Those portions of glass-melting plant which come into contact with molten glass are almost universally made of some form of fire-clay. To discuss in detail the composition and properties of the varieties of fire-clay best suited to this purpose would exceed the entire limits of this book, so that only a few leading principles can be stated. Taking first the clays intended for the production of crucibles or “pots,” we find that for the purposes of the production of such objects the prepared clay must possess a certain degree of plasticity while damp and a considerable degree of strength when dried. The dried and burnt material must be so refractory as to resist the high temperatures used in glass-melting without undergoing fusion or even serious softening. Clays of various composition and physical nature also differ very widely in their power of resisting the chemical attack of molten glass; all clays are more or less dissolved under these circumstances, but not only the rate, but also the manner, of dissolution is of importance, so that frequently a clay which dissolves rapidly but uniformly is preferred to one which dissolves more slowly but in such an irregular manner as to throw off particles of undissolved material which contaminate the glass in the form of opaque enclosures or “stones.” It is also to be noted that the best results in this direction can only be obtained by careful adaptation of the clay employed to the particular kind of glass which is to be melted in the crucibles in question. In England this question has not received the amount of attention it deserves, but in Germany and America the available fire-clays of the country have been systematically studied and exploited. As a result the glass-maker has at his disposal a large selection of materials of accurately known physical and chemical properties. By carefully correlating these with the performance of his “pots” in the furnaces, the manufacturer is able to select the most suitable material, and is, moreover, in a position to know in what direction to look for improvement or for replacement if the supply of a satisfactory brand should cease.

We may now follow briefly the process of manufacture of a fire-clay pot or crucible. The size and shape of the crucible will depend upon the particular purpose for which it is intended. Crucibles varying in capacity from 4 cwt. to 2½ tons of glass are used for various kinds of glass, but the more usual sizes lie between 30 in. and 50 in. in diameter. For many kinds of glass the shape of the pot is simply that of an open basin, circular or oval in plan and larger in diameter at the brim than at the base (Fig. 1), but for the production of flint glass, and of other glasses which are to be protected from contact with the flame and gases of the furnace, so-called “covered” pots are used. In these the basin—here of a more nearly cylindrical shape—is covered over by a dome, and access is allowed only by a relatively small hooded opening (Fig. 2). Covered pots are built up on wooden moulds, which are made collapsible, and are removed before the drying of the pot is begun.

The material for pot-making is first prepared with great care. The proper variety of clay having been selected, it is ground to a fine powder in suitable mills and carefully sieved; with this fine clay powder is mixed, in accurately determined proportions, a quantity of crushed burnt fire-clay. In some works this burnt material is obtained by simply grinding up fragments of old used pots, but the better practice is to burn specially-selected fire-clay separately for this purpose. The quantity of such burnt material added to the mixture depends upon the chemical nature and especially on the plasticity of the virgin clay employed; with so-called “fat” or very plastic clays up to 50 per cent. of burnt material is added, but with the leaner clays, such as those of the Stourbridge district in England, very much smaller proportions are used. The object of this addition of burnt material is to facilitate the safe drying of the finished pots and to diminish—by dilution—the total amount of contraction which takes place both when plastic clay is allowed to dry, and further when the dry mass is subsequently burnt; the burnt material or “chamotte,” having already undergone these shrinking processes, acts both as a neutral diluent and also as a skeleton strengthening the whole mass and reducing the tendency to form cracks.

The virgin clay and chamotte having been intimately mixed, the whole mass is “wet up” by the addition of a proper proportion of water and prolonged and vigorous kneading, usually in a suitable pug mill. The mass leaves this mill as a fairly stiff, plastic dough, but the full toughness and plasticity of such clay mixtures can only be developed by prolonged storage of the damp mass. In the next stage of the process, the plastic clay is passed to the “pot maker” in the form of thick rolls, and with these he gradually builds up the pots or crucibles from day to day, allowing the lowest parts to dry sufficiently to enable them to bear the weight of the upper parts without giving way. The building of large pots in this way occupies several weeks, and during this time the premature drying of any part of the pot must be carefully avoided. After the completion of the pot, drying is allowed to take place, slowly at first, but more vigorously after a time when the risk of cracking is smaller; when it is taken into use, the pot is usually many months old and is thoroughly air-dry. The clay, however, is still hydrated, i.e., contains chemically combined water, and this is only expelled during the early stages of the burning process. This process is carried out in smaller furnaces or kilns placed near the melting furnaces. In these the pot or pots are exposed to a very gradually increasing temperature, until a bright red heat is finally attained. This is a delicate process in which great care is required to secure gradual and uniform heating, especially during the earlier stages, otherwise the pots are apt to crack and become useless. Finally, when a bright red heat has been maintained for at least a day, the pots are ready to be placed in the furnace, and this is ordinarily done while both pots and furnace are at a red heat, the pots never being allowed to cool down again once they have been burnt.

Fire-clay is also used in the manufacture of bricks and blocks of various sizes required for the construction of glass-melting furnaces. Here fire-clay is only used in positions where contact with molten glass is expected, as in the walls of the basin or tank proper in “tank” furnaces, or at a level below that of the pot or crucible in pot furnaces; in the latter position leakage of glass from broken pots or overflow being liable to result in an accumulation of molten glass on the floors or walls of the furnace and passages. The fire-bricks used in these latter positions are usually of a much poorer quality of fire-clay than that used for the manufacture of pots, and this is justified in so far as certain of the requirements that apply to crucibles do not apply here—but on the other hand the use of more refractory bricks would result in a longer life for the furnace. Such bricks, it should be noted, are not laid in mortar when used for furnace construction, but are set in a thin paste of fire-clay in water, and these joints are kept as thin as possible. The part of the furnace known as the “siege” (French “siège”), i.e., the floor of the furnace upon which the pots are placed, is usually built of very large blocks of fire-clay, made of coarse materials calculated to give great strength. At or near the points where the flame enters the furnace, these blocks rapidly wear away, partly by melting but chiefly by a process of abrasion, for it seems that a rapidly moving flame has an abrading action of a very marked kind.

The actual tanks or basins which contain the molten glass in tank furnaces are also built of large blocks of fire-clay, but these are made of the best procurable materials, and should receive at least as much care in every respect as crucibles; it is true that their shape and size gives them greater strength, but on the other hand these blocks are expected to resist the contact of molten glass for very much longer periods of time than the average crucible. To understand the requirements for tank-blocks it is necessary to anticipate the next section to the extent of stating that in tank furnaces the glass is contained, during melting, refining and working, in a basin built up of large blocks. These blocks are not cemented together in any way, but are built up “dry” and are supported on the outside by a system of iron bars and rods. The molten glass penetrates between the blocks to a certain extent, but as the outside of all such blocks is intentionally kept as cold as possible the glass rapidly stiffens as it penetrates further into these interstices, and this stiffened glass effectually binds the blocks together and prevents all leakage. It will thus be seen that the blocks are exposed to the full heat of the furnace and to the corroding action of the glass on the inner side, but are kept cold on the outer side. As this state of affairs tends to produce cracks, these blocks are necessarily made of rather coarse material. On the other hand, the material of a block never gets so hot as the wall of a crucible, which is heated from both sides, so that extreme refractoriness is not so essential.

It is impossible, within the limits of this chapter, to go into the details of the choice of materials for tank-blocks; it is a subject upon which no finally satisfactory conclusion has yet been reached, and what has been said above will suffice to show the nature of the considerations upon which such choice must be based.

We now turn to the second class of refractory materials used in the construction of glass-melting furnaces, viz., those which are so placed as not to come into contact with molten glass. Here mechanical strength and refractoriness are almost the only considerations, but in the roof-vaults or “crowns” of tank furnaces and also of furnaces in which glass is melted in open pots, there is the further consideration that the material of the bricks used shall not contain notable quantities of any colouring oxide, since small flakes, etc., are apt to drop down into the molten glass, and would thus be liable to cause serious discolouration. Such a material as chrome-ore brick is therefore excluded. As a matter of fact, some form of “silica brick” is in universal use. Bricks of this material, otherwise known as “Dinas bricks” from the place of their first origin, in Wales, consist of about 98 per cent. of silica (SiO₂). Pure silica cannot be baked or burnt into coherent bricks entirely by itself, since it possesses neither plasticity when wet nor any binding power when burnt, but an admixture of about 2 per cent. of lime and a little alumina makes it possible first to mould the bricks when wet and then to burn them so as to form fairly strong, coherent blocks. These are of amply adequate refractoriness for the highest temperatures that can be attained in industrial gas-fired furnaces, and their mechanical strength is sufficient to make it possible to build vaults of considerable span, but on the other hand this material requires very gradual heating and constant watching while the temperature is rising or falling to any considerable extent; the reason for this difficulty lies in the fact that silica bricks swell very markedly during heating, so that unless a vault built of this material is given room to spread somewhat, it will rise seriously and may even break up completely. This risk is avoided by gradually slackening the tie-bolts that hold the vault together, and correspondingly “taking up the slack” as the vault cools when the furnace is let out. Sudden local heat also has a disastrous effect on this material, producing serious flaking. For positions where intense heat is to be borne, and at the same time mechanical strength is required, silica brick is a most valuable material, but owing to its chemical composition it is rapidly attacked by molten glass or by any material containing a notable proportion of basic constituents, so that the silica bricks can only be employed out of contact with glass.

We now turn to consider, very briefly, the general design and arrangement of some typical glass-melting furnaces. The oldest and simplest form of furnace is, in effect, simply a box built of fire-brick, in the centre of which stands the crucible, while a fire of wood or coal is placed upon either side. To attain any great degree of heat by such means, however, the size of the box or chamber and especially of the grates in which the fires are maintained must be properly proportioned both to the dimensions of the crucible and to each other. The grates are generally wide and deep, while draught is provided by means of a tall conical chimney which stands over the entire chamber and communicates with it by a number of small openings. In a more refined furnace, the chamber itself is double, and the flame, after playing around the crucible in the inside of the chamber, is made to pass through the space between the outer and inner chamber before passing to the chimney or cone. We need not give any greater attention to these primitive furnaces, since they are practically obsolete at the present time. In modern furnaces the process of combustion is carried on in two distinct stages; the first stage takes place in a subsidiary appliance known as a “gas producer,” where part of the heat which the fuel is capable of generating is utilised for the production of a combustible gas; this gas passes into the furnace proper, either direct, while it is still hot from the producer, or after being conveyed some distance, when it is again heated up by the waste heat of the furnace. In either case the gas is hot when it enters the furnace proper, and there it meets a current of air, also heated by the aid of the waste heat of the furnace. Hot gas and hot air burn rapidly and completely, and if properly proportioned yield exceedingly high temperatures. Seeing that in this process a part of the heat of combustion yielded by the fuel is generated in a subsidiary appliance and is thus lost to the furnace, it appears at first sight somewhat surprising that this system of firing is very considerably more efficient than the old “direct” system where the whole of the fuel is burnt in the furnace itself. But the advantage arises from the fact that in the newer system the fuel is handled in the gaseous form. This has the advantage, first and most important, that the heat escaping from the furnace in the hot products of combustion (chimney gases) can be transferred to the incoming unburnt gas and air and can thus be returned to the furnace. The manner in which this is accomplished will be considered below, but it may be noted here that in some furnaces the escaping products of combustion are so thoroughly cooled that they are unable to produce an effective draught in the chimney of the furnace. Another advantage of the use of gaseous fuel is the fact that complete combustion can be obtained without the use of so great excess of air, such as is required when solid fuels are to be burnt completely. For this reason much higher temperatures can be readily obtained with gaseous fuel, while the pre-heating of both gas and air also facilitates the attainment of high temperatures; further, the great facility with which the flow of either gas or air can be regulated by means of suitable valves, makes it possible to secure much greater regularity in the working of the furnaces. Finally, in modern gas-producers, the amount of sensible heat generated and therefore lost to the furnace, is kept very low, the greater part of the heat set free by the partial combustion of coal in the producer being absorbed by the decomposition of a corresponding quantity of steam into hydrogen and carbonic oxide gas. The gas as it leaves one of these producers is not very hot, and the percentage of heat lost in this way is therefore much smaller than in the older forms of gas-producer.

It is again impossible, within the limits of this chapter, to enter into the details of construction and working of gas-producers. We must content ourselves with saying that most modern producers are of the form of a tower in which a thick bed of fuel is partially burnt and partly gasified under the action of a blast of air mixed with steam. The chemical actions that take place are complicated, but the final result is the production of a gas containing from 2 to 8 or 10 per cent. of carbonic acid, 10 to 20 per cent. of hydrogen, 8 to 25 per cent. of carbonic oxide (CO), 1 to 3 per cent. methane (CH₄), and 45 to 60 per cent. of nitrogen, with varying quantities of moisture, tarry matter, and ammonia. In good producer gas, the combustible constituents (hydrogen, carbonic oxide and methane) should total from 30 to 48 per cent. of the whole by volume, but the exact composition to be expected depends very much on the type of producer and the class of fuel used. Some producers are capable of dealing with exceedingly low-grade fuels, and the gas which they yield can still be utilised for obtaining the highest temperatures—a proceeding that would have been impossible if it had been attempted to burn these fuels directly in the furnace.

The gas on leaving the producer passes along fire-brick flues or passages to the furnace proper; the path which it is now caused to take varies somewhat according to the arrangement of the furnace in question. Modern gas-fired furnaces usually belong to one of two distinct types according to the manner in which the heat of the escaping products of combustion is utilised for heating the incoming gas and air; these two types are known as the “regenerative” and the “recuperative” respectively. In regenerative furnaces the hot products of combustion, after leaving the furnace chamber proper, and before reaching the chimney, pass through chambers which are loosely stacked with fire-bricks; these chambers absorb the heat of the escaping gases, and thus rapidly become hot. As soon as a sufficiently high temperature is attained in these chambers or “regenerators,” the path of the gas-currents is altered; the escaping products of combustion are made to pass through, and thus to heat a second set of regenerating chambers, while the incoming gas and air are drawn through the heated regenerator chambers before entering the furnace chamber proper. The incoming gas and air are thus heated, absorbing in turn the heat stored in the brickwork of the regenerators. It is evident that two sets of such regenerators are sufficient, the one set undergoing the heating process at the hands of the escaping products of combustion, while the other set is giving up its heat to the incoming gas and air; when this process has gone far enough, it is only necessary to interchange the two sets of chambers, by the operation of suitable valves, and this series of alternations may be continued indefinitely. The arrangement is shown diagrammatically in Fig. 3.

In recuperative furnaces the same principle is utilised in a somewhat different manner; the outgoing products of combustion pass through tubular channels formed in fire-clay blocks, while the ingoing gas and air pass around the outside of these same blocks; the heat of the outgoing gases is thus transferred to the incoming gases by the process of conduction through the fire-clay walls of the recuperator tubes. The relative merits of the two systems have been hotly contested; the regenerative system has the advantage of avoiding all reliance on the heat conductivity of fire-clay, while it also avoids the somewhat complicated special tubular blocks required for the other system; on the other hand, the recuperative system avoids the necessity for all “reversing” valves and their regular periodical working, while it also occupies somewhat less space. Temperatures sufficiently high for all glass-melting purposes can be attained by both means.

In both systems of furnace, heated gas and heated air are admitted to the furnace by separate fire-brick flues or passages, air and gas being allowed to mix just before they enter the furnace chamber proper. The economy and efficiency of the furnace depend to a very great extent upon the manner in which this mixing is accomplished. Rapid and complete mixing of air and gas results in an intensely hot, but short and local flame, while slower mixing tends to lengthen the flame and spread the heat through the entire furnace chamber; on the other hand if the mixing of gas and air is too slow, combustion may not have been completed in the short time occupied by the gases in passing through the furnace, and combustion may either continue in the outflow flues and regenerators, or it may be prevented by the narrowness of these passages, and unburnt gases may pass to the chimney. When the openings or “ports” are properly proportioned, and the draught of the chimney is properly regulated, combustion should be just complete as the gases leave the furnace chamber, and under these circumstances small tongues of keen flame will escape from every opening in the furnace; large smoky flames issuing from a gas-fired furnace indicate incomplete combustion.

As has already been indicated, glass is melted either in pots or crucibles of various shapes and sizes, or in open tank furnaces. The general arrangement of a pot furnace working with closed or “covered” crucibles is shown in Fig. 4. In this particular furnace, the “ports” or apertures by which the gas and air enter the furnace chamber, are placed in the floor of the chamber, but these apertures are often placed in the side or end walls, or even in a central column, the object being in all cases to heat all the pots as uniformly as possible and to avoid any intense local heating, which would merely endanger the particular crucible exposed to it, without greatly aiding the real work of the furnace. In pot furnaces, however, in which the more refractory kinds of glass are to be melted, it is generally considered desirable that the flame should be made to play about the pots in such a way as to heat the lower parts of the pots most strongly. In connection with the question of the uniformity of heat distribution in a gas-fired furnace it must further be borne in mind that in the case of regenerative furnaces the direction of the flame is reversed every time the valves are thrown over, and in practice this is done about once every half-hour; this proceeding, of course, tends very much to equalise the temperature of the two sides of the furnace. In recuperative furnaces, on the other hand, the direction of the flame is not changed, and for that reason a flame returning upon itself, usually called a horse-shoe flame, is often employed; this is obtained by placing the entry and exit ports side by side at one end of the furnace; the impetus of the flame gases and their rapid expansion during combustion carry the flame out across the furnace, while the chimney draught ultimately sucks it back to the exit ports, the shape of the flame being shown in Fig. 5.

In general arrangement, a tank furnace for glass melting resembles an open-hearth steel furnace. The tank or basin, as already indicated, is built up of a number of large fire-clay blocks, forming a bath varying in depth from 20 in. to 42 in. according to the design of the furnace and the kind of glass to be melted in it. The ports for entry of gas and air and for exit of the products of combustion are in most modern furnaces placed in the side walls of the furnace just above the level of the glass, the whole being covered by a vault built of silica brick. Figs. 6 and 7 show the general arrangement of a simple form of tank-furnace such as that used in the manufacture of rolled plate glass. The furnace indicated in the diagram is intended for regenerative working with alternating directions of flame; in recuperative furnaces the horse-shoe flame is always used in tanks, while this arrangement of ports is sometimes adopted for regenerative tanks also, particularly in the manufacture of bottles. For the production of sheet glass, tank furnaces are generally sub-divided into two compartments and are also provided with various constrictions intended to arrest impurities and to allow only clear glass to pass, but as regards the arrangement of flues and ports there is a very general similarity between various furnaces of this type.

It is beyond the scope of this book to discuss the relative merits of tank and pot melting furnaces; wherever the former can be made to produce glass of adequate quality for the purpose desired, the great economy of the tank furnace inevitably carries all before it, so that bottle glass, for example, is now made exclusively in tanks, and the same applies also to rolled plate of the ordinary kind, and to the great majority of sheet glass. On the other hand, where special qualities of glass are required in relatively small quantities, or where the requirements as to quality are very extreme, the pot furnace remains indispensable. Optical glass and coloured glasses are examples of this kind, although some tinted glasses are used in sufficient quantity to justify the use of small tank furnaces for their production. The causes of the greater economy of the tank furnace are numerous, and complicated by the detailed requirements of each particular manufacture, but the most important factors in the question may be summed up thus:—

(1) The tank furnace utilises the heat of the flame more efficiently, as the glass is exposed to the heat in a basin whose surface covers the entire area of the furnace, while in a pot furnace there is much vacant, unused space.

(2) The tank furnace permits of continuous working, the raw materials being introduced at one end while the glass is being withdrawn and worked at the other end. There are thus no idle periods, and each part of the furnace remains at or near the same temperature during the whole time that a furnace is alight. For a given size of plant, therefore, a tank furnace yields a much larger output, with a relatively smaller fuel consumption.

(3) The tank furnace obviates the need for pots or crucibles, which are not only costly and troublesome to produce, but are liable to premature failure and require periodical renewal, which involves a serious loss of time for the furnace.

(4) Finally, the molten glass in a tank furnace can be always maintained at or near one constant level and is, therefore, always convenient for withdrawal by means of the gatherer’s pipe or the ladle.

In pot furnaces, on the other hand, the composition of the glass can be more accurately regulated, and the molten glass itself can be more effectively protected from contamination either by matter dropping into it or by the action of the furnace gases, while in pots it is also possible to effectually melt together materials which, in the open basin of a tank, could not be kept together long enough to combine.