March 31, 1916. THE COLLIERY GUARDIAN. 605 REFRACTORY MATERIALS AND SALTY COALS.* By Prof. John Cobb. A very important item in the running costs of a by-product coking plant is that of repairs and renewal of oven walls; and before the reason was grasped, costly miscalculations were made under this head, when the coal to be carbonised happened to contain common salt (NaCl). The walls wasted rapidly, and required fre- quent renewal; and the wasting was not from high temperature, occurring as it did on the inner faces of the oven walls, where the temperature was lowest. The difficulty is emphasised in the coals from certain dis- tricts. South Yorkshire, North Derbyshire, Stafford- shire (Stoke district) are notorious. Wales, on the other hand, is favoured. There has not been sufficient syste- matic geological work done to afford an explanation of these regional peculiarities. Some observations of my own may throw light on certain aspects of the problem as it affects carbonisation industries through the refrac- tory materials used in their plants. Experiments on Coals. In these days, when the washing of coal before coking has become so very general, it would appear at first sight that the soluble common salt would be washed out as a preliminary. But this is not so. The admixture of the salt with the coal substance is apparently very inti- mate, and coal is not porous in any ordinary sense. Even when the coal is finely divided, extraction by water is very imperfect. The following results were obtained with coals which were ground, put through a sieve 12 meshes to the inch, and then over one of 50 meshes to the inch to remove dust. The salts were extracted with hot water, and afterwards with 10 per cent, nitric acid. Chlorine Extracted as Chloride from 100 Grms. of Coal. Coal 1. Coal 2. Coal 3. Water extraction ... 0’38 ...Slight trace ... 0’36 4 hrs. 22 hrs. 4 hrs. Acid extractions— 1 0’22 ...Slight trace ... 0’148 18 hrs. 23 hrs. 18 hrs. 2 0-116 — 0’068 6 hrs. 6 hrs. 3 0’054 — 0-028 29 hrs. 29 hrs. 4 0’024 — 0-014 40 hrs. 40 hrs. 5 Traces — Traces 2 days. 2 days. Total extraction 0’794 Trace 0 618 The slowness of extraction, and particularly the incompleteness of the first extraction, even when the coal was finely divided, and the time of contact extended over hours, are significant. Analyses of Salts. Analyses of the soluble salts extracted by water from finely divided coal have given :— Extract from 100 Grms. SO3. Cl. Ca. Mg. Na(K). Total- Coal A .... 0-0064... 0 121... 0’014 ... 0-0034... 0-165... 0*3098 CoalB...... 0’0125... 0’352... 0’0514.. 0’013 ... 0'161... 0’5899 It is evident that in these cases sodium chloride was the dominant soluble constituent; and since soda would not be volatilised from any such insoluble salt as a soda feldspar during carbonisation, it is permissible to restrict the consideration of the influences arising from the volatilisation of sodium salts to this constituent. Salt goes, then, into the oven with the coal, and experience shows that, if the charged coal contains more than a small quantity (say, 0’05 per cent.) of NaCl, the oven walls suffer, though a rather larger amount does very little damage if the temperature of coking is low. When the temperature of the oven rises, the salt begins to volatilise. At 800 degs. Cent., sodium chloride melts; but it volatilises in some quantity at lower temperatures. A recent laboratory experiment showed that at 800 degs. Cent’, the-volatilisation occurred at a rate of 6’9 per cent, per hour, or with relation to the surface of sodium chloride exposed at a rate of 0'17 grm. per sq. in. per hour, or nearly 1 oz. per sq. ft. Analysis of the residue showed it to consist of sodium and chlorine in the ratio proper to NaCl; so that volatilisation had occurred in the same ratio. It was a pure volatilisation of salt as such, neglecting dissociation which would be negligible at this temperature. Calcium chloride, on the other hand, heated under the same .conditions, did not volatilise. It lost chlorine, but no calcium. A mixture of lime and sodium chloride simply lost sodium chloride, i.e., sodium equivalent to chlorine, at a rate roughly the same as with sodium chloride alone. The following table indicates the analytical results on which these conclusions are based. Apart from their bearing on the volatilisation of salt, the results with the lime-salt mixture would oppose the idea sometimes expressed that the addition of lime fixes to an appreciable extent the chlorine of any sodium chloride present, and so lessens the ammonium chloride vapour in the coke oven gas. In this respect they agree with my own experience. * From a paper read before the Coke Oven Managers’ Association. Losses on Heating for Two Hours at 800 Degs. Cent. Material used. W eight of material. Loss of weight in gms. Analysis of residue after heating. Per cent, loss on heating- on original weight. Sodium... Chloride. NaCl 2-5 gms. 0’346 Gms. Na. Gms. Ca. Gms. cv 1’27 13’82 5’8* 0’83 equivalent to 1’28 gms. Cl. 1’19 equivalent to 3’31 CaCl3. Calcium . Chloride. CaCl3 ... 3’30 gms. containing 2’11 gms. Cl3. O’J91 1’62 CaO ) NaClj - fl’25 gms 11’25 gms. 0’142$ 0’42 equivalent to 0’65 Cl. 0’89 equivalent to 1’25 CaO. 0’64 * 11’6 per cent, of the NaCl. Volatilised Salt and Fireclay Walls. With regard to the interaction that occurs between the volatilised salt and the fireclay walls of the oven, it may be stated that the action—even repeated action—■ of sodium chloride on the surface of fireclay has not necessarily any very detrimental effect. In the manu- facture of what are known as salt-glazed bricks, the walls and arches in the kilns are subjected repeatedly, over many years, to an atmosphere very much richer in sodium chloride vapour than that of 'the coke oven during carbonisation. What happens is that the fire- brick fireclay surface is attacked, and a glaze is formed of sodium-aluminium silicate—the so-called salt-glaze. This coating having once been applied to the whole exposed face, any further attack of salt vapour proceeds slowly, and with little or no penetration to any part of the firebrick which lies at the back of the face in the interior of the brick. The essential condition for securing this comparative immunity from further destruction is that the salt vapour shall not come into contact with the fireclay face until the temperature of the latter is so high that any compound formed is in a state of fusion. This tempera- ture is about 1,200 degs. Cent. At temperatures below this, the result is quite different, and, in its practical effect, much more serious. It has not been realised in the past with sufficient clearness that combination can occur between substances which have for each other sufficient chemical affinity, even when both substances are in the solid state. The combination of lime with silica, or with alumina, at a red heat far below the melt- ing point of either combining constituent, or of the com- pound formed, is perhaps the most striking example of this phenomenon, and has been completely demon- strated.* So far as concerns the formation of sodium silicate, sodium aluminate, or sodium-aluminium silicate, the same general principle applies. The interaction of soda and alumina has been traced by experiments, in which a mixture of sodium carbonate alumina and silica corre- sponding with the formula—■ Na2CO3 + A12O3 + 10 SiO2 was heated to successively higher temperatures, and the degree of interaction determined at each of these temperatures by suitable chemical examination. The results are given in the next table. Soda, alumina, and silica (Na2CO» + A12O3 + 10 SiO2. Na2Co3 + A12O3 + 10 SiO2 contains 7’7 per cent. 12’6 74’2 5*5 I Na2O NaoO + ALO3 + 10 SiOo contains AL03 (SiOoJ 8'1 per cent. 13-4 78’5 Where heated. Tempera- ture DegspC. Soluble in HC1. SiO2 in Na2CO3. SiO2 total soluble. Total soluble. Total insoluble. Appearance. Notes. Na2O. Al2Oh. SiO2. Muffle 1 Blank 800 6’1 6’9 0-2 0’3 0’5 3’6 0’3 4’1 6’4 11’2 93’6 88’8 Hardened Na2O less than theo- retical 1 1,000 0’8 1’0 0’6 0’9 i D5 3’3 96-7# Hardened * By difference >> Works furnace 1 1 1,150 1,300 0’6 0’6 0’8 1-8 0’6 1’3 0-8 0’7 1’4 2’0 2’8 4’4 96 7f 97’2 ; 95’6 97’8* ; 97-2f 1 Very hard, not fused Stony fusion t Determined Na2CO3 24 1,150 i-o 0-4 0-2 0’6 0’8 2’2 Stony fusion Slightly honey- combed * By difference f Determined SO3 = 0’4 per cent. i — It was clear that chemical interaction had occurred at comparatively low temperatures, and that compounds were formed, not in fusion at the time of their forma- tion, but ready to fuse when higher temperatures were attained. Applying this result to our analogous case, we may be quite prepared to find that, below the tem- perature of salt-glazing, there is not only a penetration of salt vapour into the interior of the porous brickwork, but a fixation‘of the salt which has so penetrated by chemical combination with the alumina and silica present. The following experiment indicates the amount of sodium chloride that can be fixed in water in soluble compounds, when extensive facilities for interaction between common salt and fireclay are afforded. Finely- ground calcined fireclay was mixed with a considerable excess of common salt, and was gradually brought up to a temperature of 1,200 degs. Cent, in a works furnace, the time occupied being 72 hours. The sintered mass, from which most of the sodium chloride had evaporated, was extracted repeatedly with water, until the washings * Journal of the Society of Chemical Industry. “ Synthesis of a Silicate.” Cobb, 1910, p. 250. contained only traces of soluble salt; and the residue was then analysed. Its formula was very approximately Na2O, 2A12O3, 7 Si02, indicating something like the same degree of fixation of soda as found in feldspar. Deduction from laboratory experiments carries us so far; but observation of another kind in large-scale w’ork has shown more. In the ordinary type of salt-glazing kiln, salt is not applied to the fires, and therefore does not reach the interior of the kiln, until the fusing temperature of the resulting glaze has been practically attained; and the excess salt fumes pass straight to the chimney. In a continuous gas-fired salt-kiln which was some years ago in my charge, some of the salt fumes passed through from the chamber in which salt-glazing was being effected to another chamber in which the temperature was lower. Examination of a test brick from this second chamber, when the salting in the first chamber was completed, disclosed no obvious change other than the formation of a slight brownish stain on the yellow face of the brick. When, however, the temperature of the second chamber was raised to make it ready for salting, the face of most of the bricks began to blister badly; and for some distance below the blister- ing the fireclay was semi-vitrified and honeycombed. Chemical examination showed that soda was present in quantity. In this case, then, as in ordinary salt-glazing, the interaction of the salt and fireclay had resulted in com- bination ; but the practical difference of a chamberful of bricks blistered and honeycombed, on the one hand, and of smoothly glazed, exactly-shaped bricks, on the other, was a clear indication that the conditions in the first case—not merely the presence of salt and fireclay —were the determining cause of the harmful effect. These conditions determined the penetration of the porous brick by the salt vapour below the temperature at which a glaze could form, the formation in the solid state of compounds between the soda and the brick substance, and, finally, the fritting and fusing in the mass, as well as on the face, when the temperature was sufficiently’raised at a later stage. From this point of view, the destruction of the wall in a coke oven by the salt in the coal carbonised takes place somewhat as follows : At a comparatively low temperature, the salt in the atmosphere of coal gas volatilises, and, penetrating into the brickwork, com- bines with its substance. In so doing, it alters, both physically and chemically, the composition of the portion of the wall nearest the face. The temperature of the brickwork is lowest near the face, highest near the com- bustion flues, and rises gradually from one to the other. The temperature near the face is not sufficiently high to fuse the soda-alumina-silica compounds formed there. But these compounds have also formed further in, where the temperature is higher; and tracing the temperature towards the inside of the brick, a point is reached where the temperature is high > enough to effect fusion — to produce liquidity, porosity, honeycombing, and rotten- ness of structure making for disintegration. Herein lies the danger. The brickwork begins to fall away in pieces—to “ chatter.” It is destroyed by a drastic process of breaking off layers of considerable dimensions. It is very likely several other phenomena also come into play. The chemical alteration in that portion of the firebrick into which the salt vapour has penetrated, is accompanied by physical changes in the coefficient of expansion. If the outer layer is expanding and contracting every time the oven is heated and cooled, to an appreciably different extent from the sub- stance at the back of it, this is, in itself, a disintegrating force, and the somewhat vitrified, and therefore more coherent, facial portion of the brick may tear away as a whole from the entirely unvitrified, and therefore less coherent, substances behind it. Another phenomenon which may be of some conse- quence is the volatilisation of iron in the form of chloride of iron, followed by the deposition of iron— probably in the metal form, because the atmosphere is reducing—when the volatilised chloride penetrates to a region where' the temperature is too high to permit of its continued existence; or the ferric chloride may penetrate to a point at which it meets with free oxygen from the combustion flue, and oxide of iron is deposited. A layer of oxide is frequently found on the plane of disintegration, when the brickwork of a failing oven is examined after cooling off. Laboratory experiments carried out by Mr. C. P. Finn demonstrated conclusively the volatilisation of iron, as chloride, from fireclay, and its after-deposition as oxide under proper conditions. I have noticed the same thing occurring on a large scale when a stream of salt vapour impinges on the bricks in the process of salt-glazing,