148 IJEAN AND WHATMOUGH: NEW METHOD OF VIIL-New Method of P w p a ~ i ~ ~ Pu~e Iodine. By BEVAN LEAN, D.Sc., B.A., and W. H. WHATMOUGH. Intmduction. IN his Nouvelles Recherches .sur les Lois des Proportions Chinaiques (p. 136), Stas says :- ‘‘ Pour effectuer une synthese complhte d’iodure d’argent, j’ai dQ nhcesssire- ment ine procurer d’abord de l’iocle pur. Tous les moyens indiqds pour arriver A ce r6sultat ne m’ont point parn presenter des garanties auffisantes. Aucnn de ces nloyens, en effet, n’exclut la possibilitd de la presence clu chlore ni cle brome. Aprks avoir mfirement examine toutes les conditions dam lesquelles on peut probablement parvenir B obtenir cle l’iode ; j e n’ai pu en ddconvrir que deux : l’une consiste dans la prbcipitation, par l’eau, de l’iode clissons dans une solution d’iodure de potassium ; l’autre &side dam la d6ccm- position de la diiodamine par la chaleur.En effet, le chlore et le brome, con- tenus dans l’iode employ&, doivent rester unis soit au potassium, soit l’am- moniuni. . . . . DQlayb clans de l’esu, l’iode a 6th introdnit ensuite dans une grande cornue tubulee, et distill6 B la vapeur cl’eau pure. Cet iode npr& avoir btk 6goutt6, a Bt6 expose sous une cloche contenant de l’azotate de cliaux desshche. Ce sel de chaux a 4th la seuIe matiere que j’ai pu dkouvrir pour s6cher l’iode sans lui communiquer des impuretds. ‘‘ L’iode, dess6che aussi bien que possible, a dt6 m6lB de cinq pour cent de son poids de protoxyde de baryum pur finement pulverisd et soumis B la distillation seche.11 a Bt6 r e p dans une cornue tubulee qui servait de r6cip- ient et qui contenait aussi du piotoxyde de baryum pur finement pulverise ; il a dt6 rectifi6 une deuxiQme fois, en condensant sa vapeur dans m e cornue vide. . . . . En distillant l’iode sur du protoxyde de baryum, j’avais un double but : j e vonlais le priver de l’eau qu’il retient avec m e grande opiniiitretd ainsi que l’acide iodhydrique. . . . L’azotate de chaux a Qtd renouvel6 tant qu’il s’est humect6.PREPARING PURE IODINE. 149 L'iode, produit itinsi, differe notablement, par soii aspect, cie celui cln commerce. Apres nvoir 6t6 fondn dans nn tube de verre, il est absolunient noir h 1'6tat liquide et solide : h la tempdrature ordinsire, il n'6met aucunc vapeur visible dam l'air. . . . . 011 admet gh6rdement qne le point cie fusion cle l'iocle est lO'i", et que soii point ci'6bullition est conipris eiitre 175' et 180".L'iode cle Is diiodaiiiine est encore solide a 1133 ; mais il est liqnide B 115", et ne bout pas encore B 200"." The chief difficulties which Stas had to overcome in the preparation of iodine mere its separation from bromine and chlorine, and subse- quently the removal of moisture and hydriodic acid. It would be of great interest t o learn in detail how Stas assured himself that his ' iodine' was free from other halogens, and that calcium nitrate was the only desiccating agent which did not introduce impurities t o the iodine. No further information on these points can, however, be gleaned from his published researches ; several of his laboratory note-books kept in the Solvay Institute at Brussels, and courteously lent to us by the Director, Dr.Paul HBger, have also been carefully examined with the same object, but without success. A few months ago, we observed incidentally that no iodine was set free when cuprous iodide was heated, even till fused, in a current of carbonic anhydride, although, as is well known, iodine is readily evolved when cuprous iodide is heated in air to a moderate temperature. A few pre- liminary experiments showed that the action was probably represented by the equation Cu2T, + o2 = 2Cn0 + I,. Now i t hase commonly been supposed that an iodide c.an be detected in the presence of other haloids by precipitation as cuprous iodide, and if so, i t should be possible to prepare cuprous iodide free from bromide or chloride, and then from it liberate pure iodine in the way indicated above.With the object of preparing iodine, free in particular from chlorine, bromine, hydriodic acid or water, the whole question was submitted to careful examination, Prepamtion of Cuprous Iodide. Action of Copper SuZplmte 8c&mzted with Sulphurous Acid on Hccloid Su1t.s.-It is well-known that cuprous iodide is immediately precipitated when a solution of copper sulphate saturated with sulphurous acid is added to a solution of an iodide (Duflos, Annalen, 39, 253). Thus :- ZCUSO, + 2KI + SO, + 2H20 = Cu,12 + K2S0, + 2H2 SO,. Cuprous iodide was prepared by this method :-TWO gram-molecules of copper sulphate were dissolved in 3 litres of water, the solution150 LEAN AND WHATMOUGH: NEW METHOD OF saturated with sulphur dioxide, and 2 gram-molecules of potassium iodide dissolved in 150 C.C.of water were added. The pale yellow precipitate which was immediately formed was allowed t o settle, the supernatant liquid poured off, and the residue washed with sulphurous acid solution by decantation until the whole of the sulphate had been removed; it was then boiled with water to render it granular, collected on linen, the product spread upon a porous tile, and finally exposed over sulphuric acid in a vacuum. Cuprous iodide retains moisture somewhat obstinately ; after exposure for three weeks over sulphuric acid, a sample still contained 0.18 per cent. of moisture, but this was removed after further exposure. The dried iodide was reserved for future experiments. One of the recognised methods of preparing cupozcs chloride is very similar to the above.A solution of copper sulphate and potassium chloride is saturated with sulphur dioxide, when a crystal- line, white precipitate of cuprous chloride is gradually deposited ; excess of sulphurous acid, however, retards the precipitation of the chloride. If the clear solution is decanted from the precipitate and boiled, a further quantity is deposited. The cuprous chloride can be purified by washing it first with a solution of sulphurous acid and afterwards with glacial acetic acid, the product being then pressed between paper and dried in a warm place (Wohler, AnnaZen, 1864, 130, 373; Rosenfeld, Bell., 1879, 954). Cuprous chloride cannot be washed by much water without under- going decomposition, for as soon as the excess of acid is removed, the following interesting action begins in the presence of air, and, as was shown by Vogel, is rendered evident by the orange or red colour which the precipitate suddenly assumes : 2Cu2CI, + 0 = Cu20 + 2CuC1,.Cuprous chloride can thus be almost completely decomposed by repeated treatment with water in the presence of air. Freshly precipitated cuprous chloride is redissolved by sulphurous acid. I n view of the knowledge that both cuprous chloride and cuprous iodide can be prepared by the action of sulphur dioxide on a mixture of copper sulphate with potassium chloride or iodide in presence of water, it might be expected that cuprous bromide could also be prepared by a similar method.This we have found to be the case, although we have not met with any mention of this method of preparation compare Dammer’s Handbuclr. der Bnorg. Chenaie, 1894 edition). About 20 grams of copper sulphate and 8 grams of sodium bromide were dissolved i n 300 C.C. of water, and sulphur dioxide passed in ; after a time, small, white crystals were deposited, These were filtered rapidly from the mother liquor, washed with sulphurous acid, spread upon a porous tile, and then exposed over potassium hydroxide in a vacuum.PREPARING PURE IODINE. 151 If the mother liquor was heated so as to expel sulphur dioxide, more crystals were deposited. The crystals, which were pale greenish-yellow, subsequently became pale bluish-grey. The copper in the cuprous bromide was determined rolumetrically.0.662 corresponded with 0.5830 iodine. Cu = 44.29. 0.6740 gave 0.8808 AgBr. Br = 56-00. Cu2Br, requires Cu = 44.21 ; Br = 55-79 per cent. Cuprous bromide, like the chloride, is decomposed by water in the presence of air, but the action does not take place so readily. It also can be dissolved by sulphurous acid. Both cuprous chloride and bromide may, therefore, be precipitated by the action of sulphur dioxide on solutions containing copper sulphate and a chloride or bromide; and if the solution is but moderately concentrated (twice decinormal in the case of the bromide) the precipitation cannot be prevented, contrary t o the statement of Fresenius, by the presence of excess of sulphurous acid. I t seemed, therefore, desirable to determine the dilution necessary t o prevent t?he precipitation of the chloride o r bromide.To ascertain this, 100 grams of copper sulphate were dissolved in 1 litre of water, and the solution saturated with sulphur dioxide. Twice normal solutions of potassium chloride, bromide, and iodide were made, and 10 C.C. of each added severally to 25 C.C. of the solution of cuprous sulphate in small stoppered flasks, and it was observed in each case whether a precipitate occurred or not. The haloid solutions were then diluted ten-fold, and again 10 C.C. of these were added severally to 25 C.C. of the same solutions of cuprous sulphate. Similar experiments were made when the haloid solutions were diluted one hundred-fold, a thousand-fold, &c. On adding silver nitrate to the last solution, containing 0.000033 gram of potassium iodide, an opalescence was produced immediately.From these experiments, it is clear that, whilst cuprous iodide, bromide, and chloride may all be precipitated by the addition of a haloid salt to a solution of cuprous sulphate, there is a great difference i n their degree of solubility, cuprous iodide being much less soluble than the bromide, and, similarly, the bromide than the chloride. This explains the anomalous results sometimes obtained by students in the practice of qualitative analysis. The results are recorded in Table I, p. 152. It is a common practice to remove iodine from a mixture of haloid salts by the addition of a solution of copper sulphate mixed wit,h ferrous sulphate, as well as by the method already examined.Experi- ments similar to those described above were, therefore, made to test the152 0-0024 gram. No opalescence after 10 days LEAN AND WHATMOUGH: NEW METHOD OF 0'0033 gram. Immediate opalescence and KCI. -- 1.5 grams. Little or no ppt. after 1 houi Crystals after 10 days. 0'0033 gram. Immediate opalescence and gradual formation of a ppt. I -I 0'00033 gram. Nu opalescence. 0.15 gram. No ppt. after 10 days. ~ - _ ~ - _ . _ _ _ _ TABLE I. I KI. KBr. 2.4 grams. Ppt. within 1 minute. 3.3 grams. Immediate ppt. 0 -24 gram. Ppt. within 3 minutes. 0.33 gram. Immediate ppt. 0.024 gram. Slight ppt. within 10 days 0-033 gram. Immediate ppt. 0.000033 gram. No opalescence after 10 days efficacy of this method; 100 grams of crystallised copper sulphate and 114 grams of ferrous sulphate were dissolved in 1 litre of water, and t o 25 C.C.of this solution, placed in small stoppered flasks, were added as before, in each case, 10 C.C. of solutions of the haloid salt. The results are summarised in the following table. TABLE TI. KCI. KBr. I KI. I I 1.5 gram. No ppt. 2.4 grams. No ppt. 3.3 grams. Immediate ppt. I I 0.33 gram. Immediate ppt. 0.033 gram. Immediate ppt.PREPARING PURE IODINE. 1.53 I n each case, after standing some time, a little ferric hydroxide was precipitated. A comparison of Tables I and I1 shows that a mixture of copper sulphate and ferrous sulphate is not nearly so liable to precipitate cuprous bromide and chloride along with iodide as a solution of copper sulphate saturated with sulphurous acid (the cuprous iodide may contain a little iron hydroxide.) Table I1 shows, moreover, that by securing a proper dilution it is very probable that cuprous iodide can be precipitated by a mixture of copper sulphate and ferrous sulphate, unaccompanied by cuprbus bromide or chloride ; if, further, the cuprous iodide, precipitated under such conditions, is collected and washed repeatedly with a solution of sulphurous acid, it is probable that every trace of cuprous bromide and chloride can be removed.A New Met?Locl of Prepcwing Cuprous Iodide.--If copper foil is torn into shreds and heated in the presence of air in a porcelain basin over a Bunsen flame, and iodoform sprinkled over it in small quantities at a time, a violent action bakes place, violet clouds of iodine being evolved, while a flame plays over the contents of the basin.The copper is then found to be coasted with a black scale which is very readily peeled off, leaving a clean copper surface, and the copper may then be re-treated with iodoform until little or no metallic copper remains. The black scale, on analysis, was found to contain a little carbon and cupric oxide, but it was almost wholly cuprous iodide (about 98 per cent.). On account of the difference in the properties of chloroform, bromoform, and iodoform, it is probable that by this method also cuprous iodide can be prepared entirely free from bromide or chloride. Prepu~ntion of Iodine frona Cuprous Iodide. As already stated, one of us observed incidentally a few months ago t h a t when cuprous iodide was heated in a current of carbonic anhy- dride no violet vapours appeared, although iodine was freely liberated if air or oxygen was substituted for the carbonic anhydride.It was found, also, that if carbonic anhydride free from air and dried by sulphuric acid was passed over cuprous iodide in a boat within a glass tube, and the iodide heated even until f u s e d , no trace of iodine could be detected in the effluent gases by means of starch paper. Cuprous iodide can, therefore, in all probability be completely freed from moisture and from hydriodic acid by fusion in a current of carbonic anhydride. Experiments showed, however, that if cuprous iodide was heated in a current of air, oxygen, nitric oxide, or nitrogen peroxide to tempera- tures between 200° and 300°, iodine was very readily liberated, the iodide at the same time becoming black.It appeared that the libera-154 LEAN AND WHATMOUGH: NEW METHOD OF tion of the iodine was dependent on the oxidation of the copper: Cu2T2 + 0, =3 2CuO + I,. We khen endeavoured to ascertain (1) whether the whole of the iodine was liberated, (2) whether the action was de- pendent on the presence of moisture, and (3) whether the iodine liberated in this way from pure cuprous iodide was pure and free from iodic anhydride or any compound of copper. A long piece of hard glass tubing, AB, heated by a gas furnace at C I), the temperature of which could be regulated by a thermostat, was connected a t E with a water pump, so $hat air could be aspirated through the tube.GH was a The following apparatus was employed. FIG. 1. glass tube through which cold water circulated, providing a condensing surface for the iodine vapour, and IK was a porcelain boat containing cuprous iodide. Action of Gases at the Ovdinarp Temperature.--The action of gases which liberate iodine from cuprous iodide with great readiness a t elevated temperatures was tried a t the ordinary temperature of the laboratory, the gases being dried by sulphuric acid. I n no case, how- ever, was sufficient iodine liberated to be condensed ; its liberation in any particular case was only detected by starch paper introduced into the tube between A and C. Air passed over cuprous iodide exposed to the light caused a slight coloration of the starch paper after 3 hours, and after 2 days the paper was quite blue. When air was passed over the iodide in the dark for 1 day, no iodine could be detected.Oxygen did not liberate any iodine in the dark. Nitric oxide and nitrogen peroxide each liberated iodine at once, both in the light and in the dark, Action of Ail* on Hot Cuprous Iodide.-It was desirable to find the lowest temperature at which iodine could be liberated in sufficient quantity to be condensed and collected, so that the chance of volatilisa- tion of cuprous iodide might be minimised (cuprous iodide boils at 759-772") ; below .200', iodine was liberated only very slowly, but between 220' and 240°, a continuous stream of violet vapour was carried forward and condensed in beautiful crystals upon the con- denser.Three experiments were made t o test whether the whole of the iodine was expelled. I n I, 2,5249 grams of cuprous iodide were heated a t 230-240" for 11 hours ; in 11, 1.7101 grams mere heatedPREPARING PURE IODINE, 156 to 400' for 18 hours, and in 1111, the cuprous iodide was heated a t 380'. I n I, 0.42, and in 11, 0.15 per cent. of the iodide remained un- decomposed. I n No, 111, the effluent vapours were passed into standard aolut'ions of sodium thiosulphate, which were afterwards titrated with standard iodine solution ; it, was found that 22 per cent. of the iodine in the cuprous iodide was liberated in the first half-hour, 54 per cent. in the second half-hour, and 19 per cent, in the third half-hour, or 95 per cent. in 18 hours ; whilst after 6 hours more a small amount of iodine was still left in combination.It is clear, therefore, that it is not easy to liberate the zuhole of the iodine from a given quantity of cuprous iodide, otherwise the relation Cu,12 : 2CuO mould connect in a way capable of exact determination the atomic weights of copper, iodine, and oxygen. Further experi- ments are being made on this point. ItaJEuence oJ Moisture on the Liberation of Iodine.--When cuprous iodide was heated a t 240' in air dried by passing it slowly through strong sulphuric acid, iodine was readily evolved. Experiments mere then made t o find whether air which had been more carefully dried FIG. 2. had the same effect. Tubes of the shape shown in the figure, made from soft glass tubing, were drier1 by heating them t o 200', and passing air dried by calcium chloride through them.A plug of glass wool, which had been dried at 200' for some hours, was placed at B, and the end A sealed. By means of a thistle funnel, about 1 gram of finely divided cuprous iodide, which had previously been fused and then allowed to cool in a current of carbonic anhydride, mas intro- duced into C, and a plug of glass wool placed in D. A layer of phosphoric anhydride was next placed in E, and another plug of glass wool at F ; the open end, G , was then drawn out, but not sealed. Two similar tubes were made ; these mere exhausted by a mercury pump and then sealed. Both tubes mere heated for 10 days in a steam oven, and allowed to remain for four weeks more at the ordinary temperature. I n one of the tubes, the bulb C, containing the cuprous iodide, was heated a t 400' for one hour, but no trace of iodine vapour could be seen.Air which had been dried over phosphoric anhydride was intro- duced into the other tube, which was then again sealed. After two weeks, this second tube mas heated to 230', when violet vapours were1.5 6 NEW METHOD OF PREPARING PURE IODINE. at once freely evolved and condensed on the cool portions of the tube. The experiments mere repeated with similar results. It seemed, there- fore, that the pres&ce of water vapour was not essential to the action of air on cuprous iodide. I n order to ascertain whether water vapour alone had any action on cuprous iodide, steam was passed over hot cuprous iodide, but no iodine was liberated even when the iodide was fused.Purity of the Iodine Prepred from Cuprous Iodide.-Iodine liberated as described from cuprous iodide a t a temperature of 240' might conceivably contain a small quantity of some substance which would not be volatile at a much lower temperature, iodic anhydride, for example. Some of the iodine, 2.7529 grams, was therefore introduced into a short glass tube open a t both ends, which was placed within a long wide tube heated to 75", and a slow current of air passed through ; after heating during four hours, a slight brown residue was left on the tube, weighing 0.3 milligram. It was found, however, that this residue was due to impurities in the air, as when the air passed over the cuprous iodide and that in which the iodine volatilised, was filtered through cotton wool, the surface of the tube remained perfectly clean, and no change of weight could be detected.In order to show that the iodine did not contain any compound of copper, the violet vapours liberated from cuprous iodide a t 240' were mixed with coal gas and air and the mixed gases burnt at the mouth of a glass tube. On examining the flame spectroscopically, no evidence of the characteristic copper bands could be detected, whilst they at once became visible when a copper wire was held in the flame. The melting point (uncorrected) of the iodine, determined in the usual way, was found to be 118*5-114". It has a blacker appearance than commercial iodine, and does not so readily emit violet vapours. Cone Zusion s. From the results of these experiments, it is evident- 1. That cuprous iodide can be prepared free from cuprous bromide or chloride. 2. That cuprous iodide can be heated without decomposition and completely dried in an atmosphere of carbonic anhydride. 3. That when dry air is passed over cuprous iodide heated at 240" the iodine is liberated and can readily be condensed. 4. That the iodine obtained in this way is free from any compound of copper. Whether such iodine is as pure or not as that prepared by Stas by other methods, in which one main difficulty was the drying of theCOHEN AND BRITTAIN: ACTION OF ALKALIS ON AMIDES. 157 iodine and the removal of hydriodic acid, it appears desirable to re- determine the atomic weight of the element prepared by this entirely distinct method. It may be added that iodine is also readily liberated, in a similar way, from palladious iodide, but owing to the costliness of this sub- stance a detailed examination of the action has not been made. OWENS COLLEGE AND ACKWORTH SCHOOL, Jn~ii~nn~, 1890,