CHEMISTRY).
second important means of deciding between possible multiples and sub-multiples of atomic weights was the discovery of Dulong and Petit (1818) that the atomic weight of an element is about equal to a constant num ber (6.3) divided by the specific heat. To be sure, this rule is not exact ; but its inexactness is not usually great enough to affect it in its office of deciding the multiple or sub-multiple of the chemical combining proportion to be taken as the atomic weight. For example, the specific heat of calcium is about 0.16; therefore its atomic weight is shown to be about 39.4, whereas the exact value found by chemical means is 40.07.
A third method of answering the question exists in the similarity of the crystal forms of similar salts of allied elements, discovered by E. Mitscherlich in 1821. If the atomic weight of one element entering into such isomor phous crystals is unknown, that multiple of the combining propor tion of this element which corresponds to the formula indicated by the known salt will be the true atomic weight. (See ISOMORPHISM.) The full significance and essential consistency of these three methods of solving Dalton's unsolved problem were not realized until 1858, when a table of atomic weights identical in principle with that used to-day was published by S. Cannizzaro. Previous doubts concerning the criteria just described had caused many chemists to reject wholly the term "atomic weights," and to call the arbitrarily selected multiples merely by some such name as "proportion numbers" or "chemical equivalents." But the num bers now used (as regards the multiples chosen) inevitably involve the atomic theory, hence the adjective "atomic" is fitting. "Weight" also is fitting, since the values are determined by means of the gravitational balance. The term "atomic mass" applies consistently only when inertia is the basis of measurement. The term "chemical equivalent" is now used to signify the atomic weight divided by the valency (q.v.).
The choice of the standard of atomic weights has varied. Dalton chose the smallest atomic weight, that of hydrogen, as his standard. Berzelius temporarily selected Oxygen= ioo as the standard of his system. Later the chemical world returned to Dalton's practice, especially because (according to early work) it was believed that the atomic weight of oxygen is nearly the whole number 16, if hydrogen is taken as I. Finally, after it had been shown by E. W. Morley and others that the ratio of the atomic weights of oxygen and hydrogen is in fact 15.878 to 1, it was decided, by general consent, in 1905, to abandon the standard H= i •000, retaining the standard 0 =16•000. The decision was based upon convenience. The permanent choice of 0=15.878 would have changed by nearly one per cent. almost every other accepted value, and would have caused much con fusion in previous quantitative statements. Besides, more atomic weights approach whole numbers when oxygen is taken as exactly 16•000 than when any other usual standard is chosen.
A more weighty reason lay in the fact that most of the values are experimentally determined by relation to oxygen, and are referred to hydrogen only through that element. Hence any sub sequent change in the accepted ratio H :0 (one of the most diffi cult to determine of all such ratios) would affect all the atomic weights, if hydrogen were chosen as the standard substance. The present unit of the system of atomic weights is therefore exactly the sixteenth part of the atomic weight of oxygen. The atomic weight of hydrogen thus becomes 1 •o0 7 7. The choice, on the whole, was a wise one; it has been justified by modern research, and has proved to be peculiarly fortunate, because probably all atoms of oxygen are alike in weight (see ISOTOPES) .
Atomic weights are numbers; that is to say, they represent ratios and are there fore devoid of physical dimensions. They are, however, very different from the quantities designated by J. A. R. Newlands and H. G. J. Moseley "atomic numbers" (q.v.), which record the serial order of the places in the periodic classification of the chemical elements. No immediate knowledge of the actual weights of indi vidual atoms is afforded by "atomic weights," unless the number of atoms in a given gross weight of some elementary substance is known. Various researches have shown that 16 grammes of oxygen contain about 6o6X
atoms; hence a single atom of oxygen must weigh o•000,000,000,000,000,000,000,026,4 gramme. The actual weights of other kinds of atoms must be in due pro portion.
The exact values of the chemical combining proportions which form the basis of the table of atomic weights are found only by experimental work. There fore, before the table is given, the necessary experimental meth ods may well be briefly described. The first and most generally useful method employed for the purpose has as its object the determination of the precise amount of one element which is necessary exactly to combine with a given amount of some other element of known atomic weight. The experimental technique is of the most refined quantitative chemical analysis. Early extensive and careful investigations of this kind were conducted by J. Ber zelius, C. de Marignac, J. B. A. Dumas, J. S. Stas and many others. Recently most of the work in this direction has been con ducted in the United States (E. W. Morley, W. A. Noyes, T. W. Richards, G. P. Baxter and others), although European investi gators (especially B. Brauner and O. Honigschmid) have made important contributions.
Experimental work of this kind naturally involves the observ ance of a number of essential conditions. Comparatively few compounds of any given element are fit to serve as a means of determining its atomic weight, for the reason that comparatively few substances may be prepared in a perfectly pure state. The choice of the compounds to be employed is in some ways the most crucial part of the whole process, for with some compounds no result worthy of consideration could be obtained, even using the greatest care possible.
Having chosen wisely, the experimenter must prepare the need ful substances, whatever they may be, in a state of very great purity. He must never forget that every precipitate carries down with it contaminating impurities absorbed or included by the sub stance as it separates from the solution. He must remember always that no receptacle necessary to contain the substance is free from the possibility of being attacked or dissolved, thus affecting the result. Moreover, precipitates are never wholly in soluble ; and most substances will volatilise if heated to an exces sive temperature. These complicating circumstances combine often in unexpected ways to introduce impurity, and the experi menter must not only guard against these dangers, but must prove by adequate tests that no such complication has occurred. Moreover, above all, he must not forget that oxygen, nitrogen and water are almost omnipresent ; and continual care must be exercised lest in some way one of these impurities may affect the substance which is serving as the basis of the work. For further statement of these and other precautions and for a brief descrip tion of apparatus suitable for avoiding many pitfalls, together with the details of an especially instructive complex case, the reader is referred to Carnegie Institution of Washington, Publica tion No. 125. A critical summary by F. W. Clarke of all investiga tions up to 1920 is to be found in the third Memoir of vol. xvi. of the Memoirs of the National Academy of Sciences (Washing ton).
A simple case may best exemplify the method. In one of many experiments, 7.59712 grammes of ferric oxide
prepared with the greatest care, were found to yield on reduction (by means of hydrogen at a high tempera ture) 5.31364 grammes of metallic iron. The loss of weight (2.28348 grammes) represents the oxygen present in the oxide. Hence, from the proportion
(16•000) :2x, the atomic weight of iron is found to be 55.848 (G. P. Baxter and C. R. Hoover). The analysis was repeated many times in order to eliminate accidental errors.
Another general method of determin ing atomic weights (applicable only to gases or vapours) depends upon Avogadro's Rule, and resolves itself into the weighing of like volumes of different gases under like conditions of tempera ture and pressure. This is the only gravimetric method applicable to the six inert gases (helium, etc.) which do not form chemical compounds. The method determines molecular weights, not atomic weights; but the number of atoms in a molecule may be inferred in other ways, and therefore the atomic weights may be calculated from the data. The method involves experimental diffi culties. The globe containing a gas inevitably weighs much more than the gas itself and is peculiarly subject to changes of buoy ancy of the air. The exact measurement of temperature and pressure is not always easy, nor is the perfect purity of the gas to be weighed a condition readily secured. Moreover, Avogadro's Rule holds only for perfect gases; no actual gas fulfils exactly its requirements, because of the bulk occupied by the molecules themselves and their mutual attraction. On the whole, making allowance for these difficulties (see STOICHEIOMETRY), the method of determining molecular (and therefore atomic weights) by comparison of the densities of gases agrees remarkably well with the results obtained from chemical analysis (Lord Rayleigh, E. W. Morley, P. A. Guye, A. Leduc, E. Moles, G. P. Baxter).
A third method of determining atomic weights (like the last, a purely physical method) is that which determines the mass (or rather the ratio of mass to electric charge) of rapidly moving charged atoms or molecules by means of the deflection by electric and magnetic fields. It appraises (by means of impressions on a photographic plate of the positions of impact of the deflected particles) the relative atomic masses pertaining to selected groups of atoms. In its original form it furnished the first experimental evidence not only that in some elements the atoms are all alike in weight, but also that in other elements this is not, the case (Sir J. J. Thomson, 1912). Different varieties of a single chemical element, similar in every respect except as regards the weights and masses of their atoms, and apparently inseparable by natural agencies when once mixed, are called iso topes (F. Soddy). Under that head will be found a full descrip tion of this method of evaluating them, which was greatly im proved by F. W. Aston, in his "mass-spectrograph." Isotopes.—Many but not all of the elementary substances have been found by this third method to be isotopic or "complex." Hence elements may be divided into two classes : simple elements, probably possessing only one variety of atom, and isotopic ele ments, containing two or more varieties. The relative propor tions of the several isotopes in a given elementary substance are shown roughly by the relative intensities of the "photographic" records; they can be shown exactly only by quantitative analysis, and then only when no more than two isotopes are present. Thus ordinary terrestrial chlorine (C1=35.46) must consist of a mix ture of about 3o atoms of C1=37 to every ioo atoms of
Although the term "atomic weight" referred originally to the elementary substances (whether simple or isotopic) which actually occur on the earth's surface, it is applicable with even greater fit ness to each isotope alone. Of all the isotopic elements only one, namely lead, has had the atomic weight of any individual isotope accurately determined by chemical analysis (Richards, Soddy, Honigschmid). The individual isotopes of this metal are unique because, so far as we can tell, they are end-products of the spon taneous disintegration of uranium, and other radioactive elements, in which the atoms of lead were segregated at the moment of their terrestrial birth and confined in the minerals producing them.
Their abnormal atomic weights (determined by chemical methods of unquestioned trustworthiness) constituted at first the most con vincing evidence of the existence of isotopes.
The following table of atomic weights of the chemical elementary substances as they exist on the surface of the earth is essentially the table issued in 1925 by the International Committee on Elements and Atomic Weights, but includes the newly discovered element hafnium, as well as two of the individual isotopes of lead which have been experi mentally investigated by chemical methods. "Atomic numbers" are also given. Usually, the larger the atomic weight the larger the atomic number; but all isotopes of a given element have the same atomic number. Except for hydrogen, the atomic number is never more and usually less than half of the atomic weight.
Redefinition of Term.—The discovery of the spontaneous disintegration of radioactive elements and the finding of isotopes have modified our theoretical interpretation of the atomic weights. Because of these discoveries, two a priori premises (of a more or less philosophical nature), namely, first, the assumption that the atoms are indivisible (the elementary substances being absolutely permanent) and, second, the assumption that the atoms of a given chemical element are all alike in weight, must to-day be aban doned, but the premises are seen on close scrutiny to be by no means an essential part of the chemical atomic theory. Neverthe less, the old definition of atomic weights must be altered in order to correspond exactly to modern knowledge. A more complete and precise definition may be worded as follows: "Primarily, atomic weights are appropriate simple multiples (decided by theory) of the relative combining proportions or relative gas densities of elementary substances calculated on a consistent basis. They represent the relative average weights of the atoms of given specimens of elementary substances referred to a common stand ard." Any such definition involves other definitions. An ele mentary substance is a substance which is not further disinte grated by ordinary chemical reactions. This definition avoids the implication that such a substance is incapable of disintegration by extra-chemical means. "Element" and "chemical elements" are sometimes used synonymously. "Atoms" are postulated as the smallest particles of such a substance under ordinary conditions. They are not necessarily incapable of disintegration under extreme conditions. Hence their name (from a privative and .roµos "divided, cut") is not now appropriate, but it will doubtless be retained ; the term "chemical atom" would perhaps be better. The qualification involved in the word "average" above is neces sary because of the discovery of isotopes. The weighted average of the atomic weights of the isotopes in any particular isotopic or "complex" elementary substance is that which is recorded as its atomic weight.
Constancy of Atomic Weights.—That the atomic weights are constant in different compounds is shown by the analysis of many pure substances containing the same element and also by H. Landolt's experiments (1907), which proved that there is no loss or gain of gravitational effect in ordinary chemical reactions within one part in ten million. Moreover, specimens of various elementary substances (e.g., sodium, calcium, copper, silver, iron, nickel, cobalt, etc.) found in different parts of the earth or even in meteorites, have been found by careful research to have con stant atomic weights independent of geographical occurrence. All the samples of terrestrial lead even, except those found in uranium or thorium minerals, show similar uniformity. That each native terrestrial mixture of isotopes is thus unvarying seems to show that each was commingled when the earth was still fluid, or else that some unknown law determines the proportion in which the isotopes are formed. If it were not for the consistency indicated in this paragraph, the table of atomic weights would be much less useful than it is. The atomic weights are precisely consistent also with the electrochemical equivalents indicated by Faraday's Law (Faraday, Rayleigh, Richards), affording thus further evidence of their fundamental nature.
Hydrogen and Other Elements.—The hypothesis of Prout (181s) that all elements are aggregates of hydrogen has been greatly strengthened by the discovery of isotopes; for it appears that the fractions in the table above are due chiefly to isotopic mixtures, in which each isotope taken separately has nearly a whole number for its individual atomic weight. The atomic weights of uranium, radium, thorium, the isotopes of lead, and helium furnish an argument in favour of the theory of the atomic disintegration in which they are concerned, and therefore support the postulate maintaining the composite nature of the elements. Nevertheless, all the simple elements and individual isotopes have atomic weights somewhat less than the appropriate multiples of that of hydrogen, as has been shown in the case of oxygen. Many theorists believe that this common deficiency is due to the actual loss of mass during the atomic coalescence of hydrogen nuclei, the expelled mass being transformed into energy. If this is true, the exact values of the simple atomic weights (and those of individual isotopes) even to the third decimal place, possess great theoretical interest, since they must furnish an essential clue to the amount of energy expended. Modern hypotheses concerning the structure of the atom (Sir E. Rutherford, Sir J. J. Thomson, N. Bohr, G. N. Lewis, I. Langmuir) assume that practically all the weight and mass of the atom (fixing, of course, its atomic weight) are con centrated in an exceedingly small nucleus in its centre.
Concord with Atomic Numbers.—For so years the atomic weights decided the arrangement of the periodic system of the elements. Recently x-ray spectra have more certainly evalu ated the atomic numbers which place the elements in this system (Moseley) ; but the agreement between the two methods is close enough to indicate a fundamental if sometimes complex relation between them.
The sun and stars ap pear spectroscopically to be made largely of the elements existing on earth. It is therefore no mere flight of fancy to infer that the vast gravitational forces which regulate the motions of the heavenly bodies are due to the collective action of countless myriads of atoms, whose individual shares in the process are recorded in the table of atomic weights. The foregoing considera tions concerning atomic weights suggest many other cosmological inferences, which are, however, beyond the scope of this article (see "Atomic Weights and Isotopes," Chemical Review, I. I., [1924] ). It is not too much to say that these unique numbers, the atomic weights, probably bear a very close relation to the unknown fundamental processes which determined the nature and evolution of the universe. (T. W. R.)
