samples of gases from Freiberg blast furnaces, and found on an average three times as much CO, as CO, besides small quantities of hydrogen and of marsh gas. Some of these analyses are given in Part ii., Chap. xv.

Conditions Favourable to Reduction.-The burning of carbon to CO or CO2 is largely influenced by the shape of the furnace and by the area of the tuyeres. A distinct contraction of the furnace at the tuyere zone (bosh) is accompanied by a much more rapid descent of charge there, and, consequently, by the accumulation of a thicker layer of glowing fuel between the liquid slag and the unmelted charge. The CO, produced before the tuyeres in its ascent has to pass through this layer of coke, and so becomes converted into CO before reaching the unmelted charge. During its slow passage upwards through the constantly increasing width of charge, the CO has ample opportunity to become oxidised to CO, by absorbing oxygen from the ore which it reduces. In the case of a furnace with vertical sides, however, the layer of glowing coke is thinner and its temperature is lower, because the heat is less concentrated; the CO, formed, therefore, largely remains as such.

Similarly, small tuyeres deliver streams of air, which are rapidly and completely absorbed by the glowing coke in their path to form CO, while the quantity delivered by a large tuyere is so great that the oxygen is largely in excess of the available carbon at any point close in front of the tuyere; hence combustion takes place almost entirely with the formation of CO., which in its passage upwards through the thin bed of glowing coke has no time to be converted into CO.

Another factor which influences the reducing action of the furnace is the blast pressure. A powerful blast forces its way to the centre of the furnace, where both any oxygen unconsumed on the way and the CO, formed by the combustion find themselves surrounded by an excess of glowing carbon and are rapidly converted into CO, which then rises through the charge, exerting its reducing effect on the metallic oxides. A weak blast, on the contrary, penetrates the charge with difficulty. Such portion as reaches the centre is, of course, converted into CO, but a large part is burnt to CO, in the immediate neighbourhood of the tuyeres, and rises through the charge, together with any oxygen which has escaped burning without being reduced by the fuel. In this way, although there is reduction in the centre, oxidising effects predominate all round the shaft.

Very small charges again, which bring about an intimate admixture of fuel and ore in the furnace, increase the reducing action; while the use of large charges results in fuel and ore reaching the smelting zone successively instead of together, and to that extent is not so favourable to reduction. Similarly, a fine-ore charge hinders the draught, and causes the furnace gases

to become more thoroughly disseminated through the ore column, so exercising a very thorough reducing action; whereas with a coarse charge a considerable portion of the blast can escape through open channels without exercising any reducing influence on the bulk of the ore.

The rapidity with which coke can be burnt in a furnace of given area with tuyeres of given size, and, therefore, the rate of smelting, depend upon the pressure of the blast, and any increase in this demands a corresponding increase in the height of the column of charge, in order to avoid waste of heat at the top. High pressure, therefore, and high furnaces mean rapid driving, especially when combined with a large tuyere

area.

The following conditions are favourable to a powerful reducing effect :

1. High column of charge.

2. Contracted tuyere zone (boshed furnaces expanding upwards).
3. Small tuyeres.

4. Small volume of blast at a high pressure.

5. Small charges of moderately fine ore.

Exactly the opposite conditions-viz., low column of charge, straight-sided furnaces, and a large number of tuyeres of large area supplying a great volume of air at low pressure to large charges of coarse material-produce an atmosphere which may be neutral or even oxidising in its effects.

"Tuyere Ratio" and "Tuyere Efficiency."-It is unfortunate that the habit of running furnaces from one general blast main, coupled with the use of the cheap and convenient so-called "positive blower," have rendered it impossible hitherto to ascertain, with any approach to accuracy, the volume of air which is supplied to any furnace, and on this account it is not possible to properly investigate the character of the work done inside the furnace under given conditions.

In the absence of satisfactory data as to the volume of air supplied, it is convenient, for the sake of comparing furnaces of different sizes, to make use of certain relations between the tonnage of ore smelted, the sectional area of the furnace at the tuyere level, and the total area of the tuyeres. The number of tons of material smelted in twenty-four hours per square foot of sectional area may be called, as first suggested by Lang, the "hearth activity," or "nearth efficiency," of the furnace; this will be affected not only by the "olume of air supplied, but also by its pressure.

As no data are obtainable in most cases with regard to the volume of air supplied, it may be assumed to be, for any given height of charge-column, roughly proportional to the total area of the tuyeres. The number of square inches of tuyere area provided for every square foot of effective sectional area between

the tuyeres may be conveniently termed the "tuyere ratio" of the furnace. With any given size of furnace and fixed "tuyere ratio," the quantity of ore smelted in twenty-four hours will be directly proportional to the pressure of blast. The number of tons smelted in twenty-four hours per square inch of tuyere area may therefore be called the "tuyere efficiency" of the furnace. The "hearth efficiency" of any given furnace will thus be the product of the volume factor, "tuyere ratio," into the pressure factor, "tuyere efficiency." These factors are all worked out for the furnaces, details of which are given in a subsequent table, as well as for the furnaces smelting silver ores without lead, described in Part ii., Chaps. xiv., &c.

Both "tuyere ratio" and "tuyere efficiency are of the utmost importance as affecting not only the capacity of the furnace, but also the quantity of matte produced from a given charge and the possibility of smelting imperfectly roasted ores, and they will be again referred to in subsequent chapters.

Composition of the Ore Charge. The composition of various lead ores before roasting has been given in Table XII., page 75; and the following Table XIX. gives the composition of certain roasted lead ores as well as of two "natural carbonate" ores added for comparison.

Reactions of the Ore Components.-Some raw ore being added to the roasted ore the chemical composition of the ore charge is often very complex. The different substances present behave as follows:

Lead oxide in powder or in porous lumps is partly reduced in the upper part of the furnace direct to metallic lead by carbonic oxide. Slagged or sintered masses are, however, impermeable to the furnace gases, and their reduction is only effected in the smelting zone by direct contact of the drops of molten oxide with glowing fuel.

Lead silicate is unaffected by CO, and even by contact with glowing carbon, except in the presence of a more stable base; basic lead silicates only give up a part of their lead, leaving an acid silicate. This is, however, readily decomposed by FeO, and at a much higher temperature by CaO also, and the PbO set free is reduced to metal by hot coke. The ferrous oxide acts equally well whether in a free condition, reduced from ferric oxide in raw or roasted ore, or combined with silica, from which combination lime will set it free in the smelting zone. In ordinary lead furnaces more or less metallic iron is reduced from its oxide in the hottest part of the furnace in contact with glowing coke, and this iron acts as a powerful reducing agent for the liberation of lead from both oxide and sulphide.

In connection with an excess of lime or other base, the sulphides of iron, calcium and barium, may act as reducing agents. These sulphides may be formed by reduction of their

TABLE XIX.-ANALYSES OF ORES FOR BLAST FURNACE SMELTING.

respective sulphates or by the action of pyrites on the bases, and their mode of action is as follows::-

2FeS + 4CaO + 3Pb.SiO4 = FeSiO4 + 2CaSiO4 + 3Pb2 + 2SO..

Lead sulphide is acted upon in the hottest part of the furnace by metallic iron, or by mixtures of ferrous silicates with carbon, and reduced to metal with formation of ferrous sulphide, but the reaction is never complete, some lead always remaining as sulphide in the matte

*

2PbS+ FeO + C = Pb + PbS. FeS + CO.

Percy gives an equation, according to which the presence of carbon is not necessary to the reduction of PbS by means of iron silicate, thus,

3(2FeO. SiO2) + 5PbS = 4FeO . 3SiO2 + 2(PbS. FeS) + 3Pb + SO2, the monosilicate becoming a sesquisilicate which has no further action upon PbS. At a higher temperature more lead can be expelled from the matte by the action of fresh iron, so that the proportion of lead in the matte varies almost entirely with the temperature. Berthier stated that lime also reduced lead from its sulphide, but, according to the experiments of Jordan,† there is no action between the two bodies at any temperature, whether in presence of carbon or not.

A certain proportion of lead is always liberated from lead sulphide above the smelting zone by reaction with oxide and sulphate, according to the characteristic equations given in Chap. ii. In some cases this reaction may become of the very highest importance, and nearly the whole of the sulphur present in the charge, even when amounting to 8 or 10 per cent., may be expelled as SO, almost without any formation of matte. The conditions upon which this reaction depends have been already touched upon in this chapter, and will be further discussed in connection with the description of Block 14 Works (Port. Adelaide). ‡

Finally, lead sulphide may be reduced to metal by the direct action of zinc vapours (themselves arising from the reduction of oxide in contact with carbon) producing zinc sulphide.

It is to be noted that most of the lead lost by volatilisation is in the condition of sulphide, as shown by the analyses of lead fume from blast furnaces run with a cold top.

Usually the largest

Lead sulphate behaves in several ways. part reacts upon lead and other sulphides above the smelting zone, liberating metallic lead and SO.. Another portion is probably reduced to PbS, and a third portion, in contact with free silica or with acid silicates, is converted into lead silicate;

*Metallurgy of Lead, p. 59. + Quoted by Percy, op. cit., p. 55.

+v. Chap. xi.