143 XX.-CONTRIBUTIONS FROM THE LABORATORY OF GONVILLE AND CAIUS COLLEGE CAMBRIDGE. No. XIII. On a C,ubical Form of Bismuthous Oxide. By M. M. PATTISON MUIR M.A. Fellow of Caius College and ARTHUR HUTCHINSON B. A, Scholar of Christ's College Cambridge. 1. WHEN 2$ to 3 parts of potassium cyanide in concentrated aqneous solution are added to 1 part of bismuth nitrate dissolved in the smallest possible quantity of cold dilute nitric acid a white precipi-tate of hydrated bismuthous oxide is obtained. About nine-tenths of the bismuth present as nitrate a t the beginning of the reaction is precipitated as hydrated oxide ; the addition of a very large excess of potassium cyanide fails to precipitate all the bismuth. If this pre-cipitate is thoroughly washed and then boiled with concentrated potassium cyanide solution it remains unchanged.2. When 2+ to 3 parts of potassium cyanide are added to a hot solution of 1 part of bismuth nitrate in the minimum quantity of nitric acid a brownish-red precipitate is obtained; and in this case also about niue-tenths of the total bismuth is precipitated. The composi-tion of this precipitate is represented by Baedeker to be Bi,0,,2Hz0 (Annalen 123,61) ; one of the present authors has described the pre-cipitate as an oxycyanide of bismuth (Proc. Roy. SOC. Edin. 11 557) ; and C. Hoffmann asserts that the substance is bismuth sulphide (Annulen 223 110). When this puce-coloured precipitate is re-peatedly treated with concentrated Iioiling potash solution a deep-red liquid is produced and a heavy greyish-black solid remains.3. Numerous specimens of this greyish-black solid have been prepared by us. In each case from 2-2$ parts of potassium cyanide were used for 1 part of bismuth nitrate ; the liquids containing the precipitates were boiled for 10-15 minutes and the precipitates allowed to settle ; the supernatant liquids were poured off and the precipitates were then heated with boiling potash (1 part of potash to 6 01' 7 of water) ; as soon as the potash became coloured brownish-red it was poured off and treatment with fresh quantities of boiling potash was continued till the liquid ceased to be coloured; the residues were then washed with boiling water until perfectly free from potash that is until the colour of slips of red litmus-paper was unaffected after the paper had remained in the liquid for about 12 hours ; they were then dried at 150-1 70".4. The substance thus prepared is a dark-grey crystalline powder 144 MUIR AND EUTClhNSOS ON A which is seen under the microscope to codsist of sniall tetrahedra belonging to the cubic system ; it is soluble with diaculty in acids ; on heating in air it turns yellow. Qualitative examination of the different specimens shewed- (I.) Absence of chlorine compounds. (2.) Absence of sulphur compounds. (3.) Presence of small traces of potassium but absence of all other metals except bismuth. (4.) Presence of very small quantities of nitrogen and carbon. (5.j Presence of a little silica. And (6.) Presence of a little water. Numerous quantitative analyses of the substance have bseb made.Bismuth was estimated (1.) By heating the specirneti in air then dissolving i a nitric acid precipitating as carbonate and weighing as Bi,O,. (2.) By fusing with potassium cyanide and weighing as Bi. (3.) By heating in air dissolving in hydrochloric acid precipitating as oxychloride and weighing as BiOCl. O q g e n was estiinated by heating in pure hydrogen ahd weighing (1) the water formed and (2) the residue left. SiEica was separated in the ordinary way during the course of the determination of bismuth by methods (1) and (3) above. Carbon was estimated (1.) By heatiag in pure air and leading thr caisbon dioxide (a) into tt potash tube weighed before add after the experiment j ( b ) into baryta-water. and determining the BaCO, formed.(Duprk and Hake Jodr; Chew 8oc. Trans. 1879 159.) (2.) By heating in connection with a Sprengel primp arld measuring the carbon dioxide evolved. The percentage of carbon found ranged from 0.03 to 0.23. Nitrogen was not estimated as the qualitative examination showed it to be present in very Minute quantities. Determination of Wuter.-On heatirig in air the substahce turned yellow and was converted intb Bi2O8 this change being accompanied by slight losx of weight which varied from 0.13 tb 0.75 per cent. the mean being about 0.3 per cent. When heated in a stream of dry nitrogen quite free ft-om oxygen and oxides of nitrogen the substance does not turn yellow but a loss of 0.5 per cent. (mean of several closely agreeing experiments) takes place; as this loss of weight was almost equal to the gain in the weight of a calcium chloride tube attached the loss was shown to be due to the removal of a little water; st the same time minute fragments of metallic bismuth were prodtlced due to the reducing action of traces of carbon present in the original substance.We do not give details of our estimations as the experiments were so numeroua ; we rather present a condensed statement of the results. I. The most trustworthy results for bismuth were obtained from a specimen prepared with great cape and which was specialIy well crystallised j several complete analyses of this were made the sub CUBIO-PL FORM OF BISMUTHOUS OXIDE. 145 stance being first heated in dry air silica and potassium being determined in the residue. BizO found varied from 97.45 to 98.10 ; mean = 97.81 per cent., which is equal to 87-67 per cent.Bi. 11. SiO varied from 1-1-1.2 Mean value = 1.13 per cent. 111. The amount of potassium found was equal to 0.45 per cent. IT. Mean peTcentnge of oxygen found by rediking in hydrogen and deducting weight of residue from that of o&ginnl: Twd different specimens were used ; the results varied from 10.58 t o 11.34. V. Mean percentage of oxygen found by redacing in hydrogen a d weighing water formed. Two different specimens were used ; the results varied from 992 t o 12.8. Some small part of the loss of weight observed when the substauce was heated in hydrogen was doubtless due to removal of water associated with the silica and potassium compounds present and a small part of the water obtained almost certainly came from this Rource; but the magnitude of t'he coi*rect,ion to he applied to the oxygen estimation on thiR account could not be accurately determined ; it certainly did not exceed 0.3 per cent.The results of the complete analysis of the greyish-black solid may then be stated as follows :-KZO. Mean = 10.84 per cent. Mean = 10.87 per cent. Oxygen . . . . . . . . . . . Bismuth . . . . . . . . . . Silica . . . . . . . . . . . . Potash (&O) . . . . . . Water (approx.) . . . . C and N (approx.) . . . 10.85 (mean of many determinations made with different specimens). 87.51 (mean of 6 determinations made with t'he same specimen). 1'13 (mean of 6 determinations made with the same specimen). 0.45 (mean of 3 determinations made with the same specimen).0.30 (mean af many determinations made with different specimens). 0.20 (N not determined ; mean per cent. of C from diflerent specimens - = 0.125). 1 0 0 9 ~ I€ a correction is made on the bismuth for the quantities of silica, potash water carbon and nitrogen we have I 146 MUIR AXl) HUTCHINSON ON A Calculated for Bi20,. Bismuth . 89.46 Bismuth. . 89.65 Oxygen. 10.85 Oxygen 10.35 100.31 100~00 5. When the greyish-black substance which we have proved to consist almost entirely of bismuthous oxide was heated to redness in air or in a stream of oxygen i t rapidly changed to the colour and appearance of ordinary bismuthous oxide. This colour change was accompanied bya loss of weight amounting to about 0.3 per cent.(see paragraph 4). Microscopical examination showed that the greyish-black compound and the yellow substance remaining on heating in air were identical in crystalline form and consisted of minute regular tetrahedra. The sharpness of the outlines and the brightness of the faces of the tetrahedrons remained unaffected by heating the only change visible t o the eye being that of colour. The change produced on heating in air or oxygen almost certainly consists in the oxidation of a minute quantity of the puce-coloured compound formed by the reaction of potassium cyanide with bismuth nitiate. This small quantity of the compound in question adheres most persistently to the tetrahedral crystals of the bismuthous oxide. When the greyish-black crystals are heated i n nitrogen the colour does not chauge but a very little bismuthous oxide is reduced to bismuth (see paragraph 4) ; in this case the adhering film of the com-pound containing carbon and nitrogen is not burnt away but some of it reacts with the bismuthous oxide which it reduces to metal.We found that i t is possible to remove the whole of the puce-coloured compound containing carbon and nitrogen by prolonged treatment wir;h boiling potash solution so concentrated that it solidifies on cooling; the residual product is the pale yellow bismuthous oxide. Specimens thus prepared and dried a t 160" did not undergo any change in weight when heated to full redness in air. 'l'he long-continued boiling with very concentrated potash tends to corrode the cubical crystals and produce needles apparently rhombic, which are contaminated with much silica.6 . We find that the best method of obt'aining cubical bismuthous oxide is to prepare the gray substance as described in paragraph 3, and then to heat this in air till it turns yellow. The oxide thus prepared is pale buff-yellow; it consists of minute but well-developed regular tetrahedra the faces of these are very smooth and bright and their edges are in some cases modified by the cube planes or in crystallographic notation the crystals exhibit the form k ( l l 1 ) and (100) of the cubic system CUBICAL FORM OF BISMUTHOUS OXIDE. 147 According to Nordenskjold ( P o g g . Ann. 114 512) bismuthous oxide crystallises in rhombic prisms; the ratio of the axes given by him is not identical with tthe ratio for either arsenious or antimonious oxide but the three oxides are probably isomorphous in the usual acceptation of this term.It is moreover well known that antiinonious and arsenious oxides are also dimorphous crystallising i n the cubic as well as in the rhombic system (Malard is however of opinion that the apparently cubic crystals of antimonious oxide are in reality built up of indi-viduals belonging to the oblique system). The cubic crystals of t'hese two oxides are usually octahedra but arsenious oxide sometimes crystallises in tetrahedra also and we may therefore consider that t,he preparation by us of tetrahedral bismuthous oxide establishes in a broad sense the isodimorphism of this group of oxides as is seen in the following table :-rhombic; prisms 0.3758 1 0.35; sp.gr. 4.2; cubic ; octahedra and tetrahedra ; sp. gr. 3.689. Arsenious oxide Antimonious oxide rhombic; prisms 0.394 1 1,414; sp. gr. 5.5 ; { cubic ; (2) octahedra ; sp. gr. 5.1. cubic ; tetrahedra ; sp. gr. 8.828. rhombic ; prisms 0.8165 1 1.064; sp. gr. 8.3 ; oxide Cubical bismuthous oxide is specifically heavier than the rhombic form. The sp. gr. of the specimen prepared by us was found to be 8.868 a t 25" referred to water a t the same temperature (mean of several closely agreeing determinations made with different speci-mens). Herapath (Phil. Mug. 64 321 [18%4]) gave the value 8.21 ; Karsten (Schweigyer's Journ. 65 320 394 [ 18321) gave 8.173 ; Playfair and Joule (Chenz. SOC. iWem.3 57 [l848]) gave 8.079 ; and Le Royer and Dumas gave 8.45 (see Playfair and Joule). The mean of these values is 8.304. 7. We have endeavoured to trace the reactions by which the cubical form of bismuthous oxide is produced but we have not yet succeeded in fully elucidating the stages of the process. Our experi-ments have shown (1.) That the tetrahedral crystals are formed in the largest quantity under the conditions described in paragraph 3. (2.) That if the precipitate is thoroughly wmhed before it is heat'ed with potash very few tetrahedra are formed. ( 3 . ) That if this precipitate is not washed but is kept in contact with very concen-trated potash solution for several weeks in the cold a few tetrahedra axe formed. We have also found that the puce-coloured precipitate produced by potassium cyanide is amorphous 148 KISIMINS PERIODATES.Repeated experiments have convinced us that the action of hot potash on this amorphous precipitate frequently results in the forma-tion not onlF of tetrahedra but also of needles which are perhap rhornbic prisms and ako of more or fewer twin-crystals somewhat of this form-In some cases the product consisted mainly of these twins in some the twins were mixed with needles and in some the product was mostly composed of tetrahedra. Analyses detailed by one of us (Proc. Roy. SOC. Edin, 11 557) led to the conclusion that the puce-coloured precipitate is an! olxycyanide of bismuth. More recent analyses and an examination of tbe action of acids on t h i s precipitate lead us to say rather that it is a compound of bismuth oxygen carbon and nitrogen. When this precipitate is boiled with concentrated potash solution 2-3 per cent. of the bismuth iri the snbstance goes into soIution along with practically all the carbon and nitrogen. Neutralisation of the deep red liquid thus obtained produces a red-brown solid which contains bismuth carbon, nitrogen and water. Analyses of this solid have led to conflicting results ; i t does not give the reactions of a cyanide and the nitrogen and carbon are present in approximately the ratio 1 1.4 whereas in cynnogen they are present in the mticr 1 0,857. We reserve m account of the puce-coloared precipitate and of the red-brown solid until we have exsmked them more fully