CHEMICAL NEWS.}

Jan. 10, 1918

THE

TRINITROTOLUENE RESIDUES AND THEIR

UTILISATION.

"ISO-TROTYL," "LIQUID T.N.T." CHLORPICRIN, AND SULPHIDE DYES.*

By MAURICE COPISAROW.

THE recent development of a great war industry-the manufacture of trinitrotoluene-has given rise to several problems, closely connected with one another, the most outstanding being :

(1) Conditions of nitration;

(2) Methods of purification; and (3) Residues and their disposal. The first two having been already dealt with by the author (CHEMICAL NEWS, 1915, cxii., 247, 283; 1916, cxiii., 37) it is intended now to deal with the problem of residues, which, owing to their gradual accumulation, require some method of utilisation.

The problems of residues, their recovery and disposal, can be regarded as part of the greater problem of the efficient and economical manufacture of trinitrotoluene. The residues are derived chiefly from two sources: (1) Those accompanying in solution or suspension the used-up acids, and

(2) Those obtained in the process of purification of trinitrotoluene by recrystallisation or washing.

The used up acid is usually considerably diluted to precipitate as much as possible of the dissolved trinitrotoluene. This procedure when carried beyond certain limits must be regarded as wasteful. as, first, a portion of the nitro bodies remain invariably dissolved or sus pended in the dilute acid, no matter what the dilution may be, and, secondly, the recovery of arid becomes a huge task involving the use of a large number of sulphuric acid

concentrators.

A better method seems to be to dilute the acid just enough for the separation and convenient removal of the bulk of the nitro body and then to extract the used-up acid with toluene. This gives a fairly concentrated acid, practically free from nitro bodies, and a toluene extract, which is transferred into the nitrators. Leitch's English Patent 15455, 1915, seems to be very much of this kind. The extraction of the nitro bodies with a solvent from the used-up acids is, of course, only necessary in case of acids to be concentrated or used for different purposes.

Whilst in the case of residues accompanying the used-up acid the problem is confined simply to their recovery, in case of residues derived from the purification processes the problem is their utilisation. Among the several methods of purification of crude T.N.T. the alcoholic treatment occupies an important position; the alcohol being applied either as a solvent and crystallising medium or as a washing agent for the crude T.N.T.

The latter method yielding a somewhat inferior product as compared with the former, gains much favour owing to the considerable technical advantage it presents, the more so as the purity of the product thus obtained answers all ordinary requirements.

On removing the alcohol from the mother-liquor in the first case and the washings in the second, a dark brown thick liquid is obtained.

This residue is slightly soluble in water, giving a

This paper, ready for publication early in 1916 and intended to follow directly the papers previously published (CHKMICAL NEWS, 1915, pp. 247, 283; 1916, p. 37), had to be kept back owing to war considerations.

13

yellowish solution which, when treated with caustic soda, gives a red coloration characteristic of a-trinitrotoluene and some other nitro aromatic bodies accompanying it.

On prolonged distillation in a current of steam the distillate consists of a yellowish aqueous solution, no oil or solid matter being present, this indicating that the residue, whilst being to some extent volatile in steam, does not contain toluene or mononitro toluenes.

The percentage of volatile matter (alcohol, water, &c.) in the residue was determined by heating a weighed quantity of the residue in a shallow dish to constant weight at 97-99° C. and noting the loss in weight.

The matter insoluble in benzene (oxidation and condensation products of the polynitrotoluenes as well as inorganic matter) was determined by extracting a weighed quantity of the residue with twelve to fifteen times its weight of benzene, and noting the weight of solid residue. The quantitative nitrogen determination was carried out with samples carefully freed from volatile matter and matter insoluble in benzene; the benzene extract being freed from benzene and kept for ten hours at 97-98° C. The results obtained are as follows:Specific gravity of the residue.. Volatile matter

Matter insoluble in benzene Percentage of nitrogen in a purified sample.. ..

1'47 0'77 0.81 per cent I'35 I'40 11

1'50 (20° C.)

15'95-16'01

Assuming the purified residue to consist of di- and trinitrotoluenes only (nitrogen in dinitrotoluenes, 15'4 per cent; nitrogen in trinitrotoluenes, 18.5 per cent), we find from the equation

0185x+0 154 (100 - x) = 15 95 (or 16:01) that the purified sample contains:

82.26 80 32 per cent 17.74 19.68 19

Dinitrotoluenes Trinitrotoluenes The residue on long standing at the ordinary temperature gradually crystallises to a semi-solid mass, which can by means of a filter-press or centrifuge be separated into two parts:

(a) Brown solid containing 1'45-155 per cent matter insoluble in benzene, and (b) a liquid resembling the original residue.

The proportion of solid matter so obtained to the total varies appreciably, depending upon the conditions of the alcoholic treatment and the duration and range of cooling, but it is seldom below 18 per cent.

The author has investigated three methods of utilisation of the T.N.T. residues :

(1) As an explosive;

(2) As various means of warfare; and (3) As dyes.

With the object of investigating the application of the residues and their products as explosives the following experiments were carried out :

1. Nitrating the crude residue at 100-120° C. with a mixture of sulphuric and nitric acid containing 6 per cent H2O and 15-17 per cent HNO3, a light yellow opaque liquid (specific gravity, 149-151 at 20° C.) is obtained containing 48:4-64'5 per cent of trinitrotoluenes (nitrogen, 16 9-17'4 per cent), and forming a gelatinous mass with gun-cotton. Considering the high percentage of nitrogen and the property of dissolving gun-cotton this liquid must be regarded as a "liquid T.N.T." of a high standard suitable for the preparation of gelatinous dynamite to be favourably compared with the Notel's explosive (D.R.P. 24235, 264503).

2. Nitrating in a manner described above a sample free from volatile and insoluble (in benzene) matter, the product is found to be practically identical as in No. 1. This method, whilst leading to a certain economy in acid, owing to diminished oxidation, involves much extra work.

3. Nitrating product from No. 1, 2, or the residue direct with a mixed acid containing 4 per cent H2O and 17 per

cent HNO3 at 120° C., a slightly yellowish solid, resembling good quality crude trinitrotoluene, is obtained. Whilst setting at 57-58° C. the nitrogen determination of this substance indicates it to contain at least 98 per cent trinitrotoluenes (N = 18.44 per cent). Therefore this product must be regarded as a mixture of trinitrotoluene isomers. As to the actual state of the isomers in the product the fact that a-T.N.T. forms additive compounds in the solid state not only with hydrocarbons, phenols, and amines, but also with nitro-compounds, such as mono-nitro- and di-nitrotoluenes (Lepsius, Chem. Zeit., 1896, xx., 839; Giua, Ber., 1914, xlvii., 1718) suggests the possibility of the product consisting of a mixture of additive compounds rather than a mixture of free isomers (see "On the Theory of the Formation of Additive Molecular Compounds by Nitro-bodies," Pfeiffer, Ann., 1916, 412, 253),

Noting the fact that a-, 3-, and y-trinitrotoluenes and probably all other modifications are practically of the same value as explosives (Will, Ber., 1914, xlvii., 704), this low melting explosive, which may be conveniently termed "iso-trotyl," must be regarded as a useful and powerful constituent for blasting powders like bellite and roburite, and for war purposes, especially in cases where a lowering of the melting point is required as in the case of trinitroxylenes.

4. The method of dissolving the residue in toluene, filtering, and then nitrating the filtrate in two or three stages is found to be unsatisfactory, as a portion of the nitrated product remains liquid (liquid similar to that described in No. 1 and 2), which is probably due to the fact that the dinitrotoluenes derived from m-mononitrotoluene are less readily nitrated than those formed from the o- and p-modifications, and also to the gradual weakening of the mixed acid as the nitration proceeds.

5. Submitting the solid and filtrate mentioned above to an exhaustive nitration it was found that both gave ultimately the same solid product setting at 57.58° C. as in No. 3.

The identity of the products of the nitration of the solid and filtrate and the fact that the nitration of the solid proceeds more readily, requiring less acid, suggest that the solid represents a higher stage of nitration than the filtrate, both leading ultimately on nitration to a mixture

of isomeric trinitrotoluenes.

It appears therefore that

1. No special purification of the residue, preliminary to further nitration, is to be recommended, as the slight excess of acid used up with the crude residue is more than compensated by the saving of extra work involved in the process of purification.

The matter insoluble in benzene present in the residue is only to a small extent transferred to the nitration product, the bulk of impurities being either oxidised or dissolved.

2. "Liquid T.N.T." prepared from the residue can be used with advantage for the preparation of gelatinous dynamite.

3. "Iso-trotyl" prepared from the residue can be usefully employed for blasting powders, or as a valuable ingredient for lowering the fusion point of explosives having a high melting point.

4. Should both the "liquid T.N.T." and the "isotrotyl" be in demand, the most economical way of preparing them would be to cool the crude residue, separate the solid, and then nitrate the solid to "iso-trotyl" and the filtrate to "liquid T.N.T."

Under the heading of utilisation for various means of warfare the author investigated the employment of the residues for the production of chlorpicrin, CC13(NO2). Applying Hofmann's method of preparing chlopicrin, originally worked out for picric acid (Ann., 1866, cxxxix., III), i.e., the treatment of the nitro-body with bleaching powder, it was found that chlorpicrin could be obtained by treating the T.N.T. residues with bleaching powder; the yield compares favourably with that obtained by the

degradation of picric acid residues, allowing for the percentage of nitrogen.

The T.N.T. residues can be also usefully employed for the making of dyes. Fusing the residues or their nitration products with sodium sulphide, with or without the addition of sulphur, dyes are obtained which dye unmordanted cotton fast colours varying from grey to khaki and brown; the particular colour depending not only on the concentration of the dye-bath but also on the temperature and duration of fusion and the addition of sulphur. These methods of disposal of the T.N.T. residues, whilst in no way exhausting the methods of their possible utilisation, suffice to indicate that in the manufacture of trinitrotoluene there are no waste products in the actual sense of the word, as even the nitrous fumes so abundantly given off during the nitration can be easily recovered.

WATER SOFTENING.

By P. E. KING.

THE hardness in water is due to soluble salts of calcium and magnesium; "temporary hardness" to carbonates of lime and magnesia, held in solution by carbonic acid; permanent hardness" to sulphates, nitrates, and chlorides of these metals, which are soluble of themselves. The greater part of the temporary hardness can be removed by boiling the water, but except in some cases where in deoiling plants exhaust steam is used for this purpose this plan is not applicable. The reaction involved is: CaCO3. H2CO3=CaCO3+CO2+H2O. Soluble. Insoluble.

The carbonic acid gas holding the carbonate in solution is driven off, and the carbonate is thrown out of solution. The permanent hardness must be removed by chemical treatment; i.e., is not removed by boiling under This distinction between ordinary atmospheric pressure. temporary and permanent hardness fails altogether as soon as the water enters a boiler, for at the very high temperature of a boiler under high steaming pressure the chloride, inert while cold, becomes actively corrosive. calcium sulphate becomes insoluble and the magnesium

The methods of water softening may be divided into two distinct groups, the first of which can be sub-divided into three :

Group I., (a) the lime processes which is used in cases where the hardness is only temporary or mainly so. (b) The soda process where either carbonate of soda or caustic soda is used.

(c) The lime and soda process, which is the method in general use.

Group II.-In the second group is found the most recent method of softening water; i.e., the Permutit system :

In Group I. the chemical reactions involved in the removal of the temporary hardness by lime are:CaCO3.H2CO3+Ca(OH)2=2CaCO3+2H2O.

Soluble.

Soluble

Insoluble.

MgCO3. H2CO3+Ca(OH)2 = MgCO3+CaCO2+2H2O. Slightly soluble. and with more lime the magnesium carbonate still in solution is converted into hydrate, which is practically

nsoluble.

MgSO4+2NaOH=Mg(OH)2+Na2SO4, CaSO4+2NaOH=Ca(OH)2+Na2SO4 ; i.e., calcium sulphate is not removed by caustic soda, but it free CO2 or bicarbonates be present sodium carbonate will be formed, and this will remove the calcium sulphate (see equation above). Caustic soda is only suitable under exceptional circumstances when the permanent hardness is chemically equivalent to the temporary hardness. Thus, by a suitable combination of reagents, waters of varying hardness and of varying composition can be successfully

softened.

Before proceeding to a description of the plants for water softening it should be pointed out that however efficient a lime or lime and soda plant may be, it cannot entirely remove all hardness from the water, this being due partly to the solubility of calcium carbonate and partly to excess of lime left in the softened water. At the very best the hardness is reduced to 2° Clark, or 3 parts per 100,000, but generally the hardness is nearer 5o.

Plants for Water Softening.

In 1841 Clark invented a process of treating water with lime to remove carbonic acid and bicarbonates of lime and magnesia. Shortly after a process of removing sulphates of lime and magnesia was invented by Porter, and a combination of these two methods is now called the Porter Clark process. Clark's plant consisted of three settling tanks, one for mixing the hard water and lime water, a second one from which softened water is being drawn, and a third kept in reserve.

The Porter-Clark process consists in the application of machinery to the ordinary Clark's process.

The Gaillet and Huet (also known as the Stanhope) process is a continuous one, lime water and sodium carbonate being employed as reagents. The clear saturated lime water is made by passing a certain definite proportion of the incoming hard water through the milk of lime contained at the bottom of the lime saturator. The saturated lime water containing solid particles in suspension gradually rises in the saturator, the solid particles meanwhile gradually falling down and clear saturated lime water is discharged at the top. The lime water is mixed with the hard water, to which also a definite quantity of soda solution is added from a separate tank, and the mixture is led to the bottom of the clarifying tower. This contains diaphragms fixed to alternate sides. As the water rises slowly in the tower the precipitated salts settle out on the diaphragms, slide down to their lower ends, and are there removed by suitable sludge cocks. The last traces of suspended matter are removed by means of a wood-wool filter placed in the upper part of the tower.

The Lassen and Hjort apparatus is one of the most efficient plants of the present day, and differs from most softening plants in its method of adding the required

15

chemicals to the water. The hard water to be softened is led info the plant by a pipe which alternately fills each compartment of a two-chambered tipper oscillating on a shaft carried in bearings. When one of these compartments is full of water the disturbance of equilibrium causes the tipper to overbalance and discharge its contents of water into the tank in which it is suspended. At the same time the other compartment of the t'pper is brought under the orifice of the inlet pipe, and filled in its turn with hard water to be discharged in the same manner when full. At each discharge of water from the tipper a definite quantity of reagent is discharged from the semi circular chemical container affixed to the side of the tipper tank. This is effected by means of the positive discharge valve placed in the bottom of the chemical container, which is opened at each movement of the tipper and caused to deliver into the reaction chamber the exact quantity of milk of lime and sodium carbonate. softening reagent required. The chemicals are generally The valve can be easily adjusted to discharge any specified quantity of reagent. The mixture of hard water and reagent passes to the bottom of the reaction chamber, where the greater part of the precipitated salts settle out, and the water rising in the chamber finally passes through a wood-wool filter before passing to the storage tank.

Another type of plant is the Paterson Osilameter By-pass water is by-passed through the "Osilameter." This is a type, in which a certain proportion of the incoming hard two-chambered tipper operating somewhat similarly to that described under "Lassen and Hjort" apparatus. The main quantity of hard water drives a water-wheel, which works agitating paddles in the chemical tank. These agitators are fitted with cups which fill the reagent measuring bucket, the excess continually flowing back into reagent bucket, the contents of which mix with both the the tank. The tipping of the Osilameter overturns the water from the Osilameter and the main supply from the is similar to that described before. water-wheel. Further purification of the softened water

Other water softening plants are those of Kennicott, Boby, Royle or Reisert, Bell Bros., and Archbutt and Deeley. These are all lime or lime and soda plants, and d fler only in mechanical working from those previously described. The Archbutt and Deeley plant is, however, not continuous, and differs from the others in that the softened water is finally carbonated. This is to neutralise the free alkali, and thus prevent the deposit of further sludge through the further action of the alkali on any remaining hardness.

[L. Archbutt states that the chief action of re-carbonating is the conversion of Mg(OH)2 into MgCO3 (Proc. Inst. Mech. Eng., 1898, p. 404; Abs. in Fourn. Soc. Chem. Ind., 1899, p. 292)].

Plants operating with saturated lime water are becoming obsolete, because as from 10 to 25 per cent of the total supply of water has to be by-passed through the lime saturator the plants have to be very large. It is also difficult to get constant saturation on large plants. Again, many waters contain impurities which coat the lime in the saturator and render it ineffective.

GROUP II.-Permutit Process.

This process is comparatively modern, but its use is rapidly extending on account of the simplicity of its action and the softness of the water produced. It is a process of softening water without the addition of chemicals or the formation of sludge. The hard water, either crude or partially softened, is passed through a bed of granular material called "Permutit." The process is not one of ordinary filtration; i.e., no solid matter is thrown out as a precipitate, and it is, in fact, most essential that the water be perfectly clear before entering the plant.

Permutit is a sodium aluminium silicate corresponding to the formula 3S 02.Al2O3. Na2O.6H2O, and to-day is manufactured at Brentford by the interaction of sodium silicate and sodium aluminate. It is marketed as hard

Na2Pm+CaCO3.H2CO3-2NaHCO3+CaPm,
Na2Pm+MgSO4 = Na2SO4+MgPm.

Similar equations will represent the reactions for mag nesium bicarbonate and calcium sulphate; i.., the bicarbonates of calcium and magnesium are converted into sodium bicarbonates, the sulphates of these metals into sodium sulphate. All traces of calcium and magnesium are removed and a water of zero hardness results. By long continued filtration of the hard water the whole of the sodium can be replaced by calcium or magnesium, or the two together, but this is not done in practice as zero water would not be obtained throughout the process. According to the hardness and the quantity of permutit employed a certain definite quantity of hard water is passed through before the process of regenerating the permutit begins. It has been found by experience that water containing much magnesium salts requires a larger proportion of permutit than one only containing salts of calcium.

The regeneration or revivification of the permutit is accomplished by a solution of common salt, which is employed in 10 per cent strength, and should be free from magnesium chloride. The following equations represent the reactions involved during regeneration :

Ca Pm+2NaCl = CaCl2+ Na2Pm,
MgPm+2NaCl = MgCl2+ Na2Pm.

Sodium permutit is regenerated and the calcium and magnesium chlorides which are formed are soluble and washed away to the drain by the wash-water. It will be noticed that no sludge is formed. The method of working is as follows::

Before introducing the salt solution the permutit is agitated by causing the water to enter the bottom of the filter. This loosens the permutit and removes any air channels which may have been formed during the softening process. After agitating for a few minutes the salt solution is allowed to flow into the permutit from the top and to percolate slowly through, the speed being arranged so that regeneration lasts about ten hours. The salt solution remaining in the filter is now washed out, and after completion the softening recommences.

In cases where the water to be softened is very hard and where it is required to have a water of zero hardness, the ideal system would be to remove the greater part of the hardness by a lime and soda plant, and the remaining hardness by the permutit system.

Although there are to-day in this country plants operating successfully by this method, there are others which have gone wrong after a period of years. The permutit has gradually become contaminated with chalk, and ultimately fails to soften the water. The cause of this is undoubtedly due in the first place to the alkalinity of the water coming from the lime and soda plant. This alkalinity, due to free lime, sodium carbonate, or caustic soda, is unavoidable, and can only be kept within limits by careful working. It affects the permutit in one or more ways.

1. Free lime passing through the permutit will produce caustic soda which will gradually dissolve it.

2. Sodium carbonate passing through the permutit already containing some calcium permutit will produce sodium permutit and calcium carbonate, the latter sometimes passing onward and giving the water an opalescent appearance,

3. Free alkali causes a further deposit of chalk to occur, this being formed from the residual hardness not completely removed in the preliminary softening. It is well known that water, even after filtration through wood-wool,

will on further filtration through sand deposit a further amount of chalk; i.e., the water is further softened. The sand acts as an accelerator of the softening action; i.e., it brings about the same effect as allowing the water to further stand for some hours. The permutit undoubtedly acts in the same way, only possibly in a more energetic manner, owing to its porous nature.

It is the opinion of the author that the reason of permutit failing is to a greater extent due to the action mentioned under (3) rather than (1) and (2), because in those cases where permutit works successfully after previous softening, the water has a more or less long period of rest before it enters the permutit system. This period would allow the reaction mentioned under (3) to take place before the water entered the permutit plant. The author is further led to this belief because the water in question did not enter the permutit plant for two to three days after previous softening with lime and yet was stili alkaline.

A method has been tried for overcoming this difficulty of the alkaline water. Calcium permutit possesses the property of absorbing free lime, the compound supposed to be formed being represented as CaPm CaO. The calcium permutit is regenerated by passing hard water through in the reverse direction, the lime softening this water and being removed as chalk which is washed away to the drain. Although this method works well for a short period the calcium permutit loses its efficiency, and free lime again passes through without being absorbed.

Assuming that a method will be found to overcome this difficulty of alkaline water it is probable that a plant consisting of three separate systems; i.e., lime and soda plant, followed by, say, calcium permutit, followed by sodium permutit, would be more expensive to work than a sodium permutit system operating directly on the crude hard water.-Society of Dyers and Colourists, xxxiv., No. 12.

THE CONFIGURATIONS OF ORGANIC COMPOUNDS AND THEIR RELATION ΤΟ CHEMICAL AND PHYSICAL PROPERTIES.* By ARTHUR MICHAEL.

In this paper the physical and chemical properties of the stereomeric, unsaturated acids will be co-ordinated with the configurations given in the first communication (1), as far as the experimeutal observations at hand permit, with a critical revision of previous views in this field.

Owing to the imperfect and inadequate experimental data, a perfectly satisfactory treatment of the subject is systematic investigations, not infrequently the members not possible at present; even in the few existing of the series most important from the viewpoint of theory were omitted, as beyond the scope of researches.

The Relations between the Physical Properties and the Configurations of Unsaturated Acids.

Density. The removal of hydrogen from contiguous carbon atoms proceeds with their greater segmentation (2), which should augment the increase in gravity already produced by the elimination of the lightest known matter. This change should be in a direct relation to the number of the removed atoms, provided substances with similar configurations and in the same or a strictly comparable state of molecular aggregation are compared (3).

That such an increase takes place with unsaturation was shown by Buff (4), and was confirmed by the later work of Schiff (5), Lossen (6), Zander (7), and Weger (8), who found that the values of the increments vary considerably in different substances. A reinvestigation of the subject is necessary, however, as the conclusions of

From the Journal of the American Chemical Society, xl., No. 11.