1610 DIVERS AND HAGA: CVII1.- The Reduction of Nity*ososulphcctes. By EDWARP DIVERS, M. D., F.R.S., and TAMEMASA HAGA, D.Sc. (Japan), F.C.S. IN 1885 (J. Chem. SOC., 47, 203), we studied the action of sodium amalgam on a solution of potassium nitrososulphate, and found that it produced sodium hyponitrite and sulphite, besides nitrous oxide,THE REDUCTION OF KlTKOSOSULPHATES. 1611 hydroxylamine, and ammonia. In 1894, Duden (Ber., 27, 3498) ex- amined this action, and found that hydraxine was produced in small quantity. This interesting result has caused us to re-examine the subject with the object of ascertaining whether what we judged to be hydroxylamine might not have been hydrczxine ; for, at the time of our former work, liydrazine had not been discovered, and it would have given the reactions we relied on as evidence of the presence of hydroxylamine. We felt it desirable also to ascertain whether hydr- oxylamine or hydrazine was the strongly reducing substance which, in very small proportion, accompanied the hyponitrite formed from a nitrite by the action of sodium.The results of our irivestigation have not only cleared up these two points, but have shown that, besides hyponitrite and nitrous oxide, sulphite, hydrazine, and ammonia,, both sul phate and amidosulphonate are also produced in large quantities. Before treating of these results, we desire to give some details which may be foand useful in the preparation of potas- sium nitrososulphate. Preparation of Potassium Nitrososu1phate.-It is a mistake to sup- pose that nitric oxide unites but slowly with potassium sulphite ; it is only the insolubility of nitric oxide in water that retards this union, as the use of suitably shaped vessels shows.We have arranged four flattened, conical bottles in series by means of corks and tubes ; the diameter of the flat bottom of each bottle being, on the inside, 19.5 cm., and the height, t o the commencement of the neck, 3 cm. only. With 100 C.C. of solution in each, the depth of the liquid is only 3-4 mm., whilst the free surface is nearly 300 sq. cm., so that the four bottles give a surface of 1200 E q . cm. to 400 C.C. of solution ; this is further increased by the circumstance that the salt, as it forms, grows up in small heaps above the level of the liquid. A con- centrated solution, containing 40 per cent.potassium sulphite and 5 per cent. of potassium hydroxide, will give 70 grams of crystals or more in three hours. In cold o r temperate seasons, external cooling is of little use ; motion of the bottles, beyond an occasional tilting, is also uncalled for. It is, of course, necessary to replace the air a t the beginning, and the nitric oxide at the end, by hydrogen ; but the pro- duction of a little oximidosulphonate or nitrite is of little moment, since these remain in solution. The salt, when drained on tiles, is pure enough for most purposes, but, for special work like the present, it can be purified and obt'ained in good crystals by dissolving it quickly at 50--60° in 4-5 times its weight of water containing 1-1-5 per cent. of potassium hydroxide, but this involves considerable loss.Reduction bg Sodium Amalgam-The amount of water present with the salt seems to be without effect on the course of the re- duction ; the extremes we hare used have been from 3 to 10 parts of1612 DIVERS AND HAGA: water to 1 part of salt, one per cent. of its weight of sodium hydr- oxide being added to the water at starting as a precaution. Wheu much less than LO parts of alkaline water are taken, some of the salt remains undissolved during the earlier stages, and less than 3 parts of water are insufficient. The amalgam we used was one contain- ingabout 2.5 per cent. sodium. The reduction goes on rapidly at the ordinary temperature, and much heat is developed, whilst if the nitro- sosulyhate solution is cooled below Oo, the amalgam acts on it only when it first comes in contact with i t ; perhaps it would not, even then, if it itself were first cooled down.At this low temperature, action is arrested for any period, but, soon after the vessel is removed from the water bath, action begins, and, once started, is not easily checked by returning the vessel to the bath, as the solution is kept warm by the heat developed by the action. A moderate rise of tem- perature to 40°, for instance, does not seem to lessen the production of either hydrazine or hyponit.rite. The sodium requisite to reduce the nitrososulphate is, as nearly as could be estimated, 3Na : K(,N,SO,-; but to destroy all the hydrazine much more is required (we nsed 2Na additional for this purpose). After the maill change is com- plete, the interaction between the amalgam and solution is very slow, the solution remaining cold, and hydrogen making its appearance, along with much ammonia.The contact of the amalgam with the solution has been maintained, in our experiments, f o r 24 hours, and for two days, but, with continuous shaking during the second stage, much less time would have sufficed. We used a stoppered vessel, the loose stopper acting effectively as a valve in keeping air out. Eydroxylamine, a Product of the Reduction of a Nitrite by Sodixirz Amalgarn.-As the testing for small quantities of hydroxylamine, alone or in presence of hydrazine, possesses some novelty, i t is well to describe the positive result. in the case of sodium nitrite before the negative one with potassium nitrososulphate is alluded to.The process we adopted consisted in shaking the solution with acetone, distilling with steam to get over the acetoxime, and evaporating the distillate with hydrochloric acid, in order to recover the hydroxyl- amine as its hydrochloride. Concerning the formation of acetoxime, we found that when the acetone was left in contact with the solution made neutral to litmus, so as to have the hydroxylamine free, but no alkali present, the action was very slow and unsatisfactory, but that it was quickly completed when the potassium hydroxide was present in some excess. The strongly alkaline solution was distilled with the '' superheated " steam. Having proved the absence OF hydrazine in the product of the action of sodium amalgam on a soluteion of sodium nitrite, we mere able, in the above WRY, t o get crystals of liydroxylamine hydrochloride from it.THE REDUCTIO~ OF NITROSOSULPHATES.1813 Hydyoxylamine not a Pyodzict of the Reduction o j Nitrososulphate by Sodium.-After removing hydrazine from the solution of reduced nitrososul?hate by means of benzaldehyde and ether, and evaporating the residual ether, m-e tested the solution for hydroxylamine by the acetone method, and failed t o find any. Moreover, after the removal of the hydrazine the solution no longer had any reducing power 011 cupric oxide. Hydruzine.-Hydrazine is quickly formed from nitrososulphate by the action of sodium amalgam, and is then slowly decomposed by it, but so long as any nitrososulphate remains, the action of the sodium is diverted from t'he hydrazine.Ainmonia can hardly be ranked as a product of the reduction of iiitrososuiphatc, beicg the result of the hydrogenisat>ion of the hydrazine. During the reduction proper, it is almost entirely absent, but makes its appearance in quantity when the sodium is able to act on the water and liberate hydrogen. Hypoizitrite.-The hyponitrite produced by the reduction of the nitrososulphate is unstable, and continuously decomposes in the alkaline solution, with evolution of gas. After treating the nitroso- Rulphate with sodium for 24 hours only, the solution was mixed with excess of barium nitrate, and filtered from the precipitate pro- duced. Silver nitrate and some nitric acid added to the filtrate gale first a little reduced silver, and then precipitated silver hypo- nitrite equivalent to almost one-fifth of the total nitrogen of the nitrososulphate.Nitrous Oxide and Nitrogeir.--No estimate of the nitrous oxide has been attempted, b u t it is formed in large quantity. I t appears to be generated along with amidosulphonic acid, as well as witoh hyponitrite. Nitrogen probably accompanies it, since hydrazine is produced. Sulphate. -There being so much sulphite produced, we expected difficiilty in determining whether sulphate mas a direct product of t'he reduction of the nitrososulphate, or only the result of incidental oxidation of sulphite by the a i r ; we experienced none, however. When the water present is not more than 3 parts to 1 of nitroso- sulphate, anhydrons sodium sulphate is deposited before the solutioii has been removed fmm the amalgam, and comes in coiitact with ail-.We hare in this way got 4 grams of anhydrous sodium sulphate (containing only R very little sulphite) from 40 grams of potassium iiitrososulphate ; but little sulphate then remained in solution, as the barium precipitate obtained from it largely dissolved in hydro- chloric acid. It should be remembered that salts are only sparingly soluble in concentrated solutions of alkali. It would seem safe to say that, on reduct,iou, one-seventh of the total sulphur appears as sulphate. The quantitative determination of the sulphate produczd1614 DIVERS AKD HhCA: in this way is liot only interfered with bay the presence of sul- phite, but by thzt of much amidosulphonate, for the latter greatly retards, even if it does not prevent, precipitation of barium sulphate and sulphite, unless excess of barium nitrate is u3ed and the solu- tion is largely diluted.In the cold, barium nitrate, in bare excess, precipitates 43 per cent. of tlie sulphur as (very impure) sulphate atid sulphite, the estimation being based on the quantity of barium nitrate used, and not on that of the precipitate. When the sulphur dioxide is rapidly removed from the acidified solution by a current of air, the precipitate, obtained on adding a very slight excess of barium nitrate, shows by its weight, after purification, that 12.5 per cent. of the sulphur is precipitated as sulphate. But whether the sulphur dioxide has been expelled or not, and whether the solution is alkaline, neutral, or acid, if it is mixed with excess of barium nitrate and allowed to stand, much more snlphate is obtained; and after renioving the amidosulphonic acid by mercuric nitrate, still more barium sulphate is slowly deposited.The sulphate pre- cipitated later is not formed by the hydrolysis of some compound in solution, for that would be accompanied with acidification, whereas the neutral solution, after it has deposited the snlphate, remains neutral. The difficulties due to the presence OE amidosulphonnte, as well as sulphite, are, doubtless, not insuperable, and although, for the present, we are not prepared with a closely approximate determina- tion of the quantity of snlphate present, we can assert that thc amount produced lies between the limits of 12 and 20 per cent.of the total sulphur, an important fact enough. Sui@hile.-The sulphite formed in tlie reduction of the nitroso- snlphate, estimated iodometrically, is equal to 31 per cent. of the sul- phur. The actual determination presented no difficulty, but as the previous neutrslisation with dilute sulphuric acid and the other unavoidable slight exposure to air in preparing the solution, mast have reduced the quantity of sulphite, it would be unjusti- fiable refinement to assert that more than about one-third of the sulphur becomes snlphite. A s hydrazine acts slowly on iodine solutioa, the sulphite solution used for the determination was f I eed from it by prolonged treatment with the amalgam. According to the first, note by oneof us to the Royal Society, on the “ Formation oE Salts of Nitrous Oxide,” the presence of hyponitrite should have interfered, but, thanks to Thum’s valuable contribution (1894) to the knowledge of hyponitrous acid, in which it is correctly pointed out that hyponitrous acid does not act on iodine (evidence to the contrary having been iiue to the presence of acid silver hyponitrite in the crude sclution), we had learned that i t was without influence.I n titrating, the sulphite solution was poured at once into the iodineTHE REDUCTION OF ISITROSOSULPHATES. 1615 solution, and the excess of iodine estimated by sodium thiosulphate. One other point about the sulphite is that, after the sulphate and sulphite have been precipitated from the alkaline solution by barium nitrate barely in excess, the clear solution, after more barium nitrate bad been added to it, continued for a day or two to deposit barium sulphite, as well as the sulphate already mentioned. Awzidoszdpphonate.-The sulphur which is not precipitated by the barium nitrate remains in solution as amidosulphouate. The pre- cipitate contains some amidosulphonate, although the barium salt of this acid is soluble in water; it can be extracted, however, from the precipitate by washing it with water. Having added a small excess of barium nitrate, and filtered off tbe precipitate after two or three days’ standing, the mother liquor, slightly acidified with nitric acid, is poured into a moderate excess of niercuric nitrate solution, in order to precipitate the amidosulphonate as the oxy- mercuric salt (see “ Amidosulphonic acid,” p.1649); this is collected, washed, and decomposed by hydrogen sulphide. The mercuric sul- phide requires much washing i n order to extract all the amido- sulpho.nic acid from it. The filtrate and washings when evaporated to dryness in a desiccator leave tho acid in an impure form, but it, can be purified without much loss by washing it with dilute sulphuric acid (see p. 1640). Another way of examinicg the mercury precipi- tate is to boil it with hydrochloric acid and potassium chlorate, or to hydrolyse it at 150°, arid then precipitate the sulphate by barium chloride. By working in these ways, we have ascertained the amount of amidosulphonic acid to be equal to nearly half the sulphur of the nitrososulphate. Collection and Analysis of Results.-Not less than essentially three independent equations have to be employed to express the results of the interaction of sodium amalgam with potassium nitrososulphate. For convenience, we write thcse equations as four, namely,- 3[2K2N,S05 + 8Na + 7H,O = 2H2NS03K + N,O + 2KOH 4[K,N,S05 + 2Na + 8NaOHI.= (NaON), + K,SO9]. = &SO4 + N, + 2NaOH K2N,S0, + 6Na + 5H20 = K2S04 + (NH,), + 6NaOH 1 * K,N,SO, + 2Na + H,O If these reactions do occur, and in the proportions indicated by the numbers prefixed to them, the products will correspond in their proportions to those found, namely : amidosulphonate equal to half the sulphur and one-fourth of the nitrogen ; one-third of the nitrogen as hyponitrite (three-fifths of this were secured as hyponitrate before further decomposition) ; one-third of the sulphur as sulphite ; and one-sixth of the salphnr as snlphate ; together with one-sixth of the 2[1616 DIVERS h S D HAGA: nitrogen, partly as hydrazine, partly as elemental nitrogen ; and one-fourth of the nitrogen as nitrous oxide (besides t,hat from the hyponitrite reaction), these being as yet unmeasured.Further, the above reactions, in the proportions marked, represent one molecule bf the nitrososulphate as being acted on by sodium in the mean propor- tion of 3 to 3$ atoms, according to the relative quantities of hydra- zine and nitrogen produced, a result which agrees well with observalion , Theoretical Considerations. -The cause of the variety and number of the products obtained in the reduction of a nitrososulphate by sodium is undonbtedly to be found in the different points a t which fission of the molecule of the salt must so easily occur, as shown by the formula we have deduced for it.Under the action of sodium, the salt shows the same disposition to give both sulphite and sulphate that it does when heated and when moistened; bitherto, it has been possible to say of the latter changes that, supposing t h e salt to be a sulphonate, it might give a sulphate by an oxidising process, bat the present observation of the generation of much sulphate by the action of wdium, in strongly alkaline solution, affords another proof of the impossibility of regarding a, nitrososulphate as a sulphoriic compound. For sodium to produce n sulphate out of a sulphite seems incredible.Nevertheless, Duden, having adopted for potassium nitrososulphate the constitution given i t by Raschig, has not hesitated to derive a sulphate from it by the action of sodium, in an equation framed to express the formation of hydrazine. Raschig's formula is and is one well suited to explain the decomposition of the salt into sulphate and nitrous oxide,--too well suited, indeed, for it is hardly conceivable how, with such a constitution, a nitrososulphate could exist at all and, if existing, could ever give back sulphite when heated, If we are to believe that sodium amalgam would act on a salt of this constitution i n such a way as to produce a sulphate, we must ignore what we know to be true of every other snlphonate. Duden detaches the OK as potassiuni hydroxide and then, armed with this alkali, puts its hydrogen in the place of the S03K, and gets sulphate outl of this and the KO.But not a single instance can be found of an organic sulphonnte in reaction with potassium hydroxidc yielding a aulphate instead of a sulphite. So also with the sulphonated hydroxylamines, both of which decompose in concentrated potassium hydroxide solution, for they, too, give potassium sulphite. When Duden obtained hydrazine from a nitrososulphate, he furnished another proof that the salt is not a sulphonate. There is a fact which we have not yet brought forward in supportTHE REDUCTION OF NITROSOSULPHATES. 1617 of the non-sulphonic constitution of nitrososulphates, which we may now give account of in this connection. It is that sodium amalga.m, as Such, is wizhont action on a true aminesulphonate or hydroxyl.aminesulphonate ; me have, indeed, just been showing in this parper that amidosulphonic acid is producible by the action of sodium; it and imidosulphonate (no doubt, also nitrilosulphonate) are entirely unaffected by sodium. Oxyamidosulphonic acid, in alkaline solution, is also untouched by it and, in acid or neiitral solution, is merely reduced to amidosulphonic acid (see p. 1636). Hantzscli and Semple (Bey., 1895, 28, 274s) have stated that Schatzmann found Fremy’s potassium sulphazilati to be reducible by sodium, but only back to the oximidosulphonate from which it is pre- pared by oxidation: as a sulphonate, it is not affected. Since, then, all aminesulphonates, oxygenated or otherwise, resist, as sul- phonates, the action of sodium amalgam, while a nitrososulphate a t Once yields to it, the latter is not of the same class, that is, is noi a sulphonat e .A similar objection to nitrososulphates being regarded a9 sul- phonic compounds has been raised by Lachmann and Thiele (AnnuZen, 1895, 288, 267). It is that, whereas all undoubted sulphonic derivatives of ammonia, when mixed with nitric and sulph- uric acids in the cold (see Divers and Haga, Trans., 1892, 61, 9G;<), give pure nitrous oxide, and sometimes even a little nitramide itself, potassium nitrososulphate does not. We come now to the formation of amidosulphonate, which is essentially that of the reduction of a sulphate to the corresponding sulphite, a thing hitherto unknown t o occur in alkaline solutioti.But we have here to do with a sulphate of the group -N20K, and it would seem that, just as EtSO3K and AgSO3K do not oxidise to sulphate, whereas KSOsK does, because it has the oxidisable pot,as- sium atom, whilst they have a, non-oxidisable atom, that is ethyl or silver; so, conversely, nitrososulphate is reducible to a virtual sulphite, because it is the sulphate of a hydrogenisable radicle, whereas other sulphates have radicles (metal, alkyl, or ammonium) which cannot be hydrogenised. If it stood alone, the conversion of nitroso- sulphate into amidosulp honate would point t o a sulphonic constitution for it, but other reactions make the acceptance OE this impossible, unless, indeed, it could be a half sulphonic, half sulphatic salt, which also seems impossible.Were it of sulphonic constitution, it ought to yield, by reduction, a hydrazinesulphonnte, but! none of this can be found. When, after reduction, the solution is acidified and all the sulphur dioxide blown out of it, it yields up all its hjdrazine to benzaldehyde and retains no discoverable hgdrazine derivative. The formation of hydrazine presents a difficulty, whatever con-1618 DIVERS AND HAQA: stitution is given to the nitrososulphates, in that it requires the reduc- tion of the KON group. That this radicle can resist the attack of sodium amalgam is shown by the formation of the hyponitrite from a nitrite and from a nitrososulphate ; besides, Dunstan and Dymond (Trans., 1887,51, 657) specialiy tested the matter, and found a hypo- nitrite to be irreducible by sodium amalgam.Two ways out of the difficulty present themselves. It may be admitted that the group KON, detached when the sodium forms alkali sulphate, is reducible to hydrazine and potassium hydroxide, although (KON), is not. Or, considering that a nitrososulphate reverts, when heated, to sulphite and nitric oxide, even in its strongly alkaline solution, as we have shown (1895) in the case of the sodium salt, we may assume that to a slight extent this reversion occurs during the heating caused by the action of the sodicm amalgam, and gives nitric oxide, or rather dinitrosyl, (NO),, ready to be reduced by the sodium aud water to hydrazine. In studying the action of alcohol on nitrososulphates, we have already had occasion t o recognise this possibility of slight rever- sion occurring at the ordinary temperatures, in order to account for the Froduction of a little aldehyde.I t was shown by us (Trans., 1895,67,1038) that potassium nitroso- sulphate decomposes more slowly when dissolved in aqueous alcohol than in water, and that from the salt and the alcohol there are formed potassium hydroxide and potassium ethyl mlpbate, besides nitrous oxide and a very little aldehyde. In a cold saturated solution of the salt in 23 per cent.. spirit, about 14 per cent. of the salt, it was then stated, interacted witth the alcohol in this way, the rest decomposing into potassium sulphate and nitrous oxide as usual. Another experi- ment, in which 14 per cent. spirit was used, seemed to indicate that weaker spirit was more effective than stronger in forming alkali and ethyl sulphate, but the experiment was quantitatively incomplete.We have since ascertained the effect of using 15, 5, and 2.5 per cent. spirit, estimating in each case, as before, the extent to which the alcohol had been active by titrating the potassium hydroxide. With the 15 per cent. spirit, 10.8 per cent. of the salt interacted with the alcohol; with 5 per cent. spirit, only 48 per cent.; and with 2-5 per cent., less than 1 per cent. of the salt. Evidently, there- fore, water lessens the power of the alcohol to form potassium ethyl sulpbate. Luxmoore (Trans., 1895, 67, 1021) has opportunely shown that a thermometer with its bulb embedded in potassium nitrososulphnte subjected to heat marks from 127" to 148O, according to circumstances,THE REDUCTION O F NITROSOSULPHATES.1619 as the temperature a t which the salt explodes ; this observation agrees with P6louze’s statement that i t does so at about 130°, which was the only part of his description we had failed to adequately justify. We had ascertained that the medium (air, oil) surrounding the salt needed to be only from 91” to lo$’, according to circumstances, to bring about the explosion, and i t has now been established by Lux- moore that then the temperature of the salt ri’ses of itself to about l S O o before it explodes. By the above observation, Luxmoore has cleared away a difficulty in P15louze’s description of the nitrososnlphates, but he has raised another without in our judgment having reason for so doing.Bemuse he has found the salt to lose 2.5 per cent. in five minutes when at a temperature a little below 105O, he considers it impossible to explain how PBlouze could have found that it did not lose weight at all. I n another of his experi- ments, it was little more than half as fast as the rate just quoted, while our observation had shown that the loss med be only 10 per cent. in ?+ hours, and t,his seems t o explain how PBlouze might have failed t o notice sufficient loss to be deemed worth recording. We have merely to assume that he exposed his salt to heat for a short time only, and in a very dry atmosphere, and that he attributed the slight loss that even then must have occurred to the presence of a little moisture i n his salt as prepared.No doubt the salt loses weight rapidly when heated in ordinarydamp air; underthose circumstnnces,itdoes so slowly even at the ordinary temperature, whilst the sodium salt loses weight rapidly. But in a weZZ dried atmosphere, either of air or hydrogen, such as was used in our experiments, the well dried, powdered salt loses weight much more slowly, so that it becomes probable that, with abso- lute dryness of salt and atmosphere, there would be no loss at all. It is surely on account of dampness of the salt that the rate of loss is most rapid a t first, as Luxmoore rightly observed, and we must therefore also assume that P6louze worked on a well desiccated salt. But the rate of loss varies greatly. I n concluding this paper, we would call attention to the compara- tively small part which Sir Humphry navy had i n the discovery of the nitrososulphat~es ; so small indeed is it that we must demur to the habit which prevails of naming him as their discoverer, for it is an injustice to the memory of Pt5louze. In the year 1800, thirty-five years before PBlonze published his work on these salts, Davy made known that apparently a combination of nitrous acid with potash was obtain- able by subjecting a mixture of potassium hydroside and sulphite in the solid state to the prolonged action of nitric oxide, dissolving the product in water, crystallising out potassium snlphate, and evapo- rating the mother liquor to dryness. The residue was a mass which,1620 DIVERS AND HhGA : IMIDO SULPHONATES. when heated, yielded about a fourth of its weight of pure nitrous oxide. There can b3 no doubt that he had obtained potassium nitroso- d p h a t e , but there can be no doubt also that he did not know it, that he did not isolate the salt, and that he thought the product t o be potassium hyponitrite, formed from nascent nitrous oxide, the nitric oxide having been deoxidised by the sulphite. It was, no doubt, Davy’s observation that led P6louze to investigate the nature of the action, and consequently to the discovery of the nitrososulph- ates ; ‘but that is all that Davy had to do with the matter. Imperial Univemit y, Tokyo, Japan.