The iodine is evidently removed from solution more rapidly but not so regularly. The solvent action of the solution becomes apparent at the higher temperatures and the penetration effect is thereby increased. Filtration was carried out as rapidly as possible and through warmed funnels.

Case iii. Potassium Permanganate, N.10.Temperature 12° C. Estimation by adding 10 cc. of filtrate from charcoal (birch dust) to 10 cc. of acidified N/10 oxalic acid at about 60° C. and estimating excess acid with N/10 potassium permanganate.

t.

(mins.).

Per cent. sorbtion.

k.

K.

I

16.2

0.174

0.012

23.5 33.8

0.090

0.012

Potassium Permanganate. 4 grms. per litre.

T= 14° C. Birch charcoal dust as before

1. (mins.).

Per cent.

Coarse charcoals were found unsuitable for velocity investigation.

Animal charcoal gave a very rapid adsorption effect, this being followed by a slow and irregular absorption effect.

(B). Velocity of Sorbtion of Gases.

The gases used were NH, and HCl. Air was slowly aspirated through strong ammonia and strong hydrochloric acid respectively and the gases dried before being passed over the charcoal. The amount of NH, adsorbed by a given weight of charcoal was determined by adding the charcoal, after adsorption, to a known excess of standard sulphuric acid. After standing half an hour the charcoal was filtered off and washed and the amount of acid remaining in filtrate and washings determined by using standard caustic soda.

The HCl adsorbed was determined by adding to a known excess of standard NaOH, the un-neutralised NaOH being estimated by a standard acid.

The charcoals used were heated for a short time before use and cooled in as high a vacuum as possible (20 cc. of mercury). This process releases adsorbed air and possibly contained hydrocarbons, thus rendering more surface available for adsorption.

I. Cocoanut Charcoal, Ammonia. T=13° C. Time (mins.) 2 3 4 6 9 15

Vol. of NH,

at N.T.P. per

I

Cc, charcoal 10.8 18.9 24 27.3 32.1 37 41.1 After 60 minutes 68.3 cc. adsorbed, Cocoanut Charcoal, HCI.

II.

Time (mins.)

Vol. HCl. per

CC. charcoal

T=13° C.

3'9 7:0 10'0 13.2 17.6 23'1 After an hour 463 cc. adsorbed. The rate at which NH, is taken in slows down much sooner than in the case of HCl. Possibly in presence of traces of moisture, the acid has a solvent effect and so opens up fresh surface in the charcoal as penetration proceeds.

Adsorption is usually ver yrapid, but in these cases very little gas was adsorbed during the first minute. This may be due to the adsorbed air not being properly driven out before admitting the gases in question. Accurate estimation was also found to be difficult.

Further work is being done on this question. The values of k and K are given :

The values are calculated assuming a=68.3 and 463 cc. respectively. Those for HCI seem to show that fresh surface is continually being opened up. In the case of ammonia the values of K are so nearly constant as to suggest that 68.3 is the true capacity per cc. of the charcoal for ainmonia under these particular conditions.

When a charcoal charged with ammonia is placed in acid most of the ammonia is neutralised at once. This is the adsorbed gas. The charcoal, if rapidly removed from the acid, soon smells of ammonia. This is due to release of abCharcoal charged with HCl behaves

sorbed gas.

similarly in presence of alkali. This effect depends on the kind of charcoal and on its state of division-the larger the particles the greater the secondary effect.

THE POROSITY AND VOLUME CHANGES OF CLAY FIREBRICKS AT FURNACE TEMPERATURES.*

By GEORGE A. LOOMIS.

THIS paper deals with the permanent changes in porosity and volume of clay fire-bricks when reburned to temperatures at or above those to which they were originally fired. These were measured for a series of temperatures to determine what relation, if any, might exist between these changes and the deformation of the same bricks under load at furnace temperatures. The possibility of such a relation is suggested by the fact that contraction of clay on heating and decrease in porosity are, to a certain extent, indications of the amount of softening of the mass due to the action of fluxes present and hence indicative of decreased resistance to deformation under pressure or decreased viscosity. Softening point determinations were also made to determine what relations these might bear to the results of the load test.

The results of tests on a large number of clay fire-bricks from various parts of the country show that bricks which withstand a load test of 40 pounds per square inch at 1350° C. without marked deformation show no marked changes in porosity or volume up to 1425° C. Bricks which do not withstand the test generally show appreciable contraction or expansion, accompanied by considerable decrease in porosity. Bricks which showed overburning and the development of vesicular structure below 1425° C., by marked expansion or increase in porosity, invariably failed under load. In general, bricks which show a decrease in porosity exceeding 5 per cent or a volume change exceeding 3 per cent (amounting to approximately I per cent in linear dimensions) when refired to 1400° C. fail to pass the load test.

No definite relation could be determined between the softening point of a brick and its ability to withstand pressure at high temperature. All bricks softening below cone 28 failed completely in the load test. Some showing quite high softening points also failed, probably due to the use of an inferior bond clay in the mixture or too small an amount of bonding material.-Journal Franklin Institute, June, 1920.

* Abstract of Technologic Paper No. 159, U.S. Bureau of Standards.

CAFFEINE FROM COFFEE SOOT. SUGGESTION FOR RECLAIMING A PORTION OF THE CONSTITUENTS WHICH ARE VOLATILISED IN THE ROASTING PROCESS.

By GEORGE E. ÉWE, Philadelphia, Pa.

THE Soot which collects in the flues and on the upper and inner surface of coffee roasters frequently contains enough caffeine to warrant its use as a raw material for the production of this valuable substance. Since there is a considerable demand for caffeine, the collection of coffee soot,

if established upon a profitable basis, would result in an added source of income to coffee roasting firms.

In order to interest roasters in the collection of the "soot," a statement regarding the probable price which might be obtained for it is pertinent. No market price has been thoroughly established for this article, because it varies greatly in caffeine content. Specimens recently examined in the Pharmaceutical Research Laboratories of the H. K. Mulford Co., Philadelphia, ranged all the way from 008 to 22.2 per cent in caffeine content. Some Analyses of Coffee Soot.

9

10

...

...

12

13

...

14

Flue of roaster.

7.10 Flue of roaster.

11.20

0.08

442 15.30

Flue of roaster.

Flue of roaster.

Ceiling of roasting room.
Flue of roaster.

(a) None of the commercial roasters from which these samples were obtained, were equipped with collectors for the express purpose of collecting the "soot," and only one (No. 5) was equipped with a dust collector. Since tea fluff, tea siftings, and damaged tea, which contain from 1 to 5 per cent of caffeine, are commonly used raw materials for the production of caffeine, it is evident that collections of coffee soot with similar caffeine contents would be excellent material for the production of caffeine and should command a price approximately equal to that of these tea materials. These tea materials were quoted around 10 cents per lb. during December, 1919. It is encouraging to report here that one-half of the coffee soot samples which we have examined possessed caffeine contents which were well above the maximum content of the starting materials obtained from tea.

A statement regarding the probable amount of soot collectable from roasters in a given time would also be pertinent, but unfortunately this cannot be offered, for the reason that no commercial roaster equipped with a soot collector was met with in this investigation. Only a properly designed and properly operated soot collector will yield figures for this.

By coffee soot is meant the smoke-like vapour which arises from the roasting barrel during the roasting process. Coffee chaff, which is also a by-product of the roasting of coffee, also contains caffeine, but its economic use for the production of caffeine has not been rendered possible up to the present time. It contains much smaller proportions of caffeine than the raw materials from tea, and in addition contains considerable pyroligneous or tarry matter which makes the production of pure white caffeine very difficult and expensive. Specimens of coffee chaff recently examined in the Mulford laboratories ranged between 6 and 11 per cent in caffeine content.

Methods of collecting Coffee Soot. Caffeine is a sublimable substance, that is, it can be made to pass into the form of a vapour by heat; and upon cooling this vapour, the caffeine will be precipitated as a crystalline "snow." It is by the principle of sublimation that caffeine is collected; since the utilisable constituent of coffee soot and flue gases from the roasting of coffee is caffeine, it is by sublimation that the caffeine containing coffee soot is best collected.

Coffee soot escapes from the roaster via the flue. In practice, a considerable proportion also escapes into the room in which the roaster is situated, and where the collection of the soot is made a practice for profit, it is a source of loss of income. This loss can be prevented by proper regulation of the draught in the flue as described later in this article. Since coffee soot escapes via the flue, it is necessary, therefore, to connect the collector with the flue. Theoretically, the conditions required are a means of cooling the coffee soot and flue gases to precipitate the caffeine contained in them; a collector to retain the precipitated caffeine and soot; and a draught regulator to control the rate of flow of the soot and gases through the collector so that it is not so fast that caffeine passes through the collector and is lost in the outer air, nor so slow that the soot is lost by being forced out into the air of the room in which the roaster is situated.

Details of a Coffee Soot Collector. The collector consists of a water-jacketed, sheetiron or cast-iron box equipped with baffle plates arranged so as to make a tortuous path for the soot from the coffee-roaster. It is connected in an upright position with the flue of the roaster.

If the resistance of the baffle plates is too great to permit the passage of the soot and flue gases, an electric fan must be installed in the exit pipe of the collector. The suction thus created must be just enough to prevent the soot and gases from coming out into the room in which the roaster is situated, and not enough to carry any of the chaff up into the collector.

The water jacket is required to be operated only in the summer, and may not be necessary at all in connection with smaller roasters.

The glass windows with incandescent lamps before them, in the sides of the collector, are required during the installation of the collector when the best conditions for operation are being established.

The door of the collector is attached very loosely so as to allow a final tight adjustment by means of the four screw clamps. A soot-tight joint is obtained by means of an asbestos or composition gasket fixed in a slot around the inner edge of the door.

The interior of the collector is painted with aluminium paint, to prevent rust from forming and becoming mixed with the soot.

The collected soot is removed by releasing the screw clamps, throwing back the door on its hinges and scraping out the soot with a long steel or wooden blade.

Only the soot, and not the chaff, possesses any degree of commercial value, therefore the collector will probably not be applicable to the collection of soot from the type of roaster in the flue of which a strong blower must be employed, for the reason that the chaff and soot are usually in

separably mixed by the blower. The collector can be applied to the collection of soot from a roaster equipped with a blower by reducing the speed of the blower, so that none of the chaff is blown up into the collector, the speed of the blower being only enough to prevent the escape of soot and flue gases into the room in which the roaster is situated. If the purpose of the blower is the removal of the chaff, reduction of its speed to collect the soot will nullify this purpose and add an operation to the roasting process, namely the removal of the chaff. This will increase the expense of coffee-roasting process, and should be debited against the returns to be expected from the collection of the soot.

For either Gas or Coal Roasters.

The collector is applicable to the collection of coffee soot from either gas-fired or coal-fired roasters, but it must be remembered that coal soot or gas soot have no value, and therefore efforts must not be directed toward the production of these types of soot.

No dimensions are indicated in the drawing of the collector, as any size can be used according to requirements; merely keeping the dimensions in approximately the same ratio as they appear in the drawing. For a single roaster, it is likely that a collector 6 feet in height will completely collect the soot.

To the writer's knowledge the collection of coffee soot is not being practised in this country, but is in Continental European countries with reputed satisfactory returns. The Continental preference for more thoroughly roasted coffee may be a factor in this respect, because the soot would consequently be richer in caffeine. Whether coffee soot can be established as a profitable source of caffeine in this country is problematical and is a challenge to our best efforts.-Tea and Coffee Trade Journal, March, 1920.

ACCORDING to the well-established RutherfordBohr theory, all the positive electricity in an atom is concentrated in a nucleus at its centre. The dimensions of this nucleus are negligibly small compared with those of the rest of the atom, its diameter being of the order of o'00001 of that of the atom. The charge on the nucleus is an integral multiple of the charge of an electron, but, of course, opposite in sign. The remainder of this atom consists of electrons arranged in space about the nucleus, the normal number of such electrons (called the atomic number) being equal to the number of unit positive charges on the nucleus, so that the atom as a whole is electrically neutral. If the number of electrons in the atom exceeds the atomic number we have a negatively charged atom or ion, while in the reverse case a positively charged atom or ion results. The atomic number of any element has been found to be equal to the ordinal number of the element in the periodic table. Thus hydrogen has the atomic number 1, helium 2, lithium 3, carbon 6, neon 10, chlorine 17, nickel 28, silver 47, cerium 58, tungFrom the Journ. of Indus. and Eng. Chem., April, 1920.

sten 74, radium 88, and uranium 92. The atomic numbers can be determined experimentally from the X-ray spectrum, so that we are not dependent upon the periodic table for our knowledge of these numbers.

Bohr, Sommerfeld, and others have developed an extensive and very successful theory of spectra upon the hypothesis that the electrons in atoms are in rapid rotation in plane orbits about the nucleus in much the same way as the planets revolve around the sun. Stark, Parson, and G. N. Lewis, on the other hand, starting from chemical evidence, have assumed that the electrons are stationary in position. It should be noted that Bohr's theory has had its greatest success when applied to atoms or ions containing only one electron and that it seems incapable of explaining the chemical or ordinary physical properties of even such simple elements as lithium, carbon, or neon.

The two theories can, however, be reconciled if we consider that the electrons, as a result of forces which they exert on one another, rotate about certain definite positions in the atom which are distributed symmetrically in three dimensions. Thus for atoms containing only a single electron the chemical theory is in agreement with Bohr's theory. But for an atom such as neon the eight electrons in the outside layer would revolve around positions which are located about the nucleus in the same way that the eight corners of a cube are arranged about the centre of the cube. This structure is not inconsistent with those parts of Bohr's theory which have received experimental confirmation. In fact, Born and Landé (Verh. d. phys. Ges., 1918, xx., 210) starting with Bohr's theory and without knowledge of Lewis' work, arrived at exactly this conception of the structure of atoms (i.e., the cubic atom) from a study of the compressibility of the salts of the alkali metals.

The atomic numbers and the properties of the inert gases furnish us with a clue to the arrangement of the electrons within atoms. The low boiling point, the high ionizing potential, the chemical inertness, &c., of helium prove that the arrangement of the electrons in the helium atom is more stable than that in any other atom. Since this atom contains two electrons we must conclude that a pair of electrons in the presence of a nucleus represents a very stable group. It is reasonable that with elements of higher atomic numbers there should be an even greater tendency for this stable pair of electrons to form about the nucleus. There are two sets of facts which furnish conclusive experimental evidence that this stable pair exists in all atoms above helium.

In the first place, the properties of lithium, beryllium, &c., show that in these elements also the first two electrons are held firmly while the remainder are easily detached. Thus, lithium readily forms a univalent positive ion by the detachment of one of the three electrons in its neutral atom. The divalence and other properties of beryllium prove that there is little or no tendency for a second stable pair of electrons to surround the first pair.

In the second place, the absence of irregularities in the observed K and L series of the X-ray spectra of the various elements proves that there are no sudden changes in the number of electrons in the innermost layers of electrons about the nucleus. From these two sets of facts, as well as

from other evidence, we may take it as a fundamental principle that the arrangement of the inner electrons undergo no change as we pass from elements of smaller to those of higher atomic number.

The properties of neon indicate that its atoms are more stable than those of any other element except helium. Since the atomic number is 10, and the first two electrons form a stable pair about the nucleus as in the helium atom, it follows directly that the other eight electrons arrange themselves in a second layer or shell possessing a very high stability. If these eight electrons revolved about the nucleus in a single circular orbit or ring, as would be suggested by Bohr's theory, there is no apparent reason why there should be any very great difference in stability between rings having 7, 8 or 9 electrons. On the other hand, we readily see that the geometrical symmetry of the arrangement of the 8 electrons at (or rotating about) the 8 corners of a cube would not only account for a high degree of stability but for the fact that an arrangement of 7 or 9 electrons would have no such stability. Chemical considerations and Born and Landé's work on compressibility also lead us to this spatial arrangement of the electrons. We shall refer to the stable group of 8 electrons by the term octet. From the principles already enunciated it is clear that in the atoms of all the elements above neon the inner electrons are arranged in the same way as those of neon.

From the atomic numbers of the inert gases we are thus able to determine the number of electrons in the various layers or shells of electrons which exist in the atoms. The results are summarised in Table 1.

TABLE 1-Distribution of Electron in the
Various Shells.

Thus the xenon atom with an atomic number 54 contains 54 electrons arranged as follows:Close to the nucleus are two electrons which constitute the first shell. This is surrounded by the second shell which contains two "layers" of 8 electrons each. The third shell, which in the xenon atom is the outside shell, contains 18 electrons.

An examination of the numbers of electrons in the layers (Table 1, 2nd column) shows that they bear a simple mathematical relation to each other, namely, that they are proportional to the squares of the successive integers 1, 2, 3, and 4. This is to be looked upon as perhaps the most fundamental fact underlying the periodic arrangement of the elements. It is significant that in Bohr's theory these same numbers, 1, 4, 9, 16, &c., play a prominent part. Thus the energies of the electron in the various "stationary states" are proportional to 1, 1/4, 1/9, 1/16, &c., and the diameters of the various possible orbits in Bohr's theory are proportional to 1, 4, 9, 16, &c. In Bohr's theory the various stationary states correspond to different number of quanta (Planck's quantum theory), the innermost orbit corresponding to one

quantum, the second orbit to two quanta, &c. We should thus consider (Table 1) that the electrons in the 1st shell are monoquantic, those in both layers of the 2nd shell are diquantic, &c. It is interesting that Born and Landé, from quite other evidence, have concluded that the outermost electrons of the chlorine atom (2nd layer of the 2nd shell) are diquantic instead of triquantic, as was at first assumed.

The foregoing theory of the arrangement of electrons in atoms explains the general features of the entire periodic system of the elements and is particularly successful in accounting for the position and the properties of the so-called 8th group and the rare earth elements. It also serves to correlate the magnetic properties of the elements.

Let us now consider the bearing of this theory of atomic structure on the phenomena of chemical valence. The outstanding feature of the theory, is that there are certain groups of electrons, such as the pair in the first shell and the octet in the second, that have a remarkable stability. Those atoms in which all the electrons form parts of such stable groups (viz., the inert gases) will have no tendency to change the arrangement of their electrons and will thus not undergo chemical change. Suppose, however, we bring together an atom of fluorine (N=9) and an atom of sodium (N = 11), (We will denote the atomic number of an element by N.). Ten electrons are needed for the stable pair in the first shell and the octet in the second shell, as in the neon atom. The sodium atom has one more electron than is needed to give this stable structure while the fluorine atom has one electron too few. It is obvious then that the extra electron of the sodium atom should pass over completely to the fluorine atom. This leaves the sodium atom with a single positive charge while the florine becomes negatively charged. If the two charged atoms or ions were alone in space they would be drawn together by the electrostatic force and would move as a unit and thus constitute a molecule. (It is convenient and it has been customary with many physicists to speak of a charged atom or molecule as an ion, irrespective of whether or not the particle is able to wander under the influence of an electric field. The writer has used the term in this way in his recent publications. This practice is very distasteful to many physical chemists and is apt to be misunderstood by them. Nevertheless, it seems to me probable, especially in view of the recent work of Milner and Ghosh, that it will be desirable to abandon the physical chemists' definition of the ion and to apply it to all charged atoms or molecules. The ion which wanders may then be referred to as a "free ion"). However, if other sodium and fluorine ions are brought into contact with the "molecule" they will be attracted as well as the first one was. There will result (at not too high a temperature) a space lattice consisting of alternate positive and negative ions and the "molecule" of sodium fluoride will have disappeared. Now this is just the structure which we find experimentally for sodium fluoride by Bragg's method of X-ray crystal analysis. There are no bonds linking individual pairs of atoms together. The salt is an electrolytic conductor only in so far as its ions are free to move. In the molten condition or when dissolved in water, therefore, it becomes a good conductor.