Surface engineering of zirconia to improve indirect composite bonding: an in vitro study

Abstract

Background

 Porcelain-layered zirconia restorations are susceptible to chipping and delamination. This study evaluated the effect of different surface treatments on the shear bond strength (SBS) of indirect composite resin to zirconia.

Methods

 Fifty zirconia specimens were randomly assigned to five groups (n = 10): Control. 50-µm aluminum oxide (Al₂O₃) sandblasting (SB50). 110-µm Al₂O₃ sandblasting (SB110). Glaze + hydrofluoric etching (Gl + HF). SB50 followed by glaze + hydrofluoric etching (SB50 + Gl + HF). After composite bonding and thermocycling (5,000 cycles), SBS was tested and failure modes were assessed. Data were analyzed with one-way ANOVA and Bonferroni post-hoc tests.

Results

 SBS differed significantly among groups (p < 0.001). SB110 (11.134 ± 0.866 MPa) and SB50 + Gl + HF (12.229 ± 1.101 MPa) showed statistically equivalent bond strength (p = 0.133, effect size -1.1) and were significantly higher than other groups (p < 0.05, effect size 0.913). The Gl + HF group had the lowest strength (3.771 ± 0.905 MPa, p < 0.05, effect size 0.913).

Conclusions

 Indirect composite could be a viable alternative to porcelain for veneering zirconia restorations in the short term. Sandblasting with 110 µm alumina particles provided high bond strength equivalent to the more complex SB50 + Gl + HF protocol, offering a clinically simpler and safer surface treatment option.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12903-025-07309-z.

Background

The advent of digital dentistry has established zirconia as a crucial material for metal-free fixed dental prostheses due to its exceptional durability, aesthetics, and biocompatibility [1–3]. Layering porcelain veneers onto the zirconia core is essential as porcelain remains the gold standard in aesthestic scenarios, however, the porcelain veneers are susceptible to chipping (a cohesive failure of porcelain) and delamination (a failure in the adhesive interface between the porcelain and the substructure), failures often attributed to thermal coefficient mismatches, residual firing stresses, and porcelain’s inherent brittleness [4–11]. As a solution, highly loaded indirect composite resins have emerged as a promising alternative veneering material for specific clinical scenarios. Beyond practical advantages like repairability and decreased wear on antagonist teeth [12], modern indirect composites offer significant biomechanical benefits. Unlike brittle ceramics, polymer-based indirect composites exhibit a lower elastic modulus and higher ductility with better impact resistance, allowing them to absorb and distribute occlusal stresses more uniformly. This stress-absorbing capability can reduce the risk of catastrophic failure at the restoration interface, making them a suitable choice for situations demanding high resilience under masticatory loads such as in bruxism patients [13, 14].

A primary challenge in veneering indirect composite is achieving a durable bond to zirconia, a task complicated by the material’s polycrystalline structure, chemical inertness, and lack of a silica phase, which prevents conventional etching [15]. To address this, two main strategies have been developed: Firstly, air abrasion and primers; micromechanical and chemical modification of the zirconia itself. Air abrasion with aluminum oxide particles effectively creates surface roughness for retention but risks surface damage and phase transformations [16–19]. This is often combined with the application of phosphate monomer-containing primers especially 10-methacryloyloxydecyl dihydrogen phosphate (10-MDP), which chemically bond to zirconia oxides [20–22]. The second strategy employs application of a silica-based glaze layer. This technique deposits a silica-rich surface on the zirconia, which can then be etched with hydrofluoric acid (HF) and silanized, facilitating a micromechanically retentive and chemically bondable interface similar to that of glass–ceramics [23–25].

While these individual techniques are well documented, the existing literature lacks a comprehensive, direct comparison of a wide range of surface treatment protocols under standardized, rigorous conditions. For instance, there is a notable absence of direct comparative evidence on the efficacy of key techniques such as air abrasion with 110 µm aluminum oxide particles versus the glazing and hydrofluoric acid etching protocol after thermocycling [23, 26]. Furthermore, the potential synergistic effect of combining techniques (e.g. sandblasting followed by glazing) is an area requiring exploration so, we hypothesized that the glazing with HF conditioning would enhance bond strength compared to sandblasting alone. Conversely, the null hypothesis was that there would be no significant difference in bond strength among the various surface treatment groups.

Therefore, this study aims to fill this gap by providing a direct comparative analysis of five distinct surface treatment strategies on the shear bond strength of indirect composite to monolithic zirconia after thermocycling to assess long-term durability. The findings will provide a valuable clinical reference for selecting the most effective and reliable bonding protocol for indirect composite-zirconia interfaces.

Methods

Sample size determination

For computing sample size, a priori power analysis was conducted using G* Statistical Power Analysis v3.1.9.7 (University of Düsseldorf, Germany) based on the large effect size (Cohen’s f = 0.52) determined in a pilot study prior to the investigation. For one-way ANOVA with five groups, an alpha level of 0.05 and a desired power of 0.80, the analysis indicated a required total sample size of 50 (n = 10 per group).

Sample preparation

The manufacturer details of the materials used in this study are presented in Table 1. Fifty pre-sintered cylindrical monolithic zirconia specimens (10 mm diameter; 2 mm thickness) (IPS e.max ZirCAD; Ivoclar Vivadent AG, Schaan, Liechtenstein) were milled using a 5-axis CAD/CAM dental milling machine (Roland DWX-52D; Roland DG Corp., Hamamatsu, Japan) [26, 27]. After milling, the zirconia disks underwent full sintering in a furnace (Programat S2; Ivoclar Vivadent AG, Schaan, Liechtenstein) according to manufacturer’s directions and then were allowed to bench-cool to room temperature.

Table 1.

Manufacturer details of materials used in this study

Material	Category	Manufacturer	LOT No.
IPS e.max ZirCAD LT	Zirconia	Ivoclar Vivadent AG, Schaan, Liechtenstein	Z06KCB
Aluminum oxide 50 µm	Sandblasting powder	Renfert, GmbH company, Hilzingen, Germany	15941205
Aluminum oxide 110 µm	Sandblasting powder	Renfert, GmbH company, Hilzingen, Germany	15831005
IPS Ivocolor glaze powder/fluo	Glaze powder	Ivoclar Vivadent AG, Schaan, Liechtenstein	Z074PT
IPS Ivocolor mixing liquid longlife	Glaze liquid	Ivoclar Vivadent AG, Schaan, Liechtenstein	Z07JHL
Z-Prime Plus	Primer	Bisco Inc., Schaumburg, IL, USA	2300112524
Porcelain etchant 9.5%	HF acid	Bisco Inc., Schaumburg, IL, USA	24000014492
Bis-Silane	Silane coupling agent	Bisco Inc., Schaumburg, IL, USA	2400001496
Porcelain Bonding Resin	HEMA free, hydrophobic bonding resin	Bisco Inc., Schaumburg, IL, USA	2400002884
Ceramage	Indirect composite	Shofu Inc., Kyoto, Japan	092390

Polishing the samples

The samples were polished in a custom-made split mold using a two-phase diamond infused rubber polisher system (Sagemax NexxZr Shine Polishing Kit; Sagemax Bioceramics, Inc, Washington, USA). Each polishing phase was performed for 60 s in one direction to the full surface (total of 2 min per sample) at a speed of 10,000 rpm as per manufacturer’s directions as shown in Figs. 1 and 2 [28, 29]. Following polishing, the specimens were cleaned ultrasonically in distilled water for 10 min in an ultrasonic bath (EasyClean MD Ultrasonic Cleaner; RenfertGmbh & Co., Hilzingen, Germany) and allowed to air-dry [26].

Fig. 1.

Standardization of polishing protocol using a modified dental surveyor

Fig. 2.

Close-up view of the polishing bur and the sample before starting polishing

Randomization and blinding

To minimize selection bias, the zirconia specimens were randomly allocated to one of the five surface treatment groups (n = 10 per group) using a computer-generated random number sequence (RAND function in Microsoft Excel). Allocation concealment was ensured using sequentially numbered, opaque sealed envelopes that contained the group assignment for each specimen. The operator performing the SBS testing was blinded to the group identity of each specimen throughout the measurement process to prevent assessment bias. However, the operator performing the surface treatments could not be blinded due to the visibly distinct nature of each protocol.

Sample grouping

The five surface treatment groups were:

• Control group: No surface treatment.

• SB50: The specimens were sandblasted using 50 µm aluminum oxide (Al₂O₃) particles (Renfert-Cobra aluminum oxide powder; Renfert, GmbH company, Hilzingen, Germany) at a pressure of 0.2 MPa which was calibrated using a factory-calibrated gauge, through a 5-mm diameter nozzle (Basic Eco Sandblaster; Renfert, GmbH company, Hilzingen, Germany), from a distance of 10 mm applied perpendicularly (90º) to the surface for 10 s [26]. After that, the samples were ultrasonically cleaned in deionized water for 5 min [30].

• SB110: The sandblasting was carried out in a similar manner to SB50 group only the particle size was changed to 110 µm [30, 31], followed by ultrasonic cleaning in the same way [30].

• Gl + HF: A small amount of feldspathic ceramic glazing powder (IPS Ivocolor glaze powder/fluo; Ivoclar Vivadent AG, Schaan, Liechtenstein) was mixed with the stain liquid (IPS Ivocolor mixing liquid longlife; Ivoclar Vivadent AG, Schaan, Liechtenstein) to the desired manufacturer’s recommended consistency and brushed as a single thin coating on the surface of the samples. Then, the samples were fired in a ceramic firing furnace (Programat EP3010/G2; Ivoclar Vivadent AG, Schaan, Liechtenstein) according to manufacturer’s directions for a total time of 13 min and 50 s (drying time of 4 min and holding time of 1 min and 30 s at a temperature of 850ºC). Finally, they were allowed to bench-cool at room temperature [26].

• SB50 + Gl + HF: The specimens were sandblasted similar to SB50 group, then ultrasonically cleaned for 10 min and allowed to air-dry. After that, glaze layer application was performed as described in Gl + HF group [26].

Fabrication of acrylic blocks

Following surface treatments, the samples were embedded in cylindrical acrylic blocks by using a custom-made cylindrical silicon mold which had a circular depression (10 mm diameter; 0.2 mm depth) in its center into which the sample was placed [27] (see Additional files 1 and 2). A finished acrylic block holding the sample is shown in Additional file 3.

Adhesive protocol for the groups

The adhesive protocol for the groups is clarified in Fig. 3.

Fig. 3.

Flowchart showing the adhesive protocol for the groups

Adhesive protocol for control, SB50 and SB110 groups

The surface of the zirconia disks was treated with one coat of Z-Prime Plus (Bisco Inc., Schaumburg, IL, USA), applied for 20 s, allowed to dwell for 30 s and then gently air dried for 5 s [26].

Adhesive protocol for Gl + HF and SB50 + Gl + HF groups

The surface of the zirconia disks was etched with 9.5% HF acid (Porcelain etchant 9.5% HF; Bisco Inc., Schaumburg, IL, USA) for 90 s and rinsed with water for 30 s then dried for 10 s. Then, the samples were treated with a two-component silane coupling agent (Bis-Silane; Bisco Inc., Schaumburg, IL, USA) which was mixed and brushed on the samples’ surfaces for 30 s, allowed to dwell for 30 s and then gently air dried for 5 s. After that, Porcelain Bonding Resin (Bisco Inc., Schaumburg, IL, USA) was applied for 20 s, gently air thinned and air dried for 5 s and then light cured for 10 s using a light curing pen (Eighteeth, Changzhou, Jiangsu, China) from a distance of 2 mm using a custom-made ring spacer and at a light curing intensity of 1300 mW/cm² [26].

Layering indirect composite

A specialized metallic split mold was fabricated for layering of indirect composite to the zirconia samples. The inferior portion of the mold had a cylindrical hole which will fit firmly over the acrylic blocks when the metal mold is locked (Additional file 4). The superior portion of the mold had a small cylindrical opening in its center measuring 5 mm in diameter and 2 mm in depth for application of the indirect composite (Additional file 5).

An indirect composite material (Ceramage; Shofu Inc., Kyoto, Japan) was applied systematically in a uniform and reproducible manner in a single increment to the center of the samples in all groups [27]. Then, it underwent preliminary curing using the light curing pen through a 0.05 mm polyester matrix strip (Polydentia SA; Mezzovico-Vira, Switzerland) for 60 s according to manufacturer’s directions at an intensity of 1300 mW/cm², as shown in Fig. 4 [32, 33]. Finally, the samples underwent final light polymerization in the laboratory light curing unit (Solidilite V; Shofu Inc., Kyoto, Japan) for 180 s as recommended by the manufacturer (Fig. 5).

Fig. 4.

Preliminary light curing of indirect composite through a polyester matrix strip

Fig. 5.

A layered sample after finishing final light polymerization

The nominal diameter of the bonded area was 5 mm. This was confirmed by measuring the diameter of a representative subset of specimens with a digital caliper (Mitutoyo Corp., Tokyo, Japan); the mean measured diameter was 5.02 ± 0.05 mm, validating the use of the nominal area for stress calculation.

Thermocycling

The samples were subjected to thermocycling in water at temperatures between 5° and 55 °C for 5,000 cycles using an automated thermocycling apparatus with 30 s dwell time and 5 s transfer time [2, 30, 34].

Shear bond strength (SBS) testing

The SBS test was performed by using a universal testing machine (Terco MT 3037 universal testing machine; TERCO Corp., Stockholm, Sweden). Each sample was mounted securely in a custom-made self-aligning jig that positioned the bonded interface parallel to the direction of the applied force in order to ensure accurate force application which was applied using a knife-edge shearing rod at a crosshead speed of 0.5mm/min until failure occurred. The movable knife-edge of the testing machine was aligned to contact the composite cylinder as close as possible to the zirconia-composite interface to minimize peel stresses [26, 30].

The SBS (MPa) was calculated by dividing the maximum force (load) at the time of fracture (N) by the adhesive area (mm²) as follows:

Surface area = πr² where, r = radius (2.5 mm) and π = 3.14

Failure mode analysis

The debonded surfaces of all the zirconia samples underwent a stereomicroscopic examination (Motic ST-39 stereomicroscope; Motic, Kowloon, Hong Kong) at 20X and 40X magnification to determine the mode of bond failure by two independent examiners and the inter-rater reliability was assessed using Cohen’s kappa statistic (κ). Failure modes were categorized as: adhesive (at the interface between monolithic zirconia and indirect composite material), cohesive (inside the monolithic zirconia or indirect composite material), or mixed [27, 34, 35].

Statistical analysis

A number of statistical tests were used in order to guarantee the reliability and validity of the results. The normality of the data distribution for each group was first evaluated using Shapiro–Wilk test, and the homogeneity of variances among the groups was confirmed using Levene’s test. To find out if there were any notable variations in SBS values among the groups, a one-way Analysis of Variance (ANOVA) was conducted under these presumptions. When an ANOVA revealed significant differences, pairwise comparisons between the groups were made using a Bonferroni post-hoc test. All results, tables and graphs were generated using SPSS software version 26 (IBM, Chicago, USA), R statistical programming language 4.4.2 and GraphPad 10. All of the statistical analyses were performed at a statistical significance level of 5% as p-value ≤ 0.05 was regarded as having a significant difference.

Results

Assessment of the shear bond strength (SBS)

Data normality and homogeneity assessment

The Shapiro–Wilk test revealed that the SBS data of all groups were normally distributed (p > 0.05). The non-significant result (p = 0.966) from Levene’s test suggests that the group variances were statistically homogeneous (Table 2). These results confirm that parametric tests, like one-way ANOVA, are adequate for comparing SBS values among the experimental groups.

Table 2.

Shapiro–Wilk test of normality and Levene’s test of homogeneity of variances for the experimental groups

Groups (surface treatments)	Shapiro–Wilk test of normality		Homogeneity of variances
	Statistic	p-value	Levene statistic	p-value
Control	0.925	0.398	0.140	0.966
SB50	0.813	0.051
SB110	0.916	0.327
Gl + HF	0.906	0.256
SB50 + Gl + HF	0.910	0.280

Descriptive statistics and comparative analysis of SBS

The impact of different surface treatment techniques on bond strength was highlighted by the mean SBS values, which showed clear performance disparities among the groups as shown in Table 3. Results showed that SB50 + Gl + HF group had the greatest mean SBS value (12.229 ± 1.101 MPa). This was followed by SB110 (11.134 ± 0.866 MPa). On the other hand, Gl + HF group had the lowest mean SBS value (3.771 ± 0.905 MPa) indicating weaker bond strength. Raw SBS data for all groups are presented in Additional file 6.

Table 3.

Mean, standard deviation, minimum and maximum values with confidence intervals of experimental groups

Groups	n	Mean ± SD	Min.	Max.	95% CI
					Lower	Upper
Control	10	9.427 ± 0.874	8.15	10.70	8.801	10.052
SB50	10	9.783 ± 0.985	8.66	12.23	9.079	10.488
SB110	10	11.134 ± 0.866	10.19	12.74	10.514	11.754
Gl + HF	10	3.771 ± 0.905	2.55	5.10	3.123	4.418
SB50 + Gl + HF	10	12.229 ± 1.101	10.70	13.76	11.442	13.017

ANOVA was used to assess the statistical significance of the variations in SBS values among the five experimental groups. ANOVA results revealed a statistically significant difference among the groups (F = 118.370, p < 0.001) (Table 4). The large effect size (ω² = 0.90, Cohen’s f = 3.24) indicates that the surface treatment technique, not chance, accounted for around 90% of variations in bond strength.

Table 4.

Results of one-way ANOVA among different surface treatment groups

Source	SS	df	MS	F	p-value	η2	ω2	Cohen’s f
Between Groups	427.615	4	106.9	118.370	0.000*	0.913	0.904	3.24
Within Groups	40.641	45	0.903
Total	468.256	49

^*Statistically significant difference at p-value ≤ 0.05

Analysis of SBS revealed that the SB50 + Gl + HF and SB110 groups produced the strongest and most reliable bonds, a finding clearly supported by their high, clustered values in the boxplot (Fig. 6A) and their narrow, right-shifted peaks in the density plot (Fig. 6B). In contrast, the Gl + HF group performed the poorest, while the control and SB50 groups showed statistically comparable, intermediate strength.

Fig. 6.

(A) Boxplot showing SBS values across surface treatment groups; (B) Density plot illustrating the spread of SBS values in each group

A Bonferroni post-hoc multiple comparison was performed and computed in Table 5 in order to investigate the significant differences found by the ANOVA test in more detail. Figure 7 also shows the pairwise results between the groups where the asterisk refers to having statistically significant differences between the pairs.

Table 5.

Bonferroni post-hoc test for pairwise comparison of SBS values between surface treatment groups

(I) groups	(J) groups	Mean diff. (I-J)	SE	Adjusted p-value	95% CI		Hedge’s g
					Lower bound	Upper bound
Control	SB50	−0.357	0.425	1.000	−1.611	0.898	−0.36
	SB110	−1.707	0.425	0.002*	−2.962	−0.452	−1.76
	Gl + HF	5.656	0.425	0.000*	4.401	6.911	5.68
	SB50 + Gl + HF	−2.803	0.425	0.000*	−4.057	−1.548	−2.8
SB50	SB110	−1.350	0.425	0.027*	−2.605	−0.096	−1.37
	Gl + HF	6.013	0.425	0.000*	4.758	7.267	6.11
	SB50 + Gl + HF	−2.446	0.425	0.000*	−3.700	−1.191	−2.47
SB110	Gl + HF	7.363	0.425	0.000*	6.108	8.618	7.39
	SB50 + Gl + HF	−1.096	0.425	0.133	−2.350	0.159	−1.1
Gl + HF	SB50 + Gl + HF	−8.459	0.425	0.000*	−9.713	−7.204	−8.5

^*Statistically significant difference at p-value ≤ 0.05

Fig. 7.

Error-bar chart for pairwise comparison illustration between groups

Post-hoc analysis revealed that SB50 + Gl + HF and SB110 groups demonstrated statistically equivalent bond strengths (mean difference = −1.096, p = 0.133), and both significantly outperformed all other groups. The Gl + HF group exhibited the lowest bond strength, which was significantly different from every other group (e.g. mean difference = −8.459 MPa vs. SB50 + Gl + HF, p < 0.001).

The control and SB50 groups were not significantly different from each other (mean difference = −0.357 MPa, p = 1.000), but both produced weaker bonds than the SB110 and SB50 + Gl + HF protocols (e.g. Control vs. SB110: mean difference = −1.707 MPa, p = 0.002).

Assessment of failure modes

The inter-rater reliability for failure mode classification showed almost perfect agreement (Cohen’s κ = 0.81) between the two examiners. The number and percentage of failure modes for each group are shown in Table 6. SB110 and SB50 + Gl + HF groups showed a high prevalence of cohesive and mixed failures, supporting their stronger bonds. In contrast, Gl + HF group presented adhesive failures exclusively, concomitant with its significantly weaker bonds. Microscopic images (20X) representing failure modes are shown in Fig. 8.

Table 6.

Prevalence of failure modes for the different groups

Groups (surface treatments)	Modes of failure
	Adhesive No. (%)	Cohesive No. (%)	Mixed No. (%)
Control	9 (90%)	-	1 (10%)
SB50	8 (80%)	-	2 (20%)
SB110	3 (30%)	2 (20%)	5 (50%)
Gl + HF	10 (100%)	-	-
SB50 + Gl + HF	2 (20%)	3 (30%)	5 (50%)

Fig. 8.

Stereomicroscopic images (20X) of different failure modes. (A) Adhesive; (B) Cohesive; (C) Mixed

Discussion

The results of this study demonstrate that the choice of surface treatment significantly influences the long-term durability of the bond between zirconia and indirect composite resin after thermocycling aging. Notably, sandblasting with 110 µm alumina particles (SB110) and the combined sandblasting-glazing protocol (SB50 + Gl + HF) yielded the highest significant bond strengths among other groups that exceeded the clinically acceptable threshold, with no significant difference between them. Therefore, the null hypothesis was rejected. The shear bond strength (SBS) test was selected for this study due to its clinical relevance in simulating masticatory stresses, its widespread adoption in assessing zirconia-composite adhesion, and its practical advantages in specimen design and alignment [36–38]. To evaluate long-term durability, all specimens underwent thermocycling aging (5,000 cycles), a method that accelerates hydrolytic degradation and induces stress at the bond interface through repeated thermal expansion and contraction of the materials. This protocol simulates approximately six months of clinical service [39] and provides critical predictive data on the bonding interface’s performance under conditions that mimic the oral environment.

The findings confirm the critical role of micromechanical retention, achieved through air abrasion, in enhancing bond strength. The SB110 protocol produced significantly higher SBS than the control and SB50 groups, a result underscored by the high incidence of cohesive and mixed failures. This can be attributed to the pronounced surface topography and microporosities created by the larger 110 µm particles, which facilitate more effective resin interlocking. This is also consistent with the findings of other studies [17, 18, 30, 31].

However, a critical consideration for clinical translation is the potential for air abrasion with larger particles to induce tetragonal-to-monoclinic phase transformation on the zirconia surface. While this transformation can create compressive stresses that may initially increase fracture toughness, it can also introduce surface microcracks and raise concerns about the long-term structural integrity and aging resistance of the zirconia framework itself [40, 41]. Therefore, while SB110 proved to be a highly effective surface treatment for bond durability in this study, its application on thin zirconia restorations or in high-stress areas should be considered with caution. Future studies incorporating X-ray diffraction (XRD) analysis are essential to quantify the phase transformation extent and correlate it with both bond strength and the structural stability of zirconia.

Conversely, the statistically insignificant improvement of SB50 over the control group indicates that 50 µm particles created insufficient roughness for effective micromechanical retention. In these groups, bonding relied primarily on chemical adhesion from the zirconia primer (Z-Prime Plus), which contains both 10-MDP and biphenyldimethacrylate (BPDM). The functional phosphate group of 10-MDP chemically bonds to hydroxyl groups on zirconia, while the methacrylate groups copolymerize with the composite resin. The BPDM monomer further enhances this chemical interaction [34, 42]. This primary reliance on chemical bonding, rather than micromechanical interlocking, explains the predominantly adhesive failure modes observed and renders the interface more vulnerable to hydrolytic degradation over time. This result comes in agreement with those of Zhao et al. [43] and Kim et al. [44] and also aligns with studies on high-strength 3Y-TZP zirconia, which recommend larger particle (i.e. 110 µm) abrasion for effective bonding [17, 40]. The discrepancy between these findings and studies showing positive effects with 50 µm particles [18, 26] may be attributed to differences in zirconia composition or primer/adhesive compositions and protocols.

The application of a silica-based glaze layer aims to create a chemically bondable surface on zirconia, yet the results were highly dependent on the pretreatment. The profoundly low performance of the Gl + HF group, despite the provision of a silica-rich, etchable surface, requires a multifactorial explanation. First, the adhesion of the glaze layer to the underlying polished zirconia is inherently weak owing to zirconia’s inertness, relying primarily on weak Van der Waals forces rather than chemical or micromechanical bonding [15, 45]. This creates a fragile interface that is highly vulnerable to stress. A significant source of this stress is the likely mismatch in the coefficient of thermal expansion (CTE) between the zirconia core and the silica-based glaze layer [46, 47]. During thermocycling, this CTE mismatch induces repeated interfacial stresses, progressively degrading the bond.

Second, the siloxane bonds formed between the silane and the etched glaze are themselves highly susceptible to hydrolytic degradation in the aqueous environment, a process dramatically accelerated by thermocycling [46, 47]. The CTE-induced stresses likely create micro-gaps that facilitate water penetration, further exacerbating this hydrolytic degradation. This combination of mechanical stress from CTE mismatch and chemical degradation explains the exclusively adhesive failure modes observed at the zirconia-glaze interface for this group. The outstanding performance of the SB50 + Gl + HF protocol demonstrates that these vulnerabilities can be overcome. The initial sandblasting with 50 µm alumina particles creates micromechanical retention that vastly improves the stability and adhesion of the overlying glaze layer, effectively locking it onto the zirconia surface and mitigating the detrimental effects of CTE mismatch and thermocycling aging. This combined approach ensures that bond failure is less likely to occur at the zirconia-glaze interface and is instead forced into the glaze-composite interface or within the materials themselves, as evidenced by the high rate of mixed and cohesive failures, resulting in significantly higher bond strength.

The stark contrast between the Gl + HF and SB50 + Gl + HF groups clarifies discrepancies in the literature. Studies reporting success with glazing alone often employed less demanding experimental conditions, such as shorter-term aging, failing to subject the vulnerable zirconia-glaze interface to a rigorous hydrolytic challenge [26] or the omission of MDP priming in sandblasted groups [46]. Furthermore, the superior performance of our combined protocol was likely aided by the use of a two-component silane system, which offers enhanced reactivity and shelf-life over prehydrolyzed one-bottle systems by ensuring fresh formation of active silanol groups immediately before application [48–50]. Ultimately, the variability in reported glaze layer performance across studies [51–53] can be attributed to a multitude of factors, including the type of glazing material, firing protocol, HF etching parameters, silane chemistry, and—most critically—the presence and rigor of artificial aging. Our findings, which align with other rigorous studies [54, 55], demonstrate that while glazing can be effective, its success is entirely contingent upon both a mechanical pretreatment to ensure adequate adhesion to the zirconia substrate and a stable silane bond. The superior performance of the combined protocol underscores the necessity of this dual mechanical-chemical strategy for achieving long-term durability.

The clinical significance of these findings is that they provide evidence-based, practical protocols for bonding indirect composite to zirconia leveraging the inherent advantages of indirect composite beyond repair scenarios and reduced wear on opposing dentition. The stress-absorbing nature of the composite, characterized by its elasticity and ability to tolerate deformation, enables it to be used in situations where stress management at the restoration interface is critical [13, 14]. Therefore, this study equips clinicians with a reliable method to expand their treatment options rather than replacing porcelain entirely. In summary, the outcomes of this study highlight the critical interplay between micromechanical and chemical bonding strategies as SBS values for the critical groups fell within 10–13 MPa which is regarded as the clinically acceptable range for composite/ceramic bond [56, 57]. The SB110 protocol relies primarily on robust micromechanical retention, optimized by larger particle abrasion, and is effectively stabilized by the chemical bond provided by the MDP primer. In contrast, the SB50 + Gl + HF protocol achieves its success through a synergistic approach: the initial sandblasting provides the essential micromechanical foundation for the glaze layer, which in turn facilitates a strong, silane-mediated chemical bond with the composite. The failure of the Gl + HF group underscores that without this mechanical foundation, chemical adhesion to a glaze layer is highly vulnerable to hydrolytic degradation. Therefore, the most durable bonds are achieved not by mechanical or chemical means alone, but by protocols that intelligently combine both mechanisms. Despite the clinically relevant findings, this study has several limitations that should be considered. Firstly, its in vitro design cannot fully replicate the complex oral environment, including the effects of saliva, pH fluctuations and variable masticatory forces on the bonding interface. Secondly, the study tested single brands of materials and lacked advanced surface characterization. The absence of scanning electron microscopy (SEM) analysis precluded a detailed topographical comparison of the treated surfaces and high-resolution imaging of the fracture modes. Furthermore, while the potential for air abrasion-induced phase transformation was discussed, it was not quantified using XRD analysis which would have provided valuable data on the long-term structural implications of the SB110 protocol. Thirdly, the thickness of the applied glaze layer was not standardized or measured using profilometry or SEM cross-sectioning. Variability in layer thickness could influence the resultant quality of the micromechanical interlock, potentially contributing to the observed variability in bond strength within these groups. Finally, the aging protocol, though rigorous, was limited to thermocycling. The absence of concomitant mechanical fatigue (chewing simulation) means the bonding interface was not tested under a full spectrum of clinically relevant stresses. Future studies should incorporate long-term aging, fractographic analysis via SEM, quantitative XRD for phase analysis, controlled glaze thickness application, and fatigue testing across a range of products to better predict clinical performance.

Conclusions

Within the limitations of this in vitro study, indirect composite resin demonstrates potential as a viable alternative to porcelain for veneering monolithic zirconia in specific clinical scenarios. From a scientific standpoint, in the short term (5,000 thermocycles), surface engineering of zirconia with either 110 µm aluminum oxide sandblasting or a combination of 50 µm sandblasting, glaze coating, and hydrofluoric acid etching yielded superior bond durability as opposed to the profound failure of glaze coating and HF etching protocol alone. These outcomes underscore that adhesion to zirconia cannot rely on chemical bonding alone; it must be underpinned by robust micromechanical retention. SB110 could be considered as the first choice protocol in terms of clinical practicality as it offers a safer, faster and less technique sensitive bonding strategy while, at the same time, showcasing clinically acceptable equivalent bond strength to the more complex SB50 + Gl + HF protocol. However, the long-term survivability of these two techniques should be investigated in clinical trials and longer term studies implementing harsher aging, fatigue simulations and advanced material characterization in order to understand their impact on zirconia’s structural integrity.

Abbreviations

CAD/CAM

Computer-aided design/computer-aided manufacturing

LT

Low Translucency

Al₂O₃

Aluminum oxide

SB

Sandblasting

HF

Hydrofluoric

Gl

Glaze

SBS

Shear bond strength

ANOVA

Analysis of Variance

SPSS

Statistical Package for the Social Sciences

10-MDP

10-Methacryloyloxydecyl dihydrogen phosphate

BPDM

Biphenyl dimethacrylate

3Y-TZP

3 mol% yttria-tetragonal zirconia polycrystal

HEMA

Hydroxyethyl methacrylate

CTE

Coefficient of thermal expansion

XRD

X-ray diffraction

SEM

Scanning electron microscope

Authors’ contributions

DS and IK contributed to the conception and design of the work. They also contributed to the acquisition, analysis and interpretation of data. DS drafted the work; and DS and IK substantively revised it. All authors read and approved the final manuscript.

Funding

This research received no external funding.

Data availability

All data generated or analysed during this study are included in this published article [and its supplementary information files].

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.