Effect of Horizontal Resolution of Printer on Trueness of 3D-Printed Provisional Crown: An In Vitro Study

Abstract Objectives  Provisional crowns play an integral role in prosthodontic therapy and need to be fabricated with high accuracy to ensure good marginal fit and proper contour. The aim of this study is to evaluate the effect of 3D printer's horizontal pixel resolution on trueness of the intaglio surface, external surface, and marginal area and the actual marginal adaptation of the interim crowns. Materials and Methods  A gypsum reference model of mandible with a prepared right first molar was scanned with a dental laboratory scanner (AutoScan-DS-MIX, Shining 3D) and a digital provisional crown was design using the computer-aided design (CAD) software (Exocad). The provisional crown was manufactured by two printers with different horizontal resolutions (Sonic Mini 4K Printer and Sonic Mini 8K Printer). The printed crowns were scanned using the aforementioned scanner. The trueness of the external surface, internal surface, and marginal area of the provisional crowns was evaluated by comparing the scanned data with the reference CAD design data using a 3D inspection software (Geomagic Control X, 3D systems). The trueness of the crown manufactured by two printers was compared by a two-sided test ( α  = 0.05). Finally, the marginal adaptation of the provisional crowns was evaluated on the printed removable dies and compared. Results  The results revealed that there was a significant difference in the trueness of the marginal area and the marginal adaptation ( p  < 0.05). Conclusions  It can be concluded that the trueness of provisional crown marginal area and the marginal adaptation was affected by the horizontal resolution of the 3D printer.


Introduction
Provisional crowns play an integral role in prosthodontic therapy. They are necessary despite the fact that early placement of a definitive restoration can be done a few days after tooth preparation. 1,2 As the template for final restoration, provisional crowns need to be fabricated with with silicone impression or thermoplastic tray. 5 Recently, provisional crown manufacture with computer-aided design (CAD)/computer-aided manufacturing (CAM) technology has become a popular option. 6,7 The CAD/CAM method reduces human and material errors that usually happen with conventional methods. 8,9 With the CAD/CAM workflows, a digital impression is adequate for manufacturing both provisional and final prostheses, which has the advantage of saving time and material. 10 Several studies have assessed the effects of printer technology, printing parameters, and post-curing parameters on the accuracy of the 3D-printed provisional crowns. The stereolithography apparatus (SLA) and digital light processing (DLP) technologies produce interim crowns with similar trueness for intaglio surface, but not for the occlusal region. 10 The post-curing time of 10 minutes is optimal for the accuracy of the provisional crowns. 2 The recommended build angle of SLA 3D printing is between 150 and 210 degrees to achieve high trueness of provisional crowns. 7 Horizontal resolution or XY resolution is a parameter representing the horizontal accuracy of the printer, which can be defined as the size of each printing pixel in the horizontal plane. In other words, the printer with a smaller size of each pixel would have higher accuracy and ability to print with more details. The recent introduction of 8K liquid crystal display (LCD) printer with better horizontal resolution is promising to improve the accuracy of 3D-printed interim crowns. The Phrozen Sonic Mini 8K Printer (Phrozen, Hsinchu, Taiwan) has the horizontal resolution of 22 µm and 1,152 ppi, and is more accurate than common dental 3D printers such as Asiga Max UV (62 µm), Ackuretta Sol (49 µm), Planmeca Creo C5 (50 µm), NextDent 5100 (65 µm), Formlabs Form 3þ (25 µm), EnvisionTEC Vida HD (50 µm), and EnvisionTEC D4K Pro (25 µm). 11 Currently, there is no article comparing the accuracy of printed provisional crowns fabricated with different horizontal resolutions of different 3D printers. Thus, this study aimed to evaluate the effect of a printer's horizontal pixel resolution on the trueness of the intaglio surface, external surface, and marginal area and the actual marginal adaptation of the interim crowns. The null hypothesis of this study was that there is no difference in the trueness and marginal adaptation of provisional dental crowns manufactured by two 3D printers with different horizontal resolutions.

Materials and Methods
The research protocol is illustrated in a flowchart in ►Fig. 1. A gypsum reference model of the mandible with a prepared right first molar was scanned with a dental laboratory scanner AutoScan-DS-MIX (Shining 3D, Hangzhou, China), which had an accuracy of 7 µm according to the manufacturer. The preparation was performed by an experienced prosthodontist with a chamfer finish line using a diamond bur (TF-11; Mani, Tochigi, Japan) with a round end taper shape.
A provisional crown was designed based on the digital scan using DentalCAD software (Exocad, Darmstadt, Germany) and exported as a Standard Tessellation Language (STL) file, which would be used as a reference model. 12,13 The STL file was imported into the Lychee Slicer printing software (Mango 3D, Bordeaux, France) and exported as Chitubox format for printing with two different printers, that is, Sonic Mini 4K Printer and Sonic Mini 8K Printer with the horizontal resolutions of 52 and 22 µm, respectively (n ¼ 10 per printer resolution). All the support designs of the crowns were identical. The building angle was 180 degrees with the occlusal surface facing the platform and the layer thickness was 30 µm. 14,15 A class II temporary resin-branded TEMP resin (Sernetek, İzmir, Turkey) was used with recommended printing and post-curing parameters by the manufacturer.
After cleaning and drying the restorations, scan spray was applied. Then the whole crown surfaces, that is, external surface, intaglio surface, and marginal area of the 20 provisional crowns were scanned using a dental laboratory scanner AutoScan-DS-MIX (Shining 3D) from multiple directions and saved as STL files. Thus, the files were divided into two groups, a group scanned from crowns that were printed with a horizontal resolution of a 52-µm printer (52-HRPPC) and a group scanned from the crowns that were printed with a horizontal resolution of a 22-µm printer (22-HRPPC). These two groups were called the 52-HRPPC group and the 22-HRPPC group, respectively. After each scan, the calibration test was performed before new scans. The unnecessary parts of the support during scanning were deleted. All these processes were performed by the same operator with 3 years of experience.
The STL files (10 per group) exported from these optical scans were then compared with the CAD reference model design file to evaluate trueness using 3D inspection software (Geomagic Control X, 3D Systems, Morrisville, North Carolina, United States). 16,17 Each test model file was directly superimposed on the reference model file using the best-fitting alignment method. Each crown was divided into three parts including the intaglio surface, external surface, and marginal area. Then, the dimensional difference between the test model and the reference model was evaluated for each part. The root mean square (RMS) value was used to assess the trueness of the interim crowns according to the following formula 18,19 : For each comparison, n is the total number of all data points measured, and X 1,i and X 2,i are the coordinates at point i in the reference data and test data, respectively. RMS is the common value used in evaluation of the trueness and precision of protheses and copings. 2,20,21 A higher RMS value indicates lower 3D trueness. Finally, to visualize the deviation analysis, colorimetric maps were generated to illustrate 3D deviation with a range from -100 µm (blue) to þ100 µm (red) and a tolerance range of AE 10 µm. In the colorimetric maps, the green range represents the allowable deviation, while the red portions illustrate positive deviation, meaning the printed provisional crown dimension was larger than the reference CAD dimension, and the blue areas represent negative deviation, indicating that the printed crown dimension was smaller than the reference CAD dimension.
After obtaining the trueness assessment, 20 removable dies of the prepared abutment were printed with the corresponding printer. The provisional crowns were placed on the dies and examined by two calibrated prosthodontists under magnifying glasses for evaluation of marginal adaptation using the same method as that of Lerner et al. 16 The score of each assessment was graded from 5 (excellent) to 1 (very poor). Three days before the real examination, the two prosthodontists were calibrated with 20 dies and 20 corresponding crowns. First, the prosthodontists made the examination separately. Any discrepancy cases were discussed until getting a 100% consensus. The calibration was repeated with the same crowns and dies in the next 2 days until reaching 100% consistency. However, since any discussion has not been done in the real examination, the scores from both examiners were collected separately and analyzed.
All data analyses were performed using a statistical software package (IBM SPSS Statistics v23.0, IBM Corp, Armonk, New York, United States). First, the normal distribution of the data was confirmed using the Shapiro-Wilk test. The differences between the two groups were analyzed using independent sample t-test (α ¼ 0.05).

Results
The trueness of the provisional crowns was evaluated according to the horizontal resolution of the LCD printer. Two-sided sample t-test results of the RMS value of each crown surface were presented in ►Table 1. For all three assessed crown areas, the average RMS values of the 22-HRPPC group were lower than those of the 52-HRPPC group, but significant difference was found only in the marginal area (p ¼ 0.008), which meant the trueness of the provisional crown's marginal area in the 22-HRPPC group was greater than that of the corresponding crown in the 52-HRPPC group (p ¼ 0.008). In addition, the RMS value of the marginal area was lower than that of the external and intaglio surfaces.
In the visual deviation analysis, the deviation results of the external surface of the two groups are illustrated in ►Fig. 2, with high positive values in the occlusal pits and fissures. In contrast, the external axial surfaces showed negative values in both groups. ►Fig. 3 displays the deviation results of the marginal area and the intaglio surfaces of the two groups in which the axial intaglio surfaces reveal positive values, the center of the occlusal intaglio surfaces represents allowable deviation, and the marginal area represents both positive and negative values with small magnitude illustrated by light color range.
On visual inspection, the crowns printed with a 22-µm horizontal resolution printer have smoother surface and less grainy lines than that of the crowns manufactured with a 52µm horizontal resolution printer (►Fig. 4).
The results of the marginal adaptation evaluation by two different prosthodontists are described in ►Table 2. Both groups had high scores (4.9 AE 0.29 in the 22-HRPPC group and 4.4 AE 0.72 in the 52-HRPPC group) indicating great actual marginal adaptation of both printed provisional crowns (►Fig. 5). These two-sided sample t-test results indicated that the marginal closure of the 22-HRPPC group was significantly higher than that of the 52-HRPPC group with prosthodontist 1's evaluation (p < 0.05) and overall evaluation (p < 0.01).

Discussion
In this study, the trueness and marginal adaptation of provisional crowns manufactured with LCD printing method were examined according to different horizontal resolutions of 3D printers. Based on the results, the research null hypothesis was rejected, indicating that the trueness of the provisional crown's marginal area and the actual marginal adaptation were affected by the printer's horizontal resolution. In general, current printers used in dentistry had various horizontal resolutions ranging from 25 to 65 µm. 22 In this in vitro study, two printers manufactured by the same company with horizontal resolutions of 22 and 52 µm were selected for comparison.
Instead of measuring the gap between the crown and the prepared tooth, we examined the trueness of the crown indirectly by comparing the reverse engineering version of the 3D printed crowns with the original CAD file. The advantage of this method over the silicone replica technique was the simplicity without wasting impression material or requiring microscope and the ability to evaluate the trueness of the external surface of the crown. 23,24 Furthermore, the generated colorimetric maps illustrated the amount of     deviation between the crowns and the reference CAD file in the entire crown surfaces. However, some errors may occur during printing; support removing, scanning, spraying, and even best-fitting alignment in the inspection software; and the total errors may affect the results. 25 Hence, actual marginal adaptation of provisional crowns was evaluated by placing them on printed removable dies and performing assessment by two different experienced prosthodontists to support the results of the trueness evaluation.
To minimize printing variables, the build orientation was 180 degrees with identical supports attached to the occlusal surface, and the layer thickness was set to 30 µm for all provisional crowns. 26,27 The laboratory scanner equipped with blue light technology has the accuracy of 7 µm; hence, the scanning error was minimal. 28 The provisional crown surfaces were also coated with an antireflection substance (IP Scan Spray, IP-Division, Haimhausen, Germany) before taking the optical impression to reduce errors due to light reflection. The thickness of this scan spray is only 2 µm according to the manufacturer, so it may have negligible effect on the scanning accuracy. [29][30][31] The areas where printing and scanning supports were placed were not selected to calculate the RMS value.
According to the results, the trueness of the marginal area was higher than that of the other surfaces in both groups. The reason may be that the marginal area was the last formed layers so it is adequately supported by previous printing layers. Furthermore, this area was superficial so it could be accurately acquired during scan. In contrast, the external surface, especially the occlusal surface, had many printing supports and these supports may affect the crown accuracy during removal. Also, due to the building angle, the first printing layers were located on the occlusal surface and these layers were insufficiently supported, affecting the trueness. The high RMS value of the intaglio surface may have resulted from the errors during scan since the crown walls interfere with the light from the scanner toward the axial portion and the junction between the axial portion and the occlusal portion of the intaglio surface. However, an exception occurred in the center of the occlusal intaglio surface, which had allowable deviation as indicated on the colorimetric maps. Less scanning light interference in this portion may be the reason for the exception.
In the present study, the average RMS value of all crown surfaces in both groups was within 100 µm, which was consistent with other studies' results. 2,7 This 100-µm limit was required for accurate fit of the restoration. 32 Both crown groups had high scores of marginal adaptation quality, which was evaluated by two experienced prosthodontists. The provisional crowns manufactured with smaller horizontal pixel size of the printer had a smoother surface, higher marginal area trueness, and better marginal closure. Currently, there are many 3D printers with different horizontal resolution on the market. The results of this study may be used as a reference for dentists in choosing a suitable printer with good horizontal resolution for making interim crowns. However, the fit of the provisional crown may be affected by other morphological factors; so future studies should be performed to measure marginal and internal gaps. 33 This study has several limitations. First, the steps of making a stone die and converting it to a digital die may have errors, so the high trueness printed provisional crown may still not have the perfect fit with the real abutment tooth. Second, only LCD 3D printing was used, and only one printing material and printers from only one company were considered. Hence, more studies should evaluate the trueness of provisional crowns manufactured by the SLA and DLP methods with different printing materials according to the horizontal resolution of the printer. Also, the actual marginal adaptation should be evaluated under a microscope for a more significant result. Furthermore, in vivo studies in the oral cavity should be conducted to confirm the marginal fit, proper contour, and occlusal as well as proximal contacts. Nonetheless, this work has shown that more trueness of the provisional crown's marginal area could be achieved with better horizontal resolution of the 3D printer.

Conclusion
Considering the limitations of the present in vitro study, it can be concluded that the trueness of the provisional crown marginal area and the marginal adaptation was influenced by the horizontal resolution of the 3D printer. For improved marginal area trueness and marginal adaptation of provisional crown, a printer with a higher horizontal pixel density should be selected over a printer with a lower density.

Note
The manuscript has been read and approved by all the authors. The data presented in this study are available on request from Dr. Nguyen Viet Anh at bsvietanhnien-grang@gmail.com.

Funding
None.