2.1

Specimen preparation

Fifteen non-carious human lower third molars were stored in a 0.5% chloramine solution after extraction. All teeth were mounted in gypsum blocks to facilitate specimen manipulation and the occlusal third of the crown was removed using a slow-speed diamond saw (Isomet 1000, Buehler, Lake Bluff, IL, USA). Dentin surfaces were checked with a stereomicroscope (Wild M5A, Heerbrugg, Switzerland) for absence of enamel and pulp tissue. On each tooth, a standard smear layer was produced using a high-speed medium-grit diamond bur (100 μm, No.: 842314014, Komet, Lemgo, Germany) mounted in the MicroSpecimen Former (The University of Iowa, Iowa City, IA, USA). Three dentin surfaces were exposed to each of the five different etching procedures:

• Group A: Dentin was etched for 15 s with 37.5% phosphoric acid (Kerr Gel, Orange, CA, USA), rinsed for 15 s and air-dried for 5 s (Control).

• Group B: Dentin was etched for 15 s with 37.5% phosphoric acid, rinsed for 15 s and air-dried for 5 s. Subsequently, dentin was etched for 15 s with 3% hydrofluoric acid (DenMat, Santa Maria, CA, USA), rinsed for 15 s and air-dried for 5 s.

• Group C: Same as group B, but a 9.6% hydrofluoric acid (Pulpdent Co., Watertown, USA) was used.

• Group D: Dentin was etched for 15 s with 3% hydrofluoric acid, rinsed for 15 s and air-dried for 5 s. Then dentin was etched for 15 s with 37.5% phosphoric acid, rinsed for 15 s and air-dried for 5 s.

• Group E: Same as group D, but a 9.6% hydrofluoric acid was used.

After etching, the three-step etch-and-rinse adhesive (OptiBond FL, Kerr) was applied to the dentin surface according to manufacturers’ instructions. Using a silicon mould, a composite (Clearfil AP-X, Kuraray Co., Osaka, Japan) build-up, 5 mm in height, was made in two horizontal increments, each separately polymerized for 20 s. After removing the mould, the build-up was post-cured from buccal and lingual for 20 s each. All light-curing was performed using a LED curing unit (The Cure, Spring Health Products, Norristown, USA; light intensity: >600 mW/cm ² ).

2.2

μTBS-testing

After 1 week of storage in a 0.5% chloramine solution at room temperature, the teeth were sectioned perpendicular to the bonding surface using a diamond saw at slow-speed (Isomet 1000, Buehler Ltd, Lake Bluff, IL, USA) under continuous water cooling to obtain beams of 1.8 mm × 1.8 mm wide. The four central specimens were mounted in the pin-chuck of the MicroSpecimen Former and trimmed at the dentin–composite interface into an hour-glass shape with a bonding surface of approximately 1 mm ² using a cylindrical fine-grit diamond bur (5835KREF, Komet, Lemgo, Germany) in a high-speed turbine under air/water coolant. Specimens were then fixed to a Ciucchi’s jig with a cyanoacrylate glue (Model Repair II Blue, Dentsply-Sankin, Tochigi, Japan) and stressed at a crosshead speed of 1 mm/min until failure using a universal testing device (LRX, Lloyd, Hampshire, UK) with a load cell of 100 N. μTBS is expressed in MPa as derived from dividing the imposed force (N) at the time of fracture by the bond area (mm ² ). The occurrence of failure prior to the actual testing was included in the calculation of the mean μTBS as 0 MPa. The highest and the lowest value in each group were excluded. Failure mode was determined at a magnification of 50× using the stereomicroscope (Wild M5A), and recorded as adhesive at interface (interfacial failure), cohesive in resin (including failures within the adhesive layer and/or composite) or cohesive in dentin. One-way analysis of variance (ANOVA) and post hoc Tukey’s HSD multiple comparisons were used to determine statistical differences in μTBS ( p < 0.05).

2.3

Feg-SEM fracture analysis

After the initial failure analysis with the stereomicroscope, representative specimens were processed for field-emission-gun scanning electron microscopy (Feg-SEM; Philips XL30, Eindhoven, The Netherlands), using common electron microscopic specimen processing techniques including fixation, dehydration, chemical drying, and gold-sputter coating .

2.4

TEM interfacial analysis

For each group, two additional dentin surfaces, prepared similarly as for μTBS-testing, were analyzed with TEM using a standardized preparation protocol, including fixation, dehydration, embedding and diamond-knife ultra-microtomy . Non-demineralized and lab-demineralized (10% formaldehyde-formic acid for 36 h) ultra-thin sections were cut (Ultracut UCT, Leica, Vienna, Austria) and examined unstained and positively stained (5% uranyl acetate for 12 min/saturated lead citrate for 13 min) using TEM (JEM-1200EX II, JEOL, Tokyo, Japan). In order to reveal potential defects, additional specimens immersed in a 50 wt% ammoniac silver nitrate solution, were prepared according to a nanoleakage detection protocol .

2.1

Specimen preparation

Fifteen non-carious human lower third molars were stored in a 0.5% chloramine solution after extraction. All teeth were mounted in gypsum blocks to facilitate specimen manipulation and the occlusal third of the crown was removed using a slow-speed diamond saw (Isomet 1000, Buehler, Lake Bluff, IL, USA). Dentin surfaces were checked with a stereomicroscope (Wild M5A, Heerbrugg, Switzerland) for absence of enamel and pulp tissue. On each tooth, a standard smear layer was produced using a high-speed medium-grit diamond bur (100 μm, No.: 842314014, Komet, Lemgo, Germany) mounted in the MicroSpecimen Former (The University of Iowa, Iowa City, IA, USA). Three dentin surfaces were exposed to each of the five different etching procedures:

• Group A: Dentin was etched for 15 s with 37.5% phosphoric acid (Kerr Gel, Orange, CA, USA), rinsed for 15 s and air-dried for 5 s (Control).

• Group B: Dentin was etched for 15 s with 37.5% phosphoric acid, rinsed for 15 s and air-dried for 5 s. Subsequently, dentin was etched for 15 s with 3% hydrofluoric acid (DenMat, Santa Maria, CA, USA), rinsed for 15 s and air-dried for 5 s.

• Group C: Same as group B, but a 9.6% hydrofluoric acid (Pulpdent Co., Watertown, USA) was used.

• Group D: Dentin was etched for 15 s with 3% hydrofluoric acid, rinsed for 15 s and air-dried for 5 s. Then dentin was etched for 15 s with 37.5% phosphoric acid, rinsed for 15 s and air-dried for 5 s.

• Group E: Same as group D, but a 9.6% hydrofluoric acid was used.

After etching, the three-step etch-and-rinse adhesive (OptiBond FL, Kerr) was applied to the dentin surface according to manufacturers’ instructions. Using a silicon mould, a composite (Clearfil AP-X, Kuraray Co., Osaka, Japan) build-up, 5 mm in height, was made in two horizontal increments, each separately polymerized for 20 s. After removing the mould, the build-up was post-cured from buccal and lingual for 20 s each. All light-curing was performed using a LED curing unit (The Cure, Spring Health Products, Norristown, USA; light intensity: >600 mW/cm ² ).

2.2

μTBS-testing

After 1 week of storage in a 0.5% chloramine solution at room temperature, the teeth were sectioned perpendicular to the bonding surface using a diamond saw at slow-speed (Isomet 1000, Buehler Ltd, Lake Bluff, IL, USA) under continuous water cooling to obtain beams of 1.8 mm × 1.8 mm wide. The four central specimens were mounted in the pin-chuck of the MicroSpecimen Former and trimmed at the dentin–composite interface into an hour-glass shape with a bonding surface of approximately 1 mm ² using a cylindrical fine-grit diamond bur (5835KREF, Komet, Lemgo, Germany) in a high-speed turbine under air/water coolant. Specimens were then fixed to a Ciucchi’s jig with a cyanoacrylate glue (Model Repair II Blue, Dentsply-Sankin, Tochigi, Japan) and stressed at a crosshead speed of 1 mm/min until failure using a universal testing device (LRX, Lloyd, Hampshire, UK) with a load cell of 100 N. μTBS is expressed in MPa as derived from dividing the imposed force (N) at the time of fracture by the bond area (mm ² ). The occurrence of failure prior to the actual testing was included in the calculation of the mean μTBS as 0 MPa. The highest and the lowest value in each group were excluded. Failure mode was determined at a magnification of 50× using the stereomicroscope (Wild M5A), and recorded as adhesive at interface (interfacial failure), cohesive in resin (including failures within the adhesive layer and/or composite) or cohesive in dentin. One-way analysis of variance (ANOVA) and post hoc Tukey’s HSD multiple comparisons were used to determine statistical differences in μTBS ( p < 0.05).

2.3

Feg-SEM fracture analysis

After the initial failure analysis with the stereomicroscope, representative specimens were processed for field-emission-gun scanning electron microscopy (Feg-SEM; Philips XL30, Eindhoven, The Netherlands), using common electron microscopic specimen processing techniques including fixation, dehydration, chemical drying, and gold-sputter coating .

2.4

TEM interfacial analysis

For each group, two additional dentin surfaces, prepared similarly as for μTBS-testing, were analyzed with TEM using a standardized preparation protocol, including fixation, dehydration, embedding and diamond-knife ultra-microtomy . Non-demineralized and lab-demineralized (10% formaldehyde-formic acid for 36 h) ultra-thin sections were cut (Ultracut UCT, Leica, Vienna, Austria) and examined unstained and positively stained (5% uranyl acetate for 12 min/saturated lead citrate for 13 min) using TEM (JEM-1200EX II, JEOL, Tokyo, Japan). In order to reveal potential defects, additional specimens immersed in a 50 wt% ammoniac silver nitrate solution, were prepared according to a nanoleakage detection protocol .