VOL. CXXII., No. 3188.

FRACTIONAL DISTILLATION WITH CONTACT RING STILL-HEADS.*

By Dr. R. LESSING.

THE author draws attention to the advantages of packing absorption towers and distilling columns with contact rings of the type described in his Brit. Pat. No. 139,880. These rings consist of cylinders of approximately equal height and diameter, having a gap in the circumference and a more or less diametrical partition connected on one side with the cylinder, but out of touch on the opposite side. They are disposed indiscriminately in the tower or column at the angles which they find when dropped promiscuously into a vessel. The provision of the central partition and openings has been found to increase the efficiency of this packing over Raschig rings by increasing the contact surface available in a given tower space and the drainage capacity.

in.

The standard rings of in, diameter by height are used in a great variety of plants of the chemical and gas industries, amongst which may be cited ammonia scrubbers, absorbers, washers, tar extractors, carbolate decomposers, carbonating and decarbonating towers, debenzolising stills, distilling columns and still-heads for many substances such as tar and petroleum products, alcohols and essential oils, condensers, coolers, heat exchangers, and are applicable wherever large interfacial surface for liquid and gas or vapour, combined with a minimum of back pressure is essential.

Smaller sizes of Lessing contact rings down to in. diameter are now available, and are particularly useful for the construction of laboratory still-heads on the lines on which the larger rings are applied in manufacturing operations. They have a surface of 8250 sq. cm. per litre (251 sq. feet. per cubic foot), and only occupy 13 per cent total space, leaving 87 per cent of free space for vapour and condensate. The resistance of such still-heads is almost negligible and they can therefore be made of considerable height.

They are simple in construction, and the rings can be transferred from one column to another, or be used in absorption towers or reaction vessels, and are therefore of general applicability. They have a remarkable fractionating efficiency, and are in this respect superior to other dephlegmators. This efficiency is due to the large surface per unit of volume, the turbulence created by frequent deflection of currents of vapour and condensate, and the consequent absence of channeling.

In order to obtain the maximum efficiency, the author recommmends lagging or even heating of the still-head to avoid undue condensation, and to provide a reflux condenser, or to use a definite length of the still-head itself as reflux condenser. Thus, return flow and irrigation of the whole length is obtained. A ring packed still-head may then be regarded as a scrubber in which the ascending vapours are scrubbed with the descending condensate.

Read at the Meeting of the London Section of the Society of Chemical Industry, Monday, May 2, 1921.

229

Still-heads can be constructed from plain glass tubes or metal pipes by merely dropping the rings into them.

The paper contains a number of graphs proving the efficiency of fractionisation obtainable. From a mixture of equal portions of benzene and toluene 47 per cent of benzene distilled within less than 1°, and a like portion of pure toluene was yielded, leaving only about 5 per cent of the total volume in the mixed middle fraction.

From a crude coke oven benzol, almost all the benzene and toluene could be distilled off each within o2° C., although the sample contained 30 per cent of naphthalene and wash oil. The effect of still-head dimensions, heat insulation and reflux condensing is shown in one set of curves. Another chart deals with the distillation of fusel oil, and a remarkably sharp separation of the various higher alcohols from each other and from water is shown in a further graph. The fractionating effect is so sharp that chemically anhydrous alcohol may be in the lower part of the still-head, while there is sufficient water in the upper portion to separate as a distinct layer from the distillate.

METHODS FOR THE COMPLETE RECOVERY OF NITRE CAKE.*

By W. H. H. NORRIS, B.A., B.Sc.

PREVIOUS to 1915 the utilisation of nitre cake in this country was on a comparatively small scale. During the war over 80 per cent of the considerably increased production was employed, but largely by methods which completely wasted the sodium sulphate contents. With ample supplies of sulphuric acid, only those methods of recovery will stand which are economically sound, and the production of by-product salt cake in a pure and marketable form will assist in this direction.

Processes of war expediency may be passed over. The oldest method for complete recovery was by the salt cake furnace, and this well-known operation calls for little comment. Expenditure on hard coke at present prices is a considerable charge, and where hydrochloric acid is the valuable product, nitre cake has little chance of competing with sulphuric acid because of the reduced yield per furnace, and the impurity of acid produced when nitre cake is employed.

Many other methods have been suggested, but the writer does not know that any of these have been carried out on a manufacturing scale with the exception of the following.

Nitre cake may be used for ammonia absorption producing sulphates of ammonia with the additional formation of sodium sulphate suitable for glass makers.

A process was described which has been worked out in conjunction with H. S. Denney and C. W. Bailey, and 220 tons of nitre cake used to produce 157 tons of salt cake and 87 tons of ammonium sulphate by it. The plant was run in conjunction with a Mond Gas Ammonia Recovery Plant, and proved successful in operation, process figures and costs for which were given.

The method is of wide application and may be used either with ammonia absorption towers and

A paper given before the Manchester Section of the Society of Chemical Industry, May 6th, 1921.

dasher washers or in the ordinary saturator as employed on coke oven and gas works liquor recovery plants. The operation works up all its own waste liquors, and requires the minimum of handling and evaporation.

Briefly, the nitre cake is digested with hot process liquor, and the whole of its contents of salt cake are precipitated in a pure form, while all the acid is taken into solution. The mixture is filtered, and washed in centrifugals, and the acid liquor resulting used for ammonia absorption, in the saturator or other absorption plant. The neutralised liquor is evaporated in the usual way, for the deposition of sulphate of ammonia containing over 24 per cent free ammonia, but concentration is checked when a definite density is reached, and the process liquor run back to the digester to extract the acidity from a fresh charge of nitre cake.

THE SEPARATION AND DETECTION OF ARSENATE AND ARSENITE.

By, GEO. W. SEARS.

OF the methods now available for the detection of arsenate and arsenite those based on the formation of arsine (Chem. Ztg., xxxiii., 1209; Z. anal. Chem., 1902, xli., 362) and those taking advantage of the oxidising power of arsenates (Am. J. Sci., [3] 1894, xlviii., 216) or the reducing power of arsenites (ibid., [4] 1896, i., 35) are perhaps the most common. The limited applicability of these methods to a general separation and detection of the more common acid radicals led the author to make a study of the relative solubility of their silver salts in various concentrations of sodium hydroxide. When silver arsenite, to which potassium hydroxide has been added, is allowed to stand in the sunlight or is heated to 50° on the water-bath, the arsenic slowly dissolves with the formation of potassium arsenate (Ber., 1894, xxvii., 1019). No information, however, was available regarding the effect of alkaline solutions other than ammonium hydroxide, on silver arsenate. Both salts are readily soluble in ammonium hydroxide ("Dictionary of Chemical Solubilities, Comey, pp. 37-42). An examination of the two salts with respect to their relative solubility in excess of sodium hydroxide leads to a study of the following equilibria. 2Ag,AsO,+2NaOH+H,02NaH,ASO,+

by drop with constant stirring until the black precipitate of silver oxide just failed to redissolve. The appearance of the black silver oxide is very definite and easy to detect, and hence makes a very satisfactory indicator. Under these conditions the solution has an acidity of about 10- to 10-. It reacts acid toward phenolphthalein and basic toward methyl orange. With congo red an orange to violet colour is obtained. Both silver arsenate and silver arsenite are completely precipitated, and can be filtered with ease. (The filtrate does not give a precipitate with H,S.) The salts obtained in this way and containing a definite weight of the acid radical were removed from the filter by means of a fine stream of water. This resulted in a rather finely divided suspension to which the solvent was added.

Solubility of Silver Arsenite.-Approximately 6 N sodium hydroxide solution was added to the precipitate of silver arsenite suspended in 10 to 20 cc. of water as indicated above, and after a thorough mixing, the mixture was filtered and both precipitate and filtrate tested for arsenic as follows. The precipitate was leached with 6 N hydrochloric acid to dissolve the arsenic and remove the silver. The clear solution was then saturated cold with hydrogen sulphide. The filtrate was strongly acidified with hydrochloric acid and saturated with hydrogen sulphide. Table I. shows the behaviour of silver arsenite toward various concentrations of sodium hydroxide. The residue invariably gave a positive test, and the filtrate a negative test, except in the last experiment, when the filtrate gave a positive test. TABLE I. Cc. of Approx.

Approx. normal

Mg. AsO".

Total volume

6 N NaOH

CC.

strength.

The results obtained show that cold sodium hydroxide up to about 15 N has no solution effect on silver arsenite, though higher concentrations seem to have a slight solvent action. In Reichard's work (Reichard, Ber., 1894, xxvii., 1019) on the action of potassium hydroxide on silver arsenite no statement is made regarding the concentration of potassium hydroxide solution used, hence it seemed advisable to repeat a part of his work, using small concentrations of sodium hydroxide solution. The results obtained from a number of experiments showed that at the temperature of the water-bath silver arsenite is slowly dissolved in 0.5 N sodium hydroxide solution. The reaction proceeds almost entirely according to Equation 3 above, as shown by the fact that only a trace of arsenic could be detected by precipitation with hydrogen sulphide from a cold hydrochloric acid solution, while a considerable amount of metallic silver was left in the

residue. On the other hand, when the mixture was kept cold no reaction was detected in concentrations of sodium hydroxide below 15 N and in higher concentrations (4-6 N) reaction according to Equation 2, seemed to predominate, although the presence of small amounts of metallic silver in the residue showed that some arsenic was dissolved according to Equation 3. Even in 6 N sodium hydroxide solution only a small fraction of the silver arsenite present was dissolved.

Solubility of Silver Arsenate.-In the investigation of the solubility of silver arsenate experiments were carried out in the same manner as those with silver arsenite except that the residue after leaching with hydrochloric acid was heated to boiling before saturating with hydrogen sulphide. The filtrate was acidified with acetic acid and the arsenate precipitated with uranyl acetate The resulting uranyl arsenate was then dissolved in hydrochloric acid and arsenic precipitated with hydrogen sulphide as before. Preliminary experiments showed that some silver arsenate dissolved even in very small concentrations (005 to 01 N) | of sodium hydroxide but seemed to indicate incomplete solution when as much as 5 mg, of the arsenate radical was present. With o5 to 1 mg. of arsenate, solution was complete as shown by a negative test on the residue and a positive test in the filtrate. When 50 to 100 mg. was used a considerable proportion seemed to dissolve, but in all cases the residue gave a positive test. Increasing the concentration of the sodium hydroxide made very little change in these results, while increased dilution of the arsenate seemed to give a greater solution effect. (Later work showed that these results were due in large measure at least to insufficient washing of the silver oxide residue).

The result of preliminary experiments indicated that an equilibrium was established in which perhaps 75 to So per cent of the arsenate was dissolved. It seemed worth while, therefore, to make a more careful study of the equilibrium

2Àg1AsO1+2NaOH+H.02NaH ̧AsO,+3Ag2O

in order to find out if possible the conditions under which it might be shifted to the right and complete solution of the arsenate be obtained. To this end several experiments were carried out as described above.

before filtering. The silver oxide residue was then washed till the wash water gave no further test for alkali with litmus paper.

Table II. shows the behaviour of silver arsenate toward various concentrations of sodium hydroxide. The filtrate gave a positive test in every experiment except the last, which was negative.

It may be seen from the above results that the solvent action of sodium hydroxide on silver arsenate increases very rapidly with increasing concentration, and is appreciable at concentrations as low as o'06 N. No action is apparent, however, in 003 A sodium hydroxide solution.

Discussion.

From the foregoing results it is evident that a quantitative separation of arsenate and arsenite may be obtained by the action of sodium hydroxide solution (05-15 V) on their silver salts, the arsenate passing into solution. Higher concentrations of sodium hydroxide react slowly with silver arsenite with the formation of a mixture of soluble trivalent and pentavalent arsenic, probably according to Equations 2 and 3 above. Even very high concentrations, however, seem to have but slight solvent action in the cold.

It would seem, therefore, that some influence other than is apparent in the equilibrium represented by Equation 1 must come into action to affect the solution of silver arsenate since similar forces seem to act in the opposite direction in Equation 2. A compound of the formula (As2O). 3 NaOH has been described by Filhol and Senderens as being formed by the action of excess sodium hydroxide on arsenic acid (Filhol and Senderens, Compt. Rend., 1881, xciii., 388). They obtained it crystalline form only after evaporation to a syrupy consistency. No evidence of a similar action with arsenious acid was found. This difference in tendency toward complex formation with sodium hydroxide may, therefore, account for the difference in solubility of the silver salts. Without the formation of a complex ion the equilibrium represented in Equation should shift toward the less soluble silver salt as is apparently the case in Equation 2.

Procedure.

(1) Precipitation.-To a nitric acid solution of the acids containing an excess of silver nitrate add sodium hydroxide drop by drop from a pipet until on shaking the dark silver oxide precipitate just fails to redissolve. Filter, wash the precipitate and transfer it to a graduate cylinder by means of 15 to 20 cc. of water from a wash bottle. Add 2 to 3 cc. of approximately 6N sodium hydroxide solution and fill up to the 25 cc. mark with water. The amount of sodium hydroxide that should be added depends on the quantity of arsenate present. More than 450 mg. of arsenate is completely dissolved in 25 cc. of solution containing 3 cc. of 6 N sodium hydroxide solution while as much as 5 cc. of the alkali in 25 cc, of solution produces no solvent action on the arsenite. Transfer the contents of the cylinder to a beaker and mix thoroughly. Filter, and wash the residue free from alkali. Test the residue for arsenite and the filtrate for arsenate.

(2) Detection of Arsenite.-Dissolve the residue of silver arsenite (1) by pouring a 5-cc. portion of 6 N hydrochloric acid repeatedly through the filter and add hydrogen sulphide to the clear solution obtained. If a dark precipitate is obtained show