Journal J. Am. Ceram. Soc.

, 82 [1] 5–16 (1999)

Nanophase Glass-Ceramics
George H. Beall* and Linda R. Pinckney*
Corning Incorporated, Corning, New York

Future applications for glass-ceramics are likely to capital- phases. In other cases, crystallization proceeds within or from
ize on designed-in, highly specialized properties for the the surfaces of the phase-separated globules themselves.
transmission, display, and storage of information. Glass- Nucleation is followed by one or more higher-temperature
ceramics with microstructures comprised of uniformly dis- treatments to promote crystallization of the primary phase or
persed crystals <100 nm in size offer promise for many phases and development of the desired microstructure. The
potential new applications as well as provide unique attri- growth process continues until neighboring crystals impinge,
butes for many current products. This paper focuses on two creating a highly crystalline body with a small amount of re-
types of nanocrystalline glass-ceramics: transparent glass- sidual glass, or until the residual glass is depleted in the crystal-
ceramics and tough, high-modulus glass-ceramics with pre- forming components. Certain glass-ceramics are specifically
cisely engineered surfaces. Transparent glass-ceramics are designed to possess a microstructure of uniformly dispersed,
formed from certain aluminosilicate glasses capable of ef- nonimpinging crystals in a matrix of continuous residual glass.
ficient crystal nucleation and slow growth. The key crys- Glass-ceramics have several advantages over conventional
talline phases include ␤-quartz solid solutions, character- powder-processed ceramics. In addition to the ease of flexibil-
ized by low-thermal-expansion behavior; spinel, with high ity of forming in the glassy state, glass-ceramics possess a
hardness and elastic modulus; and mullite, which shows uniformity of microstructure and subsequent reproducibility of
unique chromium-luminescence behavior. properties that result from the homogeneity of the starting
glass. Furthermore, glass-ceramics provide a wide range of
physical properties—such as coefficients of thermal expansion
I. Introduction (CTEs) from −75 × 10−7/°C to +200 × 10−7/°C—that often are
not readily obtained in glasses or ceramics. Many commercial
G LASS-CERAMIC technology is based on the controlled nucle-
ation and crystallization of glass. Although glass-ceramic
articles can be fabricated through the internal nucleation of
glass-ceramics are valued primarily for their near-zero thermal
expansion, while the combination of high mechanical strength
and zero porosity has been exploited by applications ranging
glass monoliths or via the sintering and crystallization of glass from architectural materials to tableware to bone implants. Of
frits, internal nucleation affords a much wider array of potential the many types of microstructures obtainable in glass-ceramics,
microstructures. Some glass compositions are self-nucleating, those based on uniformly dispersed crystals <100 nm in size
but, more commonly, certain components known as nucleating provide unique attributes for current products and offer prom-
agents are added to the batch to promote phase separation and ise for many potential new applications. Such microstructures
internal nucleation. The nucleating agents melt homogeneously can be termed ‘‘nanocrystalline.’’
into the glass but promote very-fine-scale phase separation on This paper focuses on two types of nanocrystalline glass-
reheating. The dispersed phase is typically structurally incom- ceramics: transparent glass-ceramics and tough, high-modulus
patible with the host glass and, therefore, precipitates tiny crys- glass-ceramics with precisely engineered surfaces. The first
talline nuclei on heating at temperatures of 30°–100°C above group has numerous consumer and technical applications,
the annealing point of the glass. These crystals serve as the while the second are of interest as magnetic memory disk sub-
sites for subsequent nucleation of the primary crystalline strates or other applications that require smooth surfaces,
chemical durability, and high use temperature.

R. A. Condrate—contributing editor II. Transparent Glass-Ceramics

Transparent glass-ceramics generally have two distinctive
properties: they are nanocrystalline, and they have greater ther-
Manuscript No. 189920. Received August 27, 1998; approved November 25, 1998. mal stability than their parent glasses, frequently possessing
*Member, American Ceramic Society. upper use temperatures of >800°C. The most important com-

centennialfeature
5
6 Journal of the American Ceramic Society—Beall and Pinckney Vol. 82, No. 1

mercial transparent glass-ceramics capitalize on their superior

thermal properties, particularly ultra-low thermal expansion
coupled with high thermal stability and thermal shock resis-
tance. Zero or near-zero expansion materials based on lithium-
stuffed ␤-quartz crystals are used in high-precision optical ap-
plications, such as telescope mirror blanks as well as stove
cooktops, cookware, woodstove windows, fire doors, and other
technical devices.
Another family of transparent glass-ceramics has thermal
expansions that closely match that of silicon. These materials
are generally based on dispersions of spinel or mullite in non-
alkali siliceous glass and have potential applications in the area
of flat panel displays. A third group of transparent glass-
ceramics possess luminescence characteristics that are signifi-
cantly different from those of glasses. Potential applications
include optical amplifiers, solar collectors, and upconversion
devices for laser application.
Fig. 1. Effect of temperature on the rates of homogeneous nucleation
(1) Requirements for Transparency and crystal growth. (After Tamann, 1925.)
Good transparency requires materials with low optical scat-
tering and low ionic/atomic absorption. Low scattering, the
more difficult to achieve, can be obtained by satisfying either and often between 50° and 100°C above the glass-transition
of two criteria. The first criterion is satisfied when all the temperature (Tg). The rate of crystal growth should be kept low
crystalline phases and the residual glass have closely matched to minimize grain size, utilizing a smaller temperature differ-
indexes of refraction and when the crystals possess low bire- ence between nucleation and growth temperatures than would
fringence. Glass-ceramics that meet this criterion are exempli- be suggested from the maxima of the peaks.
fied by magnesium- and zinc-stuffed ␤-quartz in the SiO2–
Al2O3–MgO–ZnO–ZrO2 system.1 Despite crystal sizes of up to (2) Transparent Glass-Ceramics Based upon Metastable
10 ␮m, the optical near-isotropy of the ␤-quartz crystals pro- ␤-Quartz Solid Solutions
vides good transparency in the glass-ceramic. Glass-ceramics with near-zero CTE are produced through
The second criterion for low scattering is satisfied in the case the precipitation of stuffed ␤-quartz (high quartz) as the major
where the crystal size is much smaller than the wavelength of crystalline phase. The idealized ␤-quartz structure comprises a
light. Within this condition, there are two types of scattering framework of interlinked helixes of SiO2 tetrahedra.7,8 Dis-
situations. One involves independent scatterers that are widely torted tetrahedral cavities, capable of accepting small cations
separated and obey the Rayleigh–Gans model.2 In this case, ␴p, with ionic radii of 0.06–0.08 nm, are created between the two
the total turbidity or attenuation due to scattering, is given as helixes. The term ‘‘stuffed’’ derives from the solid solution
wherein Al3+ replaces Si4+ in the quartz tetrahedral positions,
2 the charge balance being maintained by ions that stuff the
␴p ≈ NVk4a3共n⌬n兲2
3 interstitial tetrahedral cavities. The ultimate Li+-stuffed deriva-
tive of ␤-quartz is the stoichiometric ␤-eucryptite, LiAlSiO4,
where N is the particle number density, V the particle volume, where fully one-half of the silicon is replaced by aluminum and
a the particle radius, k ⳱ 2␲/␭ (where ␭ is the wavelength), n lithium that preferably occupy tetrahedral cavities on the SiO4
the refractive index of the crystal, and ⌬n the index difference tetrahedral plane of alternating SiO4 and AlO4 tetrahedra par-
between the crystal and the host. For practical purposes, trans- allel to (001).7 Divalent magnesium and zinc ions can partially
parency is achieved here with particle radii of <15 nm and a or completely replace lithium in some of these ␤-quartz struc-
refractive index difference of <0.1 between the glass and crys- tures. In this case, the Mg2+ and Zn2+ ions preferably occupy
tals. This is a fairly stringent condition, but it can be achieved. octahedral sites between the two tetrahedral sites described
The other scattering model involves small particles that are above. To a lesser degree, the cavities also accept small
more closely spaced. This model requires that the distance amounts of Fe2+, Mn2+, and Co2+.
between particles be no smaller than the particle radius but can A typical commercial composition may comprise ∼60 mol%
be at least up to 6 times the particle radius. Under these con- SiO2 and ∼40 mol% eucryptite components, with some of the
ditions, a quasi-continuum model developed independently by Li2O replaced by MgO and zinc. The solid solutions that have
Andreev3 and Hopper4 describes interfering fields from the been studied most are described by the general formula9
individual particles that produce a large compensating effect.
Hopper gives the formula for turbidity as Li2−2(x+y)MgxZnyO⭈Al2O3⭈zSiO2

␴c ≈ 冋冉 2
3 冊 册
× 10−3 k4␪3 共n⌬n兲2
where x + y ⱕ 1 and z > 2. In addition, phosphatic ␤-quartz
solid solutions can be formed where AlPO4 substitutes in part
for the SiO2. An example of a phosphatic ␤-quartz composition
where ␪ is the mean phase width (a + W/2) and W the inter- is the commercial Schott Zerodur威. Although all of these
particle spacing. In this case, improved transparency is allowed ␤-quartz solid solutions are believed to be metastable, with the
with particle sizes <30 nm at larger refractive index differ- exception of near the ␤-eucryptite stoichiometry, many of the
ences, up to ⌬n ⳱ 0.3. compounds retain the ␤-quartz structure even when heated at
The rate of nucleation must be close to optimal to achieve 1200°C for 100 h.8 Representative ␤-quartz glass-ceramic
such small crystals. Figure 1 shows the classic diagram of compositions are given in Table I.
Tamann5,6 that illustrates the effect of temperature on the rates Crystals of stuffed ␤-quartz typically exhibit low or negative
of homogeneous nucleation and crystal growth. The optimum volume expansion with increasing temperature. The CTE gen-
nucleation temperature lies below the optimum rate of crystal erally decreases with increasing amounts of Li+ and Zn2+ in the
growth. Both rates are limited by viscosity at low temperatures ␤-quartz structure and increases with increasing Mg2+. It is
and by lack of thermodynamic driving forces at or near the possible to design materials with essentially zero expansion
melting temperatures. Although heterogeneous nucleation rates over fairly wide temperature intervals by tailoring the compo-
can be more complex, the general rule holds that optimal nucle- sition and amounts of ␤-quartz crystals and positive-expansion
ation temperatures occur well below those of crystal growth residual glass. The measured CTE of various ␤-quartz solid
January 1999 Nanophase Glass-Ceramics 7

Table I. Transparent Glass-Ceramics Based on ␤-Quartz

Solid Solution†
Composition (wt%)
Q1 Q2 Q3 Q4
Visions Zerodur Ceran Neoceram
Oxide (Corning) (Schott) (Schott) (NEG)
SiO2 68.8 55.4 63.4 65.1
Al2O3 19.2 25.4 22.7 22.6
Li2O 2.7 3.7 3.3 4.2
MgO 1.8 1.0 0.5
ZnO 1.0 1.6 1.3
BaO 0.8 2.2
P2O5 7.2 1.2
TiO2 2.7 2.3 2.7 2.0
ZrO2 1.8 1.8 1.5 2.3
Fig. 3. Nucleation efficiency in a commercial, stuffed ␤-quartz
†
Excludes <1% Na2O, K2O, As2O3, and F.
glass-ceramic.

solutions along the join SiO2–LiAlO2 is shown in Fig. 2.10

Most of the solid solution range between 50% and 80% SiO2 10 min of nucleation, there are many thousand crystals per
by weight yields a steady CTE of −5 × 10−7/°C. Above 80% cubic millimeter.
SiO2, there apparently is insufficient lithium to prop open the Figure 4 shows a transmission electron microscopy (TEM)
␤-quartz structure, and, therefore, it transforms displacively to micrograph of the Zerodur威 glass-ceramic. The ␤-quartz crys-
the trigonal ␣-quartz form. Because of the very high CTE of tals are <0.1 ␮m in size, and, at the center, one can observe the
␣-quartz, expansions up to 200 × 10−7/°C are observed. On the tiny bricklike ZrTiO4 nuclei. Many more titanate crystals have
other hand, below 50% SiO2, toward stoichiometric ␤-eucryp- precipitated along grain boundaries.
tite, microcracking tends to be common because of the great A current commercial application for low-CTE ␤-quartz
anisotropy in expansions along the a- and c-axes and the strain glass-ceramics is the smooth, radiant cooktop for electric
developing from large crystal interfaces. In this case, micro- stoves. In this case, the glass-ceramic typically is doped with
cracking will bias the thermal expansion toward the c-axis, and 0.1% V2O5 to render the material black in general appearance.
glass-ceramics with bulk negative expansions can be achieved. However, the vanadium does allow transmission in the red and
Glass-ceramics based on either ␣-quartz or ␤-eucryptite, how- near infrared, where the tungsten halogen lamp and other re-
ever, are generally too coarse grained to be transparent. sistive elements radiate energy. Figure 5 illustrates the trans-
TiO2 and ZrO2 are well-known as effective nucleating agents mission of a Corning cooktop versus the radiation from a typi-
for stuffed ␤-quartz crystals. Studies have shown, moreover, cal tungsten halogen lamp. Although the glass-ceramic absorbs
that certain combinations of TiO2 and ZrO2, particularly at the most visible radiation, it transmits very effectively in the red
two-thirds TiO2 level, allow crystallization at lower tempera- and near infrared. Other consumer applications for transparent
tures and, therefore, at higher viscosities.11 The mixture of stuffed ␤-quartz glass-ceramics include transparent cookware,
TiO2 and ZrO2 provides an optimum number of nuclei for very such as VISIONS威, woodstove windows, and fire doors.
small grain sizes. Because zirconium titanate (ZrTiO4) is an Glass-ceramics with ultra-low thermal expansion are par-
extremely effective nucleation crystal for stuffed ␤-quartz, one ticularly valuable in applications for which thermal dimen-
might expect that 50% TiO2 and 50% ZrO2 would be more sional stability is critically important. Technical applications
effective than two-thirds TiO2 and one-third ZrO2. It is be- for which precise dimensional control is crucial include mirror
lieved, however, that TiO2 enters the ␤-quartz structure to a blanks for telescopes and reflective optics and for ring laser
certain extent in replacement of SiO2, and, thus, excess TiO2
beyond what would be expected for the ZrTiO4 stoichiometry
is required. The nucleation efficiency in a commercial stuffed
␤-quartz glass-ceramic is shown in Fig. 3. Here the crystal
density is plotted against nucleation temperature, and it is eas-
ily observed that the maximum nucleation efficiency is at
∼765°C. In fact, with only 1 min of nucleation, the crystal
density is already approaching 100 per cubic millimeter. With

Fig. 2. CTE of solid solution of ␤-quartz crystallized from glasses in Fig. 4. TEM micrograph of the Zerodur威 glass-ceramic. Scale bar
the SiO2–LiAlO2 system. indicates 0.1 ␮m.
8 Journal of the American Ceramic Society—Beall and Pinckney Vol. 82, No. 1

crystal structure is derived from that of sillimanite, with chains

of AlO 6 octahedra crosslinked by (Si,Al)O 4 tetrahedra.
Whereas sillimanite has silicon and aluminum cations ordered
on the tetrahedral sites, mullite has excess Al3+ for Si4+, with
charge balance provided by oxygen vacancies. In other words,
mullite has a defect sillimanite structure derived by 2Si4+ +
O2 ⳱ 2Al3+ + 䊐, with a typical oxygen vacancy of 2%–6%.
These substitutions create a tribridging oxygen position: oxy-
gen ions that are bonded to three aluminum ions. These oxygen
ions have a net local charge of +e/4 but are nevertheless
stable.14
In addition to the Al3+ for Si4+ solid solution, B3+ commonly
enters the tetrahedral sites. In fact, the compound 9Al2O3⭈
2B2O3 has the mullite structure. The six-cation formula for the
borate end member is Al4.91B1.09O9, with a full 10% oxygen
vacancy. This maximum vacancy also is observed in the meta-
Fig. 5. Transmission of a Corning glass-ceramic cooktop versus stable synthetic form of alumina, ␥-Al2O3, or Al6O9.
tungsten–halogen lamp. (After Partridge, 1994.) Metastable solid solutions containing ZnO and/or MgO also
have been postulated.15 These solid solutions involve partial or
complete replacement of the M2 tetrahedral aluminum by zinc
gyroscopes. The best known low-expansion optical material is or magnesium ions with a double substitution in 3:2 mullite of
the glass-ceramic Zerodur威 manufactured by Schott. Its com- either
position has been designed specifically for maximum tempera-
ture–time stability of the metastable ␤-quartz phase and for Zn2+(M2) + Si4+(M1) → Al3+(M2) + Al3+(M1)
constant thermal expansion characteristics within this range.12 or
Ambient temperature changes from −50° to +50°C produce
length changes of only a few parts per million. This property is 2Zn2+(M2) + 䊐(O3) → 2Al3+(M2) + O2−(O3)
critical for applications, such as telescope mirrors, in which
large volumes are cast, to ensure constant thermal expansion The former substitution gives a maximum zinc unit-cell for-
over the entire volume. This glass-ceramic can be readily pol- mula of Al3.5Zn0.5Si2.0O9.75 (7Al2O3⭈2ZnO⭈8SiO2 stoichi-
ished to optical quality as required for reflective mirrors. Ze- ometry), whereas the latter gives a maximum formula of
rodur威 also is used as the base block for ring laser gyroscopes, Al4.0Zn0.5Si1.5O9.5 (4Al2O3⭈ZnO⭈3SiO2 stoichiometry). These
which have generally replaced mechanical gyroscopes in air- zinc substitutions would be favored because the excess charge
planes. The gyroscopes operate on the principle of interfering on the tribridging oxygen would be removed.
clockwise and counterclockwise laser beams. The resulting Transparent mullite glass-ceramics can be produced from
beat frequency is carefully monitored and provides excellent simple binary glasses in the Al2O3–SiO2 system.16 It long has
accuracy. Recently, a giant ring gyroscope cast in a 1.2 m been known that binary Al2O3–SiO2 glasses with >10 mol%
square slab of Zerodur威 was installed underground in New Al2O3 spontaneously phase separate on cooling to two amor-
Zealand to record minute fluctuations in the earth’s rotation.13 phous phases, with mullite crystallizing spontaneously within
the Al2O3-rich regions. Additions of components—such as
(3) Transparent Mullite Glass-Ceramics B2O3, MgO, ZnO, BaO, and alkali—improve the melting char-
Mullite is a nonstoichiometric compound whose composi- acteristics of these otherwise refractory glasses as well as sup-
tion generally lies between 3Al2O3⭈2SiO2 and 2Al2O3⭈SiO2. Its press their spontaneous phase separation, yielding stable

Radiant stovetop composed of nanocrystalline ␤-quartz solid-solution glass-ceramic with near-zero coefficient of thermal expansion (photo-
graph courtesy of Eurokera, S.N.C.).
January 1999 Nanophase Glass-Ceramics 9

Table II. Transparent Glass-Ceramics Based on

Mullite Crystals
Composition (wt%)
Oxide M1 M2 M3 M4 M5
SiO2 50 60 48 48 53
Al2O3 40 20 29 25 25
BaO 10
B2O3 17.5 11 15 15
Na2O 2.5
ZnO 10 10
K2O 2 2 3
MgO 4

glasses that phase separate controllably on reheating to Al2O3-

rich droplets within a continuous siliceous glass. Fine mullite
spherulites then crystallize in a controlled manner within these
droplets, yielding a transparent glass-ceramic.
Representative mullite glass-ceramic compositions are given
in Table II. X-ray diffractometry (XRD) patterns correspond-
ing to glass-ceramics based on standard mullite (composition
M1), boron-doped mullite (M2), and metastable zinc- and mag-
nesium-doped mullites (M3 and M5, respectively), are shown
in Fig. 6. Note the shifts in intensities of major peaks, particu-
larly the 0.54 nm (2␪ ⳱ 16°) peak.
Transparent glass-ceramics, including those based on mul-
lite, spinel, and mixed ␤-quartz/spinel/magnesium-petalite,
have been studied as host media for luminescent transition- Fig. 6. Powder XRD patterns of glass-ceramics containing (A) stan-
metal ions, particularly Cr3+ ions.17 The Cr3+ ion is considered dard mullite (M1) (B) B2O3-doped mullite (M2), (C) metastable ZnO-
a key luminescent ion for potential use in tunable lasers and doped mullite (M3), and (D) metastable MgO-doped mullite (M5).
solar concentrators, because, in certain ligand field environ-
ments, the ion absorbs broadly in the visible wavelength range
yet fluoresces in the near infrared. Glasses have a sufficiently parasitic absorption and, therefore, lower quantum efficien-
large wavelength gap between Cr3+ absorption and fluores- cies. Mullite glass-ceramics also are compatible with sheet
cence (Stokes shift), but their quantum efficiency is generally manufacturing.
poor, <15%. (Quantum efficiency is defined as the number of Table III gives the composition and optical properties of a
photons emitted by Cr3+ ions divided by the number of photons transparent mullite glass-ceramic that was developed for pos-
absorbed by these ions.) Although certain crystalline phases, sible use as a luminescent solar concentrator.15,18 The product
such as aluminum phosphate (AlPO 4 ) and alexandrite concept comprised a flat sheet of chromium-doped glass-
(BeAl2O4), have similar shifts with efficiencies of ∼90%, crys- ceramic that would absorb visible light and luminesce in the
tals are not generally amenable to large-scale manufacturing in near infrared. Two edges of the sheet would be covered with
sheets similar to those necessary for solar collectors. Glass- silicon photovoltaic cells. The infrared radiation would be
ceramics can combine the best of the crystal and glass worlds. guided and concentrated by multiple internal reflections to the
Numerous studies using absorption and steady-state lumines- photovoltaic cells, where it would couple efficiently with ap-
cence have shown that Cr3+ in glass-ceramics exhibits excep- propriately doped silicon to produce electric energy. This
tionally high quantum efficiency of luminescence compared to would require a material with low scattering coefficient at high
that in glasses of the same composition. quantum efficiency. There also should be little or no overlap
Of these materials, chromium-doped mullite glass-ceramics between the absorption bands and the luminescence, which
have shown the most promise, combining good transparency should be Stoke-shifted far enough into the infrared to mini-
with broad visible absorption, wide emission in the infrared, mize overlap.
and high quantum efficiencies. These results are caused in part Figure 7 shows TEM micrographs of a mullite precursor
because the mullite crystal structure provides a suitable crystal glass (composition M3, Table II) that has fine phase separation
field environment for Cr3+, where Cr3+ substitutes for Al3+ in and of heat-treated materials between maximum temperatures
the octahedral sites. There are no suitable octahedral sites in of 800°–875°C. The rodlike mullite crystals are ∼30 × 10 nm
␤-quartz glass-ceramics; for example, Cr3+ in the crystal and in size at the lowest heat treatment. Although the parent glass
glass exhibit similar efficiencies. Furthermore, because mul- is green, the glass-ceramic appears gray with a pink tint from
lite glass-ceramics are self-nucleating, they do not require the the luminescence in the red and near infrared. Figure 8 shows
addition of nucleating agents. The use of TiO2 as a nucleating the difference in absorption between the M4 parent glass and
agent is particularly disadvantageous in that the small amount the glass-ceramic. The glass-ceramic has better transparency in
of Ti3+ that almost invariably accompanies the Ti4+ causes the key near-infrared region from 700° to 900 nm. The lumi-

Table III. Composition† and Optical Properties of a Transparent Mullite Glass-Ceramic

Scattering coefficient, Quantum efficiency
␴ (×103 cm−1) (0.1% Cr2O3)
Heat treatment (°C/h) 458 nm 633 nm 458 nm 514 nm 633 nm
Precursor 9±1
750–4800/2 h 68 ± 5 36 ± 3 0.23 0.28 0.23
750–4850/2 h 143 ± 5 55 ± 3 0.29 0.33 0.28
750–4875/4 h 204 ± 9 76 ± 3 0.33 0.38 0.31
†
48.0% SiO2, 29.0% Al2O3, 11% B2O3, 10.0 ZnO, 2.0% K2O, 0.0%–0.1% Cr2O3, and 0.4% As2O5.
10 Journal of the American Ceramic Society—Beall and Pinckney Vol. 82, No. 1

by using very pure batch ingredients. Divalent iron is a noto-

rious absorber of near-infrared radiation.

(4) Transparent Spinel Glass-Ceramics

Spinels have cubic crystal structures with the general chemi-
cal formula AB2O4, where A is a tetrahedrally coordinated,
typically divalent metal, such as zinc, magnesium, iron, or
manganese, and B is an octahedrally coordinated, usually tri-
valent metal, such as aluminum, iron, or chromium. Transpar-
ent glass-ceramics based on spinel compositions ranging from
gahnite (ZnAl2O4) toward spinel (MgAl2O4) can be crystal-
lized from glasses in the SiO2–Al2O3–ZnO–MgO system, with
ZrO2 and/or TiO2 as nucleating agents.1,19,20 These glass-
ceramics can be made highly transparent, with spinel crystals
on the order of 10–50 nm in size. The compositions of some
transparent spinel glass-ceramics are listed in Table IV.
Figure 9 shows the microstructure of a typical ZrO2-
nucleated spinel glass-ceramic (composition S1 in Table IV).
Fig. 7. TEM micrographs of the mullite precursor glass (top) and The phase assemblage consists of crystals of spinel solid so-
glass-ceramic materials after three heat treatments (maximum tem- lution and tetragonal-ZrO2 dispersed throughout a continuous
peratures of 800°–875°C). Scale bar indicates 0.1 ␮m. residual glass. Transparent glass-ceramics based on spinel solid
solution are typically 30%–40% crystalline by volume. Be-
cause of the relatively high thermal expansion of the spinel
nescence curves of the glass and glass-ceramic also are given phase (MgAl2O4 crystals have a CTE of 75 × 10−7/°C), these
in Fig. 8. (The glass curve has been multiplied by 6). The glass-ceramics generally have bulk CTEs of 30 × 10−7/°C to
glass-ceramic emits widely from 680 to >1000 nm. The glass- 50 × 10−7/°C.
ceramic is far more efficient, although the glass emits further The use of TiO2 or mixtures of TiO2 and ZrO2 for nucleation
into the infrared, and there is very little limited overlap be- rather than all ZrO2 allows for lower melting temperatures of
tween the glass-ceramic absorption and its emission. the precursor glass. TiO2 also is more efficient than ZrO2 in
The scattering coefficients and quantum efficiency as a func- nucleating spinel in glasses that contain significant levels of
tion of heat treatment with 0.1% Cr2O3 doping are given in MgO. Although ZrO2 acts as a nucleating agent in the classic
Table III. The scattering coefficient approaches 0.04 cm−1 in sense, with precipitated crystals of tetragonal-ZrO2 serving as
the material that was given the 800°C heat treatment. Although the nucleating sites for spinel crystals, no evidence of a crys-
this is good, it remains an order of magnitude too high for talline titanate phase is generally observed in TiO2-nucleated
practical considerations, because, for each scattering event, a spinel glass-ceramics. Instead, nucleation is preceded by a
cone of escaping radiation that is not internally reflected can highly uniform, ultra-fine-scale phase separation into SiO2-rich
cause dramatic deterioration of the collector efficiency. The and TiO2/Al2O3-rich areas. Spinel then crystallizes and grows
quantum efficiency of this glass-ceramic is ∼30%, which is far within the latter globular-shaped areas. TiO2 also serves as an
better than any known glass but is believed to be a factor of 2 integral component of the spinel structure, with Ti4+ entering
lower than what is required from figure-of-merit consider- the octahedral (B) site by means of a charge-balanced substi-
ations. Perhaps significant improvements can be made here by tution, such as Ti4+ + (Zn,Mg)2+ for 2Al3+ on the octahedral
carefully reducing the level of iron oxide in the glass, such as sites.

Transparent glass-ceramic cookware (VISION威) based on nanocrystals of low-thermal-expansion-coefficient ␤-quartz solid solution (photograph
courtesy of Corning Consumer Products Company).
January 1999 Nanophase Glass-Ceramics 11

Fig. 9. TEM micrograph of a transparent, ZrO2-nucleated spinel

glass-ceramic. Scale bar indicates 0.1 ␮m.

(Table IV), for example, is designed to be slightly peralumi-

nous; i.e., given a spinel formula of (Zn,Mg)Al2O4, or 1 mol of
ZnO + MgO for each mole of Al2O3, there is slightly extra
Al2O3, which remains in the residual glass. (This simplified
calculation ignores the role of titanium in the spinel phase.) All
of the fluxes, in this case ZnO and MgO, are assumed to enter
the crystal phase, leaving a residual glass with a minimum of
nonbridging oxygen atoms. This continuous siliceous glass en-
sures a high use temperature—often >900°C—in these glass-
ceramics, as well as excellent chemical durability.
The thermal expansion characteristics of these glass-
ceramics are very similar to those of elemental silicon. Figure
Fig. 8. Absorption (optical density) and luminescence as a function 10 shows the expansion curves from 0° to 1000°C of silicon21
of wavelength for mullite parent glass (dashed line) and glass-ceramic and of glass-ceramic S6 (Table IV). Potential applications for
(solid line). Glass luminescence curve has been multiplied by 6.
these glass-ceramics include those where transparency and
high use temperature and a thermal expansion match to silicon
are important, including flat panel displays and certain photo-
Spinel glass-ceramics generally have excellent thermal sta- voltaic substrates.22
bility. Glass-ceramics can have strain point temperatures sig- Given the structural similarities (well-defined octahedral and
nificantly higher than those of their precursor glasses, because tetrahedral sites) in spinel and mullite, one would expect simi-
many of the fluxes in the glass partition into the crystal phases lar crystal chemical behavior with respect to transition ele-
during the crystallization process, leaving a residual glass with ments. The crystal field relationships of transition elements in
fewer fluxes and, therefore, a glass ‘‘harder’’ than the precur- spinel have been studied extensively.23 Most of the common
sor glass. Spinel glass-ceramics, in particular, can be designed transition-element ions, including Cr3+, Ni2+, V3+, and Cu2+,
such that all the fluxes partition into the crystals, leaving a have high octahedral site preference energies and strongly par-
hard, highly siliceous continuous glassy phase. Composition S1 tition into the octahedral sites of spinel. Evidence of such par-

Table IV. Transparent Glass-Ceramics Based on Spinel Solid Solution

Composition (wt%)
Oxide S1 S2 S3 S4 S5 S6
SiO2 64.8 59.9 50.0 57.3 57.0 58.3
Al2O3 18.5 17.9 20.0 18.5 18.0 20.2
ZnO 4.6 14.3 13.0 11.6 14.0 8.4
MgO 4.6 7.0 2.4 4.2
Rb2O 1.5
Cs2O 1.5 4.0
PbO 2.0
BaO 2.0 2.0
ZrO2 7.5 3.0 7.0 5.0
TiO2 5.0 5.7 5.0 3.0
As2O5 0.5 0.5 0.5 0.5 0.5 0.5
Heat treat 900/6 800/1 700/1 800/1 800/1 800/1
(°C/h) 1060/6 950/2 850/3 950/2 950/2 950/2
CTE, 0°–300°C 34 33 48 38 37 38
(×10−7/°C)
12 Journal of the American Ceramic Society—Beall and Pinckney Vol. 82, No. 1

Table V. Composition and Properties of Some Transparent Oxyfluoride Glass-Ceramics

Composition (mol%)
Tick et al. Dejneka
Component Wang–Ohwaki No. 1 No. 2 No. 1 No. 2
SiO2 32.5 30 30 48.5 65.0
Al2O3 8.1 15 15 25.1 16.5
CdF2 21.6 29 29
PbF2 25.9 17 22
ZnF2 5
YF3 4 4
YbF3 10.8
ErF3 1.1 0.1
PrF3 0.01 0.01 0.07
LaF3 13.1 12.6
AlF3 2.5 1.6
Na2O 10.7 4.3
Refractive index 1.75 1.75 1.55
CTE (×10−7/°C) 110
Tg (°C) 400 395 408 570
Cubic lattice edge (nm) 0.572 0.575 0.575
(Pb, Cd, RE)F2,3

titioning in spinel glass-ceramics is readily observed when energies and broad transparency in the infrared region of the
these materials are crystallized. A Cr3+-doped spinel material is spectrum used for amplification of light in telecommunications
emerald green as a precursor glass and pink as a glass-ceramic, systems. Fluoride glasses have relatively poor chemical and
for example, whereas a Ni2+-doped material is brown and blue mechanical stability compared with oxide-based glasses, how-
before and after crystallization. Compared with the mullite ever, and, as a result, are difficult to prepare and handle. Re-
glass-ceramics, however, transparent spinel glass-ceramics cently, transparent oxyfluoride glass-ceramics, based on fluo-
have offered fewer advantages for the luminescent solar col- ride crystals dispersed throughout a continuous silicate glass,
lector application. Unlike mullite, for example, spinel glass- have been shown to combine the optical advantages of rare-
ceramics require a nucleating agent. Moreover, Cr3+-doped earth-doped fluoride crystals with the ease of forming and han-
spinel glass-ceramics do not absorb as broadly as mullite glass- dling of conventional oxide glasses. Table V gives the compo-
ceramics across the entire range of solar radiation (compare the sitions and properties of several of these oxyfluoride materials.
pink color of Cr3+-doped spinel with the gray color of Cr3+- Wang and Ohwaki25 discovered that 20 nm cadmium lead
doped mullite). It is possible that compositional and process fluoride crystals codoped with erbium and ytterbium could be
research can overcome these barriers. precipitated from aluminosilicate glasses, thereby yielding
The potential use of transparent, Cr3+-doped gahnite glass- transparent oxyfluoride glass-ceramics. They were able to use
ceramics as a laser medium has been evaluated recently using the resonant energy transfer between ytterbium and erbium to
excited-state absorption. Although the material is characterized produce upconversion and fluorescence in the blue-green. With
by a very structured and wide emission band from 685 to 750 an excitation wavelength of 0.97 ␮m, the measured upconver-
nm, the question of its potential as an efficient laser medium sion emission intensity from the glass-ceramic was 100 times
has not been resolved.24 that of the precursor oxyfluoride glass and 2–10 times that of
(5) Transparent Oxyfluoride Glass-Ceramics
Fluoride crystals and glasses are desirable hosts for optically
active rare-earth (RE) cations because of their low phonon

Fig. 10. Thermal expansion curves of silicon and transparent spinel Fig. 11. TEM micrograph of a transparent LaF3 glass-ceramic. Scale
glass-ceramic. bar indicates 0.1 ␮m.
January 1999 Nanophase Glass-Ceramics 13

Table VI. Fluorescent Properties of Praseodymium-Doped Oxyfluoride Glass-Ceramics versus ZBLAN

Value
Tick et al. No. 1† Tick et al. No. 2†
Dejneka No. 2 ZBLAN
Property Glass Gl-Cer Glass Gl-Cer Gl-Cer Glass
Crystal phase (Cd,Pb,Zn,Y)F2,3 (Cd,Pb,Y)F2,3 LaF3
Radiative lifetime, ␶R, 2169 2036 2052 1920 3000 2919
Judd–Ofelt (␮s)
Measured lifetime, ␶m, 0 115 0 130 240 110
at 1.3 (␮s)
Calculated quantum 0 7 0 7 8 4
efficiency (%)
Measured quantum 0 8 0 9 4
efficiency (%)
†
Tick et al., No. 1 and No. 2: 125 ppm Pr3+.

an aluminofluoride glass, indicating excellent upconversion ef- is 7%–8% efficiency for the glass-ceramics compared with 4%
ficiency. These oxyfluoride glass-ceramics have potential ap- for ZBLAN. The quantum efficiencies have been measured
plications in blue-green laser devices. directly by Quimby et al.30 for the two glass-ceramics of Tick
Tick et al.26 subsequently described ytterbium-free oxyfluo- et al. with results of 8%–9% quantum efficiency. It is, there-
ride glass-ceramics based on praseodymium-doped nanocrys- fore, apparent that isolating the characteristic RE, in this case
tals of cadmium lead fluorides containing yttrium and zinc. praseodymium, in the fluoride crystal can produce efficiencies
Their objective was to produce materials with efficient lumi- that rival or surpass those of fluoride glasses.
nescence at 1.3 ␮m that could be used as amplifier hosts for In addition to their interesting fluorescence properties, the
telecommunications. XRD data and TEM indicate that these glass-ceramics described by Tick et al. have unusual clarity
glass-ceramics are 15%–30% by volume crystal sizes of 9–18 and extraordinarily low optical scattering. The theoretical
nm. Fluorescence measurements indicate significant partition- losses of scattering in these materials are described in Fig. 12,
ing of the optically active RE cations into the fluoride crystals, in which particle separation is plotted against the particle ra-
with their behavior essentially determined by their local fluo- dius,26 defining several lines of volume percent crystallinity
ride, low-phonon-energy environment. Fluorescence lifetimes, radiating from the origin. There is a major difference in the
however, are only slightly better than those observed in the scattering losses according to the independent particle model
fluorozirconate glass ZBLAN, probably because of clustering. and the quasi-continuum model. The dashed lines show the
Recently, Dejneka27 described another family of transparent predicted losses in square brackets in dB/km according to the
oxyfluoride glass-ceramics, based on 15 nm crystals of LaF3 in Rayleigh–Gans model. The solid lines show much smaller
an aluminosilicate glass (Fig. 11), that provides the desirable losses using the assumptions of the quasi-continuum model.
fluoride environment for active RE ions while maintaining the The data are for a wavelength of 1300 nm and an assumed
processing advantages of an oxide glass. LaF3 is an ideal host index difference ⌬n ⳱ 0.1 between the crystals and the re-
for RE cations, because it has extensive solid solution with all sidual glass. The shaded area in the diagram indicates the typi-
RE ions. Eu3+ emission spectroscopy has demonstrated that cal microstructure—particle size and separation—as observed
active RE ions partition into the low-phonon-energy LaF3 crys- by TEM. The actual turbidity and measured scattering coeffi-
tals without clustering. Fluorescence and lifetime measure- cient for the oxyfluoride glass-ceramic as well as the parent
ments indicate that these materials are superior to ZBLAN both glass and a typical high-quality Schott optical glass are shown
for Er3+ amplifiers because of greater width and gain flatness of in Table VII. The scattering losses for the glass-ceramic range
the 1530 nm emission band and for Pr3+ amplifiers because of between 73 and 166 dB/km, which lies in the range where the
their higher quantum efficiency. quasi-continuum model is accurately predictive. The oxyfluo-
Table VI compares the quantum efficiencies of the Pr3+- ride glass-ceramic is only very slightly more scattering than the
doped oxyfluoride glass-ceramics of Tick et al. and Dejneka parent glass and the Schott optical glass.
with those of their parent glasses and that of the conventional
fluoride glass ZBLAN. The quantum efficiency can be calcu-
lated by comparison of the radiative lifetime, calculated from III. Tough Glass-Ceramics for
Judd–Ofelt theory, with the measured lifetime.28,29 The result Microengineered Surfaces
Nanocrystalline glass-ceramics with high elastic modulus
and moderate strength and toughness have been developed re-
cently for potential use as magnetic disk substrates in computer
hard drives.31 These glass-ceramics, based on spinel and en-
statite phase assemblages in the SiO2–Al2O3–MgO–ZnO–TiO2
system, have grain sizes of <100 nm and can be polished to an
average roughness of 0.5–1.0 nm. The majority of disk drives
are built with disks of NiP-plated aluminum, with chemically
strengthened glass disks also being used in certain products.32
High-modulus glass-ceramics offer many different advantages
for the disk substrate application, including improved flatness
and smoothness, better wear, resistance to flutter at high rota-
tional speeds, and the ability to be manufactured thinner than
glass or aluminum disks. They also offer excellent chemical
durability and are devoid of alkali ions that can cause contami-
nation problems under extreme use conditions.
Compositions within the cordierite liquidus field of the ter-
nary MgO–Al2O3–SiO2 system are easily nucleated with TiO2
Fig. 12. Theoretical scattering losses in nanocrystalline glass- to give a wide range of possible phase assemblages based on
ceramics. (After Tick et al., 1995.) various mixtures of cordierite, ␤-quartz solid solution,
14 Journal of the American Ceramic Society—Beall and Pinckney Vol. 82, No. 1

Table VII. Optical Scattering in Transparent Oxyfluoride Glass-Ceramics†

Turbidity, ␴ Scattering Particle
(measured scattering loss size, a
Sample coefficient, cm−1) (dB/km) (radius, nm) Volume (%) W/a‡
−4
SF-6 Schott optical glass 1.25 × 10 54.3
Oxyfluoride No. 1 1.52 × 10−4
glass
Oxyfluoride No. 1 1.68 × 10−4 73.0 17.0 35 2.4
glass-ceramic
(475°C/45 min)
Oxyfluoride No. 1 3.82 × 10−4 166.0 14.7 30 2.6
glass-ceramic
(460°C/120 min)
†
␭ ⳱ 467 nm. ‡W is interparticle spacing. After Tick et al., 1995.

Fig. 13. TEM micrograph of a spinel–enstatite glass-ceramic, with

minor magnesium dititanate and magnesium-petalite. Scale bar indi-
cates 0.1 ␮m.

␣-quartz, magnesium-petalite, spinel, and enstatite. Given a

sufficiently high level of TiO2, a very-fine-grained glass-
ceramic based on a metastable assemblage of spinel, enstatite,
and magnesium dititanate can be obtained. All three phases
occur as crystals <100 nm in size, with spinel crystals <25 nm.
Experimental ring laser gyroscope fabricated from zero-expansion
Spinel provides the glass-ceramic with high elastic modulus transparent glass-ceramic Zerodur威 by Schot Glaswerke.
and surface hardness, and the enstatite improves fracture
toughness, because of the well-known lamellar twinning that
develops during the martensitic transition from protoenstatite
to clinoenstatite on cooling.33 Such twinning is known to de- tures to deflect around the crystals, producing a tortuous frac-
flect cracks and increase fracture toughness. ture pattern.
Nucleation in these materials is preceded by a highly uni- Figure 13 shows the microstructure of a glass-ceramic con-
form, fine-scale phase separation into SiO2-rich and TiO2/ taining spinel and enstatite, with minor magnesium dititanate
Al2O3-rich areas, with the initial crystalline phase consisting of and magnesium-petalite, whose composition in weight per cent
magnesium aluminum titanate. The morphology of this initial is 47.1 SiO2, 22.1 Al2O3, 16.9 MgO, 1.7 ZnO, and 12.3 TiO2.
phase separation provides a template for the ultimate crystal- This material has a Young’s modulus value of 20.9 × 106 psi
line microstructure, essentially locking in its nanocrystalline (144 GPa), an abraded modulus of rupture value of 15 500 psi
nature such that, regardless of heat treatment, the crystals grow (107 MPa), and a fracture toughness of 1.3 MPa⭈m1/2. This
no larger than 100–200 nm. fracture toughness, although lower than that of aluminum or
The strength and toughness values of these glass-ceramics, coarser-grained glass-ceramics, continues to be ∼60% higher
even those with very little enstatite, are higher than might than that of glass. Table VIII provides a comparison of the
otherwise be expected from such a fine-grained microstructure, properties of this glass-ceramic with those of aluminum and of
possibly because of the phenomenon known as dispersion chemically strengthened glass. The material can be polished to
strengthening.34 Spinel crystals have both a higher density and very fine tolerances, in many cases to surfaces that are
a higher thermal expansion than the surrounding residual glass. smoother than those of alkali-containing glass, which is subject
The residual glass upon cooling is, therefore, brought into tan- both to fracture propagation through stress corrosion and ease
gential compression around the crystals and into radial tension of fracture propagation through low toughness. With a strain
perpendicular to the interface. These stresses cause any frac- point of 1000°C, these glass-ceramics also have upper use tem-
January 1999 Nanophase Glass-Ceramics 15

Table VIII. Comparison of Properties of Spinel–Enstatite Glass-Ceramics, Aluminum,

and Chemically Strengthened Glass
Value
Aluminosilicate Spinel–enstatite
Property Aluminum glass glass-ceramic
Alkali in composition Yes No
Hardness, Knoop (kg/mm2) 100 650 790
Young’s modulus (×106 psi)† 10 11 21
Strength, ring-on-ring (×103 psi)† 17 55‡ 32
Fracture toughness (MPa⭈m1/2) >0.2 0.8 1.3
Density (g/cm3) 2.7 2.6 3.0
CTE (×10−7/°C) 240 90 70
Upper use temperature (°C) 270§ ∼400 950
Thermal conductivity (W/(m⭈K)) 27 ∼3 ∼3
Surface roughness, Ra (nm) 1.0 1.0 0.5
Crystal size (nm) >10000 10–100
†
106 psi ⳱ 6.9 × GPa, and 103 psi ⳱ 6.9 MPa. ‡Ion-exchanged strength, Na+ ↔ Li+. §Nickel in NiP layer becomes magnetic
above 270°C.

9
peratures much higher than those of glass or aluminum, en- Z. Strnad, Glass-Ceramic Materials. Elsevier, Amsterdam, The Netherlands,
abling the use of the high-temperature sputtering processes 1986.
10
J. Petzoldt, ‘‘Metastabile Mischkristalle mit Quarzstruktur mit Oxidsystem
required to use ultra-high coercivity barium ferrite films. Li2O–MgO–ZnO–Al2O3–SiO2,’’ Glastech. Ber., 40, 385–96 (1967).
11
D. R. Stewart, ‘‘TiO2 and ZrO2 as Nucleants in a Lithia Aluminosilicate
IV. Summary Glass-Ceramic’’; pp. 83–90 in Advances in Nucleation and Crystallization in
Glasses. Edited by L. L. Hench and S. W. Freiman. American Ceramic Society,
Nanophase microstructures are achieved in glass-ceramic Columbus, OH, 1971.
12
J. Petzoldt and W. Pannhorst, ‘‘Chemistry and Structure of Glass-Ceramic
materials through efficient nucleation and slow crystal growth. Materials for High-Precision Optical Applications,’’ J. Non-Cryst. Solids, 129,
Transparency occurs when the crystals are small enough, low 191–98 (1991).
enough in birefringence, or well matched in refractive index 13
R. W. Hardin, ‘‘Ring Laser Measures Earth’s Rotation,’’ Photonics Spec-
with the residual glass. Highly crystalline glass-ceramics based tra, [Dec.] 35 (1997).
14
W. E. Cameron, ‘‘Mullite: A Substituted Alumina,’’ Am. Mineral., 62,
on stuffed derivatives of ␤-quartz can meet these requirements 747–55 (1977).
and display excellent thermal shock resistance because of their 15
G. H. Beall, D. R. Cole, D. W. Hall, H. J. Holland, J. W. H. Schreurs, M.
very low coefficients of thermal expansion. Less crystalline but Edwards, R. Gilles, and A. Lempicki, ‘‘Physical and Optical Properties of
just as transparent are nanocrystalline spinel and mullite glass- Chromium-Doped Glass-Ceramics,’’ Report to D. O. E. under Grant No. DE-
FG02-84ER45103, 1987.
ceramics whose grain size is extremely fine, <50 nm. These 16
J. F. MacDowell and G. H. Beall, ‘‘Immiscibility and Crystallization of
materials have increased strain points over their parent glasses Al2O3–SiO2 Glasses,’’ J. Am. Ceram. Soc., 52, 17–25 (1969).
and selectively partition transition metals such as chromium 17
R. Reisfeld, A. Kisilev, A. Buch, and M. Ish-Shalom, ‘‘Transparent Glass-
into their structures, thereby producing unique, near-infrared Ceramics Doped by Chromium(III): Spectroscopic Properties and Characteriza-
tion of Crystalline Phases,’’ J. Non-Cryst. Solids, 91, 333–50 (1987).
luminescent materials. Rare-earth ions that fluoresce farther 18
L. J. Andrews, G. H. Beall, and A. Lempicki, ‘‘Luminescence of Cr3+ in
into the infrared can be selectively partitioned into fluoride Mullite Transparent Glass-Ceramics,’’ J. Lumin., 36, 65–74 (1986).
crystals in transparent oxyfluoride glass-ceramics. Other mul- 19
L. R. Pinckney, ‘‘Transparent Glass-Ceramics Containing Gahnite,’’ U.S.
tiphase nanocrystalline glass-ceramics, although not transpar- Pat. No. 4 687 750, Aug. 18, 1987.
20
N. Gotoh and M. Hayasaka, ‘‘Glass-Ceramics,’’ U.S. Pat. No. 5 028 567,
ent, possess excellent hardness and high stiffness and can be July 2, 1991.
polished to extreme smoothness. 21
C. A. Swansen, ‘‘Recommended Values for the Thermal Expansivity of
Current applications for nanocrystalline glass-ceramics in- Silicon from 0 to 1000 K,’’ J. Phys. Chem. Ref. Data, 12, 179–82 (1983).
22
clude telescope mirrors, laser gyroscopes, radiant stovetops, L. R. Pinckney, ‘‘Transparent, High Strain Point Glass-Ceramics’’; in Pro-
ceedings of the XVIII International Congress on Glass, Section E4. American
cookware, woodstove windows, and fire doors. Future appli- Ceramic Society, Westerville, OH, 1998.
cations for these materials are likely to capitalize on designed- 23
R. G. Burns, Mineralogical Applications of Crystal Field Theory; pp. 109–
in, highly specialized properties for the transmission, display, 35. Cambridge University Press, Cambridge, U.K., 1970.
24
and storage of information. Potential products include ad- C. Koepke, K. Wisniewski, M. Grinberg, D. L. Russell, K. Holliday, and
G. H. Beall, ‘‘Excited State Absorption in Cr3+-Doped Gahnite Glass-
vanced flat-panel display substrates, solar panels, optical and Ceramics,’’ J. Luminescence, 78, 135–46 (1998).
optoelectronic components and devices, and data storage 25
Y. Wang and J. Ohwaki, ‘‘New Transparent Vitroceramics Codoped with
devices. Er3+ and Yb3+ for Efficient Frequency Upconversion,’’ Appl. Phys. Lett., 63
[24] 3268–70 (1993).
26
Acknowledgments: The authors would like to thank M. Dejneka and P. A. Tick, N. F. Borrelli, L. K. Cornelius, and M. A. Newhouse, ‘‘Trans-
P. Tick for helpful discussions and N. Foster for preparing many of the parent Glass-Ceramics for 1300 nm Amplifier Applications,’’ J. Appl. Phys., 78,
figures. 93–100 (1995).
27
M. J. Dejneka, ‘‘Rare-Earth Fluorescence in Novel Oxyfluoride Glasses
and Glass-Ceramics,’’ Proc. Soc. Photo-Opt. Instrum. Eng., 3280, 132–37
References (1998).
1 28
G. H. Beall and D. A. Duke, ‘‘Transparent Glass-Ceramics,’’ J. Mater. Sci., B. R. Judd, ‘‘Optical Absorption Intensities of Rare-Earth Ions,’’ Phys.
4, 340–52 (1969). Rev., 127 [3] 750–61 (1962).
2 29
M. Kerker, The Scattering of Light. Academic Press, New York, 1969. G. S. Ofelt, ‘‘Intensities of Crystal Spectra of Rare-Earth Ions,’’ J. Chem.
3 Phys., 37, 511–20 (1962).
N. S. Andreev, ‘‘Scattering of Visible Light by Glasses Undergoing Phase
30
Separation and Homogenization,’’ J. Non-Cryst. Solids, 30, 99–126 (1978). R. S. Quimby, P. A. Tick, N. F. Borrelli, and L. K. Cornelius, ‘‘Quantum
4
R. W. Hopper, ‘‘Stochastic Theory of Scattering from Idealized Spinodal Efficiency of Pr3+ Doped Transparent Glass-Ceramics,’’ J. Appl. Phys., 83,
Structures: II. Scattering in General and for the Basic Late Stage Model,’’ J. 1649–53 (1998).
31
Non-Cryst. Solids, 70, 111–42 (1985). L. R. Pinckney and G. H. Beall, ‘‘Nanocrystalline, Non-Alkali Glass-
5 Ceramics,’’ J. Non-Cryst. Solids, 219, 219–27 (1997).
G. Tamman, The States of Aggregation. Van Nostrand Reinhold, New York,
32
1925. C. Krishnan, ‘‘Glass Platters for Small HDDs Provide Reliable High-
6 Density Storage,’’ Data Storage, [Mar./Apr.] 55–60 (1995).
P. W. McMillan, Glass-Ceramics; p. 29. Academic Press, London, U.K.,
33
1979. G. H. Beall, ‘‘Chain Silicate Glass-Ceramics,’’ J. Non-Cryst. Solids, 129,
7 163–73 (1991).
D. C. Palmer, ‘‘Stuffed Derivatives of the Silica Polymorphs’’; pp. 83–122
34
in Reviews in Mineralogy, Vol. 29. Mineralogical Society of America, Wash- D. P. H. Hasselman and R. M. Fulrath, ‘‘Proposed Fracture Theory of a
ington, DC, 1994. Dispersion-Strengthened Glass Matrix,’’ J. Am. Ceram. Soc., 49, 68–72
8
G. H. Beall, ‘‘Industrial Applications of Silica’’; see Ref. 7, pp. 469–505. (1966). 䊐
16 Journal of the American Ceramic Society—Beall and Pinckney Vol. 82, No. 1

George H. Beall is a Research Fellow in the Science and Technology Division of

Corning Incorporated. He jointed Corning after receiving his Ph.D. in geology from
the Massachusetts Institute of Technology in 1962. He also holds degrees (B.Sc. and
M.Sc.) in physics and geology from McGill University in Montreal. Dr. Beall has
published some 70 technical papers, has edited one book, and holds 81 patents. He
is a Courtesy Professor in the Department of Materials Science and Engineering at
Cornell University. Dr. Beall is a Fellow of The American Ceramic Society and has
received the Morey Award in 1988, The John Jeppson Award in 1993, the Samuel
Geijsbeek Award in 1993, and delivered the Edward Orton Memorial Lecture in
1996. He also has been awarded the American Chemical Society’s Chemistry of
Materials Award in 1995 and the Corning Section’s Sullivan Award in 1997. He has
conducted investigations in the areas of transparent glass-ceramics, machinable glass-
ceramics, strong and tough glass-ceramics, low-temperature glasses for glass–
polymer blends, low- and negative-CTE materials, and photosensitive and polychro-
matic glasses. His continuing interest involves the relation between microstructure
and properties of multiphase glass-derived materials.

Linda R. Pinckney is a Senior Research Associate in the Science and Technology

Division of Corning Incorporated. She received her B.S. degree in earth and space
sciences from the State University of New York at Stony Brook in 1976 and her A.M.
and Ph.D. in Mineralogy from Harvard University in 1978 and 1982, respectively,
joining Corning in 1982. She holds 12 U.S. patents in glass and glass-ceramic
materials, and she is a member of The American Ceramic Society, the Materials
Research Society, and the Mineralogical Society of America. Her current research
interests include the thermal and optical properties of and new applications for
transparent glass-ceramics.