[Über den Nachweis und das Verhalten der bei der Bestrahlung des Urans mittels Neutronen entstehenden Erdalkalimetalle]
O. HAHN AND F. STRASSMANN
Die Naturwissenschaften 27, p. 11-15 (January 1939).
[Translation in American Journal of Physics, January 1964, p. 9-15]
This article in the original German.
[N.B. the placement of the figures bear no relation to their placement in the original. You may click each figure for a larger image. The istopes discussed below are as follows: "Ra I" = unclear, "Ra II = Ba-141, "Ra III" = Ba-139, "Ra IV" = Ba-140 and "Ac I" = unclear, "Ac II" = La-141, "Ac III" = La-139, "Ac IV" = La-140.]
In a recent preliminary article in this journal2 it was reported that when uranium is irradiated by neutrons, there are several new radioisotopes produced, other than the transuranic elements -- from 93 to 96 -- previously described by Meitner, Hahn, and Strassmann. These new radioactive products are apparently due to the decay of U239 by the successive emission of two alpha particles. By this process the element with a nuclear charge of 92 must decay to a nuclear charge of 88; that is, to radium. In the previously mentioned article a tentative decay scheme was proposed. The three radium isotopes with their approximate half-lives given, decay to actinium, which in turn decay to thorium isotopes.
A rather unexpected observation was pointed out, namely that these radium isotopes, which are produced by alpha emission and which in turn decay to thorium, are obtained not only with fast but also with slow neutrons.
The evidence that these three new parent isomers are actually radium was that they can be separated together with barium salts, and that they have all the chemical reactions which are characteristic of the element barium. All the other known elements, from the transuranic ones down through uranium, protactinium, thorium, and actinium have different chemical properties than barium and are easily separated from it. The same thing holds true for the elements below radium, that is, bismuth, lead, polonium, and ekacesium (now called francium). Therefore, radium is the only possibility, if one eliminates barium itself.
In the following, the separation of the mixture of isotopes and the isolation of each species is described. From the changes in the activity of the various isotopes their half-lives can be found, and also the decay products can be determined. The half-lives of the daughter decay products cannot be fully described in this article however because of the complexity of the process. There are at least three, and probably four, isomeric decay chains, each one with three species. The half-lives of all the daughter products could not be thoroughly investigated so far.
Barium was of course always used as a carrier for the "radium isotope." As a first step, one can precipitate the barium as barium sulfate, which is the least soluble barium salt after the chromate. However, due to previous experience and some preliminary work, this method of separating the "radium isotope" by means of barium sulfate was not used. The reason was that this precipitate also carries with it a small amount of uranium, and also a not negligible quantity of actinium and thorium isotopes. These are the supposed decay products of the "radium isotope," and therefore they would prevent making a pure preparation of the primary decay products. Instead of the sulfate precipitate, barium chloride was chosen as precipitating agent, which is only very slightly soluble in strong hydrochloric acid. This method worked very well.
When uranium is bombarded with slow neutrons, it is not easy to understand from energy considerations how radium isotopes can be produced. Therefore, a very careful determination of the chemical properties of the new artificially made radioelements was necessary. Various analytic groups of elements were separated from a solution containing the irradiated uranium. Besides the large group of transuranic elements, some radioactivity was always found in the alkaline-earth group (barium carrier), the rare-earth group (lanthanum carrier), and also with elements in group IV of the periodic table (zirconium carrier). The barium precipitate was the first to be investigated more thoroughly, since it apparently contains the parent isotopes of the observed isomeric series. The goal was to show that the transuranic elements, and also U, Pa, Th, and Ac
could always be separated easily and completely from the activity which precipitates with barium.
1. For this reason, the irradiated uranium was treated with hydrogen sulfide, and the transuranic group was separated with platinum sulfide and dissolved in aqua regia. Barium chloride was precipitated from this solution with hydrochloric acid. From the remaining filtrate, the platinum was precipitated again with hydrogen sulfide. The barium chloride was inactive, but the platinum sulfide still had an activity of about 500 particles/minute. Similar experiments with the longer-lived transuranic elements gave the same result.
2. A precipitate with barium chloride was made using 10 g of unirradiated uranium nitrate. The U was in radioactive equilibrium with UX1 + UX2 (thorium and protactinium isotopes) and had an activity of about 400 000 particles/minute. The precipitate showed an activity of 14 particles/min; that is, practically no activity. That means neither U, nor Pa, nor Th, comes down out of solution when the barium chloride crystallizes.
3. Finally, using a solution of actinium (MsTh2) having an activity of about 2500 particles/min, a barium chloride precipitate was separated. This gave only about 3 particles/min which is also practically inactive.
In a similar way, the barium chloride precipitates obtained from the irradiated uranium solution were carefully investigated. However, sulfide precipitates made from the radioactive barium solution were practically inactive. Also, lanthanum and zirconium precipitates had only slight activities whose origin could easily be traced to the activity of the barium precipitates.
A simple precipitate with BaCl2 from a strong hydrochloric acid solution naturally does not allow one to distinguish between barium and radium. According to the reactions very briefly summarized above, the radioactivity which precipitates with the barium salts can only come from radium, if one eliminates barium itself for the time being as altogether too unlikely.
We now discuss briefly the graphs of the activities obtained with the barium chloride. They enable us to determine the number of "radium isotopes" present, and also their half-lives.
Figure I shows the activity of the radioactive barium chloride after a 4-day irradiation of uranium. Curve a gives the measurements for the first 70 h; curve b gives the measurements on the same sample continued for 800 h. The lower curve is plotted on 1/10 the scale of the upper one. At first there is a rapid decrease of activity, which becomes a gradual increase after about 12 h. After about 120 h, a very gradual exponential decrease of activity begins again, with a half-life of about 13 days. The shape of the curves shows clearly that there must be several radioactive species present. However one cannot tell for sure what they are. They might be several "radium isotopes," or one "radium isotope" with a series of radioactive daughter products.
The three "radium isotopes" which were previously reported in the earlier article were confirmed here. They are designated for the time being as Ra II, Ra III, and Ra IV (because of a presumed Ra I reported below). Their identification and the determination of their half-lives is explained briefly with the help of the figures. Figure 2 shows the radioactive decay of the "radium" after a 6-min irradiation of uranium. Curve a is the total activity, measured for 215 min. This curve is a composite of the activity from two "radium isotopes," Ra II and Ra III (compare Fig. 3), and also a small amount of actinium, which is formed by decay of Ra II. This latter substance, which is designated as Ac II, has a half-life of about 2 1/2 h. This was shown in another experiment, which is not described here. The theoretical growth curve for such an actinium isotope resulting from Ra II is shown in the figure as curve b. Here the half-life of Ra II is taken to be 14 min, in anticipation of later results. When curve b is subtracted from curve a, then curve c in Fig. 2 is obtained. This remaining activity must now come from the radium isotopes, mostly being due to the short-lived Ra II, and with a slight contribution from Ra III with its longer half-life. The latter has a half-life of about 86 min, as is seen in Fig. 3 later on. Curve d in
Fig. 2 shows the activity due to Ra III. When d is subtracted from c, one finally obtains curve e, which is the activity due to pure Ra II. It has an exponential decrease with a half-life of 14 min. This value should be correct within ±2 min.
Now we come to the identification and halflife determination of Ra III. A uranium sample is irradiated for one hour or several hours. One finds a rapid decrease in activity at first, then a still rather intense activity which decreases to one-half in about 100-110 min, and then a further decrease. In order to show that this activity was also mostly due to a radium isotope, the following procedure was used. The "radium" was separated from the irradiated uranium sample with barium chloride; after 2 1/2 h, the barium chloride was dissolved again, and reprecipitated. The short-lived Ra II has completely decayed during this time, and the Ac II (2 1/2 h half-life) which was formed from Ra II in the barium chloride is removed in the recrystallization process. The barium chloride still has considerable activity, so a "radium isotope" must still be present. The procedure here is like that used by Meitner, Strassmann, and Hahn3 for the investigation of the artificial radioactive daughter products of thorium. The resulting activity which remains is shown in Fig. 3, curve a.
During the first hour, the rate of decrease is almost exactly exponential, with a half-life of about 86 min. A small residual activity remains, which is no doubt due to a long-lived "actinium isotope" formed by the decay of Ra III. The decay of the actinium activity can be roughly determined by the departure of curve a from a pure exponential. This is shown in Fig. 3 as curve b. (It was also shown chemically that the decay of Ra III leads to an "actinium isotope" with a relatively long life.) If one subtracts b from a, one obtains curve c for Ra III alone. It shows a very nice exponential decrease with a half-life of 86 min. This value should be correct within ±6 min.
Now we come to the third "radium isotope," which is designated here as Ra IV. In Fig. 1, curve b, a substance with a half-life of about 12 to 13 days was indicated. In a manner quite similar to that used for Ra III, it was shown that this more slowly decreasing activity must be practically all due to a "radium isotope." A lengthy irradiation of uranium was made, then the neutron source was removed, and by waiting about one day the isotopes Ra II and Ra III were allowed to decay completely. If one makes a barium precipitate now, and carefully recrystallizes again, then any activity found with the barium chloride can only be due to another "radium isotope." Such an activity was always found, even after several days of waiting. The decay rate follows a characteristic pattern. It increases gradually for several days, reaches a maximum, and then decreases with a half-life of about 300 h (12.5 days).
In Fig. 4 are shown several such curves. The sample for curve c was prepared from a uranium solution which had been irradiated with low intensity, and the other curves are due to barium precipitates from more intensely irradiated uranium solutions. (The curves cannot be used to determine the relative intensity factor directly, since the geometrical arrangement was not identical. Under identical conditions, such as equal amounts of uranium being irradiated, etc., we found that the relative intensity factor was about 7.) The shapes of the three curves are very similar. The growth of activity has a half-life of less than 40 h, and the decay about 300 h. However, the long-lived "Ra IV" doubtlessly has a half-life of somewhat less than 300 h, because the Ac IV which is mostly responsible for the initial growth of activity probably decays to a long-lived "thorium isotope." Therefore the half-life of Ra IV cannot be determined precisely, but a value of 250-300 h is probably close to being correct. From curves a, b, and c, one can see clearly that the beta rays of Ra IV are much less penetrating than those from its daughter product, since otherwise such a sharp increase would not occur.
To summarize our results, we have identified three alkaline earth metals which are designated as Ra II, Ra III, and Ra IV. Their half-lives are 14±2 min, 86±6 min, and 250-300 h. It should be noted that the 14-min activity was not designated as Ra I nor the other isomers as Ra II and Ra III. The reason is that we believe there is an even more unstable "Ra" isotope, although it has not been possible to observe it so far. In our first article about these new radioactive decay products we reported an actinium isotope with a half-life of about 40 min. Our initial assumption was that this least stable actinium had resulted from the decay of the least stable radium isotope. In the meantime, we have determined that the 14-min radium (previously given as 25 min) decays to actinium with a 2.5-h half-life (previously given as 4 h). However the less stable actinium isotope mentioned above is also present. Its half-life is somewhat shorter than previously reported, perhaps a little less than 30 min. This "actinium isotope" cannot result from the decay of the 14-min, 86-min, or the long-lived "Ra." Also this "actinium isotope" can be shown to be present after only a 5-min irradiation of uranium. The simplest explanation is to assume the formation of a "radium isotope" whose half-life must be shorter than 1 min. If it had a half-life longer than 1 min, we should have been able to detect it. We searched for it very carefully. Therefore we designate this heretofore unknown parent of the least stable "actinium isotope" as "Ra I." With a more intense neutron source it should no doubt be detectable.
The decay scheme which was given in our previous article must now be corrected. The following scheme takes into account the needed changes, and also gives the more accurately determined half-lives for the parent of each series:
The large group of transuranic elements so far bears no known relation to these isomeric series.
The four decay series listed above can be regarded as doubtlessly correct in their genetic relationship. We have already been able to verify some of the "thorium" end products of the isomeric series. However, since the half-lives have not been determined with any accuracy yet, we have decided to refrain altogether from reporting them at the present time.
Now we still have to discuss some newer experiments, which we publish rather hesitantly due to their peculiar results. We wanted to identify beyond any doubt the chemical properties of the parent members of the radioactive series which were separated with the barium and which have been designated as "radium isotopes." We have carried out fractional crystallizations and fractional precipitations, a method which is well-known for concentrating (or diluting) radium in barium salt solutions.
Barium bromide increases the radium concentration greatly in a fractional crystallization process and barium chromate even more so when the crystals are allowed to form slowly. Barium chloride increases the concentration less than the bromide, and barium carbonate decreases it slightly. When we made appropriate tests with radioactive barium samples which were free of any later decay products, the results were always negative. The activity was distributed evenly among all the barium fractions, at least to the extent that we could determine it within an appreciable experimental error. Next a pair of fractionation experiments were done, using the radium isotope ThX and also the radium isotope MsTh1. These results were exactly as expected from all previous experience with radium. Next the "indicator (i.e., tracer) method" was applied to a mixture of purified long-lived "Ra IV" and pure MsTh1; this mixture with barium bromide as a carrier was subjected to fractional crystallization. The concentration of MsTh1 was increased, and the concentration of "Ra IV" was not, but rather its activity remained the same for fractions having an equivalent barium content. We come to the conclusion that our "radium isotopes" have the properties of barium. As chemists we should actually state that the new products are not radium, but rather barium itself;
other elements besides radium or barium are out of the question.
Finally we have made a tracer experiment with our pure separated "Ac II" (half-life about 2.5 h) and the pure actinium isotope MsTh2. If our "Ra isotopes" are not radium, then the "Ac isotopes" are not actinium either, but rather should be lanthanum. Using the technique of Curie,4 we carried out a fractionation of lanthanum oxalate, which contained both of the active substances, in a nitric acid solution. just as Mine. Curie reported, the MsTh2 became greatly concentrated in the end fractions. Withour "Ac II" there was no observable increase in concentration at the end. We agree, with the findings of Curie and Savitch5 for their 3.5-h activity (which was however not just a single species) that the product resulting from the beta decay of our radioactive alkaline earth metal is not actinium. We want to make a more careful experimental test of the statement made by Curie and Savitch that they increased the concentration in lanthanum (which would argue against an identity with lanthanum) since in the mixture with which they were working there may have been a false indication of enrichment.
It has not been shown yet if the end product of the "Ac-La sample," which was designated as "thorium" in our isomeric series, will turn out to be cerium.
The "transuranic group" of elements are chemically related but not identical to their lower homologs, rhenium, osmium, iridium, and platinum. Experiments have not been made yet to see if they might be chemically identical with the even lower homologs, technetium, ruthenium, rhodium, and palladium. After all one could not even consider this as a possibility earlier. The sum of the mass numbers of barium+technetium, 138+101, gives 239!
As chemists we really ought to revise the decay scheme given above and insert the symbols Ba, La, Ce, in place of Ra, Ac, Th. However as "nuclear chemists," working very close to the field of physics, we cannot bring ourselves yet to take such a drastic step which goes against all previous experience in nuclear physics. There could perhaps be a series of unusual coincidences which has given us false indications.
It is intended to carry out further tracer experiments with the new radioactive decay products. In particular a combined fractionation will be attempted, using the radium isotope resulting from fast neutron irradiation of thorium (investigated by Meitner, Strassmann, and Hahn6) together with our alkaline earth metals resulting from uranium. At places where strong neutron sources are available, this project could actually be carried out much more easily.
In conclusion we would like to thank Miss Cl. Lieber and Miss I. Bohne for their efficient help in the numerous precipitations and measurements.
1 From the Kaiser Wilhelm Institute for Chemistry, at Berlin-Dahlem. Received 22 December 1938.
2 O. Hahn and F. Strassmann, Naturwissenschaften 26, 756 (1938).
3 L. Meitner, F. Strassmann, and O. Hahn, Z. Physik 109, 538 (1938).
4 Mme. Pierre Curie, J. Chim. Phys, 27, 1 (1930).
5 I. Curie and P. Savitch, Compt. Rend. 206, 1643 (1938).
6 L. Meitner, F. Strassmann, and O. Hahn, loc. cit.