ELEMENT 87

M. Perey*)

HISTORICAL SURVEY

In 1879 Mendeleev¹ predicted, on the basis of his periodic system of the elements, the existence of an alkali element “eka-cesium” with an atomic weight higher than that of cesium. According to the law discovered by Moseley, this element has atomic number 87. Many attempts were made to isolate this element; investigators who thought that they had discovered it gave it various names: “Russium” (Dobroserdov²), “Alkalinium” (Loring³), “Virginium” (Allison, Bishop, Sommer, and Christensen⁴), and “Moldavium” (Hulubei⁵). Element 87 may be contained in cesium minerals, and may also have a radioactive origin.

1. Search for a stable isotope of element 87. These investigations were carried out with cesium minerals.

In 1926 Bainbridge⁶ gave an estimate of the sensitivity of the various methods used for the discovery of eka-cesium: determination of atomic masses, X-ray spectroscopy, and methods based on the study of optical spectra—flame, arc, and spark spectra. Bainbridge himself, working by the most sensitive of these methods—the mass-spectrographic method—showed that if eka-cesium is present in cesium minerals, then it is at a concentration of less than \(3.5 \cdot 10^{-7}\) relative to cesium—in pollucite, and less than \(7.3 \cdot 10^{-6}\)—in lepidolite.

Using the magneto-optical method, Allison and his co-workers⁴, ⁷ came to the conclusion that element 87 is contained in certain minerals—pollucite, lepidolite, samarskite, and monazite—as well as in seawater and in the brine of Lake Searles at concentrations from \(10^{-12}\) to \(10^{-8}\). These results were subjected to very thorough criticism, and some authors⁸ even deny the existence of the Allison magneto-optical effect altogether.

In an X-ray spectral analysis carried out by Papish and Wainer⁹ in 1931 of samarskite containing uranium, niobium, tantalum, potassium, rubidium, and cesium, the authors found wavelengths \(M_{\alpha_1} = 4.517\ \text{Å}\); \(L_{\alpha_1} = 1.026\ \text{Å}\); \(L_{\alpha_2} = 1.035\ \text{Å}\); \(L_{\beta_2} = 0.853\ \text{Å}\) and \(L = 0.944\ \text{Å}\), belong-

*) Journal de Chimie Physique, 43, 152 (1946). Translated by N. Fuks.

...which, according to Moseley’s diagram, belonged to element 87. Hirsch^10, who repeated these experiments with the same analyzer crystal (calcite), found defects in the latter that caused a shift of the spectral lines, and therefore cast doubt on the previous results.

In 1936, studying preparations made from pollucite (a double aluminum–cesium silicate), Hulubei^11 found in the radiographs he obtained weak lines corresponding to the following wavelengths predicted by Moseley’s diagram: \(87L_{\alpha_1}=1.032\ \text{\AA}\); \(87L_{\alpha_2}=1.043\ \text{\AA}\). In 1937 Hulubei^12 discovered still weaker lines: \(87L_{\beta_1}=0.838\ \text{\AA}\); \(87L_{\beta_2}=0.856\ \text{\AA}\) and \(87L_{\gamma}=0.715\ \text{\AA}\). In 1937 Hirsch^10 unsuccessfully attempted to obtain the X-ray spectrum of element 87 with a sample of cesium sulfate extracted from lepidolite, and suggested that the rays indicated by Hulubei belonged to mercury, which gives wavelengths very close to those stated above. In 1943 Hirsch^13 confirmed, by new experiments, the objections he had earlier raised to Hulubei’s work. In Hirsch’s opinion, his own experiments prove conclusively that no stable element 87 exists.

1. Searches for a radioactive isotope of element 87. Kreiston^14 detected \(\alpha\)-radiation in a freshly prepared preparation of mesothorium 2; however, Hahn^15, Hahn and Erbacher^16, and Hevesy^17 refuted this observation and came to the conclusion that eka-cesium cannot be formed from the actinium isotope—mesothorium 2.

In 1933, Goben^18, who resumed the study of mesothorium 2, reported the \(\alpha\)-radiation he had observed, with a range of \(3.03\ \text{cm}\). Goben considered the source of these \(\alpha\)-rays to be mesothorium 2, which in this process is transformed into eka-cesium, but he gave no data whatever on the eka-cesium thus formed. The ratio of \(\alpha\)-decay to \(\beta\)-decay is, according to Goben’s data, \(10^{-6}\) or \(10^{-10}\). Since these results have received no further confirmation, it seems premature to us to draw any conclusions from them, especially given the very small value of the indicated ratio.

Thus, neither in cesium minerals nor in mesothorium 2 has anyone succeeded in proving more or less convincingly the presence of element 87. As the history of the discovery of new elements in recent times shows, when some physical method leads one to assume the presence of minute traces of a new element, this assumption cannot be considered proven until it has been confirmed by other physical and chemical methods.

DISCOVERY OF ELEMENT 87^19

Introduction

In order to study accurately the change with time of the activity of the \(\beta\)-radiation emitted by actinium purified from decay products, I began to measure the increase of this activity as

could be earlier, following the final purification of actinium, in order thereby to exclude the influence of the activity of the decay products.

During the first 2 hours after the completion of the purification I found that the activity of the β-radiation at first increases with a period of about 20 min., then remains constant for some time, and finally increases again owing to the formation of the decay products of actinium—radioactinium, actinium \(X_2\), and the active deposit. At first I supposed that the initial increase of activity might be caused by the formation of active deposit AcB, an isotope of lead, from traces of actinium \(X\) not removed during the purification. However, an attempt to get rid of this activity by precipitating lead sulphide was unsuccessful.

The indicated β-radiation is half absorbed by an aluminium layer 0.18 mm thick \((\mu = 39\ \mathrm{cm}^{-1})\), whereas the radiation of the active deposit is weakened by a factor of two in passing through 0.23 mm of aluminium \((\mu = 30\ \mathrm{cm}^{-1})\). Thus it seemed logical to ascribe the above-mentioned initial increase of the β-activity of actinium, as yet detected by no one, to the formation from it of a new element removed in the purification of actinium. In order to determine the chemical nature of this new natural radioactive element, I tried to ascertain at what stage of the purification it is separated from actinium. The mother liquor obtained at the last stage of the separation can contain only alkali and ammonium salts and, as it seemed at first, should not possess radioactivity. However, on evaporation to dryness it gives a residue possessing β-activity with the above-mentioned decay period of 21 min. Subsequent precipitation of cerium hydroxide (removal of RAc), lead sulphide (removal of the active deposit), and barium carbonate (removal of AcX) does not eliminate the indicated activity, so that, apparently, it is connected with some alkali element.

If one assumes that the latter is formed as a result of the emission of α-particles from actinium, it must occupy the 87th place in the periodic system. For proof I tried to detect the chemical similarity of this element to caesium by joint crystallization of both; for this purpose caesium perchlorate was chosen, since in solubility it differs from the readily soluble perchlorates of the non-alkali metals. Having added caesium chloride to the mother liquor remaining from the final purification of actinium, and carrying out precipitation with a solution of sodium perchlorate, I obtained radioactive crystals of caesium perchlorate; their activity decreased exponentially with a period of \(21 \pm 1\) minutes.

Thus we are indeed dealing here with a higher homologue of caesium with atomic number 87. Its formation from actinium is possible only through the emission of α-particles; for this reason I began to look for α-radiation in actinium purified from the products of its decay, and I did indeed discover such radiation with a range of 3.5 cm in air, reduced to normal conditions. This radiation already

had been noted earlier[^20], but had been ascribed to traces of protactinium[^21]. I have proved that there was no protactinium in my preparation.

It followed from the experiments, therefore, that this natural radioactive element, possessing β-activity with a period of 21 min., has atomic number 87 and is formed in the α-decay of actinium.

The new element was named actinium K (AcK). However, I propose, while retaining this name for the given isotope, to give the general name “francium” (Fa) to the element with ordinal number 87.

Since actinium K is a decay product of actinium, knowledge of the properties of the latter is necessary in the investigation of actinium K. I therefore found it advisable to begin with a systematic account of the properties of actinium, and only then to proceed to actinium K.

The following questions are set forth below: the chemical properties of actinium; methods for the quantitative determination of actinium; methods for the enrichment of actinium in the rare earths containing it; the radioactive properties of actinium—β-, γ-, and α-radiation.

In my next article the chemical properties of AcK will be presented, and in a third article—its physical properties, a new method for the quantitative determination of actinium by means of AcK, and general conclusions.

ACTINIUM, ORIGIN OF ELEMENT 87

As indicated in the introduction, element 87—actinium K—is formed from actinium.

Chemical properties of actinium

The natural radioactive element actinium, discovered by Debierne[^22], occupies place 89 in Mendeleev’s table, i.e., the last place in the group of trivalent elements. It is a homologue of lanthanum and is close to the rare elements in its chemical properties.

Actinium is extracted from uranium minerals together with the rare earths; to obtain preparations enriched in actinium, repeated fractionation is required. Pitchblende, in which actinium was discovered, contains a comparatively small quantity of rare earths, mainly cerium earths: lanthanum, cerium, praseodymium, neodymium, samarium, and also traces of thorium. The latter is the least basic element in this mixture, and lanthanum the most basic. Actinium somewhat surpasses lanthanum in its basic properties. Actinium is precipitated together with thorium and the rare-earth elements in the form of the hydroxide, carbonate, fluoride, and oxalate (the last, provided that the precipitation is carried out in a medium with an acidity not greater than 0.3N HCl or HNO₃).

The separation of actinium from thorium and cerium presents no particular difficulty. Thorium is tetravalent, and for its separation use is made of the insolubility of its pyrophosphate, precipitated by a solution of sodium pyrophosphate in 0.3N HCl.

Cerium can be tri- and tetravalent. To separate it from other rare-earth elements, cerium is converted into insoluble tetravalent compounds by means of potassium permanganate:

\[ 6\,\mathrm{CeCl}_3 + 2\,\mathrm{MnO}_4 + 8\mathrm{H}_2\mathrm{O} = 6\mathrm{CeO}_2 + 2\mathrm{MnO}_2 + 2\mathrm{KCl} + 16\mathrm{HCl}. \]

Sodium carbonate should be added to the permanganate solution (at the rate of 4 moles per mole of \(\mathrm{KMnO}_4\)) in order to neutralize the hydrochloric acid that is formed, which otherwise would convert \(\mathrm{CeO}_2\) into solution.

Methods for the quantitative determination of actinium

As is known, the actinium content cannot be determined directly from its radiation, but only from the \(\beta\)- or \(\gamma\)-radiation of its active precipitate, for which about 3 months are required—the time necessary for equilibrium to be reached between actinium and the active precipitate.

\[ \begin{aligned} &\mathrm{Ac}\xrightarrow[22\ \text{yr}]{\alpha} \begin{matrix} \mathrm{AcK}\\[-2pt] \mathrm{RAc} \end{matrix} \xrightarrow[ \begin{array}{c} 21\ \text{min}\\ 18.9\ \text{days} \end{array} ]{\beta,\alpha} \mathrm{AcX} \xrightarrow[11.2\ \text{days}]{\alpha} \mathrm{An} \xrightarrow[3.9\ \text{s}]{\alpha} \mathrm{AcA} \xrightarrow[2\cdot10^{-3}\ \text{s}]{\alpha} \mathrm{Ac}\,' \longrightarrow \\[6pt] &\qquad\qquad \xrightarrow[38\ \text{min}]{\beta} \mathrm{AcC} \begin{cases} \xrightarrow[2.16\ \text{min}]{\beta}\mathrm{AcC}\xrightarrow[10^{-3}\ \text{s}]{\alpha}\mathrm{AcD}\ \text{stable},\\ \xrightarrow{\alpha}\mathrm{AcC}'\xrightarrow[4.7\ \text{min}]{\beta}\mathrm{AcD}\ \text{stable}. \end{cases} \end{aligned} \]

The determination of actinium from \(\beta\)-radiation was carried out in an ionization chamber of the type adopted at the Paris Radium Institute, closed with aluminum foil \(0.02\ \mathrm{mm}\) thick. Actinium-containing preparations obtained from rare earths were placed in flat copper cuvettes \(3.5\ \mathrm{cm}\) in diameter and \(0.15\ \mathrm{cm}\) deep, which were then hermetically sealed with calibrated cover glasses by means of molten paraffin.

If the preparations were too active for measurement of \(\beta\)-radiation (at 5 or 40 cm from the ionization chamber), they were measured by \(\gamma\)-radiation by means of a large condenser shielded with a lead plate \(1\ \mathrm{cm}\) thick. The measurements were made by comparison with a radium standard measured under the same conditions. I shall not dwell on the method developed by us for measuring actinium activity\(^{23}\), but shall confine myself to giving the results obtained in this way. Taking into account the diffusion of radiation in the cuvette, we had:

\[ \begin{aligned} 100\ \text{esu/sec for }\beta\text{-rays at a distance of }5\ \mathrm{cm} &\text{ from the chamber}\ \longrightarrow 0.665 \pm 0.040\ \mathrm{mg}\ \mathrm{Ra},\\ 100\ \text{esu/sec for }\gamma\text{-rays through a }1\ \mathrm{cm}\ \text{layer of Pb} &\longrightarrow 20.6 \pm 2\ \text{millicuries Ac}. \end{aligned} \]

If the diffusion of radiation is not taken into account, the corresponding figures will be: 0.51 mg Ra and 15.8 millicuries Ac.

The determination of the magnitude of the activity in millicuries was made by means of a calorimeter.

In these measurements it is necessary to work with pure actinium preparations, since the presence of other radioactive substances would distort the results.

Enrichment of rare earths containing actinium

To obtain enriched preparations of actinium from rare earths, several methods may be used:

a) The method of fractional precipitation of rare-earth oxalates in a nitric-acid medium.^24 In this case actinium is concentrated in the mother liquors, as shown by the scheme I below. At the same time I determined the content of lanthanum and didymium in the various fractions. The weakening of the absorption bands of didymium in the fractions enriched with actinium proves that the latter approaches lanthanum in its properties.

Diagram and results of fractional precipitation of rare-earth oxalates containing actinium in a nitric-acid medium

Oxalate → 3075 g La₂O₃
+ HNO₃ → nitrates
+ (COOH)₂ + H₂O

[3] 549 g La₂O₃ solution → 2.5 h → precipitate — 2526 g La₂O₃ + HNO₃ → nitrates + (COOH)₂ 1, 2

[5] 2095 g solution → 2 h → precipitate — 2322 g La₂O₃ + HNO₃ + (COOH)₂ 4

7 673 g solution → 3 h → precipitate — 1573 g La₂O₃ + HNO₃ + (COOH)₂ 6

9 178 g solution → 7+9 → 2 h → precipitate — 1441 g La₂O₃ [8]

[11] 150 g solution → 2.5 h → precipitate — 721 g La₂O₃ [10]

□ — Products obtained during fractionation

Distribution of fractions by weight:

Distribution of fractions by activity:

Fractions 3, 5, 11 → 908 g La₂O₃ → 1754 el. st. units combined
and again fractionated according to the same scheme.
Fractions 8 and 10 → 2168 g La₂O₃ → 53 el. st. units discarded.

Scheme I.

b) The method of fractional precipitation with sodium carbonate. According to the data of Jovanović^25 on mesothorium 2, an isotope of actinium, Basile^26 found that the first portions of the precipitate are only slightly active and that the last portion is 10 times more active than the first, provided that work is carried out only with dilute solutions containing actinium lanthanum. In concentrated solutions different results are obtained; in this case sodium carbonate precipitates lanthanum in an amount exceeding the theoretical.

c) The method of fractionating the double nitrates of the rare earths and ammonium in nitric-acid solution.^27 Actinium is concentrated together with lanthanum in the least soluble fractions.

d) The method of fractionation of double nitrates of rare earths and magnesium. As has already been indicated by Debierne and Urbain and M. Curie and Tacvorian,^28 and as I have convinced myself from my own experience, actinium behaves in this separation method in a somewhat unusual way (see scheme II). Lanthanum is separated in the less soluble fractions, as in work by method (c), but without actinium; the latter is concentrated in the fractions between neodymium and samarium and is thus separated from lanthanum, in contrast to the other methods.

Fractionation of double nitrates of magnesium and of actinium-containing rare earths

□ Preparations obtained during fractionation

Scheme II.

I consider the best fractionation method, giving identical results both with preparations poor in actinium and with those very rich in actinium, to be the method recommended by Marie Curie of precipitation in the form of oxalate in a nitric-acid medium. The fractions enriched in actinium obtained by this method may be combined, as scheme I shows, and subjected to repeated fractionation.

The indicated methods of fractionation made it possible to obtain very concentrated actinium preparations (10 millicuries of actinium in 5 mg of La₂O₃); however, the latter has not yet been isolated in the pure state.

Radioactive properties of actinium

Let us recall that the initial element in the actinium family is actino-uranium, an isotope of uranium with a period of \(7.1 \cdot 10^8\) years, contained in natural uranium in an amount of 0.7%. From it uranium Y is formed.

with a period of 1.02 days, then protactinium with a period of 32,000 years, from which actinium is formed in the process of α-decay. The decay product of actinium is radiothorium, an isotope of thorium with a period of 18.9 days, which, with the emission of an α-particle, is transformed into actinium X—an isotope of radium with a period of 11.2 days. The latter in turn decays into an α-particle and actinon (the emanation of actinium), which gives an entire series of elements of the actinium deposit, ending with actinium lead.

\[ \mathrm{AcU}\xrightarrow{\alpha}\mathrm{UY}\xrightarrow{\beta}\mathrm{Pa}\xrightarrow{\alpha}\mathrm{Ac} \begin{cases} \xrightarrow{\alpha}\mathrm{AcK}\xrightarrow{\beta}\mathrm{AcX}\\ \xrightarrow{\beta}\mathrm{RAc}\xrightarrow{\alpha}\mathrm{AcX} \end{cases} \xrightarrow{\alpha}\mathrm{An}\longrightarrow \]

\[ \mathrm{AcA}\xrightarrow{\alpha}\mathrm{AcB}\xrightarrow{\beta}\mathrm{AcC} \begin{cases} \xrightarrow{\alpha}\mathrm{AcC''}\xrightarrow{\beta}\mathrm{AcPb}\\ \xrightarrow{\beta}\mathrm{AcC'}\xrightarrow{\alpha}\mathrm{AcPb} \end{cases} \]

The period of actinium was for a long time known only approximately: different authors gave for it a value from 7 to 22 years. Irène Curie-Joliot and Bouissières\({}^{29}\), on the basis of their measurements, arrived at the value 21.7 years, very close to that found by Marie Curie—21 years.

Radiation of actinium

To determine the radiation of actinium it is necessary to prepare a preparation purified of all decay products. Radiothorium, an isotope of thorium, is removed by coprecipitation with cerium oxide hydrate in the presence of oxygen\({}^{29}\).

Actinium B, an isotope of lead contained in the active deposit and having a period of 36 min., is removed by coprecipitation with lead sulfide (the action of hydrogen sulfide in the presence of a lead salt).

Then the lanthanum containing the actinium is precipitated by carbonate-free ammonia in the presence of barium chloride. In this case AcX, which is an isotope of radium and a homologue of barium, remains in solution. Immediately before the measurements, the final purification of the preparation is rapidly carried out so that an appreciable quantity of decay products does not have time to form again.

Observing the radiation of the preparations during the first several hours after purification, I found that the activity of β-radiation at first increases (Fig. 1), the magnitude of the period being approximately 20 min., i.e. different from all previously known periods. Then the activity remains almost constant for some time and, finally, increases again owing to the formation of decay products: RAc, AcX, and the active deposit. At the same time the theoretical curve of the radiation of actinium itself (Fig. 2), which cannot be observed experimentally, touches the axis of abscissas.

Thus, the formation of a new element can be detected only with very rapid work. On the other hand, extrapolating,

the β-radiation curve, showing the formation of the element, to the initial moment of time, coinciding with the end of the last purification of the preparation, one can verify that at this moment the activity of the β-radiation was equal to zero. It follows from this that the β-radiation of actinium itself could not have distorted the results obtained.

Fig. 1. Observed β-activity of the actinium preparation as a function of the time elapsed from the moment of its purification.

The measurements were carried out with an ionization chamber covered with aluminum foil \(0.01\ \mathrm{mm}\) thick and located \(5\ \mathrm{cm}\) from the preparation under study.

Thus, the β-rays of actinium have a range of less than \(7\ \mathrm{cm}\), i.e., an energy of less than \(80\) electron-kilovolts. This contradicts the data of Hall, Libby, and Latimer \(^{30}\), who, even before the discovery of actinium K, obtained for the upper limit of the actinium spectrum the value \(220\ \mathrm{KeV}\).

Together with Lequeux \(^{31,32}\), I began an investigation of the β-radiation of actinium by means of a low-pressure Wilson chamber constructed by Joliot \(^{33}\). The preparation used in these experiments was lanthanum with a very high content of actinium, purified from the products of its decay.

The preparation was applied to a cellophane sheet in a very thin layer, the weight of which was less than \(0.1\ \mathrm{mg}/\mathrm{cm}^{2}\). To check the purity of the actinium in each experiment, the cellophane sheet was divided into 2 parts, and each part was separately measured in a total-ionization chamber. Then one part was placed in the Wilson chamber, while observation of the increase in activity of the other part was continued at the same time. The curve of increase of the ionization current made it possible to verify the initial purity of the preparation, and by extrapolation to the initial moment (the end of purification) it was possible to determine the value of the branching coefficient for the process \(\mathrm{Ac} \xrightarrow{\alpha} \mathrm{AcK}\), to which we shall return below.

Fig. 2. Theoretical curve of the change in the β-activity of actinium itself after purification.

From the measurements made in the ionization chamber with the part of the preparation later placed in the Wilson chamber, and from the curve of increase in the activity of the other part, it is possible to determine the number of actinium atoms...

... contained in the preparation. In all the experiments mentioned, AcK was in equilibrium with actinium.

Measurements in a Wilson chamber at ordinary pressure show that the number of β-particles with energy exceeding 50 KeV (particles with a range greater than 4 cm) is equal to the number of α-particles emitted in the transformation Ac ⟶ AcK. The maximum energy of these fast β-particles is 200 KeV, and they may be ascribed to the β-radiation of actinium K.

As already indicated, Hahn, Libby, and Latimer, even before the discovery of actinium K, found for the upper limit of the actinium spectrum the value 220 KeV. Since these authors undoubtedly carried out their measurements too soon after the completion of the purification, they could not have observed the formation of AcK, which was already present in equilibrium concentration by the beginning of the measurements. Thus, they were in fact measuring not the β-radiation of actinium, but that of actinium K (Fig. 3). The number of β-particles with energies less than 50 KeV is approximately 4 times greater than the number of fast β-particles.

Fig. 73. β-activity of an actinium preparation as a function of the time elapsed since its purification.

At an initial pressure in the Wilson chamber of less than 200 mm Hg, β-rays with energies above 100 KeV are no longer detected. In 200 photographs we found a large number of soft β-particles, whose energy can be determined from the range. Statistical treatment of the photographs showed that, of 1150 measured β-particles, 240 had a range greater than 4 cm, i.e. energy \(>20\) KeV; 650 had a range between 2 and 4 cm, i.e. energy from 10 to 20 KeV; and 240 had a range less than 2 cm, i.e. energy \(<10\) KeV. Among the latter, more than 200 particles had a range between 1 and 2 cm, and very few—less than 1 cm. The mean activity of the preparations in these experiments corresponded to the emission of 190 α-particles per second at the beginning of the experiment, and the branching coefficient for α-decay was \(1.2\%\). Thus, the number of actinium atoms undergoing β-decay is \(15.8 \cdot 10^3\ \mathrm{sec}^{-1}\), or 70 during the expansion time in the Wilson chamber, whose effective duration was 0.0045 sec. On average, 5.75 β-particles are observed on each photograph. For 12 disintegrating actinium atoms there is evidently the emission of only one β-particle.

Attempts to detect β-rays with energy less than 5 KeV are doomed to failure, since a layer of water 1 μ thick on the surface of the preparation is sufficient for complete absorption of these rays.

Thus, the method described cannot provide precise information about the continuous spectrum of the β-radiation of actinium itself. Nevertheless, the method makes it possible—

allows one to conclude that this spectrum has a very low upper limit, apparently below 10 KeV, and that the β-particles whose energy distribution was discussed above have a secondary origin; for example, they were formed as a result of the ejection of an electron from the \(L\) level by a γ-ray with an energy of 30 to 40 KeV.

γ-radiation of actinium. Proceeding from the hypothesis of the secondary origin of the fast β-rays observed in the transformation \(Ac \longrightarrow RAc\), Lecoin, Perey, and Tsien-San-Tsin \(^{34,35}\) investigated the γ-radiation of actinium by means of a xenon ionization chamber possessing maximum sensitivity to soft γ-rays with energies from 20 to 30 KeV.

Fig. 4. \(\alpha\)- and β-activity of an actinium preparation as a function of the time elapsed since its purification.

Fig. 5. Penetrating power of the \(\alpha\)-radiation of an actinium preparation in air.

The β-rays emitted by the preparation were removed by means of a sufficiently intense transverse magnetic field.

The agreement of the value of the γ-ray energy found in this way with the results obtained in a Wilson chamber at reduced pressure permits the conclusion that actinium emits γ-rays with an energy of 37 KeV.

α-radiation of actinium. If a small quantity of a lanthanum preparation containing actinium, just thoroughly purified of all decay products, is taken, applied in a thin layer, and the change of its activity with time is studied, then entirely different results are obtained for the \(\alpha\)- and β-radiation (Fig. 4). The curve of the increase of \(\alpha\)-radiation differs sharply from the curve of β-radiation; the formation of a new element is not reflected in it, and the increase in activity is due to the formation of radioactinium, actinium X, and the active deposit. The initial point on the curve corresponds to the end of the operation of removing \(AcK\) from the preparation; the removal of \(RAc\) was completed 2 hours earlier. As is seen from the curve, actinium purified of all decay products possesses \(\alpha\)-activity.

Investigation of the $\alpha$-radiation emitted by a thin layer of a purified actinium preparation, by means of a proportional amplifier, indicates the presence of $\alpha$-rays with a range of $3.5 \pm 0.2\ \text{cm}$ under normal conditions. These rays are emitted, in my opinion, by actinium itself (Fig. 5). The range is not specified more precisely, since I intend in the future to carry out a more detailed investigation of this radiation, which I was unable to do in the work presented here.

In view of the fact that the range of these $\alpha$-particles is very close to the corresponding value for the $\alpha$-particles of protactinium ($3.54\ \text{cm}^{30}$), it was necessary to make sure that the latter was absent from our preparations. For this purpose I added a known quantity of protactinium to an inactive preparation consisting of the same rare earths as the active preparations, and measured the distribution of protactinium in the various fractions obtained at the different stages of purification of the actinium preparations from decay products.

After removal of RAc by means of cerium hydroxide, precipitated with traces of ammonia in an oxygen-containing solution, only $0.1\%$ of the protactinium taken remains in the latter. Since, in purifying actinium, about ten such precipitations are carried out, as a result no traces of protactinium should remain in the purified preparation.

As early as 1914 Meyer, Hess, and Paneth$^{20}$ detected very weak $\alpha$-radiation in actinium with a range of $3.4\ \text{cm}$. One of these authors$^{21}$ subsequently ascribed this radiation to the presence of traces of protactinium, discovered at that time,$^{31}$ and possessing $\alpha$-radiation very close to that observed by the authors mentioned. However, in the well-known handbook of Hevesy and Paneth,$^{35}$ published in 1938, the $\alpha$-radiation of actinium is not mentioned at all.

Branching coefficient of Ac

\[ \mathrm{Ac}\begin{cases} \xrightarrow{\alpha} \mathrm{AcK},\\ \xrightarrow{\beta} \mathrm{RAc}. \end{cases} \]

The ratio of the numbers of actinium atoms decaying with the emission of $\alpha$- and $\beta$-particles can, to a first approximation, be determined from the change with time in the $\alpha$- or $\beta$-activity of a preparation purified from decay products. Taking into account that the removal of radioactinium is carried out 2 hours before the removal of AcK, i.e. before the zero point in Fig. 4, the initial $\alpha$-activity of the preparation due to the radiation of actinium itself is equal to $0.2$ electrostatic units. The increase in the activity of the preparation over 24 hours, caused by the formation of RAc, amounts to $1.2$ electrostatic units (Fig. 6). The decay constant of radioactinium has the value

\[ \lambda_{\mathrm{RAc}}=\frac{\ln 2}{T}=\frac{0.693}{18.9}=0.0366\ \text{day}^{-1}. \]

Thus the branching coefficient is equal to

\[ \frac{0.2 \times 0.0366}{1.2}=0.6\%. \]

The number obtained represents a lower limit, since we have not taken into account the difference in the ranges of the $\alpha$-particles from RAc (4.5 cm) and Ac (3.5 cm).

On the other hand, the $\beta$-radiation of actinium in equilibrium with the products of its decay is due to AcB and AcC′. If it is assumed that each of these elements is the source of approximately 50% of the ionization and that, in respect to the magnitude of the mean penetrating power of their radiation, they differ little from AcK, then from the fact that the activity in the horizontal part of the curve of increase of the $\beta$-radiation of actinium is equal to 0.005 of the total activity (in the equilibrium state), there follows a value of the indicated ratio of about 1%.

Fig. 6. $\alpha$-activity of an actinium preparation, caused by the formation of RAc and by the intrinsic activity of Ac.

In the work carried out by me together with Lecoq12, we succeeded in refining the value of the branching coefficient. Details concerning the preparation of the actinium preparations have been set forth above. The measurements were made in a total-ionization chamber. Taking into account the degree of accuracy of our measurements and calculations, it may be considered that the ionization was produced practically only by $\alpha$-rays.

Let $I$ be the activity of the preparation, measured under these conditions, at the moment $t$.

$I_0$ is the same at the initial moment $t_0$.

$r$ is the ratio of the number $n_0$ of Ac atoms decaying per unit time with the emission of $\alpha$-particles to the total number $N_0$ of decaying atoms (branching coefficient).

Since at the beginning of the experiment the preparation contains neither AcX nor RAc, the current $I_0$ is due to the $\alpha$-rays of actinium. These rays, with a range of 3.5 cm, give in the total-ionization chamber $7.1 \cdot 10^{-5}$ electrostatic units of quantity of electricity for each $\alpha$-particle, so that $I_0 = n_0 \cdot 7.1 \cdot 10^{-5}$. At the moment $t$ the current $I$ will be equal to $I_0 + I'$, where $I'$ is the current caused by the $\alpha$-rays from RAc and AcX formed from Ac during the time $t$. To calculate $I'$ we shall write out the values of the decay constants of Ac, RAc, and AcX:

\[ \text{for Ac } \lambda = 0.00009\ \text{day}^{-1}\quad (T = 21.7\ \text{years}), \]

\[ \text{for RAc } \lambda' = 0.0366\ \text{day}^{-1}\quad (T' = 18.9\ \text{days}), \]

\[ \text{for AcX } \lambda'' = 0.0617\ \text{day}^{-1}\quad (T'' = 11.2\ \text{days}). \]

The numbers of atoms of RAc and AcX that have decayed during time \(t\), respectively, are equal to:

\[ a=\frac{\lambda' N_{\mathrm{RAc}}}{(\lambda N_{\mathrm{Ac}})_{t_0}} =1{,}0038t\left(e^{-0{,}00009t}-e^{-0{,}0366t}\right), \]

\[ b=\frac{\lambda'' N_{\mathrm{AcX}}}{(\lambda N_{\mathrm{Ac}})_{t_0}} =1{,}00612\left(e^{-\lambda' t}-2{,}4526\,e^{-\lambda'' t} +1{,}4526\,e^{-\lambda''' t}\right). \]

One \(\alpha\)-particle from RAc gives in the chamber of total ionization \(8{,}1\cdot10^{-5}\) electrostatic units of electricity, while an \(\alpha\)-particle from AcX together with the almost simultaneously emitted \(\alpha\)-particles from Ac, AcA, and AcC gives \(36{,}8\cdot10^{-5}\) electrostatic units. Thus, at the moment \(t\) the total current caused by the decay of \(N_0\) atoms of actinium is:

\[ I=I_0+(N_0-n_0)\left(a\cdot8{,}1\cdot10^{-5} +b\cdot36{,}8\cdot10^{-5}\right). \]

Extrapolating the curve of the increase in activity to the moment \(t_0\), one can determine \(I_0\), and, knowing \(I_0\), easily calculate the branching coefficient \(r=n_0/N_0\). In all our experiments we obtained for it the value \(1{,}2\pm0{,}1\%\), which indicates the purity of the ten preparations with which we worked.

The indicated number is the most accurate value of the branching coefficient of actinium: \(\mathrm{Ac}\xrightarrow{\alpha}\mathrm{AcK}\). The decay process of actinium with emission of an \(\alpha\)-particle, discovered by me, leads to the formation of a new element, a \(\beta\)-emitter, with a period of 21 min.

In subsequent articles the physical and chemical properties of this element and its identification with element 87 will be described.

CITED LITERATURE

1. Менделеев, Lieb. Ann. Suppl. 8, 196 (1872).
2. Добросердов, Укр. Хим. Журн. 8, 497 (1925).
3. Loring, Chem. News 140, 180 (1929).
4. Allison, Bishop, Sommer, Christensen, J. Am. chem. Soc. 54, 613 (1932).
5. Hulubei, Comptes Rend. 202, 1927 (1936).
6. Bainbridge, Phys. Rev. 34, 752 (1929).
7. Allison, Murphey, Phys. Rev. 35, 285 (1930).
8. Comstock, Phys. Rev. 51, 776 (1937).
9. Papish, Wainer, J. Am. Chem. Soc. 53, 3818 (1931).
10. Hirsch, Phys. Rev. 51, 584 (1937).
11. Hulubei, Comptes Rend. 202, 1927 (1936).
12. Hulubei, Comptes Rend. 205, 854 (1937).
13. Hirsch, Phys. Rev. 63, 93 (1943).
14. Cranston, Phil. Mag. 25, 712 (1913).
15. Hahn, Phys. Zschr. 9, 245—392 (1908).
16. Hahn, Erbacher, Phys. Zschr. 27, 531 (1926).
17. Hevesy, Danske Vid. Selsk. Math. Fys. Meddl. 7, 11 (1926).
18. Gueben, Thèse. Univ. Liège (1933).
19. M. Perey, Comptes Rend. 208, 97 (1939); J. Phys. 10, 435 (1939).
20. Meyer, Hess, Paneth, Wien—Ber. 123, 1459 (1914).

1. Meyer, Schweidler, Radioaktivität, 471 (1927).
2. A. Debierne, Comptes Rend. 129, 593 (1899); 130, 206 (1900).
3. M. Lecoin, M. Perey, A. Pompei, J. Phys. (1945). (In press).
4. M. Curie, Radioactivité, p. 138, 1935.
5. Iowanovitch, J. Chim. Phys. 18, (1927).
6. M. Bachelet, Thèse Fac. Sc. Paris, p. 47 (1939).
7. M. Curie, Traité de Radioactivité, p. 138 (1935).
8. Maurice Curie, Takvorian, Comptes Rend. 198, 1687 (1934).
9. I. Joliot-Curie, Bouissières, Cahiers de Phys. No. 26 (1944).
     Joliot-Curie, J. Phys. 3, 57 (1932).
10. Hull, Libby, Latimer, J. Am. Chem. Soc. 57, 1649 (1935).
11. M. Perey, M. Lecoin, Comptes Rend. 217, 106 (1943).
12. M. Perey, M. Lecoin, Cahiers de Phys. No. 20 (1944).
13. F. Joliot, J. Phys. 5, 216 (1934).
14. M. Lecoin, M. Perey, Tsien San-Tsiang, Comptes Rend. 146, 217 (1943).
15. M. Lecoin, M. Perey, Tsien San-Tsiang, Cahiers de Phys. No. 26 (1944).
16. Bachelet, Bouissières, Tsien San-Tsiang (1945). (In press).
17. Hahn, Meitner, Phys. Zschr. 19, 203 (1918); Soddy, Cranston, Proc. Roy. Soc. 49 A, 384 (1918).
18. Paneth, Hevesy, Manual of Radioactivity (1938).