Flexural strength, surface roughness, micro-CT analysis, and microbiological adhesion of a 3D-printed temporary crown material

To evaluate the thermocycling effect of 3D-printed resins on flexural strength, surface roughness, microbiological adhesion, and porosity. 150 bars (8 × 2 × 2 mm) and 100 blocks (8 × 8 × 2 mm) were made and divided into 5 groups, according to two factors: “material” (AR: acrylic resin, CR: composite resin, BIS: bis-acryl resin, CAD: CAD/CAM resin, and PRINT: 3D-printed resin) and “aging” (non-aged and aged – TC). Half of them were subjected to thermocycling (10,000 cycles). The bars were subjected to mini-flexural strength (σ) test (1 mm/min). All the blocks were subjected to roughness analysis (Ra/Rq/Rz). The non-aged blocks were subjected to porosity analysis (micro-CT; n = 5) and fungal adherence (n = 10). Data were statistically analyzed (one-way ANOVA, two-way ANOVA; Tukey’s test, α = 0.05). For σ, “material” and “aging” factors were statistically significant (p < 0.0001). The BIS (118.23 ± 16.26A) presented a higher σ and the PRINT group (49.87 ± 7.55E) had the lowest mean σ. All groups showed a decrease in σ after TC, except for PRINT. The CRTC showed the lowest Weibull modulus. The AR showed higher roughness than BIS. Porosity revealed that the AR (1.369%) and BIS (6.339%) presented the highest porosity, and the CAD (0.002%) had the lowest porosity. Cell adhesion was significantly different between the CR (6.81) and CAD (6.37). Thermocycling reduced the flexural strength of most provisional materials, except for 3D-printed resin. However, it did not influence the surface roughness. The CR showed higher microbiological adherence than CAD group. The BIS group reached the highest porosity while the CAD group had the lowest values. 3D-printed resins are promising materials for clinical applications because they have good mechanical properties and low fungal adhesion.


Introduction
Provisional restorations are essential for successful rehabilitation with fixed dental prostheses [1]. They ensure the health of periodontal tissues [2], protect the preparation against chemical and physical agents [3], allow comfort and patient satisfaction [4], restore masticatory function [5] and esthetics [6], and provide adjustments when required [7] until the cementation of the final restoration. In clinical practice, they may be used in short-term procedures while waiting for the final dental prosthesis [8], or in long-term procedures, when it concerns orthodontic, endodontic, or complex surgical therapies [9].

3
To achieve the success and predictability of rehabilitation treatment, biological, esthetic, and mechanical parameters must be followed from the provisional restoration stage [10], depending on material properties such as polymerization shrinkage, wear resistance, tensile strength, and color stability [11], fracture resistance [12][13][14], surface roughness [15], and biofilm adhesion [16][17][18]. In this context, chemically and thermally activated acrylic resin (AR: powder/liquid) remains the most commonly used material in daily clinical practice [19]. However, new temporary materials have emerged to improve the mechanical and esthetic properties of provisional restorations [20,21]. These include bis-acrylic resins, polymethylmethacrylate (PMMA) blocks milled in computer-aided design and computer-assisted manufacturing (CAD/CAM), and 3D-printed light-curing resins [22].
The advent of CAD/CAM technology has enabled provisional restorations to be performed via digital workflow using additive techniques (rapid prototyping) or subtractive techniques (milling) [2,23]. With these systems, rehabilitation may be conducted in terms of time efficiency, providing quality treatment to patients [24]. In addition, it allows sharing and digital data storage [25] and reduces the risk of mistakes [26]. Studies have shown that dental prostheses obtained by additive and subtractive methods present a higher level of precision and marginal adaptation when compared to conventional methods [27,28]. Previous studies have evaluated the mechanical properties of several provisional materials by comparing materials manufactured by CAD/CAM technology with those obtained by a conventional method. Simultaneously, these studies have considered wear resistance and surface roughness simulating masticatory cycles [22], fracture resistance and surface hardness [29], flexural strength [9,[30][31][32], microhardness [31], degree of conversion, and modulus of elasticity [20].
Several factors can influence the mechanical and biological behaviors of these materials, such as processing and polishing techniques [33][34][35] and humidity and temperature variations [36,37]. The thermocycling effect on these temporary materials has been evaluated; results indicate that surface roughness [37,38] and flexural strength [30,39,40] undergo changes due to the artificial aging process. However, direct comparisons of the mechanical properties between conventional temporary materials and 3D-printed resins simulating behavior in the oral environment are rarely conducted.
Therefore, this study aimed to evaluate the mechanical properties of 3D-printed biocompatible resins, with and without aging, and their surface microbiological adhesion, compared to dental materials commonly used in provisional crown fabrication to determine the applicability and predictability of these materials. The following were the hypotheses tested: (1) the provisional material will influence the flexural strength and surface roughness, (2) cell adhesion will be influenced by the materials used, and (3) thermocycling will reduce the flexural strength and surface roughness of temporary materials. Table 1 lists the trade names, manufacturers, chemical compositions, and batch numbers used in this study.
The samples were prepared according to the protocol described below for each resin material: Acrylic resin (AR) The AR group was considered the control group. The powder and liquid of self-curing AR (Dencor, São Paulo, Brazil) were provided according to the manufacturer's recommendations and introduced into a silicone mold measuring 10 × 2.5 × 2.5 mm. To achieve uniformly smooth surfaces, a glass plate was placed on the mold and left for 5 min until the final polymerization of the resin.

Bis-acryl resin (BIS)
Bis-acryl resin (Protemp 4, 3 M ESPE, CA, USA) was applied in a silicone mold with the same dimensions as described previously, in accordance with the manufacturer' mixing tip. Subsequently, a glass plate was supported to smooth the surface, fill the mold, and wait for 2 min until complete polymerization.
Composite resin (CR) Using a silicone mold with the same dimensions described previously, 2.0-mm increments of composite resin (Z350XT, 3 M ESPE, Irvine, CA, USA) were inserted into the mold using a suprafill spatula n. 1 (Golgran, São Caetano do Sul, SP, Brazil). Each increment was light-cured for 20 s in the standard mode power energy 1200 mW/cm 2 (Radii Cal, SDI, Australia). The power of the light-curing unit was measured by a single operator with a radiometer for LEDs (LED Radiometer -Kondortech, SP, Brazil). After the last increment of the resin was inserted, a glass plate was supported to smooth the surface, filling the mold, and light curing for 60 s was carried out.

3D-printed resin (PRINT)
The samples measuring 10 × 2.5 × 2.5 mm were designed by modeling a bar in mesh mixer software (Autodesk, Inc.) and files in standard triangle language were obtained and exported to printer software for slicing and drawing the print sprues. The samples were then printed using the digital light processing technique on a 3D printer (Miicraft Ultra 125, NY, USA) at a low printing speed, with a layer thickness of 65 μm and vertical position in Cosmos Temp photosensitive resin (A1; Yller Biomaterials, Pelotas, RS, Brazil). After processing, the samples were immersed in isopropyl alcohol for 3 min and post-cured for 7 min in an Anycubic Wash and Cure Plus ultraviolet light machine (Anycubic, China). The structural supports were removed.
All the bars were finished with #1200 grit SiC abrasive papers (3 M ESPE, Irvine, CA, USA) to dimensions of 8 × 2 × 2 mm, using a digital caliper (150/6 in.; Velleman 3472 B, Fort Worth, TX, USA). The #1200 grit was used because provides results similar to dental polishing disks commonly used to finish dental restorations [41]. Figure 1 shows a schematic representation of the preparation and fabrication of the bars from the silicone mold. The nonaged samples were stored in distilled water (37 °C) for 24 h and subjected to flexural strength and roughness tests. The aged groups were subjected to the thermocycling protocol described below.

Mini-flexural strength test
The bars were subjected to a three-point mini-flexural strength test with reduced dimensions, unlike those recommended by ISO 6872/2015 [42,43]. The tests were performed using a universal testing machine (ODEME-ISO150; Anchieta, SC, Brazil). The two rollers were positioned at 6 mm, and the tip of the device was placed at the center of the upper surface at a displacement speed of 1.0 mm/min and a load cell of 100 kgF. Based on this protocol, pressure was applied perpendicular to the center of the bar until the material fractured. The load/deflection curve was recorded, along with the maximum force applied, which represented the flexural strength value in MPa. Flexural strength (σ) was recorded according to the Eq. 3lF/2wh 2 . Here, l is the distance (mm) between the lower pins, F is the critical load (N) applied at the time of specimen failure, h is the thickness (mm) of the specimen, and w is the width of the specimen.

Preparation of blocks
A total of additional blocks (n = 20) of each experimental group were made with final dimensions of 8 × 8 × 2 mm, following the recommendations of each material as previously described for the preparation of the bars. Figure 2 presents a flowchart of the study design for the manufacturing of blocks and bars.

Roughness
All blocks were subjected to a roughness test. Ten samples from each group, aged (AR TC , CR TC , BIS TC , CAD TC , and PRINT TC ) and non-aged (AR, CR, BIS, CAD, and PRINT) were used for surface roughness analysis. The test was performed using a digital rugosimeter (Surftest Model SJ-2010; Mitutoyo, Japan). The reading was determined from the arithmetic mean of the roughness (R a ) between the minimum (valleys) and maximum (peaks) values reached by the active tip of the device, covering a distance of 0.25 mm, with a speed of 0.5 mm/s. Moreover, the values of the maximum average roughness (R z ) were obtained, indicating the maximum height, calculated by the maximum sum of the peaks with the maximum depth of the valleys and (R q ), demonstrating the effect of irregularities that deviated from the average of the roughness. For these three parameters, three readings were obtained at 2.0-mm intervals, in different surface areas. The data were averaged to obtain the values for each block.

Porosity
Porosity evaluation (%) was performed using high-resolution micro-CT on a non-aged sample from each group using a scanner (SkyScan 1172; SkyScan, Kontich, Belgium) with an accelerating voltage of 80 kV, a current beam of 124 μA, 0.5-mm Al filter, pixel of 20.03 μm with a 1000 × 666 resolution, a 180° rotation with 0.4 angle, and an exposure time per step of 27 min. For each specimen, 401 vertical slices were recorded and reconstructed using 2D and 3D software (CTAn and SkyScan, respectively). Then, the image was binarized to differentiate the pores of acrylic, composite, bis-acryl, CAD/CAM PMMA, and 3D-printed resins, determining threshold 61 in the "grayscale threshold" (0-255), a value that was determined from the sample analysis of the control group (i.e., AR). With this porosity limit presented, it would be 0% in the group taken as a reference, thus being able to differentiate between the samples of the other groups. This binarization transforms the image to grayscale, wherein gray values above the threshold of 61 are judged as gaps and smaller as compact AR. Subsequently, calculations were performed to determine porosity.

Cell adhesion
Eight blocks of 8 × 8 × 2 mm from each non-aged group (the same samples used in the roughness analysis) were fixed to the bottom of a 24-compartment plate with wax, sterilized by ethylene oxide, and kept at room temperature. The microbiological adherence to the specimens was evaluated using a microbial cell viability assay. Reference strains of Candida albicans (ATCC 90,028) were reactivated in Sabouraud dextrose agar medium (Difco, Detroit, USA) and incubated at 37 °C for 24 h under an aerobic atmosphere. Three to five colonies were collected and suspended in 5 mL sterile saline solution (NaCl 0.9%). A spectrophotometer (λ = 600 nm) was used to determine the cell concentration. Cell density was determined at an absorbance of 0.1 (1.0 × 10 6 cells/ml). The samples were subjected to biofilm formation in artificial saliva (2 mL saliva/specimen) and incubated for 60 min at 37 °C [44].

Statistical analysis
The sample estimate was performed based on previous studies [42,43]. The OpenEpi website (https:// opene pi. com) was used to calculate the power of the study. Comparisons between means and standard deviations were carried out in the experimental groups, for each provisional material type, considering a 95% confidence interval and the sample size per group of 15. Statistical assumptions were assessed before statistical analysis. Statistical analysis was performed using Statistix software (version 8.0, 2003; Analytical Software Inc.). Two-way analysis of variance (ANOVA) and Tukey test (5%) were performed to compare flexural strength (MPa) and surface roughness values (μm) between groups according to "material" and "aging" factors. ANOVA was conducted to assess C. albicans adherence (log CFU/mL

Flexural strength
A sample power of 100% was obtained for the CR group. For AR and CAD groups, the power was 92.50% while for BIS and PRINT was 99.93%. Results of the two-way ANOVA revealed that "material" and "aging" factors were significant (p = 0.0001). The interaction between these factors was also statistically significant (p = 0.0005) (

Weibull analysis
The results of the Weibull analysis for each provisional material with and without aging are presented in Table 3 and Fig. 3. The characteristic strength (σ 0 ) was statistically significant (p < 0.0001). The BIS (124.52 a ) and PRINT (53.47 e ) groups were significantly higher and lower, respectively, than the other experimental groups without aging. Conversely, the CR TC (38.88 f ) groups were statistically lower than those of the other groups subjected to thermocycling (p < 0.0001). The results of the chi-square test showed that the Weibull modulus (m) was significantly lower in the CR group than in the AR, BIS, and CAD

Roughness
The "material" factor was significant (p < 0.0001) for the roughness values (R a , R q , and R z ). The mean of the groups ranged within 0.16-0.35 μm for R a , 0.20-0.46 μm for R q , and 0.98-2.10 μm for R z . The AR and AR TC groups showed significantly higher R a values than the BIS and BIS TC groups, whereas the BIS and CR TC groups showed lower R a , R q , and R z values than the AR and AR TC groups (p < 0.0001). Table 4 lists the mean roughness values.

Porosity
Porosity analysis showed that the BIS (6.339%) and AR (1.369%) groups had the highest porosity values among the experimental groups, while the CAD group (0.002%) had the lowest percentage. Figure 4 presents a two-dimensional representation of the porosity of the tested groups.

Cell adhesion
The Tukey test revealed that, for the microbiological adherence of the viability of Candida albicans strains (log CFU/ mL), the "material" factor was significant between the groups, ranging from the log CFU/mL from 6.00 to 7.07 (p = 0.036). The mean log CFU/mL values were significantly higher in the CR group than that in the CAD group (p < 0.05) (Fig. 5).

Discussion
Currently, there is a tendency to perform rehabilitation based on minimally invasive dentistry. However, it is still frequent rehabilitation with tooth wear that demands the management of provisional restorations aiming at the diagnosis and functional and esthetic predictability, a fundamental step for the success of definitive prosthetic treatment. Specifically, understanding the mechanical, esthetic, and biological requirements of temporary resin materials is crucial for predicting their clinical performance. Therefore, this study aimed to investigate whether the type of temporary restorative material influences flexural strength and surface roughness, based on the effect of the aging protocol. Hence, samples were subjected to a thermocycling process (10,000 thermal cycles) simulating the conditions of the intraoral environment for approximately 1 year [45]. In addition, the porosity and surface microbiological adhesion were analyzed considering the provisional resin material. In the present study, the specimens were made smaller than those recommended by the ISO 6872/2015, and the analysis was performed using a miniature three-point bending jig, so the mini-flexural strength test was applied. The methodology was based on previous studies [42,43]. The hypothesis that the provisional resin material type affects the flexural strength was accepted, as the BIS and PRINT groups presented higher and lower flexural strength values, respectively, which were statistically different from the other groups, even when compared to the control group (i.e., AR). Little is known about the mechanical properties of 3D-printed resin materials because they are innovative products. Therefore, samples obtained by rapid prototyping can be influenced by the manufacturing technique [46], modified during post-curing, or suffer shrinkage because of the minimum layer thickness [47], presenting a lower strength compared to milled ones. Materials used to fabricate provisional materials have different chemical compositions and physical properties. Due to the filler content and dimethacrylates in their composition, bis-acryl resins have a higher flexural strength than conventional resins [48][49][50]. Results of Weibull analysis showed that the characteristic strength values tended to follow the same pattern as the Tukey test results, demonstrating the greater reliability of the results.  The first hypothesis tested in the study (i.e., the AR group showed the greatest surface roughness (R a )) was partially accepted as it was statistically different only in comparison to the BIS group. Specifically, R a is the parameter that determines the surface quality, considering the amplitude deviations in the sample center [51]. This result corroborates with that of Gantz et al. [52] who reported that the surface roughness of AR (0.40 ± 0.19 μm) after processing was higher than that of bis-acryl (0.12 ± 0.06 μm). The higher R a scores in the AR group can be explained by the bubble formation during the powder/liquid hand mixing of the material, while bis-acryl has a reduced risk because it is manipulated using a dispenser and mixing tips, favoring the homogeneity of the material [53]. In this study, the mechanical properties of roughness and flexural strength were inversely proportional. The BIS group had higher flexural strength and, consequently, lower surface roughness.
Although R a is considered the only roughness parameter evaluated in most previous studies [31,52,54], this study also measured R q (root mean square roughness) and R z (maximum height of the roughness profile based on peaks and valleys) to increase precision. Therefore, when R a values are equal, the restorative material with lower R q and R z values has a lower surface roughness and, consequently, would be more efficient. The results revealed that the R q and R z values for the AR group were significantly higher than those for BIS and PRINT groups. These findings are similar to the those of Giti et al. [55], who reported higher roughness for conventional compounds based on PMMA (12.89 ± 6.8 μm) regarding additive manufacturing (4.09 ± 2.14 μm; p = 0.006).
Micro-CT can provide a powerful, nondestructive 3D approach for quantifying the porosity of materials [56]. Porosity is multifactorial and may be influenced by monomer shrinkage during polymerization [57], air trapping during material mixing [58], or monomer vaporization associated with exothermic reactions [59]. The porosity analysis showed better results for the CAD/CAM PMMA resin, with a lower percentage of porosity and, consequently, a lower mean surface roughness. Porosity values above 11% have The cerium nitrate appear as radiolucent areas similar to the black arrows. The cerium nitrate appear as radiopaque areas similar to the white arrows been associated with reduced mechanical properties and retention of microorganisms [60], and levels lower than this may be clinically acceptable [61]. Although pores were noticeable in the study samples, all the porosity values were less than 6.33%, which is within the acceptable range.
Assessment of the roughness of provisional crowns is essential because it is directly associated with the health of periodontal tissues. A rough surface not only causes discomfort to the patient but also favors the accumulation of biofilm [22], predisposing the patient to bacterial colonization [62], which can lead to infection and periodontal inflammation [63]. Several studies have reported that, even when provisional materials do not have adequate roughness rates, finishing and polishing can improve this property [64][65][66]. However, this study did not evaluate these parameters regarding provisional materials, which may explain the high porosity values of acrylic and bis-acrylic resins.
The second hypothesis that the provisional material was influenced by Candida albicans adherence was partially rejected. Candida albicans was used in this study as it is the most common opportunistic intraoral pathogen isolated from the oral cavity [67]. Once C. albicans infection process is directly associated with adhesion on the material's surface, the properties of the surface roughness and hydrophobicity may play a relevant role in the fungal colonization and adhesion [68,69]. In this situation, no statistical difference was observed between the AR, BIS, and PRINT groups, which can probably be explained by the hydrophobic and anti-adhesive properties of these materials, which hamper microorganism adherence to the surface. The adherence of C. albicans to provisional crowns was previously described by Ozel et al. [67]. The authors reported that the PMMA group showed greater fungal adhesion than the bis-acrylic group and adhesion was not significantly associated with surface roughness. Although the material composition can influence the roughness, affecting microbial colonization and adhesion, and surface roughness was considerably higher for the AR group, no correlation with fungal adhesion was observed.
Thermocycling was used to investigate the behavior of temporary restorative materials, simulating conditions in the intraoral environment. Thermocycling is one of the most effective aging mechanisms [70]. The temperature variation used to perform the test (5-55 °C) was similar to the food temperature ingested during daily meals, without causing damage to oral tissues [71]. The results showed that all evaluated resins were influenced by thermocycling. The BIS TC resin had significantly higher flexural strength values than those of the CR TC and AR TC groups, and the third hypothesis was partially accepted. The flexural strength of all temporary materials with respect to the effect of thermocycling was reduced, except for the PRINT TC group, wherein this mechanical property increased after aging. Mehrpour et al. [39] showed that the flexural strength of bis-acrylic materials is higher than that of other conventional restorative materials, even after thermal cycling, thereby corroborating our findings. However, Haselton et al. [48] compared the flexural strength of bisacrylates and PMMA after 10 days in saliva and reported that only some bis-acrylic resins had similar characteristics to those in our study, which may be related to the specific characteristics of each material. Thus, direct comparisons between materials may not be feasible because of the variety of methodologies and specific designs for obtaining samples.
Regarding the roughness and aging protocol, the hypothesis was partially rejected, as only the CR TC and BIS TC groups had significantly lower roughness values than the AR TC group. After thermocycling, the roughness increased for all experimental groups, except for the CR TC group, which was reduced. In accordance with Oliveira et al. [38], the roughness results showed no significant differences between different manufacturers for acrylic and composite groups. However, the authors showed that, after thermocycling, the surface roughness of these materials increased. The absorption of water and solubility leads to the formation of products by hydrolysis and the release of non-reactive substances, thereby interfering with biocompatibility. Bollen et al. [72] reported that roughness values greater than 0.2 μm favor microbiological adhesion. In this study, the roughness ranged from 0.14 to 0.36 μm, and the evaluated materials had low roughness, being in a clinically acceptable range. Provisional restorations must be used for a considerable amount of time to minimize the effects of microbiological colonization on rough surfaces. Finishing and polishing techniques are highly recommended to improve this property. Therefore, the results of this study suggest that provisional materials produced by rapid prototyping may be promising because they have similar or improved mechanical properties, when compared to conventional materials (i.e., being within the clinically acceptable ranges). However, additional studies should be conducted to evaluate the mechanical properties of temporary materials regarding finishing and polishing techniques and to investigate the biological behavior in situations that simulate the oral environment. Moreover, diversified methodologies regarding the manufacturers of the products studied, variability of thermal cycles, and sample-conditioning methods (distilled water, artificial saliva, or different solutions) hamper direct comparisons between material types. Clinical trials to evaluate the long-term behavior of conventional resin materials, compared to innovative materials (printed and milled), are fundamental for determining their predictability in daily clinics.

Conclusions
Based on the limitations of this study, it can be concluded that thermocycling reduced the flexural strength of most of the evaluated provisional materials. However, surface roughness was not affected when each material was evaluated. Moreover, fungal adhesion by Candida albicans and porosity varied among materials. In addition, 3D-printed resins are more promising materials for clinical applications, when compared with conventional provisional dental materials, because of their good mechanical properties and low fungal adhesion.