Influence of surface-conditioning strategies on the shear bond strength of veneering ceramic to aged and non-aged zirconia frameworks fabricated with milling and 3D-printing

Abstract

PURPOSE

The purpose of this in vitro study was to investigate how various surface-conditioning strategies and thermal aging affect the shear bond strength (SBS) of aged and non-aged zirconia frameworks fabricated through different techniques.

MATERIALS AND METHODS

A total of 208 rectangular frameworks were fabricated by computer numerical control (CNC) and lithography-based ceramic manufacturing (LCM) techniques. The specimens were randomly allocated into 4 groups according to the surface-conditioning strategy applied on the veneering surface: no surface conditioning (C), airborne-particle abrasion (APA), tribochemical silica-coating (TSC), and grinding (GRD). Surface topographies were scrutinized under a scanning electron microscope, surface roughness (Ra) was measured, and the SBS test was performed. Failure patterns were evaluated. Data were statistically analyzed (α = 0.05).

RESULTS

Three-way ANOVA showed significant main effects of fabrication technique, surface-conditioning, and thermal aging (P < .001 for all) on SBS, with no significant three-way interaction (P = .225). LCM showed higher SBS than CNC (24.02 vs 22.04 MPa), TSC yielded the highest value (32.28 MPa), and aging reduced overall SBS from 27.20 to 18.85 MPa. Two-way ANOVA showed significant main effects of fabrication technique and surface-conditioning on Ra, with significant interaction between them (P < .001 for all).

CONCLUSION

Surface-conditioning strategy had the greatest influence on zirconia-veneer bond performance, with TSC producing the highest SBS values regardless of fabrication technique. Although LCM exhibited statistically higher SBS values than CNC, this difference was modest and both techniques demonstrated comparable bond performance from a practical perspective. LCM specimens also exhibited higher surface roughness than CNC specimens.

INTRODUCTION

Zirconia is a polymorphic ceramic material known for its exceptional mechanical properties.¹,²,³,⁴,⁵,⁶ Unique among ceramics, zirconia can transition among monoclinic, tetragonal, and cubic phases depending on temperature. The incorporation of approximately 3 mol% yttrium oxide into zirconium dioxide (3Y-TZP) stabilizes the tetragonal phase, resulting in a microstructure composed predominantly of fine tetragonal zirconia grains.¹,⁷,⁸,⁹ While this configuration enhances the material’s mechanical properties through transformation toughening mechanism,¹,¹⁰ it also induces significant birefringence, thereby diminishing light transmittance and contributing to the inherent low translucency.¹⁰,¹¹,¹² Therefore, zirconia frameworks are veneered to enhance visual appeal,²,⁶,¹³,¹⁴,¹⁵,¹⁶,¹⁷ commonly with low-fusing feldspathic ceramics.¹⁸

Traditionally, zirconia framework has been veneered using a layering technique, wherein a slurry composed of ceramic powder and modeling liquid is meticulously applied to the framework with a brush, followed by sintering at temperatures between 850℃ and 950℃.²,¹⁵ However, it is technique sensitive.¹⁹ An alternative approach, the heat-pressing technique, utilizes the lost-wax method to fabricate the veneer, resulting in a material with greater homogeneity and reduced porosity. To streamline the manufacturing process and reduce variability introduced by operator skill, the file-splitting techniques (CAD-on and Rapid Layer Technology) have emerged. This digitization enables the 3D design and milling of veneer components through CAD-CAM technology, which can then be either adhesively bonded using resin-based luting agents or fused to the zirconia framework with fusion glass ceramics²,²⁰ Lithium disilicate has emerged as a high-strength glass-ceramic option for veneering zirconia frameworks using the CAD-on technique.²

Clinical outcomes have revealed a high incidence of catastrophic failures such as chipping (a cohesive bond failure occurring within the veneering ceramic)²,³,²¹,²²,²³ and delamination (an adhesive bond failure occurring at the interface).²,⁶,⁸,²¹,²²,²⁴,²⁵,²⁶ These failures may arise from (i) an incompatibility between the thermal expansion coefficients of zirconia framework and the veneering ceramic, (ii) improper framework design, (iii) the fluidity, volumetric shrinkage, and wettability of the superstructure, (iv) insufficient mechanical strength of the veneer layer, (v) irregular layer thicknesses, (vi) improper veneering technique, and (vii) poor interfacial bonding.²,⁶,⁸,²¹,²⁷,²⁸ Accordingly, it can be emphasized that the interfacial bonding between the veneering ceramic and zirconia framework plays a pivotal role in the long-term clinical success of bi-layered all-ceramic restorations.²¹ A strong and durable bond ensures mechanical integrity at the framework-veneer interface,²⁰ hinders interfacial gap formation, and prevents the above-mentioned failure modes under intraoral conditions, where cyclic occlusal forces and thermal fluctuations are constantly at play. Moreover, sufficient shear bond strength (SBS) contributes to uniform stress distribution across the restoration,²,²⁵ reducing the likelihood of crack initiation and propagation at the interface.

Due to its polycrystalline microstructure and lack of glassy phase, zirconia is resistant to conventional bonding techniques like acid etching and silane application, which are generally effective for silica-based ceramics.²⁹,³⁰ Owing to its chemically inert surface and low surface energy, various interfacial surface conditioning strategies have been explored to alter zirconia’s topography and chemistry, aiming to enhance surface roughness and energy and thus improve micromechanical interlocking and chemical bonding with veneering ceramic.⁶,⁷,²⁴,²⁹,³¹

3D-printing has emerged as a promising alternative to the conventional computer numerical control (CNC) technique. It constructs structures through a layer-by-layer deposition of material, allowing for the fabrication of complex geometries that are often unachievable with traditional CNC tools.¹⁰,³² This approach not only facilitates the simultaneous production of multiple customized components but also enhances design flexibility and material efficiency.³²,³³ Among the various 3D-printing technologies, lithography-based ceramic manufacturing (LCM) stands out for its ability to produce ceramic components with high dimensional accuracy and favorable mechanical properties, positioning it as a valuable technique. LCM is a vat photopolymerization-based process that utilizes a highly loaded ceramic slurry composed of fine ceramic particles dispersed in a photosensitive resin. During printing, each layer is selectively polymerized through light at 460 nm wavelength, producing a green body with high resolution and dimensional accuracy. The green part then undergoes thermal debinding to remove the organic binder and is sintered at elevated temperatures to achieve full densification and restore the intrinsic mechanical and optical properties of the ceramic.³²,³³,³⁴,³⁵ The manufacturer has reported that an embedded wide quad extended graphics array projector provides superior precision.

To the best of the authors’ knowledge, studies investigating the SBS of veneering ceramics to 3D-printed zirconia frameworks are scarce. Given the differences in surface morphology, density, and microstructure and surface roughness introduced by the 3D-printing process, it is critical to explore how these factors affect the adhesion between zirconia framework and veneering ceramic. Therefore, it was aimed in this in vitro study to evaluate and compare the influence of various interfacial surface conditioning strategies (control [C], air-borne particle abrasion with aluminum oxide (Al₂O₃) particles [APA], tribochemical silica coating [TSC], and grinding [GRD]) on the SBS of veneering ceramic to aged and non-aged zirconia frameworks fabricated with both LCM and CNC techniques. The null hypotheses were that there would be no significant difference in the SBS of veneering ceramic to zirconia frameworks between LCM and CNC techniques, that neither surface conditioning strategies nor thermal aging would have significant effects on the SBS values, and that there would be no significant difference in the surface roughness (Ra) of zirconia frameworks between LCM and CNC techniques or among the different surface conditioning strategies.

MATERIALS AND METHODS

A rectangular specimen (14 × 12 × 1 mm) was virtually designed using a software program (Tinkercad; Autodesk, San Francisco, CA, USA), and the geometry was saved in a standard tessellation language (STL) file format for the fabrication of all specimens. Using a software program (G*power, v.3.1.9.7; Heinrich-Heine-University, Düsseldorf, Germany), the minimum required sample size was determined to be 13, assuming 85% study power, an effect size of 0.25, and a significance level of 0.05. A total of 208 rectangular zirconia frameworks were fabricated by LCM and CNC techniques (n = 104 per technique). The experimental setup is illustrated in Fig. 1.

Fig. 1. Experimental design (CNC, Computer Numerical Control; LCM, Lithography-based Ceramic Manufacturing; APA, Airborne-particle abrasion; C, control; TSC, tribochemical silica-coating; GRD, grinding; SEM, scanning electron microscope; Ra, surface roughness; TA[-], non-aged thermally; TA[+], aged thermally).

Fabrication of specimens

CNC specimens were milled (Redon GTR; Redon, İstanbul, Türkiye) from a pre-sintered zirconia block (GC Initial Zirconia HT; GC) and subsequently sintered (TABEO-1/M/ZIRKON-100; Ador Dental GmbH, Tokyo, Japan) to achieve full density at 1500℃ according to the manufacturer’s instructions.

LCM specimens were 3D-printed (Cerafab S65 Medical; Lithoz, Vienna, Austria) from a zirconia slurry (LithaCon 3Y 210; Lithoz, Vienna, Austria). The layer thickness was adjusted to 25 µm, with a printing resolution of 40 × 40 µm. Following fabrication, the green-state 3D-printed specimens were cleaned to remove residual unpolymerized slurry using a solution (LithaSol 30; Lithoz, Vienna, Austria) in combination with an air-pressure brush (Iwata Airbrush; Anest Iwata-Medea, Yokohama, Japan). Subsequently, the samples underwent 2 successive thermal processing stages, namely debinding and sintering. During the sintering stage (LHTCT; Nabertherm GmbH, Lilienthal, Germany), the temperature increased to a peak of 1450℃ and was maintained for 2 hours (Fig. 2).

Fig. 2. Thermal post-processing of 3D-printed specimens.

The bonding surfaces were sequentially wet polished using silicon carbide abrasive papers with grit sizes of 600, 800, and 1000. Thereafter, all specimens were ultrasonically cleaned (Sonorex Digital; Bandelin Electronic, Berlin, Germany) for 10 minutes and dried with oil-free air to eliminate surface residues (Fig. 3).

Fig. 3. Zirconia specimens (A) Milled [CNC]; (B) 3D-printed [LCM].

Implementation of surface conditioning strategies

The specimens for each above-mentioned fabrication group were randomly allocated into 4 groups in accordance with the surface conditioning strategy implemented on the veneering surface: C, APA, TSC, and GRD (n = 26 per strategy).

• • The veneering surfaces of the C group specimens were left untreated.

  • The veneering surfaces of the APA group specimens were abraded (Basic master; Renfert) with 110-µm Al₂O₃ particles (Cobra Aluoxyd; Renfert) at 0.2 MPa pressure perpendicularly from a 10-mm distance for 15 seconds. A metal apparatus (Fig. 4) was employed to maintain a constant distance of 10 mm.

  • The veneering surfaces of the TSC group specimens were abraded (Rocatector Delta device; 3M ESPE, Seefeld, Germany) with 110-µm Al₂O₃ particles modified with silica (Rocatec Plus; 3M ESPE, Seefeld, Germany), following the same protocol used for the APA group. The veneering surfaces were then silanized with the aid of a micro-brush (3M ESPE Sil; 3M ESPE, Seefeld, Germany), allowed to polymerize for 60 seconds, and air-dried with compressed air for 5 seconds.

  • The veneering surfaces of the GRD group specimens were ground with a coarse-grit (120 µm) cylindrical diamond rotary instrument mounted in a high-speed handpiece (Extratorque 506C; KaVo, Biberach, Germany) operating at 350,000 rpm under water cooling. The instrument was held parallel to the horizontal plane and applied to the surface for 10 seconds.

Fig. 4. Metal apparatus used to standardize 10-mm nozzle-to-specimen distance during airborne-particle abrasion.

Scanning electron microscope analysis

After surface conditioning strategies, 2 specimens from each group were randomly selected, dried, sputter-coated with palladium (Palaron, Sputter Coater; Quorum Technologies, Sussex, UK), and examined under a scanning electron microscope (FEI Quanta 400 ESEM FEG; Thermo Fisher Scientific, Hillsboro, OR, USA) at ×500 magnification with an accelerating voltage of 30 kV (Fig. 5).

Fig. 5. Scanning electron micrographs (×500). (A) Unconditioned CNC; (B) APA-conditioned CNC; (C) TSC-conditioned CNC; (D) GRD-conditioned CNC; (E) Unconditioned LCM; (F) APA-conditioned LCM; (G) TSC-conditioned LCM; (H) GRD-conditioned LCM).

Surface roughness measurement

Surface roughness (Ra, µm) values were measured using a contact profilometer (TIME^®3200/3202 [TR200]; Beijing Time High Technology Ltd, Beijing, China). For each specimen, 3 measurements were obtained at different locations, and the mean value was calculated.

Layering veneer porcelain

For applying veneer porcelain, the layering technique was used. To ensure standardization, an alumina ring mold was attached to zirconia specimens and filled with veneering porcelain (IPS Emax Ceram Dentin Porcelain; Ivoclar Vivadent, Schaan, Liechtenstein). After condensation and drying processes, the mold was removed, and the specimen was sintered in a porcelain furnace at 750℃ (Programat EP5000 G2 press furnace; Ivoclar Vivadent, Schaan, Liechtenstein) according to the manufacturer’s instructions (Table 1). After firing, the diameter of the veneered ceramic was measured as Ø3 × 3 mm. To offset firing shrinkage and avoid the need for additional firing, the mold dimensions were enlarged beyond the intended final dimensions. The bonded surface area was calculated according to the diameter of the cylindrical veneering porcelain. Since the bonding diameter was standardized at 3 mm, the bonding area was determined using the formula A = πr², where r = 1.5 mm, resulting in a bonded surface area of 7.07 mm².

Table 1. Firing parameters for IPS e.max Ceram Dentin veneering ceramic.

Firing cycle	Stand-by temperature ℃/°F	Closing time / min ℃/°F	Heating rate ℃/°F/min	Firing temperature ℃/°F	Holding time / min	Vacuum 1 ℃/°F	Vacuum 2 ℃/°F
IPS e-max	403 / 757	4:00	50 / 90	750 / 1382	1:00	450 / 842	749 / 1380

Thermal aging

In each surface conditioning strategy, the specimens were further divided into two subgroups as non-aged (TA[-]) and thermally-aged (TA[+]) (n = 13 per group). Relevant specimens were thermally-aged in a thermal cycle device (SD Mechatronik Thermocycler, Westerham, Germany) with a bath temperature of 5 – 55℃, a transfer time of 10 s between baths, and a dwell time of 30 s in the bath for 10,000 cycles which simulates 1 year of use in the oral environment (Fig. 6).²⁷

Fig. 6. Thermal aging of specimens.

Shear bond strength test

All specimens were embedded in auto-polymerizing polymethyl methacrylate resin blocks (Integra; BG Dental, Ankara, Türkiye). The SBS test was carried out using a universal testing machine (Marestek; Mares Engineering, İstanbul, Türkiye). A knife-edge chisel was aligned parallel to the interface, and shear load was applied at a crosshead speed of 1.0 mm/min until specimen failure occurred. Each SBS value was calculated in megapascal according to the following formula: SBS = F/S where F is maximum load force (Newton) when failure occurred, and S is bonded area (mm²).

Determination of failure patterns

The interfaces were examined under a stereomicroscope (Leica S8 APO; Leica Microsystems, Wetzlar, Germany) at × 40 magnification to identify the failure patterns which are categorized as adhesive failure at the zirconia surface with no residues of veneer ceramic, cohesive failure within the veneer ceramic, and mixed failure consisting of both cohesive and adhesive failures.

Statistical analysis

All data were statistically analyzed (IBM SPSS Statistics, v23; IBM, Armonk, NY, USA). Both for SBS and Ra data, the Shapiro-Wilk test validated the supposition of normal distribution (P > .05), and homogeneity of the variance of the groups was checked using Levene’s test. Three-way analysis of variance (ANOVA) was conducted to investigate the influence of 3 variables (fabrication technique, surface-conditioning strategy, and thermal aging) on SBS data. Two-way analysis of variance (ANOVA) was conducted to investigate the influence of 2 variables (fabrication technique and surface-conditioning strategy) on Ra data. The Tukey post hoc test was used for multiple comparisons (α = 0.05).

RESULTS

A 3-way ANOVA was conducted to examine the effects of fabrication technique (CNC vs LCM), surface-conditioning strategy (C, APA, TSC, GRD), and thermal aging (TA[-] vs TA[+]) on SBS (Table 2). Descriptive statistics and pairwise comparisons are illustrated in Table 3 and Table 4. The overall model was statistically significant and explained a very large proportion of the variance in SBS (R² = 0.964, adjusted R² = 0.961). The 3-way interaction among fabrication technique, surface-conditioning strategy, and thermal aging was not statistically significant, F(3,192) = 1.467, P = .225, ηp² = 0.022, indicating that the combined effects of fabrication technique and surface conditioning on SBS were consistent across aging conditions. Significant 2-way interactions were observed between fabrication technique and surface-conditioning strategy, F(3,192) = 5.166, P = .002, ηp² = 0.075, and between surface-conditioning strategy and thermal aging, F(3,192) = 4.331, P = .006, ηp² = 0.063. However, the interaction between fabrication technique and thermal aging was not statistically significant, F(1,192) = 0.770, P = .381, ηp² = 0.004.

Table 2. Three-way ANOVA results for SBS data.

Source	Type III Sum of squares	df	Mean square	F	P
Fabrication Technique (A)	202.391	1	202.931	99.133	< .001
Surface-conditioning Strategy (B)	6692.229	3	2230.743	1089.726	< .001
Aging (C)	3625.988	1	3625.988	1771.308	< .001
A * B	31.724	3	10.575	5.166	.002
A * C	1.577	1	1.577	0.770	.381
B * C	26.598	3	8.866	4.331	.006
A * B * C	9.006	3	3.002	1.467	.225

R squared = 0.964 (Adjusted R squared = 0.961)

Table 3. Descriptive statistics for mean SBS ± SD values with pairwise comparisons according to main effects and two-way interaction levels.

Surface conditioning strategy	Fabrication technique		Aging		Total
	CNC	LCM	TA[-]	TA[+]
Control	16.77 ± 4.12a,A	18.09 ± 3.35b,A	20.98 ± 1.12b,A	13.86 ± 1.22a,A	17.43 ± 3.77A
APA	22.06 ± 4.55a,C	23.53 ± 4.84b,C	27.21 ± 2.01b,C	18.38 ± 0.82a,C	22.80 ± 4.71C
TSC	30.63 ± 4.96a,D	33.93 ± 4.78b,D	36.67 ± 3.08b,D	27.89 ± 1.86a,D	32.28 ± 5.10D
GRD	18.71 ± 4.55a,B	20.52 ± 4.67b,B	23.95 ± 2.03b,B	15.28 ± 1.01a,B	19.61 ± 4.66B
Total	22.04 ± 6.97a	24.02 ± 7.49b	27.20 ± 6.30b	18.85 ± 5.64a	23.03 ± 7.28

C, Control – no treatment; APA, air-borne particle abrasion with aluminum oxide particles; TSC, tribochemical silica-coating; GRD, grinding; TA[-], non-aged thermally; TA[+], aged thermally; CNC, computer numerical control; LCM, lithography-based ceramic manufacturing. Superscript uppercase letters indicate statistically significant differences within the same column, while superscript lowercase letters denote significant differences within the same row.

Table 4. Descriptive statistics for mean SBS ± SD values with pairwise comparisons according to three-way interaction combinations.

Conditioning strategy	Thermal aging	CNC	LCM
C	TA[+]	20.71 ± 1.00	21.26 ± 1.21
	TA[-]	12.82 ± 0.72	14.92 ± 0.40
APA	TA[-]	26.28 ± 2.03	28.14 ± 1.58
	TA[+]	17.84 ± 0.69	18.92 ± 0.55
TSC	TA[-]	35.12 ± 2.66	38.22 ± 2.73
	TA[+]	26.14 ± 0.71	29.63 ± 0.33
GRD	TA[-]	23.10 ± 1.12	24.80 ± 2.39
	TA[+]	14.32 ± 0.31	16.24 ± 0.12

C, Control – no treatment; APA, air-borne particle abrasion with aluminum oxide particles; TSC, tribochemical silica-coating; GRD, grinding; TA[-], non-aged thermally; TA[+], aged thermally; CNC, computer numerical control; LCM, lithography-based ceramic manufacturing.

Significant main effects were found for all 3 factors. Fabrication technique had a significant effect on SBS, F(1,192) = 99.133, P < .001, ηp² = 0.34, with LCM specimens demonstrating higher overall SBS than CNC specimens (24.02 ± 7.49 MPa vs 22.04 ± 6.97 MPa). Surface-conditioning strategy showed the largest effect on SBS, F(3,192) = 1089.726, P < .001, ηp² = 0.94. TSC produced the highest SBS values (32.28 ± 5.10 MPa), followed by APA (22.80 ± 4.71 MPa), GRD (19.61 ± 4.66 MPa), and the untreated C group (17.43 ± 3.77 MPa). TA[+] also had a strong negative effect on SBS, F(1,192) = 1771.308, P < .001, ηp² = 0.90, with TA[-] specimens showing substantially higher SBS than TA[+] specimens (27.20 ± 6.30 MPa vs 18.85 ± 5.64 MPa).

Descriptive analyses demonstrated that LCM consistently produced higher SBS values than CNC across all surface-conditioning strategies and aging conditions. TA[+] resulted in a pronounced reduction in SBS within each conditioning group. Pairwise comparisons confirmed that all surface-conditioning strategies differed significantly from each other (P < .05), with TSC showing superior performance and the C group demonstrating the lowest SBS. Despite these differences, the non-significant three-way interaction indicates that the pattern of differences between fabrication techniques and conditioning strategies remained parallel across aging conditions.

Fig. 5 presents the scanning electron micrographs of the tested zirconia groups. C specimens exhibited relatively smooth surfaces with minimal irregularities, while APA- and TSC-conditioned groups showed rougher, irregular morphologies with distinct micro-retentive features and no evident directional pattern. In contrast, the GRD group revealed clear parallel scratches and grooves consistent with the direction of bur movement, producing an anisotropic texture. The LCM-fabricated groups demonstrated generally more irregular and heterogeneous surface topographies compared to their CNC-fabricated counterparts.

Two-way ANOVA revealed that fabrication technique significantly influenced surface roughness (Ra), F(1,96) = 742.211, P < .001, ηp² = 0.885. Surface-conditioning strategy also showed a highly significant main effect, F(3,96) = 1004.571, P < .001, ηp² = 0.969. In addition, a statistically significant interaction was found between fabrication technique and surface-conditioning strategy, F(3,96) = 75.977, P < .001, ηp² = 0.704, indicating that the effect of surface treatment varied according to manufacturing method (Table 5). The overall model was statistically significant and explained a very large proportion of the variance in surface roughness values (R² = 0.976, adjusted R² = 0.975).

Table 5. Two-way ANOVA results for Ra data.

Source	Type III Sum of squares	df	Mean square	F	P
Fabrication Technique (A)	2.990	1	2.990	742.211	< .001
Surface-conditioning Strategy (B)	12.139	3	4.046	1004.571	< .001
A * B	0.918	3	0.306	75.977	< .001

a. R squared = 0.976 (Adjusted R squared = 0.975)

Descriptive statistics and pairwise comparisons are illustrated in Table 6. Pairwise comparisons revealed that all conditioning strategies produced significantly different Ra values within each manufacturing technique (P < .001). In both CNC and LCM groups, surface roughness increased progressively from control to APA, TSC, and GRD. The lowest Ra values were observed in the control groups (CNC: 0.67 ± 0.04; LCM: 0.71 ± 0.04), whereas grinding generated the highest roughness values (CNC: 1.37 ± 0.05; LCM: 1.82 ± 0.02). For APA, TSC, and GRD conditions, LCM specimens demonstrated significantly higher Ra values than CNC specimens (P < .001), whereas no statistically significant difference was found between fabrication techniques in the untreated control condition (P = .096).

Table 6. Descriptive statistics for mean Ra ±SD values with pairwise comparisons.

Conditioning strategy	CNC	LCM	Total
C	0.67 ± 0.04a,A	0.71 ± 0.04a,A	0.69 ± 0.04A
APA	0.99 ± 0.05a,B	1.32 ± 0.09b,B	1.16 ± 0.18B
TSC	1.16 ± 0.09a,C	1.70 ± 0.08b,C	1.43 ± 0.29C
GRD	1.37 ± 0.05a,D	1.82 ± 0.02b,D	1.59 ± 0.23D
Total	1.05 ± 0.26a	1.34 ± 0.44b	1.22 ± 0.40

C, Control – no treatment; APA, air-borne particle abrasion with aluminum oxide particles; TSC, tribochemical silica-coating; GRD, grinding; CNC, computer numerical control; LCM, lithography-based ceramic manufacturing. Superscript uppercase letters indicate statistically significant differences within the same column, while superscript lowercase letters denote significant differences within the same row.

Fig. 7 illustrates the failure patterns. In the TA[-] group, CNC specimens showed 32 adhesive, 2 cohesive failures out of 52 specimens; the remaining specimens showed mixed failures. LCM specimens showed 29 adhesive, 19 mixed and 4 cohesive failures. In the TA[+] group, CNC specimens showed 38 adhesive, 1 cohesive, 13 mixed failures. LCM specimens showed 36 adhesive, 2 cohesive, 14 mixed failures.

Fig. 7. Failure patterns. (A) Adhesive; (B) Mixed; (C) Cohesive.

DISCUSSION

This study assessed the effect of different conditioning strategies (APA, TSC, and GRD) on the SBS of veneering ceramic to aged and non-aged zirconia frameworks fabricated through CNC and LCM techniques. The results indicated that SBS were significantly influenced by the variables. Moreover, fabrication technique and surface-conditioning strategy significantly influenced the Ra values. Accordingly, all null hypotheses were rejected.

Among tested conditioning strategies, TSC achieved the highest mean bond strengths (TA[-] 35.12 ± 2.66, TA[+] 26.14 ± 0.71 for the CNC group and TA[-] 38.22 ± 2.73, TA[+] 29.63 ± 0.33 for the LCM group). This is in accordance with the previous studies.²⁷,³⁰ However, earlier research³⁶ compared different strategies, including TSC, and reported that neither strategies increased bond strength. In TSC, the surface is blasted with silica-coated alumina particles. Kinetic energy causes micro-abrasion, increasing surface roughness, which improves mechanical interlocking. Simultaneously, due to the high-velocity impact, silica particles become embedded in the zirconia surface to a depth of 15 µm.³¹,³⁶,³⁷,³⁸ The tribochemical reaction facilitates a thin silica layer forming on the roughened zirconia surface. This silica-coated surface can then react with silane coupling agents, which act as molecular bridges. When veneering ceramic is applied, the silane-modified zirconia surface improves chemical bonding with the silica-rich veneering ceramic. This results in increased wettability and chemical adhesion.³⁷,³⁸ This dual mechanism, which includes mechanical interlocking from increased roughness and chemical adhesion through silanization, might explain why TSC performed better in this study. However, when TSC is not available, APA may serve as the nearest practical alternative in clinical settings, although it may not achieve the same SBS. This is because APA provides only micromechanical retention without surface chemical activation.³¹ Surface roughness is related to adhesion and is a key factor influencing the long-term durability of restorations.³⁹,⁴⁰ Rough surfaces increase the available surface area for bonding and enhance surface free energy, thereby improving wettability.³⁸,³⁹,⁴⁰ Surface conditioning protocols generally increase the surface roughness. The findings of this study are consistent with previous reports, demonstrating that surface conditioning strategies increased surface roughness (Ra) values.⁴¹,⁴²,⁴³,⁴⁴

Although GRD exhibited higher SBS values compared to the C groups, its values (TA[-] 23.10 ± 1.12, TA[+] 14.32 ± 0.31 for the CNC group and TA[-] 24.80 ± 2.39, TA[+] 16.24 ± 0.12 for the LCM group) were lower than those obtained with the other surface conditioning strategies. On the other hand, it showed the highest roughness values among all groups. Previous studies have indicated that GRD either failed to improve the SBS between framework and veneer⁴⁵ or exerted a detrimental effect,⁴⁶ contradicting the assumption that greater surface roughness of Y-TZP enhances bonding. The performance of the GRD group in this study can be attributed to a few reasons. First, GRD does not enhance the surface chemically, as seen in TSC.⁴⁷ Second, GRD may introduce surface flaws, microcracks, and phase transformation (tetragonal → monoclinic) on the zirconia surface due to its abrasive nature.¹,² These flaws may act as stress concentrators and initiate crack propagation at the interface under loading. Third, while GRD increases surface roughness, the morphology may be irregular⁵ or contaminated as seen after milling. This may reduce wettability and limit the intimate adaptation of the veneering ceramic.

Despite broadly comparable performance between fabrication techniques, LCM specimens exhibited slightly but statistically significantly higher SBS and Ra values than CNC specimens. This finding holds implications for 3D-printed zirconia, indicating an enhanced responsiveness to surface conditioning protocols. Such behavior may be attributed to alterations in surface porosity or compositional characteristics inherent to the layer-by-layer fabrication process of 3D-printing.⁴⁸ Previous studies⁴,²⁹ have shown that 3D-printed specimens have increased surface roughness compared to CNC ones. Although this surface characteristic is mainly linked to the nature of 3D-printing; the building orientation, layer thickness, light exposure intensity, and firing protocols can become influential in the background.⁴⁸,⁴⁹,⁵⁰ Higher SBS values observed in LCM can be correlated with the microporous nature of 3D-printed specimens. This may allow the penetration of veneering ceramic into those indentations and thereby enhance micromechanical retention. However, some studies in the literature have reported that CNC-fabricated specimens may exhibit higher surface roughness values compared to 3D-printed ones.¹³,⁴⁹,⁵¹ These discrepancies may be attributed to variations in additive manufacturing techniques or differences in the measurement methods used to evaluate surface roughness.

The compatibility of thermal expansion coefficients (CTE) between core and veneering materials plays a critical role in the long-term success of layered restorations.¹³,¹⁵,²³ In the present study, all materials exhibited closely matched CTE values, ranging approximately from 10.0 to 11.0 × 10^-6/K. This thermal expansion harmony minimizes the risk of interfacial stresses during firing and intraoral thermal cycling; thereby, challenges deriving from CTE mismatch were effectively eliminated. As a result, CTE was not considered a dependent variable in this study, allowing the focus to remain on other influential factors affecting the veneer-core interface.

In this study, TA[+] specimens demonstrated a substantial reduction in SBS across all fabrication techniques and surface-conditioning strategies, indicating that TA[+] acts as a dominant and consistent degrading factor at the ceramic-ceramic interface. Similar findings have been reported by Özdemir et al.,²⁷ whereas Guess et al.⁵² reported no significant difference between aged and non-aged specimens. This reduction can be attributed to degradation mechanisms associated with thermal aging. Under humid conditions, zirconia undergoes tetragonal-to-monoclinic (t→m) phase transformation, known as low-temperature degradation (LTD), which leads to volumetric expansion, microcracking, and deterioration of mechanical properties.⁵³,⁵⁴,⁵⁵ In addition, water diffusion and thermal fluctuations promote interfacial stress accumulation and slow crack growth, weakening both micromechanical interlocking and chemical bonding at the zirconia-ceramic interface.⁵³,⁵⁴,⁵⁵ Aging-related microstructural changes, such as surface defects and grain pull-out, may further compromise structural integrity.⁵³,⁵⁴,⁵⁵ The consistent TA[+] effect across all groups indicates that, although surface treatments improve initial SBS, they cannot fully prevent long-term degradation under simulated oral conditions. Clinically, this suggests that the durability of zirconia-ceramic bonding depends more on resistance to hydrothermal aging than on initial bond strength alone.

Failure mode analysis provides insight into the bond quality between zirconia and veneering ceramic. Adhesive failure occurs at the interface, whereas cohesive failure occurs within the veneering ceramic, mixed failure represents a combination of both.¹³,⁵⁶ In zirconia-based systems, the zirconia core-ceramic veneering interface is considered the weakest part.⁵⁶ In the present study, failure patterns were predominantly adhesive, suggesting that interfacial failure remains the main fracture mode, particularly after thermal aging. The predominance of adhesive failures can be attributed to the inherent limitations of SBS testing, which concentrates stresses at the zirconia-ceramic interface and promotes crack initiation at this region.⁵⁶ This phenomenon does not necessarily indicate weak bonding, but rather reflects the test configuration and crack propagation behavior within zirconia-based systems.¹⁵ Moreover, surface conditions play a critical role, as untreated specimens tend to result in adhesive failures, whereas improved surface conditioning promotes a shift toward mixed and cohesive failures.²⁸,⁵⁷,⁵⁸ Moreover, LCM specimens exhibited higher SBS values and a lower proportion of adhesive failures compared to CNC specimens, indicating improved interfacial bonding performance.¹⁰

Another important consideration in interpreting the findings is the difference in sintering protocols between the groups. Although both groups were sintered at similar final temperature and dwell time, differences in the overall thermal processing protocols should be acknowledged as a limitation. In particular, the more complex and prolonged thermal process of additively manufactured zirconia, including debinding and stepwise heating stages, may influence microstructural features such as grain structure and porosity.⁵⁹,⁶⁰,⁶¹,⁶² These variations could affect surface characteristics and, consequently, bonding behavior. Therefore, the observed differences should not be attributed solely to the manufacturing technique, as thermal processing may act as a confounding factor.

In addition to the findings discussed above, it is essential to consider that the SBS testing method has inherent characteristics that may influence the results. SBS testing is sensitive to experimental variables and may show limited correlation with clinical outcomes. In addition, shear-based testing methods are associated with non-uniform stress distribution and stress concentration at the bonding interface, which may not accurately reflect intraoral loading conditions.⁶³,⁶⁴,⁶⁵ Therefore, the present findings should be interpreted as comparative laboratory data, while their clinical implications should be considered with caution.

This study has several limitations. Only a single type of veneering ceramic and one veneering technique were employed, which may limit the generalizability of the findings. A limited number of surface treatments were applied, and a single zirconia generation (3Y) and a single surface roughness measuring device (contact profilometer) were used. Future studies should consider the use of various veneering ceramics, layering techniques, surface treatments, zirconia generations, non-contact surface roughness measurement methods (confocal profilometry or atomic force microscopy) to evaluate the stability of the bond strength and surface characterization.

CONCLUSION

Within the limitations of this in vitro study, the following conclusions were drawn:

• 1. Surface conditioning strategy had the greatest influence on zirconia-veneer bond performance, with tribochemical silica coating producing the highest shear bond strength values regardless of fabrication technique.

  2. Although LCM-fabricated zirconia specimens demonstrated statistically higher SBS values than CNC-fabricated specimens, this difference was modest and both techniques demonstrated comparable bond performance from a practical perspective.

  3. Surface roughness was higher in LCM-fabricated specimens than in CNC-fabricated specimens.

  4. Thermal aging significantly reduced SBS in all groups, indicating deterioration of interfacial durability irrespective of fabrication method.

  5. Optimization of surface treatment plays a more critical role than fabrication technique in improving zirconia bonding performance.