Effect of surface topography and wettability on shear bond strength of Y-TZP ceramic

Zirconia ceramics have been widely used as dental restorations due to their esthetic appearance and high flexural strength. The bonding of zirconia with resin cement should rely on both mechanical and chemical bonds. This study was performed to investigate the effect of zirconia surface topography and its wettability after surface pretreatments on the microshear bond strength (μSBS) of a resin cement. Zirconia slabs were prepared and randomly divided into 5 groups based on the surface treatment as follows: no treatment (control), air abrasion (AB), etching with hydrofluoric acid (F), the mixture of hydrofluoric acid and nitric acid (FN), or the mixture of hydrochloric acid and nitric acid (CN) for 10 min. The specimens were subjected to investigation of surface roughness characteristics [average roughness (Ra), peak-to-valley average distance (Rpv), skewness (Rsk), and kurtosis (Rku)] using atomic force microscopy (AFM) and measurements of surface contact angle (θc) and μSBS of a resin cement. In addition, the area % of the nanoscale surface irregularity (nSI%) was calculated from the AFM images. The effects of nSI%, Ra and θc on the μSBS were analyzed by multiple linear regression analysis (p < 0.05). Multiple regression analysis revealed that the nSI% was the most predominant factor for the μSBS (p < 0.001). A surface with larger nSI%, higher Ra and relatively lower θc was essential for establishing a reliable resin-zirconia bond.


Surface topography assessment using atomic force microscopy (AFM)
Surface roughness characteristics measurement All the specimens were investigated for surface roughness characteristics using AFM (PARK XE7, Park Systems, Gyeonggi, South Korea) in noncontact mode.The parameters Ra, Rpv, Rsk, and Rku were calculated using analysis software (XEI 4.3, Park Systems, Gyeonggi, South Korea).The Ra, Rpv, Rsk, and Rku data were analyzed statistically with one-way ANOVA and Tukey's test (p < 0.05).
Rsk is a measure of the asymmetry of the profile about the mean line calculated over the evaluation length.Profiles with predominant deep valleys have negative skewness, whereas those with several high peaks show positive values.Rku is a measure of the peakedness of the profile about the mean line calculated over the evaluation length.It describes the sharpness of the probability density of the profile, of which Rku > 3 indicates sharp peaks and valleys, and Rku < 3 refers to relatively flat peaks and valleys.If Rku = 3, then the curve is Gaussian (Fig. 1) 14 .

nSI assessment
To observe nanoscale surface topography, a 5 × 5 µm-AFM image with 256 × 256 pixels was taken with a scan rate of 0.5 Hz to assess nSI on the treated zirconia surface.The nSI area was quantitatively defined as a percentage of the whole surface (nSI%) by manual counting using a simple hundred-square grid from a 2D image.If a clearly smooth grain surface presented more than 50% of the square grid, it was counted as a smooth grain surface grid (Fig. 2a).Meanwhile, if the rough or irregular grain surface presented more than 50% of the square grid, it was counted as an nSI grid (Fig. 2b).
Twelve images from each group were measured by 2 observers, 2 times with a 1-week interval.A validation of the method used in this study was performed by both intraobserver and interobserver reliability analysis using the intraclass correlation coefficient (ICC) test, whose correlation ratios were 0.988 and 0.972, respectively (p < 0.05).The nSI% was analyzed statistically with one-way ANOVA and Tukey's test (p < 0.05).

Surface contact angle ( θ c ) assessment
All the treated specimens were used for contact angle measurement by the sessile drop technique.One microliter of deionized water was dropped on the specimen surface and left for 1 min prior to the measurement using a goniometer (CMU-PHYS™ Goniometer, CMU-PHYSnanolabs, Chiang Mai, Thailand).The θ c data were analyzed statistically with one-way ANOVA and Tukey's test (p < 0.05).www.nature.com/scientificreports/

μSBS test
After the investigations, all the specimen surfaces were applied with a 10-MDP containing primer (Clearfil™ Ceramic Primer Plus, Kuraray Noritake Inc., Okayama, Japan).After placing polyethelene tubes (Tygon ® , Norton Performance Plastic, OH, USA) with an internal diameter of 0.8 mm and a height of 0.5 mm on the zirconia surface, resin cement (Panavia™ V5, Kuraray Noritake Inc., Okayama, Japan) was subsequently injected into the tube and light cured for 20 s from the top surface using a light-curing unit (Bluephase ® N, Ivoclar Vivadent, Schaan, Liechtenstein).All specimens were stored in 37 °C distilled water for 24 h.After carefully removing the tubes, the specimens were subjected to μSBS testing using a universal testing machine (Instron® 5566, Illinois Tool Works, MA, USA) using the wire-loop technique at a crosshead speed of 0.5 mm/min.The materials used are listed in Table 1.The μSBS data were analyzed statistically with one-way ANOVA and Tukey's test (p < 0.05).

Failure mode analysis
After the μSBS test, failure mode analysis was carried out using a stereoscopic microscope (Olympus™ SZX7, Olympus, Tokyo, Japan) at 56 × magnification.Failure modes were categorized into 3 types as follows 15 : adhesive failure (80-100% of the failure occurred at the interface between resin cement and zirconia), cohesive failure in resin cement or zirconia (80-100% of the failure occurred within resin cement or zirconia), and mixed failure (mixed with adhesive and cohesive failure patterns in the same specimen).The failure mode data were analyzed statistically with the chi-square test (p < 0.05).

Multiple linear regression analysis
To investigate the effect of nSI%, Ra and θ c on μSBS, multiple regression analysis was performed by using IBM SPSS Statistics version 27.0 (IBM, Chicago, IL, USA) (p < 0.05).Model fit was assessed using the coefficient of determination (R squared).

Surface topography assessment using AFM
The means of Ra, Rpv, Rsk, Rku and their standard deviations are presented in Table 2.All the surface treatments exhibited significant increases in Ra and Rpv values compared with the control group (p < 0.05).The AB and F groups exhibited the highest Ra and Rpv values among the treatments.The Rsk values were negative in the control and CN groups, whereas they were near zero in the AB group and positive in the F and FN groups.The Rku value of the control group was more than 3.All the surface treatments caused an insignificant change in Rku compared with the control group (p > 0.05).The only significant difference in Rku was found between the CN group (highest Rku, more than 3) and the F group (lowest Rku, less than 3) (p < 0.05).
Representative AFM images exhibiting the surface topography of nontreated and surface-treated zirconia are shown in Fig. 3. Smooth surface grains and grain boundaries were clearly observed in the control group.On the other hand, the F and FN groups clearly exhibited nSI on grains with unclear grain boundaries.In the AB group, nSI with unclear grain boundaries was observed in the lesser amount.For the CN group, small protrusions were observed on the smooth surface grains with traceable boundaries.
The means and standard deviations of nSI% are shown in Table 2.All the treatments significantly increased nSI% compared with the control group (p < 0.05), and the F and FN groups showed significantly higher nSI% than the AB and CN groups (p < 0.05).

θ c assessment
The mean θ c and standard deviations are shown in Table 2.All the surface treatments significantly decreased θ c compared with the control group (p < 0.05), in which the FN group exhibited significantly lower θ c among the treatment groups (p < 0.05).

μSBS test
The means and standard deviations of μSBS in each group are shown in Table 2.All the treatments significantly increased μSBS compared with the control group (p < 0.05), and the F group exhibited significantly higher μSBS among the treatment groups (p < 0.05).

Failure mode analysis
The failure mode distribution is displayed in Fig. 4. The results of the chi-square test revealed that there were significant differences in the failure mode distribution among the groups (p < 0.05).The majority of the failure mode of the control was adhesive failure, whereas it was mixed failure in the AB, F, FN, and CN groups.

Discussion
The results of this study revealed that the air-abrasive and acid-etching treatments used in this study caused variations in the surface topography (Ra, Rpv, Rsk, Rku and nSI%) of the zirconia surface and affected the θ c and bonding performance of the resin cement.Additionally, nSI%, Ra and θ c significantly affected μSBS between resin and zirconia.Thus, the null hypothesis was rejected.www.nature.com/scientificreports/AFM was employed in this study since it is an essential tool for characterizing surface topography, particularly on the nanoscale, providing a direct view of surface features at high resolution without surface coating being required 16 .
With the aim of creating zirconia surfaces with various micro-and nanoscale irregularities, the treatments of airborne particle abrasion or acid etching with HF and its mixture with HNO 3 , including a mixture of HCl with HNO 3, were used.Air-pressurized abrasion can roughen the surface with large surface defects and mechanically damage the zirconia grains in the top 1-to 1.5-µm zone 17 .HF etching using a high concentration can corrode both intergranular and intragranular parts of the zirconia surface, being slightly faster at the grain boundaries 18 , leading to a size reduction and dislodgement of zirconia grains 12 .These two methods increased Ra and Rpv to the highest values, whereas the mixture of HF with HNO 3 (FN group) and the mixture of HCl with HNO 3 (CN group) caused the lesser increase in Ra and Rpv.The results of this study were consistent with previous studies, which reported that the mixture of HF with HNO 3 was capable of roughening a zirconia surface 2,19,20 but to a lesser extent than HF etching alone 21 .It was speculated that adding HNO 3 reduced the penetration of the mixture into grain boundaries, no longer dislodging the deeper-part grains.Last, the mixture of HCl with HNO 3 barely etched the zirconia substrate, although its strong oxidative effect 22 and its ability to readily dissolve common metallic oxides and hydroxides 23 have been addressed.
In adhesive dentistry, adequate surface wetting of the adhesive on the substrate is the primary requirement to achieve good adhesion 24 .Good surface wetting, determined by a low contact angle, leads to intimate contact between adhesive and adherend and enhances mechanical interlocking 25,26 .Thus, surface wettability could be one of the parameters used to predict bonding efficacy.It has been documented that the contact angle is the parameter manifested from the substrate's surface roughness and surface chemistry 27 .To eliminate the effect of chemical heterogeneity on the treated surfaces, each specimen was stored in a closed container for 21 days before contact angle measurement.This aging condition possibly exerted airborne hydrocarbon adsorption, which subsequently increased the contact angle 27,28 .Although this method overestimated the values of measured θ c , it aided the surface chemistry of all treated groups to be homogenized, and hence, the wettability derived from the effect of surface topography could be predominant 29 .
The Wenzel theory states that an increase in surface roughness could decrease the contact angle for a hydrophilic surface 30,31 .However, several studies demonstrated that the contact angle decreases following increasing roughness to a certain extent, where no greater improvement in wetting is expected for highly rough surfaces 32,33 .It should be noted that the wettability was determined not only by the average surface roughness but also by the surface geometry in three dimensions, such as the distance between peaks compared to their height, of which  www.nature.com/scientificreports/ the proper distance and height could facilitate the liquid to better wet the surface by forming a capillary action 32 .
Excessively high peaks and symmetrically distributed high peaks and deep valleys (near-zero Rsk) conversely decrease wettability (increase θ c ) due to trapped air in surface cavities 33,34 , whereas flat-shaped peaks and valleys (Rku < 3) are more favourable for enhancing surface wettability 35 .From the results of this study, the FN group exhibited the lowest θ c , which could be attributed to an optimal height and a predominance of peaks (positive Rsk) having a symmetric Gaussian shape (Rku = 3).On the other hand, both the AB and F groups created higher rough surfaces, which might be too high, with a more symmetric topography (near-zero Rsk).Thus, the wettability of the AB and F groups was hampered compared to that of the FN group.When compared to each other, the shape of peaks and valleys was relatively sharp in the AB group (Rku > 3) and flat in the F group (Rku < 3), leading to the better wettability of the F group than that of the AB group.Last, the θ c values of the control and CN groups were the highest among the studied groups due to their low surface roughness and predominance of sharp valleys (negative Rsk and Rku > 3).Presumably, the moderately high roughness, with flat peaks (greater skewness and small kurtosis), might be preferable to enhance liquid wettability on the zirconia surface.Interestingly, nSI was obviously observed on the grain surfaces of treated zirconia, especially in the F and FN groups.Since this nSI could not be quantified by the roughness parameters described above, the nSI% was evaluated manually by counting under the hundred-square grid method.It was speculated that a zirconia surface with a higher nSI% could increase the surface area for adhesion; thus, a higher bond strength could be expected.Accordingly, the F and FN groups, including the AB group, which had a high nSI%, demonstrated a higher µSBS than the groups with a lower nSI%.
Regarding the µSBS results, the highest µSBS was found in the F group, although its wettability was not the highest.The AB group had comparable µSBS to the FN group, despite its lower wettability and lower nSI% than that of FN.These results indicated that not only the wettability, defined by θ c , and the nSI%, but also the surface roughness (Ra), which was highest in the F and AB groups, could cooperatively contribute to improving the bond strength of resin cement to zirconia.This speculation was confirmed by the multiple regression analysis, revealing that nSI%, Ra and θ c significantly affected the µSBS (p < 0.001, p < 0.001 and p = 0.035, respectively), in which nSI% was the most predominant factor contributing to the bonding to zirconia (standardized coefficient beta = 0.499), and Ra was a more significant predictor than θ c (standardized coefficient beta = 0.367 and − 0.174, respectively) (Table 3).Therefore, the surface roughness would be an essential parameter for bonding with mechanical interlocking by resin tags to the zirconia surface, and the surface wettability would assist their optimal formation.In this analysis, the variance inflation factor (VIF) value of each predictor was less than 5 (Table 3), which indicates that the multicollinearity between the variables of nSI%, Ra and θ c was small.Therefore, the nSI% would not be strongly associated with the wettability on the surface.
It is noteworthy that all the surface treatments could increase bond strengths compared with the control group.These findings are in agreement with other studies 21,36,37 , indicating that roughness is indispensable for the mechanical bond of resin-zirconia.Unlike the results from this study, a higher bond strength of air abrasion over HF treatment has been reported 21,38 .Additionally, some studies have demonstrated that the use of HF/ HNO 3 could increase bond strength to a larger degree than air abrasion 19,37 .These inconsistent results could be attributed to the differences in acid concentration, etching time, or resin cement types 19,39,40 .In particular, the 3-week waiting period prior to cementation in this study might affect the chemical bonds of each treatment to different degrees.
Within the methodology in the current study, there are some clinical limitations to be considered.First, it would be impractical to treat the intaglio surface of a crown and wait 3 weeks before proceeding with cementation.Second, the hazards of HF are well recognized 41 .Working with highly concentrated HF requires an isolated workspace with adequate ventilation, which is unfeasible in general clinical practice.Finally, the tetragonal-tomonoclinic (t-m) phase transformation, which is a critical consideration for the surface treatment of Y-TZP ceramics, was not evaluated in this study.Previous studies demonstrated that gentle air abrasion protocols, as used in this study, generally trigger t-m phase transformation [42][43][44] , whereas HF etching with conditions close to the current study was found to induce phase transformation to the lesser degree than air abrasion 45 .Nevertheless, such a small amount of monoclinic phase presented on the zirconia surfaces prior to the bonding procedure was found to not influence the resin bond strength [45][46][47] .
After aging, the increase in monoclinic contents 48 and the decrease in bond strengths of resin-zirconia 45,49,50 were widely reported.However, the study revealing the negative correlation between monoclinic contents and bond strengths after aging is scarce.On the other hand, some previous studies showed increased monoclinic contents without a significant drop in resin bonding 51,52 .It was concluded that the effect of resin hydrolysis is responsible for the decreased bond strength more than the phase transformation 21,52 .Moreover, it is still controversial whether the development of residual stresses at the interface following phase transformation may negatively affect surface adhesion 47,50 .Therefore, further study investigating the effect of t-m phase transformation on both immediate and long-term bond strengths is needed.
It could be concluded based on this in vitro study that appropriate surface topography and surface wettability are essential for establishing reliable resin-zirconia bond strength.Of all the factors evaluated, nSI% was the most predominant factor, whereas Ra and θ c could concomitantly affect the bond strengths of resin cement to zirconia.The surface roughness properties Rsk and Rku are important parameters for facilitating surface wettability.

Figure 1 .
Figure 1.Schematic illustrations demonstrate the meaning of parameters Rsk (a) and Rku (b).

Figure 2 .
Figure 2. The representative 5 × 5 µm zirconia surface in a 2D image, showing the counting from a hundredsquare grid.Blue squares (a) show smooth grain surface grids, and red squares (b) show nanoscale surface irregularity (nSI) grids.

Figure 3 .
Figure 3. Representative AFM images (upper) and line-scan profiles (lower) of the control (a), AB (b), F (c), FN (d), and CN (e) groups.Grain boundaries (white arrow) were clearly observed in the control group, whereas surface irregularities on grains (black arrows) were observed in all treated groups.

Figure 4 .
Figure 4. Frequency of failure modes after µSBS testing in each group (n = 12).The numbers in each bar are the numbers of specimens in each failure mode.There were significant differences in the failure mode distribution among the groups (p < 0.05).The majority of the failure mode of the control was adhesive failure, whereas it was mixed failure in the AB, F, FN, and CN groups.