Hydrogen Isotope with Mass 3

V. I. Chernyaev, Leningrad

In 1934 experimental data were published on the basis of which the conclusion was drawn that a hydrogen isotope with mass 3 exists in natural sources.[^1]

This assertion was based chiefly on the experiments of two groups of authors. First, Tuve, Hafstad, and Dahl[^2] supposedly found in a beam of ions obtained from 98% heavy water a certain number of particles with mass 3, which the authors identified with the nucleus of the isotope \({}^{3}_{1}\mathrm{H}\); and second, Lozier, Smith, and Bleakney,[^3] with the aid of a mass spectrograph, were supposedly able to detect the presence of this isotope in a portion of nearly pure deuterium (the heavy isotope of hydrogen with mass 2). At the present time, however, the results of these experiments must be called into question. The experiments of Bleakney and collaborators, who attempted to detect the isotope \({}^{3}_{1}\mathrm{H}\) with the aid of a mass spectrograph, at first[^4] led to negative results, then supposedly gave a positive result,[^3][^5] and, finally, a careful repetition of the experiments[^6] again led to the conclusion that the isotope \({}^{3}_{1}\mathrm{H}\), in any case in appreciable quantities, is not contained in natural hydrogen.

Bleakney’s method, however, is of independent interest and, if used rationally, can give very valuable results. Thus, for example, using this method,[^7] the author was able—though only after the hydrogen isotope with mass 2 (deuterium) had been discovered—to detect its presence in natural hydrogen, where it occurs in very small quantities; to estimate the changes in the relative amounts of the isotopes \(\mathrm{H}\) and \(\mathrm{D}\) after partial separation of the isotopes; and finally, with known precautions, to determine with great accuracy the ratio of the amounts of the two isotopes in natural hydrogen.[^8]

It is therefore of interest to become acquainted with this method in greater detail; this will also help to clarify the cause of the contradictory results obtained with it.

As is known, the basic idea of the mass spectrograph is that a beam of positive ions passes through a space subjected to the action of electric and magnetic fields. Under the action of these fields the ions undergo deflections from their initial direction of motion, and it is clear that these deflections will depend both on the charge of the given ion and on its mass. The ions

with identical charge-to-mass ratios \(\frac{e}{m}\) undergo identical deflections. Therefore they can be, for example, “focused” into a narrow beam and, as was done in Dempster’s experiments, led from one part of the mass spectrograph to another if a narrow slit is placed at the “focus” of these ions in the partition separating the two parts of the mass spectrograph from one another.^9 Behind the slit there is an ion trap connected to an electrometer, by which the ion current is measured, and consequently also the intensity of the ions (i.e., their number entering the trap per second). If different ions are present in the ion beam entering the mass spectrograph, then, by varying the electric field while keeping the magnetic field unchanged, one can bring to the exit slit beams of ions with different values of the ratio \(\frac{e}{m}\). In the case of ions of the same charge but different masses, the relative values of the ion masses are obtained directly from the values of the applied potential difference. By changing the field in this way, one observes for which of its values the electrometer gives the maximum deflection, and calculates the ratios of the ion masses. The magnitudes of the maximum deflections of the electrometer (peaks) make it possible to judge the relative intensities of the different groups of ions.

Bleakney developed an improved type of mass spectrograph, allowing operation at extremely low gas pressures, and applied it to determining the presence of deuterium in ordinary hydrogen.^7

From work on mass spectrography it was known that if ions are produced from gaseous hydrogen (for example, by electron bombardment of hydrogen), then molecular ions \(\mathrm{H}_2^+\) (a hydrogen molecule without one electron) are formed chiefly, and in much smaller quantity—monatomic ions \(\mathrm{H}^+\) (protons) and triatomic ions \(\mathrm{H}_3^+\). All these ions possess a single positive charge. Since gaseous hydrogen is diatomic, it is evident that \(\mathrm{H}_2^+\) ions are formed by a simple electron impact. As long as collisions with other particles that can lead, for example, to the decomposition of \(\mathrm{H}_2^+\) ions occur rarely, i.e., at sufficiently low pressures, the number of \(\mathrm{H}_2^+\) ions increases proportionally to the pressure. Thus the intensity of \(\mathrm{H}_2^+\) ions at low pressures can be expressed as a linear function of the pressure \(p\):

\[ I\left(\mathrm{H}_2^+\right)=a_1p. \tag{1} \]

Under the same conditions (low pressures) the intensity of monatomic ions as a function of pressure is expressed by a function having both a linear and a quadratic term:

\[ I\left(\mathrm{H}^+\right)=a_2p+b_1p^2, \tag{2} \]

and the intensity of triatomic ions will be expressed by a quadratic function:

\[ I(\mathrm{H}_3^+) = b_2 p^2. \tag{3} \]

\(a_1, a_2, b_1\) and \(b_2\) are constants.

Indeed, \(\mathrm{H}^+\) ions may be formed, first, directly from \(\mathrm{H}_2\) molecules in a single electron impact, when the impacting electron causes, in one act, dissociation of the \(\mathrm{H}_2\) molecule and ionization of one of the atoms. The number of ions formed by such a mechanism is proportional to the pressure; this gives the linear term of formula (2). On the other hand, ions may also be formed in two stages. First the impacting electron, tearing an electron from the \(\mathrm{H}_2\) molecule, forms the molecular ion \(\mathrm{H}_2^+\), which then decomposes, under the influence of a second electron impact or upon collision with other particles, into the ion \(\mathrm{H}^+\) and an H atom. The probability of each of these two acts is proportional to the pressure; the probability of the composite process is equal to the product of the probabilities of the elementary processes whose sequence constitutes the composite one; therefore the probability of the latter (and hence also the intensity of the \(\mathrm{H}^+\) ions formed in such a process) is proportional to the square of the pressure. Hence the quadratic term of formula (2) is obtained. Owing to the insignificant number of \(\mathrm{H}^+\) ions in comparison with the number of \(\mathrm{H}_2^+\) ions in these processes, the number of \(\mathrm{H}_2^+\) ions practically does not change, and consequently the intensity of the \(\mathrm{H}_2^+\) ions remains proportional to the pressure.

Finally, because there are no stable \(\mathrm{H}_3\) molecules in the normal state, triatomic ions \(\mathrm{H}_3^+\) cannot be formed in a single electron impact, and the linear term in formula (3) must be absent. The \(\mathrm{H}_3^+\) ion may be formed if, for example, an \(\mathrm{H}_2^+\) ion is first formed, to which an H atom is then attached as a result of a collision. A purely quadratic dependence on pressure is obtained.

Let us now suppose that instead of pure hydrogen we have a mixture of hydrogen and deuterium (an isotope of hydrogen with mass 2), the amount of deuterium being very small in comparison with the amount of hydrogen, as is the case for hydrogen obtained from natural sources. In such a case the mixture must contain the molecules \(\mathrm{H}_2\), HD, and \(\mathrm{D}_2\). Obviously, \(\mathrm{H}_2\) molecules will be the overwhelming majority, HD a small amount, and \(\mathrm{D}_2\) a negligibly small amount. If this mixture is subjected to electron bombardment, we shall obtain various monatomic, diatomic, and triatomic ions, whose intensities are expressed (at low pressures) respectively by formulas of the type (1), (2), and (3). The possible ions thereby obtained are given in Table 1. In this table the second column gives the nature of the ion, the third—the ratio of the mass of the ion to its charge (if \(m/e\) for the proton is taken equal to unity), than

TABLE 1

determines the behavior of the ion in the mass spectrograph, the fourth—the character of the dependence of the intensity $I(p)$ of the given group of ions on the pressure, and the fifth (qualitatively) the relative intensity of the ions, if one takes into account both the ratio of the total amounts of H and D and the probabilities of formation of ions with different numbers of atoms.

In Table 1, $a_1, a_2, a_3,\ldots$ and $b_1, b_2, b_3,\ldots$ are different constants. Let us emphasize once again that the pressure $p$ is so small that the probability of secondary reactions (which are proportional to the square of the pressure) is considerably smaller than the probability of primary ones, which vary linearly with pressure.

From the column $\dfrac{m}{e}$ we see that in the mass spectrograph six groups of ions should be obtained, for which $\dfrac{m}{e}$ is equal to $1,2,\ldots,6$. However, configurations 6, 7, 8, 9 $\left(\dfrac{m}{e}=4,5,6\right)$ are too weak for them to be observed. It is also evident that the $\mathrm{D}^{+}$ ions, having a very small intensity, are completely masked by the most intense group of $\mathrm{H}_2^{+}$ ions with the same value of $\dfrac{m}{e}$. The ions $\mathrm{H}_3^{+}$ and $[\mathrm{HD}]^{+}$ have intensities comparable with one another; since both these groups have the same value of $\dfrac{m}{e}$, they are not separated in the mass spectrograph. Their total intensity as a function of pressure evidently has the form

\[ I\bigl[\mathrm{H}_3^{+}+(\mathrm{HD})^{+}\bigr]=ap+bp^2. \tag{4} \]

On the other hand, the intensity of ions of mass 2 (here, as indicated, practically only diatomic ions $\mathrm{H}_2^{+}$ occur) is proportional to the pressure. Therefore, instead of measuring the pressure, for example in millimeters of mercury, one may measure it by the intensity of the $\mathrm{H}_2^{+}$ ions, or directly

deflection of the electrometer corresponding to the peak of ions of mass 2 (i.e., when the beam of ions of mass 2 just enters the slit of the ion trap). Let the deflection corresponding to ions of mass 2 be equal to \(I_2\), and the deflection of the electrometer when ions of mass 3 are brought to the slit be equal to \(I_3\). In this case, passing from intensities and pressures to the corresponding deflections of the electrometer, formula (4) may be represented as

\[ I_3 = a I_2 + b I_2^2 . \tag{5} \]

Bleakney\(^7\) investigated two portions of hydrogen. One of them was obtained by electrolysis of ordinary water, while the second was a portion somewhat enriched with deuterium, prepared by Brickwedde by evaporating hydrogen at the triple point.\(^ {10}\) The results of investigating the dependence of ion intensity on pressure by means of a mass spectrograph are presented in Fig. 1. Along the abscissa axis are plotted, in arbitrary units, the pressure (or the intensity \(I_2\) of ions of mass 2), and along the ordinate axis—the corresponding intensities of ions of mass 3 (\(I_3\)). Curve \(I\) gives the result for hydrogen obtained by electrolysis from ordinary water, while curve \(III\) gives that for hydrogen enriched with deuterium by Brickwedde. As was to be expected, curve \(III\) is closer to a straight line than curve \(I\) (relative predominance of diatomic ions with mass 3). Further, it turns out that both of these curves can be represented by formula (5), if different values are used for the constant \(a\), but one and the same value for the constant \(b\). Namely, in the units of Fig. 1: curve \(I\)

\[ I_I = (6.6 I_2 - I_2^2)\cdot 10^{-5}, \tag{6} \]

curve \(III\)

\[ I_{III} = (190 I_2 - I_2^2)\cdot 10^{-5}. \tag{7} \]

Fig. 1

The difference between curves \(III\) and \(I\) is represented in Fig. 1 by the straight line \(II\) and is expressed analytically in the form:

curve \(II\)

\[ I_{II} = I_{III} - I_I = 183.4 \cdot 10^{-5} I_2 . \tag{8} \]

Since the deuterium content in both portions is small, it is evident that the enrichment does not appreciably affect the number of triatomic ions of mass 3 (\(\mathrm{H}_3^+\)), which also leads to the invariability of the quadratic

ISOTOPE OF HYDROGEN WITH MASS 3

member in formulas (6) and (7). Therefore the difference between the number of isotopic ions \((\mathrm{HD})^+\) in the two portions of hydrogen is measured by formula (8). Since \(I_2\) is measured by the number of \(\mathrm{H}_2^+\) ions, it is obvious that the angular coefficient of curve II gives the increase in the ratio of the amount of HD to \(\mathrm{H}_2\), due to the growth of the concentration of D

\[ \frac{\mathrm{HD}}{\mathrm{H}_2}=183.4\cdot 10^{-5}=\frac{1}{546}. \]

The ratio of the concentration of HD molecules to the concentration of \(\mathrm{H}_2\) molecules is given by the coefficient \(a\) of equation (5). For curve III, consequently, the relative concentration is

\[ \frac{\mathrm{HD}}{\mathrm{H}_2}=190\cdot 10^{-5}=\frac{1}{526}, \tag{9} \]

and for ordinary hydrogen (curve I)

\[ \frac{\mathrm{HD}}{\mathrm{H}_2}=6.6\cdot 10^{-5}=\frac{1}{15100}. \tag{10} \]

Passing from molecular concentrations to atomic ones, for Brickwedde’s portion, Bleakney obtained

\[ \frac{\mathrm{H}}{\mathrm{D}}=1050\pm 5\% \tag{11} \]

and for electrolytic hydrogen

\[ \frac{\mathrm{H}}{\mathrm{D}}=30\,000\pm 20\%. \tag{12} \]

It is more convenient that the same results can be obtained in a somewhat different way. Dividing equation (5) by the intensity \(I_2\) of the \(\mathrm{H}_2^+\) ions, we obtain

\[ \frac{I_3}{I_2}=a+bI_2. \tag{13} \]

Fig. 2

This is the equation of a straight line, and the intercept at the origin \(a\) gives directly the ratio of the number of diatomic ions with mass 3 \((\mathrm{HD})^+\) to the number of diatomic ions of mass 2 \((\mathrm{H}_2^+)\), since the linear term in (5) comes only from \((\mathrm{HD})^+\) ions. In Fig. 2 the corresponding results are presented. Along the abscissa axis, as before, are plotted the intensities \(I_2\) (pressures), and along the ordinate axis—the ratios of intensities \(\dfrac{I_3}{I_2}\). Curve I refers to electrolytic hydrogen, curve III to Brickwedde’s portion.

The number of triatomic ions \(H_3^+\) depends not only on the pressure, but also on the fields used in the mass spectrograph and on the geometry of the instruments, which may be more or less favorable for the formation of these ions. The coefficient \(b\) in equation (5) refers to the intensity of these triatomic ions and, as is clear from equation (13), is given by the slope of the curves in Fig. 2. The relative number of isotopic molecules

\[ \frac{HD}{H_2} \]

(the constant \(a\)) does not depend on these conditions, because the ions \((HD)^+\) and \(H_2^+\) are diatomic; the mechanism of their formation is the same, and changing external conditions lead to proportional changes in their intensities.

Curve \(IV\) in Fig. 2 was taken with different electric and magnetic fields for electrolytic hydrogen. Therefore its slope \(b\) changed greatly, while the initial ordinate \(a\), within the experimental error, coincides with the value of \(a\) for curve \(I\). Finally, curve \(V\) refers to a portion of hydrogen somewhat enriched in deuterium by diffusion with the light isotope. Accordingly, its initial ordinate \(a\) is smaller than for ordinary hydrogen.

It is known that, during the electrolysis of water, the light isotope of hydrogen \(H\) is preferentially liberated. Therefore the estimate of the content of \(D\) in ordinary hydrogen from curve \(I\) (Figs. 1 and 2) is clearly too small. Subsequently, using the same method, Bleakney and Gould\(^{8}\), having obtained hydrogen from a portion of ordinary water by its complete decomposition by passing its vapors over hot iron in a vacuum, were able to establish the true value of the deuterium concentration in natural hydrogen. It proved to be equal to \(D : H = 1 : 5\,000\).

The same procedure may be used to determine whether the hydrogen isotope of mass 3 \((H_1^3\) or \(T)\) is present in natural hydrogen. Its amount in hydrogen obtained from natural sources must be quite insignificant. However, if during electrolysis the water residue is enriched in deuterium, then it must also, and probably to an even greater degree, be enriched in the isotope \(T\). If one takes a portion of hydrogen highly enriched in deuterium (for example, with a content of \(99\%\), or about that), and assumes that, besides the isotopes \(H\) and \(D\), the isotope \(T\) is also present, then the quantity of \(D\) is significantly greater than the quantity of \(H\), and the quantity of \(H\), in turn, must be significantly greater than the quantity of \(T\). The pressure can be measured by the intensity of the diatomic ions \(D_2^+\),

\[ \left(\frac{m}{e} = 4\right), \]

which are present in overwhelming abundance. The ions \((DT)^+\) have the ratio

\[ \frac{m}{e} = 5. \]

In addition to them, the triatomic ions \((DDH)^+\), as well as \((HHT)^+\), have the same ratio

\[ \frac{m}{e}. \]

However, the latter are extremely rare, because there are few \(H\) atoms, and an utterly negligible number of \(T\) atoms. Ions of mass 5 thus consist of the diatomic \((DT)^+\) and the triatomic ions \((DDH)^+\); consequently, as before, their intensity will be represented by the sum of a linear and a quadratic function of the pressure or intens—

ness \(I_4\) of \(D_2^+\) ions. The linear term is given by the intensity of \((DT)^+\) ions, and the quadratic term by the intensity of \((DDH)^+\) ions. Dividing an equation analogous to equation (5) by the pressure (the intensity \(I_4\) of \(D_2^+\) ions), we obtain

\[ \frac{I_5}{I_4}=a+bI_4, \tag{14} \]

i.e. a straight line analogous to the straight lines in Fig. 2, whose initial ordinate \(a\) at once gives the ratio of the quantities \(\frac{DT}{D_2}\). Thus, by varying the ratio of the intensity of ions of mass 5 to ions of mass 4 at different pressures (different intensities of \(D_2^+\) ions), one can still determine the concentration \(\frac{DT}{D_2}\).

The first attempt by Bleakney and Gould\(^4\) led, within the errors of the experiment, to the value \(a=0\). From this, taking into account the experimental errors, the authors concluded that the content of \(T:D\) in their portions of water (91 and 98% heavy) was less than \(1:10^5\). Taking \(D:H=5000\) for ordinary hydrogen, the authors concluded that in natural hydrogen the ratio \(T:H<1:5\cdot10^8\), if it is assumed that during electrolysis no separation between \(D\) and \(T\) occurs. In reality, some increase in the concentration of \(T\) during electrolysis is to be expected; therefore \(T:H\) should be still smaller in nature.

Subsequently Bleakney, with collaborators, having constructed more sensitive apparatus\(^3\) and using a portion of water with 99% heavy-water content, established that the initial ordinate \(a\) has a small, but nevertheless nonzero, value; from this the authors obtained for their portion the value \(T:D=5\cdot10^6\), or for ordinary hydrogen \(T:H=1:10^9\) or less, if the relative enrichment of \(T\) during electrolysis is taken into account. By that time Rutherford’s experiments with collaborators\(^1\) had already been carried out; they had obtained artificial isotopes \(T\) by means of nuclear disintegration, and, as it seemed, evidence had also been obtained for the existence of \(T\) in the natural state by the already mentioned experiments of Tuve, Hafstad, and Dahl\(^2\). In addition, the experiments of Leitmeyer and Jung\(^11\), which are admittedly very doubtful, lead to the same result; they made use of the magneto-optical effect of Allison, still not yet understood.

The described experiments of Bleakney, which gave a positive result, concerned hydrogen obtained from the residue of electrolytic water, obtained from an initial volume 225,000 times larger. Later the volume was reduced to \(1:150\,000\,000\) of the original, i.e. \(75\,m\) of water were reduced\(^5\) to \(0.5\ \mathrm{cm}^3\). In this case the experiment showed that the intensity of the ion of mass 5 varied proportionally to the pressure, i.e. it had to be only diatomic \((DT)^+\) (in other words, the concentration of \(H\) is too insignificant for triatomic ions of mass 5 to be obtained in appreciable numbers).

The concentration \(T:D\) for this portion proved to be \(1:10^4\). Taking into account the enrichment that the isotope \(T\) must undergo during the various stages of electrolysis, the authors estimated the amount of \(T\) in ordinary water as \(T:H = 7:10^{10}\).

The results obtained were, naturally, of great interest to physicists engaged in nuclear processes. If it had been possible to obtain the isotope \(T\) in sufficient quantity, this would have given them yet another elementary “projectile,” with the aid of which—alongside the “projectiles” already at the disposal of physicists, the proton, deuteron, neutron, and \(\alpha\)-particles—it would have been possible to obtain interesting nuclear reactions, not to mention the fact that investigation of the nucleus of the isotope \(T\) itself would have led to important results. In view of this, Rutherford undertook an attempt to enrich water strongly with the isotope \(T\) by means of prolonged electrolysis. The results of the investigation were published in 1937.^12 At Rutherford’s request, from \(13000\ \text{l}\) of ordinary water in Norway an electrolytic residue of \(11\ \text{cm}^3\) was obtained, i.e. the volume was reduced by more than \(10^9\) times. Aston investigated this final portion on a mass spectrograph. In doing so, however, he used not Bleakney’s indirect method, but a much more direct one. In Aston’s mass spectrograph, beams of various ions are focused by means of magnetic and electric fields onto a photographic plate. The focusing in Aston’s instrument is so precise that ions of mass 5, \((DDH)^+\) and \((DT)^+\), owing to the fact that they possess, though very slightly, nevertheless different masses (the mass of the nuclei of \(T\) could be determined from the nuclear reactions of Rutherford and his collaborators), should have given separate lines on the photographic plate, \(0.5\ \text{mm}\) apart. The experiments gave a line of mass 5, which undoubtedly was the line of the triatomic ions \((DDH)^+\), but no traces of a line from \((DT)^+\) ions, which should have appeared as a satellite of the \((DDH)^+\) line, were obtained. Taking into account the exposure time and the resulting intensity of the \((DDH)^+\) ion line, Aston concluded that in this portion of heavy water the ratio \(T:D\) was in any case less than \(2:10^5\). This result plainly contradicts the data of Bleakney and collaborators, who obtained a ratio of \(1:10^4\) for a portion decomposed by electrolysis by approximately a factor of 10 less. Rutherford also tried to establish the presence of \(T\) in this portion of heavy water by using certain nuclear reactions that should have indicated its presence, but again with a negative result. On the basis of these data Rutherford concluded that the isotope \(T\) is probably unstable, which may explain its negligible quantities in nature.

The estimate of the amount of isotope \(T\) in Bleakney’s experiments, as we have seen, was based on the assumption that triatomic ions \((DDH)^+\) of mass 5 are formed only in two successive collisions (a quadratic dependence on pressure). However, Smith’s work^13 in 1937 showed that the primary ion,

HYDROGEN ISOTOPE WITH MASS 3

consisting of three hydrogen atoms, can be formed in a single electron impact from hydrocarbon molecules. Thus, if any hydrocarbons were present as impurities in the discharge, as happens quite often (vapors of vacuum grease, etc.), then, by means of a substitution reaction, they readily exchange their hydrogen for the deuterium surrounding them, and from the “heavy” hydrocarbons thus formed, in a single impact, triatomic ions \((\mathrm{DDH})^+\) may be obtained, whose number varies linearly with pressure, and hence they are mistakenly counted as diatomic ions \((\mathrm{DT})^+\) of the same mass 5. The mechanism by which triatomic hydrogen ions are obtained from hydrocarbons in a single electron impact is not yet clear; however, the experimental fact obtained by Smith is apparently correct.

To verify the results of the earlier experiments, and also, incidentally, the correctness of Smith’s conclusion, Bleakney and co-workers undertook new experiments \(^{6}\). In their work the authors used a diffusion apparatus for isotope separation, operating according to the principle proposed in its time by Hertz \(^{14}\). Later, Hertz proposed a more perfected method \(^{15}\), which Sherr \(^{16}\) used and which was employed in the experiments described by Sherr, Smith, and Bleakney \(^{6}\). We shall indicate here only the principle of operation of Hertz’s apparatus. A portion of gas is admitted into a system consisting of a series of diffusion pumps (in Sherr’s case, twenty-nine), and, as a result of the preferential diffusion of the light constituents of the gas either through the walls of porous tubes or through mercury vapor in the pumps themselves, with continuous pumping of portions of gas by the pumps from one place in the apparatus to another, in the end a certain equilibrium is reached, at which, at one end of the apparatus, the maximum concentration of the heaviest constituents of the mixture is established, and at the other, of the lightest. The magnitude of these maximum concentrations is the greater, the greater the difference in the masses of the heavy and light constituents of the gas, and it also depends on the type of apparatus.

Sherr, Smith, and Bleakney decomposed the purest heavy water that could be obtained over a hot tungsten filament in an evacuated isotope-separation apparatus. Before the pumps began operating, a small portion was taken from the contents and examined in the usual way on the mass spectrograph; it gave essentially the previous results, namely the formation of a primary ion with mass 5. Then the apparatus was set in motion, and, after equilibrium had been established, portions were again taken from the “light” and “heavy” ends of the system. The gas taken from the “light” vessel contained approximately 10% ordinary hydrogen and 90% deuterium, while the other impurities were extremely insignificant, obviously because the impurities, having a much greater molecular weight than \(\mathrm{H}_2\) or \(\mathrm{D}_2\), had almost entirely gone into the “heavy” vessel of the apparatus. In the mass spectrograph this portion gave a peak with mass 5, the relative height of which disappeared if the results were extrapolated to a pressure equal to zero, i.e.

the initial ordinate \(d\) for it is equal to zero, and hence the ions are exclusively triatomic—\((\mathrm{DDH})^+\). The gas from the “heavy” end of the apparatus contained many impurities and gave a much higher peak with mass 5, although it would seem that the molecules of light hydrogen should have passed practically entirely into the light fraction.

Then the entire heavy fraction was removed from the apparatus, and the remaining gas was again subjected to separation. Since the apparatus separated other gases from hydrogen with very great efficiency, this procedure practically removed from the instrument all the heavier compounds. Calculations, confirmed by measurements on various hydrogen gases, showed that if molecules of mass 5 were present, then in the heavy fraction there would be a noticeable enrichment in them as compared with gas that had not been subjected to the separation process. Analysis of this fraction showed that it was the purest deuterium that had ever been in the hands of an investigator. Experiments in the mass spectrograph showed that the ratio of the number of ions of mass 5 to the number of ions of mass 4 is \(3 \cdot 10^{-7}\), i.e., incomparably smaller than in portions of gas not subjected to the diffusion process. Since, with the indicated sequence of operations, the heavier gases were preferentially removed, while the number of molecules containing \(T\) in the heavy fraction should even have increased somewhat as compared with a fresh portion of gas, it seems very probable that the peak of primary ions with mass 5 was due to the presence of “heavy hydrocarbons.” Estimating the total enrichment that was to be expected both from electrolysis and from diffusion, the authors came to the conclusion that the ratio of \(T\) in ordinary hydrogen to the amount of \(H\) must be less than \(1 : 10^{12}\).

In the light of these new results it would be interesting to repeat the experiments of Tuve, Hafstad, and Dahl, in order to establish whether they had made any error.

It should, of course, be noted that the very low experimentally established upper limit for the abundance in nature of the hydrogen isotope with mass 3 does not necessarily lead to the conclusion that it is unstable, although some calculations \(^{17}\) speak in favor of this conclusion.

As a result of the analysis of the experiments described, it may thus be said that if the hydrogen isotope with mass 3 is present in natural sources, it is present in such an insignificant quantity that its separation is associated with incredible, if not altogether insurmountable, difficulties.

LITERATURE

1. V. I. Cherniaev, Priroda, 23, No. 9, 68, 1934.
2. M. A. Tuve, L. R. Hafstad a. O. Dahl, Phys. Rev., 45, 840, 1934.
3. W. Wallace Lozier, Ph. T. Smith a. Walker Bleakney. Phys. Rev., 45, 655, 1934.
4. W. Bleakney a. A. J. Gould, Phys. Rev., 45, 281, 1934.

1. P. W. Selwood, H. S. Taylor, W. W. Lozier, W. Bleakney, J. Am. Chem. Soc., 57, 780, 1935.
2. R. Sherr, L. G. Smith, W. Bleakney, Phys. Rev., 54, 388, 1938.
3. W. Bleakney, Phys. Rev., 41, 32, 1932.
4. W. Bleakney and A. Gould, Phys. Rev., 44, 265, 1933.
5. J. P. Tarnell and J. J. Livingood, Experimental Atomic Physics, ONTI, 1936, ch. IV.
6. V. I. Chernyaev, Priroda, 23, no. 2, 31, 1934; Advances in the Physical Sciences, 14, 711, 1934.
7. W. Laytimer and H. Young, Phys. Rev., 44, 690, 1933.
8. E. Rutherford, Nature, 140, 303, 1937.
9. L. G. Smith, Phys. Rev., 51, 263, 1937.
10. G. Hertz, Z. Physik, 79, 108, 1932; V. I. Chernyaev, Advances in the Physical Sciences, 14, 685, 1934.
11. G. Hertz, Z. Physik, 91, 810, 1934.
12. R. Sherr, J. Chem. Phys., 6, 251, 1938.
13. T. W. Bonner, Phys. Rev., 53, 711, 1938.