. A., Bergstrom, T. S., and Walker, A. (2003). 137, 485–490. doi: 10.1002/ajpa.20928 Quantification of dental microwear by tandem scanning confocal microscopy Krueger, K. L., Ungar, P. S., Guatelli-Steinberg, D., Hublin, J., Pérez-Pérez, A., and scale-sensitive fractal analyses. Scanning 25, 185–193. doi: 10.1002/sca. Trinkaus, E., et al. (2017). Anterior dental microwear textures show habitat-driven 4950250405 Frontiers in Ecology and Evolution 12 frontiersin.org