CHEMICAL NEWS,

July 21, 1911

Two ratios

ON THE CONNECTION BETWEEN

29

Ag

=

=

35'457.

=

The Helium-Argon Group. Watson has determined the weight of a normal litre of helium as 0.17814 and 0.17830 grm. (Journ. Chem. Soc., xcvii., 810). Hence He 3.994. For neon, eleven determinations of the normal litre ranged from 0.8997 to o'9006 grm.; in mean, o'9002. This, by the method of limiting densities, gives Ne = 20'200. By the same method Watson (Journ. Chem. Soc., xcvii., 833) has reduced the densities of krypton and xenon, as measured in 1908 by Moore (Proc. Chem. Soc., xxiv., 273). The final 82.92 and Xe The density of argon has been re-determined by Fischer and Hähnel (Ber., xliii., 1435). The mean of seven determinations is 19.945, referred to 0 16. Hence A 39.89. Ramsay and Gray, by means of the microbalance, have been able to weigh the gaseous emanation of radium, for which they propose the name niton (Comptes Rendus, cli., 126). The values thus found for its molecular weight are 222, 216, 227, 218, 217, in mean 220. The same value was also found by Debierne by a different method (Comptes Rendus, cl., 1740).

values are Kr

=

22

=

130'22.

Miscellaneous Notes.

=

Richards and Baxter have investigated the subject of density corrections, or, in other words, the reduction of weights to a vacuum (Journ. Am. Chem. Soc., xxxii., 507). They regard the validity of the corrections, as applied at Harvard, as well established. Relations between the atomic weights have been studied by Howard (CHEMICAL NEWS, ci., 181, 265). Dubreuil has continued his recalculation of the determinations of Stas (Bull. Soc. Chem., [4], vii., 119).

OF COMPOUNDS,

AND THE CHEMICAL FORCES AT PLAY WITHIN THEIR MOLECULES.

By GEOFFREY MARTIN, B.Sc., M.Sc., Ph.D.

PROF. RICHARDS points out in his suggestive Faraday Lecture to the Chemical Society (published in abstract in CHEMICAL NEWS, July 7) that the more strongly the atoms are chemically united or attracted together in the molecule (as judged by the heat of formation), the denser (i.e., more contracted) is the resulting compound, and he goes on to suggest that the forces of chemical attraction actually compress the atom themselves.

May I be allowed to point out that this is but a particular case of a wider generalisation published by me some years ago in my book "Researches on the Affinities of the Elements" (J. and A. Churchill, 1905). For example, I showed in that work that there exists a similar relationship between the chemical forces uniting the atoms in the molecules of a compound and the volatility, fusibility, and hardness of the compound. It is this: the more strongly attracted together are the atoms in the molecule (and therefore the more stable the molecule) the more strongly attracted together are the molecules themselves, and consequently the greater is their cohesive force, involatility, and infusibility. By taking practically all the known data I showed that as a class chemically unstable bodies (i.e., bodies in which the atoms in the molecules are but feebly attracted together) are characterised by their volatility, fusibility, and lack of hardness-properties depending upon molecular attraction (cohesive force)whereas chemically stable compounds (i.e., compounds in which the chemical forces uniting the atoms in the molecule is great) are characterised by their involatility, infusi bility, and hardness. In other words, it is the internal chemical forces which the atoms exert on each other in the molecule which decides the external attractions with which the molecules themselves are attracted together and consequently properties arising out of this molecular attraction, such as volatility, fusibility, hardness, and density of the compound (see "Researches on the Affinities of the Elements," pp. 68 et seq., 111, 120, 123; also CHEMICAL NEWS, 1904, Ixxxix., 241, "The Connection between the Volatility of Compounds and the Chemical Forces at Play within the Molecule "). It is therefore very doubtful whether the fact quoted by Prof. Richards is a good reason for assuming that the atoms themselves are really compressible. The fact of the matter is that the chemical forces exerted by the atoms so enormously transcend any other forces that they exert, that they almost completely decide all the properties, both chemical and physical, exhibited by any molecular aggregation of matter-a conclusion also arrived at in my essay, and there examined in detail.

Ureïns of p-Oxyphenylaminoacetic and p Methoxyphenylaminoacetic Acids.-J. Aloy and Ch. Rabaut.By applying the method of Hugounenq and Morel to p-oxyphenylaminoacetic acid and the corresponding þ. methoxy acid the authors have prepared the mixed ureas and the symmetrical ureïns of both acids. The mixed urea of the former exists in the form of white crystals, fusing at 198°, very slightly soluble in water and soluble in alcohol. The properties of the other mixed urea are very similar; its melting-point is 193°. When heated the mixed ureas decompose, setting free aniline. The symmetrical ureïns are prepared by the action of COC12 on the sodic solutions of the corresponding acids. They are both yellow substances; that derived from p-oxyphenylaminoacetic acid decomposes when heated, but the other compound is more stable,-Bull. Soc. Chim. France, ix., No. 6,

THE experiments which form the subject of the present communication have been already briefly described in Proceedings of the Royal Society, 1909, A, lxxxii., 396; and 1910, A, lxxxiii., 483. It was thought, however, that as the subject possesses not only scientific interest, but has also important bearings on metallurgy and its allied industries, a description of the investigation giving more experimental details might be of interest.

The primary object of the experiments was to fix approximate values for the boiling-points of a number of metals, and thus clear away the very considerable uncertainty which existed in most cases. They were afterwards extended to a consideration of the dependency of the boiling-points on the pressure. The existence of such a dependency is in itself an important criterion of any method for fixing the boiling-points, particularly when dealing with the high temperatures necessary to effect the ebullition of most metals.

The problem first received attention in 1879 from Carnelley and Williams (Trans. Chem. Soc., 1879, p. 563), who attempted to determine the temperature of the vapours of a few metals by suspending tubes containing metals of known melting-point above the surface. To take an example of their results, there is little doubt that the value for tin is at least 700° C. too low. Since these experiments, very little has been done in the way of a systematic determination of metallic boiling-points. Féry (Annales de Chimie et de Physique, Sér. F, xxviii., 428), however, determined the boiling point of copper in 1903 by optical measurements in an arc furnace and also effected the fractional distillation of brass. The arc furnace is unsuitable for such measurements on account of the difficulties of power regulation and of temperature estimation. Moissan (Comptes Rendus, 1906, cxlii., 425), in the course of an extensive investigation on the vaporisation of metals in his arc furnace, obtained values giving the weight volatilised in a certain time under given conditions. From these experimental results Watts has attempted to deduce approximate values for the boiling-points of the metals ((Trans. Am. Electrochem. Soc., 1907, xii., 141). Great uncertainty is due to the considerable vapour tension possessed by many metals at temperatures well below their boiling. points, to the total lack of data giving the latent heat of

* A Paper read before the Faraday Society, May 23, 1911,

30

2000 390

drops.

Drops sent up decidedly.
Drops vigorous, half-way up
crucible.

Drops projected out of crucible.

Vaporisation, and further to the fact that the actual energy expenditure in the furnace was apparently not measured by Moissan. Moreover, this is greatly influenced by the conductivity of the vapours surrounding the arc, and many of the results are of little value by reason of the carburisation which took place. No other data were available with the exception of some values arrived at in vapour density determinations (cf., Mensching and Meyer, Annalen, 1887, CCxl., 317; and Wartenberg, Zeit. Anorg. Chem., 1908, lvi., 320), and the general position was so unsatisfactory and the data actually available so discordant that further investigation was urgently needed. With regard to the effect of pressure, only ebullition under very low pressures (Kraft, Ber., 1903, xxxvi., 1690) had been studied, except for metals like zinc of comparatively low boiling-point (Barus, Bull. U.S. Geolog. Survey, No. 103; Phil. Mag., 1890, [5], xxix., 141). The main reason for this was the

16

1770

260

18

1815

260

1860

260

1875

290

22

1880

290

23

24

25

26

Surface quiet.

agitation, no drops

thrown up.

Decided agitation, drops thrown

up gently.

Drops half-way up crucible.

Vigorous, drops projected out

of crucible.

21,

↑ C

B

difficulty of finding any material capable of withstanding reduced pressures at temperatures above 1400° C., and the consequent necessity for limiting the temperature by the use of very high vacua. This difficulty was avoided in the present experiments by arranging the whole furnace inside a vacuum enclosure.

The investigations may be divided into three main sections:

1. A study at atmospheric pressure of the boiling-points of a number of metals which do not combine appreciably with carbon (antimony, bismuth, copper, lead, magnesium, silver, tin).

2. A study at atmospheric pressure of the boiling-points of some metals which are readily attacked by carbon with the formation of carbides (aluminium, chromium, iron, manganese).

3. The influence of diminished and increased pressures on the boiling-points of certain of the metals under I (bismuth, copper, lead, silver, tin, zinc).

PART I.-Non-carburisable Metals.

In some earlier experiments carried out by Dr. L. Bradshaw the loss in weight of the metal maintained at fixed points over a wide range of temperature was determined. The loss by volatilisation, however, was found to take place over such a large temperature interval that in the present investigation other methods were adopted. Experiments were carried out in which a considerable quantity of metal, contained in a thin-walled graphite crucible, was heated in a carbon tube resistance furnace so chosen that the temperature could be quite rapidly raised, it being hoped that when the metal entered into ebullition a discontinuity in the rise of temperature would be observed, as shown by optical temperature readings on the outside of the crucible. Such, however, was not the case, the temperature of the walls rising considerably above the boiling-point of the metal. It was eventually found that the best results were to be obtained by gradually raising the temperature of the metal and taking observations of the surface from above through an absorbing glass. The surface at first remains perfectly still, but on approaching

the boiling-point a slight agitation is observed to set in, which rapidly becomes vigorous. In most cases the dif ference between the temperatures indicated, when a gentle agitation is first apparent and when the ebullition is so vigorous that globules of metal are projected to a height of over 10 cm., does not exceed 100°. By taking the boiling-point as that temperature at which ebullition first becomes decided, quite concordant results were obtained in different experiments (cf., records of experiments given in Table I.).

From observations of a number of metals in this manner it seems probable that the boiling-point may with fair approximation be taken as the temperature at which the vaporisation becomes sufficiently vigorous to cause a decided projection of drops from the surface (e.g., 1955° in the above cited three experiments with silver).

In most of the experiments the furnace employed was of the vertical carbon tube resistance type (see Fig. 1), tubes 25 cm. long and of 20 and 30 mm. internal diameter respectively being used. The ends were electro-coppered and soldered into brass castings provided with water circulation at A and B, and a side tube of carbon was fitted to the centre of the resistor, and continued at the other end by a brass fitting furnished with a window. The whole was packed in crushed wood charcoal, this being a good heat insulator, and also desirable as regards absence of vapours at high temperatures. Suspended and fitting fairly closely inside the resistance tube was a long graphite crucible, made as thin as possible (about 1 mm.) in the walls, and containing the charge of metal. This form of crucible acts as a reflux condenser, and the amount of metal does not rapidly decrease. Graphite being an excellent conductor of heat, the error due to temperature difference between the outer and inner surfaces of the crucible walls was thus rendered negligible. A depth of about 30 mm. of metal was ordinarily used, and tempera ture readings were taken on the walls of the portion thus occupied (this being arranged immediately opposite the side tube) by means of a Wanner optical pyrometer. Freedom from fumes in the side tube was secured by the passage of a current of hydrogen admitted at c. With

CHEMICAL NEWS,} July 21,

33

Using the two sizes of carbon tube mentioned, currents of about 500 and 900 ampères at 8 and 10 volts respectively were required to raise the temperature to 2500° C. moderately quickly. The temperature was under delicate control by manipulation of a rheostat in the field circuit of the separately excited dynamo supplying the current. Although experiments made up to 1500° in comparing the optical temperature readings with those of a thermo-couple immersed in the metal indicated that the difference between the temperature of the metal and that of the outer walls was negligible, it was thought desirable to obtain further proof of this. With this object in view and to make use of a somewhat different method of measurement, the apparatus shown in Fig. 2 was devised, the metal being con

this current were stopped when the metal was gently boiling ebullition ceased. On substitution of nitrogen for hydrogen the temperature readings agreed among themselves in similar experiments, but were always considerably higher (50° to 100°) than those obtained in a hydrogen atmosphere. This curious effect is probably to be explained by the ease with which hydrogen diffuses through the crucible walls, dislodging the heavy vapours above the surface of the metal.

That this action actually occurred could be distinctly seen by interrupting and then suddenly restarting the current of hydrogen whilst looking down the crucible. Measurements were in most cases made with both nitrogen and hydrogen, but it seems probable that the results obtained with the use of the latter approximate more closely to the true boiling points at atmospheric pressure.

Accuracy of the temperature measurements was ensured by the insertion of an ammeter and rheostat in the standard lamp circuit of the optical pyrometer. Comparison with a thermoelement up to 1500° showed that the readings closely agreed with the thermoelectric temperatures (due allowance being made for the difference between the two scales). A further check was secured by determining the "black body" melting-points of strips of platinum, rhodium, and iridium specially prepared in a high state of purity by Messrs. Johnson, Matthey, and Co. These strips, which were 4 mm. wide and 8 cm. long, were mounted horizontally and heated electrically. A series of results obtained for the melting-points of these metals is given below.

Platinum (deg.).

TABLE II. Rhodium (deg.).

Iridium (deg.).

FIG. 2.

tained in the annular portion of a graphite crucible of the form shown, heating being effected internally by radiation from the intensely heated central carbon rod supplied with current from the two thick graphite rods. To attain the boiling-point of lead, i.e., about 1500° C., a current of about 200 ampères at 20 volts was necessary, using a central carbon rod of 9 mm. diameter. The arrangement was fixed inside a wide carbon tube lagged with Kieselguhr, and temperature readings taken on the outside of the crucible down a side tube exactly similar to that used in the apparatus shown in Fig. 1. Here the effect of temperature errors due to imperfect thermal conductivity of the crucible walls is reversed, and the possibility of reflexion from the resistor altogether removed. Experiments with lead gave results in agreement with those obtained by the other method, but it was equally impossible to obtain limitation of the temperature of the outer surface of the crucible when the contents were caused to vigorously boil. It was therefore concluded that the apparatus shown in Fig. I was introducing no error due to this source. A curious effect was, however, traced to the current of

The values given by Holborn and Henning (Sitzungsber. K. Akad. Wiss. Berlin, 1905, xii., 311) for the "black body" melting-points are as follows: Platinum 1545°, rhodium 1650°, iridium 2000°. Seeing that the readings obtained with this optical pyrometer are now capable of reference to some definite standard, it was thought best to publish the results without further correction, all temperatures being given on the optical scale.

In the following table will be found a list of the boilingpoints experimentally determined, and also of the data previously available.

In the case of magnesium the boiling-point was determined by the use of a protected thermocouple immersed in the metal, the temperature remaining steady while the metal distilled into the upper portion of the crucible.