VELOCITY DISTRIBUTIONS IN K AND TI BEAMS 1319 BEAM PURITY Ke is negligible in the present work. A search for TI The vendor of the potassium reported that it was made with a molecular beam apparatus. It was ined 2 percent sodium and 0.5 percent oncluded that there was less than 0.05 percent Tl2 Because of the low vapor pressure of these ty an amount wholly negligible in the present analysis stituents compared to that of potassium, no meas amount of sodium or lithium is expected to be present COMPARISON WITH OTHER RESULTS in the potassium beam with a magnetic resonance The experimental techniques used in two early atomic beam apparatus, but no sodium transitions were attempts to measure velocity distributions.were not sufficiently refined to permit an accurate test of the make it unlikely that an amount of sodium sufficient Maxwell velocity distribution law to be made. The to give a observable distortion of the velocity distribu- work of both Eldridge and Lammert shows deviations The thallium metal was reported to be 99.95 percent greater than the experimental uncertainties. Zartmand thallium,exclusive of the oxide on the surface. The investigated the velocity distributions in bismuth remaining 0.05 percent was largely lead, copper, and beams. Unfortunately, no critical test of the Maxwell cadmium, all of which are undetectable. Since no beam distribution could be made since bismuth vapor con- as detected with an unoxygenated detector wire, it is tains significant amounts of molecular species other certain that there were no measurable amounts of than atomic Bi. Thus zartman actually measured the esIum, rubidium, or potassium in the thallium beam. superpositions of distributions of Bi, Big, and Big e There is the possibility that the beam may also When he adjusted the amount of Big so that the agree- contain dimers of the atomic species. In the case of ment was good on the low-velocity side, there was a thallium and potassium the beam in question is in- pronounced excess of molecules on the high-velocity vestigated with an atomic beam apparatus that will side. The amount of Bis was small so that it little atoms with magnetic moments of the order of a effect on the major part of the distribution. Bohr magneton, which characterize the atomic states Ko used essentially the same apparatus as Zartman of potassium and thallium. Molecules of K, are in to measure the fraction of Big in bismuth beams as a a 2 states as are, presumably, those of the Tl, if they function of the oven temperature and pressure.From exist. They have magnetic moments of the order of a his data, he was able to calculate the dissociation nuclear magneton and are not significantly deflected energy of the Bis molecule. When the amount of dimerization was adjusted to give good agreement on by inhomogeneous magnetic fields which deflect atoms the low-velocity side of the distribution, there were too rough great excursions The molecule of potassium, K,, is known to be many molecules on the high-velocity side. He also found present in a potassium beam, but the data on the frac- evidence that Bis was present in the beam tion of molecules is inconsistent. In a molecular beam Cohen and Ellett measured velocity distributions in experimentl in which the nuclear magnetic moment of sodium and potassium beams. As a velocity selector potassium was measured, it was found that the ion current produced at the detector by the molecular perpendicular to the direction of propagation of the component may be as great as one percent of that beam. Since the deflection of beam atoms is proportional to the gradient of the magnetic field, it is important favorable experimental conditions, Rosenbergl4 ob- ent to be constant if the observed distribution could be tained 0.25 percent k, in a potassium beam at 5o0%k. fitted to the theoretical distribution after the adjust The fraction of K, may be shown from thermodynamic ment of constants not determinable from a priori considerations to increase with increasing oven tem- considerations. The nature of the constants is such peratures. Thus the distortion of the velocity distribu- that an independent check of the observed maximum and that deduced from the temperature cannot be ture cannot be the result of a larger fraction of K, made. Subject to these limitations, they observed no as the dimer would shift the velocity distribution to the systematic deviation from Maxwells theory at low low velocity side. If 0.25 percent K, is assumed to be oven pressures. At high oven pressures, they observed present in the beam, the maximum increase in the serious deviations from theory, notably a deficiency experimental velocity distribution at any velocity at knife-edge slits would give better results than their which observations have been made would be only 2-mm thick rectangular slits. However, on the basis of 0.5 percent of the maximum intensity. The effect of the that the 1s Kusch, Millman, and Rabi, Phys. Rev. 55, 1176(1939) ical distributions coul IR. Rosenberg, Phys. Re even at low pressures