884
Vol. 58
COMMUNICATIONS TO THE EDITOR
prising when we consider that the chief processes Calculations of the change of potential energy of interchange are essentially irreversible. Web- of a hydrogen bridge, connecting two like oxygen ster, Wahl and Urey [iba.,3, 129 (1935)] have atoms, as the hydrogen moves along the 0-0 found that water made from the oxygen in carbon centerline show in fact two minima, if the 0-0 disdioxide is 10.2 p. p. m. heavier than water con- tance is taken as the experimental value (2.76 8.) taining the same hydrogen but with its oxygen in ice, when one uses for the 0-H interaction enisotopes in equilibrium with the carbon dioxide. ergy a modified Morse potential function of the There can be little doubt that the carbon dioxide form [Huggins, J . Chem. Phytys., 3, 473 (1935); of the atmosphere is in equilibrium with the water also another article in press] of the earth. Plants, taking in two atoms of carL’ = e - a ( r - C‘c-a’(r - la) bon dioxide oxygen and one atom of water oxygen return to the atmosphere a molecule of oxygen and obtaining the constants from band spectrum data build into plant substance the remaining atom of for the OH molecule. The minima are about 1.10 8.from each oxyoxygen. If we assume that there is no discrimination between the isotopes of oxygen in this process, gen. The hump between them is about 0.15 X ergs high-nough larger than ‘/2 kT at we find that the air oxygen produced by plants should be approximately 6.8 p. p. m. heavy. The ordinary temperatures (-0.02 X ergs) to small difference between this figure and the value make jumps of the hydrogen over or through this found by Dole may be due either to other slower barrier very rare. One concludes then that the processes tending toward a state of true equilib- bridges in HzO are of the unsymmetrical type, each rium or to a preferential selection of 0l8for the H nucleus, in effect, bridging between an OH- ion and an Ha0 molecule. The potential curve for synthesis of plant substance. We hazard the prediction that water prepared such a case has but a single minimum, close to the from the oxygen in carbohydrates either will have OH- end. In other words, two hydrogens of the the same density as water from atmospheric oxy- four surrounding each oxygen are close (-1.1 B.) gen or will be slightly heavier. Since we do not and two more distant (-1.65 A.). A very slight expect to be able to verify this prediction experi- unsymmetrical Coulomb force added to the 0-H mentally for some time we have no desire to re- interactions assumed above suffices to eliminate serve the field. We hope that others will be inter- entirely the central hump. Although a single hydrogen cannot jump to the ested in examining the isotopic composition of the other end of its bridge by itself, jumps in unison of oxygen in natural organic substances. BYERLY LABORATORY CHARLES H. GREENE all of the hydrogens in a ring are possible. This RADCLIPFE COLLEGE ROGER J. VOSKUYL mechanism accounts for the dielectric behavior CAMBRIDGE, MASSACHUSETTS quite as well as the less likely assumption of rotaRECEIVED FEBRUARY 13, 1936 tion of the HzO molecules. Consideration of the effect of the additional THE STRUCTURES OF ICE AND LIQUID WATER Coulomb forces in the cases of hydrogen and hySir: droxide ions leads one to conclude that the former Although the assumption of four “hydrogen are best considered as (H20HOH2)+ions with a bridges” or “hydrogen bonds” connecting each relatively strong symmetrical bridge in the center oxygen atom to adjacent oxygens in ice and liquid while the latter are merely OH-. Either can conwater accounts satisfactorily for many of their temporaneously add an H 2 0 molecule a t one end properties, certain properties such as the low and lose one a t the other, the positive ion doing conductivity of pure water and the high dielectric this more readily than the negative. This afconstants (for low frequencies) for both solid and fords a chain mechanism satisfactorily accounting liquid near 0’ seem to require that these bridges for the relative mobilities of these ions as combe unsymmetrical [Huggins, “Chemistry, the pared with others. Science of Atoms and Molecules,’’ Preliminary These subjects and others related to them will edition, 1933; Bernal and Fowler, J . Chem. Phys., be discussed in detail elsewhere. 1, 515 (1933)l. Pauling’s recent calculation THEJOHNS HOPKINS UNIVERSITY MAURICE L. HUGGINS [THISJOURNAL, 57, %SO (19:33)] of t h e entropy B A L 7 IMORR, MARYLAND RECFHVED FEBRUARY 2-1, 1936 of ice also favors this coriclusiun. Y12)
