ABSTRACTS

HEAVY HYDROGEN AND HEAVY WATER*

The presence of \(H^2\) in ordinary hydrogen was proved by Urey, Brickwedde, and Murphy\(^4\), who studied the optical spectrum of the residues from the evaporation of liquid hydrogen and found weak lines in the positions calculated for \(H^2\).

The width and doublet of \(H^2\) was resolved, and the separation of the lines proved to be in agreement with theory. Calculation of the ratio (abundance ratio) \(H^1/H^2\) gave 4000, which should have explained the above-mentioned discrepancy very well, but this result proved to be inaccurate. The phenomenon of isotopy was also found in the Lyman series of hydrogen obtained in the discharge in vapors of “heavy” water\(^5\).

The presence of \(H^2\) was further shown by Bainbridge\(^6\) with the aid of a mass spectrograph; he found an atomic mass of \(2.01351 \pm 0.00006\) relative to He and \(2.01351 \pm 0.0018\) relative to \(O^{16}\); the corresponding packing effect was \(0.675\%\). Mass-spectrographic evidence for the existence of \(H^2\) was also found by Kalman and Lazarev\(^7\). Tracy\(^8\) theoretically calculated the atomic weight of \(H^2 = 2.0113 \pm 0.0012\).

The nucleus \(H^2\), at first considered as a compound of two protons and one electron \((p_2 e)\), must, it is understood, play an essential role in the construction of nuclear structure. The discovery of the neutron and the positron (positive electron) gave several alternative possibilities for the structure of the \(H^2\) nucleus (which was named deuteron or deutron**); for example, it may consist of two neutrons, which are primary uncharged particles of mass 1, and one positron. According to this scheme the proton loses its former fundamental significance as the basic particle in the structure of the nucleus and proves to be a compound of a neutron and a positron\(^ {10}\).

Partial separation of the isotopes \(H^1\) and \(H^2\) was achieved by E. W. Washburn and Urey\(^ {11}\), who examined water from industrial electrolytic cells that had operated from two to three years. Study of the optical spectrum of hydrogen from this water showed an increase in the relative content of \(H^2\). Oxygen from this water was combined with nitrogen, and the spectrum of nitrogen oxide was studied. It turned out that there had been a decrease in the concentration of the isotope \(O^{18}\) by approximately \(8\%\). The ratio \(H^1/H^2\) in natural terrestrial hydrogen must depend on the properties of the samples studied, since there are several fractionation reactions during the preparation of hydrogen. It was also found that \(H^2\) tends to be absorbed (“clean up”) in a discharge tube\(^ {12}\); thus an estimate of the relative content based on the optical emission spectrum is, in all probability, erroneous. A noticeable increase was found\(^ {13}\) in the specific gravity, freezing point, and boiling point, as well as a decrease in the refractive index, of water that had undergone prolonged electrolysis.

Lewis and Macdonald started with 20 liters of water taken from an old electrolytic cell and formerly half-normal in alkali, and carried out electrolysis between nickel electrodes with a current of 250 A until the volume had been reduced by 90%. Nine-tenths of the amount obtained was neutralized

* From Nature, October 7, 1933, translated by M. Wolkenstein. For an account of the first works on the isotope of hydrogen \(H^2\), see G. S. Landsberg, Uspekhi Fizicheskikh Nauk, vol. XII, nos. 2–3, p. 343.

* According to the terminology now accepted in American literature, the isotope \(H^2\) is called deuterium. (Translator’s note*.)

with carbon dioxide and distilled off; the water driven off was mixed with the remaining \(1/10\) of the alkaline solution. Then the process was repeated until the volume had been brought to half a cubic centimeter, the electrolysis being carried out at a temperature below \(35^\circ\), and preferably around \(0^\circ\), in order to reduce losses by evaporation; the current was also reduced as the volume of liquid decreased. The difference in the cathode polarization between \(H^1\) and \(H^2\) reached 0.004 V. Accumulation of the heavy isotope of oxygen \(O^{18}\) did not take place. In the final result, water was obtained with a specific gravity of 1.073 and containing 65.7% \(H^2\). The losses of \(H^1\) and \(H^2\) were related as five to one, and a further reduction of the volume by electrolysis to \(1/4\) of the final volume should, as is supposed, yield water containing 99% of its hydrogen in the form \(H^2\). The estimate of the concentration of \(H^2\) in ordinary water as 1 : 6500, given in this article, is too high.

Newell and Ficklen\(^ {14}\) measured the specific gravity of water from baths used for chroming and operated for three years; samples from nine baths have specific gravities varying from 1.00002 to 1.00064.

In the distillation of water through a distillation column, considerable separation of the isotopes of hydrogen and oxygen may be achieved, especially if the distillation is carried out under reduced pressure\(^ {15}\). A column 6 m high was used in two ways: a) an ordinary mixture of isotopes was obtained at the base of the distillation apparatus; when the stationary state was reached after two days, the density of the water at the top of the column decreased by 60 millionths; b) enriched isotope mixture was obtained at the top of the column, and samples were taken daily from its base—the density increased by 70–80 millionths, in comparison with ordinary water. Although the vapor pressure of heavy water is noticeably lower than that of ordinary water, in separation by distillation this difference is sufficiently apparent, since the water molecules are continuously exchanging \(H^1\) and \(H^2\), and the corresponding calculation shows that it is not the vapor pressure itself, but the square root of it, that is proportional to the atomic concentration.

Distillation under reduced pressure should be more effective. Water from the bottom of a large distillation apparatus, operated for two months, showed a constant increase in its density\(^ {16}\).

The method of separating hydrogen and oxygen in a given sample of water and determining what part of the increase in density is due to \(H^2\) and what part to \(O^{18}\), used by G. N. Lewis and Macdonald\(^ {17}\), consists in passing steam over heated iron. But this is difficult, and Lewis\(^ {18}\) indicated a method reducible to isotope exchange in an aqueous solution, to the reaction of ammonia with water:

\[ H^1H^2O + NH^1H^1H^1 = H^1H^1O + NH^1H^1H^2. \]

Ammonia in water must form ammonium hydroxide hydrate \(NH_4OH\) and again dehydrate—both reactions proceed at high speed, and since the fourth hydrogen is exactly like the others, each hydrogen has equal chances of being split off during dehydration. Rapid exchange of positions of such hydrogen isotopes gives an approximately random distribution between \(NH_3\) and \(H_2O\). One mole of water at \(0^\circ\) absorbs about one mole of \(NH_3\), and thus ammonia contains three hydrogen atoms, more than half of all \(H^2\) present; it is removed when the ammonia is pumped off. A sample from the distillation apparatus, with a density exceeding that of ordinary water by 0.000182, was saturated with ammonia at \(0^\circ\), and then the ammonia was pumped off at room temperature. After sixfold repetition of the process, the water had an excess density of 0.000025, so that the increase in density by at least 0.000098 was due to \(H^2\). Another experiment consisted in using sulfurous gas instead of ammonia:

\[ H_2O^{18} + SO^{16}O^{16} = H_2O^{16} + SO^{16}O^{18}. \]

In this case, the initial excess density decreased to 0.000109, so that due to \(O^{18}\) the density increases by at least 0.000073. In these

in unfinished experiments the excess density by 0.000170, out of a total value of 0.000182.*

In precise work the isotope content in ammonia must be determined, in particular if it has been obtained from electrolytic hydrogen, and precautions must be taken against large losses of water during evaporation.

It is interesting to note that, according to Golmq’s report, electrolytic hydrogen is more active in ammonia synthesis than another, equally pure hydrogen obtained by reducing water vapor with iron: the reaction rate with electrolytic hydrogen was 30–40% greater.

A possible method for separating H² from electrolytic hydrogen was proposed by Bleakney, Gould, and Taylor²⁰, who report that enrichment of H¹H² occurs when the gas is gradually removed from the charcoal on which it had been adsorbed, in agreement with some theoretical results of Eyring.

Practically pure “heavy water,” H₂²O, was obtained by Lewis and Macdonald; moreover, their investigation of its physical properties was carried out with 0.12 cm³ of liquid, in which the content of the isotope H¹ was, in all probability, no more than 0.01%. The freezing point was determined as +3.8° and the boiling point as 101.42°. The vapor-pressure curve was established and the following values were obtained for the ratio \(p_2/p_1\), where \(p_1\) is the vapor pressure of ordinary water and \(p_2\) that of heavy water.

The usual calculation shows that the latent heat of vaporization is greater than that of ordinary water by \(259 + 3\) or 4.2 cal/mole. The density at 25° was determined as 1.1056 and the temperature of maximum density as 11.6°. The values of the ratio

\[ \frac{\text{volume}}{\text{volume at }4^\circ} \]

were obtained as follows:

Water is customarily regarded as a liquid anomalous in various respects; heavy water (H₂²O) is apparently still more anomalous, but the difference between it and ordinary water decreases with increasing temperature.

Recently, results have been reported from various investigations of H² and H₂²O. The spin of the H² nucleus²² is probably \(3/2\), or twice the spin of the H¹ nucleus. Studies²²,²³ were made of the emission spectra of hydrogen containing 25% H¹H¹, 50% H¹H², and 25% H²H², and of the electronic band spectrum of the neutral hydroxyl OH².

The mobility of ions in H₂²O was studied²⁴ by means of determining the electrical conductivities of hydrogen and potassium chlorides in ordinary water and in almost pure H₂²O. A special pipette cell was used, containing 0.25 cm³ of liquid, with electrodes made of platinum wire. The concentration of H₂²O in the electrolytic liquid was determined from the density, and the resistances were extrapolated from 97% H₂²O to 100%. The ratios of the specific electrical conductivities in ordinary water and in H₂²O were obtained. Results were found for five different temperatures from 5° to 18°. It is assumed that the ratio of the specific electrical conductivity to the specific electrical conductivity at infinite dilution is the same in both electrolytes and that the ratio of the mobilities of the ions K⁺ and Cl⁻ is the same in both solvents. Then it follows—

* Recently, indications have appeared in the literature of the possibility of exchange of hydrogen atoms between molecules of heavy hydrogen and ordinary water. This reaction must be of great theoretical interest, since it evidently proceeds without activation energy. There are also indications of an exchange reaction between atoms of heavy hydrogen and hydrogen in cane sugar (K. F. Bonhoeffer and G. W. Brown, Zs. Phys. Chem. B 23, 171, 1933). (Translator’s note.)

The calculation of mobilities at 18° for H²⁺, K⁺, and Cl⁻ in pure H₂²O gives 213.7, 54.5, and 55.3, if the corresponding mobilities of H¹⁺, K⁺, and Cl⁻ are 315.2, 64.2, and 65.2. It turns out that the mobilities of ions in heavy water are noticeably lower than in ordinary water. The values of the conductivity ratios were determined with an accuracy of up to 0.5%.

Lewis, even before he succeeded in concentrating H₂, predicted that water H²H¹O would not support life and must be lethal for higher organisms. He then showed experimentally²⁸ that tobacco seeds (Nicotina tabacum or purpurea) respond to the test as had been predicted. Twelve seeds were placed in pairs in six identical glass tubes, and 0.02 cm³ of ordinary distilled water was added to each of three tubes, and 0.02 cm³ of pure H₂²O to each of the remaining three. All six tubes were hermetically sealed and placed in a thermostat at 25°. Three pairs of seeds in ordinary water began to germinate after two days and, after two weeks, produced well-developed shoots. The seeds in H₂²O showed no development macroscopically; they were then placed in ordinary water, but the result of this was not reported. Six entirely similar tubes, each containing two seeds, were filled either with ordinary distilled water or with water in which half the hydrogen was H². At the end of four days all six seeds in the ordinary water had produced well-developed shoots, whereas the degree of germination of the seeds in the heavier water was approximately the same as in ordinary water after two days.

The toxic effect of ordinary distilled water is well known²⁶. The strange influence of thermal treatment of water on its ability to stimulate cell development, indicated by Lloyd and Barnes²⁶, is attributed to the varying content of polymerized water molecules in freshly condensed vapor and in newly melted ice. Ice²⁷ was considered as (H₂O)ₙ, perhaps (H₂O)₃, liquid water as for the most part (H₂O)₂ with some amount of (H₂O)₃ and H₂O, and dry steam as H₂O. It was calculated that at 20° liquid water contains 31.1% liquid ice, and that newly melted ice should contain much more (H₂O)₃ than newly condensed steam, unless the changes in polymerization occur very rapidly. In newly melted ice, rich in trihydrol, the cells of Spirogyra developed normally, whereas newly condensed steam, rich in dihydrol and monohydrol, killed the cells. Barnes and Jahn²⁸ reported that Englena develops much more rapidly in water from newly melted ice than in condensed steam.

However, the data concerning the influence of thermal treatment on the physical properties of water are contradictory. Wills and Becker²² say that the diamagnetic susceptibility changes under such treatment, whereas Menzies³⁰ finds no change in vapor pressure, and La Mer and Miller³¹ find no change in the refractive index. The above-mentioned assumptions about the structure of liquid water postulate tetrahedral arrangement of water molecules (five H₂O molecules in one group) in ice, while liquid water³² is composed of two other forms.

Until 1894 chemists did not suspect that the atmosphere contains any constituent parts other than those Lavoisier had recognized in the preceding century. It took two and a half centuries to establish the fact that water, which, like air, is one of the most common substances used in chemistry, contains, besides the two “elements” hydrogen and oxygen, each of which, as is now known, consists of mixtures of at least two kinds of atoms. It is highly probable that the chemical properties of heavy hydrogen isotope differ from the properties of ordinary hydrogen. The creation of a new organic chemistry, in which every compound containing carbon (which likewise consists of at least two isotopes, C¹² and C¹³) and hydrogen is doubled through the synthesis of a “heavy” partner—at present seemingly only a mirage—will after some time become, without doubt, an accomplished fact. The use of heavy water in medicine still awaits investigation.

Literature

1. Mecke and Childs, Phys. Rev. 36, 330, 1930; cf. Naudé, Zs. f. Phys. 88, 362, 1931.
2. Birge and Menzel, Phys. Rev. 37, 1669, 1931.
3. Bleakney, Phys. Rev. 41, 32, 1932.
4. Urey, Brickwedde and Murphy, Phys. Rev. 41, 32, 1932.
5. Ballardlow and White, Phys. Rev. 43, 941, 1933.
6. Bainbridge, Phys. Rev. 41, 115, 1932.
7. Kallmann and Lazarew, Naturwiss. 20, 206, 472, 1932; possibility of H³. Cf. also Conrad, Zs. f. Phys. 75, 504, 932, neutral particles H³ in canal rays; Lewis and Spedding, Phys. Rev. 43, 946, 1933, do not find spectroscopic evidence for the existence of H³ in an amount of 1 in \(10^6\) in almost completely pure \(H_2^2\); the relative fine structure of \(H^1\) and \(H^2\) is given.
8. Grace, Journ. Am. Chem. Soc. 54, 2562, 1932.
9. Harkins, ibid, 1254.
10. Sōxl, Nature 132, 174, 1933.
11. E. W. Washburn and Urey, Proc. Nat. Acad. Sci. 18, 496, 1932.
12. G. N. Lewis and Spedding, Phys. Rev. 43, 964, 1933.
13. Washburn, Smith and Frandsen, Journ. Chem. Phys. 1, 288, 1933; G. N. Lewis and Mcdonald, ibid 341.
14. Nowell and Ficklen, Journ. Am. Chem. Soc. 55, 2167, 1933.
15. G. N. Lewis and Cornish, Journ. Am. Chem. Soc. 55, 2616, 1933.
16. G. N. Lewis, ibid, 55, 3502; 1933.
17. G. N. Lewis and Mcdonald, Journ. Chem. Phys. 1, 341, 1933.
18. G. N. Lewis, Journ. Am. Chem. Soc. 55, 3502, 1933.
19. Pincass, “Die industrielle Herstellung von Wasserstoff,” 53, 1933.
20. Bleakney, Gould and Taylor, Phys. Rev. 43, 497, 1933.
21. G. N. Lewis and Mcdonald, Journ. Am. Chem. Soc. 55, 1933, Nature 132, 248, 1933.
22. Lewis and Ashley, Phys. Rev. 43, 837, 1933; Ashley, ibid. 770.
23. Chamberlain and Cutter, Phys. Rev. 43, 772, 1933.
24. Lewis and Doody, Journ. Am. Chem. Soc. 55, 3504, 1933.
25. Lloyd and Barnes, Proc. Nat. Acad. Sci. 18, 426, 1932; Nature 129, 891, 1933.
26. Barnes, Proc. Roy. Soc. A. 125, 670, 1929.
27. Barnes and Jahn, Proc. Nat. Acad. Sci. 19, 638, 1933.
28. Wills and Bocker, Phys. Rev. 42, 687, 1932.
29. Menzies, Proc. Nat. Acad. Sci. 18, 567, 1932.
30. La Mer and Miller, Phys. Rev. 43, 207, 1933.
31. Fowler and Bernal, Journ. Chem. Phys. Aug. 1933; Trans. Farad. Soc. 29, 1049, 1933.