A simple solution to recycle and reuse dental CAD/CAM zirconia block from its waste residuals

Purpose: To seek a simple solution that can recycle and regenerate dental CAD/CAM zirconia green blanks from its waste residuals. Methods: Waste residuals (3M® LavaTM Plus HT) were pulverized after dry milling and cutting, and subsequently sieved before pickling in a 0.5 M nitric acid. These powders were then dry-pressed and pre-sintered into blocks at seven different temperatures in the range 800−1100 °C. New zirconia blocks flagged with the same batch numbers were used as control. These blocks were cut into bars before subjected them to manufacturer-recommended sintering at 1450 °C. Crystalline phases (by XRD), elemental compositions (by EDX), surface morphologies (by SEM), machinability, linear shrinkage rate, relative density, and Knoop microhardness were evaluated before and after sintering, and four-point flexural strengths were also evaluated for the sintered zirconia bars. Results: Only tetragonal phases were found in both preand fully-sintered recycled zirconia blocks. SEM results showed that pre-sintered samples at 950 °C had smooth and flat surfaces with evenly distributed particles. Recycled and control zirconia blocks had similar elemental compositions. Results from machined surface, linear shrinkage rate, relative density, and Knoop microhardness established that 950 °C and 1000 °C were suitable pre-sintering temperatures for recycling zirconia. Pre-sintered recycled zirconia had no significant differences in flexural strengths, however, samples pre-sintered at 1000 °C exhibited the closest value (897 MPa) compared to that of the control (904 MPa). Conclusions: Dental CAD/CAM zirconia can be recycled and reused from its waste residuals by adopting a simple method that requires a pre-sintering at 950


Introduction
With many recent developments in modern science and technology, the overall quality of life, medical treatments, and life expectancy have been improved significantly. Consequently, the number and ratio of aging populations in the world are increasing [1]. The global percentages of people aged ≥ 60 years have increased from 9.2% in 1990 to 11.7% in 2013, and it is expected to reach 21.1% by 2050 [2]. In addition, with the improvements in living conditions and increasing competition in modern society, people are paying more attention to their appearances, and especially to their teeth. Hence, aesthetic dentistry is becoming increasingly popular, and there is a keen demand for better dental restoration materials. In the field of dental restorations, porcelain-fusedto-metal (PFM) restorations are commonly used whereas PFM exhibits satisfactory mechanical strength from the bottom metal layer. However, PFM manifests two prominent drawbacks: the metal part restricts the light transmittance of porcelain and can cause allergic reactions in the human body [3]. Subsequently, all-ceramic restorations have the advantages of aesthetics and biocompatibility, and thus, represent ideal materials for dental restoration.
High-performance ceramics provide good foundations for developing high strength all-ceramic dental materials, of which yttria-stabilized zirconia (YSZ) has been under intensive focus recently in the field of dentistry [4][5][6][7][8][9][10][11][12][13]. This material can be toughened by stress-induced transformations and has relatively higher strength and toughness than alumina ceramics. Besides, its mechanical properties, color, and stability make it an ideal material for dental restorations. Nevertheless, zirconia cannot be processed in a dental laboratory by traditional methods (e.g. powder or liquid application), and the processing of robust compact sintered zirconia is a time-consuming procedure that leads to high wear of the milling instruments [14]. Therefore, the fabrication of zirconia are commonly done on prefabricated blanks that require computer-aided design/computer-aided manufacturing (CAD/CAM) to design different dental structures, such as inlays, onlays, veneers, crowns, fixed partial dentures, and implant abutments [15,16], and machine into individual forms by milling processes. Subsequently, the structures are sintered at specific high temperatures to achieve the optimum mechanical strength.
Since the sintering process of zirconia leads to shrinkage, specialized computer software [17] is necessary to compensate for the shrinkage effect by enlarging the original form prior to machining.
Similar to all CAD/CAM milling processes, dental zirconia milling dust (~30%) and block residuals (~80%) as wastes have been reported [18] that incur significant economic and environmental losses [19]. Therefore, an appropriate reuse of such zirconia wastes would certainly increase its ecofriendliness and cost efficiency. Gouveia et al. [19] attempted to recycle dental zirconia, but the recycled product can only be used in the jewelry industry. Therefore, in this study, we aim to seek an easy and clean method to recycle dental CAD/CAM zirconia green blank wastes that can be reused effectively for dental prostheses.

Recycling pre-treatment of zirconia discs
Wastes of 3M® Lava™ Plus High Translucency Zirconia discs (Lot: 579006, diameter: 98 mm, thickness: 25 mm) were collected from inhouse laboratories. These residues were first powdered by a pulverizer (FS-100, Xin Li, China) using a speed of 10,000 rpm, and then sieved through a 325 mesh (48 μm) screen. The powders were soaked, mixed, and stirred properly in a 0.5 mol/L nitric acid for 5 min at 55 °C. Finally, the powders were cleaned thoroughly with distilled water and stored in a drying oven for future uses.

Dry Press of Green Zirconia Body
The dry pressing method, which is a simple process to form green blanks [20], was adopted in this study. Approximately 10.0 g of powder were weighed using an analytical balance (AT201, Mettler-Toledo, Switzerland) and poured into a custom-made rectangular stainless steel mold of dimensions 35.0 mm × 30.0 mm × 10.0 mm ( Figure 1). The mold was compacted under a hydraulic press (Silfradent 660, Italy) using a force of 80 kN to achieve the final size of the recycled zirconia green body (35.0 mm × 30.0 mm × 3.0 mm). Finally, the recycled green body was dried in an oven for 24 h. A total of 21 green body blanks were produced and the dimensions of each pressed green body were recorded. Appropriate new zirconia discs with the same lot were also produced to use as control.

Sintering of Green Body
The recycled green body blocks were subjected to a pre-sintering process in a box furnace (XD-1700M, Brother, China). First, binders and other organic additives were removed by heating the samples to 600 °C using a slow heating ramp of 5 °C / min from room temperature. Then, for each pre-sintering temperature (800, 850, 900, 950, 100, 1050 and 1100 °C), three dry-pressed blanks were heated to the desired temperature from 600 °C at a rate of 10 °C / min. The heating duration was 2 h in each case before the sample was cooled naturally inside the furnace to room temperature. After that, the dimensions of each pre-sintered green body block were measured. Finally, the pre-sintered and control zirconia blocks were cut into 18 bars with dimensions of ~30.0 mm × 5.0 mm × 3.0 mm by using a cutting machine at dry conditions (ISOMET5000, Buehler, USA). Cross sectional pictures of cutting faces from all zirconia bar samples were taken in a 6500K lightbox.
The final stage of sintering was conducted following the manufacturer guidelines as follows. Both pre-sintered recycled green bodies and control zirconia bars were subjected to the same furnace heating: First, from room temperature to 800 °C at a rate of 20 °C/min, and then from 800 to 1450 °C at a rate of 10 °C/min with a holding time of 2 h. In the end, the samples were naturally cooled in the furnace. After that, the dimensions of each sintered sample were measured. Figure 2 shows the full heating curves of pre-sintering and sintering schemes implemented in this work.

Hardness, Linear Shrinkage, and Relative Density
The Knoop hardness of three randomly selected pre-sintered zirconia green bodies sintered at different temperatures as well as the respective fully sintered bar samples were tested by Leitz Micro-Hardness Tester (Leitz Inc. New York, USA) using 25 g and 50 g of loading forces with the help of a software (Leica QGo Software by Leica Microsystems Imaging Solutions Ltd., Wetzlar, Germany). The loading time of each sample was set as 10 s. Three different areas were tested on each bar specimen.
For the determination of linear shrinkages, length, width and thickness of a specimen at dry-pressed (n = 3), pre-sintered (n = 7), and fully sintered (n = 7) stages were measured by a slide caliper for three times and recorded. The linear shrinkage (L) was calculated by the following equation [21]: Where: L 1 = average length, width, or thickness of the sample before heating. L 2 = average length, width, or thickness of the sample after presintering/ sintering.
To estimate the relative density, three bar-shaped samples were randomly selected from the groups of zirconia green bodies pre-sintered at different temperatures as well as the respective fully sintered samples. In general, the following method was followed [22]: (1) weighing the sample dried in an oven, (2) using the boiling absorption method where the sample was put in a container completely submerged in distilled water, and boiled for 2 h, (3) the sample was then taken out of the boiling water and the surfaces were wiped off carefully to avoid absorption of water through the ceramic pores, and (4) weighing the sample immediately in air. The relative density (d) was calculated as follows: Where: m 1 = sample's weight in air. m 2 = sample's weight in air after fully submerged in water. d 0 = theoretical density of the tetragonal phase of zirconia (d 0 = 6.10 g/ cm 3 ).

Flexural Strength Test
Four-point bending tests were conducted in a universal testing machine (ElectroPuls™ E3000, Instron Industrial Products, Grove City, PA, USA) to determine the flexural strengths of seven randomly selected bar specimens from each group after the completion of the full sintering process. Before the test, all samples were polished by SiC abrasive papers (200-, 600-, and 2400-grit papers) to achieve the dimensions of ~24.0 mm × 4.0 mm × 2.0 mm, and the chamfer edges were made in accordance with ISO 6872 [23]. The compressive extension rate was set as 1.0 mm/min, the center-to-center distance between outer support rollers was 20.0 mm and the center-to-center distance between inner support rollers was 10.0 mm. The data was recorded by the console software (Instron Industrial Products, PA, USA).
The flexural strength (σ) was calculated as follows: Where: P = maximum load recorded in the software in Newton. L = center-to-center distance between outer support rollers in millimeters. w = width of the sample in millimeters (the dimension of the side at right angles to the direction of the applied load). b = the thickness of the sample in millimeters (the dimension of the side parallel to the direction of the applied load).

X-ray Diffraction
X-ray diffractometer (SmartLab, Rigaku, Japan) was used to determine the crystalline phases (tetragonal and monoclinic) present in the recycled zirconia powder, recycled pre-sintered zirconia blocks and in the sintered zirconia blocks. Scans were performed at a rate of 1.00°/min in two thetas with a step-size of 0.02° using a source voltage of 40 kV and a current of 40 mA.

SEM and EDX
SEM (SU1510, Hitachi, Tokyo, Japan) and EDX (IXRF System Inc., Austin, TX, USA) were conducted on recycled zirconia powder and zirconia blocks for all different pre-sintering temperatures as well as on the respective fully sintered zirconia blocks. A thin gold layer was coated on each sample by an MSP-2S magnetron sputter. The coating duration was 3 min for all the samples.

Statistical Analysis
Mean values of the results were analyzed by one-way ANOVA using the software SPSS 23 (Chicago, IL, USA) with linear shrinkage, relative density, and Knoop hardness of pre-sintered and fully sintered samples including the flexural strength of fully sintered samples as dependent variables. Independent variables were different pre-sintering temperatures and the parameters of the control zirconia block. The level for all statistical significances was set as 0.05.
It should be noted that the percentile data were arcsine-transformed prior to ANOVA analyses. It is a typical data processing method for percentages ranges from 0 to 100% because the variances of means near 0 and 100% tend to be smaller than the variances of means in the middle of the range (30-70%) [24]. Hence, the percentages based on counts are discontinuous and have a binomial distribution. The arcsine-transformation is needed for error reductions in these types of data [25]. Arcsine-transformation converts the percentage to angle for which the equations are listed below: Where: Y = percentages in decimal fractions Y' = transformed data in degree or radian Figure 3 shows the standard XRD peaks of monoclinic zirconia and tetragonal zirconia together with the measured diffraction patterns of recycled zirconia powder, pre-sintered and fully sintered zirconia blocks at 1000 °C. It can be concluded that the monoclinic phase exists only in pulverized zirconia, while no monoclinic phases can be detected for presintered and fully sintered zirconia blocks.

SEM
Surface morphologies of the blocks from each group were examined using an SEM. For pre-sintered samples, ×250 magnification images were captured to observe their surface roughness, while ×10000 magnification images were captured for the fully sintered samples to evaluate the particle size distributions. The serial SEM images (Figure 4) show the differences in surface topography among samples pre-sintered at different temperatures. Figure 4(a) shows the control zirconia block. It can be observed that the surface of the original block was very flat and smooth. The image on the right depicts the control block after it was fully sintered at 1450 °C that evidently has a uniform distribution of particles. Figures  4 (b)-(h) show the surface morphologies of recycled zirconia blocks presintered at respective temperatures of 800, 850, 900, 950, 1000, 1050 and 1100 °C. For the sample pre-sintered at 800 °C (Figure 4(b)), the surface appeared quite rough compared to the control sample surface and the quality of particledistribution was poor after it was fully sintered at 1450 °C. In addition to those non-uniform particle distributions, several small cracks or cleavages between particles could also be seen. These bad qualities of the surface structure can be directly related to the poor mechanical properties of the fully sintered products.
The samples pre-sintered at 850, 900, and 950 °C are shown respectively in Figures 4(c), 4(d), and 4(e). As the pre-sintering temperature increased,  surfaces of the zirconia blocks became smoother, and more evenly distributed particles with a fewer number of cleavages between the particles were found in the corresponding fully sintered bodies for which the images are shown on the right.
However, for the samples pre-sintered at temperatures ≥ 1000 °C as presented in Figures 4(f)-(h), the surfaces look denser which may lead to more difficulties in the further machining process. In summary, it can be concluded from the SEM results that 950 °C seems to be a suitable temperature for pre-sintering of the samples.

Cutting surface morphologies
The cross-sectional pictures of the pre-sintered zirconia blocks are shown in Figure 5. It is evident that as the pre-sintering temperature increased, more scratches appeared on the cross-sectional surface areas, meaning that the machining became more challenging for them. By comparing the images with the control sample, it can be concluded that the wear degree of the control block was similar to the samples pre-sintered at 950 and 1000 °C.

Energy Dispersive X-ray Spectroscopy
EDX measurements were done on selected areas of zirconia powder and block samples after the imaging in the SEM. The magnification of the selected areas was set as 100 times. The results are shown in Figure 6.
The quantitative analysis of the EDX results established that the recycled blocks had a similar proportion of key elements such as zirconium, yttria, aluminum and oxygen compared to the control block and the powder. As each key element has a distinct role in the system, a different proportion of elements may induce different properties in a recycled block compared to the control block. Figure 7 shows the linear shrinkage of pre-sintered zirconia blocks and fully sintered zirconia blocks. The black and red solid curves show the linear shrinkage rate of the pre-sintered and fully sintered zirconia blocks, respectively, while the red dashed line indicates the linear shrinkage rate of the control zirconia block.

Linear Shrinkage
The black solid curve shows the linear shrinkage rate of the pre-sintered zirconia block. Although the rate initially increased slowly from 800 to 950 °C, there was a sharp increase in the rate starting from the temperature 1000 °C. However, one-way ANOVA results (Table 1) revealed that there were significant differences (p < 0.001) among the samples pre-sintered at different temperatures in the range from 800 to 1100 °C in terms of their linear shrinkage rates.
For the fully sintered systems, recycled zirconia blocks showed a higher linear shrinkage rate compared to the control block. The red solid curve shows that once they were fully sintered, the linear shrinkage rate fluctuation was ~20%. The statistical analyses are shown in Table 2. It could be concluded that the result was significant at 0.05 level on post hoc multiple comparisons between fully sintered control blocks and recycled zirconia blocks with different pre-sintering temperatures. All recycled zirconia blocks showed a significantly higher value in linear shrinkage rate than the control sample (p < 0.001). Figure 8 shows the relative densities of pre-sintered zirconia blocks and fully sintered zirconia blocks. The black and red solid curves show the relative densities of pre-sintered and fully sintered recycled zirconia blocks, respectively. The black dashed line shows the relative density of the control zirconia block before full sintering, while the red dashed line indicates the relative density of the fully sintered control zirconia block.

Relative Density
Similar to linear shrinkage rates, the relative densities of pre-sintered samples increased relatively slowly in the range from 800 to 950 °C. However, for temperatures ≥ 1000 °C, it exhibited a sharp increase. The one-way ANOVA results ( Table 3) also indicate that there were significant differences in relative densities among samples pre-sintered at various temperatures (p < 0.001). This phenomenon can be related to the toughening mechanisms occurring primarily in the temperature range from 1000 to 1100 °C, which could lead to poorer machinable properties.
The red solid curve in Figure 8 shows that the relative density undulates around 95% for the fully sintered samples. From the statistical analysis results (Table 4), it could be concluded that with the level of significance set as 0.05, recycled zirconia samples pre-sintered at 900, 950, 1000, and 1050 °C had no significant difference in their relative densities after the full sintering process, indicating they had similar relative densities compared to what the control zirconia block showed. Figure 9 shows the results of Knoop hardness tests performed for different samples. The black solid curve and the black dashed line represent the hardness of the pre-sintered blocks and the control zirconia block, respectively. The red solid curve and the red dashed line show the hardness of fully sintered samples and fully sintered control block, respectively. In general, it shows nearly a linear correlation between pre-sintering temperatures and Knoop hardness values. The hardness of the control zirconia block before the full sintering was ~95 HK, which was very similar to the value obtained in the case of recycled zirconia block pre-sintered at 1000 °C.

Hardness
One-way ANOVA analysis (Table 5) also proves the above observations. It shows that only the samples pre-sintered at 1000 °C had no significant differences at p < 0.05 compared to the control before full sintering. However, the results of one-way ANOVA (Table 6) shows that after full sintering the control block manifested a significantly higher Knoop hardness value than all the different samples pre-sintered at different temperatures. In addition, the red solid curve in Figure 8 shows that their final hardness after full sintering fluctuates around 980 HK. Although it is lower than the hardness of the control, it is large enough to be suitable for dental uses.   Table 7) also proved that there was no significant difference at the level of p < 0.05 between the flexural strengths of fully sintered control zirconia block and the recycled blocks pre-sintered at 950, 1000, and 1050 °C.

Discussion
In this study, as the experiment results indicate, dental zirconia waste residual blocks were successfully recycled with satisfactory properties. In particular, pre-sintering of the dry-pressed blocks at temperatures in the range 950−1000 °C deemed necessary to be suitable for dental applications requiring a certain level of flexural strength and the quality in surface topography.
The choice of the dry pressing method in this work was driven by the following two aspects: (1) its high production rate, and (2) its small dimensional tolerances leading to a better tolerance control. Dry pressing is one of the easiest techniques for powder compacting [20]. The most significant disadvantage of this method is the development of non-uniform density distributions as described in the literature.
However, this phenomenon mainly appears when a big mold is used for pressing. Hence, a smaller mold may not be much affected in this process.
It is well known that as different additives or impurities may exist in the green body, melting, decomposing, and volatilization of those additives or impurities could result in the deformation or craze in the ceramics which should be avoided. In fact, various impurities have been effectively removed by an acid pretreatment that was not detected from the EDX result. It is worth noting that the 5.7−8.7 wt.% of carbon content detected from the EDX were from the carbon tape used to attach the zirconia sample on the SEM/EDX mount. In addition, Han et al. [26] has shown in a recent study that carbon can always be found on zirconia surfaces due to the adsorption of atmospheric CO 2 . Besides the temperature, heating rate and heating duration should have a major influence not only on the binder   removal process but also in determining the final quality of the ceramics . Ding [27] have reported that during the sintering of zirconia, lower heating rates resulted in higher flexural strengths and hardness of the system. This may be attributed to the fact that a higher heating rate could shorten the crystallization period, and as a result, pores in between the grains cannot be excreted. Stages of mean weight loss appeared while the sintering temperature was below 600 °C, followed by a severe physicochemical reaction and a change of state. Binders or other additives were also removed during this period [28]. To prevent the occurrence of volatilizing gases expanding too fast and affecting the mechanical properties of the ceramics, a slower heating rate of 5 °C / min was set from the room temperature to 600 °C.
With regards to the linear shrinkage rate, relative density, and hardness test results, fully sintered recycled zirconia blocks exhibited very similar results compared to the control zirconia block. In addition, the above three physical parameters were greatly affected by the powder formation processes as well as by the sintering conditions. A higher and more evenly distributed pressure during the dry pressing method could make the ceramics have a smaller linear shrinkage rate, higher relative density, and greater hardness value, while a slower heating and cooling rate including a longer heating duration can further improve those properties. However, these adjustments in the processing mean an increased recycling cost with longer production times. Nonetheless, as the improvements in properties may not that obvious after achieving a certain level, further investigation are certainly needed to find an appropriate balance between the material's performance and the production cost.
It should be noted that no standard evaluation method exists for measuring the machinable properties. Boccaccini introduced a brittleness index (BI) to evaluate ceramic machinability [29], where : Hardness is used to evaluate the ability to resist deformations of a material and fracture toughness is used to evaluate the ability to resist cracking in a material. The higher a material's brittleness index is, the poorer its machinability. Despite BI has been shown as a reliable evaluator with a marginal chipping factor in the cases of dental glass ceramics [30], determinations of HV and KIC rely on indentation mechanics and BI is not suitable for zirconia [31] because zirconia is toughened by transformations that would resist crack formations and growth. Therefore, median cracks may not form during the indentation process, or the cracks would be much smaller than those naturally occurred in zirconia, or the cracks can also be very small that polishing would eliminate or severely reduced its size. Indeed, the ISO standard for dental ceramics (ISO 6872) has indicated that the determination of KIC should not be relying on indentation hardness. Therefore, BI should not be utilized to determine the machinability of dental ceramics. Even though studies have shown machinability of a material can be determined by using removal rates during cutting, grinding force (P) measurement and cutting energy, where P ∝ [KIC 1/2 H 5/8 /(E/H) 4/5 ] 8/9 (E is elastic modulus) [32], those studies relied on indentation fracture mechanics and therefore it is useless in dental ceramics. In particular, chipping is one of the biggest problems in any dental zirconia restoration process [33]. Hence, a proper development to evaluate the machinability of dental CAD/CAM zirconia is indeed necessary.
The results of flexural strength tests should be stressed as it indicates that after the pre-sintering temperature reaches 1000 °C, the flexuralstrength was on a declining curve. Flexural strength is closely related to crack propagations, such that the initial crack of ceramic depends on the grain size because at initial stage the crack could only propagate into the nearer grains [34]. In addition, according the R-curve behavior in zirconia, the resistances of crack propagations would be greater for longer cracks [35]. For the hard-brittle materials like ceramics, intergranular fracture is the main fracture mode. Small and irregular grain shapes make the propagation path longer, needing more energy for propagations which in fact help to improve the fractural strength of ceramics. However, as the ceramics were first pre-sintered at a relative high temperature, secondary crystallization may lead to unusual grain to growth. The larger grain size makes the total length of grain boundaries smaller, and as a result the effect of R-curve behavior gets reduced. In addition, from the linear shrinkage rate and the relative density results it was seen that the linear shrinkage rates and relative densities of zirconia samples pre-sintered below 1000 °C did not change much compared to the green bodies, which indicate that the grains did not start diffusing fully below 1000 °C, hence the abnormal grain growth appears only for the samples pre-sintered at temperatures > 1000 °C, which explains the drop in flexural strengths for the respective samples.
In summary, this paper describes a recycling method for dental zirconia ceramic wastes from the used blocks, especially suitable for the use in dental laboratories and for zirconia manufacturers. The advantages of this recycling method could be its short production cycle, high efficiency and cost effectiveness. In addition, it entails a simple manufacturing procedure and the recycled blocks exhibit very similar physical and chemical properties compared to the control sample. However, further improvements are required to produce larger and thicker recycled samples. Besides, the development of a universal sintering scheme suitable for different kinds of dental zirconia ceramics. Moreover, a standard quantitative measurement method should be developed for the cut surfaces. Further studies are also needed to examine aging behavior, bulk material's content of flaws, cracks and clinical trials of the restorations made from the recycled products.

Conclusion
We have developed a relatively simple, easy and clean method to recycle dental zirconia. It can be concluded that recycling dental zirconia CAD/ CAM waste residuals from the used blocks is feasible with optimized presintering temperatures in the range from 950 to 1000 °C.   Table 4. One-way ANOVA analysis for fully sintered relative density (percentile data were arcsine-transformed before ANOVA analysis). -.1230996278 *. The mean difference is significant at the 0.05 level. Table 2. One-way ANOVA analysis for fully sintered linear shrinkage rate (percentile data were arcsine-transformed before ANOVA analysis).  Table 7. One-way ANOVA analysis for fully sintered flexural strength.