1. The Apparatus. By H. J. S. SAND. IN previous publications apparatus for the rapid electroanalytical determination and separation of metals has been described, the principal advantages of which have been set forth in detail (a, Trans. Chem. Soc., 1907, xci., 373; b, Trans. Chem. Soc., 1908, xciii., 1572; c, Trans. Faraday Soc., 1909, v., 159). Attention is here only drawn to the ready adaptability of the apparatus to varying quantities of electrolyte, and to the fact that the efficiency of stirring is independent of the strength of

by making the anode wherever possible of glass, and the An anode made largely of glass has already been described on a former occasion (loc. cit. a), but the use of this has been discarded owing to the fact that it was too fragile.

Figs. 1a and 1b show the apparatus as now employed. The inner electrode a* is designed to be gripped at the top by the clutch described (loc. cit. b), its wide glass stem rotates inside the glass tube A, which latter is gripped by the collar P of the outer electrode, and held by the clamp of one of the stands previously described. In this electrode platinum is only employed in the gauze cylinder B, in the strips of gauze c which lead into the interior of the electrode to the copper wire D, and in the strips E leading out at the top, which are wound over by the platinum wire F. This latter forms a ring against which contact is made by the blocks of the clutch previously referred to. Attention may be drawn to the following points :-All connections through the glass are made in the manner previously repeatedly referred to, by strips of platinum gauze. In the present case these strips are taken in threefold thickness at the top, and have a width of not less than I cm. below, so that it is possible to pass currents up to 10 ampères through the electrode without endangering it.

4

A

5

6

7

8

9

10 84

FIGS. Ia AND 1b.

current employed. These advantages are absent from all those methods of working in which the current itself is utilised to stir the liquid, and which might otherwise appear attractive from the circumstance that they do away with the expense of a motor.

It was felt, however, that the amount of platinum employed essentially made the apparatus somewhat expensive, and the present communication contains a description of electrodes which may be used in a very considerable number of cases, and which, while retaining as much as possible the features of the former apparatus, only require a minimum amount of platinum. This has been achieved

⚫ A Paper read before the Faraday Society, December 13, 1910,

FIG. 2.

The gauze cylinder B is fitted to the bottom of the electrode by blowing the glass against it. At the top, the cylinder stands off from the glass by 2 or 3 mm. in order to allow the gas bubbles which are thrown inwards in consequence of the action of centrifugal force on the liquid, to escape readily. It is held securely by four small protuberances g on the glass. The copper wire is held in the interior of the tube H by the enamel support i. The platinum wire F is wound in such a manner as to allow its ends to be twisted together at the bottom so that there is no risk of its being pulled off. The stem is widened at K just enough to prevent the tube A from being pulled off, but not enough to hinder the collar P of the outer electrode from being slipped over it. The contraction of the stem at L serves for the purpose of allowing the space between the latter

The inner electrode was made to my instructions by Universitätsmechaniker a. D. Fritz Köhler, of Leipzig.

CHEMICAL NEWS,

Jan. 13, 1911

and the guide-tube A to be washed down readily with a few drops of water before disconnecting. The total weight of platinum in this electrode is about 5 grms.

For copper determinations a cathode of silver was used; for zinc determinations one of nickel was employed. In the former case the electrolytic deposit was readily removed at the end of an experiment by means of a hot dilute sulphuric acid solution (1:10) to which some hydrogen peroxide (2 cc. perhydrol per 100 cc.) had been added. The same stripping solution was used a very considerable number of times. The use of a nickel cathode for zinc determinations is not novel, and needs no further comment. The cathodes do not differ very essentially from those previously described, but great attention has been paid to design them in such a manner that they may be constructed without much trouble with the appliances usually available in a chemical laboratory. The electrode is made only of two principal pieces, viz., (1) the frame, which is cut from a single piece of metal sheet of about half a mm. thickness, and bent into the shape shown, and (2) the gauze. From the latter a cylinder is first constructed, and the final form is obtained by introducing eight small pleats at the top, the size of which can be easily calculated beforehand. These pleats also serve to hold the frame in its place at the top of the gauze, whereas it is fastened below by means of four stout rivets. For the silver cathode fairly coarse gauze (Kahlbaum's coarse gauze) was employed. Owing to the softness and brittleness of the metal it was found desirable to leave no loose ends, but to finish off all the edges in the oxygen blowpipe. It was found easiest to do this by removing a few horizontal threads, and then melting down the projecting ends of the vertical wires. In a similar way the vertical joint in the gauze cylinder was made by removing a few of the vertical wires so that about 5 mm. of the horizontal threads protruded from each end. These latter were then twisted together, and the ends thus obtained were fused down and flattened, thus giving to the joint the appearance shown in the figure. In the case of the nickel cathodes fusing of the loose wires was obviously not feasible, but it was here unnecessary owing to the much stiffer nature of the material. The ends were simply pleated over, and joints made where necessary by fastening with either nickel rivets or wire. The mesh used was II per linear cm.

The electrodes may conveniently be formed on suitable glass dummies, which may also be employed to return them to their proper shape if they should accidentally be bent out of form. Copper electrodes may also be made by the identical process employed for those of silver.

The

With regard to the practical application of these electrodes the following remarks may be made. numbers given by Mr. Smalley prove that in copper determinations it is possible to obtain as good results with them as with platinum electrodes. There appears, however, to be an appreciable difference in favour of the platinum with regard to the slowness with which the copper goes into solution during the process of disconnecting. Rapid work is therefore desirable during this operation, and, if possible, a large current should be kept passing,

This experience is easy to explain when we remember how very much greater the polarisation capacity of platinum is than that of silver, in consequence of the greater solubility of hydrogen in the former. Harmful couples between the copper and the metal of the cathode will thus arise more quickly in the case of silver than in that of platinum.

A slight blackening at the top of the frame of the silver electrode generally takes place after electrolysis owing to the action of ozone, but apparently does not affect the results appreciably.

The zinc determinations caused no more difficulty with the nickel electrode than with that of platinum. The zinc was removed by warm dilute hydrochloric acid, and the electrode became slightly discoloured after some time without any effect on the results, The glass of the stem of the anode and of the guide-tube is ground down slightly

15

during use, and is found as a turbidity in the exhausted liquids, but no effect of this on the results was noticed. In conclusion, attention may be drawn to a stand illustrated in Fig. 2, which has been designed to take all the auxiliary apparatus required during analysis. It will be seen that there is a compartment at the back in which the circular block of wood used for raising the tripod is visible. In front of it there is a compartment for the thermometer, and in front of this a compartment for the split cover-glasses. The alcohol and ether jars are held in recesses, by means of a wooden bar, in such a manner that it is impossible to upset them. This bar is held by two round-headed screws (not visible), and has two vertical slits which allow it to be slipped into position or removed without undoing the screws. This base of the stand is extended in front to give it more stability, and the washbottle and spare beaker required in analysis may conveniently be placed on the extension.

Mention may also be made of the fact that Mr. Smalley has often found it convenient to replace the glass coverglasses by such cut from pieces of mica.

2.

Results obtained with the foregoing Apparatus in the Analysis of Solutions containing Copper and Zinc. The Determination of Zinc from Citric Acid Solutions. By W. M. SMALLEY.

Copper. A standard solution of copper sulphate was made by dissolving a known weight of purest electrolytic copper (Kahlbaum) in nitric acid, evaporating down to dryness with a small excess of sulphuric acid, and diluting to 1 litre. Twenty-five cc. of the solution thus obtained were calculated to contain o'3025 grm. of copper, and the electro-analysis with platinum electrodes confirmed 0.3025 grm. Known quantities of this solution were taken for each determination.

To obtain good results warm solutions were always employed. The difficulty of washing the deposit without loss was minimised by having only very small quantities of nitric acid in the electrolyte. At the same time it was found better not to do away with this acid altogether, since, as is well known, it improves the density of the deposit very much. The general mode of working consisted in depositing the bulk of the copper with 10 ampères as long as the solution showed a blue colour, then removing the last traces with 5 ampères until the ferrocyanide or the hydrogen sulphide test showed the liquid to be free from copper, and then increasing the current to 10 ampères during the process of disconnecting. The beaker employed was a narrow form 200 cc. beaker, and the amount of electrolyte used was 110 to 120 cc. (Since these experiments were completed a smaller cathode has been made which may be employed in a 130 cc. beaker with 75 cc. of electrolyte, and which would thus allow the determinations to be completed a little more rapidly than those recorded). Table I. will show that the results obtained under these conditions with varying quantities of copper are practically as good as those usually got by means of platinum electrodes.

Zinc. -For the examination of the deposition of zinc a standard solution was made up by dissolving a known quantity of purest zinc (Kahlbaum) in nitric acid, evaporating off the latter with a small excess of sulphuric acid on the water-bath, and then making up to a known volume. The strength of the solution was checked by electro-analysis according to the acetate method with platinum electrodes, and found correct within the usual limits of error.

Table II. will show that equally accurate results are obtained with the present electrodes by the acetate method as with those of platinum.

The usual procedure was adopted. The solution was generally first made fairly acid with sulphuric acid to cause it to correspond with conditions actually met with in analysis, it was then made alkaline with caustic soda until the precipitate first formed had re-dissolved, and finally acid once more with acetic acid. During analysis a drop

was repeatedly tested with litmus-paper, and, if necessary, | the solution was maintained acid by the addition of more acetic acid. The temperature was kept low by allowing the beaker to stand in a large crystallising dish containing cold water. The liquid was tested for freedom from zinc after the experiment by means of potassium ferrocyanide. In a few cases it was also electrolysed to constant weight of the deposit.

Citric Acid Solutions.-As there is a distinct tendency with acetate solution for the results to be very slightly high when the liquid is completely exhausted, the following experiments were carried out, in which the acetic acid was replaced by citric (Table III.). The tendency to high results appears to be smaller in this case, but a longer time was usually required to exhaust the electrolyte completely. Solutions containing both Copper and Zinc.-In Table IV. two experiments solutions were analysed containing both copper and zinc, the identical method being employed which was used for these metals singly.

tion of zinc by evaporating down on a water-bath with excess of sulphuric acid to expel the nitric; the residue was dissolved in water made strongly alkaline with caustic soda solution until the precipitate first formed had re-dissolved (8 grms. NaOH), and then acid again with citric acid (15 grms.).

The following additional remarks may be made regarding the use of the electrodes. When it was necessary to clean them thoroughly before a first determination, this was done by boiling first with caustic soda solution, and afterwards by treatment with nascent hydrogen, obtained by electrolysing dilute sulphuric acid with a high current density. In the case of the silver electrode an appreciable amount of blackening of the upper part of the frame. took place during this process, owing to the production of ozone at the anode. It was found best only to remove as much of this as could be rubbed off with a silk handkerchief, but it was not desirable to remove the black deposit completely by ignition, since this leaves the surface of the silver in a somewhat porous state, in which it is specially liable to further attack by ozone in future determinations.

the greater part passes into solution. If a small quantity of a barium compound is mixed with the monazite sand before treatment, practically the whole of the mesothorium is separated with the insoluble residue, none passing into the solution. The part of the residue containing the mesothorium can readily be separated from the heavy unattacked sand by decantation. From this part the mesothorium can be separated by methods similar to those employed in the working up of pitchblende residues for radium. The cases of complete chemical identity between groups of radioactive elements, differing presumably in atomic weight, which now appear to include all the radio-active elements of period of average life longer than a year, are in direct conflict with the principle of the periodic law.

338. "Traube's Molecular Volume Method Applied to Binary Mixtures of Organic Substances." By WILLIAM

RINGROSE GELSTON ATKINS.

The compounds and eutectic mixtures described by Philip (Trans., 1903, lxxxiii., 814) were prepared, and their densities determined at or slightly above their melting, or eutectic-points respectively. Chloral hydrate was also studied.

In dealing with a compound, the molecular weight is given, according to Traube, by M = p(2 atomic volumes + co-volume), whereas if the calculation of the sum of the molecular weights of the molecules in the material being studied be made, one obtains M = p(2 atomic volumes +n co-volumes), where n is the number of molecules going to form the supposed compound. From the measure of agreement between the values thus obtained and the theoretical molecular weight of the assumed compound, it is possible to ascertain whether a compound exists or whether the fused mass is merely a mixture, eutectic or otherwise.

In the accompanying table are given the calculated molecular weights obtained from the relation :

M = pt [ at. vol, +co.-vol. at o° (I + at)]

for a compound, and—

M=pt [ at. vol. +n co.-vol. at o° (1+at)], using the values for the constants given by Reychler ("Outlines of Physical Chemistry," p. 68), and by Smiles ("Chemical Constitution and Some Physical Properties,' p. 143), where a is the coefficient of expansion of a gas (1). Freezing-point curve shows eutectic point at 15.6°. Traube's method reveals, at the most, slight association (2). Both methods point to a considerably dissociated compound, m. p. 28.8° (3).

Both methods show a considerably dissociated compound, m. p. 28.5° or 30'0°, according to which crystalline form was present (4).

Freezing-point curve shows eutectic-point at 18.1°. Traube's method points to the opposite conclusion (5). Both methods point to a considerably dissociated compound, m. p. 61° (6).

Both methods show a greatly dissociated pound (7).

com

A highly dissociated compound is shown by the freezingpoint curve. Traube's method agrees with this, but the results might almost equally well be taken as pointing to slight association among the like molecules (8).

The magnetic rotation and chemical evidence point to a compound, as does the above calculation (9).

Finlay's determinations of the freezing-point curves show that this is a mixture. Traube's method indicates a compound.

Thus in seven of the examples the evidence afforded by the various methods is not contradictory, whilst in two cases the results are absolutely at variance. It must, however, be remembered that the values obtained by Traube's method indicate association in the fused mass without distinguishing, for example, whether in No. 2 it is the phenol molecules which are associated with each other and the a-naphthylamine similarly, or whether the phenol and a-naphthylamine molecules combine in a certain percentage

of cases. If, however, the latter be the case, it is not easy to see why the association of the phenol should be between one and two when mixed with a-naphthylamine, and between two and three when mixed with carbamide. For the latter conclusion cannot be avoided unless it is supposed that the association of the carbamide in phenol solution is sufficiently high to raise the average of the mixture (phenol, 66 per cent, carbamide 33 per cent, by molecules) to the value found. This is rendered very unlikely, as a determination of the molecular weight of thiocarbamide in phenol (about 17 per cent of thiocarbamide by weight) gave a nearly normal value, 87 instead of the theoretical, 76. This result was obtained in an attempt to trace the freezing-point curve of phenol and thiocarbamide to see if a compound similar to the phenol-carbamide compound could be obtained. However, the separation of crystals of what had all the appearance of being thiocarbamide show an eutectic-point at 36.8°, only 3.74° below the meltingpoint of the pure phenol used. The thiocarbamide present at the eutectic-point amounted to only 4.2 per cent.

n

Thus it seems fair to conclude that whilst by Traube's method alone it is not possible to decide whether the association is between the like molecules or the unlike, taking other facts into consideration the method may be considered trustworthy for temperatures up to 60° at any considerable. In view, however, of the very good agreerate. Above this the effect of the coefficient becomes ment in the case of chloral hydrate (163.5 calc., and 165.5 theoretical) at 66°, it seems to be quite allowable to use the method at 105° for the thiocarbamide-thiocyanate fused The above evidence for the existence of Philip's compounds in the liquid state receives confirmation from Treitschke's viscosity measurements (Zeit. Phys. Chem., 1907, lviii., 425). These show that the compounds exist as such, but in a largely dissociated condition. Rise of temperature increases the dissociation.

mass.

=

(1 mol.).. .. .. P105/15-1231 298 423 304 (a) M as compound. (c) M theoretical. (b) M as mixture.

339. "Hydroxycodeine; a New Alkaloid from Opium." By JAMES JOHNSTON DOBBIE and ALEXANDER Lauder.

This alkaloid was discovered by Messrs. T. and H. Smith, Edinburgh, in the mother-liquors obtained in the working up of the opium alkaloids after the ordinary alkaloids have been eliminated. It has the formula C18H21O4N, and is a tertiary base. It is soluble in water and the usual organic solvents. So far it has not been obtained in the crystalline condition. On heating, it softens about 40°, and is completely melted at 51°. It is slightly dextrorotatory, and contains one methoxyl group. The hydrobromide, С18H21O4N,HBг, crystallises readily from water in large hard prismatic crystals. The platinichloride has the formula (C18H21O4N)2, H2PtCl6. The alkaloid appears to be a hydroxy-derivative of codeine; its absorption spectra agree very closely with those of codeine, and the colour reactions of the two alkaloids are practically identical.