The Crystallization of An Aluminosilicate Glass Inthek O-Al O - Sio System
The Crystallization of An Aluminosilicate Glass Inthek O-Al O - Sio System
The Crystallization of An Aluminosilicate Glass Inthek O-Al O - Sio System
www.intl.elsevierhealth.com/journals/dema
Centre for Adult Oral Health, Barts and the London, Queen Marys School of Medicine and Dentistry,
Turner Street, Whitechapel, London, E1 2AD, UK
b
Department of Biomaterials in Relation to Dentistry, Queen Mary, University of London,
Mile End Road, London E1 4NS, UK
c
Division of Biomaterials and Tissue Engineering, Eastman Dental Institute, University College London,
256 Grays Inn Road, London WC1X 8LD, UK
d
496 Alegre Avenue, Nipomo, CA 93444, USA
Received 16 December 2004; accepted 2 February 2005
KEYWORDS
Ceramics;
Leucite;
Crystallization;
Microscopy;
X-ray diffraction
Summary Objectives. The aims of the study were to explore the nucleation and
crystallization kinetics of an aluminosilicate glass in K2OAl2O3SiO2 system and to
characterize it.
Objectives. A starting glass composition of wt%; 64.2% SiO2, 16.1% Al2O3, 10.9% K2O,
4.3% Na2O, 1.7% CaO, 0.5% LiO and 0.4% TiO2 was heated in an electric furnace and
later quenched to produce glasses. The glass powders were heat treated using
differing heat treatment schedules and quenched. Dta, Xrd, Eds and Sem analyses
were used to characterize and explore the crystallization kinetics of the glasses.
Results. Phase separation of the glasses was identified and characterized in the
glasses. Tetragonal leucite, cubic leucite and sanadine glass-ceramics were produced.
Fine leucite crystals (1 mm2) were crystallized with minimal matrix microcracking.
Significance. Amorphous phase separation appeared to be an important precursor to
nucleation and crystal growth in the alkali aluminosilicate glasses explored. It was
possible to control the crystallization of tetragonal leucite and sanidine phases by
selected heat treatment of glass powders and monoliths, resulting in the production
of fine grained tetragonal leucite glass-ceramics.
Q 2005 Academy of Dental Materials. Published by Elsevier Ltd. All rights reserved.
Introduction
* Corresponding author. Tel.: C44 2073777000x2160; fax: C44
2088101254.
E-mail address: m.cattell@qmul.ac.uk (M.J. Cattell).
0109-5641/$ - see front matter Q 2005 Academy of Dental Materials. Published by Elsevier Ltd. All rights reserved.
doi:10.1016/j.dental.2005.02.001
812
substructures and as a reinforcing agent in all
ceramic restorations. Leucite glass-ceramics are
extremely versatile as they can be processed into
dental prostheses via many mechanisms including,
sintering, [1] heat pressing [2] and Computer aided
design and machining [3].
Depending on the composition, feldspars can be
heated to form a eutectic mixture of leucite and
liquid according to the K2OAl2O3SiO2 ternary
phase diagram [4]. The crystallization of leucite
has also been achieved in these glasses by changes
to the compositional ratios and an increase in K2O
content, together with the addition of suitable
nucleating agents [5]. Significant crystallization
was thought to be related to a minimum 1 wt%
CaO and at least 12 wt% K2O content in the glasses
[6]. Leucite glass-ceramics can be produced by
controlled crystallization of a glass via heat treatments; [7] the addition of synthetic leucite to
powdered glass [8] and the blending of a high
expansion leucite containing frit with a low expansion glass frit [9].
Leucite (KAlSi2O6) is a naturally occurring potassium alumino-silicate that has corner linked four
and six-membered rings of SiO4 tetrahedra, forming
the structure of a (Si, Al)-O framework silicate.
Within the framework formed exists two different
cation sites. The larger site contains 16 positions
coordinated by 12 oxygens, organized in line with
channels formed by six-membered rings. The
second site contains 24 positions coordinated by
six oxygens, although only 16 are filled [10]. The
larger sites are usually occupied by potassium,
rubidium or cesium ions and smaller sites by water
molecules or left vacant. Cubic and tetragonal
leucite both contain 16 (KAlSi2O6) in the unit cell,
with potassium ions occupying the centre of
channels aligned in rows parallel to the channel
axes in cubic leucite. A reversible temperature
dependent cubic to tetragonal leucite phase transformation occurs around 605625 8C [10,11]. At
lower temperatures the potassium ions are too
small to fill the cavities in the cubic structure and
are moved away from their positions [12]. There is a
collapse of the framework about them resulting in a
change in symmetry and strain induced crystal
twinning [13]. Leucite transformation leads to a
reversible 1.2% discontinuous volume change, from
low temperature tetragonal to the high temperature cubic leucite starting at 400 8C, in respect to
crystallized feldspathic glasses. Lattice parameters
for room temperature tetragonal leucite were
(13.197!10K10 m) and
reported at aZ13.197 A
(13.819!10K10 m) and at 400 8C latcZ13.819 A
(13.565!10K10 m)
tice constraints of aZ13.565 A
for cubic leucite. A change in thermal expansion
813
Ka2Z1.5444 A). Data was collected using a scintillation counter and a graphite diffracted beam
monochromator.
Structure refinement
Data were refined using General Structure Analysis
Software (GSAS, Los Alamos National Laboratory,
USA). Three potential phases could be present and a
starting model for each was determined from the
Inorganic Crystal Structure Database at the Daresbury Laboratories (Daresbury, UK). The model used
for tetragonal leucite was based on the structural
determination of Palmer et al. [29]. The starting
model used was a space group I41/a with lattice
and cZ13.75 A
(1 A
Z
parameters of aZ13.06 A
10K10 m). For the cubic leucite phase a space group
of IA-3D was used, with starting unit cell dimensions
[30]. The sanidine K (Si3 Al) O8
of aZ13.5 A
(disordered) phase was modelled using a space
group of C2/m (no12) with lattice parameters of
, bZ13.035 A
, cZ7.175 A
and betaZ116
aZ8.604 A
degree angle [31]. For the refinement, peak shapes
were modeled on a pseudo Voigt distribution and an
asymmetry parameter was refined. Scattering
factors for neutral atoms were assumed. Because
of the complex nature of the background, due to
amorphous material, twenty background parameters were used. A scale factor, four peak shape
variables, cell parameters and a zero point correction were refined. Structural parameters were not
refined. The unit cell dimensions were determined
for each sample.
814
for secondary electron imaging (SEI). All specimens
were polished with 1 mm followed by 0.3 mm alumina
micropolish (Buehler, USA) and cleaned. Cleaning in
alcohol for 30 s and water rinsing was carried out
before ultrasonic cleaning for 3 min at 50 8C. Specimen etching was achieved using 0.1% hydrofluoric
acid for 1 min, followed by water rinsing. Etched
specimens were mounted on coded brass stubs and
gold coated using a sputter coater (Balzers ScDo5o
Bal-Tec, Liechtenstein), for 100 s at 40 mA. Secondary electron imaging was carried out with a field
emission scanning electron microscope (JSM 6300F,
Joel, Ltd, UK), using an accelerating voltage of 5 or
10 kV. One photomicrograph was taken per etched
specimen so that the microstructures could be
Figure 1 (a) SEM photomicrograph of the quenched glass block prior to heat treatment showing elliptical domains
(!18000). (b) SEM photomicrograph of the quenched glass block withdrawn at 650 8C showing dendritic growth
(!6500). (c) SEM photomicrograph of the quenched glass block withdrawn at 750 8C showing phase coarsening and
crystallization (!19000). (d) SEM photomicrograph of the quenched glass block withdrawn at 850 8C showing sanidine
crystals (!1300). (e) SEM photomicrograph of the quenched glass block withdrawn at 850 8C showing leucite and
sanidine crystals (!5000).
Figure 2 Light microscope image of the 950 8C specimen displaying the morphology of leucite and sanidine
crystals (!50).
Characteristic X-rays were acquired using a Pentafet detector and a beryllium window to give
quantitative results, via an X-ray analysis program
(Link eXLII, Oxford instruments, High Wycombe,
UK). A dot map was used to quantify the elemental
areas in several instances. Cobalt was used as a gain
calibration during elemental quantitative analysis.
Light microscopy
Specimens were viewed in a light microscope
(Olympus BX60, Olympus optical Co., Ltd, UK), before
etching and gold coating and the images collected via
a digital camera connected to a computer.
Results
Secondary electron imaging of the
crystallization process
The 355105WQ glass (block) crystallization
The 355105WQ glass frit appeared transparent prior
to crystallization. Photomicrographs illustrating
the crystallization process for the 355105WQ
Table 1
815
polished glass blocks withdrawn over the temperature range 650950 8C are shown in Figs. 1ae and 2.
Spherical and elliptical (mean (SD)) 0.1 (0.1) mm2
phase separated areas were present in the
quenched glass blocks (355105WQ), before heat
treatment (Fig. 1a). The glass block withdrawn at
650 8C revealed tiny inclusions and areas of coarsening and dendritic growth (Fig. 1b). Signs of
domain coarsening and primary crystallization were
present on the 700 and 750 8C specimens (Fig. 1c).
Crystallization was present at the grain boundaries
or in the vicinity of coarsened phase separated
areas (Fig. 1c). Increased signs of crystallization
and coarsening were present for the 800 8C specimen. Extensive sanidine crystallization (platelets)
occurred in the 850 8C specimen (Fig. 1d), together
with a lower area fraction of leucite crystals
(Fig. 1e). The 900 and 950 8C specimens produced
leucite crystals together with a lower area fraction
of sanidine crystals. Curling of the sanidine platelets was observed in the 850950 8C temperature
range. The morphology of these curved platelets
can be clearly observed using light microscopy
(Fig. 2). Extensive leucite crystallization was present for the 1000 8C specimen.
The 355105WQ and 355105 glass powder
crystallization
Sanidine crystals (platelets) were visible in the
355105WQ powder samples for all holds at 850, 900
and 950 8C. Leucite crystals (mean (SD)) 4.2
(8.8) mm2 (Table 1) and crystal and circumferential
matrix microcracking were present for all of the
1120 8C samples (Fig. 3). The leucite crystals
frequently appeared coarsened or coalesced, leading to an irregular shaped morphology. Possible
phase separated areas were also present intermittently in the glassy matrix.
The 355105 glass frit appeared semi-transparent
before the crystallization heat treatments. The
glass powder (355105) withdrawn and quenched at
650 8C had signs of spherical phase separated areas
and possible primary crystallization (Fig. 4a).
Energy dispersive X-ray analysis (EDS) of the
Crystalline content and leucite particle size of the 355105 and 355105WQ experimental glass-ceramics.
Experimental
glass-ceramic
Crystalline
component
Leucite area
fraction (%)
355105 &
355105 :
355105WQ &
Tetragonal leucitea
Tetragonal leucite
Tetragonal leucite
15.9
13.7
20.3
1.1G0.7
0.8G0.5
4.2G8.8
0.023.9
0.073.6
0.535.9
All Samples were ramped from 23 8C. & One-step heat treatment: (3 8C/min ramp to 1120/1 h hold); : One-step heat treatment:
(3 8C/min ramp to 1120/3 h hold).
a
Trace amounts of cubic leucite were also present.
816
Figure 4 (a) SEM photomicrograph of the 355105 powder specimen quenched at 650 8C (!12000). (b) Elemental dot
map showing calcium rich phase separated areas in the 355105 powder specimen quenched at 650 8C. (c) EDS spot
analysis of the phase separated areas in the 355105 glass powder quenched at 650 8C. (d) EDS glass matrix spot analysis of
the 355105 glass powder quenched at 650 8C.
817
Figure 5 (a) SEM photomicrograph of leucite crystals and phase separation in a 355105 powder specimen heated from
23 8C at 3 8C/min and withdrawn at 850 8C (!5000). (b) SEM photomicrograph of leucite crystals and sanidine platelets in
the 950 8C/1 h 355105 powder specimen (!3500). (c) SEM photomicrograph showing leucite crystals in the 1120 8C/1 h
hold 355105 powder specimen (!3500).
The results of the particle size analysis for 355105 and 355105WQ glass powders.
Glass powder
5.39
5.88
1.11
1.15
23.47
28.18
818
Table 3
Glass-ceramic
powder
Glass-ceramic
Mean beta
degree angle
(SD)
355105WQ
(850 8C)
355105WQ
(900 8C)
355105WQ
(950 8C)
355105WQ
(1120 8C)
355105
(1120 8C)
Sanidine
8.5264 (0.0007)
12.9961(0.0012)
7.1458 (0.0006)
115.91 (0.007)
Sanidine
8.535 (0.0008)
13.0044 (0.0011)
701452 (0.0006)
115.89 (0.007)
Sanidine
8.5217(0.001)
12.9807 (0.002)
7.1445 (0.001)
115.93 (0.009)
Tetragonal
leucite
Tetragonal
leucite
13.012 (0.003)
13.012 (0.003)
13.6344 (0.006)
90
13.1879 (0.0006)
13.1879 (0.0006)
13.7489 (0.009)
90
Z10K10 m).
All samples displayed were ramped from 23 8C at 3 8C/min and held for 1 h at the stated holding temperatures (1 A
X-ray diffraction analyses of the crystallization process for the 355105 glass powder.
Temperature (8C)
850
850
850
850
850
900
900
900
900
900
950
950
950
950
950
1120
1120
1120
1120
1120
10
30
60
120
180
10
30
60
120
180
10
30
60
120
180
10
30
60
120
180
Tetragonal leucite
Tetragonal leucite (MP)CSanidine (M)
Tetragonal leucite (MP)CSanidine (M)
Sanidine
Sanidine
Tetragonal leucite
Tetragonal leucite(MP)CSanidine (M)
Sanidine (MP) CTetragonal leucite (M)
Sanidine (MP) CTetragonal leucite (M)
Sanidine (MP)Ctrace Tetragonal leucite (M)
Tetragonal leucite (MP)CSanidine (M)
Tetragonal leuciteCSanidine
Tetragonal leuciteCSanidine
Tetragonal leuciteCSanidine
Sanidine (MP)Ctrace leucite (M)
Tetragonal leucite
Tetragonal leucite
Tetragonal leucitea
Tetragonal leucite
Tetragonal leucite
All samples were ramped from 23 8C to the holding temperatures at 3 8C/min before holding.
a
A trace component that fitted cubic leucite was also identified.
(a)
819
raw data
(b)1750
2500
counts /a.u.
counts /a.u.
1500
2000
1500
1000
1250
1000
750
500
500
250
10
15
20
25
30
35
40
45
10
50
15
20
(d)
raw data
(c)
2500
counts /a.u.
counts /a.u.
2000
1500
1000
500
10
15
20
25
30
35
2 theta /deg.
25
30
35
40
45
50
25
30
35
2 theta /deg.
40
45
50
2 theta /deg.
2 theta /deg.
40
45
50
raw data
1100
1000
900
800
700
600
500
400
300
200
100
10
15
20
Figure 6 (a) The X-ray diffraction trace for the 900 8C/1 h hold, 355105WQ powder specimen. (b) The X-ray diffraction
trace for the 1120 8C/1 h hold, 355105WQ powder specimen. (c) The X-ray diffraction trace for the 900 8C/1 h hold,
355105 powder specimen. (d) The X-ray diffraction trace for the 1120 8C/1 h hold, 355105 powder specimen.
Discussion
Scanning electron microscopy of the quenched
glasses showed discreet elliptical and spherical
domains in both the 355105WQ polished glass blocks
(Fig. 1a), and both the 355105WQ and 355105
powder specimens (Fig. 4a). Energy dispersive
X-ray analysis and an elemental dot map (Fig. 4b
d) revealed the presence of calcium rich spherical
domains surrounded by more silica rich areas in the
355105 glass powder specimen. Amorphous phase
separation may therefore be present, with high
calcium aluminosilicate type phase separated
areas. Immiscibility has been characterized in
silicate glass forming systems, with the degree,
connectivity, and composition of phase separated
areas being related to the position on the immiscibility gap [32]. Further heat treatment caused the
coalescence and coarsening of these areas (Ostwald
ripening) and evidence of crystallization at the
grain boundaries between the two glass phases. In
particular, the precipitation of a platelet phase was
preferential at this interface (Fig. 1c), which could
have lowered the interfacial energy and provided
an increased driving force for crystal nucleation
[33]. The crystallization of a primary Sanidine
platelet phase appeared to be associated with
820
leucite crystallization at high (48 wt%) TiO2 content [6]. Titanium dioxide was added to the 355105
(0.4 wt%) and 355105WQ (0.4 wt%) glass compositions and has been suggested as a sub-microscopic
catalyst allowing the heterogeneous nucleation and
growth of the major crystalline phase [40]. Amorphous phase separation may however, also be a
consequence of high valence transition oxides like
TiO2 being displaced from the aluminosilicate network to form a separate phase with a divalent
cation [41,38]. Evidence of this was not apparent in
the studied glasses, but was identified by the
present authors in compositions with higher TiO2
content in subsequent work.
Leucite has previously been crystallized in alkali
aluminosilicate glasses via controlled surface crystallization [39] and this was thought to be the
dominant crystallization mechanism in the present
glasses studied. Leucite crystal sizes in the range
0.11 mm were initiated from the surface in high
(8 wt%) titanium containing alkali aluminosilicate
glasses [16]. Simultaneous surface and volume
leucite crystallization was however, also discussed
in glass compositions with low K2O content [16].
Tosic
et al. [42] indicated surface leucite crystallization at particle sizes less than 0.075 mm and
volume crystallization above this particle size in
aluminosilicate glasses. The finding was complicated by the presence of secondary phases whose
crystallization behavior was unknown. Further work
is therefore necessary to confirm the above
conclusions.
The early stages of bulk leucite growth have been
observed as dendrites growing in four preferred
crystallographic directions [42]. A diffusion controlled growth process that evolved at the smooth
atomic-scale faceted crystal-glass interface was
suggested. A change in dendrite shape due to the
growth of secondary and tertiary fibrils and their
ripening resulted in a highly organized tetragonal
leucite structure. The 355105WQ glass block withdrawn at 650 8C (Fig. 1b) exhibited the growth of
dendrites in six directions, which are crystallographically available for a cubic crystal. Further lateral
growth and thickening of fibrils into uncrystallized
areas between the main fibres were seen on the
glass blocks withdrawn at 750 8C, 850 8C (Fig. 1e)
and 950 (Fig. 2). Freiman et al. [43] described the
slow lateral growth of arms into adjacent areas of
uncrystallized melt as a secondary crystallization
process. Freiman and Onada [44] confirmed LiC or
NaC ions in glass maybe responsible for lowered
viscosity and increased mobility affecting
crystal morphology and growth. The alkali concentration at the grain boundaries and between the
growing arms leading to secondary crystallization.
821
References
[1] Nathanson D. Principles of porcelain use as an inlay/onlay
material. In: Garber DA, Goldstein RE, editors. Porcelain
and composite inlays and onlays, aesthetic posterior
restorations. Carol Stream, IL: Quintessence Publishing
Co, Inc; 1994. p. 326.
[2] Wohlwend A, Schaerer P. The Empress technique for the
fabrication of full ceramic crowns, inlays and veneers.
Quintessenz Zahntech 1990;16:96678.
[3] Mormann WH, Bindl A. The cerec 3a quantum leap for
computer-aided restorations: initial results. Quintessence
Int 2000;31:699712.
[4] Schairer JF, Bowen NL. The system K2OAl2O3SiO2. Am J Sci
1955;253:681746.
[5] Binns DB. The chemical and physical properties of dental
porcelain. In: McLean JW, editor. Dental ceramics, Proceedings of the 1st international symposium on dental
ceramics. Chicago: Quintessence Pub., Co.; 1983. p. 4156.
[6] Hermansson L, Carlsson R. On the crystallization of the
glassy phase in whitewares. Trans J Br Ceram Soc 1978;77:
325.
[7] Ibsen RL, Chadwick TC, Pritchard SA. Strong dental
porcelain and method for its manufacture. US patent
5,009,709; 1991.
[8] Burk B, Burnett AP. Leucite containing porcelains and
method of making same. US Patent 4101330; 1978.
[9] Weinstein M, Katz S, Weinstein AB. Fused porcelain to metal
teeth. US Patent No. 3,052,982; 1962.
[10] Deer WA, Howie RA, Zussman J. Framework silicates. In: An
introduction to the rock-forming minerals, 11th ed.
London:Wiley; 1966, p. 281317.
[11] Mackert Jr JR. Effects of thermally induced changes on
porcelain-metal compatibility. In: Preston JD, editor.
Perspectives in dental ceramics, Proceedings of the fourth
international
symposium
on
ceramics.
Chicago:
Quintessence Pub., Co.; 1988. p. 5364.
[12] Mazzi F, Galli E, Gottardi G. The crystal structure of
tetragonal leucite. Am Mineral 1976;61:10815.
[13] Palmer C, Putnis A, Salje EKH. Twinning in tetragonal
leucite. Phys Chem Miner 1988;16:298303.
[14] Mackert JR, Butts MB, Fairhurst CW. The effect of the
leucite transformation on dental porcelain expansion. Dent
Mater 1986;2:326.
[15] Hermansson L, Carlsson R. High and low temperature forms
of leucite. In: Proceedings, eighth international symposium
on the reactivity of solids. Gothenberg, Sweden, Swedish
Institute for Silicate Research; 1976. p. 5415.
[16] Rouf MA, Hermansson L, Carlsson R. Crystallisation of
glasses in the primary phase field of leucite in the K2O
Al2O3SiO2 system. Trans J Br Ceram Soc 1978;77:369.
[17] Denry IL, Mackert Jr JR, Holloway JA, Rosenstiel SF. Effect
of cubic leucite stabilisation on the flexural strength of
feldspathic dental porcelain. J Dent Res 1996;75:192835.
[18] Piche
PW, Obrian WJ, Groh CL, Boenke KM. Leucite content
of selected porcelains. J Biomed Mater Res 1994;28:6039.
[19] Mackert Jr JR, Russell CM. Leucite crystallisation during
processing of a heat pressed dental ceramic. Int
J Prosthodont 1996;9:2615.
[20] Mackert Jr JR, Evans AL. Quantitative X-ray diffraction
determination of leucite thermal instability in dental
porcelain. J Am Ceram Soc 1991;74:4503.
[21] Mackert Jr JR, Evans AL. Effect of cooling rate on leucite
volume fraction in dental porcelains. J Dent Res 1991;70:
37139.
822
[22] Barreiro MM, Riesgo O, Vicente EE. Phase identification in
dental porcelains for ceramo-metallic restorations. Dent
Mater 1989;5:517.
[23] Shareef MY, Van Noort R, Messer PF, Piddock V. The effect
of microstructural features on the biaxial flexural strength
of leucite reinforced glass ceramics. J Mater Sci: Mater Med
1994;5:1138.
[24] Cattell MJ, Knowles JC, Clarke RL, Lynch E. The biaxial
flexural strength of two pressable ceramic systems. J Dent
1999;27:18396.
[25] Fairhurst CW, Lockwood PE, Dingle RD, Thompson WO. The
effect of glaze on porcelain strength. Dent Mater 1992;8:
2037.
[26] Mackert Jr JR, Williams AL. Microcracks in dental porcelain
and their behaviour during multiple firing. J Dent Res 1996;
75:148490.
[27] Cattell MJ, Clarke RL, Lynch E. The transverse strength,
reliability and microstructural features of four dental
ceramics-part 1. J Dent 1997;25:399407.
[28] Metzler KT, Woody RD, Miller AW, Miller BH. In vitro
investigation of the wear of human enamel by dental
porcelain. J Prosthet Dent 1999;81:35664.
[29] Palmer DC, Dove MT, Ibberson RM, Powell BM. Structural
behavior, crystal chemistry and phase transitions in
substituted leucite: high resolution neutron diffraction
studies. Am Miner 1997;82:1629.
[30] Dove MT, Cool T, Palmer DC, Putnis A, Salje EKH, Winkler B.
On the role of AlSi ordering in the cubic-tetragonal
transition of leucite. Am Miner 1993;78:48692.
r kristallographie
[31] Scheel HJ. Lead feldspar. Zeitschrift fu
1971;133:26472.
[32] Seward TP, Uhlmann DR, Turnbull D. Development of twophase structures in glasses with special reference to the
system BaOSiO2. J Am Ceram Soc 1968;51:63443.
[33] Uhlmann DR. General discussion. Trans Faraday Soc 1970;
50:22334.
[34] Szabo
I, Nagy B, Vo
lksch G, Ho
land WH. Structure, chemical
durability and microhardness of glass-ceramics containing
apatite and leucite crystals. J Non-Cryst Solids 2000;272:
1919.
[35] Ohlberg SM, Golob HR, Strickler DW. Crystal nucleation by
glass in glass separation. In: Symposium on nucleation and
crystallisation in glasses and melts, Columbus, OH: Ceramic
Society; 1962. p. 5562.
[36] James PF. Nucleation in glass forming systemsa review. In:
Simmons JH, Uhlmann DR, Beall GH, editors. In: Advances in
ceramics, nucleation and crystallisation in glasses, 4.
[37]
[38]
[39]
[40]
[41]
[42]
[43]
[44]
[45]
[46]
[47]
[48]
[49]
[50]
[51]