Journal of the Mechanical Behavior of Biomedical Materials 74 (2017) 251–260

Contents lists available at ScienceDirect

Journal of the Mechanical Behavior of

Biomedical Materials
journal homepage: www.elsevier.com/locate/jmbbm

Fracture, roughness and phase transformation in CAD/CAM milling and MARK

subsequent surface treatments of lithium metasilicate/disilicate glass-
ceramics
Abdur-Rasheed Alaoa, Richard Stollb, Xiao-Fei Songc, John R. Abbottd, Yu Zhange, Jaafar Abduof,
⁎
Ling Yina,
a
Mechanical Engineering, College of Science and Engineering, James Cook University, Townsville QLD 4811, Australia
b
Restorative Dentistry, College of Medicine and Dentistry, James Cook University, Cairns QLD 4870, Australia
c
Key Laboratory of Advanced Ceramics and Machining Technology of Ministry of Education, School of Mechanical Engineering, Tianjin University, Tianjin 300072, China
d
JCU Dental, James Cook University, Cairns QLD 4870, Australia
e
Department of Biomaterials and Biomimetics, New York University College of Dentistry, New York, NY 10010, USA
f
Melbourne Dental School, Melbourne University, 720 Swanston Street, Melbourne, Victoria 3010, Australia

A R T I C L E I N F O A B S T R A C T

Keywords: This paper studied surface fracture, roughness and morphology, phase transformations, and material removal
CAD/CAM milling mechanisms of lithium metasilicate/disilicate glass ceramics (LMGC/LDGC) in CAD/CAM-milling and sub-
Fracture sequent surface treatments. LMGC (IPS e.max CAD) blocks were milled using a chairside dental CAD/CAM
Lithium metasilicate/disilicate glass-ceramics milling unit and then treated in sintering, polishing and glazing processes. X-ray diffraction was performed on all
Material removal mechanisms
processed surfaces. Scanning electron microscopy (SEM) was applied to analyse surface fracture and mor-
Roughness
phology. Surface roughness was quantitatively characterized by the arithmetic average surface roughness Ra and
Surface treatments
the maximum roughness Rz using desktop SEM-assisted morphology analytical software. The CAD/CAM milling
induced extensive brittle cracks and crystal pulverization on LMGC surfaces, which indicate that the dominant
removal mechanism was the fracture mode. Polishing and sintering of the milled LMGC lowered the surface
roughness (ANOVA, p < 0.05), respectively, while sintering also fully transformed the weak LMGC to the strong
LDGC. However, polishing and glazing of LDGC did not significantly improve the roughness (ANOVA, p > 0.05).
In comparison of all applied fabrication process routes, it is found that CAD/CAM milling followed by polishing
and sintering produced the smoothest surface with Ra = 0.12 ± 0.08 µm and Rz = 0.89 ± 0.26 µm. Thus, it is
proposed as the optimized process route for LMGC/LDGC in dental restorations. This route enables to manu-
facture LMGC/LDGC restorations with cost effectiveness, time efficiency, and improved surface quality for better
occlusal functions and reduced bacterial plaque accumulation.

1. Introduction Denry and Holloway, 2010; Höland et al., 2006a; Kelly, 2008). These
differences result in compressive stresses in LDGC, which can deflect
Monolithic ceramic crowns and bridges are proven to be more advancing cracks (Apel et al., 2008; Denry, 2013; Denry and Holloway,
durable than veneered core restorations where brittle fractures fre- 2010; Serbena and Zanotto, 2012).
quently occur in the weak porcelain veneers and the veneer-core in- Due to the high strength of LDGC and the brittleness of its glassy
terfaces (Beuer et al., 2009; Guess et al., 2010; Swain, 2009; Zhang phase, it is very difficult to machine using chairside or laboratory CAD/
et al., 2009; 2013a). Ideal ceramic restorations should be made from CAM milling systems. Alternatively, LDGC restorations are made from
durable and highly aesthetic materials, such as lithium disilicate low-strength lithium metasilicate (Li2SiO3) glass ceramic (LMGC)
(Li2Si2O5) glass ceramics (LDGC) (Reich et al., 2014). The high strength blocks, which can be easily CAD/CAM-milled to form basic full-contour
and toughness of LDGC arise from ~70 vol% of interlocking needle-like crowns and bridges (Höland et al., 2000). Meanwhile, milled LMGC
lithium disilicate crystals, which have different thermal coefficients and blocks require sintering to transform lithium metasilicate to lithium
elastic moduli from their glassy matrix (Apel et al., 2008; Denry, 2013; disilicate for formation of strong LDGC. However, the milling process

⁎
Corresponding author.
E-mail address: ling.yin@jcu.edu.au (L. Yin).

http://dx.doi.org/10.1016/j.jmbbm.2017.06.015
Received 19 May 2017; Received in revised form 9 June 2017; Accepted 12 June 2017
Available online 15 June 2017
1751-6161/ © 2017 Elsevier Ltd. All rights reserved.
A.-R. Alao et al. Journal of the Mechanical Behavior of Biomedical Materials 74 (2017) 251–260

induces surface and subsurface flaws in LMGC, which are difficult to fabricated by the manufacturer via melting a base glass consisting of
diminish by the subsequent heat treatment, and may compromise the 69.3 wt% SiO2, 15.4 wt% Li2O, 6.05 wt% K2O, 4 wt% ZnO2, 3.38 wt%
strength of LDGC restorations and shorten their lifespans (Denry, 2013; Al2O3, and 3.84 wt% P2O5 at 1450 °C (El-Meliegy and van Noort, 2012;
Rekow et al., 2011; Rekow and Thompson, 2005). Thus, minimization Höland et al., 2006a). This was followed by subsequent annealing at
of milling-induced flaws in LMGC is necessary for quality assurance for 480 °C for 1 h to precipitate lithium metasilicate crystals (El-Meliegy
LDGC restorations. and van Noort, 2012). After cooling to room temperature, the glass
Further, polishing and glazing are also applied to finalize surface ceramic contains approximately 40 vol%, 0.5–1 µm lithium metasilicate
texture, reduce roughness and enhance light reflection (Boaventura crystals (Bühler-Zemp and Völkel, 2005; El-Meliegy and van Noort,
et al., 2013; Höland et al., 2006b). In fact, at many dental laboratories 2012). It has the biaxial strength of 130 ± 30 MPa, fracture toughness
or clinics, sintering, polishing and glazing of CAD/CAM-milled LMGC of 1 ± 0.1 MPa m1/2 and Vickers hardness of 5.4 ± 0.1 GPa (Bühler-
contours are arbitrary procedures (Lin et al., 2012), which can result in Zemp and Völkel, 2005).
variable surface quality. Currently, there is lack of optimized fabrica-
tion process selection and sequence to ensure the reliability of LDGC 2.2. Chairside CAD/CAM milling
restorations.
Surface quality, such as phase transformation, surface roughness LMGC blocks were milled using a chairside CAD/CAM milling unit
and fracture morphology, plays a critical role in determining the wear (CEREC MC XL, Sirona, Germany) with a step bur 12S (Ref 6240167,
and fatigue performance of dental restorations (Alao et al., 2017; Sirona, Germany) and a cylindrical pointed bur 12S (Ref 6240159,
Currana et al., 2017; Denry, 2013; Peng et al., 2016; Rekow and Sirona, Germany), both of which have the same composition and
Thompson, 2005; Rekow et al., 2011; Ulutan and Ozel, 2011; Zhang properties. The step bur consists of three cutting faces with lengths of
et al. 2013b). Clinical studies have shown that CAD/CAM-processed 3 mm, 3 mm, and 6 mm, and diameters of 2.1 mm, 1.7 mm and 1.3 mm,
single LDGC restorations achieved 100% cumulative survival rate up to respectively. The cylindrical pointed bur comprises of two cutting faces
two years (Fasbinder et al., 2010) and 96.3% after four years according with lengths of 4 mm and 8 mm and diameters of 2.1 mm and 1.8 mm,
to the Kaplan-Meier survival analysis (Reich and Schierz, 2013). The respectively. Both burs are electro-plated with diamond abrasives, and
survival rate for three-unit LDGC partial fixed dentures was 93% up to are used to generate flat surfaces as schematically shown in Fig. 1. Wet
four years (Reich et al., 2014) and 87.9% for up to ten years (Kern et al., milling was conducted following the program recommended by the
2012). A five-year clinical study indicates that nearly 100% survival manufacturer, which simulates surface milling of crowns, the most
rate for LDGC crowns but 70% for fixed partial LDGC dentures challenging step in the CAD/CAM process (Luthardt et al., 2004). A new
(Marquardt and Strub, 2006). The longest clinical observation of LDGC step bur was gold-coated and observed using scanning electron micro-
posterior crowns after 15 years reveals 81.9% survival rate (van den scopy (SEM) (Jeol JSM5410V, Japan). Fig. 2(a) shows the SEM mi-
Breemer et al., 2017). Clinical analyses of failed LDGC restorations have crograph of the step diamond bur morphology. Fig. 2(b) reveals the
found that fracture and chipping were the root cause of failure, which diamond cutting edges with an average grit size of approximately
originated from surface damage and flaws (Della Bona and Kelly, 2008; 50–60 µm.
Mores et al., 2017; Valenti and Valenti, 2009; van den Breemer et al.,
2017; Zhang et al., 2013b). Therefore, the diminishment of surface 2.3. Surface process protocols
flaws in LDGC restorations is essential to prolong their lifespans. In
addition, surface quality also critically affects cell adhesion, prolifera- After milling of LMGC blocks, the samples were cleaned in acetone
tion and protein adsorption (Brunot-Gohin et al., 2013). and treated by sintering, polishing and glazing to simulate various
In clinical practice, external surfaces of restorations must be fin- clinical fabrication processes. These process routes are schematically
ished to a high surface luster to reduce fracture risk, bacterial plaque shown in Fig. 3 and designated as CAD/CAM (i.e., CAD/CAM milling),
accumulation, tooth stains, and wear on antagonist/adjacent teeth (De CAD/CAM-polish, CAD/CAM-sinter, CAD/CAM-polish-sinter, CAD/
Jager et al., 2000; Jefferies, 2007; Kou et al., 2006; Steiner et al., 2015; CAM-sinter-polish, CAD/CAM-sinter-glaze, and CAD/CAM-polish-
Whitehead et al., 1995). Intaglio surfaces of restorations are often sinter-glaze processes.
roughened to improve bonding to adhesives (Brunot-Gohin et al., Sintering of milled LMGC samples was carried out in a programed
2013). The surface quality of ground LDGC using dental handpieces and dental furnace (P300, Ivoclar Vivadent, Liechtenstein) at a stand-by
diamond burs (Song et al., 2016), and glazed, polished and ground temperature of 403 °C. Then, the samples were heated to 770 °C at a
LDGC (Boaventura et al., 2013; Kou et al., 2006; Tholt et al., 2006) have heating rate of 60 °C/min and held at the temperature for 10 min. After
been individually studied. However, it is unclear how the process se- that, they were heated again to 850 °C at a heating rate of 30 °C/min
lection and sequence will influence the surface quality of LMGC/LDGC and held for another 10 min before cooling to 700 °C. Finally, they were
in CAD/CAM milling, and subsequent sintering, glazing and polishing.
This paper, therefore, aimed to investigate the process-quality re-
lation to determine the optimized processing protocol for LMGC/LDGC.
X-ray diffraction was used to analyse crystalline phases and phase
transformations. Surface roughness was measured in terms of the ar-
ithmetic average roughness Ra and the maximum roughness Rz using a
desk-top SEM-assisted morphology analytical software. Scanning elec-
tron microscopy (SEM) was applied to analyse the material removal
mechanisms, surface fracture and morphology. Finally, an optimal
fabrication process for LDGC restorations was proposed to achieve the
improved surface integrity.

2. Experimental procedure

2.1. Materials

LMGC blocks of 14.5 mm × 12.4 mm × 18 mm (IPS e.max CAD,

Ivoclar Vivadent, Liechtenstein) were selected. The material is Fig. 1. Chairside CAD/CAM milling of a LMGC block using two diamond burs.

252
A.-R. Alao et al. Journal of the Mechanical Behavior of Biomedical Materials 74 (2017) 251–260

Fig. 2. SEM micrographs showing (a) the morphology of a step diamond bur 12S, (b) diamond abrasives on the bur with approximately 64-µm grit size.

Fig. 3. Schematic diagram of all fabrication process routes.

cooled to room temperature.

Dry polishing was conducted manually using a clinical dental
handpiece with a grey white rubber diamond bur (Exa Cerapol UM, ISO
658.900.114.525.060, Edenta, Switzerland) operated by a dental clin-
ician. The bur contains dispersed diamond abrasives embedded in a
softer elastic matrix (Jefferies, 2007) and is used for intermediate fin- Fig. 4. SEM micrograph of the morphology of the grey white rubber diamond bur with
ishing, eliminating scratches and smoothening surfaces as re- approximately 60–70-µm grit size.
commended by the manufacturer. Fig. 4 shows the SEM micrograph of
the grey-white bur morphology, revealing the average diamond grit size characterized to obtain the arithmetic average roughness, Ra, and the
of approximately 60–70 µm. Surfaces were polished at a force of ap- maximum roughness, Rz (De Chiffre et al., 2000) using desktop SEM-
proximately 1 N with the bur rotating at 5000 rpm to obtain clinically assisted morphology analytical software (Phenom-World BV, the
acceptable surface quality. Netherlands). The software draws lines perpendicular to machining
Glazing was conducted by firing samples without glass beads in the traces and calculates a profile for each measurement line. Fig. 5(a)
programed dental furnace at a stand-by temperature of 403 °C. Then the shows the 3D surface roughness measurement process where mea-
samples were reheated to 820 °C at a heating rate 90 °C/min and held surements were made across three profiles and their average values
for 10 min. Next, they were heated to 840 °C at 30 °C/min heating rate computed. The cut-off and evaluation lengths were 0.2 mm and 1 mm
and held for 7 min. Finally, the samples were cooled to 700 °C and then respectively. Fig. 5(b) shows the 2D surface morphology of Fig. 5(a). On
to room temperature. each surface, three roughness measurements were performed to obtain
mean values and standard deviations. Each processed surface was also
2.4. Surface characterization viewed under SEM (JEOL JSM 5410LV, Tokyo, Japan) to study surface
fracture, morphology and removal mechanisms.
X-ray diffraction was performed to all processed surfaces using an x-
ray diffractometer (D/MAX-2500, Rigaku, Japan) with a Cu Kα radia- 2.5. Statistical analysis
tion of 0.15406 nm wave length. The scanning rate employed was
4° min−1 over the incident angles (2θ) in the range of 10–80° at 0.02° One-way analysis of variance (ANOVA) was conducted at 95%
scanning step. confidence interval to examine the significance of fabrication processes
All processed surfaces were gold-coated and quantitatively on the average surface roughness, Ra, and the maximum height, Rz.

253
A.-R. Alao et al. Journal of the Mechanical Behavior of Biomedical Materials 74 (2017) 251–260

Fig. 5. A demonstration of surface roughness measurement using the desktop SEM-as-

sisted morphology analytical software, in which (a) 3D surface roughness measurement
with three profiles to be analyzed and (b) 2D surface morphology of (a).

Paired t-test was also performed at 95% confidence to examine the

influence of each surface treatment on surface roughness values.

3. Results

3.1. X-ray diffraction

Fig. 6(a) shows the approximately identical x-ray diffraction pattern

for the LMGC surfaces produced in the CAD/CAM and CAD/CAM-polish
processes, which is similar to the pattern for the original LMGC surface Fig. 6. (a) Approximately identical x-ray diffraction pattern for the LMGC surfaces pro-
from the manufacturer (as a reference). All peaks were identified as the duced in the CAD/CAM and CAD/CAM-polish processes and for the original LMGC sur-
face from the manufacturer (as a reference), (b) Approximately identical x-ray diffraction
lithium catena-silicate (Li2SiO3) or lithium metasilicate crystalline
pattern for the surfaces produced in the CAD/CAM-polish-sinter, CAD/CAM-sinter, and
phase, matching the published patterns for LMGC in Alemi et al. (2013) CAD/CAM-sinter-polish processes, (c) Approximately identical x-ray diffraction pattern
and the Joint Committee on Powder Diffraction Standards (JCPDS)-file for the surfaces produced in the CAD/CAM-polish-sinter-glaze and CAD/CAM-sinter-glaze
No. 30–0766 for lithium metasilicate crystals (Thieme and Rüssel, processes.
2015).
Fig. 6(b) shows the approximately identical x-ray diffraction pattern 3.2. Surface Roughness
for the surfaces produced in the CAD/CAM-polish-sinter, CAD/CAM-
sinter, and CAD/CAM-sinter-polish processes. All main peaks were Fig. 7 shows the results of the arithmetic average roughness, Ra, and
found identical to the theoretical pattern of lithium disilicate crystals the maximum roughness, Rz, for different processes and sequences. The
published in the JCPDS-file No. 82–2396 (Thieme and Rüssel, 2015). smoothest surface with Ra 0.12 ± 0.08 µm and Rz 0.89 ± 0.26 µm was
This indicates that LMGC was transformed to LDGC in all three surfaces. produced in the CAD/CAM-polish-sinter process while the roughest
Fig. 6(c) shows the approximately identical x-ray diffraction pattern surface with Ra 1.11 ± 0.09 µm and Rz 6.46 ± 0.61 µm was generated
for the surfaces produced in the CAD/CAM-polish-sinter-glaze and in the CAD/CAM- milling process. The moderate roughness with Ra
CAD/CAM-sinter-glaze processes. These patterns are similar to Fig. 6(b) 0.76 ± 0.08 µm and Rz 4.03 ± 0.27 µm was yielded in the CAD/CAM-
and the theoretical pattern of lithium disilicate crystals in the JCPDS- polish process, which is similar to the surface with Ra 0.70 ± 0.1 µm
file No. 82–2396 (Thieme and Rüssel, 2015), indicating that the phase and Rz 3.75 ± 0.09 µm obtained in the CAD/CAM-sinter-glaze process.
transformation from LMGC to LDGC also occurred in these surfaces.

254
A.-R. Alao et al. Journal of the Mechanical Behavior of Biomedical Materials 74 (2017) 251–260

Fig. 7. Arithmetic mean roughness, Ra, and maximum roughness, Rz, versus different
process routes. Each data point is the mean value of three profiles on each processed
surface; the error bars correspond to ± one standard deviation for the three profiles.

The improved surfaces were generated in the CAD/CAM-sinter process

with Ra 0.58 ± 0.07 µm and Rz 0.89 ± 0.26 µm. It was followed by the
CAD/CAM-sinter-polish process with Ra 0.42 ± 0.12 µm and Rz
2.54 ± 0.88 µm and the CAD/CAM-polish-sinter-glaze process with Ra
0.25 ± 0.01 µm and Rz 1.55 ± 0.08 µm.
Both the average roughness Ra and the maximum roughness Rz were
significantly influenced by fabrication processes (p < 0.05). Table 1
summarizes all p-values of paired t-tests for all fabrication processes.
Both polishing and sintering significantly improved the roughness Ra
and Rz values obtained in milling (p < 0.05). Further sintering of the
polished surface also significantly reduced both Ra and Rz values
(p < 0.05). Further polishing of the sintered surface only significantly Fig. 8. SEM micrographs of the LMGC surface produced in the CAD/CAM process re-
improved the Ra values (p < 0.05) but did not significantly change Rz vealing (a) milling traces, fractures and shear band-induced pile-ups leading to the ma-
values (p > 0.05). Glazing did not significantly improve either Ra terial pulverization, (b) details of surface fractures, and (c) details of pulverized/smeared
areas.

Table 1 (p > 0.05) or Rz (p > 0.05) obtained in the sintering and the polishing-
Summary of the p-values of all paired t-tests for Ra and Rz values of all paired fabrication sintering processes.
processes.

Surface treatments p-Value

3.3. Surface morphology
Ra (µm) Rz (µm)
Fig. 8 shows the CAD/CAM-milled LMGC surface morphology.
CAD/CAMed versus CAD/CAMed-polished p < 0.05 p < 0.05 Machining grooves and traces scratched by diamond abrasives are de-
CAD/CAMed-versus CAD/CAMed-sintered p < 0.05 p < 0.05
CAD/CAMed-polished versus CAD/CAMed-polished- p < 0.05 p < 0.05
tailed in Fig. 8(a). Localized irregular fractures and cracks, and pul-
sintered verized/smeared areas caused by shear bands are shown in Fig. 8(b).
CAD/CAMed-sintered versus CAD/CAMed-sintered- p < 0.05 p > 0.05 Pulverized or smeared areas with accumulated microchips on the highly
polished fractured areas are revealed in Fig. 8(c). Fig. 9 shows the LMGC surface
CAD/CAMed-sintered versus CAD/CAMed-sintered-glazed p > 0.05 p > 0.05
morphology produced in the CAD/CAM-polish process. Fig. 9(a) shows
CAD/CAMed-polished-sintered versus CAD/CAMed- p > 0.05 p > 0.05
polished-sintered-glazed that polishing smoothened milling traces and scratches, and produced
residual pulverized debris and glassy networks. Fig. 9(b) details the

255
A.-R. Alao et al. Journal of the Mechanical Behavior of Biomedical Materials 74 (2017) 251–260

Fig. 9. SEM micrographs of the LMGC surface produced in the CAD/CAM-polish process Fig. 10. SEM micrographs of the LDGC surface produced in the CAD/CAM-sinter process,
revealing (a) smootherned milling traces and residual debris, and (b) details of glassy showing (a) the bulging of the milled surface and (b) details of the bulged surface.
networks.
4. Discussion
glassy networks.
Fig. 10 shows the LDGC surface morphology produced in the CAD/ This study focused on the phase transformation, roughness and
CAM-sinter process, on which the large scale fractures on the milled surface morphology in CAD/CAM-milling, sintering, polishing and
surface in Fig. 8 disappeared. Fig. 10(a) shows the bulging of the milled glazing of LMGC and LDGC for dental restorations.
surface due to the sintering-induced phase transformation from LMGC The identical x-ray diffraction patterns (Fig. 6(a)) on the surfaces
to LDGC, resulting in significantly improved surface roughness than the produced in CAD/CAM and CAD/CAM-polish processes indicate that
CAD/CAM-milled surface in Fig. 8. Details of the bulged surface in simply milling and subsequent polishing did not cause any phase
Fig. 10(b) reveal microcracks, which indicate that the sintering cannot transformation of the original main crystal phase of lithium metasili-
completely heal milling-induced surface damage. cate. The identical x-ray diffraction patterns (Fig. 6(b) and (c)) on the
Fig. 11 shows the LDGC surface morphology produced in the CAD/ surfaces produced in CAD/CAM-polish-sinter, CAD/CAM-sinter, CAD/
CAM-sinter-glaze process. Fig. 11(a) shows that the glazing has resulted CAM-sinter-polish, CAD/CAM-polish-sinter-glaze and CAD/CAM-sinter-
in the recovery of the bulged surface produced in the CAD/CAM-sinter glaze processes prove that only sintering can transform lithium meta-
process in Fig. 10, and made the surface look smoother. Fig. 11(b) re- silicate to lithium disilicate. Glazing did not cause new phase trans-
veals the residual pulverized debris in the form of very fine particles at formations, suggesting that sintering of LMGC produced highly stabi-
the micro or submicro scales. lized lithium disilicate crystals in LDGC, which has the improved
Fig. 12 shows the LDGC surface morphology produced in the CAD/ strength to 400 GPa (Denry and Holloway, 2010).
CAM-sinter-polish process. Polishing traces, scratches and fractures are The choice of average and maximum surface roughness parameters
visible in Fig. 12(a). Localized fractures, enlarged debris and smeared (Ra and Rz) in Fig. 7 was based on their frequent use as surface texture
areas are detailed in Fig. 12(b). These indicate the polishing after sin- characterization parameters that allow comparison with previously
tering did not significantly improve surface morphology. published work (De Chiffre et al., 2000). Different process routes
Fig. 13 shows the LDGC surface morphology fabricated in the CAD/ yielded a large range of roughness values with Ra 0.12–1.11 µm and Rz
CAM-polish-sinter process, which has the very smooth surface texture 0.89–6.46 µm. Such values are comparable with those of the ground-
without visible machining traces and phase-transformation induced polished surfaces of LDGC with Ra 0.7 µm (Kou et al., 2006) and glazed
crystal bulging. This may be attributable to the combined action of LDGC with Ra 0.64 ± 0.014 µm (Boaventura et al., 2013).
polishing and sintering processes on the milled CAD/CAMed surface. The surface fracture on the CAD/CAM-milled LMGC was featured by
Fig. 14 shows the LDGC surface morphology produced in the CAD/ the SEM micrographs in Fig. 8, which was associated with greatest
CAM-polish-sinter-glaze process. It was similar to the very fine surface surface roughness in Fig. 7. This can be attributed to the diamond pe-
produced in the CAD/CAM-polish-sinter process in Fig. 13. However, netrating, ploughing and scratching of LMGC surfaces in milling, re-
adhesion of fine particle debris was observed on the finally glazed sulting in lateral fracture and localized heat (Jefferies, 2007). Repeated
surface (Fig. 14(b)). surface fractures also resulted in the material pulverization and created

256
A.-R. Alao et al. Journal of the Mechanical Behavior of Biomedical Materials 74 (2017) 251–260

Fig. 11. SEM micrographs of the LDGC surface produced in the CAD/CAM-sinter-glaze Fig. 12. SEM micrographs of the LDGC surface produced in the CAD/CAM-sinter-polish
process demonstrating (a) residual pulverized debris and (b) details of these powdered process showing (a) machining traces, scratches and fractures and (b) details of localized
debris. fracture, enlarged debris and smeared areas.

smeared surface defects, which were the consequence of the inter- LMGC-LDGC phase transformation (Fig. 6). Sintering might have also
granular and/or transgranular microcracking. These microcracks ori- caused the relaxation of machining-induced mechanical residual
ginated from localized shear stress fields, which were superimposed by stresses (Zhang and Howes, 1994), which also contributed to the
hydrostatic compressive stresses in diamond-material contact zones smoothening of milled surfaces (Fig. 10). It might have also generated
(Zhang and Howes, 1994). The evidence of the transgranular micro- thermal stresses (Denry, 2013), that can restrict the thermal expansion
cracking was observed clearly as lithium metasilicate crystals frag- of milled surfaces and lead to the surface bulging (Fig. 10).
mented into micro-sized ceramic grains in Fig. 8(c). The pulverized The insignificantly different Rz but significantly different Ra values
surface in Fig. 8(a) emanated from surface and subsurface shear bands (Table 1) produced in CAD/CAM-sinter (Fig. 10) and CAD/CAM-sinter-
that were reported in our earlier investigation as the main deformation polish (Fig. 12) processes need an in-depth analysis. The maximum
in nanoindentation of LMGC (Alao and Yin, 2015). Shear bands can roughness Rz generally represents the largest peak-to-valley surface
form the plastic instability that localizes large shear strains in relatively defect size (Cook, 2006) and can be used as a measure of the subsurface
thin bands during the material deformation process (Greer et al., 2013). damage depth in ceramic grinding and lapping processes (Li et al.,
They can represent the material plastic deformation or serve as crack- 2008). Consequently, polishing after sintering was unable to remove
initiating sites (Sergueeva et al., 2005). In milling, shear bands might surface peaks and valleys of milling-induced ploughing grooves or
have initiated microcracking through extensions of subsurface shear scratches due to the high strength of sintering-transformed LDGC. In-
bands into underlying tensile matrices or from stress concentration sites stead, it only smoothened the average profile to decrease average Ra.
at intersections of shear bands (Alao and Yin, 2015). Thus, the ma- Therefore, polishing needs to be conducted prior to sintering. Other-
chining grooves in Fig. 8(a) might represent large fragmented shear wise, polishing of sintered surfaces requires the usage of high-stiffness
strains which melted unto the surface due to machining-induced high polishing media in multiple-step processes with finer abrasives to im-
temperatures, roughening the surface (Fig. 7). prove the roughness and transit the brittle-ductile removal (Schmidt
Polishing of CAD/CAM-milled LMGC resulted in significant reduc- and Weigl, 2000; Yin et al., 2007). Interactions of finer abrasives with
tions of both surface roughness Ra and Rz (Fig. 7, Table 1). This in- work materials are dominated by normal forces resulting from negative
dicates that polishing effectively smoothened milling traces and scrat- rake angles and small undeformed chip thicknesses in abrasive–ma-
ches of LMGC by abrasive wear mechanism (Fig. 9). The mechanical terial contact regions (Alao and Konneh, 2012; Evans et al., 2003).
energy was transferred from abrasives to the LMGC material, removing These high rake angles permit necessary hydrostatic pressures/tem-
or displacing the material (Evans et al., 2003; Jefferies, 2007). Mean- peratures to enable plastic deformation to occur (Evans et al., 2003).
while, polishing formed debris and produced glassy networks shown in However, these multiple procedures would remarkably increase man-
Fig. 9. ufacturing time and cost for restorations (Steiner et al., 2015).
Sintering of CAD/CAM-milled surfaces (Fig. 10) also significantly The CAD/CAM-polish-sinter process (Fig. 13) produced the lowest
reduced the Ra and Rz roughness (Fig. 7 and Table 1), ascribing to the roughness, in which polishing easily removed milling traces and

257
A.-R. Alao et al. Journal of the Mechanical Behavior of Biomedical Materials 74 (2017) 251–260

Fig. 13. SEM micrographs of the LDGC surface produced in the CAD/CAM-polish-sinter Fig. 14. SEM micrographs of the LDGC surface produced in the CAD/CAM-polish-sinter-
process revealing (a) a very smooth surface texture and (b) details of the smooth surface glaze process showing (a) a very fine surface similar to the surface in Fig. 13(a) and (b)
without visible defects. details of the surface with fine particle debris.

scratches of weak LMGC and the subsequent sintering complemented

the polishing further for roughness reduction and phase-transformed
strength increase. The yielded roughness (Ra = 0.12 ± 0.08 µm) was
lower than the threshold roughness (Ra = 0.20 µm) for bacterial plaque
retention (Bollen et al., 1997), suggesting that the bacteria accumula-
tion is unlikely to occur. Further, no second step polishing with finer
grits is needed for structure retention and shininess. Thus, this process
route appears to be the most practical and cost effective.
The subsequent glazing of the CAD/CAMed-polished-sintered sur-
faces (Fig. 14) could not further improve the roughness (Table 1 and
Fig. 7), which is in agreement with that some finishing techniques can
obtain surfaces equivalent to glaze-fired ceramic surfaces (Tholt et al.,
2006). Meanwhile, glazing at a high temperature might have induced
the thermal deformation in LDGC (Denry and Holloway, 2004), which
might not be detrimental to the mechanical strength (Cattell et al.,
2002; Denry and Holloway, 2004; Wen et al., 2007). Nevertheless,
glazing slightly increased the surface roughness above the bacterial
retention threshold (Bollen et al., 1997), implying an increased possi-
bility of bacteria accumulation. Thus, glazing should be eliminated to
polished-sintered surfaces. In addition, thermal stresses induced by Fig. 15. The optimized fabrication procedure for occlusal/facial/lingual and intaglio
glazing (Fig. 11) might have restricted the thermal expansion of bulged (cementation) surfaces in LDGC restorations.
sintered surfaces. This could have caused sufficient hydrostatic com-
pressive stresses to recompact surfaces, which generated powder debris LDGC restorations as schematically described in Fig. 15. The CAD/
to increase the roughness (Fig. 7). Thus, glazing improved neither CAM-milling of LMGC first generates rough restorative profiles. Pol-
roughness Ra nor Rz of CAD/CAMed-sintered surfaces (Table 1) because ishing is only conducted on occlusal, facial and lingual surfaces. Then,
of the bursting of bulged surfaces to powders. sintering is applied to strengthen the material. Intaglio surfaces pro-
Comparing all processes, the CAD/CAM-polish-sinter process route duced in the CAD/CAM-sinter process can be etched by hydrofluoric
proved to be the optimized procedure, meeting clinical dental standards acid to roughen surfaces for cementation (Borges et al., 2003; Kelly,
and eliminating secondary polishing and glazing for time and cost ef- 1997). It is likely that this optimized procedure may benefit both dental
fectiveness. Therefore, this route is proposed for the production of patients and clinicians who routinely use LDGC restorations and expect

258
A.-R. Alao et al. Journal of the Mechanical Behavior of Biomedical Materials 74 (2017) 251–260

them long-lasting. Given that the long-term (after 15 years) survival a review of the literature. Dent. Mater. 13, 258–269.
Borges, G.A., Sophr, A.M., de Goes, M.F., Sobrinho, L.C., Chan, D.C.N., 2003. Effect of
rate for LDGC crowns achieved only 81.9% (van den Breemer et al., etching and airborne abrasion on the microstructure of different dental ceramics. J.
2017), continuing efforts for improvement are needed. An important Prosthet. Dent. 89, 479–488.
thrust for these efforts is to foster a more analytical engineering ap- Brunot-Gohin, C., Duval, J.-L., Azogui, E.-E., Jannetta, R., Pezron, I., Laurent-Maquin, D.,
Gangloff, S.C., Egles, C., 2013. Soft tissue adhesion of polished versus glazed lithium
proach to clinical manufacturing of restorations. This study has pro- disilicate ceramic for dental applications. Dent. Mater. 29, e205–e212.
vided a rational understanding of the optimized clinical procedure for Bühler-Zemp, P., Völkel, T., 2005. IPS e.max CAD Scientific Documentation. Ivoclar
LDGC restorations which can be practically utilized in restorative Vivadentpp. 1–16.
Cattell, M.J., Palumbo, R.P., Knowles, J.C., Clarke, R.L., Samarawickrama, D.Y.D., 2002.
dentistry. The effect of veneering and heat treatment on the flexural strength of Empress® 2
ceramics. J. Dent. 30, 161–169.
5. Conclusions Cook, R.F., 2006. Strength and sharp contact fracture of silicon. J. Mater. Sci. 41,
841–872.
Currana, P., Cattani-Lorenteb, M., Wiskottb, H.W.A., Durualb, S., Scherrerb, S.S., 2017.
This work investigated the surface fracture, roughness and mor- Grinding damage assessment for CAD-CAM restorative materials. Dent. Mater. 33,
phology, phase transformations, and material removal mechanisms of 294–308.
LMGC/LDGC produced in CAD/CAM milling, sintering, polishing and De Chiffre, L., Lonardo, P., Trumpold, H., Lucca, D.A., Goch, G., Brown, C.A., Raja, J.,
Hansen, H.N., 2000. Quantitative characterization of surface texture. CIRP Ann.-
glazing. The CAD/CAM-milling process produced the coarsest surface Manuf. Technol. 49, 635–652.
roughness due to extensive brittle cracking and pulverization. Polishing Jager, De, Feilzer, A.J., Davidson, C.L., 2000. The influence of surface roughness on
of CAD/CAM-milled LMGC efficiently improved roughness by wearing porcelain strength. Dent. Mater. 16, 381–388.
Della Bona, A., Kelly, J.R., 2008. The clinical success of all-ceramic restorations. J. Am.
the milled traces and scratches through abrasive mechanical actions. Dent. Assoc. 139 (9 suppl), 8S–13S.
Sintering transformed LMGC to LDGC with increased strength. The Denry, I., 2013. How and when does fabrication damage adversely affect the clinical
subsequent glazing did not change the stabilized LDGC but might have performance of ceramic restorations? Dent. Mater. 29, 85–96.
Denry, I., Holloway, J.A., 2004. Effect of post-processing heat treatment on the fracture
induced additional thermal stresses leading to further deformation. strength of a heat-pressed dental ceramic. J. Biomed. Mater. Res. Part B: Appl.
Polishing of LDGC did not significantly reduce surface roughness be- Biomater. 68B, 174–179.
cause the high-strength LDGC makes the material difficult to machine. Denry, I., Holloway, J.A., 2010. Ceramics for dental applications: a review. Materials 3,
351–368.
The CAD/CAM-polish-sinter process produced the best roughness,
El-Meliegy, E., van Noort, R., 2012. Glasses and Glass Ceramics for Medical Applications.
which fulfilled clinical requirements for cost effectiveness, time effi- Springer, New York.
ciency, and surface quality for occlusal functions and bacterial plaque Evans, C.J., Paul, E., Dornfeld, D., Lucca, D.A., Byrne, G., Tricard, M., Klocke, F., Dambon,
O., Mullany, B.A., 2003. Material removal mechanisms in lapping and polishing.
retention. Further studies will be conducted to determine the reliability
CIRP Ann.-Manuf. Technol. 52, 611–633.
of the proposed process route for LDGC restorations and establish the Fasbinder, D.J., Dennison, J.B., Heys, D., Neiva, G., 2010. A clinical evaluation of
relations between their surface integrity and fatigue behavior. chairside lithium disilicate CAD/CAM crowns: a two-year report. J. Am. Dent. Assoc.
141, 10s–14s.
Greer, A.L., Cheng, Y.Q., Ma, E., 2013. Shear bands in metallic glasses. Mater. Sci. Eng. R:
Acknowledgements Rep. 74, 71–132.
Guess, P.C., Zhang, Y., Kim, J.-W., Rekow, E.D., Thompson, V.P., 2010. Damage and
The authors thank Drs. Shane Askew and Kevin Blake of the reliability of Y-TZP after cementation surface treatment. J. Dent. Res. 89, 592–596.
Höland, W., Apel, E., van ‘t Hoen, Ch, Rheinberger, V., 2006a. Studies of crystal phase
Advanced Analytical Center and Mr. Kevin Chong of College of formations in high-strength lithium disilicate glass–ceramics. J. Non-Cryst. Solid 352,
Medicine & Dentistry at James Cook University (JCU) for experimental 4041–4050.
assistance; Mr. Jim Ruddy of Ivoclar Vivadent, Australia for providing Höland, W., Schweiger, M., Frank, M., Rheinberger, V., 2000. A comparison of the mi-
crostructure and properties of the IPS Empress 2 and IPS Empress glass-ceramics. J.
e.max CAD blocks; and Mr. Matthew Batty of Sirona, Australia for of- Biomed. Mater. Res. B Appl. Biomater. 53, 297–303.
fering CAD/CAM technical support. A.-R.A acknowledges the JCU PhD Höland, W., Rheinberger, V., Apel, E., van’t Hoen, C., Höland, M., Dommann, A., Obrecht,
Scholarship (JCU IPRS). The work was supported by the JCU M., Mauth, C., Graf-Hausner, U., 2006b. Clinical applications of glass-ceramics in
dentistry. J. Mater. Sci: Mater. Med. 17, 1037–1042.
Collaboration Grants Scheme; the Australia-China Science and Research Jefferies, S.R., 2007. Abrasive finishing and polishing in restorative dentistry: a state-of-
Fund from the Department of Industry, Innovation, Climate Change, the-art review. Dent. Clin. North Am. 51, 379–397.
Science, Research and Tertiary Education, Australia (Grant No. ACSRF Kelly, J.R., 1997. Ceramics in restorative and prosthetic dentistry. Ann. Rev. Mater. Sci.
27, 443–468.
GMB 12029); the National Natural Science Foundation of China (Grant
Kelly, J.R., 2008. Dental ceramics: what is this stuff anyway? J. Am. Dent. Assoc. 139,
No. 51375335); and the United States National Institutes of Health, 4s–7s.
National Institute of Dental and Craniofacial Research (Grant No. Kern, M., Sasse, M., Wolffart, S., 2012. Ten-year outcome of three-unit fixed dental
R01DE017925 and Grant No. R01DE026772). prostheses made from monolithic lithium disilicate ceramic. J. Am. Dent. Assoc. 143,
234–240.
Kou, W., Molin, M., Sjögren, G., 2006. Surface roughness of five different dental ceramic
References core materials after grinding and polishing. J. Oral Rehabil. 33, 117–124.
Li, S., Wang, Z., Wu, Y., 2008. Relationship between subsurface damage and surface
roughness of optical materials in grinding and lapping processes. J. Mater. Process.
Alao, A.-R., Konneh, M., 2012. Surface finish prediction models for precision grinding of Technol. 205, 34–41.
silicon. Int. J. Adv. Manuf. Technol. 58, 949–967. Lin, W.-S., Harris, B.T., Morton, D., 2012. Trial insertion procedure for milled lithium
Alao, A.-R., Stoll, R., Song, X.-F., Miyazaki, T., Hotta, Y., Shibata, Y., Yin, L., 2017. disilicate restorations in the precrystallized state. J. Prosthet. Dent. 107, 59–62.
Surface quality of yttria-stabilized tetragonal zirconia polycrystal in CAD/CAM mil- Luthardt, R.G., Holzhüter, M.S., Rudolph, H., Herold, V., Walter, M.H., 2004. CAD/CAM-
ling, sintering, polishing and sandblasting processes. J. Mech. Behav. Biomed. Mater. machining effects on Y-TZP zirconia. Dent. Mater. 20, 655–662.
65, 102–116. Marquardt, P., Strub, J.R., 2006. Survival rates of IPS empress 2 all-ceramic crowns and
Alao, A.-R., Yin, L., 2015. Nano-mechanical behavior of lithium metasilicate glass- fixed partial dentures: results of a 5-year prospective clinical study. Quintessence Int.
ceramic. J. Mech. Behav. Biomed. Mater. 49, 162–174. 37, 253–259.
Alemi, A., Khademinia, S., Joo, S.W., Dolatyari, M., Bakhtiari, A., 2013. Lithium meta- Mores, R.T., Borba, M., Corazza, P.H., Della Bona, A., Benetti, P., 2017. Influence of
silicate nanomaterials: optical properties and density functional theory calculations. surface finishing on fracture load and failure mode of glass ceramic crowns. J.
Int. Nano Lett. 3, 1–11. Prosthet. Dent. http://dx.doi.org/10.1016/j.prosdent.2016.12.012.
Apel, E., Deubener, J., Bernard, A., Höland, M., Müller, R., Kappert, H., Rheinberger, V., Peng, Z., Rahman, M.I.A., Zhang, Y., Yin, L., 2016. Wear behavior of pressable lithium
Höland, W., 2008. Phenomena and mechanisms of crack propagation in glass-cera- disilicate glass ceramic. J. Biomed. Mater. Res. Part B Appl. Biomater. 104B,
mics. J. Mech. Behav. Biomed. Mater. 1, 313–325. 968–978.
Beuer, F., Schweiger, J., Eichberger, M., Kappert, H.F., Gernet, W., Edelhoff, D., 2009. Reich, S., Endres, L., Weber, C., Wiedhahn, K., Neumann, P., Schneider, O., Rafai, N.,
High-strength CAD/CAM-fabricated veneering material sintered to zirconia copings- Wolfart, S., 2014. Three-unit CAD/CAM-generated lithium disilicate FDPs after a
A new fabrication mode for all-ceramic restorations. Dent. Mater. 25, 121–128. mean observation time of 46 months. Clin. Oral Investig. 17, 2171–2178.
Boaventura, J.M.C., Nishida, R., Elossais, A.A., Lima, D.M., Reis, J.M.S.N., Campos, E.A., Reich, S., Schierz, O., 2013. Chair-side generated posterior lithium disilicate crowns after
de Andrade, M.F., 2013. Effect finishing and polishing procedures on the surface four years. Clin. Oral Investig. 17, 1765–1772.
roughness of IPS Empress 2 ceramic. Acta Odontol. Scand. 71, 438–443. Rekow, E.D., Silva, N.R.F.A., Coelho, P.G., Zhang, Y., Guess, P., Thompson, V.P., 2011.
Bollen, C.M.L., Lambrechts, P., Quirynen, M., 1997. Comparison of surface roughness of Performance of dental ceramics: challenges for improvements. J. Dent. Res. 90 (8),
oral hard materials to the threshold surface roughness for bacterial plaque retention:

259
A.-R. Alao et al. Journal of the Mechanical Behavior of Biomedical Materials 74 (2017) 251–260

937–952. Ulutan, D., Ozel, T., 2011. Machining induced surface integrity in titanium and nickel
Rekow, D., Thompson, V.P., 2005. Near-surface damage-a persistent problem in crowns alloys: a review. Int. J. Mach. Tools Manuf. 51, 250–280.
obtained by computer-aided design and manufacturing. Proc. Inst. Mech. Eng. H 219, Valenti, M., Valenti, A., 2009. Retrospective survival analysis of 261 lithium disilicate
233–243. crowns in a private general practice. Quintessence Int. 40, 573–579.
Schmidt, C., Weigl, P., 2000. Machinability of IPS Empress 2 framework ceramic: po- van den Breemer, C.R.G., Vinkenborg, C., van Pelt, H., Edelhoff, D., Cune, M.S., 2017. The
tential for ceramic dentures. J. Biomed. Mater. Res. (Appl. Biomater.) 53, 348–352. clinical performance of monolithic lithium disilicate posterior restorations after 5, 10,
Serbena, F.C., Zanotto, E.D., 2012. Internal residual stresses in glass-ceramics: a review. J. and 15 years: a retrospective case series. Int. J. Prosthodont. 30, 62–65.
Non-Cryst. Solids 358, 975–984. Wen, G., Zheng, X., Song, L., 2007. Effects of P2O5 and sintering temperature on micro-
Sergueeva, A.V., Mara, N.A., Kuntz, J.D., Lavernia, E.J., Mukherjee, A.K., 2005. Shear structure and mechanical properties of lithium disilicate glass-ceramics. Acta Mater.
band formation and ductility in bulk metallic glass. Philos. Mag. 85, 2671–2687. 55, 3583–3591.
Song, X.-F., Ren, H.-T., Yin, L., 2016. Machinability of lithium disilicate glass ceramic in Whitehead, S.A., Shearer, A.C., Watts, D.C., Wilson, N.H.F., 1995. Comparison of methods
in vitro dental diamond bur adjusting process. J. Mech. Behav. Biomed. Mater. 53, for measuring surface roughness of ceramic. J. Oral. Rehabil. 22, 421–427.
78–92. Yin, L., Han, Y.-G., Song, X.-F., Wang, H., 2007. Effect of diamond burs on process and
Steiner, R., Beier, U.S., Heiss-Kisielewsky, I., Engelmeier, R., Dumfahrt, H., Dhima, M., damage involving in vitro dental resurfacing of a restorative porcelain. J. Phys. D:
2015. Adjusting dental ceramics: an in vitro evaluation of the ability of various Appl. Phys. 40, 5291–5300.
ceramic polishing kits to mimic glazed dental ceramic surface. J. Prosthet. Dent. 113, Zhang, B., Howes, T.D., 1994. Material-removal mechanisms in grinding ceramics. CIRP
616–622. Ann.-Manuf. Technol. 43, 305–308.
Swain, M.V., 2009. Unstable cracking (chipping) of veneering porcelain on all-ceramic Zhang, Y., Kim, J.-W., Bhowmick, S., Thompson, V.P., Rekow, E.D., 2009. Competition of
crowns and fixed partial dentures. Acta Biomater. 5, 1668–1677. fracture mechanisms in monolithic dental ceramics: flat model system. J. Biomed.
Thieme, K., Rüssel, C., 2015. Nucleation and growth kinetics and phase analysis in zir- Mater. Res. Part B: Appl. Biomater. 88B, 402–411.
conia-containing lithium disilicate glass. J. Mater. Sci. 50, 1488–1499. Zhang, Y., Lee, J.J.W., Srikanth, R., Lawn, B.R., 2013a. Edge chipping and flexural re-
Tholt, B., Miranda-Júnior, W.G., Prioli, R., Thompson, J., Oda, M., 2006. Surface sistance of monolithic ceramics. Dent. Mater. 29, 1201–1208.
roughness in ceramics with different finishing techniques using atomic force micro- Zhang, Y., Sailer, I., Lawn, B.R., 2013b. Fatigue of dental ceramics. J. Dent. 41,
scope and profilometer. Oper. Dent. 31–4, 442–449. 1135–1147.

260