Frederick Soddy Nobel Lecture

382

    1921 F.SODDY

mesothorium from the radium in the fractional crystallization of chlorides

of mesothorium, radium and barium, which I so obtained. The radioactive

constituents were concentrated from the inactive barium, to an extent of

several hundred times, without the slightest change in the ratio between

them, so far as could be ascertained by the most careful radioactive measure-

ment. Thorium X was also found to be non-separable from mesothorium and

radium in agreement with Strömholm and Svedberg’s conclusion. In the

fractional crystallization Ramsay and Hahn found the radiothorium to be

separated with the barium. Their discovery of this element was entirely due

to the fact that between the separation of the radium with the mesothorium,

a certain, probably considerable, interval of time must have elapsed.

From this date I was convinced that this non-separability of the radio-

elements was a totally new phenomenon, quite distinct from that of the

most closely related pairs, or groups, of elements, hitherto observed in chem-

istry, and that the relationship was not, as usually supposed, one of close

similarity but of complete chemical identity.

"When it is considered what a powerful means radioactive methods of

measurement afford for detecting the least change in the concentration of a

pair of active substances, and the completeness and persistence of some of the

attempts at separation, which have been made, the conclusion is scarcely to

be resisted that we have in these examples no mere chemical analogues but

chemical identities." (1910, p. 285).

The case of the rare-earth elements has often been cited as an example of

mixtures of elements, once thought homogeneous, but with the application

of more refined methods, or more repeated separations, proving to be sep-

arable. But the case is not really comparable. The difficulty in the rare

earths is not so much in the actual separation, but in ascertaining what the

effect of an attempted separation has been. Every process has to be labori-

ously repeated a large number of times before a separation detectable by

the insensitive methods available, such as the determination of the chemical

equivalent, is apparent. So soon as any difference in properties is detected,

separation follows. Thus, didymium was resolved owing to the difference in

colour of the salts of its two constituents, and now the spectra afford a more

general cxiterion of the same character. But for the radio-elements the sub-

stances are known severally as individuals from the start, unrivalled methods

are available for detecting the least beginnings of separation so soon as it

occurs, and yet the constituents once mixed remain inseparable by chemical

processes

3

. Today, of course, no argument as to the complete chemical iden-

---

    O R I G I N S   O F   T H E   C O N C E P T I O N S   O F   I S O T O P E S

383

tity of isotopes is required, since half the common elements examined by the

positive-ray methods have been found to be mixtures of isotopes.

On broader and quite general philosophic grounds, and without in the

least postulating a continuation of the genetic series of the radio-elements

throughout the Periodic System, I arrived at the conclusion reached by

Strömholm and Svedberg. It was certain that among the groups of chemi-

cally identical radio-elements differences of atomic weight of whole units

must exist. Thus from the atomic weights of the parent elements and the

number of 

α

-particles expelled, the atomic weights of ionium (230) and

radiothorium (228) must differ by two and four units, respectively, from

that of the chemically identical thorium (232). Once one enquired what

evidence the chemist had for the real homogeneity of the elements as dis-

tinct from their chemical homogeneity, the conclusion followed at once

that if all the elements were mixtures in constant proportions of chemical

identities differing step-wise by whole units in atomic mass, the chemist with

his methods must have remained unaware of it.

"The recognition that elements of different atomic weight may possess

identical chemical properties seems destined to have its most important

application in the region of the inactive elements, where the absence of a

second radioactive nature, totally unconnected with the chemical nature,

makes it impossible for chemical identities to be individually detected.

Chemical homogeneity is no longer a guarantee that any supposed element

is not a mixture of several of different atomic weight or that any atomic

weight is not merely a mean number." (1910, p. 286).

The spectrum of ionium

In 1912 Exner and Haschek, using the ionium-thorium preparation of Auer

von Welsbach, already alluded to, and A. S. Russell and Rossi, using an ac-

tive ionium-thorium preparation prepared for the Royal Society, independ-

ently attempted to determine the spectrum of ionium, without, in either

case, succeeding in finding a single new line. The spectra were those of

thorium simply.

The interpretation of this remarkable result was, at the time, dependent

upon the incompletely known period of ionium. This constant fixes the

proportion of ionium by weight in the materials used. All the evidence was

in favour of ionium having a very long period, so that the proportion of

---

384

   1921 F.SODDY

ionium by weight in both preparations would be so high that the failure for

it to show any spectrum due to ionium would be offundamental importance.

The low range of the 

α

-rays from ionium indicated a period for this ele-

ment of the order of from 2 to 5 x 10

5

 years, but this evidence is but slight.

On the other hand I had been engaged continuously since 1905 in an experi-

mental effort to detect the growth of radium in carefully purified uranium.

The failure to detect this after many years led to the minimum estimate for

the average life of ionium of 10

5

 years. This depended on one condition,

that ionium was the only intermediate of long period between uranium and

radium (uranium I and II for this purpose being considered a single element).

This has now been established. Over sufficiently long periods the growth of

radium from uranium has been found to proceed according to the square of

the time. This proves that ionium is the only long-lived intermediate. Also

the actual period of ionium to an accuracy of two per cent has been found to

be 10

5

 years. So that a considerable percentage of the ionium-thorium prep-

arations must have been pure ionium. Exner and Haschek concluded that

the period of ionium must have been much overestimated, and that this ap-

peared from their result to be less than that of radium. But they realized the

impossibility of reconciling this with the other evidence. Russell and Rossi,

after reviewing all possible explanations, remarked: "But the possibility

that they (thorium and ionium) are identical in all physical and chemical

properties, and differ only in atomic weight and radioactive properties,

should not be lost sight of."

I endorsed this view:

"The simplest, if somewhat heterodox, view to take is that ionium is a

long-period element, and that its spectrum, as well as its whole chemical

behaviour, is identical with that of thorium. It is clear that the conception of

the chemical elements as necessarily homogeneous is undetermined and that

different elements with different atomic weights are chemically identical

not as an exception but as a consequence of the way in which the known dis-

integration series have been shown to run their course. In non-radioactive

matter this heterogeneity cannot be distinguished, but it is not so in the case

of matter actually in process of evolution." (1912, p. 322).

The radio-elements and the periodic law

I now pass on to the second phase of this history, the fitting of the radio-

---