THE CHEMICAL NEWS

VOL. CII., No. 2664.

THE ATOMIC WEIGHT OF TANTALUM.* By CLARENCE W. BALKE.

Introduction.

Of the previous determinations of the atomic weight of this element, those by Berzelius (Pogg. Ann., 1825, iv., 14). Rose (Pogg. Ann., 1856, xcix., 80), Hermann (Journ. Prakt. Chem., 1857, Ixx., 193), and Blomstrand (Acta Univ. Lund, 1864) need not be considered, as their material was too impure to give results of value. The determinations of Marignac, which appeared in 1866 (Arch. Sci. Phys. Nat., 1866, [2], xxvi., 89; also Journ. Prakt. Chem., xcix., 23), were far more satisfactory. He analysed potassium fuotantalate, K2TaF7, and the corresponding ammonium salt, (NH4)2TaF7. In the case of the potassium salt he obtained the following percentages :-Ta205, 56'50, 56.75, 56′55, 56'56; K2SO4, 44′37, 44 35, 44 22,

44°24.

The ammonium salt gave the following percentages of tantalum oxide :-63.08, 63°24, 6327, 63°42. These numbers were corrected for a small amount of potassium sulphate which was present in the oxide, due to the fact that he was unable to prepare the ammonium fluotantalate free from traces of the potassium salt.

If in the calculations the atomic weights of the elements involved are taken as given in the 1910 International Table the following values for the atomic weight of tantalum are obtained :

From the ratio K2TaF7: Ta205

182.4 to 184 6 K2TaF7: K2SO4 181 6 to 1829 K2SO4 Ta2O5 · 1819 to 183.0 (NH4)2TaF7: Ta205 .. 1805 to 183.8 It will be observed that there is a variation of several whole units between the individual determinations, yet for about forty years the accepted value of this constant was based on these results.

In 1906 Hinrichsen and Sahlborn published a series of determinations in which metallic tantalum was burned to the oxide (Ber., 1906, xxxix., 2600). The method has great simplicity, but it appears very doubtful to the writer if this metal can be prepared sufficiently pure for the purposes of atomic weight work. Their results are given below.

Here again there is a variation of over one unit between the individual determinations. Clarke, after reviewing these various determinations of the value of this constant, remarks that its uncertainty probably amounts to as much as a unit ("A Re-calculation of the Atomic Weights,' Smithsonian Miscellaneous Collections, 1910, liv., No. 3). Wt. of Ta. Wt. of Tagos.

O'45437

Per cent O. 22.12

0'37200

0'41278

0'33558

180.98 181.14 180.59 Mean .... 181.03 While carrying out an investigation on columbium (Balke and Smith, Journ. Am. Chem. Soc., 1908, xxx., 1637) it was found possible to make a satisfactory determination of the atomic weight of that element through the conversion of its pentachloride into the corresponding oxide. Consequently the writer decided to carry out a similar investigation with tantalum chloride, and, in general, the methods used were very similar to those employed in

the former work.

Presented at the San Francisco meeting of the American Chemical

Preparation of Materials.

295

In the preparation of all materials used during the investigation great care was exercised in the cleaning of all containing vessels, and especial care was taken to exclude dust. Whenever possible the apparatus used was made continuous by fusing the various parts together, and this was always the case where chlorine was being used. In the case of chemicals not especially purified, those of the highest purity obtainable were employed.

Chlorine. The chlorine used, including that used in the preparation of the sulphur monochloride, was made by allowing hydrochloric acid to drip upon potassium permanagate contained in a large flask. The acid used was prepared by distilling the constant boiling mixture from large Jena retorts. During the first part of the distillation small quantities of potassium permanganate were added from time to time, and the first portions of the distillate were rejected. The chlorine was passed through a wash bottle containing water, two containing sulphuric acid, and finally through a tower 1 metre long containing glass pearls, and so arranged that the latter could be moistened from time to time with concentrated sulphuric acid.

Sulphur Monochloride. This compound was prepared by passing a stream of chlorine over molten sulphur which had been crystallised previously from carbon disulphide. The apparatus was so constructed that the liquid could be re-distilled from the receiver without transferring the material. The first portions of the distillate were rejected, and the final product was collected directly in sealing flasks through which a stream of dry filtered air had been allowed to flow.

Nitric Acid.-The nitric acid used in the determinations was distilled twice from a platinum retort, the middle third of the distillate being retained in each case. The acid was preserved in quartz asks.

Water. The water used, including that employed in the final re-crystallisations of the potassium fluotantalate, was prepared by re-distilling the ordinary distilled water of the laboratory, after the addition of alkaline permanganate, from a still ordinarily used in the preparation of conductivity water.

Tantalum Chloride. -This compound was prepared by the ignition of tantalum oxide in a current of chlorine, and the vapours of sulphur monochloride. The oxides used for preparations I. and II. was prepared from some of the and Smith from the columbite of South Dakota (Journ. potassium fluotantalate which had been obtained by Hall Am. Chem. Soc., 1905, xxvii., 1369), and which was very kindly furnished me by Prof. Edgar F. Smith. (The author wishes to take this opportunity to express his appreciation for this and many other kindnesses shown him by Prof. Smith). This salt was re-crystallised four times from water containing appreciable amounts of re-distilled hydrofluoric acid, and only the first crops of crystals were used in the next crystallisation. These operations were carried out in platinum and hard rubber dishes, and hard rubber funnels were used in the filtration of the solutions. The final crop of crystals was heated in platinum dishes with concentrated sulphuric acid, which had been distilled from platinum, until most of the acid had been expelled. The residue was washed by decantation with large amounts of pure water. After this washing was complete, the mass was strongly ignited in platinum or quartz crucibles. Tantalum oxide prepared in this way has been shown to be pure (Balke, Journ. Am. Chem. Soc., xxvii., 1140).

The oxide used for the third preparation of the chloride was obtained from the columbite of South Dakota as

follows:-The finely ground mineral was fused with potassium hydroxide in large iron crucibles; the mass was dissolved in water, and the solution, after filtration, was treated with sulphuric acid, which precipitated the metallic acids. These were washed with water by decantation, and then dissolved in an excess of hydrofluoric acid. addition of potassium hydroxide to this solution gave a

The

No. 10.

to one of my students, Mr. Lloyd Almy, who carried out

this preliminary work). The final purifications were carried out as previously outlined.

The apparatus used for the preparation of the chloride is shown in Fig. 1. The tantalum oxide (50-100 grms.) was introduced into the hard Jena glass tube H, the ends of which were connected with the bulbs E and I by means of carefully ground joints which were lubricated with a trace of graphite. These joints were wrapped with several thicknesses of asbestos paper, and the bulb of a thermometer was inserted in the latter at G, and the temperature regulated so as to prevent a deposition of the chloride at this point. A stream of chlorine was passed through the apparatus by means of the tube B for three to four hours, the entire apparatus being heated with a free flame in order to remove any traces of moisture which might be present. At this time the tube H containing the oxide was heated to a low red heat for the same purpose. Sulphur monochloride was then run into the bulb E through c, and heated to boiling. The oxide was slowly converted directly into the chloride, which together with the excess of sulphur monochloride collected in bulb 1. When all of the oxide had been converted into chloride, leaving practically no residue in the tube H, the sulphur monochloride was driven on through the entire apparatus, and collected in o. The tantalum chloride was then distilled over into the bulb, and the apparatus sealed off at L. The two-way stopcock

Density of Tantalum Oxide.-As was found in the case of columbium oxide, the density of the tantalum oxide obtained from the present determinations varied quite appreciably, the numbers 8.62, 7.91, and 8:06 having been obtained. The density of this substance is not far from that of the weights used, which were taken as having a density of 84, and it was found unnecessary to make a vacuum correction in the case of this substance.

Analysis of Tantalum Chloride.

The ratio studied was that of tantalum chloride to tantalum oxide, the transformation being made in quartz bulbs provided with tightly ground stoppers of quartz or glass. One of Ruprecht's best balances was used in the

was then turned so as to deliver chlorine through D. The tantalum chloride was heated to boiling, and the entire apparatus heated, so that the residue of sulphur monochloride and some of the tantalum chloride was driven over into o. The entire amount of chloride was then distilled over into K, and finally into the small bulbs N. The chloride in all the bulbs was then brought to the boiling-point, and a small amount distilled over into the receiver o to insure the removal of the last traces of sulphur monochloride. When cold the apparatus was sealed off at P and м, and finally the individual bulbs were sealed off. These con tained from 5 to 20 grms. of chloride.

The tantalum chloride so prepared was absolutely white in colour, having none of the pale yellow tint which this compound is usually said to possess. When fused the chloride had a pale straw-yellow colour. Five grms. of this chloride were thrown into a flask containing bromine water, and the contents heated for several hours. The tantalum hydroxide was filtered out, and the filtrate was evaporated to a very small bulk. This remained perfectly clear after the addition of a solution of barium chloride, which was taken to indicate that no sulphur remained in the preparation.

Density of Tantalum Chloride.-The density of this chloride was determined in carbon tetrachloride, and the number 3.62 obtained. Tantalum chloride is somewhat

work, and the weights were carefully standardised. All weighings were made by substitution, using a tare of the same material and very nearly the same weight as the object being weighed. The excess weight of the quartz bulb and its stopper over the tare was determined. By carefully tapping one of the small bulbs containing the tantalum chloride the latter was broken up into small pieces, and by means of a file and a hot rod a crack was drawn about two-thirds of the way around the neck of the bulb. This bulb, together with the weighed quartz bulb, into the neck of which had been slipped a wide-neck funnel tube, were introduced into a gas-tight box supplied with a removable plate-glass top. Two holes had been cut in the sides of the box, and rubber gloves were attached on the inside. This box contained a large amount of calcium chloride, and a stream of air, which had been passed through large amounts of sulphuric acid, and then filtered through glass-wool, was passed through the apparatus for several hours. By slipping the hands into the gloves it was possible to remove the top of the bulb containing the chloride, and introduce the latter into the quartz bulb without soiling the neck of the bulb. In this way it was possible to effect the only transfer of material involved in the determinations in an atmosphere of dry air. replacing the stopper in the quartz bulb the latter was removed from the box, and the excess of weight over the

After

Before entering upon a discussion of the theoretical aspects of the matter, certain other facts must be considered, namely:

1. In the detonation of a dried mixture of cyanogen with twice its own volume of oxygen the formation of carbon dioxide is complete; moreover, under such conditions it has been proved that carbon monoxide is primarily formed in the wave itself, the second stage of the combustion—namely, 2CO + O2 =2CO2-taking place in the rear of the wave (H. B. Dixon).

2. A well-dried mixture of carbon monoxide (36 per cent), ozone (8 per cent), and oxygen cannot be fired with a powerful electric spark; also, on sparking a well-dried mixture of carbon monoxide (60 per cent), chlorine peroxide (29 per cent), and oxygen (11 per cent), although a flame is propagated through the gases, as much as 76 per cent of the original carbon monoxide may remain unburnt H. B. Dixon and E. J. Russell, Trans. Chem. Soc., 1897, xxi., 605).

3. Dried carbon monoxide and oxygen completely combine without flame in contact with a heated platinum wire; moreover, the writer has recently proved that the most careful drying possible greatly accelerates the rate at which the gases combine in contact with a hot fire-clay surface at 500°.

4. There are certain well-established instances in which combustion is not determined by the presence of moisture -namely, the combustion of cyanogen, of carbon disulphide, and of hydrocarbons (ethane, ethylene, and acetylene).

Theories Respecting the Function of Moisture. 1. H. B. Dixon has consistently maintained that, in the ombustion of carbon monoxide, steam merely acts as the carrier of oxygen; he contends that in explosions the ormation of carbon dioxide is always limited by its dissociation, and that at the highest temperature (e.g., in the wave-front when the mixture C2N2 +202 is detonated) it is not formed at all by direct interaction of the monoxide and oxygen, because the internal energy which would thereby be imparted momentarily to the newly-born dioxide molecule would bring about its dissociation. For the same reason flame is not propagated through a dried mixture of carbon monoxide and oxygen. But, if steam be present, the interaction of CO +OH2 = CO2 + H2 would bring molecules of the dioxide into existence with a much less degree of internal agitation, and therefore capable of continued existence, whilst the hydrogen liberated would immediately combine with oxygen, forming steam, which is less easily dissociated than carbon dioxide. This explanation, whilst consistent with many of the facts connected with the combustion of carbon monoxide, cannot be extended to other well-known instances, and is particularly inapplicable to the case of hydrogen.

2. Mendeleeff, in his "Principles of Chemistry," ascribed the mutual inertness of carbon monoxide and oxygen to the circumstance that gases combine according to a supposed "law of equal volumes," or, in other words, that from the kinetic standpoint the primary changes in all cases must be considered as involving the collision of two molecules only. In the case of carbon monoxide he postulated the following cycle of changes :(i.) CO + OH2 = CO2 + H2 ; (ii.) H2 + O2 = H2O2 ; (iii.) CO + O2H2 = CO2 + OH2. But, according to this supposed "law of equal volumes," a well-dried mixture of carbon monoxide and nitrous oxide, or of carbon monoxide and ozone, should be active, whereas Dixon has proved them to be as non-explosive as a dried mixture of the monoxide and oxygen.

3. H. E. Armstrong has always contended that chemical actions cannot occur between two perfectly pure substances, but require the conjunction of an electrolyte in order to form a closed conducting system. The presence of steam, which he supposes may always be regarded as r.ndered "conducting" by association with some traces of

On the other hand, H. B. Dixon has urged that a rate of mixture of carbon monoxide and oxygen is incompatible explosion of nearly 1700 metres per second for a moist with any interaction of the complexity thus postulated. There is doubtless prima facie much force in this objection, but it is by no means fatal, seeing that the dimensions of units, and the duration of chemical action, though exthe explosion wave are incomparably greater than molecular tremely short when measured in terms of ordinary gross units of time, is at least many thousands of times greater than the intervals between successive molecular collisions A more serious objection to Armstrong's (see Note). theory is the fact that there are several well-established cases in which combustion apparently does not depend upon the presence of moisture.

(NOTE.-As the writer understands Dixon's objection to Armstrong's view, it is that whilst chemical action in the explosion wave may last a comparatively long time (i.e., during many molecular collisions), and that therefore a quintuple molecular collision might happen in that period, it is impossible for the wave to be propagated as a soundwave through quintuple collisions. Ordinary sound-waves may be many molecules thick, but they are propagated through bi-molecular collisions).

4. In 1893 Sir J. J. Thomson (Phil. Mag., xxxvi., 321) pointed out that if the forces holding the atoms together in a molecule are electrical in character, the presence of drops of any liquid (such as water) of high specific inductive capacity would probably cause a sufficient loosening of the bands between the atoms to render the molecule much more reactive. He showed that the complete drying of a gas renders it non-conductive. H. B. Baker, in his Wilde Lecture before the Manchester Literary and Philosophical Society (Man. Mem., liii., part 3), described a number of new experiments which led him to tentatively put forward the theory that chemical interchanges in gaseous systems depend upon the presence of both "ions" and water-vapour; the "ions" act as nuclei for the condensation of steam, and the liquid drops of water so formed, by virtue of their high specific inductive capacity, facilitate chemical change in the layer of gas immediately in contact with them. Chemists will await with the greatest interest the further development of this hypothesis, but the idea that such rapid changes are met with in gaseous explosions are dependent upon the formation of aggregates of steam molecules in an atmosphere containing less than four of them per 1000 millions, and that such aggregates approximate to liquid drops at the high temperatures of flames, makes large demands upon the imagination, and it will require to be supported by the strongest experimental evidence.

(To be continued).

Behaviour of Iron towards Solutions of Stannous Salts.-A. Thiel and K. Keller.-The authors have performed a series of experiments to discover whether iron actually precipitates tin from solutions of stannous salts. These experiments included both determinations of the rate of the reaction between iron and an acid solution of a tin salt, and also determinations of electromotive force, and in both cases it was found that tin is separated when iron is immersed in a solution which is not too acid and not too poor in tin. The same result was obtained analytically, and thus tin and iron behave as would be expected from their position in the tension series.-Zeit. Anorg. Chem., Ixviii., No. 3.

Sir ARCHIBALD GEIKIE, K.C.B., President, followed by Mr. A. B. KEMPE, Vice-President, in the Chair.

PAPERS were read as follows:

"Sequence of Chemical Forms in Stellar Spectra." By Sir NORMAN LOCKYER, K.C.B., F.R.S.

This paper gives in the first place a résumé of the facts obtained from a study of various celestial and laboratory spectra from 1873 to 1905, in so far as they relate to the working hypothesis formulated in 1873, that in the reversing layers of the sun and stars various degrees of celestial dissociation are at work.

It is shown that each step in the study of celestial and laboratory spectra and their inter-relation has strengthened that hypothesis, and that it has enabled a classification of stars-based on their chemistry and heat conditions-to

be made.

This classification has been recently checked by studying the physical conditions of stars, using a calcite-quartz optical train to obtain photographs of the extension of the spectra of each chemical group of stars into the ultraviolet on the same plate and under equal conditions of atmosphere and altitude. This confirmed the prior chemical results, absolute parallelism being found between the two sets of photographs.

The recent researches detailed in the paper have been made with better photographs and with increased know. ledge of the changes in spectra. This work has enabled the classification to be carried into finer details, and the heat levels at which various chemical forms are predominant have been observed with greater certainty. Thus the Alnitamian group of stars in the Kensington classification has been divided into four species. These indicate the effects of four different temperatures, associated with observed changes in spectral lines. When in the laboratory special groups of lines appear at different temperatures, each group is predominant in turn in stars of different temperatures. In the laboratory the effects of all temperatures are integrated; in stars the low temperature and

therefore its effects are absent.

Diagrams are given (1) showing the range through stars of rising temperature of the different chemical forms, and also stellar level at which each chemical form predominates; (2) relation of the four sub-divisions of the Alnitamian group to groups immediately above and below it on stellar temperature curve.

The chemical forms predominant in spectra of hottest stars are hydrogen, photohydrogen, silicium, carbon, nitrogen, oxygen, sulphur, and the cleveite gases. These researches into the finer details of stellar and laboratory spectra have afforded help in dealing with stars showing peculiar spectra. It has been shown that Ursa Majoris, a Sirian star with peculiarities, has the lines of protochromium more developed than any other star yet examined, either higher or lower on the stellar temperature scale. Other stars with peculiar spectra have been found to have sets of lines which have not yet been traced in terrestrial spectra.

"Influence of Viscosity on the Stability of the Flow of Fluids." By A. MALLOCK, F.R.S.

The effect of this paper is to call attention to an observation made by the late Mr. W. Froude, F.R.S., with regard to an experiment on fluid jets, and to its application in explaining some of the phenomena presented by the flow of viscous fluids.

The experiment referred to was shown by Froude at the Bristol Meeting of the British Association in 1875, and one

299

of the deductions drawn from it was that in a viscous flow the character of the stream differed according as to whether the flow was towards decreasing or increasing pressure.

"Atmospheric Oscillations." By HORACE LAMB,

F.R.S.

The paper treats of the free oscillations of an atmosphere, the temperature being a function of the altitude, and the adiabatic laws of expansion being assumed. In particular the case of a uniform temperature gradient is discussed in some detail. The possible oscillations are of various types, of which the most important is of the character of a longitudinal wave. The results are simplest when the equilibrium state is one of convective equilibrium, and the velocity of the longitudinal wave is then equal to the Newtonian velocity of sound, (g H), corresponding to the temperature of the lowest stratum. The bearing of the results on Kelvin's theory of the semidiurnal barometric oscillation is examined; and it appears that the existence of a free period of the earth's atmosphere, of somewhat less than twelve mean solar hours, is highly probable. Other types of oscillation depend for their frequency on the degree of stability of the atmosphere, and may under circumstances be comparatively slow. It is possible that these may account for some of the minor fluctuations of the barometer. The paper includes also an examination of the theory of waves at a surface of discontinuity.

charge in Electrolytic Gas and other Gases." By P. J. "A Theory of the Chemical Action of the Electric Dis

KIRKBY.

The complicated results attending the passage of an electric discharge through electrolytic gas, described in previous papers, and in particular the chemical effects observed in the positive columns of long discharges, were explained by the theory that the chemical action is due to molecular dissociation effected by the collisions of gaseous ions constituting the current with the molecules of the gas, the atoms of which are thus set free to enter into new combinations.

In this paper an account is given of experiments designed to determine both the number of molecules of water (w) formed by the passage of the atomic charge through 1 cm. of the positive columns of various discharges passed through electrolytic gas (2H2 + O2), and also the electric force (Y) within these columns-the results providing data for testing and developing the theory. The chemical effects of the positive column are attributed to the motion of the electrons alone. The number of atoms of oxygen set free by the collisions of an electron moving through I cm. in electrolytic gas at mm. pressure under the electric force Y is of the form cpe-bp/Y where c, bare Certain constants. Hence, if an atom of oxygen can unite directly with a molecule of hydrogen w is proportional to cpe-bp/Y.

All the experimental results satisfy this condition within error - limits in the particular form w/p=79e-427 plY, and thus support the above theory as well as the hypothesis that water vapour is formed by the collision of an atom of oxygen with a molecule of hydrogen.

This equation involves that the energy of formation of an oxygen-molecule is less than, and probably nearly equal to, 6.1 x 10-12 ergs; and experiments of Berthelot upon the heat of transformation of ozone into oxygen are shown to be in fair agreement. An estimate, 7 X 10-12 ergs, is also given for the energy of formation of a molecule of water-vapour.

These experiments also prove, independently of theory, that dissociated atoms of oxygen are not charged electrically.

Similar experiments were made with the gaseous mixture CO+H2. The chemical effects observed in the positive column reproduce the main features observed with electro lytic gas, and are explainable by a similar theory.