Content uploaded by Charles Robert Kurkjian
Author content
All content in this area was uploaded by Charles Robert Kurkjian on Oct 22, 2017
Content may be subject to copyright.
Perspectives on the History of Glass Composition
Charles R. Kurkjian*
Bell Communications Research (Bellcore), Morristown, NewJersey 07960
William R. Prindle*
,†
Santa Barbara, California 93105
The 100th anniversary of The American Ceramic Society
corresponds approximately with the 100th anniversary of
what might be considered the start of the age of glass sci-
ence, i.e., the publication, in Germany, in 1886, of the cata-
log of Schott und Genossen, containing 44 optical glass
compositions. The American Ceramic Society centennial
seems, accordingly, to be an appropriate occasion to exam-
ine the history of glass composition that both preceded and
followed the seminal work of Schott and to survey some of
the major discoveries and changes in glass composition as
well as the reasons that led to them. Although it is certainly
of interest to consider a more complete history of the glass
industry, we have opted to attempt the more modest task
just described. The history of glass structure theories is
considered, particularly with regard to the effects of com-
position on structure, and how these relate to glass prop-
erties. The article then continues with a discussion of recent
special glasses and concludes with a description of light-
guide glasses, the discovery of which has changed the na-
ture of glass science and the glass industry.
I. Introduction
THE 100th anniversary of The American Ceramic Society
corresponds approximately with the 100th anniversary of
what might be considered the start of the age of glass science—
the disclosure, in 1886, of the work of Otto Schott and Ernst
Abbe in Germany. This was the publication of the catalog of
the ‘‘Glastechnisches Laboratorium, Schott und Genossen’’
(see Fig. 1).
1
Up to this time, very little real glass science had
been done, although, with the limited tools at their disposal,
earlier workers did quite remarkable things. Most work was
done in an attempt to understand what soda–lime–silica glasses
were and to improve their quality. Schott
2
conducted detailed
studies of the effects of various additions and substitutions to
the basic soda–lime glass composition. He and Winkelmann
3,4
were the first to attempt to model glass behavior development
by means of a set of factors with which properties could be
calculated.
As a result of the coincidental natural occurrence of alkali,
alkaline-earth ‘‘impurities,’’ and sand, soda–lime–silica glass
became the ‘‘staple’’ glass composition very early in time, and
it has continued to be so with only slight changes. Variations in
production techniques and specific use requirements have led
to the deliberate addition of a variety of other oxides, so that
most commercial ‘‘soda–lime’’ glasses now contain six or
more constituents. Over the years, the bulk of commercial
glasses for most purposes has continued to be based on silica as
the primary glass former.
Research (by X-ray technologies, optical spectroscopy,
physical property measurements, etc.) during the 20th century
has been conducted on simple glass compositions to attempt to
understand glasses as materials as well as to attempt to under-
stand their structure and properties well enough to predict prop-
erties from composition and to design a glass from a list of
requirements. Although glasses with rather remarkable proper-
ties ranging from infrared transmission and superionic conduc-
tivity to biological activity have been discovered, it is probably
not entirely accurate to say that we can design a glass for a
given purpose. Available commercial silicate glasses do their
job quite admirably, but they are rather complicated glasses
that fulfill rather simple tasks. In 1970, the discovery that a
simple (titania–silica)–silica compound glass fiber could con-
duct light over rather long distances without requiring ampli-
fication has resulted in a ‘‘new glass industry’’—the ‘‘light-
guide industry.’’ Since then, in somewhat of a turnabout,
scientists have discovered that these simple glasses display a
wide range of unexpected, new, complicated, and often incom-
H. A. Anderson—contributing editor
Manuscript No. 190479. Received December 30, 1997; approved February 16,
1998.
*Member, American Ceramic Society.
†
Retired from Corning Incorporated, Corning, NY.
J. Am. Ceram. Soc., 81 [4] 795–813 (1998)
Journal
centennialfeature
795
pletely explained properties. We attempt in this review to il-
lustrate some of the interesting events in this long history.
In this article we
●Provide a brief review of the early history of glass;
●Review the work of Abbe and Schott, i.e., the start of glass
science;
●Review the development of more-modern glass composi-
tions;
●Provide a brief review of the history of glass structure and
property relations;
●Bring the history up-to-date by discussing some new special
glasses and the new era of optical fibers.
Our purpose here is to provide an overview of the very large
field of inorganic glasses for the benefit, perhaps, of a re-
searcher new to glass. We attempt to provide a sense of the
present and future of glasses–properties–understanding as well
Fig. 1. Photograph of Otto Schott and pages from glass catalog.
796 Journal of the American Ceramic Society—Kurkjian and Prindle Vol. 81, No. 4
as why and how we arrived at this position. Because of the
many glasses that have been studied, we are forced to limit our
comments to compositions that illustrate major discoveries or
changes; accordingly, some important glass compositions are
not discussed.
Besides the apparent limitations imposed by practical con-
siderations and economics, physics appears to impose a real
limitation to the variation of properties available in materials
that lack a periodic crystalline lattice. Inorganic glasses are
generally considered to be isotropic; dielectric; transparent, un-
less colorants are added; chemically durable and, therefore,
chemically inert; and brittle.
These general properties, however, are constrained only in
those glasses that are normally thought of when we think of
glasses. If we broaden our chemical viewpoint to include chal-
cogenide, halide, and, especially, metallic glasses, a wide va-
riety of properties becomes available. Although the lack of a
crystalline lattice appears to impose a rather severe constraint
with regard to some properties at the moment, in most cases,
we are not in a position to state unequivocally whether these
constraints are absolute. For instance, it had been thought to be
impossible to make bulk metallic glasses because their hard
sphere structure results in rather simple dense packing, which
is, in itself, not conducive to extensive supercooling. However,
this recently has been shown to be untrue. Also, it normally is
considered that the lack of a crystalline lattice means that plas-
tic flow is not possible because the dislocations that result in
plastic flow cannot form. We have yet to determine whether
this is completely true, or if it can be sensibly modified.
There were few modern books in English on glass until the
publication of George Morey’s book, The Properties of Glass,
5
in 1938. After World War II, and following the publication of
the second edition of Morey in 1954, many other books ap-
peared. Recently, many edited books and edited proceedings
have appeared regularly.
6–12
The more general books usually
present a short history of glass, a definition and description of
glass, and chapters that present properties and compositions of
simple glasses, where the chapters are arranged either by prop-
erty or by glass composition. These books are important and
useful, especially as texts on the science and technology of
glass. There also have been many excellent review articles and
book chapters that present special subjects. Examples are the
two excellent series by Doremus and Tomozawa
12
and Uhl-
mann and Kreidl.
13
In particular, in the Uhlmann and Kreidl
series, the Kreidl chapter on glass-forming systems is very
useful. It historically, scientifically, and technologically dis-
cusses almost every known glass-forming system. Here we
attempt, by perhaps rather extreme simplification, to illustrate
some of the issues having to do with property–composition
development. We present our simplified and personal view of
some glass compositions–structures in order to make some
simple generalizations. This hopefully leads to a better general
understanding of what has been done, in many cases, empiri-
cally, and hopefully leads to the possibility of predicting what
remains possible. Such predictions were attempted at a meeting
to celebrate Kreidl’s 80th birthday.
14
The sections that follow immediately have to do with the
early history of glass. The reader is directed to the papers of
Cable
15–17
and symposia arranged by Kingery
18,19
for other
interesting insights into this history.
II. Early Glasses
(1) Middle Eastern Origins and Roman Growth
The earliest known synthetic glasses were created in Asia
Minor several millennia ago. Some isolated examples may be
as early as 7000 BC, but it is clear that, by 2500 BC, there were
many sources, probably first in Mesopotamia, then in Egypt.
The first glassmakers were motivated to create decorative ob-
jects, possibly to simulate gems and semiprecious stones, using
sintered bodies of silica and desert soda (natron) with appro-
priate colorants, such as copper, manganese, and iron salts.
There was no demonstrated interest in transparency at this
time. Beads also were made, and, later, small vessels were
constructed by coating sand cores with a glassy skin—the cores
were removed after forming.
The earlier and sometimes parallel development of ceramic
and metallurgical processes undoubtedly influenced and aided
the growth of early glass technology; some furnace improve-
ments and raw materials were applicable to glassmaking. The
extensive Egyptian tradition of faience making also must have
had an effect, contributing knowledge of raw materials and
sintering techniques.
About 2000 years ago, the blowing of glass articles with a
pipe was invented, probably in Syria, and this advance in tech-
nology was followed by a rapid increase in the use of glass-
ware. Glassblowing spread quickly through the Roman Empire,
and soon glass bowls and drinking vessels were in use through-
out society, in both ordinary households and among the ruling
classes. A desire for clear and transparent vessels came with
this remarkable growth of blown glass production. Accord-
ingly, strong efforts were directed toward the elimination of
iron and other contaminants, particularly for the higher-quality
glassware.
5
(2) Raw-Material Preparation—The Search for
Transparency
Early glasses in the western world were almost all soda–
lime–silica compositions that varied depending upon the avail-
ability of raw materials, but generally differed little from pres-
ent-day commercial glasses (Table I). Beach sand and a crude
source of alkali were typical ingredients, with both the sand
and the alkali containing enough lime or magnesia to give
chemical durability adequate for the purposes of the time. In
the case of the sand, fragments of shells provided some lime,
and plant ash generally brought some magnesia along with the
soda and potash.
Two different sources of alkali affected the composition of
early glasses. On the Eastern Mediterranean littoral natron (hy-
drated Na
2
CO
3
) was usually the favored alkali, because it was
available from northern Egypt (see, e.g., glass 2 in Table I).
20
Further east, in Mesopotamia and Persia, the alkali was usually
provided by plant ash that contained more K
2
O (2%–4%) and
MgO (2%–6%) (see, e.g., glass 3 in Table I).
21a
The alkali
content of the ash was influenced by the soil in which the plants
grew: plants that grew in salty soil or near the sea were high in
soda, whereas those that grew inland had higher potash con-
tents.
22
Agricola (1556)
23
refers to the use of salts made from
the ashes of salty herbs as well as to natron and ‘‘rock-salt.’’
When these were not available, he suggested the ashes of oak
could be used, or, as a last resort, the ashes of beech or pine.
The practice of using natron to produce higher-soda glasses
continued in the Mediterranean region through early and me-
dieval times. However, there was a surge in the use of potash
in glassmaking during the 9th through 13th centuries, before
soda again became the predominant alkali.
24
Much glass made in the Middle Ages was dark green, dark
brown, or almost black as a result of the impurities present.
This ‘‘waldglas,’’ or forest glass, often was used for bottles and
drinking vessels, but interest grew in preparing clearer, more-
transparent glass. Although little is known about glass technol-
ogy in the middle ages, we do know that some attention was
given to the purification of raw materials. One of the major
sources of glass technology information in this period comes
from L’Arte Vetraria,
25
written by Antonio Neri, an Italian
priest and glassworker, in 1612, and translated to English in
1662 by Christopher Merrett, an English physician and one of
the founders of the Royal Society. (It also was translated by
Johann Kunckel in 1679; both Merrett and Kunckel added
valuable personal observations on glassmaking.) Agricola and
Neri devoted considerable space to raw-material preparation,
discussing the careful selection of crystals (quartz) and clean
‘‘white stones free of black or yellow veins’’ to be used in
April 1998 Perspectives on the History of Glass Composition 797
place of sand if high clarity was desired. The stones were
reduced to fine particles by pounding in a mortar, and the silica
powder then often was fritted with the alkali salts. Neri gave
considerable attention to alkali preparation, discussing in some
detail the washing of various plant ashes to prepare alkali salts
for clear crystal glass. The purification process consisted of
repeated sieving of the raw salt, dissolving it in boiling water,
filtering, and evaporating. Thus the impurities causing color,
such as iron compounds, were left behind.
Unfortunately, much of the alkaline-earth and alumina of the
ashes were left behind as well; therefore, many clear glasses
prepared from the purified raw materials had relatively poor
resistance to attack by moisture. The much-admired clear
‘‘cristallo’’ glass produced in Venice–Murano in the early
1500s suffered from low lime and magnesia content. As a
result, many of the elegant examples of the elaborate Venetian
glass of that period now in museums have developed surface
crizzling (a multitude of fine surface fractures) because of their
poor chemical durability; some extreme examples are sticky to
the touch and appear to sweat.
26
These cristallo glasses provide
an example of an unintended consequence of the desire to
optimize one glass property, colorless clarity in this case, caus-
ing a deterioration in another property, durability.
(3) Colored Glasses
Although the preceding section described colorless glasses
purposely made free of unwanted color, other glasses were
colored purposely for decoration since the earliest days of
glassmaking. Glassworkers in Egypt, the Middle East, and the
Roman Empire knew that small amounts of certain salts could
be incorporated in the melt to produce strongly colored glasses,
some transparent, some opaque. This addition of colorants was
probably the first example of the use of minor ingredients to
change glass properties to produce a desired effect.
The earliest and most widely used solution colorants were
salts of copper (blue-green from the presence of Cu
2+
), iron
(blue to green from Fe
2+
, yellow to brown from Fe
3+
), and
manganese (amethyst or purple from Mn
3+
).
27
The use of small
quantities of manganese as a decolorizer to compensate for iron
colors, also known in the Middle Ages, was referred to by
Agricola and Neri and was used by the Venetians in the pro-
duction of cristallo. Cobalt was first used in the 14th century
BC (deep blue from Co
2+
). The use of chromium as a solution
colorant probably began early in the 19th century.
27
Copper and gold ruby glasses were prized highly for their
beauty and for their scarcity, the latter a result of the difficulty
of producing these colloidal colors. Both glasses presented se-
rious challenges to the glassmaker because of their sensitivity
to composition, melting conditions, and subsequent thermal
history. The ruby color is caused by the selective absorption of
light by small gold or copper crystals (∼50 nm in diameter) that
are formed by the precipitation of the metals in their atomic
state. The formation of the metal crystals is enhanced by re-
heating the glass (‘‘striking’’) and by the presence of reducing
agents, e.g., stannous chloride. The red glasses found in old
church windows are most likely copper reds, either copper
rubies, suspensions of cuprous oxide, or copper stains, because
gold rubies do not seem to have been made with any certainty
until the 17th century.
25,28
Opaque glasses colored by suspensions of relatively large
crystals (with diameters in the micrometer range), where the
crystals behave essentially as color pigments, have been known
since antiquity. The pigments are generally insoluble or of
limited solubility in the matrix glass. Some of the opaque col-
ors formed in this way are white glasses containing suspensions
of tin oxide, arsenic pentoxide, or calcium antimonate, and
yellow glasses colored by lead antimonate. Opaque blue
glasses colored by copper calcium silicate or cobalt alumi-
nate, green glasses colored by chromic oxide, and brown or
red-browns from iron or iron-manganese oxide mixtures also
are used.
The most dramatic examples of colored glass are probably
the church windows of the Middle Ages, with the greatest
created during the 10th through 14th centuries. Most of these
windows also contain much stained glass, wherein a colorant is
diffused into the glass surface at temperatures well below that
of molten glass. Copper reds and silver yellows are perhaps the
best-known examples of surface stains.
(4) Lead Glasses
The first major departure from alkali–lime–silica glasses
came during the 17th century with the commercial introduction
of lead flint glasses. Lead had long been a minor constituent in
glazes, mosaics, and artificial gems. It was introduced as cal-
cined lead or lead oxide, primarily for its fluxing effect. Neri
discussed lead glasses at some length in L’Arte Vetraria and
emphasized that great care must be taken to thoroughly calcine
the lead to avoid the formation of molten lead because ‘‘the
least lead remaining breaks out the bottom of the pots and lets
all the metall run into the fire.’’
25
Shortly after the publication of Merrett’s English translation
of Neri’s work in 1662, George Ravenscroft, an English glass
merchant, turned glassmaker to develop a clear glass based on
English ingredients.
29,30
This latter requirement was motivated
by the difficulty English glassmakers were experiencing in
obtaining raw materials at acceptable cost, because a monopoly
controlled the import of plant ashes for soda.
31
The glass mer-
chants also were struggling with unresponsive foreign suppli-
ers, much breakage in transit, and oppressive tariffs.
32
After a
series of experiments, Ravenscroft introduced a clear potash
Table I. Glass Compositions
Glass
†
Oxide content (wt%)
SiO
2
B
2
O
3
Na
2
OK
2
O CaO MgO Al
2
O
3
Fe
2
O
3
(1) Egypt, 1500 BC
‡
67.8 16.08 2.08 3.8 2.89 3.22 0.92
(2) Palestine, 4th Century 70.5 15.7 0.8 8.7 0.6 2.7 0.4
(3) Sudan, 3rd century 64.2 15.9 2.65 10.2 2.73 2.06 2.3
(4) Italy, 9th–10th centuries 77.8 6.4 8.7 2.1 0.7 2.2 0.8
(5) Container glass, 1980 73.0 13.7 0.4 10.6 0.3 1.8
(6) 1:1:6 soda–lime–silica 75.3 13.0 11.7
(7) Faraday ‘‘heavy glass’’
‡
10.6 15.6
(8) ‘‘Jena Standard Glass’’
‡
67.2 2.0 14.0 7.0 2.5
(9) Schott thermometer glass 72.0 12.0 11.0 5.0
(10) Schott utensil glass 73.7 6.2 6.6 5.5 3.3
(11) Schott Welsbach chimney
‡
75.8 15.2 4.0
(12) Nonex
‡
73.0 16.5 4.25
(13) Pyrex 80.5 12.9 3.8 0.4 2.2
(14) E-glass, typical 54.0 10.0 17.5 4.5 14.0
†
(1) Morey,
5
Table I-1 (10); (2) Brill,
20
Jalame glass, (low potassium, high sodium); (3) Brill,
21a
Sedeinga tomb glass, (high potassium, high magnesium); (4) Brill,
21b
Frattesina
glass, (mixed alkali); (5) Ryder and Poole
43
; (6) by calculation; (7) Faraday;
37
(8) Hovestadt,
40
Jena glass 16
III
, 1884; (9) Hovestadt,
40
p. 246, Jena glass 59
III
, 1889, ‘‘ideal
thermometer glass’’; (10) Steiner,
42
p. 172, Jena glass 202
III
, 1893, recalculated from batch; (11) Steiner,
42
p. 172, Auer von Welsbach gas light chimney, Jena glass 276
III
, 1895,
recalculated from batch; (12) Corning code 7720; (13) Corning code 7740, Morey;
5
(14) Aubourg and Wolf,
46
typical composition, can vary, depending upon manufacturer and
materials.
‡
Glass contains other oxides: (1) 0.54% Mn
2
O
3
, 1.51% CuO, and 1.0% SO
3
; (7) 70% PbO; (8) 7.0% ZnO; (11) 4.0% Sb
2
O
3
and 0.9% As
2
O
3
; (12) 6.25% PbO.
798 Journal of the American Ceramic Society—Kurkjian and Prindle Vol. 81, No. 4
lead glass that was the ancestor of English lead crystal. The
Worshipful Company of Glass Sellers of London, a trade guild,
quickly recognized the potential of Ravenscroft’s invention and
negotiated to buy his entire production. The first glasses suf-
fered from poor chemical durability and crizzling, and it was a
few years before a truly moisture-resistant lead crystal was
produced. The glass was called ‘‘crystal,’’ and the fact that lead
was the key ingredient was kept secret by Ravenscroft and his
immediate successors. These glasses also were called flint
glasses, because they were based on high-purity silica from the
flint nodules found commonly in the Cretaceous chalk deposits
of southeast England, plus calcined lead oxide, niter (potassium
nitrate), and potash from wood ashes (good quality potash had
become more readily available in the latter part of the 17th
century). A substantial business grew in the manufacture of
lead crystal articles that took advantage of the higher refractive
index and the ease of cutting and polishing of the lead flint to
create sparkling goblets, bowls, and vases.
The 17th century also was a period of growing interest in
science, and glass improvements became driven by scientists
seeking better optical instruments, particularly telescopes. Ga-
lileo and Kepler made a number of discoveries in optics that
made possible considerable improvement in telescopes, using
the soda–lime–silica crown glasses of the time. (Crown glass
was the name given to window glass of the period that was
made by the crown process, wherein a large blown bubble of
glass was transferred to a pontil, opened, and spun into a cir-
cular disk by centrifugal force.) However, later optical physi-
cists and astronomers found themselves increasingly frustrated
by poor glass quality and by the difficulty imposed by chro-
matic aberration in obtaining a clear, sharp focus. After New-
ton explained the refraction of light by prisms, he examined
many glasses and studied their dispersion (the variation in re-
fractive index with wavelength). Because the glasses were
probably all reasonably similar in composition, given the lim-
ited variety of glasses available, he concluded, incorrectly, that
all glasses had the same dispersion, and, therefore, that chro-
matic aberration was an uncorrectable fault in lenses. Accord-
ingly, Newton then decided that it was useless to attempt to
build a better refracting telescope and switched his energies to
reflecting telescopes. Others did the same, and refracting tele-
scopes went into eclipse until well into the 18th century.
33
During the early 1730s, Chester Moor Hall, an English law-
yer with an amateur interest in telescopes, recognized that lead
flint glasses had higher dispersion than soda–lime crown
glasses. He reasoned that chromatic aberration could be cor-
rected by an objective lens with two elements: a convex crown
element and a concave flint element. (‘‘Crown’’ and ‘‘flint’’
became the terms used to describe, respectively, low refractive
index (low dispersion) and high refractive index (high disper-
sion) optical glasses, respectively.) This doublet worked, and
some telescopes were built using this first achromatic lens. The
invention was not patented or publicized, however, and was
rediscovered by John Dollond, who patented it in 1758. Dol-
lond and his son Peter were skillful, well-known opticians, and
they were quite successful in marketing the achromats.
34
These
doublets were largely successful in bringing the red light and
blue light to focus in a common image, although a secondary
spectrum remained. This improvement should have encouraged
investigations of the effects of composition on the optical
properties of glass, but progress was slow because of the prob-
lems of making homogeneous glass. Poor-quality optical glass
persisted until stirring of the melt was introduced by Pierre
Louis Guinand and his successors in the beginning of the 19th
century.
Joseph Fraunhofer entered optical physics from the practical
side, working for a time in an optical institute where Guinand
was employed both as a glassmaker and as an optician grinding
and polishing lenses. Fraunhofer made some excellent achro-
mats, which helped revive refracting telescopes. In the process,
he experimented with glass compositions and recognized that
more choices were needed in refractive index and dispersion
‘‘Deceptively like a Solid’’
Roald Hoffman
The conference is on Glass, in Montreal. Wintry light declines
to penetrate windows, and soon will be lit glass-enclosed glows
so that we may talk, talk into the night (fortified by bottled
mineral waters), of the metric of order trespassing on prevailing
chaos that gives this warder of our warmed up air, clinker,
its viscious, transparent strength.
The beginning was, is silica, this peon stuff
of the earth, in quartz, cristobalite, coesite,
stishovite. Pristine marching bands of atoms
(surpassing the names we give them)
build crystalline lattices from chains, rings, of Si
alternating with oxygen, each silicon tetrhedrally
coordinated by O’s, each oxygen
ion, so different from the life-giving, inflaming
diatomic gas, joining two silicons; on to rings
in diamondoid perfection in cristobalite;
helical O-Si-O chains in quartz, handed in
coiling, mirror images of each other, hard, ionic SiO
2
.
There must be reasons for such perfection—
time lent to the earth: then lava flowed, the air
blew thicker, still no compound or simple eye to
fret defect into the unliquid from which silica
crystallized. But in time we did come, handy,
set to garner sand, limestone, soda ash, to break
the still witness of silica. Heat disrupts. Not the
warmth of Alabama midsummer evenings, not
your hand but formless wonder of prolonged fire,
the blast of air drawn in, controlled fire storms.
Sand, which is silica, melts. To a liquid, where
order is local but not long-range. Atoms wander
from their places, bonds break, tetrahedra in a
tizzy, juxtapose, chains tilt, bump and stretch—
Jaggerwalky.
The restive structures in microscopic turmoil
meld to gross flow, bubbling eddies of the melt.
Peace in crystal meshes, peace
in hot yellow flux. But the gloved
men who hold the ladies get nervy volcanoes
on their minds. So—tilt, pour . .. douse,
so quench, freeze in that micro lurch.
Glass forms, and who would have thought it clear?
We posit that the chanced, in its innards so upset,
ought not be transparent. Light scattered from
entangled polymer blocks, adventitious dirt,
owes it to us— oh, we see it so clearly—to
lose its way, come awash in black or at least
in the muddy browns of spring run-off, another flux.
But light’s submicroscopic tap dance is done in place.
The crossed fields shimmer, resonant, they plink
electron orbits of O and Si.
Atoms matter, their neighbors less, the tangle of the locked-in
liquid irrelevant in the birthing of color, or lack of it.
Optical fibers Crystal Palace
The Worshipful Company of Glass Sellers
recycled Millefiori
prone to shattering Prince Rupert’s drops
Chartres, Rouen, Amiens float
Pyrex Vycor glass wool
network modifiers the Palomar mirror
smoked for viewing eclipses thermos
lead glass microcrack
etched with hydrofluoric acid spun
frustration bull’s eyes annealed
borosilicate softening point
High winds on Etna or Kilauea
spin off the surface of a lava lake thin fibers.
Pele’s hair. The Goddesses’ hair, here black.
From The Metamict State; pp. 44–48.
University of Central Florida Press, Orlando, FL, 1987.
April 1998 Perspectives on the History of Glass Composition 799
beyond those of ordinary crown and flint glasses if better lenses
were to be made that corrected the secondary spectrum. Fraun-
hofer also wrote about the chemical durability of glasses and
was probably the discoverer of the mixed-alkali effect, noting
that glasses with mixed alkalis had superior durability. (The
mixed-alkali effect is the distinctly nonlinear behavior ob-
served in some properties when one alkali ion is gradually
substituted for another alkali ion. This phenomenon is observed
in properties affected by transport mechanisms, such as elec-
trical conductivity, dielectric loss, internal friction, and self-
diffusion.) Later, Fraunhofer’s spectral studies enabled him to
make observations on dispersion for the principal glass com-
ponents of the day.
17,22,25,35,36
Impressed by Fraunhofer’s results, the Royal Society estab-
lished, in 1824, a commission consisting of Michael Faraday,
John Herschel (the astronomer), and George Dollond (another
of the famed clan of opticians), to study the possibility of
making superior glasses for telescope objectives. Faraday be-
came interested in glassmelting and made some prolonged in-
vestigations during 1825–1830 that demonstrated the benefits
of melting glasses in platinum containers and the importance of
stirring melts to improve homogeneity. His experiments, un-
fortunately, did not contribute much to the knowledge of glass
composition, although he did demonstrate that boron could be
used in glassmaking to make a passable lead borosilicate flint
glass. Faraday later (1845) conducted some experiments of
significance with his ‘‘heavy glass’’ (see glass 7 in Table I), in
which he demonstrated the Faraday effect (rotation of the plane
of polarization of light in a magnetic field).
15,33,35,37,38
It may seem surprising today that eminent scientists and
intellectuals of the time were deeply interested in finding so-
lutions to glass composition problems. For example, Vogel
states, ‘‘Goethe, then Prime Minister of a German duchy . . . in
1829 wrote to his friend, the noted chemist Do¨breiner at the
University of Jena, ‘it would be most important to determine
the relation of refraction and dispersion in your [barium and
strontium] glasses . . . I should be pleased to contribute the
modest funding . . .’.’’
36,‡
The first, however, to make an extensive study of the effects
of a wide range of elements on the properties of glass was the
Rev. William Vernon Harcourt, an English clergyman. The late
18th century and the early 19th century was a period of highly
significant advances in the discovery and isolation of new el-
ements, and, in the 1830s, Harcourt began investigations of the
effects of many of these new elements on the optical properties
of glass. Among the elements he first used in glass were be-
ryllium, cadmium, fluorine, lithium, magnesium, molybdenum,
nickel, tungsten, uranium, and vanadium. He also studied the
effects of other elements, including antimony, arsenic, barium,
boron, phosphorus, tin, and zinc, first introduced into glass by
others. Harcourt did not confine his studies to silicate glasses,
but also melted some phosphates, borates, and titanates, in part
because he found it difficult to fuse the silicates to a homoge-
neous glass. He did not widely publicize his findings, but Sir
George Stokes, the noted mathematician and physicist, learned
of his work, collaborated with him, and helped to bring the
results to the attention of the scientific community in 1871, the
year of Harcourt’s death. In 1874, Stokes made a small, three-
component lens that was largely free of the secondary spectrum
from some of Harcourt’s glasses. Therefore, even though Har-
court’s glasses were not completely homogeneous, his work
demonstrated that different glassmaking ingredients did bring
changes in dispersion and refractive index that could yield
glasses that began to solve the optical problems of the
time.
16,31,39
Although the work of Harcourt should have encouraged the
British glass industry to investigate further the effects of dif-
ferent glass constituents, it did little beyond the production of
some standard optical crowns and flints by Chance Brothers in
Birmingham. Experimentation to develop new glass composi-
tions was severely constrained at that time in England by ex-
orbitant taxes on all glassmelting. Therefore, Chance Brothers
concentrated instead on improving the quality of the standard
glasses by stirring the melt. Accordingly, the initiative in glass
composition research passed to German glassmakers, who built
on the work of Fraunhofer and Harcourt.
31
As is evident from the foregoing, until the late 19th century,
the development of new glasses was largely a matter of an
occasional fortuitous discovery. These early investigations, al-
though often motivated by a need, were not pursued system-
atically, had difficulty in yielding a homogeneous product, and,
with the exception of Harcourt’s work, usually used the same
few ingredients. Hovestadt wrote, in 1900, ‘‘. . . the develop-
ment of the art of glassmaking in response to optical require-
ments kept, for a long time, to one narrow groove, and no new
fluxes broke the monotony of a uniform series of crowns and
flints.’’
40
III. Abbe and Schott
Ernst Abbe, professor of physics at Jena University, became
interested in optical glasses through his work with Carl Zeiss,
a microscope maker at the university. Similar to the telescope
makers, Abbe soon realized that a wider variation in dispersion
for a given refractive index was needed to remove completely
the secondary spectrum from optical images. He wrote on the
subject in the late 1870s, and his remarks attracted the interest
of Otto Schott, a young German chemist who had been explor-
ing glassmelting phenomena in connection with his family’s
glassworks in Westphalia. Schott contacted Abbe and sent him
some lithium glasses he had prepared with the thought the
samples might aid Abbe in his research for glasses with dif-
ferent optical properties. By 1881 Abbe and Schott were col-
laborating and thus was born one of the greatest and most
productive associations in the history of glass composition.
31,40
Schott moved to Jena in 1882 to be closer to Abbe and Zeiss.
Abbe (the scientist), Schott (the glassmaker), and Zeiss (the
instrument builder) worked together in a synergistic manner
that bore dramatic results. Abbe and Schott would discuss the
composition changes to be made, Schott would then prepare
homogeneous glass melts, and Abbe would measure the results.
If the properties appeared to be an improvement, Zeiss would
grind and polish lenses and observe the performance of the
‡
The poem on the previous page was authored by Roald Hoffman, the Nobel
Laureate in chemistry in 1981 with Kenichi Fukui for ‘‘His application of molecular
orbital theory to chemical reactions.’’ Also, the 1977 Nobel Prize in physics was
awarded to P. W. Anderson, Sir N.F. Mott, and J. H. van Vleck for ‘‘Their funda-
mental theoretical investigations of the electronic structure of magnetic and disor-
dered systems.’’ The eminent scientists continue to find glassy systems of interest.
‘‘Glass . . . is much
more gentile, graceful,
and noble than any
Metall,...itismore
delightful, polite, and
sightly than any other
material at this day
known to the world,’’
Antonio Neri, 1612
800 Journal of the American Ceramic Society—Kurkjian and Prindle Vol. 81, No. 4
finished pieces, then feed his observations back to Abbe and
Schott. In this manner they started with silica, soda, potash,
lime, and lead oxide and eventually added 28 other elements in
quantities of at least 10% to produce glasses of refractive in-
dexes and dispersions substantially different from those made
previously.
The techniques used by Abbe and Schott in their studies
were based on careful observation and measurement, although
almost totally empirical, because no reasonable theories existed
to guide their work. Additions of new minor ingredients were
made to correct or offset faults in the original compositions.
For example, Schott found that, in borate and phosphate
glasses, alkalis had to be used very sparingly, if at all; other-
wise surface staining resulted on exposure to air. However,
when alumina, zinc oxide, and barium oxide were added, the
surface durability could be improved enough to make the
glasses serviceable. Schott soon learned that the addition of
some elements would have no effect on optical properties but
would have a favorable effect on other properties.
In an effort to at least make their results usable, Schott, and
Winkelmann developed what probably was the first composi-
tion–property model.
4
They produced a series of oxide factors
that allowed them to calculate the value of a property knowing
the composition. Today, many such models are available be-
cause of computers (see Cable
41
).
Early useful results were obtained with boron, barium, and
fluorine, leading to families of borosilicate crowns, barium
flints, and fluor crowns. (The demarcation between crowns and
flints is arbitrarily defined by their dispersion and is shown
in Fig. 2.) The government was quite impressed by the prog-
ress and made some large grants to support the work of the
laboratory that became, in 1884, the Jena firm of Schott und
Genossen.
The discoveries of Abbe and Schott were not confined to
optical glasses. In the 1890s, the group at Jena analyzed the
problem of the secular rise of the zero in the aging of glass
thermometers. It was noted in one of the early observations of
the mixed alkali effect that the zero rise was particularly pro-
nounced (more than one celsius degree) in glasses with ap-
proximately equal quantities of soda and potash. Glasses made
with either only soda or only potash as the alkali suffered only
one tenth or less the secular rise as the mixed alkali glasses.
The most stable glass was found to be a borosilicate (see glass
9 in Table I).
40
Improved laboratory glassware also resulted from Schott’s
further pursuit of boron in glass with the discovery that boro-
silicate glasses had exceptional resistance to attack by boiling
water. Accordingly, these glasses also made excellent boiler
gauge glasses. It also was noted that boric oxide was the most
effective addition to silicate glasses in reducing the coefficient
of thermal expansion, and this discovery led to laboratory
glassware with improved resistance to thermal shock.
40
In the remarkably short period from 1879 to 1886, Otto
Schott, with the assistance of Abbe and Zeiss, created and
offered commercially a surprising array of optical glasses. Be-
sides using a systematic approach to glass composition re-
search, Schott had mastered the small-scale melt-stirring pro-
cess so as to be able to make a homogeneous product. The
glasses also had been carefully characterized, so they were sold
with exact measured values of refractive index and dispersion.
This work was a watershed in the history of glass composition
in that it demonstrated for the first time the ability to tailor the
properties of a glass by judicious adjustments in composition
based on a composition–property model.
42
IV. Modern Glasses
(1) Soda–Lime–Silica Glasses
Although sand and alkali were known from the earliest days
of glass to be necessary ingredients, the role of lime was not
apparent until much later times. Lime was not recognized as an
important glass constituent by early glassmakers, because ad-
equate amounts of lime were generally added unknowingly as
an impurity in the sand and alkali. Lime appears to have been
added consciously to glass batches in Roman times, but Neri
mentioned lime only casually in suggesting that small quanti-
ties could be added ‘‘. . . to make a very fair and beautiful
Crystall.’’
12
Only in the 17th, 18th, and 19th centuries did the
increase in chemical durability brought about by the addition of
lime to alkali silicate glasses become understood. Bohemian
glassmakers added lime to their fine crystal in the 17th century,
and, during the late 1700s, P. D. Deslandes added up to 6%
lime to increase the resistance of Saint-Gobain’s plate glass to
attack by moisture.
31
Guinand and Fraunhofer observed that it
was necessary to add lime to increase glass durability, and, in
1830, J. B. Dumas, a French glass technologist, noted that the
chemical durability of glass was improved by adding one part
of lime to one part of soda and six parts of silica. The addition
of lime to the batch became essential in practical glassmaking
Schott und Genossen
As stated by Douglas and Frank,
31
‘‘The success of the
firm [Schott und Genossen] was spectacular. Its first price
list of 1886 contained forty-four optical glasses of which
nineteen were essentially new compositions. The first
supplement of 1888 added twenty-four glasses, including
eight new barium light flints which were remarkable for
their small dispersion compared with refractive index. They
contained so little lead oxide that the usual light absorption
shown by flint glasses was greatly reduced. New glasses
were added to the list every few years, and the effect on the
manufacture of optical systems was so great that Germany,
which had previously imported ninety percent of its optical
systems from England and France, started to export to these
countries. Thus, an industrial development which was ac-
complished in less than ten years virtually eliminated exist-
ing manufacturers and, for about 30 years, until the outbreak
of the World War I, Jena held an effective world monopoly
in the manufacture of optical glass.’’
Fig. 2. Refractive index, n, versus reciprocal dispersion, , showing
historical development of optical glasses.
34
(White area within curve
represents modern glasses (Morey et al.); hatched area represents ear-
lier glasses, i.e., 1880–1934 (Schott et al.); and black area represents
glasses before 1880.)
April 1998 Perspectives on the History of Glass Composition 801
to preserve durability as the use of synthetic soda ash (pure
Na
2
CO
3
, with no CaCO
3
) became wide-spread. This practice
began early in the 19th century, because soda ash from the
Leblanc process became available and its use was common
practice after the 1860s when the Solvay process became the
principal source of soda.
Most of the current commercial glasses are soda–lime–silica
glasses. It is significant that the compositions of these glasses,
used typically for containers and flat glass, has changed little
over the centuries, ranging from 65% to 75% silica, with alkali
ranging from 10% to 20%, and lime as the balance. Although
most older soda–lime–silica glasses contained a few percent
alumina from raw-material impurities and from refractories, a
similar amount has been added customarily since Schott, based
on his observations of Thuringian glasses, demonstrated, in the
late 1880s, that it benefited durability and resistance to devit-
rification. Glasses with these compositions are relatively easy
to melt and form, do not devitrify easily, and generally have
reasonable resistance to attack by moisture. They can be made
quite color-free and nontoxic with pure raw materials that are
available worldwide at acceptable cost.
5,40
Improvements in melting technology—e.g., more-resistant
refractories and higher temperatures—have increased chemical
durability through lower alkali and higher lime contents. This
trend also has been encouraged by the comparative costs of
soda and lime, and, currently, economic factors are the princi-
pal determinant for soda–lime–silica compositions. Enough is
now known about the effects of composition on properties to
permit major glass constituents to be adjusted several percent-
age points to compensate for differences in raw-material prices
to reach the lowest-cost composition.
43
(2) Borosilicate Glasses
The first new major glass system to be explored beyond the
soda–lime–silica glasses utilized the other great glass-former,
boric oxide, important for its many commercial applications.
Although borax was known and used in the Middle Ages as an
exotic flux, its use in practical glassmaking became a realistic
possibility with the discovery of extensive deposits in Turkey,
Chile, and, in particular, California, in the latter part of the 19th
century.
42
As noted in an earlier section, Abbe and Schott, during the
1880s, were the first to use boron compounds in glass in sig-
nificant amounts, first in optical glasses, then in glasses for
laboratory apparatus. Both Faraday and Harcourt had made
some use of boron in glass, but Abbe and Schott clearly estab-
lished that borosilicate glasses had superior resistance than
soda–lime–silica glasses to chemical attack and had better ther-
mal shock resistance because of their lower coefficient of ther-
mal expansion. The introduction of the Auer von Welsbach
mantle in gas lamps in 1887 created a need for a lamp cylinder
or chimney with improved resistance to thermal shock. Schott
met this need with a glass containing 15% boric oxide (see
glass 11 in Table I) having a very low coefficient of expan-
sion.
36,40,42
At about this time in the United States, cracking and break-
age of the globes of railroad brakemen’s lanterns was a prob-
lem when rain showers struck the hot glass. Why Schott heat-
resistant glasses were not used to solve this problem is not
known; perhaps it was due to the poor international technical
communications of the times. Corning Glass Works, the lead-
ing U.S. manufacturer of lamp chimneys and bulbs for electric
lamps, was asked to investigate the matter, and it became the
first assignment of the newly formed research laboratory in
1908. In 1909, Corning introduced a borosilicate glass that
solved the lantern globe thermal shock problem, but had poor
chemical durability. Corning’s first research director, Eugene
C. Sullivan, a chemist hired from the U.S. Geological Survey,
and William C. Taylor, a young chemist colleague recently
from Massachusetts Institute of Technology, worked further on
the problem.
44
By 1912, they had perfected a chemically du-
rable, shock-resistant, lead borosilicate glass marketed under
the name Nonex威(for nonexpanding) that reduced lantern
globe breakage by 60% (see glass 12 in Table I). Nonex also
proved to double the life of the battery jars used by the rail-
roads in their new electrically powered signal systems.
Fig. 3. Soda–lime–silica phase diagram (R. S. Roth, T. Negas, and L. P. Cook; Fig. 5321 in Phase Diagrams for Ceramists, Vol. IV. Edited by
G. Smith. American Ceramic Society, Columbus, OH, 1981).
802 Journal of the American Ceramic Society—Kurkjian and Prindle Vol. 81, No. 4
Corning continued to explore borosilicate glass composi-
tions and their applications. In 1913, Corning physicist Jesse T.
Littleton suggested that Nonex glass vessels might be used for
baking pans, and a cake baked by his wife in the bottom of a
battery jar demonstrated that the idea was sound. However,
Nonex contained too much lead for food preparation. There-
fore, a lead-free borosilicate composition was developed by
Sullivan and Taylor which was named Pyrex威(see glass 13 in
Table I). A pressed Pyrex ovenware line was introduced in
1915, and it became an immediate sales success. At about this
time, the supply of glassware to U.S. laboratories from Schott
and other European suppliers was interrupted by World War I,
and it soon became apparent that blown Pyrex borosilicate
glass was an extremely good glass for laboratory apparatus
because of its low expansion and its chemical durability. Ac-
cording to Morey,
5
‘‘. . . [this] glass has the lowest liquidus
temperature of any known mixture having so high a silica
content and in this fact doubtless is to be found a clue to its
exceptional ability to withstand devitrification.’’
Other variations of borosilicate glass were explored by Corn-
ing for special applications that would benefit from low ex-
pansion and corrosion-resistant glasses. One noteworthy ex-
ample was the development of a glass with an expansion 25%
lower than that of Pyrex ovenware that was used for casting the
200 in. mirror for Mt. Palomar’s Hale telescope in 1934. The
astronomers had sought the lowest possible expansion in their
quest for a mirror that would show the least distortion with
temperature changes. This glass also had the low liquidus tem-
perature necessary to survive the extremely slow cooling rate
of the cast blank without devitrification or phase separa-
tion.
22,45
Glass chemists Harrison P. Hood and Martin E. Nordberg
observed during experimentation with borosilicates at Corning
in the 1930s that very large changes in properties would occa-
sionally result when some compositions were heat-treated. Fur-
ther investigation revealed that these glasses were separating
into two intermingled glassy phases, one of which was silica-
rich. A composition was developed that separated upon heat
treatment into a very-high-silica phase and a very-alkali-rich
phase, with the latter being easily dissolved and leached out by
hot nitric acid. The remaining porous, silica-rich skeleton could
then be consolidated into a solid, pore-free form by heating.
This glass, given the name Vycor威, in 1939, could be melted
and formed easily before the leaching and consolidation steps.
The 96% silica composition of the finished glass meant that it
had physical and chemical properties closely approaching those
of pure-silica glass, yet it could be fashioned before heat treat-
ment into shapes that were impossible to form with fused silica
because of the extremely high viscosity of silica glass at even
very high temperatures.
45
This is an example of a new glass
composition being created by carefully exploiting the observa-
tion of a new phenomenon.
Another important step in borosilicate glass composition de-
velopment also occurred in the 1930s with the introduction of
E-glass fibers. A fiberglass material was needed for electrical
insulation applications, and standard soda–lime–silica compo-
sitions were not suitable, because their conductivity was too
high. Soda–lime glasses also proved generally unsuitable for
fiberglass, because the great surface area to volume ratio made
them particularly vulnerable to attack and dissolution by water.
Accordingly, Urban Bowes, R. A. Schoenlaub, and others at
Owens–Corning Fiberglas reduced and ultimately removed the
alkalis and added boric oxide and alumina and increased the
lime.
46
The boric oxide and lime reduced viscosity and in-
creased durability, and the alumina also helped durability and
lowered the liquidus. The resulting lime–alumina–borosilicate
glass (see glass 14 in Table I) had excellent electrical resistiv-
ity, superior resistance to attack by moisture, and good me-
chanical properties. Therefore, it is not surprising that glasses
of this general composition make up 90% of the continuous
glass fiber currently produced.
46
(Historical note: In this same
publication, Dumbaugh and Danielson
47
reference a 1925 pat-
ent issued to F. M. and F. J. Locke
48
for a series of alumino-
silicates that is broad enough to include these alkaline-earth
aluminoborosilicates.) Also, most commercial fiberglass com-
positions are based on ternary or quaternary eutectics, thus
taking advantage of the well-known fact that the most stable
glasses are often close to eutectic compositions.
(3) Photosensitive Glasses and Glass-Ceramics
No account of the history of glass compositions would be
complete without some reference to the discovery of glass-
ceramics. Accordingly, a brief description of the early work on
glass-ceramics is presented here.
The natural tendency of glasses to devitrify must have led
glassmakers occasionally over the ages to consider pursuing
the phenomenon to its logical conclusion—a completely crys-
talline product. The French chemist Re´aumur is known to have
attempted to produce crystalline vessels by holding glass
bottles packed in gypsum at a red heat for several days. Al-
though these did devitrify to a crystalline form, he was unable
to control the process, and the bottles were deformed and of
low strength.
49
Others experimented with the limited precipi-
tation of crystals from glass to create ruby or opal glasses, but
the development of a process to control the massive crystalli-
zation of bulk glass did not occur until the middle of the 20th
century.
There was a series of glass composition inventions during
the 1940s and 1950s based upon nucleation and controlled
growth of crystals experiments conducted at Corning Glass
Works. The work was led by S. Donald Stookey, but ultimately
involved several others, including Armistead, Beall, Mac-
Dowell, Araujo, Rittler, and Grossman.
Stookey had been investigating nucleation and precipitation
of crystals in ruby and opal glasses and found, in 1942, that
copper, gold, and silver could be deposited as tiny particles of
metal through photo-induced heterogeneous nucleation. The
process was aided by the presence of a ‘‘sensitizer,’’ such as
cerium or tin. Taking this a step further in 1951, he learned he
could photo-induce a sodium fluoride opal in a silicate glass
nucleated with silver (Fotalite威).
50
Stookey made an unanticipated discovery, in 1954, of the
controlled crystallization of glass ceramics. He had invented, in
the late 1940s, a process for preparing lithium silicate crystal
images in glass with a photosensitive technique (Fotoform威).
Very small quantities of silver (∼0.06 mol%) were introduced
in the lithium silicate glass as a nucleating agent, and, after
selective exposure to ultraviolet radiation, a heat treatment at
∼600°C caused lithium metasilicate crystals to form that then
could be leached out of the unexposed glass.
One day, Stookey placed a plate of preexposed lithium sili-
cate glass in a laboratory oven to perform the 600°C heat
treatment. The temperature controller stuck in the ‘‘on’’ posi-
tion and the glass was heated to 900°C, where it normally
would be quite soft and fluid. Stookey was alarmed by the
overheating, and he was certain that he had ruined the oven. He
knew that this glass melted and flowed below 700°C, and he
believed that it would flow on to the floor of the oven. How-
ever, in his own words, ‘‘Imagine my astonishment on opening
the door to see an undeformed, opaque solid plate! Snatching a
pair of tongs, I immediately pulled the plate out of the hot
furnace, but it slipped from the tongs and fell to the tile-
covered concrete floor, clanging like a piece of steel but re-
maining unbroken! It took no great imagination to realize that
this piece of Fotoform was not glass, but something new and
different. It must have crystallized so completely that it could
not flow, even though the temperature was more than 200°C
above the softening temperature of the glass. And obviously it
was much stronger than ordinary glass.’’
50
An examination of
the fine-grained glass-ceramic that had been formed revealed
that it was composed of lithium disilicate and quartz crystals,
and was much harder and higher in electrical resistivity than
April 1998 Perspectives on the History of Glass Composition 803
regular glass. This episode demonstrates the power of the pre-
pared mind in recognizing a seminal event: Stookey quickly
realized that, theoretically, all glasses can be converted to crys-
talline bodies having new properties that depend on the nature
of the particular crystals formed.
51
A vigorous research and
development program was then followed by Stookey, Beall,
and others at Corning that continues to this day and has pro-
duced many very useful glass-ceramics.
Stookey later (1959) made another discovery based on pho-
tosensitive precipitation work with the invention of photochro-
mic glasses, i.e., glasses that darken when exposed to sunlight
and regain their clarity when the ultraviolet radiation is re-
moved. Stookey, acting on a suggestion from Armistead, in-
troduced silver halides (chlorides and bromides) in small quan-
tities (∼0.5 wt%) to glasses that had been doped with a copper
sensitizer. The glasses were heat-treated at 600°C to precipitate
supersaturated microcrystals of silver halide. When photons
strike the microcrystals, some of the silver is reduced to the
metallic Ag
0
state, with the electron being borrowed from a
chloride ion. The metallic silver particles color the glass gray,
or darken it. When the ultraviolet source is removed, the me-
tallic silver is oxidized back to Ag
+
, and the glass clears.
52
V. Structure and Properties
As stated earlier, the early soda–lime–silica glass composi-
tions were very close to the standard soda–lime–silica glasses
that are currently in use. The fact that they were first discov-
ered by accident, and yet are fortuitously close to the ideal
commercial composition, remains somewhat of a surprise.
However, the raw materials were relatively available, and com-
positions that were very different would probably either crys-
tallize, dissolve, or be unmeltable, and, therefore, in a sense,
this may be considered a Darwinian result. Shortly after the
start of the glass science era, i.e., the early 1920s, after Schott’s
work, the opening of the Corning Research Laboratory in 1908,
and the formation of the Department of Glass Technology (ini-
tially named Department of Glass Manufacture) in 1915 under
the direction of Professor W. E. S. Turner in Sheffield, En-
gland, the new tool of X-ray diffractometry was first applied to
the study of silicate crystals, and then to silicate glasses.
53
This
led to an overall attack by the scientific community on the
problem of the understanding of this unusual material.
(1) History of Glass Structure Studies
Although it was very early realized by Tammann
54
that glass
formation was a kinetic phenomenon, it also was clear that the
kinetic processes involved were controlled by the details of the
structure of the materials involved. Although, in 1926, Gold-
schmidt
55
indicated that SiO
2
, GeO
2
,B
2
O
3
,P
2
O
5
,As
2
O
3
,
As
2
S
3
,Sb
2
O
3
, as well as the (silica) model, BeF
2
, were all able
to form glasses by themselves (i.e., they were glass-formers),
the history of the discovery of the existence of these simple
inorganic glass-formers was not clear. Even in the case of
silica, the first recorded instance of the recognition of its ability
to form a glass on its own is not clear. Sosman
56
points out that,
in 1813, Marcel formed glassy silica by heating small quartz
crystals in an oxygen-injected alcohol lamp. Again, a check of
the literature shows that, according to Rawson,
57
in 1834, Ber-
zelius melted glasses in several tellurite systems, while, in
1868, Roscoe investigated a series of BaO–V
2
O
5
glasses. As
indicated earlier, in the mid-1800s, Harcourt
39
melted mainly
phosphate glasses, because he found that silicate glasses were
‘‘pasty.’
The first detailed descriptions of the expected ‘‘crystal’’
structures and the reasons that such structures formed glasses
were proposed by Ha¨gg,
58
Goldschmidt,
55
and Zacharia-
sen.
59,60
They were crystallographers/chemists who, since the
discovery of X-rays had studied crystal structures with this tool
and, therefore, quite understandably approached the problem of
glass formation and structure from the results of those studies.
Arguably, Zachariasen’s rules constitute the most famous work
in glass science to date. These rules describe what he consid-
ered to be the necessary conditions for glass formation:
●An oxygen is linked to not more than two central atoms;
●The number of oxygen atoms surrounding a central atom
must be small;
●Oxygen polyhedra share corners—not edges or faces—with
each other;
●At least three corners in each oxygen polyhedra must be
shared.
Over the years, the validity of these rules has been debated.
Recently, many workers have argued for,
61
against,
62
and
about these rules and the ‘‘continuous random network’’ to
which they imply. The rules have been discussed and updated
by Cooper.
61,63
They also have been modernized in the sense
that the topological basis of these rules has been extended
(Gupta and Cooper
64
and others
65
).
That such structures (as proposed by Zachariasen) were in-
deed found in simple glasses was first demonstrated by Warren
and co-workers (Fig. 4(a))
66,67
using X-ray diffractometry. The
Fig. 4. (a) Two-dimensional representation of a disordered sodium
silicate network.
66
(b) Continuous random network model of vitreous
silica.
74
804 Journal of the American Ceramic Society—Kurkjian and Prindle Vol. 81, No. 4
details of these experiments and the controversy over the de-
tails of the continuous random network (CRN) model proposed
by Warren continues. Indeed, in 1921, prior to the work of
Ha¨gg, Goldschmidt, and Zachariasen, Lebedev
68
had proposed
a microcrystallite model that was apparently completely at
odds with the CRN model. The initial, basic controversy had to
do with the amount of medium- and long-range order that
exists in normal glassy (e.g., silicate) solids. The controversy
appeared to have started as a result of the X-ray studies of
Warren (1932) in the United States on the one hand and Leb-
edev and his co-workers in Russia (1922) on the other hand.
However, Nemilov
69
has recently suggested that Frankenheim
was perhaps the first to propose a model for the structure of
glasses early in the 19th century. Frankenheim
70
states in his
1835 book, The Teaching of Cohesion, ‘‘It seems possible that
glasses, which are seemingly quite noncrystalline, are, never-
theless, aggregates of finely disperse crystals, which are partly
sintered with each other, and partly combined like a mortar.’’
Since that time, the popularity of a given model for glass struc-
ture has been found to be cyclic. Indeed, at almost the same
time as Frankenheim’s book appeared, Faraday characterized
glass ‘‘. . . rather as [a] solution of different substances one in
another rather than as a strong chemical compound.’’
37
Therefore, we see that various models for the structure of
glass have been proposed over the years, and, as illustrated by
one of the present authors,
72
these proposals seem to show a
cyclic character. One of the early proponents of the microcrys-
tallite model (Porai-Koshits
73
) commented, ‘‘. .. investigators
who, like Bell and Dean
74
(Fig. 4(b)), construct and utilize the
random network models, come to the same conclusion about
the validity of Zachariasen’s hypothesis. But those investiga-
tors who use the quasi-crystalline models, find crystallinity or
crystal-like regions in glasses.’’
Over the years there have been discussions between different
investigators, often because of the advent of a new or improved
experimental technique, of the reinforced validity of one or
another of the old models, or indeed, new evidence that sup-
ports still another new model. In particular, with some new
experimental tools (especially magic angle sample spinning
(MASS) nuclear magnetic resonance (NMR) and other NMR
techniques
75
) that allow, for instance, the determination of the
numbers of n-coordinated silicons (Q
n
) and computational
techniques, additional structural information, even in terms
of the illusive medium range order, are now becoming more
available.
In this regard, Gaskell
76
recently has reviewed structural
analysis techniques and results. He concludes ‘‘. . . all our ac-
cessible information may be inadequate to solve the structure in
any real sense since the set of possible structures that ad-
equately fits all experimental data within statistical errors may
still be so broad that subjective arguments alone distinguish the
final choice.’’ Wright
77
has stated ‘‘. . . the modern crystallite
and random network theories both involve continuous net-
works and differ mainly in respect of the magnitude of the
fluctuations in the degree of intermediate-range order which
must occur in any such network and the spatial distribution of
defects. Certainly the appearance of new experimental and
theoretical techniques holds the promise of eventual break-
throughs in the measurement of the position of each of the
atoms in a glass structure. Even so, the question of how such
information is to be conveyed may remain. Perhaps we may
have to be satisfied with a collection of distribution functions.
78
In addition, while the details are of scientific, and also probably
of practical interest and importance, the overall view is prob-
ably clear enough to be stated without too much fear of dis-
agreement.’’ Indeed, many different ways of looking at, ana-
lyzing, and describing these structures have been suggested
over the years, with the general picture remaining about the
same. As indicated above, at the level of our discussion here, it
is not possible, nor is it necessary, to resolve this issue.
Although a more complete understanding of the fundamental
structure of any glass is desirable both scientifically and tech-
nologically, as Cooper
79
pointed out ‘‘. . . a subjective view of
the history of glass composition development has led to the
conclusion that, up to the present, the influence of structure on
composition development has been meager. Yet, we may ex-
pect that, in the future, a greater benefit will accrue from a
structural approach.’’ As indicated earlier, Cable
41
has re-
viewed the efforts that have been made to model the compo-
sition–property behavior of glasses since the first attempts of
Winkelmann and Schott. In this regard, the application of com-
puters and statistical techniques may prove more useful in the
future.
78
(2) Anomalous and Normal Glasses—Silica versus
Silicates
Both historically and commercially, silica-based glasses
make up almost all of the glasses in current use. The present
value of glass products produced today is 42.5% container,
35% flat, 11% fiber, and 11.5% special (i.e., 43% TV, 29%
other, 11.5% chemical/pharmaceutical, and 7% optical).
80
Therefore, most of our brief discussion of structure relates to
silicate glasses.
The early Egyptians discovered that fluxes were required for
(silicate) glasses to melt readily. They were necessary because
the production of silica by ordinary melting techniques requires
temperatures >2000°C. The naturally occurring additions that
were made—sodium, potassium, calcium, etc.—bonded ioni-
cally to the oxygen ions by breaking Si—O—Si bonds in some
way, thus breaking down the continuously corner-connected
three-dimensional silica network. This was very useful, be-
cause it resulted in lower melting and processing temperatures
(Fig. 5). On the other hand, if too much alkali and/or not
‘‘To understand
science, it is necessary
to know its history,’’
Auguste Compte, 1831
Fig. 5. Effect of modifiers on the viscosity of silica and germania
8
((---) K
2
O, (—) Na
2
O, and (⭈⭈⭈)Li
2
O).
April 1998 Perspectives on the History of Glass Composition 805
enough alkaline-earth were inadvertently used, the product
would not be satisfactory.
Although there currently are disagreements about the details
of the randomness of alkali and alkaline-earth silicate glasses,
especially at intermediate ranges (∼20–200 nm), there is no
disagreement over the fact that these additions to silica result in
a glass with strikingly different behavior. Clearly, the critical
temperatures (liquidus, processing, glass, etc.) are all quantita-
tively different, but, in addition, most of the properties are
qualitatively different as well. In some respects, it is not unlike
any pure material with no or few defects: changes in the fre-
quency of these defects or dopants may have a pronounced
effect. As indicated above, silica is completely coordinated in
three dimensions. Although attempts have been made to model
the flow in such a structure, no satisfactory detailed model
exists. Although there is three-dimensional coordination with
the addition of network modifiers (sodium and calcium), the
uniformity and continuity are not retained. The ionic bonding
to the modifiers provides weak points for flow, in the liquid
and in the solid.
As discussed above, in pure silica, such weak points are not
available. Thus, qualitatively, the flow processes are different.
This means that comparing either the viscous behavior at
1400°C or the hardness at room temperature of two glasses is
not appropriate.
Silica (and to some extent other similar glass formers that are
continuously fourfold, three-dimensionally coordinated, e.g.,
germania and beryllium fluoride) is unique, although very
similar to water in many ways. Because of silica’s continuous,
tetrahedral, open, three-dimensional network, it has many
anomalous properties—low and even negative thermal expan-
sion coefficient, as well as positive temperature and negative
pressure coefficients of elastic constants. All of these anoma-
lous properties disappear when silica is sufficiently broken
down as described above. There is an added problem in at-
tempting to understand the continual change from anomalous
to normal behavior, because phase separation makes glasses
with moderate concentrations of network modifier inacces-
sible. (Although, recently, sol–gel techniques appear to have
overcome this problem.
81
) Although it is known that germania
and beryllium fluoride have structures that are very similar to
silica (corner-joined tetrahedra) and also show the various
anomalies that make silica unique, attempts to use these glasses
as models usually fail for one or more reasons. Perhaps the
most important characteristics to examine in comparing these
glasses are their viscosity–temperature behavior, thermal ex-
pansion, and chemical durability. These are strikingly different
in silica glasses on the one hand and silicate glasses on the
other. If we have an idea of how these properties work and can
Cable on Neri/Merrett
In an interesting early proposal of a structural model,
Cable
71
recently has suggested that Neri/Merrett
25
was a
proto-Zachariasen: ‘‘So far as structural models are con-
cerned, there is a really intriguing comment in Merret that
shows him thinking that forming a glass was a matter of the
geometry (the ‘‘Figure’’) of the separate oxides and how
they could fit together: ‘Besides diaphaneity is a property
not communicated to any thing malleable, and who would
call that Glass, that were not transparent which are incon-
sistent with the nature of Glass. For the materials of Glass,
Sand, and salts, have such figures as seem incapable of such
adhaesion in every part one to another. For all salts have
their determinate Figure which they keep too, in their great-
est solutions and actions of the fire upon them, unless a total
destruction be wrought upon them, as many instances might
evince, and that Figure is various according to the salts.
Salt-peter, and all Alcalizate-salts are pointed, and by their
pungency, and caustickness seem to be made up of infinite
sharp pointed needles, And as for Sand the Figure thereof is
various, nay, infinite it appears in Microscopes . . . . Now
how can any man imagine that such variety of Figures in
Sand can so comply with the determinate Figures of salt as
to touch one another in minimis which is necessary to make
it malleable? Whereas to make it Glass ’tis enough that
those two touch one another at certain points onely,
whereby such an union it formed, which is necessary to
denominate Glass but wholly incompatible with malleabil-
ity. And this union is that which makes in Glass Pores, from
whence comes it’s diaphaneity as you have heard from Lu-
cret. Besides something said before, declares that they both
remain the same in the compound [as] they were before. I
shall conclude this argument and say, that I conceive that
nothing but the Elixir will perform this effect, and that both
of them will come into the world together.’ ’’
25
Cable thus argues that Neri/Merrett suggested that the
differences between the opaque and plastic nature of, say,
metallic solids on the one hand and transparent brittle glassy
solids on the other hand is due to their microstructure, or
‘‘Figures’’ and that, in glasses, the microstructure is perhaps
open as a result of bonding only at points, i.e., corners,
whereas, in metals, the very compact, close-packing, is re-
sponsible for their behavior.
Fig. 6. Variation of the logarithm of viscosity versus Tg/Tfor strong
and fragile liquids ((䊊) SiO
2
,(䉭) GeO
2
,(䊉)Na
2
O⭈2SiO
2
, (+) metal-
lic glass, and (䉮) heavy-metal fluoride glass) (modified from Ref. 83).
806 Journal of the American Ceramic Society—Kurkjian and Prindle Vol. 81, No. 4
be controlled, in a general sense, we may be able to design a
glass with some degree of success. The qualitative difference in
the viscosity of silica and other normal glasses is shown in Fig.
6. Silica demonstrates Arrhenius behavior over the logarithm
of viscosity range from ∼1 to 14, whereas normal glasses dem-
onstrate almost a continuously changing logarithm of viscosity,
or one that can be approximated by two straight lines—one
with E≈35 kJ/mol and one with E≈10 kJ/mol.
82
These
viscosity behaviors are controlled by continuous Si—O—Si
bonding in silica (an anomalous or strong liquid) and the
breakup of this bonding by network modifiers (Na—O or
Ca—O) in a normal or fragile liquid. At some high temperature
(low viscosity) in the normal liquid, the flow occurs by the
motion of silicate units, whereas, at lower temperature for the
normal liquid and at nearly all temperatures in silica, the flow
is presumably by some type of bond-interchange mechanism.
Angell
83
has noted this behavior, and his plot of viscosity
versus the reciprocal temperature normalized by the glass tran-
sition temperature (T
g
) illustrates in a striking manner the dif-
ferences just discussed (Fig. 6). His terminology for his behav-
ior is ‘‘strong’’ (or anomalous for silica, germania, and
beryllium fluoride) and ‘‘fragile’’ (for other normal glass-
forming liquids).
The diffusion and, thus, the electrical conductivity also are
qualitatively different for these anomalous and normal glasses.
The conduction in silica is by impurities, because the diffusion
coefficients of silicon and oxygen are extremely small,
whereas, in a sodium-glass or sodium–calcium-glass, these
ions have appreciable mobility and, thus, result in substantial
conductivity, although the glass remains basically an insulator.
The mobility of these ions obviously can be either helpful or
detrimental. In the case of a glass tube for the processing of
semiconductors, or for an insulating film for the same semi-
conductor, the use of a sodium-containing glass can result in
undesirable contamination with sodium. On the other hand, the
mobility of the cations makes these glasses suitable for ion
exchange, whereby the glass can be strengthened or can be-
come a waveguide by virtue of the changes in index brought
about by the ion exchange. This ion exchange, however, also
results in at least one mechanism responsible for the poorer
chemical durability of soda–lime–silica glasses.
(3) Invert Glasses
The use of too great a content of flux (sodium, potassium,
calcium, magnesium, etc.) results in glasses that are prone to
devitrification, or that are so nondurable that they either dis-
solve or become undesirably clouded. This is understandable
when one considers the CRN model. If too much network
modifier is added, the network is no longer a network. The
network and ion mobilities are so great that rearrangements can
occur readily, and the possibility of crystallization is enhanced.
However, in one of the many surprises in glass composition–
property behavior, Trapp and Stevels
84
have developed a series
of so-called invert glasses (although it has been pointed out by
Weyl and Marboe
85
that Ha¨nlein
86
had studied glasses of a
very similar type many years before that had a composition (in
wt%) of 35K
2
O⭈15Na
2
O⭈10CaO⭈40SiO
2
). This description in
terms Y(the number of bridging oxygens) is essentially the
form used earlier by Stevels:
●For Y⳱4, if O/Si ⳱2, this is a three-dimensional network,
all four corners bridging;
●For Y⳱3, if O/Si ⳱2.5, this is a two-dimensional sheet,
three corners bridging;
●For Y⳱2, if O/Si ⳱3, this is a one-dimensional chain, one
corner bridging;
●For Y< 2, if O/Si ⳱4, this is an individual SiO
4
tetrahedra,
all oxygens nonbridging.
If the stiffness of the network is decreased gradually by replac-
ing the bridging oxygens by nonbridging oxygens until only
individual, isolated tetrahedra remain, crystallization occurs
very readily. However, Trapp and Stevels, as Ha¨nlein before
them, has shown that, by the incorporation of many modifiers,
crystallization can be frustrated. In fact, as shown in Fig. 7 and
as is evident from the ‘‘invert’’ name, some properties show an
inversion in their trend at Y≈2. In such cases, although the
kinetics of crystallization are enhanced, the probability of crys-
tallization may be made more difficult because of the lack of a
simple, most stable structure. More-recent work in glasses with
mixed anions shows that they increase the glass-forming range
and produce a variety of behaviors.
(4) Metallic Glasses
Only recently has it been possible to produce bulk amor-
phous metals, i.e., metals that can be retained in the glassy state
while cooling at rates of the order of degrees per minute. This
means that samples centimeters in all dimensions can be pro-
duced. This is possible using essentially the same technique
used in the case of the invert glasses. The liquidus temperature
is decreased at the same time that the compositional compli-
cation is increased. As a result, crystallization is frustrated.
The first glassy metals were produced at California Institute
of Technology by a group directed by Duwez,
87
who was
studying the formation of metastable metal phases by splat
cooling. For the following many years, this group and others
led by Turnbull
88
at Harvard University produced a variety of
amorphous metals and studied their properties and structure.
However, all of these glass compositions required cooling rates
of at least hundreds of degrees celsius per second. Recently,
Inoue
89
and co-workers in Japan (La−Ni−Cu, Mg−Y−Cu−Ni,
and Zr−Ni−Cu−Al) and Johnson
90
and co-workers at Cali-
fornia Institute of Technology (Zr−Ti−Cu−Be) succeeded in
producing the bulk glassy metals described above. These
Fig. 7. Viscosity isotherms of an invert glass system (Na
2
O–K
2
O–
CaO–SrO–BaO silicate).
84
April 1998 Perspectives on the History of Glass Composition 807
glasses could be maintained in the glassy state by cooling at
moderate rates, and their viscosity behavior was more remi-
niscent of silica than of other less-stable glasses. In fact, these
new glasses were as strong as normal soda–lime–silica or
Na
2
O⭈2SiO
2
glasses—i.e., a network glass with three of the
four doubly bonded oxygens intact (Fig. 8).
(5) Nonnetwork and Noninvert Glasses—Some New and
Interesting Optical Glasses
Most of the commercial glass compositions that we have
considered so far contain a traditional glass former. They have
been developed for rather general purposes—windows, con-
tainers, high-temperature and thermal shock applications, and
optical applications. Although properties have to be tailored so
that optical glasses can be worked, the main consideration has
been the optical properties—refractive index and dispersion. It
is somewhat understood what types of atoms and/or ions are
required to obtain the desired property—the refractive index is
controlled by both the electronic and atomic contributions to
refraction, whereas the dispersion is controlled by the position
and magnitude of the intrinsic ultraviolet absorptions. There-
fore, the idea essentially has been to find a glass solvent that
can dissolve the required quantity and type of solute. Until
1936, according to Weyl and Marboe,
85
‘‘. . . all glasses made
commercially contained, as a major constituent, a compound
which, by itself, could form a glass.’’
Morey
91
then discovered and patented the Kodak glasses.
While developing a series of rare-earth borate glasses that
greatly expanded the dispersion–refractive index region (see
Fig. 2), Morey learned that some glasses can be produced with-
out B
2
O
3
. An example is (in wt%) 15ZrO
2
⭈45Ta
2
O
5
⭈40TiO
2
.
These glasses were described after Ha¨nlein published on his
invert glasses, and, even so, they were equally as unexpected,
if not more so, because, although some of them contained a
small amount of B
2
O
3
as the glass former, many of these par-
ticular compositions contained no traditional glass former, per
se. Somewhat after the development of these Kodak glasses,
K. H. Sun,
92
a collaborator of Morey’s, patented an equally
unexpected type of glass. Again, although some of these
glasses used beryllium fluoride, which was Goldschmidt’s and
Zachariasen’s model glass former, some of them contained
only aluminum fluoride as the apparent glass former (in
mol%): 34A1F
3
⭈50PbF
2
⭈10MgF
2
⭈10SrF
2
⭈2BaF
2
⭈1CaF
2
⭈
1LaF
3
⭈1CeF
3
⭈1ThF
4
.
Besides the unusual glasses described above, glasses in sev-
eral aluminate, gallate, tungstate, molybdate, titanate, niobate,
and tantalate systems, i.e., glasses with no real glass former,
have been produced. These glasses often have unexpected
properties, and it is interesting that glasses can be formed in
such systems.
(6) Heavy-Metal Fluoride Glasses
There is an obvious similarity between the aluminum fluo-
ride glasses of Sun and the heavy-metal fluoride (HMF) glasses
discovered accidentally at the University of Rennes, in France,
in 1975.
93
At the time of their discovery, indeed even until
now, although there has been speculation about the structure,
etc., of the two types of glasses discovered by Morey and Sun,
little real structural work has been done. On the other hand,
because of the possibility of a major use for the HMF glasses
as lightguides and also because of the much larger glass sci-
ence/technology population in 1974, an enormous amount of
research of all types has been conducted on these glasses. De-
spite the very substantial efforts expended on this topic, the
issue of the structure of HMF glasses is yet more elusive than
that of most oxide glasses. This is caused in part because there
is no really fixed structural unit, as, for instance, a SiO
4
tetra-
hedron. Although it generally is accepted that zirconium is the
network former, it has variable coordination with fluorine (7–
8). Also, as opposed to the corner sharing of silica tetrahedra
that is known in both crystalline and glassy silicates, both edge
and corner sharing are known in crystalline HMF crystals.
(7) Phosphate Glasses
Phosphate glasses have an interesting history. It was discov-
ered very early that phosphorus pentoxide is a glass former. It
differs from the other classic glass formers in that, although it
forms a tetrahedral three-dimensional network (because phos-
phorus is pentavalent), one of the four oxygen ions is doubly
bonded, that is, not coupled to the network. Although a silica-
like structure is possible, as in aluminum phosphate, normal
alkali and alkaline-earth glasses tend to have chainlike struc-
tures because of this pentavalency.
In 1941, Kreidl and Weyl
94
discussed the history of these
glasses. As indicated above, phosphates were added by many
early investigators, and, in fact, it appears that Vernon Har-
court
39
did most of his work on these glasses. The first Schott
catalog, in 1886,
1
included so-called phosphate crown glasses.
In 1894, Elder and Valenta
95
found that phosphate crowns had
much better ultraviolet transmission than other optical glasses,
and, when the beneficial physiological effects of ultraviolet
radiation were discovered in 1925, research increased. At about
the same time, it was discovered that phosphates were only
slightly soluble in hydrofluoric acid. American Optical
96
de-
Neri/Merrett on the Brittleness of Glass
Although other Neri/Merrett comments about the struc-
ture of glass might be subject to some differences in inter-
pretation, the following statements are quite unequivocal:
‘‘tis of its own nature the most brittle thing in the world, and
to make it malleable a quality quite contrary to it’s nature
must be introduced. Besides, diaphaneity is a property not
communicated to anything malleable, and who would call
that Glass, that were not transparent? As well may one name
that Gold which is not ponderous nor malleable, as that
Glass which is malleable and not transparent. Add hereunto,
that the nature of malleability consists in a close and
throughout adhaesion of parts to parts, and a capacity to the
change of Figure in the minutest parts. Both which are
inconsistent with the nature of Glass.’’
25
Fig. 8. Probability of crack initiation in a (䊐) standard and a (䊏)
less-brittle soda–lime–silica glass.
103
808 Journal of the American Ceramic Society—Kurkjian and Prindle Vol. 81, No. 4
veloped glasses containing high concentrations of iron, but
which were relatively colorless in the visible, while absorbing
harmful infrared wavelengths. The low melting and processing
temperatures and low viscosity of phosphate glasses made
them useful as solder glasses.
A very important use of phosphate glasses emerged in 1988.
Although silicate glasses were usually preferred for laser
glasses because of their superior thermomechanical properties,
phosphates were found to have superior laser characteristics—
line widths, laser cross section, etc., as well as low, nonlinear
refractive index and essentially zero dependence of optical path
length on temperature. However, perhaps their most interesting
property, in this case, is their ability to dissolve ionic plati-
num,
97
and, thus, such phosphate glasses could be produced
essentially free of inclusions. This is a very necessary quality
of a high-power laser glass.
98
Michael Faraday was probably
the first glassmaker to use platinum to contain molten glass:
‘‘Platina also was ultimately found to answer perfectly the
purpose of retaining the glass, . .. . Neither the glass nor any
of the substances entering into its composition, separate or
mixed, had the slightest action on it.’’
37
Often it is desirable to have a glass with a very low softening
point as well as a high thermal expansion coefficient, for in-
stance, for sealing glass to aluminum. Phosphate glasses have
been developed for such purposes. The high expansion coeffi-
cient with the associated low T
s
is quite reasonable, because it
is expected that, by increasing the amount of flux and, thus,
decreasing the melting or glass temperature (decreasing the
bond strength in the solid), the thermal expansion coefficient
also increases for the same reason. Unfortunately, the usual
additional consequence of this reduced bond strength is a de-
crease in chemical stability—an increase in solubility or cor-
rosion rate.
Perhaps even more interesting, then, is the set of glasses
recently discovered by Tick.
99a
Incredibly, Tick discovered a
fairly wide range of glass formation in the lead tin fluorophos-
phate system. Certain compositions in this field, developed for
ophthalmic applications, show a valuable combination of prop-
erties—low melting temperature, T
m
, low glass temperature,
T
g
, and high expansion coefficient, while retaining respectable
chemical durability, less than 50 times greater weight loss than
soda–lime float glass (0.02 versus 1 mg/(cm
2
⭈d)). The durabil-
ity depends strongly on the amount of lead and tin, but not
strongly on the amount of fluorine. Small changes in lead con-
tent have a very strong impact on the durability. It is, first of all,
surprising that, in such a system where there are probably at
least two glass formers, the system had not been discovered
and studied earlier. Even more surprising is that these glasses
can be melted at T≈400°C with T
g
≈100°C, and yet have
chemical durabilities comparable to soda–lime–silica glasses.
Although several structural models have been proposed, that of
Brow et al.
99b
seems most likely. In a modification of Tick’s
original model, they find extensive P—F bonding and suggest
that the reason for the special properties of these glasses is the
result of unusual electron lone-pair structures associated with
the Sn
2+
and Pb
2+
sites.
(8) New Transparent Glass-Ceramics
Another exciting, unexpected, and, to some extent, unex-
plained area is that of transparent glass-ceramics. Although
such materials were discovered by Stookey quite early in the
history of glass-ceramics, these new materials show transpar-
encies that are of a completely different order. In 1993, Wang
and Ohwaki
100
reported efficient infrared-to-green up-conver-
sion using Er
3+
in a transparent oxyfluoride glass-ceramic. The
rare earth (Er
3+
) apparently dissolves preferentially in these
crystals, and, thus, the absorption–emission behavior is that of
a crystalline rather than a glassy host. Although this is useful,
the more important effect is that the transmission of the com-
bined glass-ceramic is comparable to that of the glass host
itself. That is, the scattering caused by the presence of these
small crystals, 20–30 vol% of crystals ∼9–18 m average size,
is small. Further work on such systems by Tick, Borelli, and
co-workers
101
has shown that this effect can be more or less
explained using a model proposed by Hopper.
102
His model
assumes that, rather than small crystals in a continuous glass
matrix, the system can be approximated as a continuum with
spatial variations of refractive index. The development of these
materials is another example of the unexpected.
(9) Less-Brittle Glasses
One of the most active areas of glass research over the years
always has been that of glass strength. It is very interesting and
surprising, therefore, that, in all of the thousands of papers
dealing with the strength of glasses, as far as the authors are
aware, not one specifically has suggested that the resistance of
the glass surface to crack initiation might be improved by
changes in the bulk glass composition. Reduction in damage to
the surface has been accomplished by the generation of a com-
pressive stress in the surface, either by thermal or chemical
means or by the application of coatings.
A recent publication, presentation, and patent from Ito, Seh-
gal, and co-workers
103
reveals the interesting possibility of a
substantial improvement in the abrasion resistance of silicate
glasses by rather modest compositional changes. Figure 8
shows that, by small compositional changes (see Table II),
marked improvements in cracking susceptibility have been pro-
duced. Although these workers have proposed a model to ex-
plain the effects that they observed, the model certainly appears
to deserve much more fundamental study.
(10) Other Special Glasses
Glass R&D over the years has produced an array of glass
compositions–properties that have been unexpected and that
initially have been used in special applications. As technology
has changed, however, occasionally these uses also have
changed. Often limited uses become, if not extensive, at least
more commercially important. Chalcogenide glasses that were
used first as index-matching liquids have become important
because of their infrared transparency. This transparency is the
result of low vibrational (phonon) frequencies. These low pho-
non frequencies, in turn, make these glasses important for fiber
lasers. Fiber lasers were discovered by Koester and Snitzer
104
in the mid-1960s and were of virtually no interest until the
development of suitable diode lasers in the mid-1990s. Diode
lasers are now a critical component in long-distance telecom-
munications systems. Although the final choice of the glass
host for diode lasers has not been made (clearly, there will be
more than one such material), heavy-metal fluoride glasses and
sodium–zinc–telluride glasses may be the host of choice in
some cases.
Finally, a special category of glasses is bio-glass. Bio-
glasses are useful and interesting because they do not fit into
the criteria described in the Introduction—they are nondurable!
Table II. Glass Compositions
†
Oxide Composition (wt%)
Standard soda–lime–silica glass
SiO
2
66–75
MgO 0–5
CaO 7–12
Na
2
O 12–20
K
2
O 0–3
Al
2
O
3
0–4
Scratch-resistant glass
SiO
2
75.5–85.5
RO 1–8
R⬘
2
O 10–23.5
RO + R⬘
2
O 11–24.5
Al
2
O
3
0–5
RO/R⬘
2
O (weight ratio) 0.5 (at most)
†
Reference 103. The authors of this work recently have indicated that, in a dry
nitrogen environment, the loads necessary to produce two cracks increases to 350 g
for the standard glass and to 3500 g for the scratch-resistant glass.
April 1998 Perspectives on the History of Glass Composition 809
Their reactivity, which allows the rapid formation of surface
gels, apparently is the reason for their suitability.
VI. New Era of Optical Fibers
The impact of optical communications technology has been
so swift and pervasive that life without it is all but forgotten.
Those who are in the glass industry are perhaps more conscious
of this development than the average person, and, therefore,
only a short discussion of this phenomenon is necessary. How-
ever, as one of the present authors has recently emphasized, the
importance of the work of early glass researchers to this in-
dustry should not be underestimated.
105
The main body of this article has been concerned with a
discussion of rather complex silicate and, in one instance, phos-
phate glasses, and the rather natural and ordinary properties
that are obtainable in such systems. In the case of lightguide
fibers, almost the opposite characteristics are seen: the compo-
sitions and structures are rather simple, natural, and ordinary,
but the properties that are observed are very unpredicted and, in
some cases, at least as yet, unexplained. The history of the
origins of the glass lightguide have been well documented.
After early theoretical and experimental work by Kao and co-
workers
106,107
at the Standard Telecommunications Laboratory
in Harlow, England, in 1970, Maurer and co-workers
108
at
Corning produced the first fiber with low optical loss (20 dB/
km). This fiber consisted of a titania–silica core glass and a
pure-silica cladding. The core material, as required for light-
guiding, had a higher refractive index than the core glass. The
primary reason that the loss is not lower than 20 dB/km is that
the titanium in the core is easily reduced from 4+ to 3+ at the
high processing temperatures. This results in ultraviolet ab-
sorption with a tail in the visible and near infrared.
Intensive R&D work followed worldwide on all aspects of
this technology. There were two primary difficulties. First, not
many organizations were familiar with the handling and pro-
cessing of fused silica. Second, even if the organization had
such a capability, it was neither immediately obvious what
glass could be substituted for the titania-doped silica first used
by Corning nor what procedures could be used to fabricate such
composite, or glass-clad glass, fibers. As indicated, the core
had to have a refractive index higher than the cladding, and
few, if any, known glasses had a refractive index less than
silica. Thus, although silica was known to be rather easily
obtained in high-purity, high-transmission form, it could not be
used as the core—the light-carrying portion of the waveguide.
For these reasons, most initial work was directed at the puri-
fication of multicomponent silicate glasses. It proved ex-
tremely difficult to purify these glasses sufficiently, and, con-
sequently, most workers eventually turned their attention to the
‘‘silica’’ glasses first shown by the Corning workers to be so
promising.
Three processes have been developed to produce lightguide
fibers, which are now mainly germania-doped silica as the core
material and pure silica as the cladding.
109
In the first process
(OVD), the glasses are built up on the outside of a ‘‘bait’’ rod
in the form of a powder or soot by depositing them from a torch
moving in the axial direction. Thus, the preform is built up
radially, layer by layer. This preform is sintered in a subsequent
operation. In a second process (MCVD), the (core or core and
deposited cladding) glasses are laid down again layer by layer,
on the inside of a silica tube, which itself may act as the
cladding glass. The layers are sintered in situ as a part of this
operation. In the third process, the preform is built up in an
axial rather than in a radial direction and is thus called vapor
axial deposition (VAD). In the first two processes, the index
gradient is produced stepwise by depositing layers, each with
constant refractive index. In the VAD process, the composition
and, thus, the desired refractive index variation, can be pro-
duced and maintained across the gas stream.
To date, ∼100×10
6
km of lightguide fiber have been de-
ployed worldwide, with ∼5×10
6
km of fiber being deployed
per year.
110
This fiber has ∼0.2 dB/km loss, essentially the
intrinsic optical loss for fused silica. The fiber usually has a
germania–silica core glass and a silica or fluorine-doped silica
cladding, and operation is at 1300 or 1550 nm. Although han-
dling, connecting, and deploying of such brittle fibers has pre-
sented a challenge, rather outstanding success has been
achieved.
The use of lightguide fibers requires a variety of devices for
the control and manipulation of the light signals. Here, we
briefly describe the role of glass fibers as active components.
As indicated in an earlier section, there are areas of rather
intense R&D that should be mentioned here, because they re-
inforce our earlier premise that these relatively simple glasses
show some quite remarkable effects. One of these effects is the
formation of fiber Bragg gratings by laser irradiation.
111,112
These devices are extremely useful for the control of optical
signals and for the detection of a variety of environmental
effects when used as sensors (the sensor area itself is one of
enormous activity at the moment, but one that we have no
space to discuss).
113
The mechanism of the formation of grat-
ings is not yet completely understood. Although the ordinary
refractive index difference produced by this irradiation is of the
order of 10
−4
, by using special techniques, such as hydrogen
loading or flame brushing, index increases of up to 10
−2
can be
realized. It initially was believed that the generation of Ge-E⬘
centers and the associated absorption resulted in the refractive
index change through the Kramers–Kronig transform. How-
ever, it now is believed that the observed increases in refractive
index are too large to be the result only of such a mechanism,
and structural changes with associated volume compaction also
have been shown to be involved.
As a final indication of the complexity of the effects that
have been demonstrated in these glasses, an effect that contin-
ues to receive attention, but that also continues to defy com-
plete understanding, is second harmonic generation (SHG).
This effect—the production of light of twice the energy (half
the wavelength) of the generating light—was first discovered
in 1986 by Osterberg and Margulis.
114
When it was first ob-
served in lightguide glasses, its discovery was quite unex-
pected, because it requires a noncentrosymmetric medium. The
mechanism that seems best to describe this effect is the freez-
ing-in of a periodic direct-current field. This field is probably
the result of the asymmetric photoionization of defects in the
glass.
115
Although this direct-current field was initially pro-
duced by the interaction of the fundamental and the induced or
applied second harmonic beams, such SHG also has been
found to be formed by poling.
116,117
In this technique, a direct-
current of >10 kV is applied to a sample at ∼200°–300°C to
form the grating. The exact model to explain this effect or the
maximum magnitude of the effect has not yet been determined
completely.
Although all of the above devices or effects have been dis-
covered more or less outside of the glass industry and basically
not by traditional glass scientists, as described by one of the
authors in another report,
105
basic work done earlier, as usual,
will form the basis for the ultimate understanding of these
effects. The study of defects in silica and in other glasses, as
well as the study of direct-current conductivity and electrode
polarization over the years will be extremely valuable.
The glass compositions used in the lightguide industry, al-
though extremely simple, had not been the subject of very
much experimental or theoretical work prior to the interest in
lightguides. As indicated in an earlier section, very little data
were available on any of the interesting properties—such as
refractive index, density, or viscosity—and this hampered
progress in the development of both waveguides and associated
devices (behavior of RE ions, defects for gratings, etc.). Al-
though few surprises have resulted in this regard from the work
that followed, it continues that very little detailed study has
810 Journal of the American Ceramic Society—Kurkjian and Prindle Vol. 81, No. 4
been conducted on these glass-former glasses, except on those
properties of direct interest to lightguides.
VII. Conclusions
The history of glass composition has been a very long and
extremely interesting one and one that has many lessons. Al-
though there are apparently many constraints on what can be
done within the framework of an amorphous inorganic material
(i.e., materials that lack a periodic lattice), we do not fully
understand the theories that govern such behavior, and, thus,
many surprises can be expected in this fascinating field. His-
torically, we have seen that progress has been cyclic and spotty
at best. Ideas often have appeared ahead of their time. For
instance, fiber lasers were invented by Snitzer in 1963 and
attracted almost no attention until the mid 1990s, when the
necessary associated electronics (appropriate diode lasers)
were developed to make them commercially viable. Metallic
glasses, on the other hand, were more or less continuously
studied for 25 years before compositions were discovered that
allowed bulk metallic glasses to be produced, and now hope-
fully to be more generally useful commercially. The tin lead
fluorophosphate glasses that defy the natural law of low T
m
,
low T
g
, and low chemical durability are not really extraordinary
compositions—in fact they are quite ordinary CRN glasses—
and yet they were discovered only recently because of the need
for such very-low-melting glasses.
In the past, glass scientists have studied the behavior of
many simple and moderately complex glasses and have a rea-
sonably good qualitative idea of the role of various oxide con-
stituents in the determination of the behavior of these glasses.
Also, we have available reasonably good empirical models that
allow quantitative prediction of some aspects of their behavior.
However, such understanding and models are able to predict
behavior only over relatively narrow compositional ranges.
Also, although many interesting properties have been dis-
covered in a variety of glass systems, perhaps only in the recent
case of lightguide fibers has such a discovery been the basis for
a major new industry. This development has had an enormous
impact on both the industry and the science of glass. A whole
new generation of glass optical fiber scientists has appeared.
An important final question is whether we can expect that
another such major development may occur in the future, or
whether we have only to look forward to small, interesting
developments in glass science and technology.
Acknowledgments: The authors would like to thank especially Michael
Cable and Prabhat K. Gupta for continuous advice and counsel. They also would
like to thank many other friends, particularly Roger Araujo, George Beall,
Robert Brill, Allan Bruce, Paul Danielson, Leonid Glebhov, R. Hilton, Kai
Karlsson, Dieter Krause, Alex Marker, Susan Morse, Eliezer Rabinovich, Ju¨r-
gen Steiner, and Paul Tick, for their help and advice, even though it may not
necessarily have been taken. The authors have written this work with the insight
and inspiration of the late Professors N. J. Kreidl and A.R. Cooper always in
mind.
References
1
Anonymous, ‘‘Production and Price Catalogue,’’ Glastechnisches Labora-
torium, Schott and Genossen, Jena, Germany, 1886.
2
O. Schott, ‘‘On Glass Works for Optical and Other Scientific Products,’’
Verein z. Beforderung des Gewerbefleisses, (1888).
3
A. Winkelmann, ‘‘On the Specific Heats of Glasses of Various Composi-
tions,’’ Ann. Phys. Chem.,49, 401 (1893).
4
A. Winkelmann and O. Shott, ‘‘On the Elasticity and the Tensional and
Compressional Strength of Various New Glasses, in Their Relation to Chemical
Composition,’’ Ann. Phys. Chem.,51, 697 (1894).
5
G. W. Morey, The Properties of Glass, 1st and 2nd eds. Reinhold, New
York, 1938 and 1954.
6
J. Zarzycki, Glasses and the Vitreous State. Cambridge University Press,
New York, 1991 (see this for a good list of early books).
7
A. Feltz, Amorphous Inorganic Materials and Glasses. VCH, New York,
1993 (see this for a good list of later books).
8
A. K. Varshneya, Fundamentals of Inorganic Glasses. Academic Press,
New York, 1994.
9
J. Shelby, Introduction to Glass Science and Technology. American Chemi-
cal Society, Washington, DC, 1997.
10
V. D. Frechette, Non-Crystalline Solids. Wiley, New York, 1960 (start of
Proceedings of the Physics of Non-Crystalline Solids Conferences).
11
J. D. Mackenzie, Modern Aspects of the Vitreous State. Butterworth, Wash-
ington, DC, 1964.
12
R. H. Doremus and M. Tomozawa, Treatise on Materials Science and
Technology, Glass. Academic Press, New York, 1985.
13
D. R. Uhlmann and N. J. Kreidl, Glass: Science and Technology. Academic
Press, New York, 1983.
14
O. El-Bayoumi and D. R. Uhlmann, ‘‘Proceedings of the Symposium on
Glass Science and Technology: Problems and Prospects for 2004’’ (dedicated to
N. J. Kreidl on the occasion of his 80th birthday), J. Non-Cryst. Solids,73, 1985.
15
M. Cable and J. W. Smedley, ‘‘Michael Faraday—Glass Maker,’’ Glass.
Technol.,30, 39 (1989).
16
M. Cable and J. W. Smedley, ‘‘William Vernon Harcourt: Pioneer Glass Sci-
entist and Founder of the British Association,’’ Glass Technol.,33, 92 (1992).
17
M. Cable, ‘‘Threads of Glass,’’ Glass Technol.,29, 181 (1988).
18
W. D. Kingery, The Social and Cultural Contexts of New Ceramic Tech-
nology. American Ceramic Society, Westerville, OH, 1995.
19
W. D. Kingery, The Changing Role of Ceramics in Society. American Ce-
ramic Society, Westerville, OH, 1997.
20
R. H. Brill, ‘‘Scientific Investigations of the Jalame Glass and Related
Finds’’; Ch. 9, p. 257 in Excavations at Jalame Site of a Glass Factory in Late
Roman Palestine. Edited by G. D. Weinberg. University of Missouri Press,
1986.
21
(a) R. H. Brill, ‘‘Scientific Investigation of Some Glasses from Sedeinga,’’
J. Glass Stud.,33, 11 (1991). (b) R. H. Brill, ‘‘Chemical Analysis of Some
Glasses from Frattesina,’’ J. Glass Stud., 34, 11 (1992).
22
G. B. Hares, ‘‘3500 Years of Glassmaking’’; p. 1 in Advances in Ceramics,
Vol. 18, Commercial Glass. Edited by D. C. Boyd and J. F. McDowell. Ameri-
can Ceramic Society, Columbus, OH, 1986.
23
G. Agricola, De Re Metallica (in Latin). Froben, Basel, Switzerland. Trans-
lated by Herbert and Lou Hoover. Dover Press, New York, 1950.
24
R. H. Brill; private communication, 1997.
25
A. Neri, L’Arte Vetraria, 1679.
26
R. H. Brill, ‘‘Crizzling—A Problem in Glass Conservation’’; p. 121 in
Conservation in Archeology and the Applied Arts. International Institute for
Conservation of Historic and Artistic Works, London, U.K., 1975.
27
W. A. Weyl, Coloured Glasses. Society of Glass Technology, Sheffield,
U.K., 1951.
28
W. Ganzenmuller, ‘‘Contributions to the History of Gold Ruby Glass,’’
Glastech. Ber.,15, 346 (1937).
29
B. E. Moody, ‘‘The Life of George Ravenscroft,’’ Glass Technol.,29, 198
(1988).
30
Anonymous, ‘‘W. E.S. Turner,’’ Glass Technol.,3, 201 (1962).
31
R. W. Douglas and S. Frank, A History of Glassmaking. G. T. Foulis,
Henley-on-Thames, U.K., 1972.
32
E. B. Haynes, Glass Through the Ages. Penguin, London, U.K., 1948.
33
F. J. T. Maloney, Glass in the Modern World. Doubleday, New York, 1967.
34
J. Strong, Concepts of Classical Optics. W. H. Freeman, San Francisco,
CA, 1958.
35
E. W. Deeg, ‘‘Optical Glasses’’; see Ref. 22, p. 9.
36
W. Vogel, Chemistry of Glass. American Ceramic Society, Columbus, OH,
1985.
37
M. Faraday, ‘‘On the Manufacturing of Glass for Optical Purposes’’; in
Experimental Researches in Chemistry and Physics. Taylor and Francis, Lon-
don, U.K., 1859.
38
L. P. Williams, Michael Faraday. Basic Books, New York, 1964.
39
G. G. Stokes, ‘‘Notice of the Researches of the Late Rev. William Vernon
Harcourt, On the Conditions of Transparency in Glass, and the Connexion Between
the Chemical Composition and Optical Properties of Different Glasses,’’ Report of
the British Association for the Advancement of Science, 1871.
40
H. Hovestadt, Jena Glass and Its Industrial Applications. Macmillan, Lon-
don, U.K., 1902.
41
M. Cable, ‘‘Classical Glass Technology’’; in Glasses and Amorphous Ma-
terials, Materials Science and Technology Series, Vol. 9. Edited by J. Zarzycki.
VCH, New York, 1991.
42
J. Steiner, ‘‘Otto Schott and the Intervention of Borosilicate Glass,’’
Glastech. Ber.,66, 165 (1993).
43
R. S. Ryder and J.P. Poole, ‘‘Container Glass’’; see Ref. 22, pp. 35–42.
44
Anonymous, ‘‘Eugene Cornelius Sullivan—Glass Scientist,’’ Corning
Glass Works, Corning, NY, 1964.
45
J. H. Munier, ‘‘Perspective of the Role of Research, Development, and
Engineering in the Corning Glass Works, the First 100 Years,’’ Corning, Inc.,
Corning, NY, 1976.
46
P. Auborg and W. W. Wolf, ‘‘Glass Fibers’’; see Ref. 22, pp. 51–64.
47
W. Dumbaugh and P. Danielson, ‘‘Aluminosilicate Glasses’’; see Ref. 22,
pp. 115–32.
48
F. M. Locke and F.J. Locke, ‘‘Glass,’’ U.S. Pat. No. 1 529 259, 1922.
49
M. Reamur, Memoires de l’Academie des Sciences, 1739.
50
S. D. Stookey, Journey to the Center of the Crystal Ball. American Ceramic
Society, Columbus, OH, 1985.
51
S. D. Stookey, ‘‘Catalyzed Crystallization of Glasses in Theory and Prac-
tice,’’ Ind. Eng. Chem.,51, 805 (1959).
52
W. H. Armistead and S. D. Stookey, ‘‘Photochromic Silicate Glasses Sen-
sitized by Silver Halides,’’ Science (Washington, DC),144, 50 (1964).
53
R. W. G. Wyckoff and G. W. Morey, ‘‘X-ray Diffraction Measurements of
Some Soda–Lime–Silica Glasses,’’ J. Soc. Glass Technol.,9, 265 (1925).
April 1998 Perspectives on the History of Glass Composition 811
54
G. Tammann, The States of Aggregation. van Nostrand, New York,
1925.
55
V. M. Goldschmidt, ‘‘Investigation Concerning Structure and Properties of
Crystals,’’ Skr. Nor. Vidensk.–Akad. Kl. 1: Mat. Naturvidensk. Kl.,1926, [8]
130 (1927).
56
R. B. Sosman, The Properties of Silica. The Chemical Catalogue Co., New
York, 1927.
57
H. Rawson, Inorganic Glass-forming Systems. Academic Press, New York,
1967.
58
G. Hagg, ‘‘The Vitreous State,’’ J. Chem. Phys.,3, 42 (1935).
59
W. H. Zachariasen, ‘‘The Vitreous State,’’ J. Chem. Phys.,3, 162 (1932).
60
W. H. Zachariasen, ‘‘The Atomic Arrangement in Glass,’’ J. Am. Chem.
Soc.,54, 3841 (1932).
61
A. R. Cooper, ‘‘Zachariasen’s Rules, Madelung Constant, and Network
Topology,’’ Phys. Chem. Glasses,19, 60 (1978).
62
J. C. Phillips, ‘‘Spectroscopic and Morphological Structure of Tetrahedral
Oxide Glasses,’’ Solid State Phys., 37, 93 (1982).
63
A. R. Cooper, ‘‘W. H. Zachariasen—The Melody Lingers On,’’ J. Non-
Cryst. Solids,49, 1 (1982).
64
P. K. Gupta and A. R. Cooper, ‘‘Topologically Disordered Networks of
Rigid Polytopes,’’ J. Non-Cryst. Solids,123, 14 (1990).
65
C. S. Marians and L. W. Hobbs, ‘‘The Phase Structure of Aperiodic SiO
2
as
a Function of Network Topology,’’ J. Non-Cryst. Solids,106, 309 (1988).
66
B. E. Warren and A. D. Loring, ‘‘X-ray Diffraction Study of the Structure
of Soda–Silica Glass,’’ J. Am. Ceram. Soc.,19, 202 (1936).
67
B. E. Warren, H. Krutter, and O. Morningstar, ‘‘Fourier Analysis of X-ray
Patterns of Vitreous Silica,’’ J. Am. Ceram. Soc.,19, 202 (1936).
68
A. A. Lebedev, ‘‘Polymorphism and Annealing of Glass,’’ Tr. Gos. Opt.
Inst.,2, 1 (1921).
69
S. N. Nemilov, ‘‘Moritz Ludvig Frankenheim (1801–1869)—Author of the
First Scientific Hypothesis of Glass Structure,’’ Sov. J. Glass Phys. Chem.,21,
148 (1995).
70
M. L. Frankenheim, The Teaching of Cohesion, Comprising the Elasticity
of Gases, the Elasticity and Coherence of Liquid and Solid Bodies, and Crystal
Science. A. Schultz, Breslau, Germany, 1835.
71
M. Cable; private communication, 1997.
72
G. E. Peterson, C. R. Kurkjian, and A. Carnevale, ‘‘Random Vector Sta-
tistical Studies of Amorphous Materials’’; p. 369 in Boron in Glass. Edited by
L. D. Pye and V.D. Frechette. Plenum Press, New York, 1977.
73
E. A. Porai-Koshits, ‘‘The Structure of Glass’’; in Proceedings of the XI
International Congress on Glass (Prague, Czechoslovakia, 1977).
74
R. J. Bell and P. Dean, ‘‘The Structure of Vitreous Silica: Validity of the
Random Network Theory,’’ Nature (London),212, 1354 (1966).
75
G. Turner, R. Kirkpatrick, S. Risbud, and D. Oldfield, ‘‘Multinuclear
Magic-Angle Sample–Spinning Nuclear Magnetic Resonance Spectroscopic
Studies of Crystalline and Amorphous Ceramic Materials,’’ Am. Ceram. Soc.
Bull.,66, 656 (1987).
76
P. H. Gaskell, ‘‘Models for the Structure of Amorphous Solids’’; see Ref.
41.
77
A. Wright, ‘‘Neutron Scattering from Vitreous Silica, V, The Structure of
Vitreous Silica: What Have We Learned from 60 Years of Diffraction Studies?’’
J. Non-Cryst. Solids,179, 84 (1984).
78
G. E. Peterson, C. R. Kurkjian, and A. Carnevale, ‘‘Random Struc-
ture Models and Spin Resonance in Glass,’’ Phys. Chem. Glasses,15,52
(1977).
79
A. R. Cooper, ‘‘Connectivity and Easy Glass Formation,’’ Mater. Sci. Fo-
rum,67&68, 385 (1991).
80
M. Spallek, ‘‘Market Prospects for Special Glasses,’’ Ind. Ceram.,13, 167
(1993).
81
E. M. Rabinovich, A. J. Bruce, and P. L. Trevor, ‘‘Multicomponent Glasses
from Particulate Gels,’’ J. Non-Cryst. Solids,160, 126 (1993).
80
M. Spallek, ‘‘Market Prospects for Special Glasses,’’ Ind. Ceram.,13, 167
(1993).
81
E. M. Rabinovich, A. J. Bruce, and P. L. Trevor, ‘‘Multicomponent Glasses
from Particulate Gels,’’ J. Non-Cryst. Solids,160, 126 (1993).
82
U. C. Paek, C. M. Schroeder, and C. R. Kurkjian, ‘‘Determination of the
Viscosity of High Silica Glasses during Fibre Drawing,’’ Glass Technol.,29,
263 (1988).
83
C. A. Angell, ‘‘Structural Instability and Relaxation in Liquid and Glassy
Phases Near the Fragile Point,’’ J. Non-Cryst. Solids,102, 205 (1988).
84
H. J. L. Trapp and J.M. Stevels, ‘‘Conventional and Invert Glasses Con-
taining Titania, Part 1,’’ Phys. Chem. Glasses,1, 107 (1960).
85
W. A. Weyl and E.C. Marboe, The Constitution of Glasses. Interscience,
New York, 1962.
86
W. Ha¨nlein, ‘‘Physical Properties of Glasses of the System SiO
2
−Na
2
O−
K
2
O−CaO System,’’ Z. Tech. Phys.,14, 418 (1933).
87
W. Klement, R. H. Willens, and P. Duwez, ‘‘Non-Crystalline Structure in
Solidified Gold–Silicon Alloys,’’ Nature (London), 187, 869 (1960).
88
H. S. Chen, ‘‘Glassy Metals,’’ Rep. Prog. Phys,43, 353 (1980).
89
A. Inoue, A. Kato, T. Zhang, S. G. Kim, and T. Masumoto, ‘‘Mg-Cu-Y
Amorphous Alloys with High Mechanical Strengths Produced by a Metallic
Mold Casting Method,’’ Mater. Trans.,JIM,32, 609 (1990).
90
E. Bakke, R. Busch, and W. L. Johnson, ‘‘The Viscosity of the
Zr
46,75
Ti
8.25
Cu
7.5
Ni
10
Be
27.5
,’’ Appl. Phys. Lett.,22, 3260 (1995).
91
G. W. Morey, ‘‘New or Improved Glasses,’’ Br. Pat. No. 462 304, 1937.
92
K. H. Sun, ‘‘Fluoride Glass,’’ U.S. Pat. No. 2 466 509, 1949.
93
M. Poulain, M. Poulain, J. Lucas, and P. Brun, ‘‘Fluorine-Containing Glass
with ZrF
3
: Optical Properties of Glass Doped with Nd
3+
,’’ Mater. Res. Bull.,10,
243 (1975).
94
N. J. Kreidl and W. A. Weyl, ‘‘Phosphates in Ceramic Ware: IV, Phosphate
Glasses,’’ J. Am. Ceram. Soc.,24, 372 (1941).
95
J. M. Elder and E. Valenta, ‘‘Absorption Spectra of Colorless and Colored
Glasses,’’ Denkschr. Mat. Naturwiss. Akad. Wien,61, 285 (1894).
96
A. G. Pincus, ‘‘Glass Compositions and Methods of Making Same,’’ U.S.
Pat. No. 2 359 789, 1944.
97
J. S. Hayden, D. L. Sapak, and A. J. Marker, ‘‘Elimination of Metallic
Platinum in Laser Glasses,’’ Proc. SPIE—Int. Soc. Opt. Eng.,895, 176 (1988).
98
M. J. Weber, ‘‘Science and Technology of Laser Glass,’’ J. Non-Cryst.
Solids,123, 208 (1990).
99
(a) P. A. Tick, ‘‘Water-Durable Glasses with Ultralow Melting Tempera-
tures,’’ Phys. Chem. Glasses,25, 149 (1984). (b) R. K. Brow, C. C. Phiefer,
X. J. Xu, and D. E. Day, ‘‘An XPS Study of Anion Bonding in Tin(II) Fluo-
rophosphate Glass,’’ Phys. Chem. Glasses,33 [2] 33–39 (1992).
100
Y. Wang and J. Ohwaki, ‘‘New Transparent Vitroceramics Codoped with
Er
3+
and Yb
3+
for Efficient Frequency Upconversion,’’ Appl. Phys. Lett.,63,
3268 (1993).
101
P. A. Tick, N. F. Borrelli, L. K. Cornelius, and M.A. Newhouse, ‘‘Trans-
parent Glass Ceramics for 1300 nm Amplifier Applications,’’ J. Appl. Phys.,78,
6367 (1995).
102
R. W. Hopper, ‘‘Stochastic Theory of Scattering from Idealized Spinodal
Structures II. Scattering in General. The Basic Late Stage Model,’’ J. Non.-
Cryst. Solids,70, 111 (1985).
103
J. Sehgal, J. Kase, A. Takada, H. Takahashi, Y. Nakao, and S. Ito,
‘‘Scratch-Resistant Glass,’’ Eur. Pat. Appl. No. EP 0 757 021 A1, 1996; per-
sonal communication, 1997; and J. Am. Ceram. Soc., in review.
104
C. J. Koester and E. Snitzer, ‘‘Amplification in a Fiber Laser,’’ Appl. Opt.,
3, 1182 (1966).
105
C. R. Kurkjian and D. M. Krol, ‘‘The Science and Technology of Silica
Lightguides for Telecommunications’’; in Structure and Imperfections in
Amorphous and Crystalline Silica. Edited by R. A. Devine. Wiley, New York,
1998.
106
C. K. Kao and G. A. Hockham, ‘‘Dielectric Fibre Surface Waveguide for
Optical Frequencies,’’ Proc. IEE,113, 1151 (1966).
107
M. W. Jones and C. K. Kao, ‘‘Spectrophotometric Studies of Ultra Low
Loss Optical Glasses II, Double Beam Method,’’ J. Sci. Instrum.,Series 2,2,
331 (1969).
108
F. Kapron, D. B. Keck, and R. D. Maurer, ‘‘Materials Development of
Optical Fibers,’’ Appl. Phys. Lett.,17, 423 (1970).
109
J. B. MacChesney and D. G. DiGiovanni, ‘‘Materials Development in Op-
tical Fiber,’’ J. Am. Ceram. Soc.,73, 3537 (1990).
110
A. J. Mocenigo, ‘‘The Competitiveness of U.S. Manufacturers in the Glob-
al Market for Fiber Optics,’’ Fiber Integr. Opt.,15, 63 (1996).
111
K. O. Hill and G. Meltz, ‘‘Fiber Bragg Grating Technology Fundamentals
and Overview,’’ J. Lightwave Technol.,15, 1263 (1997).
112
B. Poumellic, P. Guenot, I. Riant, P. Sansonetti, P. Niay, P. Bernage, and
J. F. Bayon, ‘‘UV-Induced Densification During Bragg Grating Inscription in
Ge:SiO Preforms,’’ Opt. Mater.,4, 441 (1995).
113
R. Kashyap, ‘‘Photosensitive Optical Fibers: Devices and Applications,’’
Opt. Fiber Technol.,1, 17 (1994).
114
U. Osterberg and W. Margulis, ‘‘Dye Laser Pumped by Nd:YAG Laser
Pulses Frequency Doubled in a Glass Optical Fiber,’’ Opt. Lett.,11, 516 (1986).
115
E. M. Dianov, and D. S. Starodubov, ‘‘Photoinduced Second-Harmonic
Generation in Glasses and Glass Optical Fibers,’’ Opt. Fiber Technol.,1,3
(1994).
116
R. A. Meyers, N. Mukerjee, and S. R. Brueck, ‘‘Large Second-Order Non-
linearities in Poled Fused Silica,’’ Opt. Lett., 16, 1732 (1991).
117
P. Kazansky, P. St. J. Russell, and H. Takebe,, ‘‘Glass Fiber Poling and
Applications,’’ J. Lightwave Technol.,15, 1484 (1997). 䊐
812 Journal of the American Ceramic Society—Kurkjian and Prindle Vol. 81, No. 4
Charles R. Kurkjian is a member of the Fiber and Component Reliability Group of
Bell Communications Research (Bellcore). He joined Bellcore in October, 1994, after
retiring from AT&T Bell Laboratories, Murray Hill, NJ, where he was a Distin-
guished Member of Technical Staff. Dr. Kurkjian received his B.Sc. in ceramic
engineering from Rutgers University in 1952 and his Sc.D. from Massachusetts
Institute of Technology in 1955. He held post-doctoral positions at MIT and the
University of Sheffield, England, before joining Bell Laboratories in 1958. Since that
time he has been a Distinguished Visiting Professor, American University in Cairo
(1971); Visiting Fellow, Cambridge University (1978); and National Academy of
Science Visiting Scholar, USSR (1986). He is a member of the National Academy of
Engineering and the Academy of Ceramics and is a Fellow of the Society of Glass
Technology and the American Ceramic Society. Dr. Kurkjian has published more
than 125 technical papers, has edited one book, and holds seven patents. He received
the Morey Award in 1987 and delivered the A. L. Friedberg Memorial Lecture in
1992. Dr. Kurkjian has been chairman and trustee of the Glass Division of the
American Ceramic Society (Glass and Optical Materials Division) and a Vice Presi-
dent of the Society. He has conducted investigations of glass structure and properties
using various types of spectroscopy—NMR/EPR, Mossbauer, optical, and acoustic.
His current efforts are in the area of mechanical properties of glasses, particularly
optical fibers.
William R. Prindle received his B.S. and M.S. in physical metallurgy from the
University of California at Berkeley and received his Sc.D in ceramics from Mas-
sachusetts Institute of Technology. In 1954 he joined the Hazel–Atlas Glass Division,
Continental Can Company, and became General Manager of R&D before leaving in
1962 to become Manager of Materials Research, American Optical Company. In
1966 he joined Ferro Corporation in Cleveland, where he became Vice President-
Research. In 1971 he returned to American Optical Corporation, where he was Vice
President and Director of Research until 1976. From 1976 to 1980 he was in Wash-
ington, DC, as Executive Director of the National Materials Advisory Board, a unit
of the National Research Council of the National Academy of Sciences. He joined
Corning Incorporated at the end of 1980 and retired in 1992 as Division Vice Presi-
dent and Associate Director, Technology Group. Prindle is a Distinguished Life
Member and Fellow of the American Ceramic Society and was President of the
Society 1980–1981. He received the Norton Award of the New England Section in
1974, the Toledo Glass and Ceramic Award of the Northern Ohio Section in 1986,
and the Bleininger Award of the Pittsburgh Section in 1989. In 1990 he presented the
Friedberg Memorial Lecture of the NICE. In 1983 he received the Phoenix Award as
Glass Man of the Year. He also was President of the International Commission on
Glass 1985–1988, and is a member of the National Academy of Engineering.
April 1998 Perspectives on the History of Glass Composition 813
