Full Text
Conversion of Hydrogen into Helium1
Fritz Paneth and Kurt Peters.
1. The Basic Idea of the Work
In the present development of Prout’s hypothesis, in astrophysical calculations of the lifetimes of fixed stars, and in explaining the origin of Hess radiation from the standpoint of radioactivity, it is theoretically necessary to allow for the possibility of the conversion of hydrogen into helium. However, it has not yet been possible to carry out this transformation of elements, although the most varied kinds of electrical discharges have been used for this purpose, with the expenditure of large amounts of energy.
This reaction itself must release enormous quantities of energy; from the decrease in mass of four gram-atoms of hydrogen in passing into helium one can calculate the thermal effect of the reaction, equal to \(6.4 \cdot 10^{11}\) calories. Therefore, in general, it cannot be regarded as established that carrying out this reaction requires the supply of energy from outside. Another possibility for carrying out this reaction on an appreciable scale may consist in catalytically accelerating a reaction that proceeds by itself immeasurably slowly. The basic idea of this work was to try whether hydrogen is partially converted into helium without the supply of energy from outside if it is brought into contact with a suitable catalyst; in this connection, as the catalyst we had in advance settled on palladium.
We had, of course, no data for a preliminary estimate of the quantities of helium that could be formed from hydrogen during the practically possible duration of an experiment under the most favorable conditions. However, the hopes of obtaining the desired effect had to be greater the more accurately we could carry out the analytical determination of helium. Attempts to increase the amount of helium formed had, on the other hand, to be supplemented
by trying to reduce the amount of helium which we could still determine; in this way we hoped, by acting from above and from below, to bring the experiment down to quantities of one and the same order. Of course, at the beginning of the work we could not judge whether this goal was attainable. Now, after we have worked on this problem for several years, we think that we have succeeded in solving it, and below we shall briefly set forth our results. A more detailed article, with drawings of the apparatus and with references to the literature, will appear elsewhere.
2. Determination of helium.
By improving the ordinary method hitherto used for the spectroscopic investigation of helium, we have lowered the limit of possible determination of it to \(10^{-8}—10^{-9}\ \mathrm{cm}^3\) (which corresponds to \(10^{-12}—10^{-13}\ \mathrm{g}\)). First of all, by the usual method we remove the more readily condensable gases, absorbing them with charcoal cooled by liquid air; then, with an excess of oxygen, we oxidize hydrogen to water. We carry out this combination not, as is usually done, by an electric discharge, but burn the hydrogen, using platinum or palladium as a catalyst. The excess oxygen is then absorbed by cooled charcoal, and the small volume of gas remaining is transferred into a glass capillary with an orifice diameter of \(0.1\ \mathrm{mm}\), wrapped on the outside with wire electrodes. Owing to the fineness of the capillary it can replace the slit of the spectroscope.
In addition to the gases neon and helium, which remain uncondensed after such treatment, at the beginning of the passage of discharges through the capillary there are usually also observed insignificant traces of hydrogen, and sometimes of other gases. A circumstance especially favorable for the determination of small quantities of helium is that, during prolonged electrical excitation of the electrodeless glass tube, all gases disappear earlier than the light noble gases—helium and neon.
By our method, the removal of neon is not achieved, and in general it is difficult to accomplish. Even if it could be carried out by some other method1, in our formulation of the problem this should be avoided. Indeed, the presence of neon is an extremely valuable indication that atmospheric air has entered. The danger of its entering through a not entirely faultless stopcock or a bad spot in the glass is very great; when small quantities of helium appear, we can be fully certain that it has not entered together with air only if the absorption is arranged so that, in that case, neon as well as helium would also have found its way into the spectral tube.
Thus, we consider the presence of helium of non-atmospheric origin to be proven only in the case when it is either completely free of neon, or when in the spectrum the neon lines appear much more weakly than would correspond to the neon content in the neon–helium mixture from air.
We could not in all our experiments obtain the complete absence of neon lines, since the spectroscopic sensitivity of helium with respect to contamination by neon is extraordinarily great; a few percent of neon added to helium already cause the appearance of most of the neon lines. Nevertheless, if the percentage content of neon is very small, then from the character of the spectrum one may conclude with certainty that the greater part of the gas consists of helium.
In order to test the sensitivity of our method for determining helium, we attempted to prove the formation of helium from active thorium, which undoubtedly occurs as a result of its $\alpha$-radiation, but which previous methods had not succeeded in proving experimentally. We succeeded in this without any difficulty. From a preparation of thorium $B$—thorium $C$, equivalent to 3 mg of radium1, we obtained a brilliant spectrum of helium. As is easily calculated, the amount of helium determined by us in this case was $10^{-7}\ \mathrm{cm}^3$.
Since in our capillary we could determine without error—qualitatively—an amount of helium at least 10 times smaller, we arrive at the limit of sensitivity of our method—at least $10^{-8}\ \mathrm{cm}^3$. We think that the weakest, yet clear appearance of the green helium line $\lambda = 5016\ \text{\AA}$, by which we determined the very slightest traces of helium, can be observed at amounts another ten times smaller; this limit of sensitivity of our method has not yet been precisely established, since in decisive experiments we did not wish to carry the determination of helium to the extreme limits attainable.
We shall also point out that the development of a method for determining helium that is as sensitive as possible and at the same time simple may prove useful in the investigation of various other questions connected with helium. Thus, this method enabled Günther to determine the helium content in iron meteorites, in which the radium content reached $5 \cdot 10^{-14}\ \mathrm{g}$ per gram of substance. This gave us the possibility of making, so far as we know, the first determination of the age of a meteorite. For the Mount Joy meteorite we obtained a minimum age of 600 million years2.
Another area in which our method, with slight modifications, proved useful is the investigation of natural gases for helium content. Whereas until now this required several liters or, at the least, several hundred \(cm^3\) of gases, Gehlen carried out such analyses with several \(cm^3\). This is a considerable simplification, especially in taking samples. In a number of investigations of gas sources for helium content the following interesting result was obtained: in one German gas source a helium content was found 10 times greater than in the richest source of natural gas hitherto known in Germany. This new source, with a helium content of 0.19%, approaches the Canadian sources with 0.33%, from which helium is extracted for technical purposes. At the present time the question is being investigated whether the technical extraction of helium from the German source is profitable.
3. Verification of earlier data on the artificial formation of helium.
Although the greater part of the assertions encountered in the literature concerning the formation of helium during electrical discharges has already been refuted by later investigations, more careful in the experimental sense, it nevertheless seemed necessary to us to test some of these indications by our extraordinarily sensitive method.
Thus, Strutt left open the question whether, upon bombardment of certain salts by cathode rays, quantities of helium are liberated sufficient for them to be detected [as J. J. Thomson asserted], if not spectroscopically, then at least by means of positive rays.
However, from Thomson’s data it can be calculated that the quantities of helium that can be detected in his apparatus are also accessible to determination in our apparatus. This test was undertaken by Gehlen; after prolonged bombardment of salts he was unable to detect spectroscopically even the slightest traces of helium.
Further, for many hours we subjected hydrogen in an ozonizer to the action of dark electrical discharges. We chose this method of supplying electrical energy because here it is possible to work without electrodes and with the careful exclusion of all traces of air, and also to vary the gas pressure, the voltage, and the current strength within fairly wide limits. In the ozonizer, too, we could not establish the appearance of helium, even in the most negligible quantities.
Finally, in connection with the indications of some authors concerning the formation of helium under these conditions, we passed for several...
hours, strong electric discharges through hydrogen in a Geissler tube with aluminum electrodes. The result of the investigation for helium here was also completely negative. Therefore, on the question of the formation of helium from hydrogen under an electric discharge, we adhere to the opinion of those authors who deny its appearance in measurable quantities.
4. The appearance of helium under the action of palladium on hydrogen.
We first tried to realize the idea, mentioned at the beginning, of the transformation of hydrogen into helium by the action of palladium in the following form. Hydrogen in the largest possible quantities—up to a liter and more—was passed through hot palladium. We expected that at the moment of emergence, perhaps, some of the chaotically arranged protons and electrons, present in a relatively high concentration, would combine not into hydrogen atoms but into helium nuclei.
The greater part of the hydrogen that entered our apparatus through a palladium capillary heated to red heat was bound by heated calcium or else released into the vacuum through another palladium capillary; the remainder was removed, as indicated above, by combustion in an excess of oxygen, and the remaining gases were examined spectroscopically. The apparatus was nevertheless very complicated, and it was very difficult to exclude air so completely that it could not be detected by the presence of neon in the spectral tube; the amount of neon contained in \(1\ \mathrm{cm^3}\) of air, i.e. \(1.8 \cdot 10^{-8}\ \mathrm{cm^3}\) of neon, already impairs the purity of the experiment. In individual experiments, however, it was possible to exclude the penetration of air completely. In these cases it was sometimes possible to see several weak helium lines; they were insufficient, however, to convince us of the suitability of our palladium capillary for the formation of helium, primarily because no proportionality was observed between the brightness of the helium lines and the amount of hydrogen passed through the capillary, i.e. the duration of the experiment. Therefore we came to the conclusion that, when hydrogen is passed through incandescent palladium, helium is not formed in such quantities as we could establish with certainty by our means.
Nevertheless, the experiments indicated were the only ones that gave us at least hints of the formation of helium. Since we had the impression that the hydrogen remaining in the apparatus after the experiment gives, in a cold capillary, the same barely perceptible effect as much larger quantities of hydrogen passing through a hot capillary, we arrived at the assumption that the formation of helium proceeds on the surface of palladium at room temperature. If this suppo-
...position is correct, the effect should increase with an increase in the surface of the palladium.
Therefore, we carried out the following experiments with various preparations of palladium black, palladium sponge, and palladium asbestos. Instead of passing hydrogen through palladium, we allowed it to be absorbed by palladium and then, after various intervals of time, burned it with oxygen on the same palladium preparation. This considerably simplified the apparatus, and it was much easier to avoid the penetration of air than in an apparatus with a palladium capillary. Along with the purity of our helium spectra, the strength of the effect also increased. We succeeded in preparing specimens which, after only 12 hours, when heated simultaneously with hydrogen, gave off an amount of helium sufficient to obtain 4–5 lines of the helium spectrum without a single neon line. With other, less successful preparations, we could establish clear helium spectra only after several days or weeks. In these experiments we had the impression from the very beginning that the previously unobserved proportionality between the effect and the duration of the experiment is realized here, at least in the roughest outline. We have not yet been able to establish this proportionality precisely, since up to now the quantities of helium, not reaching \(10^{-7}\,\mathrm{cm}^3\), could only be estimated by order of magnitude.
In addition, we observed that not only do two palladium preparations apparently prepared in exactly the same way often differ in their action, but one and the same preparation also tends to diminish its activity with the passage of time.
It is known that the positive capacity of palladium with respect to hydrogen is subject to inexplicable fluctuations, and “active” palladium by itself gradually becomes “inactive.” We tried the methods recommended for restoring the activity of palladium: heating in hydrogen or oxygen, or in a mixture of both, or in vacuum, and in fact in many cases achieved success; palladium which no longer gave off helium in noticeable quantities became, after such treatment, again “active” in this respect. Our experiments showed that a parallelism between the activity of palladium in the sense of binding hydrogen and its activity with respect to the formation of helium exists only insofar as no preparation that did not bind hydrogen gave helium either; on the other hand, preparations that absorbed hydrogen well sometimes gave very little helium or none at all, precisely in those cases when the saturation of palladium with hydrogen was carried out with heating.
Since helium can be completely removed from palladium preparations with a large surface only at high temperature,
we assumed that helium could be detected in all such preparations that had lain for a prolonged time unused at room temperature—since under these conditions they contain a certain amount of bound hydrogen. The various old palladium preparations investigated by us in this direction all, without exception, confirmed this view; since, without heating, it is also impossible to remove completely the adsorbed and occluded air, we of course did not obtain pure helium, but a mixture of helium with neon, in which, however, enrichment in helium could clearly be established. The amounts of helium, in accordance with the duration of the experiment, were relatively large; thus, 1 g of 50% palladium asbestos, which we had obtained 2 years earlier from the firm Kahlbaum and which had been prepared, possibly, still earlier, gave \(10^{-6}\ \mathrm{cm}^3\) of practically pure helium.
Since this amount, in comparison with those obtained from other, older preparations, was abnormally large, we concluded that this palladium asbestos was still, even now, unusually active with respect to helium. We investigated the formation of helium on it and on preparations made by us. It proved, indeed, to be much better than all our preparations, and also than other samples of palladium asbestos from the same firm. With it it proved possible to carry out a series of experiments to which we attach especially great importance. For several days we kept it in our air-tight apparatus and alternately saturated it with hydrogen and oxygen. Before each change of the saturating gas we investigated the amount of helium contained in it. After standing for 12 hours in oxygen, only barely perceptible traces of helium appeared, the presence of which is readily explained by their formation from residues of hydrogen. After the palladium asbestos had immediately thereafter been in an atmosphere of hydrogen for 5 hours, analysis gave a 10–100 times greater amount of helium. We repeated this experiment 3 times in succession with one and the same result. When we began the investigation of this palladium asbestos, helium was formed on it from hydrogen in the amount of \(10^{-8}\)—\(10^{-7}\ \mathrm{cm}^3\) per day; after 20-fold heating this asbestos became inactive. By alternate treatment in oxygen and hydrogen at various temperatures we restored to it the ability to form helium, but no longer to the extent as at first: it gave per day only \(10^{-9}\)—\(10^{-8}\ \mathrm{cm}^3\) of helium—an amount that was also given by some of the preparations made by us.
We investigated the formation of helium not only on palladium preparations, but also on platinum asbestos, platinum black, platinum sponge, and on pyrophoric nickel powder. Some individual platinum preparations gave clearly positive results, although the effect was always smaller than with the best
palladium preparations. The activity of nickel seems to be still less, but nevertheless, on the basis of the results obtained so far, we are inclined to the view that it too lies within the limits of measurability.
5. Discussion of sources of error.
We have, of course, repeatedly asked ourselves whether the presence of helium in our apparatus could not be explained by the fact that it was already there before the experiment.
The simplest assumption is that the helium comes from the air. As was mentioned above, a simple defect in the apparatus admitting air is not sufficient as an explanation, since in that case helium and neon would have had to be present in the same quantitative ratio as that in which they occur in air. Helium from the air can be invoked as an explanation only if one can indicate a mechanism that could produce a fractionation of the air enriching it in helium.
Such a mechanism can, under certain circumstances, be realized. For the duration of our experiments and at room temperature, glass is sufficiently impermeable to helium, but at high temperature it lets appreciable quantities of this gas pass through1.
Neon passes through glass so much more slowly than helium that from atmospheric air one can obtain helium almost free of neon by a single passage through hot glass2. Therefore we took care that the tube with the palladium preparation—the only part of our apparatus which, for a short time, is subjected to heating (for the burning of hydrogen and the removal of helium)—was surrounded by a vacuum jacket immersed in water. As numerous control experiments showed, in this way it is possible completely to prevent the penetration of helium from the air.
The second possibility for the fractionation of helium from neon is based on the ability of glass selectively to adsorb helium from a mixture of helium with neon and then dissolve it. We carried out experiments with glass tubes which, for several days, were filled with a mixture of neon and helium in the same ratio
as in the atmosphere, at a total pressure of \(1/3\) of an atmosphere. When tubes prepared in this way are heated, almost pure helium is liberated. However, the amounts of helium that can be extracted by glass from atmospheric air are so negligible that the glass tubes in which the palladium preparations were kept could not have appreciably changed the amounts of helium found in our experiments. This was also confirmed by special control experiments.
The fact that palladium preparations act best soon after preparation may give rise to the suspicion that, during their preparation, these preparations become saturated with helium from the air, and give up this helium only after repeated heating and thus become exhausted. However, it can be shown that palladium, in contrast to glass, does not have the ability selectively to adsorb helium; likewise, it neither dissolves helium nor allows it to pass by diffusion. By special experiments we convinced ourselves that the amounts of helium adsorbed by palladium simultaneously with neon from a neon—helium mixture under a pressure of \(1/3\) of an atmosphere are given off by palladium upon a single heating so completely that their secondary liberation cannot be demonstrated by our method. In this respect it is immaterial whether the preparations are made by thermal decomposition of palladosammine chloride in the mixture mentioned (in which case metallic palladium in statu nascendi comes into contact with helium), or whether palladium asbestos, obtained by precipitating palladium from solution, is then placed in an atmosphere of helium and neon. The partial pressure of helium in air constitutes only one 20,000th part of the partial pressure of helium in our experiments; therefore palladium preparations that had been in contact not with helium in high concentration, but only with air, could not have adsorbed helium in any significant amount and then given it up with great difficulty, as would have to be assumed in order to explain our results from this point of view. Asbestos also does not have the ability to extract helium from a neon—helium mixture, as our experiments showed; crystalline asbestos behaves in this respect quite differently from amorphous substances—glass and fused quartz.
During the experiments only hydrogen and oxygen were introduced into our apparatus. Both were taken from a specially constructed electrolyzer, which before use was freed from the last traces of air by evacuation and many hours of operation. Each of these gases, according to our measurements, contained less than 0.001 of air. But even if it were assumed that, owing to a suddenly occurring leak, contamination of the gases by air took place, this could not explain the appearance of helium free from neon, since here again—
Thus there is no mechanism that could produce fractionation; one might also suppose that palladium selectively absorbs neon and in this way increases the helium content in the remaining gas. We showed that this supposition is incorrect as follows: palladium preparations that had given off helium upon heating were completely dissolved by us in our closed apparatus, and we analyzed the gases thereby evolved. In these gases, too, the ratio of the quantity of helium to neon was greater than in atmospheric air.
Dissolving old pieces of palladium sheet and examining the gases obtained thereby, we became convinced that palladium itself does not give helium, which might occur as the result of a special kind of radioactive decay. Since palladium is impermeable to helium, the latter would have had to accumulate inside the palladium. Even in palladium ten years old we could not establish the presence of helium, except for traces lying at the limit of detectability, which may be attributed to hydrogen bound on the surface.
Of the old samples of platinum sheet investigated by us, the greater part likewise contained no helium; in one sample, however, we found appreciable quantities of it. What caused this difference from the other samples we cannot say until new experiments have been carried out; it is possible that this sheet, which had long been in use, contained cavities, so that over the years helium could have formed in them but had no possibility of escaping.
6. Discussion of the Results.
Thus, we found that all sources of error that had to be taken into account are insufficient to explain the observed effects. This led us to the conclusion that the helium appearing in our apparatus is formed by the action of palladium on hydrogen. The fact that helium does not appear when palladium is saturated with oxygen and is formed again when it is saturated with hydrogen admits, in our opinion, no other interpretation. Even if one assumes that palladium contains helium (which contradicts what was shown in the preceding paragraph), one would still have to add another hypothesis—that this helium can be liberated from the palladium only by hydrogen. That hydrogen does not possess the ability to make palladium permeable to helium we proved by a special experiment: a palladium capillary, ignited in a mixture of pure helium (from monazite sand) and hydrogen, passed just as little helium as in the case when the ignition was carried out in helium devoid of hydrogen.
From our experiments one cannot draw the direct conclusion that helium is formed through the interaction of hydrogen and palladium because catalysis promotes the tendency, inherent in hydrogen, to transform into helium (this was our basic premise). But if one is at all prepared to admit the formation of helium, this remains the most probable explanation. Palladium may be assigned the role of a catalyst; in favor of this is the fact that platinum acts in the same way qualitatively, though more weakly quantitatively.
The continuation of the work must chiefly clarify on what the different activity of different palladium preparations, and of one and the same preparation at different times, is based. It seems to us that it is still too early to seek a connection between this activity and allotropic modifications of palladium, whose existence is assumed on the basis of other data. The suggestion that only the adsorption of hydrogen produces the effect, while its absorption has no effect, we can express only with great caution, although there are some facts speaking in favor of this hypothesis. Only with an increase in the amount of experimental material will it be possible to say how well founded are the hopes for the possibility of increasing the rate of helium formation.
A further important question, which we had to leave unresolved, concerns the determination of the large quantities of energy which, in all probability, are liberated in the formation of helium. It has repeatedly been pointed out that any formation of helium must immediately manifest itself in an enormous evolution of heat. However, the expected quantities of energy, though incredibly large relatively—in comparison with those liberated in chemical reactions—are, in absolute magnitude, even for the quantities of helium formed in our experiments, still so small that they cannot be detected by our method. In the formation of \(10^{-8}\ \mathrm{cm}^3\) of helium one may expect a heat effect of \(0.28\ \mathrm{cal}\); but when palladium is saturated with hydrogen, much more considerable thermal effects are observed, which may be attributed partly to the heat of absorption, partly to the heat of formation of a chemical compound, and partly also to the heat of combustion of residual oxygen. Alongside these thermal effects, the heat effect of helium formation in such small quantities is very difficult to establish.
Moreover, there exists the possibility—and even, on theoretical grounds, the probability—of the initial transformation of the energy released in helium formation not into heat, but into radiation. We repeatedly attempted to detect \(\gamma\)-radiation near palladium preparations saturated with hydrogen, or the emission of electrons when hydrogen was passed over palladium, but in neither case (despite indications existing in the literature) did we obtain any-
any effect on the electroscope. These experiments will be continued with improved means, but their success is doubtful: even in the event that the union of hydrogen atoms into helium is accompanied by radiation, we shall in all probability be unable to detect it. It must be so penetrating that we shall not be able to catch it with our instruments until it has undergone repeated scattering by electrons (the Compton effect).
-
By special experiments we established that even at room temperature glass cannot be regarded as completely impermeable to helium, if one uses so sensitive a method as ours; nevertheless, in our apparatus helium penetrating through the glass could barely be detected after a week. In experiments lasting several days or weeks we immersed the corresponding glass parts in water. With this precaution, the amount of helium that penetrated into the apparatus over several months remained below the limit of sensitivity of our method. ↩↩↩↩