~o r e generally, if there i s an appreciable spread in the magnitude of the moments of the different clouds, a s recent experiments (see Ref. 4) indicate, the saturation magnetization should be equated to 3kCcw/rllt* with rllZ* = in2] /Ln], where the bracl\-ets signify averages over the moment distribution. 'B. Mozer, D. T. Keating, and S. C. Moss, Phys. Rev. G, 868 (1968). 8~. W. Cable, E. 0. Wollan, and H. R. Child, Phys. Rev. Letters g, 1256 (1969). $~h e possibility that many of the

more »

... ation clouds a r e nucleated at Fe o r other magnetic impurity sites is prccludcd by the low impurity level of -30-40 ppm in our samples. 'ON. D. Lang and H. Ehrenreich, Phys. Rev. 168, 605 (1968). "By assuming that the magnitude of each Ni-atom moment in Ni-Cu depends only on its local chemical environmeilt, C. G. Robbins, H. Claus, and P. A. Beck [Phys. Rev. Letters 22, 1307 (1969)l obtained an empirical fit to their data for bulk ferromagnetic moment versus alloy composition. However, this assumption ignores purely magnetic correlations, i.e., factors in the local environment that a r e explicitly magnetic and not uniquely determined by the chemical surroundings. The distailce Erom a Ni-rich nucleating site must be such a factor. Thus, the polarization clouds in Ni-Cu a r e basically analogous in origin to those in Pd-Fe A two-center shell model with oscillator potentials, 5 S forces, and i2 t e r m s is developed. The shell structures of the original spherical nucleus and those of the final fragments a r e reproduced. F o r small separation of the two centers the level structure resembles the Nilsson scheme. This two-center shell model might be of importance in problems of nuclear fission. F r o m physical intuition it is evident that it is not possible to describe the process of nuclear fission all the way from the ground state of the fissioning nucleus up to the final stage of two separated fragments by means of a one-center shell model like, e.g., the Nilsson model. It is, instead, essential to allow for a preformation of both final fragments in the deformed shell model. This is in fact an additional degree of freedom. In a recent paper this type of single-particle potential was proposed.' This potential consists of two connected oscillator potentials, including a spin-orbit force and T2 term.2 Only the angular-momentum-independent t e r m s have been treated in Ref. 1. In this paper we report the results of a m o r e realistic calculation including all the T-dependent t e r m s and obtain consequently the c o r r e c t asymptotic single-particle levels for a symmetric two-cent e r shell model. The Hamiltonian for this potential is where 5 is the single-particle momentum and V is the momentum-independent p a r t of the potential. I t i s noticed that V describes two connected oscillators: while the momentum-dependent t e r m s becoine V($) = -K~W ,~{ Z~. S + g [ c 2 -S~(~ +3)]}, = -K E w~{~~. s + p [ g 2 -$~(~+ 3 ) ] ) , (3) where Tl and Tz describe the angular momenta I with respect to the two centers at z = -2, and z = + z " respectively. It follows from (2) and (3) that the Harniltonian (1) indeed contains for z, = 0 the c a s e of a spherical Nilsson potential and f o r z , = R (R being the nuclear radius) the c a s e of two identical and well-separated potentials of the Same type. This behavior is due to the special Ansatz for V and is, therefore, automatically present also for the 6-dependent t e r m s in (1). The structure of those latter t e r m s i s determined by quite general invariance requirements. The shape of the two connected (amalgamated) nuclei described by (1) i s spherical. It is also