Of the various glass families of commercial interest, most are based on silica, or silicon dioxide (SiO₂), a mineral that is found in great abundance in nature—particularly in quartz and beach sands. Glass made exclusively of silica is known as silica glass, or vitreous silica (lechatelierite). (It is also called fused quartz if derived from the melting of quartz crystals.) Silica glass is used where high service temperature, very high thermal shock resistance, high chemical durability, very low electrical conductivity, and good ultraviolet transparency are desired. However, for most glass products, such as containers, windows, and lightbulbs, the primary criteria are low cost and good durability, and the glasses that best meet these criteria are based on the soda-lime-silica (soda-lime glass) system. Examples of these glasses are shown in Table 1-->.

The introduction to this article referred to W.H. Zachariasen's classic definition of glass as a three-dimensional network of atoms forming a solid that lacks periodicity, or ordered pattern. Just such a random atomic arrangement as would appear in a sodium silicate glass is shown schematically in Figure 2-->. Here the building blocks of the glass network are polyhedra formed around what is known as a network-forming (NWF) cation—that is, a positively charged ion such as, in this case, silicon (Si⁴⁺). The four positive charges of the silicon ion lead it to form bonds with four oxygen atoms, forming SiO₄ tetrahedra, or four-sided pyramidal shapes, connected to each other at the corners. An oxygen atom that connects two tetrahedra is known as a bridging oxygen. An oxygen atom joined to only one silicon atom is a nonbridging oxygen; its one remaining negative charge is satisfied by bonding to a network-modifying (NWM) cation—in this case, a univalent sodium ion (Na⁺)—which occupies an interstice adjacent to the SiO₄ tetrahedron. This corner-sharing tetrahedral structure achieves a liquidlike randomness, rather than a crystalline regularity, because there is a bending of the Si-O-Si bond at the bridging oxygen. In addition, there are twist angles arising between two connecting tetrahedra, as shown in Figure 3-->.

On the scale of several atoms, the structure of multicomponent glasses usually is not as random as that shown in Figure 2-->. This is because the various components of a molten mixture may display liquid-liquid immiscibility during cooling; that is, the components may separate into two or more disordered glassy phases that eventually are quenched in as glass inside glass when the substance becomes rigid. Two distinct mechanisms of phase separation exist, the nucleated droplet and the spinodal; the microstructures produced by these two mechanisms, as revealed by electron microscopy, are shown in . In the interface between the droplets and the matrix is sharp, owing to a sharp change in composition. With time the droplets increase in size until thermodynamic equilibrium compositions are achieved. In , on the other hand, spinodal structures, often wormlike in appearance, represent minor fluctuations in composition. With time the difference in composition becomes greater. An understanding of such variations in glass microstructure is vital to many industrial glassmaking processes (see Industrial glassmaking: Phase-separation techniques (industrial glass)).

Most properties of glass—except for elastic and strength behaviour in the solid state—are sensitive to its chemical composition and, hence, its atomic structure. (The role of composition and structure in the formation of the glassy state is described in Glass formation: Atomic structure (industrial glass).) In oxide glasses, the specific composition-structure-property relationships are based upon the following factors: (1) the coordination number of the network-forming (NWF) ion, (2) the connectivity of the structure, as determined by the concentration of nonbridging oxygens, which, in turn, is determined by the concentration and nature of network-modifying (NWM) ions, (3) the openness of the structure, determined, again, by the concentration of NWM ions, and (4) the mobility of the NWM ions. Thus, tetrahedrally connected networks, such as those formed by silicates and illustrated in Figure 2-->, are more viscous (viscosity) than triangularly connected networks, such as those formed by borates. In silicates, the addition of network-modifying alkali ions would raise the concentration of nonbridging oxygens, and the resulting lowered connectivity would lead to a lowering of viscosity. Networks in which the interstitial spaces are less filled with NWM ions possess lower density and allow greater permeation of gases through them. Since alkali ions are the most mobile species through interstices of oxide glasses, the higher the alkali concentration, the lower the chemical durability and electrical resistivity of the material.

In the random atomic order of a glassy solid, the atoms are packed less densely than in a corresponding crystal, leaving larger interstitial spaces, or holes between atoms. These interstitial spaces collectively make up what is known as free volume, and they are responsible for the lower density of a glass as opposed to a crystal. For example, the density of silica glass is about 2 percent lower than that of its closest crystalline counterpart, the silica mineral low-cristobalite. The addition of alkali and lime, however, would cause the density of the glass to increase steadily as the network-modifying sodium and calcium ions filled the interstitial spaces. Thus, commercial soda-lime-silica glasses (soda-lime glass) have a density greater than that of low-cristobalite. Density follows additivity behaviour closely. (The densities of representative oxide glasses are shown in Table 2-->.)

As can be seen from Figure 5-->, the viscosity of glass, as measured in centimetre-gram-second units known as poise, decreases with rising temperature. Figure 5--> also indicates the temperatures at which certain glasses reach standard viscosity reference points that are important in glassmaking. For instance, the working point, the temperature at which a gob of molten glass may be delivered to a forming machine, is equivalent to the temperature at which viscosity is 10⁴ poise. The softening point, at which the glass may slump under its own weight, is defined by a viscosity of 10^7.65 poise, the annealing point by 10¹³ poise, and finally the strain point by 10^14.5 poise. Upon further cooling, viscosity increases rapidly to well beyond 10¹⁸ poise, where it can no longer be measured meaningfully. (The softening point and working point of the major oxide glass families are indicated in Table 2-->.)

The annealing point and the strain point lie in the glass transformation range shown in Figure 1; often, the glass transition temperature (T_g) and the annealing point are used synonymously, and the strain point marks the low-temperature end of the range. The T_g may also be considered the maximum temperature for intermittent service. It is evident from Figure 5--> that the T_g of vitreous silica is the highest of the commercial glasses and that increasing the amount of alkali additions (and therefore the concentration of NWM ions) lowers T_g. Of all the various factors affecting viscosity, water, in the form of hydroxyl ions or molecular water, lowers viscosity the most.

As is evident from Figure 1, glass normally expands when heated and shrinks when cooled. This thermal expansion of glass is critical to its thermal shock performance (that is, its performance when subjected suddenly to a temperature change). When a hot specimen of glass is suddenly cooled—for example, by plunging it in iced water—great tension may develop in the outside layers owing to their shrinking relative to the inner layers. This tension may lead to cracking. Resistance to such thermal shock is known as the thermal endurance of a glass; it is inversely related to the thermal-expansion coefficient and the thickness of the piece. The thermal expansion coefficients of various oxide glasses are shown in Table 2-->.

The primary determinant of chemical durability in glass is an ion exchange reaction (ion-exchange reaction) in which alkali ions in the glass are exchanged with hydrogen atoms or hydronium ions present in atmospheric humidity or water. The alkali ions thus leached out of the glass further react with carbon dioxide and water in the atmosphere to produce alkali carbonates and bicarbonates. These are seen as the white deposits that form on a glassy surface in dishwashing tests or after extended humidity exposure (often called weathering). The relative “weatherabilities” of the major oxide glasses are indicated in Table 2-->. In addition, the weathering resistance of several commercial glasses is shown in Figure 6. In general, glasses that are low in alkali offer increased weathering resistance. Vitreous silica is the most resistant, but borosilicates and aluminosilicates also offer excellent weathering resistance.

Since univalent alkali ions have the greatest mobility through the glassy structure, they are the primary charge carriers of a glass and therefore determine its electrical conductivity. In general, the higher the concentration of alkalis, the higher the electrical conductivity—as can be seen in Table 2-->, where soda-lime silicates and sodium borosilicates are shown to have the highest conductivities of the glasses listed. The most noted exception from the additivity relationship here is the mixed-alkali effect, in which glasses containing two or more different types of alkali ions have a significantly lower electrical conductivity than linear additivity would suggest. In applications such as high-wattage lamps, where low electrical conductivity is desired, mixed-alkali glasses are useful.

The dielectric, or nonconducting, property of glass is important for its use either as a medium separating the plates of a capacitor or as a substrate in integrated circuits. For capacitor usage, the dielectric constant must be high, whereas for substrates it must be low enough to allow high signal speeds between semiconductor chips. In general, the dielectric constant of glass generally increases with the concentration of NWM ions. Therefore, as is shown in Table 2-->, vitreous silica has one of the lowest dielectric constants, while most soda-lime-silicates have high dielectric constants.

A ray of light, on passing from one transparent medium to another transparent medium of different density, will be transmitted through the second medium with no loss of intensity or change in direction if it strikes the boundary between the two mediums at a right angle (90°). In geometric terms, the right angle at which the light ray meets the boundary is called the normal. If the light ray meets the boundary at an angle other than the normal, then it will be partially reflected back into the first medium and partially refracted, or deflected, in its path through the second medium. The extent to which the light is reflected and refracted depends on the relative densities of the two mediums involved (usually glass and air) and on the angle at which the light ray meets the boundary (known as the angle of incidence). As is shown in Figure 7-->, if the light ray meets the boundary at less than a certain critical angle (θ_c), most of the light will be refracted while a small amount is reflected. If it arrives at the boundary at the critical angle, then the emerging light will be of diminished intensity and will assume a direction parallel and close to the boundary; most of the light will be reflected. Finally, if the critical angle is exceeded, all of the light will be reflected back into the glass without suffering any loss of intensity. Known as total internal reflection, this phenomenon is widely exploited in single-lens reflex cameras and in fibre optics, in which light signals are piped for great distances before signal boosting is required.

Refraction can be expressed as a constant, known as the refractive index, which is derived mathematically from the ratio of the sine of the angle of incidence on the medium to the sine of the angle of refraction within the medium. The refractive index of a particular type of glass depends on its composition and on the wavelength of the light. The refractive indices of various oxide glasses are shown in Table 2-->.

Glasses of commercial importance are composed of a variety of chemical compounds; these are described in Glass compositions and applications (industrial glass), and the constituents of the major oxide glasses are listed in Table 1-->. For glass manufacture on an industrial scale, these chemical compounds must be obtained from properly sized, cleaned, and treated minerals that have been preanalyzed for impurity. silica is obtained from clean sand. Appropriate mineral sources for soda are soda ash (sodium carbonate) and sodium hydroxide. Lime is obtained from limestone (calcium carbonate) or from dolomite (calcium magnesium carbonate) when magnesium oxide is also needed. In the past it was customary to add about 0.25 percent arsenic oxide and 0.5 percent sodium nitrate to aid in glass fining, or removal of bubbles. These chemicals are no longer recommended in view of hazards to the individual and the environment; instead, less noxious compounds such as sodium chloride, sodium sulfate, or sodium nitrate are recommended.

After a glass batch is mixed in blenders, it is conveyed to the doghouse, a sort of hopper located at the back of the melting chamber of a glass-melting furnace (see Figure 8-->). The batch is often lightly moistened to discourage segregation of the ingredients by vibrations from the conveyor system, or it may be pressed into pellets or briquettes to improve contact between the particles. The batch is inserted into the melting chamber by mechanized shovels, screw conveyors, or blanket feeders. Continuous glass-melting chambers are 6 to 12 metres wide and as much as 30 metres long (20 to 40 feet wide by 100 feet long). They may hold as much as 1,000 tons of glass and produce as much as 50 to 500 tons per day. For smaller production rates, day tanks or unit melters are used. In the large melting chambers, the tank is made of high-density, highly corrosion-resistant refractory materials, such as electrocast alumina-zirconia-silica, to ensure a trouble-free service life of 5 to 10 years.

Natural gas, oil, or electricity may be used to generate the heat of melting. For fossil-fuel firing, the furnaces are often of the regenerative type (see Figure 8-->). In regenerative ovens, firing is carried out in cycles. For half of the cycle (10 to 15 minutes), fuel and air are passed through a hot checker-brick arrangement in a set of regenerator chambers on one side of the oven. The heated mixture is then directed through ports to the melting chamber, where it is burned over the glass melt. The hot flue gases, after exiting the chamber through another set of ports, are directed through another set of regenerators, where they impart much of their heat to the checker-brick arrangement there. For the second half of the cycle, the firing sequence is reversed: combustible mixture is brought in through the second regenerator and is preheated by the checker-brick; this increases the thermodynamic efficiency of the combustion.

In the melting chamber, temperatures reach a peak of 1,475° C (2,685° F) for a soda-lime-silicate glass. At these temperatures, large quantities of gas are generated by the decomposition of raw materials in the batch. These gases, together with trapped air, form bubbles in the glass melt. Large bubbles rise to the surface, but, especially as the glass becomes more viscous, small bubbles are trapped in the melt in such numbers that they threaten the quality of the final product. They are removed in a process called fining, which takes place mostly in another section of the furnace known as the conditioning chamber (see Figure 8-->). From the melting chamber, the molten glass is allowed to pass through a throat in a divider wall, or bridge wall, into the conditioning chamber, where temperatures are held at about 1,300° C (2,375° F). Here the fine bubbles are removed by being dissolved back into the glass. In addition, the glass is homogenized by diffusive mixing. In order to ensure that the composition of the melt is uniform throughout, mechanical mixers or nitrogen or air bubblers can be installed in the bottom of the melting chamber. Special challenges to homogenization can be posed by residual unmelted material from the batch, particularly sand grains, as well as devitrification products, material from the refractory lining of the melting tank, foreign matter such as bottle caps, and vaporization of the various glass constituents, particularly of boric oxide and alkali. Glass homogenization, in fact, is the rate-limiting step in the entire glass-melting process.

All traditional glass products such as tableware, containers, tubes and rods, flat glass, and fibreglass are formed of glass made by the melting process. viscosity is the key property in glass forming. After melting and conditioning (described in Industrial glassmaking (industrial glass)), glass is delivered to a forming machine in a manageable shape at a viscosity of approximately 10⁴ poise. At this viscosity, indicated in Figure 5--> as the working point, the glass can be worked on to form the desired object and then released in a near-solid condition. All through the process, heat is extracted in a controlled manner in order to allow the viscosity to increase from the levels typical of a liquid to those of a solid.

The modern method of producing flat glass for such products as windows (window) and mirrors (mirror) is the float process, in which molten glass is brought over the lip of a broad spout, allowed to pass between rollers, and floated over a bath of molten tin in a steel container (see Figure 10-->). Glass enters the container at approximately 10^3.5 poise—a viscosity that, for soda-lime-silica glass, is present at a temperature greater than 1,000° C (1,800° F). It is cooled over the length of the tin bath, which has a melting point of 232° C (450° F), and exits in a nearly solidified sheet form with a viscosity of about 10¹³ poise. Under such conditions glass spreads by gravity to a thickness of 7 millimetres (0.28 inch), but, if it is compressed with graphite paddles or stretched with knurled rollers, glass may be made in thicknesses of 2 to 25 millimetres and in widths up to 4 metres.

The two types of OWG are fabricated in different types of processes. Stepped-index OWGs are made by fusing a glass rod of core composition inside a glass tube of clad composition and then drawing them together. They also are made by the double-crucible technique (see Figure 11-->), in which two concentric compartments of a platinum crucible are fed with glass rods, and a composite stream is allowed to exit a bottom orifice. In either case, the glass fibre is attenuated to its proper dimensions by a high-speed mechanical winder. Between the orifice and the winder, the fibre passes through a laser-monitored diameter-control feedback mechanism and is coated with a polymer, usually containing ultraviolet-cured polyacrylates, to provide protection from surface abrasion.

Graded-index OWGs are made by one of several vapour-deposition processes. The most popular version is called modified chemical vapour deposition (MCVD). In this method, an example of which is shown in Figure 12-->, silicon tetrachloride (SiCl₄) vapours are mixed with varying quantities of phosphorus oxychloride (POCl₃) and either germanium tetrachloride (GeCl₄) or boron trichloride (BC1₃). Heated to 1,300°–1,600° C (2,375°–2,900° F), the vapours undergo an oxidation reaction that produces a soot containing silica and other oxides. This soot is allowed to condense on a cooler substrate such as a tube or rod. The condensed material, which may be composed of several hundred layers, is then dehydrated at approximately 1,200° C (2,200° F) and sintered at 1,500°–1,700° C (2,700°–3,100° F). By controlling the ratio of the chloride mix, the desired composition (and hence the desired refractive index) of the layers are obtained—germanium increasing the refractive index and boron decreasing it. The glass preform is subsequently mounted on a fibre draw tower, and its end is heated to 1,900°–2,200° C (3,450°–4,000° F) inside an induction or resistance furnace or by means of an oxy-hydrogen flame or a carbon dioxide laser beam. As in the double-crucible technique shown in Figure 11-->, the diameter of the attenuated fibre is scanned and used as a feedback signal to control the preform feed rate, and the fibre is subsequently coated with polymer.

During the glass-forming process, glasses often develop permanent stresses because various regions of the material pass through the glass transition range at varying cooling rates and at varying times. In order to ensure dimensional stability (for instance, for space-based telescope mirrors) and to avoid the development of excessive tension in critical regions, these stresses must be reduced by the process of annealing. As is explained in Properties of glass (industrial glass), the atomic structure of a glassy solid undergoes a process of relaxation as it is cooled through the transition range. The time required for relaxation to be sufficient to reduce internal stresses can range from only a few minutes when the glass is held at its annealing point to a few hours when it is held at the lower temperatures of its strain point. (The strain point and annealing point of several oxide glasses are shown in Figure 5-->.) Practical annealing is achieved generally by holding the product approximately 5° C (9° F) above its annealing point for 5 to 15 minutes, followed by slowly cooling it through the glass-transition range and the strain point and finally to room temperature. In dead-annealing, glass is so well annealed that the internal tension is almost undetectable. As is explained in Optical properties (industrial glass), internal stresses are examined by using the photoelastic property of glass.

By the 15th century, Venetians had learned to make a fine rock crystal known as Venetian cristallo. Beautiful coloured glasses and techniques for decoration, such as gilding, also were developed. The Venetian product was exported to all over Europe and the Byzantine Empire. After Constantinople was captured by the Turks in 1453, however, the Venetian glass trade fell. Glassmakers emigrated from Venice and helped other Europeans set up their glass houses (see Figure 13-->). Germans and Bohemians (Bohemian glass) concentrated on the preparation and selection of newer, purer raw materials. Glass quality improved by the addition of low-iron potash and purer quartz. By 1680, Bohemian “crystal”—basically a potash-lime glass—was developed. In 1674 George Ravenscroft (Ravenscroft, George) of London experimented with “flint” glass, a lead-crystal (flint glass) composition made with a large proportion of calcined flints and potash. By using 15 percent lead oxide, quartz pebbles imported from the Po River in Italy, and purer potash, he produced a fine, lustrous glass, soft enough to be cut and engraved easily and of greater refractive power than the common soda-lime glass. By the end of the century, there were 11 houses in London producing leaded crystal. During the same period, Johann Kunckel (Kunckel von Löwenstjern, Johann) in Germany developed a reliable formula for producing ruby-red glass (ruby glass) using gold chloride. Gold was dissolved in aqua regia and mixed with the batch, which was then melted, formed, and subsequently reheated to “strike” the precipitation of the ruby-red colour. Kunckel also developed a phosphate opal glass, also called porcelain glass or Milchglas (milk glass), by adding burned bone or horn to the soda-lime batch.

D.R. Uhlmann and N.J. Kreidl (eds.), Glass—Science and Technology (1980– ), provides an in-depth discussion of glass for the advanced professional. Narottam P. Bansal and R.H. Doremus, Handbook of Glass Properties (1986), is a good compilation excerpted from published literature. Arun K. Varshneya, Fundamentals of Inorganic Glasses (1994), comprehensively covers glass composition, structure, properties, melting technology, and the forming of glass products and includes a chapter on glass transition-range behaviour and annealing/tempering.Samuel R. Scholes, Modern Glass Practice, 7th rev. ed. by Charles H. Greene (1975, reprinted 1992); and P.J. Doyle (compiler and ed.), Glass-Making Today (1979), discuss various aspects of glass manufacturing. Fay V. Tooley (compiler and ed.), The Handbook of Glass Manufacture: A Book of Reference for the Plant Executive, Technologist, and Engineer, 3rd ed., 2 vol. (1984), treats such topics as glass composition, raw materials, furnaces, refractories, fuels, and glasshouse instrumentation. George W. McLellan and Errol B. Shand, Glass Engineering Handbook, 3rd ed. (1984), describes the engineering and technology of glass products, detailing chemical durability, strength, and annealing techniques. ASM International, Handbook Committee, Engineered Materials Handbook, vol. 4, Ceramics and Glasses, ed. by Samuel J. Schneider, Jr. (1991), is an excellent updated compilation of articles written by various experts on a variety of topics covering glass melting and forming technologies.Harry Heltman Holscher, Hollow and Speciality Glass: Background and Challenge (1965); and R.W. Douglas and Susan Frank, A History of Glassmaking (1972), chronicle major historical developments.Arun Kumar Varshneya