1

CHAPTER 1
INTRODUCTION TO SILICATE GLASS TECHNOLOGY

1.1 GLASS AND ANTIQUITY

G L A S S is one of man's most useful and versatile materials. It is also one of

the oldest. Obsidian, a glass formed by volcanic action, had widespread usage
during the Stone Age because it was easily flaked to form sharp cutting edges
for arrowheads, spearheads, knives and the like. Its use has been documented
in East Africa as early as 150,000 years ago and it was used at least to the end
of the Stone Age in the New World, about 1500 years ago. Indeed, obsidian
has been known as the "Dating Stone" (1) because of the ease to which it
lends itself in age determination. Wherever prehistoric peoples and volcanic
glass occurred together in nature, humans made use of this natural product as
a raw material for manufacture of tools. This has become evident from the
widespread discovery of detritus, in the form of worked- fragments and
shards, found in the vicinity of present- day deposits of eruptive rocks in
many parts of the world.

Obsidian varies in appearance from nearly clear colorless to pinkish or even

gray-green, brown, red, and black. Its usefulness as a tool- making material
derives from the ease with which it furnishes the sharp edges required for
cutting or scraping. Obsidian is readily flaked by percussion and pressure
m e t h o d s similar to those used for w o r k i n g flint. However, the edges
produced are much sharper than any other material, except perhaps that of
worked metals.

The obsidian dating method depends upon the fact that, while the glass
originally formed contains about 0.1 to 0.3% water by weight, water from the
atmosphere continues to slowly diffuse into the body of the glass over a long
period of time. Obsidian itself is a complex silicate containing about twelve
major components, including silica. Depth of penetration of the water is a
measure of how long it has been since the surface of obsidian was created, or
2

when the stone was last chipped or cracked. If one removes an edge of glass
from a given obsidian artifact and views it under magnification in polarized
light, the edge that has been penetrated by water has a higher refractive
index than that of the pristine glass. This allows one to observe an interface
b e t w e e n the h y d r a t e d p a r t and the n o n - h y d r a t e d part, seen in the
microscope as a thin line. The uptake of water also changes the volume of the
affected glass, and causes a mechanical strain which renders it visible as a
bright band under polarized light due to an effect called strain- birefringence.
The thickness of the rind may vary from 1.0 micron (0.00004 inches) to
more than 50 microns. Dates have been established as far back as 200,000
years, and have been authenticated, in many cases, by radioactive Krypton/
Argon studies. It is thus clear that man (Homo Sapiens) has used natural glass
from dates close to that of his origination.

Perhaps the earliest known artificial glass is the glaze found on stone beads
of the B a r b a r i a n A g e ( 1 2 , 0 0 0 B . C . ) . T h e oldest known articles made
completely from glass have been dated to about 7000 B.C. There is a tale that
glass manufacture was d i s c o v e r e d in a n t i q u i t y by s a i l o r s w h o w e r e
transporting large blocks of soda-ash to a distant port. They happened to land
on a beach and used the blocks as a wind- screen for a bonfire. The heat of
the fire caused the combination of sand and soda-ash to form a melt. This,
when cooled, produced a hard product similar to obsidian. Thus, artificial
glassmaking was born.

"Modern" glassmaking began in Egypt about 3500 years ago. This took the
form of fibers, flattened plates, dishes and utensils, and bottles of varied
shapes. Glass fibers were used as early as 1500 B.C. after the discovery that
glass could be drawn into fine filaments from a melt. It was also discovered
that additions of manganese, nickel, cobalt, copper, iron, and chromium
could be used to produce "colored" glasses. In the Egyptian XVIIIth Dynasty,
glassmakers created small vases for cosmetics by attaching either a small clay
form or a bag filled with sand to a stick. Rotating this form in molten glass
resulted in a small vase of the desired shape. Coarse fibers of colored glass
were then heated and pressed into the external walls of the vessel to make
decorative patterns. Similar decoration, using colored glass fibers, was used
in the Rhineland of Middle Europe during the Roman Empire to make glass
dishes, cups and bottles, i.e.- up until about 100 A.D.
 3

However, the production of "flat" glass posed a seemingly insurmountable

obstacle for nearly 2000 years. Somewhere around 200 A.D., the invention of
the g l a s s - b l o w i n g pipe in S y r i a permitted glassmakers to produce a
forerunner of flat glass. Glass was blown into a large cylindrical shape, then
cut and either "spun" or ironed flat before it hardened. The first colorless
glass had been made about 100 years earlier by adding manganous oxide to
the melt. Although attempts to manufacture acceptable flat glass went on for
centuries, it was not until the 1600's that the first plate glass was produced in
France by cutting and grinding. Until the late eighteenth century, flat glass
was an expensive handmade luxury. Crown spinning, that is- formation of a
blown bulb of hot glass which was cut and then spun to form a flat disc of
glass, was the predominant method of making flat glass in this era. And it is
only in recent times that man has learned how to make plate glass by a "float"
method. This will be described in more detail below.

During the past 3000 years, man has learned how to make controlled melts of
glass. Glass technology has advanced slowly, with more progress in the past
80 years than in all the preceding years that glassmaking had existed. This
p r o c e s s b e g a n in the 1800's w i t h the I n d u s t r i a l R e v o l u t i o n and the
establishment of scientific research in Industry. Systematic studies of many
thousands of compositions, the development of special manufacturing
techniques, and the discovery of processes for attaining new and unusual
properties, forms, and uses, have made glass an extremely versatile material.
At present, about 700 different glass compositions are in use which have been
fabricated into tens of thousands of different articles for varied uses.

Noteworthy among these modern developments are:

1. Air tempering to improve strength (some standard- temper products

are strong enough to withstand severe abuse such as use as a hammer to
drive nails).

2. Chemical tempering has produced glasses with strengths of more

than 100,000 psi. in tensile strength.

3. Photosensitive glasses which reversibly darken and fade in

response to light intensity exposure.
4

4. Glass-ceramics, which are partly crystalline bodies formed by

additives which cause crystals to grow in the final glass product.

5. The development of continuous melting and high-speed automatic

forming of high quality glass articles such as bottles, jars, electric light
bulbs, window panes, and the like, in millions of tons.

6. Development of extensive knowledge of glass composition ranges,

and the c h e m i c a l - p h y s i c a l p r o p e r t i e s which determine whether
specific compositions can be melted to form a useful glassy state.

7. Preparation of glass by a "sol-gel" method. This method consists of

precipitating a silicic- acid as a gelatinous composite, drying the gel,
and then consolidating the dried mass into a solid glass by heating. The
advantage lies in the use of much lower temperatures to achieve glass
articles of complex composition. We will describe this method in more
detail later on in this Chapter.

Nevertheless, most of the prior-known glasses are silicate- based. That is, the
main ingredient is sand (which is silicon dioxide, S1O2 , containing small
3
amounts of impurities). When other glass-formers such as phosphate, P O 4 -,
have been used, the glass products obtained were found to be extremely
unstable to hydrolysis by boiling water. Indeed, because of perceived
commercial advantages, Schott Glaswerke of Jena, Germany spent many years
of research during the early 1900's attempting to obtain stable phosphate
based g l a s s e s . T h e r e are many chemical and structural reasons why
p h o s p h a t e - based g l a s s e s are not easily o b t a i n e d in a stable state.
Notwithstanding, they can be prepared so that their stability to boiling water
exceeds that of many of the silicate-based glasses. This is the subject of this
manuscript. Indeed, the method to be described involves a completely new
approach to non- silicate glass-making. We will describe the unique "rules"
required for preparing such glasses, which contrast diametrically to those
developed over many thousands of years for silicate-based glasses.

But, as we shall see, the main reason that Schott Glaswerke did not succeed is
that they used the known glass-technology, based on rules developed for
silicate glasses, to attempt to force the phosphate molecules to conform to
 5

preconceived ideas of g l a s s - p r e p a r a t i o n . W e will describe the correct

approach, based on the peculiarities of the PO4- tetrahedron, which utilizes
the tendency of the tetrahedra to form long chains of polymeric molecules of
specific coordination.

But first, let us examine the structure of glasses and then summarize silicate-
glass technology so as to be able to compare the newest polymerization
technology with that already developed for silicate glasses over a period of
many thousands of years.

1.2 THE GLASSY OR VITREOUS STATE

Glass is a non-crystalline solid. That is, it does not consist of an ordered

arrangement of its constituent cations and anions. The American Society for
Testing and Materials (2) defines glass as "an inorganic product of fusion
which has cooled to a rigid condition without crystallizing". It has also been
defined as "an inorganic substance analogous to a rigid liquid".

If one obtains an x-ray analysis of glass, the resulting diffraction pattern

consists of a broad featureless band, in contrast to a crystal of the same
composition. For example, sodium silicate is soluble and can be dissolved to
form a solution. If one then adds acid and precipitates a silicic-acid, i.e.- H2
S1O4 or S1O2 x H2O, and separates it from the solution (note that the sodium
ion is lost), a silica- gel is first obtained, containing considerable water.
Heating causes a loss of some of the water. Further heating causes the gel to
crystallize in one of several forms of silica such as crystobalite.

If the crystalline phase is then melted and cooled more rapidly, a glassy phase
results. As shown in Figure 1.2.1., given on the next page, the x-ray patterns
of these three forms of silica have very specific, but differing, features.

Note that crystobalite shows specific diffraction lines indicating a crystalline

structure, while the diffraction pattern of the other two forms exhibits only a
broad featureless band. Yet, all three have identical chemical compositions,
i.e.- S i 0 2 .
6

Figure 1.2.1.-

X-ray Patterns of Three Forms of S i l i c a

—I 1 1 1 1 Γ
Crystobalite I

I I I I I I I I
0 .04 .08 .12 .16 .20 .24 .28
Sinßf/λ

For many years, the structure of glass remained a mystery. In 1932,

Zachariasen (3) proposed that atoms in silicate glass form an extended three-
dimensional network which, although lacking the symmetry usually found in a
crystalline structure, has an energy content comparable to that of the crystal.
This means that the coordination numbers are approximately equal in both
phases. Using these concepts, Zachariasen proposed a set of "rules"
regarding the vitreous state. These are concisely stated in the following:

1.2.2.- Zachariasen's Theorems Relating to the Amorphous State

a. Each oxygen atom must be linked to no more than two

cations.
 7

b. The number of oxygen atoms around any one cation must

be small, i.e. - three to four.

c. The oxygen polyhedra must share corners, not faces or

edges, in order to form a three-dimensional network.

d. At least 3 corners of each polyhedra must be shared.

Zachariasen was aware that the S1O4 - tetrahedron was the structural basis for
silicate- glasses, but attempted to expand his concepts to include all possible
types of glass compositions. Work by Warren (4) and others supported
Zachariasens concepts. Based on the S1O4 - tetrahedron, it was obvious that
the network would have to be three dimensional. From a Fourier analysis of
the diffraction pattern of simple glasses composed of either: S 1 O 2 , B 2 O 3 , or
sodium silicates, radial distribution curves were obtained which showed the
distribution of neighboring atoms about each atom in the glass. It was the
combination of this work and that of Zachariasen which led to the random
network theory for the structure of glass. Additionally, early workers
postulated that glass-structure might well consist of "building blocks" which,
when joined in a random configuration, represented the structure of glass.
These represented "short-range order", but not "long-range order" in the
amorphous solid.

The following diagram, given as Figure 1.2.3. on the next page, shows a
representation, in two dimensions, of the differences between crystalline and
amorphous silica, the latter a simple g l a s s . K e e p in m i n d t h a t in a
4+
tetrahedron, each S i atom is surrounded by four (4) 0= atoms.

It should be apparent in this diagram that both structures are three-

dimensional with the next layer of tetrahedra attached to the top apex
(oxygen atom) of the tetrahedral arrangement of atoms comprising the
structures. T h u s , considerable differences in structure exist between
crystalline and amorphous phases, even though they may have identical
chemical compositions.
8

Figure 1.2.3.

Structures Based on Tetrahedra

Glassy or Amorphous S t a t e

The conspicuous difference to be noted is that the volume occupied by the

same number of tetrahedra is greater for the amorphous state as compared to
the crystalline state. Indeed, such a volume-temperature relationship had
long been known, where the amorphous state occupies a larger specific
volume than the crystalline state.

This is illustrated in the following diagram, given as Figure 1.2.4. on the next
page. In this Figure is shown the volume-temperature relation when a solid
transforms into a liquid, and then cools to an amorphous solid. It is easy to
see that a definitive change in volume occurs, as the nature of the phase
changes.
 9

Figure 1.2.4.-

Volume-Temperature Relation Between the

C r y s t a l l i n e , Glassy and Liquid States

Transformation Range | Liquid

Melt1ng||
Point

Crystal

Temperature

Glass is not a supercooled liquid. This is apparent in the above diagram where
it is evident that a crystal, heated to its melting point, forms a liquid of
greater volume. When this liquid is cooled, it will either recrystallize to form
an ordered solid of decreased volume, or it may form an amorphous solid,
particularly if the composition is comprised of so-called "glass-formers". In
the latter case, the liquid becomes supercooled until it reaches a certain
temperature, generally called the transformation point (or temperature), T g ,
where a "glassy-state" is formed. This is seen in the above diagram as the
"break-point" point in the cooling curve. For the amorphous state, this
behavior is related to the absence of crystal nuclei and/or a melt viscosity too
h i g h to a l l o w s i g n i f i c a n t c r y s t a l - n u c l e i f o r m a t i o n r a t e s . T h u s , a
transformation range exists, which would not be present if glass were
merely a supercooled liquid.

It was originally thought that the break in slope of the cooling curve given
above (transformation range) was due to a kinetic factor. The structural
configuration frozen-in at this temperature would then persist approximately
10

unchanged at all lower t e m p e r a t u r e s . If one measures the change in

expansion of the glassy state below the transition temperature, one finds that
it is a linear function of temperature, as shown in the following Figure:

Figure 1.2.5.-

Expansion of Glass as a Function

of Temperature

Softening Point- Τ

T
9 *

y / ^ Slope = Expansion
Coefficient

Temperature in °C.

For most solids, expansion is usually linear. However, for glass, the expansion
is linear up to the glass transformation point, T g , Then an anomalous
expansion occurs, corresponding to the transformation range given above in
Figure 1.2.4., i.e.- [ T g 2 - T g lJ , which includes the softening p o i n t , T s p.

It is this type of data (and others) which support the present concept of the
amorphous state, but not that of the supercooled liquid. The work of Warren
(1942) and other authors, who used x-ray diffraction to study the structure of
g l a s s , s t r e n g t h e n e d Z a c h a r i a s e n s c o n c e p t s of a three dimensional,
amorphous network.

An idealized multicomponent glassy structure is shown in the following

diagram, given as Figure 1.2.6. on the next page, wherein several cations are
illustrated in the diagram:
 11

Figure 1.2.6.-

A Schematic Representation of a Multicomponent Glass

This structure is quite similar in concept to that already given in Figure 1.2.2.
for silica, except that the structure is modified by the presence of several
cations, each having its own degree of coordination. Note that three different
c a t i o n s a r e d e p i c t e d in t h i s t w o - d i m e n s i o n a l r e p r e s e n t a t i o n of a
multicomponent glass based on tetrahedra. Each c a t i o n has its own
distinctive coordination in the glassy structure.

Diffraction patterns of glass have a very diffuse nature, and are similar to those
obtained from liquids. Originally, it was thought that the diffuse nature of the
x- ray patterns might due to very small crystals. However, there is no small
12

angle scattering from glass. To confirm this, examine the curves of Figure
1.2.1. again. Note the small-angle scattering present in silica gel, which has
been shown to be due to the existence of discrete particles of some 10-100 À
in size, with gaps or voids between them. The absence of such scattering in
the x-ray pattern for glass is perhaps the most persuasive experimental fact
concerning this phenomenon and indicates that the bonding must be
essentially continuous. It must therefore be concluded that glass is not a
supercooled liquid. Because the position of the main diffraction bands of glass
and gel in Figure 1.2.2. are comparable to that of the principle diffraction line
for crystobalite, the conclusion can be drawn that the interatomic distances
are about equal for all three materials. From a Fourier analysis of such
diffraction patterns, radial distribution curves can be obtained which show the
distribution of neighboring atoms about each atom. It is the combination of
this work and that of Zachariasen which led to the random network theory
for the structure of glasses.

The most recent work on glass structure has strengthened this concept.
Although glasses typically display short-range order among nearest atomic
neighbors, no observations had been made where longer range order was
discerned. Using a special method of neutron-diffraction, Gaskell et al (5)
observed the first evidence for medium structural order in an oxide glass.
Their study of the distribution of calcium ions in a simple calcium silicate
g l a s s r e v e a l e d that the i m m e d i a t e v i c i n i t y of a c a t i o n - that is, that
encompassing its first few neighbors- looks very similar to the environment of
a crystal. Thus, the cations shown in Figure 1.2.6. as an example of a multi-
component glass do not have a readily apparent arrangement, and the major
structural arrangement appears to be that of the silicate tetrahedra which
comprise the amorphous structure. However, the work of Gaskell et al was
the first to reveal that the cations do have an intermediate-range order.

This is seen in the following diagram, shown as Figure 1.2.7 on the next
page. Here, we have shown but ten (10) octahedra, i.e.- CaOß - comprised of
2+
the C a cation surrounded by six (6) oxygens, derived from the surrounding
tetrahedra, arranged in a pattern having a medium range of order. The
coordinate-covalent bonds of the Ca-O pairs arise from the "lone-pair"
electrons always associated with the oxide-atom.
 13

Figure 1.2.7.-

Order in Glass: Local Order in a Calcium Silicate Glass

Thus the cations apparently impose an "order" on the tetrahedra, even though
the overall structure is amorphous in nature.

However, it is easily seen, in the above Figure, that the amorphous structure
comprised by the silica-tetrahedra is the major contributor to the overall
vitreous structure. Thus, there is a "block" arrangement of the tetrahedra
which form a short-range order along with an arrangement of cations having a
medium-range order. This sequence comprises the overall vitreous structure.
Such a structure would have a three-dimensional form where the basic
structure is formed from tetrahedra. Additionally, cations having a medium-
14

range order would be present, but would be buried within the block-
structure. Such an arrangement is illustrated in the following Figure:

Figure 1.2.8.-

Three-Dimensional Arrangement of Tetrahedra in an

Amorphous Structure

Herein is shown an arrangement of tetrahedra with short-range order, but not

with long range order. It should thus be evident from this discussion that the
structure of glass does have a degree of order, but not a completely arranged
order of atoms like a crystalline material.

1.3. GLASS FORMERS AND GLASS COMPOSITIONS

On the basis of his rules, Zachariasen divided the simple oxides into those
which should form glasses, and those that were structure modifiers. These
are shown in Table 1.1. Both the oxides and the structural units are given,
along with the oxygen- metal single bond strength. This general correlation
was first made by Sun in 1947 (6) who suggested that glass forming capability
was dependent upon the central bonding strength of the structural unit, as
shown in the following Table:
 15

TABLE 1,1
Glass Formers and Structural Modifiers
Glass Pormers
SINGLE
GLASS COORDINATION VALENCE STRUCTURAL BOND
FORMERS UNIT OF CATION UNIT ENERGY
{KCAL/MOLI

B2O3 BO2 +3 TRIANGULAR 119

4
S1O2 S1O4 - +4 TETRAHEDRAL 106

4
Ge02 Ge04 " +4 TETRAHEDRAL 108

AI2O3 A10 2- +3 TRIANGULAR 101

P205 PO4 -
3 +5 TETRAHEDRAL 111

3
V2O5 VO4 - +5 TETRAHEDRAL 112

3
AS2O5 ASO4 - +5 TETRAHEDRAL 87

3
Sb 05 Sb0 " +5 TETRAHEDRAL 85
2 4

Glass Modifiers
COORDINATION IN GLASS
B2O3 B- 0 + 3 4 89

T1O2 Ti - Ο +4 6 73

Zn Zn - Ο +2 2 72
4 36
Pb Pb-O +2 2 73
+4 6 39

Al Al-O +3 6 67
16

TABLE 1.1 (Continued)

Glass Formers and Structural Modifiers SINGLE
GLASS NETWORK VALENCE COORDINATION BOND
FORMERS MODIFIERS OF CATION IN GLASS ENERGY

Zr Zr - Ο + 4 8 61

La La-O + 3 7 58

Y Y-O +3 8 64

Ga Ga-O + 3 6 45

Mg Mg - Ο + 2 6 37

Ca Ca-Ο + 2 8 32

Sr Sr-O + 2 8 32

Ba Ba - Ο + 2 8 33

Cd Cd-Ο + 2 2 60
4 30
6 20

Li Li-Ο + 1 4 36

Na Na - Ο + 1 6 20

Κ Κ- Ο + 1 9 13

Rb Rb - Ο + 1 10 12

Cs Cs-O + 1 12 10

Bond strengths of glass formers are all above 80 Kcal/mole. All of these form
covalently bonded anions to which the network modifiers are ionically
 17

bonded in the glassy n e t w o r k . It is, perhaps, easier to v i s u a l i z e the

relationship of glass-formers to each other, and to glass- modifiers, by
considering the following Figure:

Figure 1.3.1.-

The Periodic Table as Related to Glass Formers and Modifers

1+ Glass Formers

H 2+ 3+ 4+ 5+ - 1 He
Glass Modifiers
Li Be Β F Ne
Na Mg Al Si Ρ S Ar
Κ Ca Sc Ti V Cr Μη Fe Co Ni Cu Cu Zn Ga Se As Se Kr
Rb Sr Y Zr Cd Sb Te xe

Cs Ba Le Hg Pb Rn
A c 104

Ce Pr Nd Sm Eu 6d Tb Dy HO Er Tm YD
(3 + )
Th U

In this Figure, the glass-formers are given in the larger bold type, while the
glass modifiers are given in normal type. It is easy to see that those elements
generally regarded as anionic , i.e.- those which form anions easily, are those
which are usually classified as glass-formers. Those elements which are
cationic in nature are those which are glass (structure) modifiers.

Thus, glass-formers include: borates, silicates, germanates, phosphates,

sulfates, and to a much lesser extent, arsenates and antimonates, as well as
sulfides, and fluorides. Note also that vanadates are included as major glass-
formers. The glass network- modifiers include: alkali metal cations, alkaline
earths, and many of the transition metal cations (of varied valencies), plus the
lanthanides, in some cases.

Sometimes in glass parlance, network modifiers are further divided into

stabilizers and fluxes, according to the functions for which they are added to
make commercial melts. For example, fluxes are added to lower melting and
18

working temperatures of silicate glasses by decreasing viscosities, e.g.- Na20 ,

K2O and B2O3 . Stabilizers are added to improve chemical durability and/or to
prevent crystallization, e.g.- CaO, MgO, and AI2O3.

In general, the following oxides are added to a specific glass formulation to

achieve the following results:

Alkali Oxides, e.g.- Na20 and K2O: added as fluxes to reduce high viscosities
of glass melts.

Alkaline earth oxides, e.g.- CaO and BaO: added to produce glasses resistant
to water and acid leaching, i.e.- high chemical durability

Boric oxide : added to increase durability and stability of glass products

Aluminum oxides: added to increase "hardening of glass", i.e.- resistance to

devitrification and improved durability.

Glass-network structures tend to be covalent. The Si-O, Ge-O, P-O, B-O, As-O
and Sb-0 bonds, according to Pauling's rules, are all 50% covalent, or greater.
Modifiers are used to alter glassy properties, including refractive index ,
electrical conductivity and expansion coefficients. An intermediate, although
not capable of forming a glass by itself, can enter into a glass-forming position
and/or in a modifying position.

Notwithstanding, it should be obvious that the network structure which each

glass former is capable of assuming is defined by the following:

1.3.2.- Network Dimensionality = Coordination of Structural Unit - 1

Thus, only those anionic structural units having a coordination number of four
can form a three-dimensional network.

The following table illustrates this aspect:

 19

TABLE 1.2.
Dimensional Networks Formed by the Various Glass - Formers

GLASS STRUCTURAL COORDINATION TYPE OF

FORMER UNIT OF UNIT DIMENSIONALITY
Cation Space

B2Q3 BO2- 3 2 Chains

3
BO3 - 4 3 Network

4
S1O2 S1O4 - 4 3 Network

4
Ge02 Ge04 " 4 3 Network

AI2O3 A10 2- 3 2 Chains

3
P2O5 PO4 - 3 2 Chains

3
V2O5 VO4 - 3 2 Chains

3
AS2O5 As04 ~ 4 3 Network

3
S b 20 5 Sb04 ' 4 3 Network

Note that many non-silicate glasses form chains (two - dimensional) as the
primary configuration. Of those which do form three-dimensional networks,
4 4_
only S 1 O 4 - and G e 0 4 have bond strengths sufficient to form glasses
possessing high three- dimensional stability, with the desired properties of
chemical durability and other properties of commercial interest. S i 0 2 is by far
the most important glass-forming system (because g e r m a n i u m is too
expensive). T h u s , most c o m m e r c i a l glass melts are based on silicate
compositions. A number of commercial formulas have arisen, and are being
modified continuously in the Industry.

Table 1.3. gives some approximate compositions of commercial silicate-based

glasses. The principle usage, and % by w e i g h t are given for the major
components used to make the melts:
20

TABLE 1.3.
APPROXIMATE COMPOSITIONS OF COMMERCIAL GLASSES
Glass* Percent by Weight
Code Usage SiOo Μ2Ω3Β2Ω3 NagO K9O MgO CaO PbO Other

0010 lamp tubes 63 1 7 22

0080 lamp bulbs 73 1 17 5 29

0120 sealing 56 2 4 9 29 0.8 As

0281 housewares 73 2 1 15 .3 4 6 .2As

1720 sealing 61 17 5 1 7 9

2475 neon signs 67 10 7 12ZnO

6720 general 60 10 1 9 2 9ZnO

7250 seal beam 78 2 15 5

lamps

7740 pyrex
1M
80 2.3 13 .5 2BaO
labware .2 F/Cl

8039 ophthalmic 35 2 7 51 4Ti02

flint 1 Zr02

9741 UV transmit 67 6 24 2 .1 .5 1.1 F

OCT insulation 59 4.5 3.5 11 .5 6 16

OC-C acid-resist 65 4.5 5.5 8 .5 3 14 8Ti02

4Zr02

LOF window 73 1 13 3.8 8.6 .lFe203

PKT windshield 73 1 14 4 8.6 .lFe203

.3S02

* refers to Corning Glass Works code numbers; OC = Owens Corning; LOF=

Libby, Owens & Ford; PKT= Pilkington Bros. Ltd.
 21

All of the above formulations are approximate but illustrate the diversity of
formulations that are a v a i l a b l e . A s stated, some 700 different glass
compositions are in usage today, variations among which are designed to take
advantage of some physical property such as expansion coefficient for a
specific commercial application.

The following describes some of the various general classes of commercial

glasses usually manufactured:

1.3.3.- Classes of Silicate-Based Glasses Usually Manufactured

1. Fused Silica 2. Alkali Silicates

3. Soda-Lime Glasses 4. Aluminosilicates

5. Borosilicate Glasses 6. Lead Glasses

T h e s e w i l l be d i s c u s s e d in m o r e d e t a i l b e l o w , after w e s u r v e y the
manufacturing methods used for producing silicate-based glasses.

1.4. THE MANUFACTURE OF SILICATE-BASED GLASSES

The general steps in manufacture of glass (here we use the term "glass" to
mean silicate-based glass) are:

1.4.1.- Steps in the Manufacture of Glass

a. Weigh out components (glass sand; soda ash; limestone;

feldspar; and other components)

b. Mix with cullet (cullet is broken glass of the same composition)

c. Melt mixture

d. Allow to "fine" and become homogeneous

22

The raw materials generally used for manufacture of silicate-based glasses are
given in the following Table:

TABLE 1- 4
SOME RAW MATERIALS USED IN GLASS MANUFACTURE
Oxide Required Materials Used

S1O2 glass sand, feldspar, petalite, nephaline

syenite, lead silicate

AI2O3 feldspar, petalite, nephaline syenite,

calcined alumina, hydrated alumina

B2Q3 boric acid, borax

ΙΛ2Ο lepidolite, petalite, spodumene, L12CO3

K2O feldspars, pearl ash, potassium carbonate

Na20 soda ash, feldspars, nephaline syenite

MgO dolomite, calcined magnesite (MgCOs)

CaO limestones

PbO litharge(PbO), lead silicates

BaO barium carbonate

ZnO zinc oxide

Note that several minerals are specified. These are defined as:

Nephaline syenite = 54% albite (Na20 - A 1 2 0 3 -6S1O2 ) , 24% microcline ( K 2 0

- AI2O3 -6S1O2 ), 22% nephalite (3Na 2 0 -4 A 1 2 0 3 - K 2 0 -9Si0 2 ).

Feldspar consists of three main types: microcline , albite and anorthite (CaO-
AI2O3 - 2S1O2 ). The effects of addition of feldspar to commercial glass batches
 23

may be summarized as follows: 1. increased resistance to scratching; 2.

increased resistance to breaking during bending; 3. increased resistance to
breakage due to impact; 4. improved thermal endurance of the glass; 5.
improved chemical durability; 6. decreased tendency of glass to devrtrify.

Petalite = ( L i 2 0 - A 1 2 0 3 -8Si0 2 )

Lead silicate = PbO- S i 0 2

Borax = ( N a 2 0 - 2 B 2 0 3 - 10H 2O)

Lepidolite = (LiF-KF - A 1 2 0 3 - 3 S i 0 2 )

Spodumene = ( L i 2 0 - A 1 2 0 3 - 4 S i 0 2 )

Dolomite = CaMg(C0 3 )

Note that most of these materials are complex silicates, which break down
under heating to form to form specific oxides required in a given formulation.

T h e m o s t i m p o r t a n t classes of g l a s s e s c u r r e n t l y manufactured are

summarized in the following:

1.4.2.- Types of Glasses Manufactured

Silica-glass: This is the familiar "quartz" glass (or "fused- silica" glass)
which is used extensively in microelectronics as a substrate for IC's. It
has excellent ultraviolet transmission capabilities to about 1850 Â, a low
coefficient of expansion and is highly resistant to chemical attack by
water or other chemicals. It also finds use in optics and as crucibles for
solid state work.

Alkali Silicate Glasses: Pure silica glass has a v e r y high melting

temperature. The addition of alkalis reduces the high viscosity inherent
in pure-silica glasses and brings glass manufacture into the range of
industrially accessible temperatures and usable refractories. It does so
by the formation the breaking of Si-O bonds, and the formation of Si-O-
24

Na linkages. The limit of glass formation is about 1.0 M2O : 1.0 S1O2 .
Alkali silicate glasses are usually soluble and sodium silicate glasses form
the basis of the soluble silicate industry. They have usage as adhesives,
cleaners, cements, defloculents and to some extent as p r o t e c t i v e
coatings.

Soda-Lime Glasses: Soda-lime glass is by far the most often used glass
composition because it is cheap and easy to manufacture. If soda, i.e.-
Na2<3, is added, it "breaks" Si - Ο bonds and "softens" the glass. When
"lime" is also added, i.e.- CaCC>3 , it "stabilizes" the glass with respect to
cost, chemical durability, and ease of manufacture. The optimum glass
usually consists of: 72% silica, 15% soda, about 10% lime and magnesia,
2% alumina, and 1% miscellaneous oxides. This composition comprises
by far the most numerous "melts" made in the Glass Industry. Soda -
lime glass is used for windows, containers and lamp envelopes. It
accounts for about 90% of all tonnage manufactured.

Borosilicate Glasses: Boric o x i d e e n t e r s the s i l i c a n e t w o r k as a

s t r u c t u r a l m o d i f i e r w h i l e l o w e r i n g the v i s c o s i t y of the m e l t . .
Consequently, it is r e g a r d e d as a flux. T h e result is the t y p i c a l
borosilicate glass, which is represented by "Pyrex™", as originally
developed by Corning Glass W o r k s . Pyrex glass is used to make
laboratory glassware because of its excellent thermal resistance, a
property not found in other types of glasses. However, too much boric
oxide causes either a phase separation to occur, or the chemical
durability to suffer.

Aluminosilicate Glasses: When alumina is added to a soda-lime melt, it

assumes a tetrahedral coordination, healing the Si-O bonds broken by
the addition of alkali. It also improves the chemical durability and
resistance to devitrification. However, a solubility limit exists for the
most part. Currently, it has proved useful in strengthening glass
compositions, typically by an ion-exchange process. Such glasses show
much higher strengths than comparable low-alumina containing glasses.
Alkaline earth aluminosilicate glasses have high temperature resistance
and good dielectric properties, and are used for cook-ware, fiber glass
and glass-ceramics.
 25

Lead Glasses: Lead oxide is a good flux for glass melts and yields glasses
with high refractive indices, high density, and a low softening point.
Commercially, The addition of PbO produces a "softening" effect and
yields glasses with high refractive index and density. Lead glasses are
used for radiation-shielding windows, optical glasses, lead-crystal art
ware, and low-melting solder glasses.

Colored Glasses: Certain minor constituents are added to glass melts for
the purpose of adding color, or other spectral properties, to the
finished product. These additives include:

1. Red Glasses - colloidal Au or Cu

2. Pink Glasses - manganese plus cerium oxide

3. Orange Glasses - cadmium sulfide

4. Amber Glasses - iron oxides

5. Yellow Glasses - uranium oxides

6. Green Glasses - chromium oxides, ferric oxide, copper oxides

7. Blue Glasses - cobalt oxides, copper oxide

8. Violet Glasses - nickel oxides

9. Gray Glasses - nickel plus cobalt oxides

10. Black Glasses - manganese plus chromium oxides

11. Ultraviolet Absorption Glasses - cerium oxide, titanium oxide,

vanadium oxides

12. Infrared Absorption Glasses - iron or copper oxides

26

13. Specialty Uses :

Decolorization of Glasses - add manganous oxide

Opacified Glasses - add calcium fluoride, zinc sulfide

3+ 2+ 3 + +
Fluorescent Glasses - add L n ions, M n ,U , Cu~ , T l ,
2+ 2 +
S n , or P b .

Radiation Protection - add cerium oxide

14. Fining Uses - add sulfate, fluoride, chloride, bromide, iodide,

antimony oxide, or arsenic oxides.

Note that nearly any color can be achieved in a finished glass article. Some
colors, particularly the yellow-green color obtained from iron oxides (present
as an impurity in the minerals used to make the glass melt), are deliberately
masked by the addition of small amounts of MnO or a selenide which
introduces a balancing absorption in the blue and the green portion of the
spectrum.

The following diagram, given as Figure 1.4.3. on the next page, shows steps
required to manufacture glass. These include:

a. Batch Mixing

b. Melting

c. Stirring, fining and conditioning

d. Forming

e. Annealing and cooling

The most important step is that of melting and fining, followed by that of
forming and then annealing. It should be clear that silicate glass manufacture
has evolved from the use of sand, with specifically chosen oxides added to
 27

Figure 1.4.3.-

Steps in the Manufacture of Glass

Glass Sand Soda A s h Limestone Feldspar-to

Si02^99& N a 2C 0 3 to yield C a O yield alumina
crus h e d _ t o y id e l + s om e M gO silica, N a 2 0
washed & U&2O -pulverized andK20
seived -pulverized

Batch Mixing < i .

FORMING - hot viscous Cullet ( broken glass)

^ _ I glass s hap e d by pressin g - sa me com posi tion
or blowing
1 r 1
I i
I Finishing ] <
^ * '

> Α π neal i ng a nd Cooli ng

> ["Finishing |

Inspection and Shipping i '

achieve certain effects in the final glass composition. These oxides are
commonly added as complex silicates, which break down during melting to
provide the final composition of the glass.

Because such silicates have a rather high melting point, it is customary to add
"cullet", which is broken glass, to the front-end of the furnace. The glass has a
28

much lower softening and melting point, and serves to form a liquid phase
which causes the silicate additives and sand to melt faster, with less heat
required.

Glass furnaces are charged automatically or manually, and continuously or

intermittently. Most glass is melted in a continuous tank, whose holding
capacity may be as high as 1400 tons. Outputs of 300 tons per day are
common. The glass-melting furnace itself is lined with a refractory material
such as mullite, which is an aluminum silicate. The refractory must be capable
of withstanding dissolution in molten glass. Most materials are soluble
therein, and the part which dissolves must not be deleterious to the finished
glass batch. Fortunately, once the surface of the refractory is exposed to
molten glass, the rate of dissolution decreases radically. Natural gas and oil
are the primary fuels used. However, since the molten glass batch is usually
electrically conductive, it can be heated by the passage of an electric current.
Electric heating is used in Europe, and in the U.S., electric heating is used as
a booster in some cases.

Once a melt is obtained, the fining operation proceeds. Fining is the process
where bubbles, obtained from melting of the raw materials, are caused to
dissipate. In some cases, special compounds are added, while in other cases,
air or nitrogen is blown in to cause the small bubbles to subside. If a
continuous furnace is used, fining continues as the glass batch proceeds down
the furnace. Figure 1.4.4., given on the next page, shows a top view of a
typical continuous furnace used for processing glass.

In this case, the furnace is cross-fired, i.e.- heated from both sides, and is
constructed with seven (7) regenerative heat exchangers to recover heat from
the flue gas before it passes to the stack. They are usually operated in pairs,
one chamber being heated by passage of flue gas from combustion while the
other is being preheated.
 29

Figure 1.4.4.-

A TYPICAL CONTINUOUS GLASS MELTING FURNACE

Stack Top View of Continuous

Regenerative, Cross-fired,
Glass Melting Furnace

I I I I I I I I I I I I
Regenerative Heat Exchangers!
X X I , 1 I.J TjJLJULi

Each regenerative heat exchanger is connected to the other. In a typical

operation, the flow of air, combustion gases and flue gases is reversed about
every 30 minutes. The melting and mixing tank is the one usually heated,
while the fining and holding tanks are usually heated by convection. A cross-
section of a typical regenerative heat exchanger is given as follows:

Figure 1.4.5.-

In Figure 1.4.6., given on the next page, is shown a cutaway of the same
continuous furnace which shows the total lengthwise cross-section:
30

Figure 1.4.6.

Stack Cutaway View of Glass Melting Furnace Showing

Heating Ports, Melting Tank & Fining Tank
Raw Material
Loading Area

Flue Outlets J NT jl ·
ËsA ESA ESA BSk BSA *S& ßSnk rHrl |ι»71 l^rl WL
WH I S - m M J a J B t FBI I tBd to! Β CV

Here, we have shown the glass tank divided into two parts by a throat. By
common consent, the area of molten glass is arbitrarily divided into the
melting zone, the fining zone and the working zone. No actual demarcation
exists in these areas, except as temperature differences exist. In the melting
zone, the lumps of batch float in a mass of foam produced by their own
melting and the mass of boiling glass about and below them. The foaming is
m o s t v i o l e n t a b o u t h a l f w a y t h r o u g h the m e l t i n g z o n e , w h e r e the
temperatures are highest. Further on, the temperature is maintained at a
level high enough to reduce the viscosity of the molten mass so that foam
bubbles burst easily and rapidly. Thus, a foam-line is established beyond which
no foam is apparent. Here begins the fining zone where the glass is quiescent
and any small bubbles present have an opportunity to rise to the surface and
burst. Since no heat is introduced into the furnace within the fining zone,
radiation is used to maintain the temperature during fining. It is here that
"fining agents" are introduced to aid in removing small bubbles and other
inhomogeneities such as unmelted particles. Typical fining agents used are:

1.4.7.- TYPICAL FINING AGENTS USED IN THE GLASS INDUSTRY

1. Agents which give off oxygen gas

nitrates peroxides chlorates Perchlorates

eerie oxide manganese dioxide A s 20 3 & S b 20 3

 31

2. Agents which give off SO9 gas

sulfates sulfites

3. Agents which depend upon evaporation from melt

Halides (NaCl, KCl, etc.) ammonium salts (NH4 CI , etc.)

These agents are added to the melt at the fining section of the tank, usually
on a continuous basis. Fining agents introduce gas bubbles which cause
existing bubbles to grow larger in the melt, thereby increasing their rate of
destruction. Artificial bubbling is also used to increase fining rates during the
final stages of the melt processing before usage.

Gas evolution during melting and convection- currents resulting from

differences in t e m p e r a t u r e w i t h i n the t a n k are t r e m e n d o u s aids to
homogenization of the melt. But, uncontrolled lengthwise waves can sweep
gaseous inclusions, foam and scum, and "stones" (unmelted particles) into the
area of glass removal from the furnace. We have shown a throat within the
furnace shown in 1.4.5. whose function is to prevent this.

1.5. FORMING AND FINISHING OPERATIONS

Glass is used to make over 10,000 different articles. Many, such as light bulbs,
are made on high speed forming lines, where the individual articles move at
40 miles per hour. Glass may be rolled between heavy iron rollers to form
patterned glass. Glass may also be cast, drawn or pressed.

Forming operations include:

1.5.1.- Pressing: an operation where glassware is formed by pressing hot

glass between a mold and a plunger.

This is illustrated in the diagram given on the next page, viz.-

32

Manufacture of Pressed Glassware

Empty Mold Loaded Mold (hot glass) Glass Pressed Finished

Piece

Care must be taken not to introduce undue strain in the finished piece since
the amount of residual strain is usually proportional to its size and thickness.
Most finished pieces are small in size. Following the pressing operation, the
glass piece must then be annealed thoroughly. Other variations include the
use of a split mold and a so-called "font-mold" for making solid pieces of glass.
Glass articles made by this method are usually those of the "cut-glass" variety,
where an artisan uses a grinding wheel to cut designs into the surface of the
glass. Lead - crystal is made by this method.

1.5.2.- Casting: hot glass is poured into a mold and allowed to cool. Then it is
annealed (sometimes in place). It is this method which is used to make large
solid articles of glass. Television tube blanks are made by a centrifugal casting
method where the hot glass is cast within the hollow mold, which is then
spun to make the hot malleable glass conform to its outer surface. Such an
apparatus is given in the following diagram, given on the next page.

In this case, hot glass is introduced into a spinning mold (Step 1). The
centrifugal force causes the molten glass to climb up the sides of the mold,
while a plunger pushes the glass-stop to the bottom of the spinning mold,
thereby forming the neck of the television tube blank (Step 2).
 33

Centrifugal Casting of Television Tube Blanks

1.5.3.- Rolling: flat glass of all kinds is made by rolling. In a batch process,
glass is poured from a pot onto a large iron table and rolled to desired
thickness by a heavy iron roller. In another process, the hot glass is passed
through a set of heavy, water-cooled, rollers to form a continuous sheet of
predetermined dimensions. This sheet is then cut to desired size for usage
thereof. Prior to the development of "float-glass", these sheets had to be
ground and polished, if they were to be used as window panes, and/or plate
glass. Patterned and wired glass is also formed by this method.

1.5.4.- Blowing: this operation refers to the use of gas pressure (usually air)
to form a shape in a mold. The glass container industry is based on this
operation. A typical operation to form a bottle includes the following steps, as
shown on the next page.
34

Manufacture of Bottles by a Blowing Process

Blob of Glass Counter Blow to

Introduced Preform Glass

Reheat Preform in Final Blow- Bottle

Bottle Mold Formed in Mold

Note that the preform is reheated for the final blow. When a hot "blob" of glass
is introduced into the mold, it settles to the bottom under air pressure.
Additional air pressure introduced from the bottom of the mold causes the
hot glass to conform to the outline of the mold, as shown in the third part of
the diagram. The "preform", as it is called, is then introduced into a bottle
mold, reheated and then caused to conform to the desired bottle shape by a
"final blow". The final step is annealing (see below). Millions of bottles and
containers are made by this method every year.
 35

When glassware made by a blowing process exits the molding stages, they
enter a "lehr" where they are annealed. Because the surface of the glass
containers is not completely smooth, they tend to rub against each other and
microscopic cracks are induced in the glass-surfaces. Breakage can be severe.
In the past, when glass containers were heavy and line speeds slow, there was
little need to prevent such microcracks. However, present day demands for
lightweight glass containers have mandated the use of "hot-end" coatings
immediately after forming.

The coating materials most often used are tin-oxide based, derived from tin
tetrachloride and/or butyltin trichloride. Application protects the pristine
glass surface from reaction with moisture, and controls microcracking of the
glass surface as well. Another benefit is reduction of breakage during the
lehring operation.

In m o s t modern glass container manufacturing facilities, bottles and

containers have become increasingly lighter in weight and speeds in the lehr
and customer's filling lines have accelerated enormously. This has dictated
the use of a "cold-end" coating as well. Basically, the reason glass is cold-end
coated is to further protect and lubricate the glassware so that it can be
moved smoothly with high speed handling equipment, without damage to the
glass surfaces. Thus, a massed group of bottles on the lehr is moved into
single lines for inspection and then on through packaging. W h e n the
containers reach the customers plant, the glass is unpacked, cleaned, filled,
labeled and repacked. Glass to glass contact occurs during all of these
processes and efforts to minimize surface damage is mandatory. Although
glass is theoretically a very strong material, resistance to scratching must be
achieved so as to keep the inherent strength of the containers high. Cold-end
coatings most often used include oleic acid and water-based polyethylene
emulsions. The former is easily removed by washing, whereas the latter is
not.

1.5.5.- Float Glass: This process was originally developed by Pilkington

Bros., Ltd. in England in the late 1 9 5 0 s . Essentially, it involves casting
molten glass upon a molten bed of tin-metal to form a sheet of glass in a
continuous manner. The atmosphere in the float zone is inert to prevent the
tin from oxidizing. This process totally eliminates any grinding and polishing
36

operations, and is used worldwide to manufacture plate glass, including

mirrors. The following diagram illustrates the process used:

T h e F l o a t Glass P r o c e s s t o M a n u f a c t u r e P l a t e Glass

Glass Heat , F i r e L Cooling

Controlled
Stack
Sheet I Gas Ink Zone P o l i s h ! Zone Atmosphere

[Loading 1

flve L

IGlass Furnace) Float Bath Annealing

In this case, the output of the glass furnace is fed as a flat sheet into the float-
bath, where it is reheated, fire-polished and and cooled, all in an inert
atmosphere to prevent the melted tin from oxidizing. Additionally, various
coatings can be applied to the plate glass.

1.5.6.- Drawing: tubing of all sorts, cane or glass-rod, and sheet are made in a
continuous operation. The operation can be vertical or horizontal. The three
major methods of drawing tubing use the Danner, Velio or Updraw machines.
All are basically the same: hot glass flows around a ceramic cone or mandrel
and is carried along asbestos-covered rollers. Air blowing through the
mandrel maintains the glass as a continuous tube. This is then cut into
manageable lengths before usage.

1.5.7.- Fiber Forming: this process makes fibers for textiles, and electrical,
thermal, and acoustical insulation. In one process, molten glass is forced
under pressure through platinum orifices, with steam jets set at right angles
to the flow, in order to form fibrous-glass wool. In another, glass is remelted,
and forced through a mandrel with small holes, to form continuous filaments
which are wound to form a glass fiber for weaving thereof. A more specific
description of this process, and the products made, will be given below.
 37

1.6. THERMAL PROCESSING AND PROPERTIES OF SILICATE GLASSES

The glassy state possesses several interesting properties in regard to its

manipulation by thermal means. Many of these properties are inherent, while
others can be induced. The following describes some of these induced
properties, including that of glass- ceramics. In that case, controlled
recrystallization is used to make various glass-ceramics which have qualities
not available in "ordinary" glass.

1.6.1.- Glass Points and Devitrification

All glasses are unstable to one or more crystalline compounds and will
recrystallize, given the proper thermal treatment. This process is called
devitrification. Such behavior is familiar to most glass manufacturers and is
usually prevented by adjustment of the starting glass- formulation.

When glass is heated from room temperature, it undergoes several thermal

changes. The easiest way to show this is to examine the DTA-thermogram of a
typical silicate glass, as shown in the following, viz-

A Typical D i f f e r e n t i a l Thermal Analysis of S i l i c a t e Glass

fSjTij = Devitrification
Point

Programmed
Τ = Glass Transition
9 Temperature
Point

Τ = Softening P o i n t L ^ ^ r
L_sß j>r
Working = Melting
ο
TD Range Point

Time — * »

In the above diagram, the abscissa is time where the temperature is

programmed in a linear manner, i.e.- 3T/3t = a constant. When the glass is
38

heated (usually as a powder for differential thermal analysis, i.e.- DTA), the
first thermal anomaly reached is T g , the glass transition point. Since DTA is
essentially a plot of differences in heat capacity, this first point appears as a
discontinuity. At the bottom of the discontinuity is the softening point, T s p .
T h e n e x t p o i n t is an exothermic peak w h i c h is the d e v i t r i f i c a t i o n
temperature, i.e.- the glass undergoes a transformation from the amorphous
to the crystalline state with evolution of heat. When these crystals are further
heated, they melt atTiyr. These thermal changes are reversible, and the glassy
state again results when the melt is cooled properly. The temperatures
between the bottom of the discontinuity, i.e.- T s p and the point of incipient
devitrification is called the "working range" of the glass. There, the viscosity
of the glass is low enough so that forming operations can be carried out.

1.6.2.- Annealing

Although we have mentioned this operation, we have not described it in

detail. When a glass article is made, either by casting, pressing or blowing,
and allowed to cool, it will develop internal strain within the glass walls. This
strain results because it is the outer surface of the glass where heat escapes,
while the inner part remains hotter. Thus, a heat differential is set up where
the outer part of the glass article cools and contracts faster th^n the inner
part, as shown in the following illustration:

DEPICTION OF WHY STRAIN RESULTS IN A HOT GLASS LAVER

NOTE: THE OUTSIDE •COMPRESSION^. . . ^

OF THE GLASS LAYER
COOLS FASTER THAN
INSIDE. THEREFORE3 COOLER
THE INSIDE, BEING
HOT, DEFORMS UNDER
TENSION, FROM THE
COMPRESSIVE FORCES
AT THE OUTSIDE AND
AN INTERNAL STRAIN
REMAINS VHEN THE
GLASS LAYER IS
COMPLETELY COOL.
rnMPPFQÇiriN"

Internal strain is universal in glass because of the amorphous structure put

into place as the glass cools below its glass- transition- point. If the strain is
not relieved, the solid glass will, at some point, explode violently. It is
 39

therefore necessary to "anneal" the glass. There is a temperature at which

strain- introduction is maximum, called the "strain-point". To anneal the
glass, one heats it slightly above the strain-point temperature, at the so-called
"annealing" (strain-relief) temperature. When both external and internal
temperatures of the glass become even, the external temperature is reduced
slowly through the strain-point, so that a "strain-free" glass article results.

Examples of these so-called "glass-points" are listed for several varieties of

silicate glasses, as given in the following table:

TABLE 1-5
Glass Points for Some Silicate Glasses in °C.
(refer to Table 1-3 for compositions)
Glass Strain Point Annealing Point Softening Point Working Point

0010 392 492 626 983

0080 473 514 696 1005

0120 395 435 630 985

0281 491 532 714 1024

1720 667 712 915 1202

6720 505 540 780 1023

7250 436 480 712 1128

7740 510 560 821 1252

8039 440 475 617 866

9741 408 450 705 1161

LOF 486 525 700 985

Ρ KT 522 551 734 1047

40

The "glass-numbers" refer to the varieties already described in Table 1- 3. To

properly anneal a glass article, it is necessary to hold the article at the
annealing temperature for the time required until the internal temperature
matches the external temperature, which in turn depends upon the thickness
of the glass article.. Cooling slowly (usually no more than 1-2 °C/minute)
through the strain point then produces a strain-free glass object. Note that
neither the glass transition point, nor the softening point, is ever exceeded.
Details on how the temperature cycle is calculated so as to produce strain-
free glass will be given in the following Chapter.

1.6.3.- Tempering

Tempering is achieved by by rapid cooling of the surface of the glass from

temperatures near to the softening point. The result is a high compressive
stress on the surface, as illustrated in the following diagram given on the next
page.

Several advantages accrue, notably that of strength which can range up to

50,000 psi. Another advantage is that the breaking pattern can be controlled.
The tempering process requires meticulous control of time and temperature
of heating. The glass is heated in electric or gas-fired ovens to a temperature
near to its softening point. It is then quickly removed and quenched by air-
blasts or by immersion in a bath of oil, or in fused inorganic salt baths.
Properly performed, there is no p e r c e p t i b l e difference in a p p e a r a n c e
between annealed and tempered glass plate. However, It is quite difficult to
temper flat glass plate less than about 1/8 inch thick.

Even though tempered glass is very strong, it is possible to break it by

excessive stress. If the stress system is unbalanced, as by disruption of the
surface even at a single, isolated point, or flaw, the whole piece ruptures into
small, interlocking, blunt-edged fragments. These fragments become smaller,
on the average, the greater the strain energy released. That is, the greater the
original degree of temper, the greater is the energy released. An effort is
often made to obtain maximum strength without exceeding an amount of
tension called the "dicing limit". At the dicing limit, the glass, when broken,
breaks, or "dices" into pieces roughly 1/4 inch square, if the tempered glass
plate has been properly manufactured.
 41

THE TEMPERING PROCESS FOR PLATE GLASS

TEMPERED GLASS PLATE

The largest usage for tempered glass plates is that of windshields for
automobiles. In this case, the thermal tempering is accomplished in a hot
bath, composed of either oils, fused salts or blasts of hot air.

Still another method for preparing tempered glass is that of "chemical-

strengthening". In this method, one places the glass article to be tempered in
+
a molten salt bath. For example, the ion, K+, will exchange with N a ion in a
sodium aluminosilicate glass at 500 °C. to produce strengths in excess of
+
100,000 psi. Sometimes the lithium ion, L i , is used as well. The volume
increase (or decrease) required to accommodate larger or smaller ions in the
surface is restrained by the interior glass structure. This exchange, which
affects only a thin surface layer, permits compressive strengths much higher
than previously possible. Removal of the exchanged ions on the surface can be
42

effected by the use of sulfur dioxide, or by heating. This results in a low-

expansion glass surface.

2
Considering the low density of glass (about 2.50 gm./crn ) these strengths are
especially advantageous, since chemically-tempered glass compares quite
close to those of commercially available materials. The following Table is a
summary of the measured strength to weight ratio of several types of glass:

TABLE 1-6
Strength to Weight Ratios for Glass and Other Materials
2 3
in lb / i n per l b / i n
Glass forms Other-Materials

4 4
Abraded glass 2 χ 10 Tin metal 1 χ 10

4
Pressed glass 5 x 10

4 5
Blown glass 9 χ 10 Wrought iron 1.5 χ 1 0

5
Tempered glass 3 x 10

Glass ceramics 4 x 10
5

5 5
Ion-exchanged glass 9 χ 10 Stainless steel 8 χ 10

6 6
Glass fibers in plastics 4 χ 10 Carbon steel wire 1.5 χ 1 0

6 7
Drawn glass fibers 6 χ 10 SiC whiskers 2 χ 10

It is easily seen that glass forms produced by the various methods compare
very favorably with other high strength materials.

1.6.4.- Thermal Shock Resistance of Glass

The application of heat to glass results in a stress caused by expansion of the

glass at the point of heating. Actually, this is a type of birefringence which is
reversible. There are upper limits of temperature usage for glass articles,
 43

determined by stresses set up by temperature differences across the glass. If

a glass plate is completely restrained, the tensile strength on the cool side is
given by:

σ = Ε α Δ Τ/2(1-ν)

where ΔΤ is the temperature difference across the glass, α is the expansion

coefficient, Ε is Y o u n g s modulus, and ν is Poissons ratio. The maximum
thermal shock resistance of glass-ceramics is about 480 ° C , which is much
higher than most glasses which range from about 100 - 300 ° C

1.6.5.- Glass Ceramics and Nucleation Processes

Glass- ceramics are made by heterogeneous nucleation and crystal growth

processes in which more than 50% of the glass is converted to a fine-grained
crystal. This process was discovered by Stookey (14) and is made possible by
the addition of a nucleating agent such as NaF. Subsequent thermal
processing, also called "ceramming", serves to cause the major crystal phases
of the glass to grow. Other nucleating agents that have been used include:
metal colloids, T i 0 2 and titanates, chromic oxide, P2O5 , Z r 0 2 , ZnS , plus
various fluorides. Obtaining such fine-grained fully dense ceramic materials is
based on the fact that one can cause the glass to fully nucleate and
recrystallize. Such a process allows the fabrication of a new class of materials
having unique and useful properties not found in glass alone.

Glass - ceramics are used as:

- Cook-ware where the utensil can be directly heated on a hot

surface without breaking due to thermal stress.

- Low expansion printed- circuit boards

- Stove-tops for cooking where the heating elements are placed
beneath the glass-ceramic surface.

- Table ware which will not break when hot food is placed in or on
the dishes.
44

Sales of smooth cook-top stoves has become big business where fully 48% of
stove- appliances sold in 1990 were of this variety, that is- incorporating a
glass-ceramic surface. Such electric or gas appliances use a resistive coil
heating element, or a radiant element such as a halogen lamp of the same
wattage, on the back side of the surface. The overall cleanability and visual
impact of the smooth- top stove has made it a leading design for the 1990s,
and was made possible by the advent of glass-ceramics.

1.6.6.- Opacification

Still another thermal nucleation process is used in the Glass Industry. Opal
glasses are translucent, that is - they are partially transparent due to multiple
internal scattering processes. Opacification is obtained by selectively
precipitating crystals, having a refractive index differing from that of the glass
matrix, within the glass structure. The amount of opacification is determined
by the difference in refractive index between the glass and the crystals, as
well as the number and size of the crystals. This allows the manufacturer to
control opacification by reheating techniques. Opal dinnerware is obtained by
the addition of zinc sulfide or calcium fluoride to the melt to form a second
phase of small crystals.

1.6.7.- Colloidal Colors

The development of colloidal colors in glass also depends upon controlled

nucleation and growth of small metallic crystals in glass. Gold, silver and
copper are most often used. Such crystals have selective absorption and
reflection properties, and impart a unique "color" to the glass article. The
metal to be precipitated is dissolved in the glass melt in conjunction with a
reducing agent such as SnO or Sb203. After melting, the glass is rapidly
cooled in a temperature region where the nucleation rate is high. It is then
reheated to a temperature where crystal growth of the metal crystals is high.
This technique allows the size and number of crystals, and therefore the
color, to be controlled. Gold imparts a flaming red color and produces a deep
red crystalware.
 45

1.6.8.- Photosensitive Glasses

The reduction step required in the formation of colloidal colors is the basis
3+
for preparation of photosensitive glasses. If an optical sensitizer such as C e
is substituted for SnO or Sb2Û3 in the glass formulation, colorless glasses can
be made that are photosensitive. Exposure to ultraviolet radiation results in
3+ 3+
absorption by the C e ion and the subsequent reduction of A u ions to form
nuclei. These nuclei grow to colloidal size upon reheating. If a photographic
negative is used during the exposure, a three-dimensional image can be
formed in the glass. In a like manner, metallic nuclei can be used to induce
heterogeneous nucleation of a c r y s t a l l i n e p h a s e in the g l a s s itself.
Photosensitive opal glass products, called F O T O - L I T E ™ , are made by
precipitation of NaF on Ag.

The use of lithium silicate glass makes possible a type of glass which can be
"chemical machined". In this process, a heterogeneous crystalline phase,
Li2Si03, can be formed, which is much more soluble in 10% HF. Once the
glass is exposed to ultraviolet and the crystal phase dissolved, the glass is
again exposed to ultraviolet and then reheated to produce a hard, nonporous,
strong ceramic. T h i s is called FOTOCERAM™. Both are trademarks of
Corning Glass Works, Elmira NY.

Another photosensitive glass was reported in 1964 (15). In this case, the
clear glass changes to a uniform gray color when exposed to either ultraviolet
or visible light. Such glass has found use as "sunglasses" and are sold widely.
Crystallites of silver halide, about 50 Â in diameter and separated by about 500
Â, are the active centers. The analogy to a photographic film is quite clear,
except that the nucleated centers are not allowed to grow in size (as they do
when the photographic film is "developed"), but remain the same size while
entrapped w i t h i n the glassy s t r u c t u r e . T h u s , the p h o t o c h r o m i s m is
reversible, that is- the glass darkens and fades in response to the amount of
light present.

The following diagrams illustrate both the rate at which such photochromism
occurs, and the wavelengths affected:
46

Darkening and Fading Rate for a Typical Photochromic Glass

0.6 ι 1

30 60 120
Time in Seconds

Spectral Transmission, for a Typical Photochromic Glass

ö
ο
••-I
w
w
•i-l Absortion of Photosensitive
a Glass when Exposed

Visible Range

J-
5000 10,000 15,000 20,000
Wavelength in Â

Note that it takes about 30 seconds for the darkening effect to become
maximum. As long as the light intensity remains the same, the absorption,
due to photochromic centers, remains. When the light intensity dims, the
absorption decreases. The transmission curve given above shows the
 47

wavelengths affected. These include most of the visible range, making the
glass almost ideal for use as reversible sunglasses.

It is also possible to use other halides such as gold, but they have not found
such widespread usage as the silver halide used for the sunglasses trade.

The above d i s c u s s i o n has c e n t e r e d on d e v i t r i f i c a t i o n and properties

associated with control of nucleation. It is clear that a large variety of glasses
can be made having specific properties much superior to those of non-
nucleated glass. Let us now examine some of the inherent properties of glass.

1.7. INHERENT PROPERTIES OF GLASS

These properties are those which appear as a result of the indigenous

structure of glass itself, as a usable medium for forming articles of commerce.
Let us first examine the strength of glass.

1.7.1. - Strength of Glass

The inherent strength of pristine glass has been estimated to be about

4,000,000 psi. The practical strength of glass is, however, much lower. Glass
itself can be considered to be a perfectly elastic material. No matter how
much stress is applied, so long as the stress is not sufficient to cause fracture,
the glass will return to its original shape when the stress is removed.
Nevertheless, nearly all glass products, including plate glass, break under
stresses less than 1% of the inherent strength. This tremendous discrepancy
is believed to be due to the presence on the glass surface of a multitude of
imperfections ranging from submicroscopic fissures to scuffs and scratches
v i s i b l e to the naked e y e . Some are introduced due to manufacturing
processes, while others may be due to subsequent handling. Several separate
systems of flaws may be present in a single piece of glass, the total number
being as many as several thousand per square inch, and most of them will be
invisible. Since flaws are randomly distributed, it is a matter of chance
whether a major or minor flaw happens to fall within a zone of maximum
stress, and the observed breaking stress will vary accordingly. In any case,
each flaw acts as a stress concentrator and a deep, narrow crack with
48

sharp apices may cause a local increase of over 1000:1. Thus, the actual stress
at the apex of such a flaw may be 1,000,000 psi or more.

Glass always fails in tension or from a tension component set up by the

applied load. The failure nearly always occurs at a flaw. For example, a normal
glass plate may have an inherent breaking strength of 15,000 psi. Abrading
the surface lowers this to about 8000 psi, while tempering raises this value to
a practical limit of about 35,000 psi. It has been said that "we measure not
the strength of glass, but the weakness of the surface". In the case of
tempered glass plate, the increased strength most certainly is due in part to
the initial softening of the surface wherein part of, if not most of, the surface
flaws are eliminated before the quenching operation is begun. Yet, if abrasion
(or scratches) of the tempered glass plate occurs subsequently, the glass plate
can be subject to failure. Since glass is brittle, completely lacking the
plasticity of metals, it is unable to relieve high local stress except by forming
new surfaces. This leads to propagation of the original flaw or fault, and
o c c u r s a l m o s t i n s t a n t a n e o u s l y . It is t h e r e f o r e o b v i o u s t h a t s t r e s s
concentrators- holes, grooves, notches and the like should be absent
whenever possible, particularly in the case of tempered glass plates.

One other factor needs to be mentioned, that of the effect of moisture. If the
glass has been subjected to high moisture over a long period of time, even if it
has been in storage, its surface will develop submicroscopic fissures which
may become the cause of failure.

1.7.2.- Chemical Durability

All silicate glasses are subject to hydrolysis losses in water, more so in boiling
water. They are also subject to loss in hot (100 °C.) hydrochloric acid, and
even greater losses in hot (100 °C.) NaOH. These losses are related to the
diffusivity of the cations from the interior of the amorphous structure, vis:

2 + 3 +
Relative Ratio of Diffusivity = M+ » M > M

Water also dissolves the surface layer of silicate-tetrahedra slowly, Hot HCl
less so and hot NaOH even more. These losses have been tabulated and
compared. Although the losses have usually been rated as A, B, C, etc. or
 49

1,2,3, we will find it more useful to use the actual weight losses experienced
in terms of:

2 2
mg/cm /hr. or g m / c m / hr

2
where c m is the actual exposed surface of the glass to water or other liquids.
In general, silicate glasses are resistant to corrosion by oxidizing and reducing
liquids and chemicals, including hydrocarbons, amines, mercaptans, oils and
fats, and metal salt solutions. A certain amount of dissolution occurs in strong
acids, strong bases and boiling water. HF is particularly corrosive because it
forms fluorosilicic acid, which is soluble in water. However, silicate glasses
can be made w h i c h are water- soluble, acid- soluble or base- soluble,
depending upon the intended end- usage.

1.7.3. - Dielectric Strength

Glass is often used as a substrate for electronic applications because of its high
dielectric strength. If a voltage is applied to a glass form, no electrical
conduction will be noted. As the voltage is increased, a slight increase in
conduction will be observed, up to the point where the applied voltage
gradient exceeds that of the inherent dielectric strength of the glass. Since
+
conduction of electrons is carried mainly by monovalent ions such as N a
within the glassy structure, it is easy to see that dielectric strength can be
tailored to its end- usage.

Moisture and other products will deposit on the surface of glass. The result is
a decreased surface resistivity. Most glasses have a surface resistivity of about
7 16 2
10 to about 1 0 ohm/cm , depending upon the relative humidity.

1.7.4.- Hardness of Glass

The hardness of glass is intermediate between apatite (5 on the Mohs scale)

and quartz (7 on the Mohs scale). Thus, it is easily scratched by most metals
and many minerals including sapphire and diamond.
50

1.7.5.- Coefficient of Friction

The static coefficient of friction is very nearly one (1.0) on fresh formed
surfaces of glass. After several hours of exposure to air, moisture or other
contaminants, the result is a much decreased coefficient. Values as low as
0.15 are observed. It is for this reason that freshly formed glass surfaces are
routinely overcoated to prevent "microcracks and chips" from forming on the
surface of glass articles by rubbing, while the articles are being manufactured.
Materials used include tin oxide, oils and waxes and sometimes antimony
oxide.

1.7.6- Elasticity

Glass is in general almost perfectly elastic. No plastic flow has been observed
even at rather high pressures. The Young's modulus (E) for most silicate
6 2+ 2+
glasses lies between about 7 and 13 χ 1 0 psi. The addition of C a or B e
ions to a glass formulation has produced compositions with Young's modulus
of up to 20 million psi. Knowing Poisson's ratio, v, and Ε allows one to
calculate other elastic properties via the following:

Typical Values

6
Modulus of Rigidity: G = E/{2(1+ v)} 4.1 χ 10 psi

6
Bulk Modulus : Κ = E / { 3 ( 1 - 2v)} 6 χ 10 psi

7 1
Compressibility: C = 1/K 1.7 χ 1 0 psi

However, these parameters are not used very often, except for applications
where dimensional stability is a must.

1.8. SILICATE GLASS PRODUCTS CURRENTLY BEING MANUFACTURED

A number of novel uses for glass products have evolved over the past 100
years, due to new properties of glass discovered and particularly to the new
methods of manufacture that have evolved during the past few years. For
example, the so-called "sol-gel" method of preparing silicate glass involves the
 51

precipitation of a gelatinous medium, which is then carefully dried and then

consolidated to form the desired a r t i c l e of g l a s s , w i t h a much lower
expenditure of energy.

1.8.1.- Fibers of Glass

Though glassmakers have known since ancient times that small fibers of glass
could be drawn out as fine filaments, the enormous number of applications for
glass fibers was not realized until early in this century. In general, there are
four major uses for glass fibers at the moment, namely:

1. Fiberglass textiles

2. Tangled mats for thermal insulation

3. Mixtures with other structural materials to provide extra

strength

4. Use as fiber-optics.

FIBER - GLASS TEXTILES

In the early 18th century, glass makers made fine fibers by drawing a thread
from a heat- softened glass rod onto a spinning wheel. As a novelty, Carlo Revä
of Venice wove some of this "spun-glass" into a sample of cloth. In 1713, René
Reaumur submitted Revas glass cloth to the Paris Academy of Science for
inspection. An entire garment was made from woven glass and given to a
Spanish princess, but the fiber was too coarse to allow the garment to be
folded.

In 1908, G. von Pazsiczky developed a method for drawing several glass fibers
from holes in the bottom of a refractory glass-melting chamber. In 1911, E.
Pick received a patent for his technique of using centrifugal force to draw
fibers through tiny holes in a rotating drum. And, during the First World War,
Germany used similar methods to make fiber-glass as a substitute for asbestos
insulation. By 1920, W. Schuller had found a new way to make fiberglass out of
glass rods, drawing strands from a hundred such rods which had been gas or
52

electrically heated, then winding the strands onto a single drum. The strands
were small enough, and elastic enough, to allow glass cloth to be woven. In
the 1930s, the Owens-Illinois Glass Company and the Corning Glass Works
c o m b i n e d their separate fiber-glass facilities to form Owens-Corning
Fiberglass. Workers there developed mass production methods that allowed
industrialization of fiber-glass to proceed during the period of 1936-1950.

To draw out a continuous single glass filament, such as for making glass yarn
or optical fibers, molten glass is fed through precious-metal spinnerets having
very accurate holes in their ends. About 400 such spinnerets is used in any
single machine, and thin glass fibers are drawn off at a rate of several miles
per minute. One piece of feed-glass about 1.9 cm. in diameter, and a meter
long, produces about 156 kilometers of filament. After being drawn, the glass
is usually coated with organic resins to prevent further degradation while
being processed further, to form glass cloth by weaving and the products
manufactured from glass cloth, including drapes for the home. Other
products can include yarns, cords, braids, cloths and tapes. Although glass
cloth itself is colorless, the organic resins used for coating can be dyed or
pigmented. Glass fibers are very strong, but have little ductility. The thinner
the fiber, the more it can bend without breaking. Fibers made from specific
types of glass are used for particular applications. Low alkali lime-alumina
borosilicate fibers, which have excellent dielectric and weathering properties,
are used for general and electrical textiles and for reinforcing plastics,
papers, rubbers and gypsum. Inexpensive soda-lime fiberglass is used for air
and liquid filters. Soda-lime borosilicate fibers, w i t h their high acid
resistance, are used for corrosion protection and as mats for storage-battery
retainers, as well as acid filter cloths and anode b a g s . T h e e x c e l l e n t
weathering properties of lime-free soda borosilicate glass fibers makes them
ideal for high efficiency filter papers and for light weight thermal and
acoustical insulation material. Finally, high-lead silicate fiberglass, with its
special x-ray opacity, is used for surgical pad strands and for x-ray protection
aprons.

FIBERGLASS INSULATION

On the other hand, production of glass insulation demands a dense cluster of

many short, fine fibers of glass. A completely different process has come into
 53

usage. In 1929, the Hager brothers and F. Rosengarth of Britain succeeded in

developing a new process for making glass wool. In this process, a stream of
molten glass falls on a fire-clay disk with radial serrations. By rotating this
refractory disc about a vertical axis, short thin glass fibers are thrown off
tangentially, along with some large blobs of glass. Both Cie de St. Gobain of
France, and Owens-Corning improved on this process, this time replacing the
ceramic disc with a metal spinner pierced with thousands of holes. The
vertical cylinder spins, drawing out a fiber through each hole via centrifugal
force. T h e fibers are deflected downward, and further attenuated by a
"blowing" ring. The fibers are then bonded with an organic resin, collected on
a conveyor belt, and then subjected to another blower that tangles the glass
fibers into a random mat.

Fibers formed by this production method range from 6 to 12 pm. in diameter

and have an excellent degree of uniformity. The natural density of glass wool
3
made this way is about 24 k g . / m and can be used at temperatures up to about
500 °C. Each strand of fiberglass, though thinner than a human hair, is still a
glass rod, and so retains the properties of glass. In general, fiberglass is heat-
resistant and non-combustible, making it a good substitute for asbestos. It
does not rot or decay, nor does it absorb moisture. It does not conduct
electricity and is resistant to ordinary weathering. Fiberglass insulation has
e x c e l l e n t t h e r m a l p r o p e r t i e s and can be u s e d , in conjunction with
appropriate resins to construct "fiberglass" forms and shapes.

COMPOSITE FIBERGLASS MATERIALS

Until the 1940s, the combined cost of the resin and the fiberglass precluded
major industrial development of composite materials. In 1947, it was found
that the reinforcing fibers need not be so fine, and that coarser materials
could be made more cheaply. This opened up new techniques such as the
diamond-pattern mat and swirled-mat fabrics and plastics. What is done is to
place shorter glass fibers in a patterned mat which is then used to form the
final object. For example, fiberglass mats are laid down on a given form, say a
boat-hull, and alternated with coats of thermosetting resin. A great number of
manufactured items are sold today, including large tanks for underground
burial at gasoline stations. The latter have evolved because of the resistance of
54

fiberglass to corrosion by gasoline aliphatic hydrocarbon fractions, and the

ease of manufacture of such tanks.

In 1951, I.A. Rubinsky in the United states investigated the use of fiberglass
as reinforcement for concrete, but found that fiberglass would not replace
steel because of its low modulus of elasticity. However, in 1957, work in
Russia determined that if the fibers were protected against the alkali in the
cement, its use in the Construction Industry was quite feasible. The quality of
commercially available fiberglass has improved dramatically in recent years,
making fiber-glass-reinforced concrete an important building material. Such
materials are now in general usage throughout the world.

FIBER OPTICS

A new and major usage of glass fibers has come about since 1970 in the field
of information transmission. Nowadays, nearly all of the new telephone lines
being installed consist of optical-fiber cables. If one forms a glass fiber, and
surrounds it with a material having a refractive index less than that of the
glass, one obtains an optical fiber capable of conducting all of the light which
enters it.

However, there is a limit to the angle to which the light is admitted, which in
turn depends upon the refractive index of the glass. Actually, this "acceptance
angle" is the Brewster angle, Θ , beyond which point all of the incident
radiation is reflected. The so-called "acceptance angle" is actually:

θ = 1/2(180-20)

and the amount of light that will actually enter the optical fiber is:

2 2 1 2
NA = ni sin θ = [ni - γ\2 ^ ^

where NA is the numerical aperture, ni is the refractive index of the glass

fiber, and ri2 is the refractive index of the material used to coat (or "clad") the
fiber. This only occurs when: ni > ri2 .

What we are discussing is made evident by the following diagram:

 55

The Front End of a Fiber Optic Showing the Angle o f

Acceptance, θ ,and Total Internal Reflection of an Optic Ray

8c = Critical Angle

Here is shown the acceptance cone, the angle of acceptance, θ , and the total
internal reflection of an optic ray of light. Note that θ is one-half of the angle
which describes the cone. The angle of refraction in the cladding can be
determined from Snell's Law, i.e.-

ni sin θι = n 2 sin Θ2

where ni is the refractive index of the core and θι is the acceptance angle of
the core, etc. When ni is greater than n 2 , a specific angle, called the critical
angle, 0 C , can be determined, and is defined as:

6 C = arcsin ( n 2 / n i )

For example, the critical angle for light incident from a quartz medium having
a refractive index of 1.46 and reflecting from a cladding surface having a
refractive index of 1.33 is equal to 65.6 °. For incident angles less than this,
propagating light does not refract into the cladding at all, but is totally
reflected.

It is patent that as a light beam enters front end of a fiber, it propagates

through a series of total internal reflections and emerges on the far end of the
fiber. However, the reflectance at each reflection point is not exactly 100%,
but some value less than this. Thus, an internal loss in the intensity of the
light beam (or attenuation) occurs through any given optical fiber. Therefore,
a major concern is the transparency of glass for this application. For short
path lengths in glass, this poses little problem. However, for lengths greater
than about 100 meters, the attenuation becomes serious. Scientists and
56

technicians have continually improved the transparency of silicate glass over

the centuries until about 1966 when attenuation became important for optical
fibers. Today, such fibers are some 100 orders of magnitude more transparent
than they were in 1966. This is shown in the following diagram, viz-

Optical Transparency in Glass

0.1
τ τ
1.0 — ω

10 _ Ο

ω
CL

HlO-10 Β
ω
ω
_ | ίο-loo I
Optical Glass ω
CÖ
L_
I-

Phoenician

Egyptian

10' 1 ι ι
3 0 0 0 BC 1000 AD 1900 1966 1979 1986
YEAR

In 1966 when the first work was being carried out on optical fibers, because
the path lengths became lengthy, the method of measurement of attenuation
of light was specified in terms of decibels, i.e.- dB, similar to that used for
sound waves. This is actually an attenuation in steps of ten
1 2
where 1 0 = 1.0 dB, 1 0 = 2.0 dB, etc. Optical loss is then specified as
dB/km. , as shown in the diagram. It is clear that the transparency of glass
fibers has improved by at least 100 orders of magnitude since the first optical
fibers were made.
 57

The idea of transmitting light through a long narrow cylinder of glass

o r i g i n a t e d as long ago as 1854 w h e n B r i t i s h p h y s i c i s t J o h n Tyndall
demonstrated that light could be guided within a j e t of water. The first
quantitative experimental investigations of light communications were made
in 1920, but it was not until the invention of the laser that an adequate light
source became available.

Today, the light source is usually infrared light from a semi-conductor laser.
Flashing off and on millions of times per second, the laser sends bundles of
light through the fibers, similar to the O's and Γ s of computer language. The
greatest attraction of fiber optic systems is their carrying capacity, or
bandwidth, as it is sometimes called. Bandwidth is the range of frequencies
that a signal can be flashed on and off. In other words, the bandwidth is the
number of pulses, in megabits or even gigabits per second, that the optical
fiber can carry. With fiber optics, two or even three pulsed frequencies can be
sent simultaneously. This is called "multiplexing". This increased bandwidth
leads to high volume, high speed information transfer. The reason that semi-
conductor lasers are used is two-fold:

1) they can be synchronized rather easily to existing audio sources.

2) they have wavelength emissions that correlate well to the glass core
transmission.

A typical transmission of a silica fiber is shown in the following diagram, given

on the next page.

Note that the loss reaches a low value at about 0.85 μ and 1.06 μ. The former
corresponds to an output wavelength emitted by a GaAlAs laser, while the
latter is that of a YAG:Nd laser. The large absorption peak at about 0.96 μ is
-
due to O H ions contained in the glass. A comparison of the measured
transmission loss with the intrinsic scattering loss shows that the fiber adds
very little excess loss between 0.5 and 0.7 μ. This is an indication of the
purity of the glass and the perfection of the waveguide.
58

Transmission of Optical Fiber as a Function of Wavelength

CD

§ 20 —
Ο

\ \. Scattering Loss / \

X ^ > ^ ^ 0.8S μ ^ ^ " ^ /

π 1 Γ L » I L 0 6
*
0.6 0.7 0.8 0.9 1.0 1.1
Wavelength in Microns

There are three (3) types of fibers currently being used. These are:

1) Step-index multimode.

2) Graded- index multimode

3) Single- mode fibers.

These are illustrated in the following diagram, as given on the next page.

The major difference shown there is that the graded index fiber consists of a
core in which the refractive index varies in a continuous manner, the middle
of the core being of higher refractive index than the outer parts of the fiber.
The method of manufacture is essentially the same as that described above for
preparing "chemically-strengthened" tempered glass. The surface of the glass
+
fiber is immersed in a molten salt bath where a smaller ion such as L i is
 59

Step Index Multimode Graded Index Multimode I Single Mode

Refractive Index Profiles

Light Paths Through Optical Fiber

gS2
introduced, causing a change in the refractive index as well. This results in a
"graded-index" glass, where the inner part of the core has a higher index than
the outer layers. In some cases, layers of glass having differing indexes are
formed, and the whole is then softened so that the layers mix together. The
step-index multimode fiber is the same as we described above where a
cladding of lower refractive index is e m p l o y e d . A s of 1991, the most
common fiber used has been a pure silica (S1O2), doped with T1O2, P2O5 , or
AI2O3 to increase the refractive index of the silica core.

As we stated, the outer cladding of a multi-mode fiber can be of a glass having

an expansion coefficient close to that of the core. Alternatively, a coating can
be placed on the core as it is being formed. The latter process has several
advantages, the main one b e i n g t h a t p r i s t i n e surface of the fiber is
immediately covered so that chemical interactions with impurities and
moisture of the air are circumvented.

An apparatus designed to manufacture such glass optical fibers is shown in the

following diagram, as given on the next page. Note that this apparatus can be
used for both step-graded and graded index fibers as well.
60

,
ι- . . - ,. = f l _ Preform Feed
Drawing and Coating y — M .
f , m
Apparatus for 1
Manufacture of _
Optical Fibers s
Induction
I — 1 | ^ Furnace
H

:;:;:;:Fiber Diar^^ter Monitor!

^ii^|Bjij^^ ^
;; :
Applicator

UV C u r i n
9

Fiber Drawing ^ ^ ^ ^ | ^ ^ Coating

^ Concentricity
l!iS;!;l;l;l:!:l;!:l;!;l;!;!:!:!:l:!;!:!:l;!:!;l
^^^ Monitor

Control of both the drawing processes and the coating has a profound effect
on the quality of glass fibers used for optical communications. A process for
coating fibers immediately after the forming process was described in 1980
(16) in which a low modulus polymer coating was used to protect the freshly
 61

formed fiber surface from abrasion or other damage. Using a laser forward-
scattering monitor, researchers were able to obtain fibers with diameters
exhibiting standard deviations of < 0.25%. An in-line temperature controlled
coating applicator, resembling a magnet-wire-coating applicator, and a
coating concentricity monitor enabled a protective layer to be applied that
could be cured quickly.

Since the fibers need to be protected as soon as they are formed, the types of
coating materials, the techniques for applying them, and the curing methods
are severely limited. As a practicality, any system that requires removal of
solvents and/or reaction by-products will be too slow to permit commercial
production of coated fibers. Additional restrictions are that the application of
the coating must not damage the surface of the fiber, and the film must be
solid by the time it reaches the capstan that guides the fiber onto spools or
drums.

It is important that the modulus of the coating be low, because it must

withstand stretching and bending after the fiber is coated. Many of the low-
modulus thermoplastics are viscous, and they remain hot (and thus soft) too
long for high- speed operations.

Among the low-modulus thermosetting plastics, urethane-acrylate and epoxy-

acrylate polymers, cured by exposure to ultraviolet light, have been used
successfully. Certain acrylics and silicones are also used.

Typical parameters of coating include:

1. Rate of draw = 5 to 10 meters per second

2. Low absorption of water = 0.3% in 21 days immersion

3. Tensile strength of coating = 500 psi.

4. Refractive index = 1 . 4 2

5. Glass transition = 1 2 6 °C.

62

6. Ultimate elongation = 145%

7. Young's modulus (psi @ 2.5%) = 140 - 390 @ 125 °C; 360 -

1240 @ - 60 °C.

The glass transition temperature for these thermoplastics is the same as we

have already described, i.e. the temperature at which the plastic coating
changes from a soft elastomer to a hard glassy state.

Present uses of optical fibers include worldwide telecommunication cables,

analytical instruments, military hardware, process controls and computer
links.

1.8.2.- The Sol-Gel Process

During the 4000 year history of silicate glass technology, materials that could
be added to make glass were limited to salts , minerals and oxides. It is only
in the past 40 years that a new method, called the "Sol-Gel Process", has
evolved in which amorphous silicates and oxides can be prepared at room
temperature via solution chemistry.

There are essentially two different kinds of sol-gel technology. The first, or
"colloidal" method, involves the dispersion of colloidal particles to form a sol
and then the destabilization of the sol to produce a gel (17). The second
method involves the polymerization of organometallic compounds such as
alkoxides to produce a gel with a continuous network (18). It is the latter that
has received the most attention, although if one wishes to add metallic ions,
including Na, Ca, A l and others, to a silicate sol, one must do so by
concurrently forming a colloidal gel of hydroxides.

In the polymerization method, a solution of alkoxides of Si, Al, Ti, and other
elements are hydrolyzed and condensation polymerization is caused to take
place. For example, tetra-ethyl silicate is converted into a silica [(Si02)nl glass
via a network of Si- 0 chains and rings. Such a reaction is illustrated as
follows:

(R0) 4Si + 4 H 2 0 4 EtOH + Si(OH) 4

 63

where R is usually either Me, Et or Isoprop, i.e.: CH3- , C2H5 - or C3H7 - . It is

the silicic acid, i.e.- Si(OH)4 t which forms the base for polymerization. This
reaction can be catalyzed by either acid or base. Thin films of the same
material can also be made using a dip-coating process. As the polymerization
takes place in solution, the monomer solution gradually thickens and becomes
a gel. On drying, the gel slowly shrinks and, when dry, becomes a rigid,
amorphous, porous mass called a xerogel. The rate at which the gel dries out
and collapses on the substrate has a major effect on the porosity, which in
turn affects the mechanical, optical and electronic properties of these
amorphous oxide glasses. Controlling the porosity has definitive effect on the
final properties of the glass so-produced. This is made more evident by the
following diagram, where various stages in formation of a xerogel are shown:

Sequence of Structural Changes during Gelation,

Drying and Densification of Sol-Derived Glasses

DRIED GEL PARTIALLY GLASS

DENSIFIED GEL

The steps in preparing a gel- derived glass, as shown in the above diagram
include:
64

A. Forming the Sol

Β. Gel Formation

C. Aging the Gel

D. Drying the Gel

E. Densifiying the Gel

F. Glass Formation

Note that a sol is first formed as individual particles which then coalesce to
entrap water to form the gel. A s the gel ages, it condenses and finally
becomes dried to form a semi-solid (dried-gel). Upon further aging and
drying , a partially densified gel results, which is the xerogel. Firing at an
elevated temperature then produces the desired glass (but usually not without
pores). Note that the porosity is influenced by a host of processing factors,
including the aging conditions and the relative rates of condensation and
drying. It is thus possible to systematically vary the pore sizes in xerogels
from less than 0.4 nanometers to greater than 3.0 nanometers. Note that the
Figure given above shows how each separate stage affects both the density
and the porosity of the product xerogel.

The characteristics of xerogels of a particular composition are related to

factors that affect the rate of hydrolysis and polycondensation reactions. So
far, pH, gelation temperature, aging time and temperature, and concentration
and molar ratio of water to the metal alkoxide have been identified as the
most important parameters. These factors are summarized in the following
diagram, where R can be any alkyl group such as methyl or ethyl, as given on
the next page.

These stages apply not only to tetraethyl silicates being hydrolyzed, but to
metal alkoxides as well. In the diagram given on the next page, five (5)
stages are shown.
 65

These are:

1. Raw material stage (monomer)

2. Hydrolysis stage
3. Polycondensation stage (rate of gelation)

4. Drying stage (rate of removal of water from gel)

5. Xerogel stage (dried gel product)

Hydroxyl Catalyzed Acid Catalyzed

High Rate of Condensation

1
High Rate of Hydrolysis

Rapid Gelation
Slow Gelation
(Growth Rate
( Nucleation Rate
Determining Process)
Determining Process)

|Drying| r

[Rate of Evacuation |Rate of Evacuation |

1 r
Slow Rate Hypercritical
Slow Rate Hypercritical
I I _ l
Porosity of Product Xerogel Porosity of Product Xerogel
ι I
Low Porosity High Porosity Low Porosity High Porosity

In general, hydroxyl (base) catalysis produces a lower degree of porosity in

the xerogel than the acid-catalyzed material. The pores of sol-gel glasses are
significant not only for the physical attributes they confer, but for what they
can contain. Dopants can be added to the alkoxide mixture undergoing
hydrolysis and polymerization. The most recent advancement has been the
addition of organic compounds which are contained within the disordered
66

Channels and pores of the g l a s s . A p p l i c a t i o n s are foreseen in optics,

electroöptics, sensors, catalysts, biosyntheses, and others (19).

+
It is well established (20) that the presence of acid |(Η3θ) } in the solution
increases the rate of the hydrolysis reaction, whereas hydroxyl ( O H ) ions
increase the polycondensation reactions. Depending upon the experimental
conditions of solvent extraction, either powders, large grains or monoliths
can be obtained.

THE GELATION AND CONDENSATION PROCESSES

T h e alkoxides mostly employed include: (ROUSi, (ROUTi, ( R O ^ B , and

(R0)4A1, where R is usually Et- or Me-. Prassas and Hench (21) studied the
g e l a t i o n , c o n d e n s a t i o n and x e r o g e l - g l a s s c o n v e r s i o n s of silica-based
compositions, using infra-red techniques to determine rates of gelation and
condensation. Whereas (EtOUSi (tetraethyl silicate = TES) is rather insoluble
in water, (MeO^Si (tetramethyl silicate = TMS) reacts rapidly to form both
Si(OH)4 and MeOH. Therefore, methyl or ethyl alcohol is often used to
increase solubility in water. If the chemical process is carefully controlled
( p H , r a t i o of c o n c e n t r a t i o n of (R0)4Si to w a t e r , t e m p e r a t u r e , etc.),
precipitation can be avoided and homogeneous gelation takes place.

As stated before, it is the Si(OH)4 which acts as a base for the polymerization-
condensation reactions taking place. The same is true for the products from
the titanium-based, aluminum-based and boron-based alkoxides as well. It is
also easy to see that the ratio of water to alkoxide present is the major factor
which controls the gelation process. Too low a water/alkoxide ratio results in
a sol. A correct proportion will result in a monolithic gel and xerogel. Note
also that the water added can be pH-controlled so as to control the rate of
gelation and polycondensation.

There is no incubation time for the polycondensation reaction. This takes

place between a silanol (and/or titanol, etc.) and an alkoxide group. The
molecular complexity is intricate and not thoroughly understood, primarily
because it depends upon several factors, including that of the ratio of alcohol
and water present, pH, temperature and total concentration of the alkoxides
 67

present. Nevertheless, from the infra-red data observed (21), linear and
cyclic polymers are known to result.

Crosslinking agents such as ethylene glycol have also been used concurrently
to control the v i s c o s i t y of the p o l y m e r solution. Such agents prevent
premature hydrolysis of the alkoxides present. This also has the advantage of
controlling the homogeneity of the gel so-produced. These agents react (22)
with the alkoxides to form glycolate intermediates so as to control the
stability and Theological properties of the gel being produced.

DRYING THE GEL

The gel consists of two phases, the network solid phase and the connected
pores filled with liquid phase. Initially, the surface is covered with the liquid
phase. As it drys, the liquid tends to impose a capillary force on the solid
network (due to surface tension forces) and the solid surface is exposed.
Because the network is compliant, the gel shrinks under capillary force as the
liquid evaporates. The gel thus becomes stiffer because the solid network is
becoming more tightly bound. When the gel is too stiff to contract under
capillary force, shrinkage stops and the meniscus of the liquid phase in the
bigger pores begin to recede into the interior of the gel. Note that the
menisci of the bigger pores have a larger radius of curvature and hence a
lower capillary force. Thus, the rate of evaporation is greater in large pores
than in smaller ones. Additionally, liquid will flow back into smaller pores, if
they are connected to the larger pores, because of the greater force of
capillarity in the small ones. When the walls of the bigger pores are dry, the
interior surface does not shrink further, even though the region around the
smaller p o r e s s h r i n k s u n d e r s u r f a c e t e n s i o n . T h i s c a u s e s a s t r e s s
concentration at the bottom of the bigger pores. The bigger pores thus act as
a microscopic flaw, and when the stress concentration is great enough, a
macroscopic crack will develop. It is for these reasons that gels with different
pore sizes are easily cracked during drying. Indeed, it has been shown (22)
that only gel films with a thickness of less than 1.0 μ can be dried on a
substrate without cracking.

Since the cracking of monoliths and films is caused by capillary forces

imposed by both the liquid phase present, and the variation in pore sizes,
68

reduction of the surface tension of the liquid and formation of uniform pore
sizes would be expected have a major effect upon the drying and aging of
monolithic gels and xerogels. Two methods are currently being practised: 1)
the use of drying control chemical additives (DCCA), and 2) hypercritical
drying. By the use of these two techniques, it is possible to optimize gelation,
aging and drying of gels to produce fully dried monolithic xerogels rapidly and
routinely.

In hypercritical drying, the monoliths are dried by heating the gel under
pressure to a point above the critical temperature and pressure of the solvent.
Thus, no capillary force is present. The hypercritically formed xerogel has the
same volume as the gel from which it was made. If it has been prepared
carefully so that the pores are nearly the same size, the xerogel monolith can
then be fired to form the glass article without fear of pores or cracking in the
final glass product.

A surfactant acts to reduce the capillary force of the liquid phase so that its
use will reduce the stress concentration in the solid phase during the drying
stage. The same may be said when DCCA's are employed. Several kinds of
chemicals have been used as a DCCA, including formamide and oxalic acid.
Both of these produce more uniform pores than can be obtained otherwise,
which helps to increase the permeability of the gel.

Both of these DCCA reagents can be used in either an acidic or basic medium.
In a basic medium, formamide slows down the hydrolysis reaction by some
five times (23). In contrast, formamide in an acidic medium results in a 3.5
times faster rate of hydrolysis. Additionally, its use leads to a larger pore
distribution, but with a narrow distribution of sizes. Thus, the gel network has
substantially larger necks, greater strength, and is more capable of resisting
drying stresses.

T h e following shows a representative pore distribution obtainable with

formamide, viz-
 69

Pore Size Distribution of Gels

Prepared with Different Amounts
of Formamide as a Drying Chemical
Control Additive (DCCA)

10D

25% Formamide
by Volume Added

50% Formamide
by Volume Added

100
Pore Radius in Â

Use of oxalic acid as a DCCA also controls the size and shape of the pore
distribution level. Either DCCA. i.e.- formamide or oxalic acid, greatly
decreases the breadth of pore distribution which decreases the magnitude of
capillary stresses during drying.

It is clear that the use of D C C A ' s in the sol-stage affects each of the
succeeding stages. Addition of a basic DCCA such as formamide produces a
large sol-gel network with uniformly larger pores. An acidic DCCA such as
oxalic acid results in a somewhat smaller scale network after gelation, but also
with a narrow distribution of pore sizes.

What we are addressing is illustrated in the following diagram, where the

three stages of the gel formation are shown, viz-
70

Step 1. - I SQL FORMATION 1

Basic DCCA I

10Â 50Â
«
100Â O O1 0Â
Large

_ • — Pore

1 Ho DCCA

^ 30 A 50 Â 100 A 200 Â * ••TJ.* A l

^ φ**·* m
10A ISA 20A 25A • • , * φ ^
Variable \ ·*·Α'# ·

·········
Size Pores φ w Φ

' 1
Acid DCCA Φ · Φ

ι 1
Uniform ^ Φ· · · · · · · ·
Small Pore · ·-.· · ·
s[ · · · ·
Step 2 - I Gelatio
n
Without DCCA With DCCA

Large Pore Distribution Small Pore Distribution

 71

Step 3 - [Aging of the Gel

Without DCCA With DCCA

Uneven Connectivity Even Connectivity

of Network and Pores of Network and Pores

Thus, either basic or acidic DCCA's can minimize differential drying stresses
by minimizing differential rates of evaporation and ensuring a uniform
thickness of the solid network which must resist the drying stress.It should
be clear, then, that the use of DCCA s in sol-gel technology represents a major
achievement in glass science, and the ability to prepare monolithic glass
articles.

THE GEL TO GLASS CONVERSION

Successful conversion of gel monoliths to glass depends upon knowledge of

the phenomena that lead to densification, and the avoidance of crystallization.
It has been shown (21) that there is a characteristic temperature at the
beginning of densification, called Tbd . Estimation of Tbd can be made by
dilatometric means. This is shown in the diagram given on the next page.

Note that Tbd in this case is about 425 °C. Sintering to form glass appears to
involve a viscous flow mechanism. An abrupt change in textural properties
above Tbd usually is observed. Thus, the original textural characteristics of a
given xerogel has a major effect on its behavior during sintering. For example,
xerogels with the same chemical composition, but with variations of 3- 25
times in their specific surface areas, exhibit marked differences in their
2
sintering behavior. A xerogel with a low surface area of, say, 9.0 M / g m could
not be densified to form an amorphous product. A xerogel with a high surface
2
area of 355 m / g m densified easily (21). Thus, the firing process is strongly
dependent on the xerogel micro structure, which is determined by conditions
of gelation, aging and drying.
72

Densification Curve of a Monolithic Xerogel Artifact

as a Function of Temperature
151

ö
Temperature in C.

Typical Thermogravimetric Analysis of a Xerogel

loo I 1 I I I Ι-
ο 10Q 200 300 400 500
Temperature in ° C.
 73

Thermogravimetric analysis can also be applied to identify some of the

processes taking place. A diagram illustating this aspect is also given on the
previous page, and may be regarded as representative of the weight changes
which occur.

Up to about 150 °C, the xerogel continues to lose weight by evaporation of

physically adsorbed water and solvent. At 150 ° C , pyrolysis of the organic
residues begins and continues to about 400 °C. From 400 - 600 ° C „ the tiny
pores collapse and the article is densified. This continues up to about 800 °C,
by a viscous flow sintering mechanism. In this range, the large pores collapse
and a dense amorphous product results. The resulting product is identical to
that made by conventional glassmaking processes.

Indeed, it is now possible to produce 100 % pure silica that has been poured
into molds of desired shapes and hardened, not ground down and polished as
in conventional glass technology. As a result, optical parts can be molded to
exacting dimensions.

Geltech, Inc., located in Alachua, Florida 32615 uses proprietary sol-gel

technology to custom-manufacture high-purity silica lenses and porous glass.
As shown in the following photograph, given on the next page, they are
available in various shapes and sizes, up to 40 mm. in diameter.

Also pictured are flanged lenses for easy positioning and alignment, as well as
a porous silica glass which presents unique, fully interconnected and uniform
porosity, this can be used as a single material, or as a basis for doping to
create new composites. Lenses are available in any combination of piano,
convex, and concave surfaces, including aspherical lenses, meniscus lenses
and prisms.

Another commercial use of the sol-gel technology has been the preparation of
lead-glass crystal (30% PbO). Aqueous Pb(NOs)2 solution is combined with
potassium silicate (water-glass) solution to form a gel, which is then sintered
@ 700 °C. , and then melted @ 1200 °C. A considerable savings in energy has
resulted, in addition to a marked improvement in handling safety and product
quality (24).
74

Still another use has resulted, that of manufacture of large mirror blanks for
telescopes by the sol-gel technique. In this case, the glass composition
required is that of fused silica or "ultra-low expansion" titanium silicate glass.
The process steps (25) include:

Manufacture of Large Telescope Blanks via the Sol-Gel Method

mix and PURE

Κ-Silicate RIGID Dealkalize
SiO sol pour gel ^ > GEL S i 0 2S H A P E '
Formamide @ < 100 °C SHAPE control pH < 100 p p m j

K20

Dry &
Prefire
to 1000 °C.

OPTICALLY Consolidate STRONG

CLEAR FUSED 1
SILICA SHAPEj POROUS
1300 - 1720 °C
Si 0 2

SHAPE
 75

In this process, each mirror segment is cast directly in the gel-form into its
final configuration, and then processed to form the final structure. Physical
properties of "gel-glass for reflective optics" was entirely equal to that of Code
#7940 (Corning Glass Works, Elmira, NY) usually employed.

1.8.3.- Architectural Glass

By architectural glass, we mean panels of glass used in the construction of

buildings and homes. The use of glass as windows to admit light dates back to
Roman times, about 2500 years ago. Originally, man lived in caves and other
natural structures. Later, rude dwellings were built from branches tied
together and perhaps covered with mud. Or, they might be formed from
animal-skins tied together over a framework of branches. The only inside
light was from the hole at the top of the structure which allowed smoke from
a fire to dissipate. More sophisticated structures began to appear near to the
end of the New Stone Age, about 6000 years ago. A t first, a window was
merely an opening to provide light and air in a dwelling. Only crude shutters
and lattices covered the first w i n d o w s . Later, linen, oiled paper, and
sometimes mica or gypsum were used for window panes, in addition to
wooden shutters.

So-called "City-States" were established about 5000 years ago where people
built complicated and fortified structures for mutual protection. Most walls
were as much as 30 feet thick and were slanted from bottom to top to
compensate for the poor strength of the sun-dried mud bricks used in their
construction. Windows were an impossible luxury. Approximately 4000 years
ago, the peoples of the Mediterranean area developed tombs and temples
built of stone blocks. The great temples and pyramids, built about 3500 years
ago, are fine examples of stone construction. However, windows were again
an impossibility. In Egypt, houses were made largely of wood, sun-dried brick
and mud plaster, with openings as windows. The hot climate dictated use of
shutters, but not closure of the windows. Walls of buildings of the Babylonian
and Assyrian era, constructed of clay and straw bricks, also had to be 30-40
feet thick in order to carry weight and not to crumble when wet. Long,
narrow rooms were the rule and doors were narrow. Again, ventilation and
internal lighting was accomplished by holes in the roof and torches hung on
the walls. Windows were virtually unknown.
76

It was the use of bricks in Babylonian and Assyrian buildings that made
possible the use of the arch and the vault. The Minoan culture on the island of
Crete used the arch to support wall construction and the column to support
the roof. Their buildings consisted of courts, surrounded by rows of columns,
and had guarded gateways. Because of the warm climate, windows were not
needed. The buildings were characterized by low, wide arches, thick walls,
heavy roof supports and virtually no windows. This architecture was carried
over into the Greek and Roman cultures, each of whom made contributions.
Following the fall of Rome in 476 AD, Romanesque style buildings began to
appear in Europe in the 900s, using the ribbed vault and pointed arch.

Houses then became more livable, with somewhat larger windows. However,
very large windows were little used, except in the churches. In this case, a
framework of light stone bars divided a large window into smaller areas so
that colored or "stained" glass panes could be more easily supported. Usually,
this took the form of tall narrow arched divisions below, with circles, cusps
and other shapes cut into the stone filling the upper part of the window.
When the great Gothic cathedrals of Europe were built in the Middle Ages,
their beautiful stained-glass windows became famous. Because of the great
difficulty in obtaining large flat pieces of stone, small panes were later
fastened together by canes of lead metal to form a larger network. Such
"windows" persist even today as "Stained-Glass". By the 1200s, small-paned
glass windows began to be used and were common from the 1500's onward.
In the present era, homes and buildings are planned so that windows will
admit as much light as possible. Sometimes, walls of glass-blocks are used.
Large picture-windows have become the norm. The disadvantage of having
many large windows in a building or home is that they permit a great amount
of heat to be lost during cold weather. This has led to "double-glazed"
windows, that is, two glass panels set together, but separated by an air-space,
in a single pane.

Today, in the 1990's, an ever increasing number of buildings, whether new or

old ones requiring renovation, are being given an easy-care glass facade all
around. The entire facade of the building is covered by sun-resistant glass
sheets and color-matched parapets which are attached to the structural mass
by means of a virtually invisible steel or aluminum support. Glass can also be
adapted to other facade elements such as polished marble, granite, or other
 77

decorative materials, producing buildings with a légère look and timeless

elegance. Depending upon a building's function and construction, there are
different facade systems available. Cold facades consist of glass sheets,
normally toughened glass with rear ventilation, which are mounted at least 2
cm. away from the actual parapet. In the case of warm facades, the glass panes
are fused onto a parapet without any air gap, by means of a heat insulating
sheet and a steam barrier which faces inward. These elements form a wall
and also protect the interior from climatic influences and heat. In waste air
facades, the glazing consists of an external glass unit and another glass sheet
attached from the inside at a distance of at least 15-20 cm. The volume
between the panes is used to discharge waste air from the building and to
heat a new supply of air. Another method is the "second-skin facade" where
the facade is made primarily of sun-resistant insulating glass. Such a building
can be seen in the f o l l o w i n g d r a w i n g w h e r e i n such glass panels are
incorporated. Note that the entire external surface of the building surface is
covered except for the columns.
78

The glass surrounds the building like a secondary skin, also covering glazed
parapet glass panes. Since the sun-resistant glass in front absorbs or reflects
the sun's rays and the residual heat in the air gap is released to the glass in
front of the facade, up to 80% of the insulating solar energy can be kept out,
reducing the overall energy transmission (g-value) to 20%. T h i s type of
construction is primarily ideal for fully-air-conditioned buildings where sun
protection is absolutely essential. It also can be assembled in front of existing
facades of buildings, lending an entirely new appearance. By using colored or
reflecting glass, the old facade is no longer visible but the view from the
inside is not impaired.

Such construction has occurred in much of the United States and Europe
because of the nature of so-called "solar-windows". There are two types of
solar-windows, "passive" and "active". We shall address passive solar-windows
first.

Passive Solar-Windows

In the early 1970's, costs of heating buildings were low, due primarily to the
low price of heating oil. Since that time, costs have risen sharply, and
impetus to reduce such costs have also greatly increased. Because of this, the
numbers of solar-windows installed in buildings have risen dramatically. Such
windows function to admit certain parts of the visible spectrum, while
reflecting other parts, including the near infra-red part of the spectrum. In
this way, heat inside the building is retained in winter. In summer, the heat
of the Sun is reflected outside, thereby keeping the building cooler and
reducing costs of air-conditioning. Although several glass manufacturers
supply solar-windows, we will describe in detail those supplied by PPG
Industries, Pittsburgh, Pa. which are marketed as SOLARCOOL BRONZE™
reflective windows. This name was chosen because of the reflective color one
sees when the windows are in place. A wide range of reflective glass windows
have become available for sun protection in a wide variety of colors, as
described below.

The construction of such windows involves two plate glass panes, separated by
an air-space. A layer of transition-metal oxides is applied to one surface of one
of the glass panes, as shown in the following diagram:
 79

1ST SURFACE FILM (OUTSIDE)

:>>>>;GLAS5 P A N E > > > > > > > > > > > >

2nd SURFACE FILM (INSIDE)

I AIR - SPACE
:
; ; > : i i g l a s s "pane
: ; : ;

The film can be on the outside of the glass pane or the inside. Usually, only
one side is coated. Note that the window is "double-glazed" to ensure that
heat losses do not occur when the window is installed in place on the
building. The film functions to admit certain parts of the solar light spectrum
and reflect others, so as to aid in climate control within the building.

The FILM is applied to the hot glass ribbon as it emerges from the "float-
glass" furnace, using a pyrolytic method. This means that a mixture of the
desired components is sprayed on the hot glass surface, where it pyrolyzes to
form a uniform film. In this case, a mixture of transition-metal oxides is
formed. This oxide film, once formed, is thermally stable and does not fade in
bright sunlight, including Equatorial light-intensities. A typical chemical
composition is:

Chrome - i.e.: Cr2 O3

Cobalt oxides - typically three valence states

i.e.: CoO, C02O3 and C0O2.

Small amount of iron oxide (< 5%)

The usual method involves spraying an ethanolic solution directly on the glass
while it is still in the furnace. Such a solution can be easily controlled as to
specific ratios of components and to total amount of components per unit
volume. When sprayed on a hot glass surface, the ethanolate ignites and burns
off, leaving a uniform dense layer of intimately mixed oxides as a film.

The Oxide-Film transmits only a part of the incident radiation from the Sun.
The requirements for such a film are:
80

¥ Uniform within all parts of the FILM

* Of uniform transmission from pane to peine

¥ Contain no impurities not deliberately added

* Uncoated, and not overlaid by external Films or

coatings.

The last requirement even applies to transparent coatings which may affect
both the transmissive and reflective optical properties of the Oxide-Film.
Even though such a film may be invisible to the naked eye, it could affect the
operating characteristics of the SolarCool Film. The following diagram shows
the reflective and transmissive properties of a plate glass window pane when
installed in a building:

STANDARD CLEAR FLOAT GLASS

It is easy to see that about 77% of the Sun's energy is transmitted through the
glass pane. A certain amount of energy is also reradiated and convected
because of the reflective optical surfaces present.

The following diagram shows the reflective properties of SOLAR-COOL

windows when installed in a building:
 81

SOLAR - COOL BRONZE COATED GLASS

Note that only about 26% of the S u n s energy is transmitted by the Solar
Bronze window (or about 1/3 as much as the uncoated glass). Additionally, 3.4
times as much of the internal energy of the building is reradiated and
convected inside the building when Solar-Bronze windows are in place. The
effect on costs of operating heating and air-conditioning equipment is
significant, and in many cases, the installation of such windows is repaid
within only a few years.

The Performance Data for 1/4" (6 mm.) glass panes are:

Visible light transmittance =21%

Outdoor Visible Reflectance = 35%

Indoor Visible Reflectance = 14%

Shading Coefficient = 0.43

Relative Heat Gain = 105 BTU/Hr./Sq. Ft.

Total Solar Energy Transmittance = 26%

Total Solar Energy Reflectance = 30%

82

PPG applies quality-control procedures to all batches of Solar Cool Bronze

glass. After installation, each window pane is expected to exhibit a uniform
coating with no optical defects apparent. When viewed from the inside, each
window is a uniform medium-grey as one views the sky line. From the
outside, each window reflects as a mirror and has a dull bronze tint.

Several new reflective coatings have recently been introduced (26). These
include Sky-Blue and Supersilver by The Glaverbel Group (Brussels, Belgium).
The following T a b l e gives some of the performance comparisons to be
expected:
TABLE 1-7
Performance Comparisons of Some Reflective Glass Products
1/4" Tinted Visible Light % Outdoor Total Solar Shading
Glass Transmitt. Reflectance Energy Trans. Coefficient

Azurlite 72 7 38 0.61
Aquamarine

Solex Green 76 7 46 0.69

Solar Bronze 52 6 43 0.71

Solar Gray 41 6 46 0.69

Blue Tint 57 6 48 0.70

Stopray Silver 43 47 27 0.65

Stopray 50 13 31 0.58
Neutral

Supersilver 64 33 23 0.67

Sky Blue 35 19 23 0.60

Note the wide Variety of colors available and the dramatic improvements
made in external reflectance of the glass panels.
 83

Active Solar Windows

Most of these windows have remained in the research stage, but deserve
mention here because of their potential usage in the near future.

One type of window can be classified as surface emissive, that is - surface

emissivity can be controlled with applied voltage. The applied film consists of
electrochromic materials formed by vacuum deposition of thin films. The
absorptivity, emissivity and reflectance of such films can be changed
reversibly by applying suitable voltage between two layers. Three versions have
been tested in which tungsten oxide was used as the insertion layer, and
lithium ions as the insertion medium to change optical properties. One
example is a six-layered structure of:

polyethylene terephthalate/Al/a-Li xW03/liquid electrolyte/a -Nb205/In203

which can be switched from a state of high solar absorbtivity with high
infrared emissivity to a state of low solar absorbtivity and low infrared
emissivity. Another structure tested is:

glass/In203/oc-Li xW03/liquid electrolyte/a -Nb205/In203

which can be switched from a state of low solar absorbtivity with low infrared
emissivity to a state of low solar absorbtivity and high infrared emissivity. Still
another combination is:

As2S3/In203/a-Li xW03/liquid electrolyte/a -Nb205/In203

w h i c h can be s w i t c h e d b e t w e e n t h e r m a l l y r e f l e c t i n g and t h e r m a l l y
transmitting states. None of these "windows" have become commercial as yet.

Finally, a "switchable" glass panel has been constructed using nematic liquid
crystals ( 2 7 ) . A t room temperature, the glass panel is translucent (like
2
"frosted" glass). About one watt of electricity is required to make 1 ft of glass
completely transparent. When the power is off, the Liquid Crystal molecules
are oriented in random fashion, giving a milky appearance. Application of
power causes them to align to transmit light and the panel becomes clear.
84

REFERENCES CITED

1. Obsidian: The Dating Stone"-1. Friedman and F.W. Trembour, Am. Set, 66
, 44-51, (1978).

2. Am. Soc. Testing Mater., ASTM C162-56- "Standard Definitions of Terms

Relating to Glass Products", ASTM Stand. 1965 Part 13, pp. 145-159 (1965).

3. W. H. Zachariasen, J. Am. chem. Soc, 54 3841 - 3851, (1932)

4. B. E. Warren, J. Appl Phys. 13 , 602-610 (1942)

5. P.H. Gaskell, M.C. Eckersley, A.C. Barnes & P. Chieux, Nature , 350, 675
(1991).

6. K. H. Sun, J. Am. Cer. S o c , 30, 277 (1947)

7. J.G. Vail, Soluble Silicates, Their Properties and Uses, Vol 1. - chemistry &
Vol 2. - Technology, Reinhold, N.Y. (1952)

8. J.H. Gibbs, Modern Aspects of the Vitreous State, - ed. J.H. Mackensie, pp.
152-187, Butterworths, Wash. D.C. (1960)

9. R.G. Ehman, Handbook of Glass Manufacture- F.V. Tooley - ed. Vol II, pp.
221-256, Ogden Pub. N.Y. (1960).

10. H.H. Holscher, loc cit, pp 299-387; see also: Glaze & Tims, loc cit, pp
57-87.

11. W. Eitel, Silicate Science, Vol 1 & II., pp 349-375, Academic Press, N.Y.
(1962).

12. R. Gunther, Glass Melting Tank Furnaces, Soc. Glass Techn. Sheffield,
England, (1958).

13. Hutchins & Harrington, Encyclopedia of Chemical Technology, Vol 10, pp

533-604, John Wiley & Sons, N.Y. (1966).
 85

14. U.S. Patent 4,920,971 (Jan. 1960) by S.D. Stookey (to Corning Glass
Works); U.S. Patent 4,933,857 (Apr. 1960) by S.D. Stookey (to Corning Glass
Works)

15. W.H. Armstead and S.D. Stookey, Science, 144 , 150 (1964).

16. L.L. Byler and F. V. DiMarcello, "Fiber-Drawing, Coating and Jacketing",

Proc. IEEE, 68 , 1194 (1980).

17. Aksay, Shih, and Sarikaya, "Colloidal Processing of Ceramics with Ultrafine
Particles", in Ultrastructure Processing of Advanced Ceramics, Edited by
MacKensie and Ulrich, p. 393, Wiley & sons, NY (1988).

18. Mehrota, "Chemistry of Alkoxide Precursors", J. Non-Crust Solids, 121 1-

6, (1990).

19. Chemical and Engineering News, pp 28-30, May 27, 1991 (Published by
the American Chemical Society, Washington, D.C.) S e e also: "Chemical
Sciences Impact on future Glass Research", by D.R. Ulrich, Cer. Bull., 64
1444 (1985).

20. L.C. Klein, "Sol-Gel Glass Technology- A Review", Glass Ind. 1981, 14-16.;
See also: Kamiya and Sakka, "Glasses Prepared from Metal Alcoholates", Res.
Rep. Fac. Eng. Mie Univ., 2 87-1-4 (1977); S.P. Mukherjee, "Sol-Gel Processes
in Glass Science and Technology", J. Non-Cryst. Solids, 42 477 (1980);
Zarzycki, Prassas and Phalippou, "Synthesis of Glasses from Gels: the Problem
of Monolithic Gels", J. Mater. Sci.,17 , 3371-3379 (1982); J.D. Mackensie,
"Glasses from Melts and Glasses from Gels, a Comparison", J. Non-Cryst.
Solids, 48 1-10 (1982); Kamiya and Sakka, "The Sol-Gel Transition in the
Hydrolysis of Metal Alkoxides in Relation to the Formation of Glass Fibers and
Films", J. Non-Cryst Solids, 48 ,31-46 (1982).

21. Prassas and Hench, "Physical Chemical Factors in Sol-Gel Processing"-

Chapter 9 in Science of Ceramics : Chemical Processing , edited by L.L. Hench
and D.R. Ulrich, Wiley, New York (1986).
86

22. Yi and Sayer, "Sol-Gel Processing of Complex Oxide Films, Cer. Bull. 70 ,
1173-1179, (1991).

23. Scherer, "Aging and Drying of Gels", J. Non-Cryst Solids, 100, 88 (1988).

24. "PERSPECTIVES- Glassmaking: "Soluble Silicates Can Save Time and

Energy", Chemistry in Britain, Vol. 21- #6 , p. 514 (June 1985).

25. "Gel-Derived Fused Silica for Large Optics", R.D. Shoup, Cer. Bull , 70
1505 (1991).

26. "Glass Innovations Reflect Architectural and Decorating Trends", Author

Unknown, Cer. Bull,70 49 (1991)

27. "Switchable Glass Gives Security and Privacy", B.E. Thompson, Laser Focus
World, June 1991, p. 55.