VOLUME 88, NUMBER 7 PHYSICAL REVIEW LETTERS 18 FEBRUARY 2002 Measurement of the Electron Affinity of Cerium V. T. Davis Department of Physics, United States Military Academy, West Point, New York 10996 J. S. Thompson Department of Physics and Chemical Physics Program, University of Nevada, Reno, Nevada 89557-0058 (Received 28 August 2001; published 5 February 2002) The electron affinity of cerium has been measured using laser photodetachment electron spectroscopy. The electron affinity of Ce 1 G 4  was determined to be 0.955 6 0.026 eV. The data also show that Ce 2 has at least two bound excited states with binding energies of 0.921 6 0.025 eV and 0.819 6 0.027 eV relative to the  1 G 4  ground state of the cerium atom. The present experimental measurements are compared to recent calculations of the energy levels of Ce 2 . Strong disagreement with the most recent theoretical atomic structure calculations highlights the complicated nature of this particular lanthanide. DOI: 10.1103/PhysRevLett.88.073003 PACS numbers: 32.10.Hq, 32.80.Gc Studies of negative ions have resulted in advances in the understanding of electron-electron interactions in many-bodied calculations designed to model the properties of atoms, molecules, and clusters. For example, accurate calculations of the binding energy of the extra electron, for even simple anionic atomic systems, requires sophis- ticated calculations that include a detailed accounting of effects that can ignored in many atomic structure calcu- lations. Several recent reviews of negative ion research [1 –3] have pointed out the computational complexity encountered by theoretical investigations of lanthanide negative ions and the paucity of experimentally derived information for these ions. The importance of experimen- tally determined parameters, such as electron affinities (EA), is crucial to the understanding of the electron- electron interactions which are responsible for the exis- tence of negative ions. Negative ion structure calculations are difficult, and approximations are typically used in the calculations to reduce the number of terms contributing to the energy levels, so as to make the calculations tractable. Further- more, as atomic Z increases, the relative contributions of electron correlation and relativistic effects become com- parable, increasing the complexity of the calculation. The lanthanides are particularly interesting physically because of their unique properties, which result from the relation- ship of their 4f , 5d, and 6s electrons to one another. Although the small radii of the 4f orbitals shield them from outside influences, their binding energies are nev- ertheless comparable to their outer neighbors. Since the spread of energies within a particular configuration is much larger than the spread in those binding energies, the various configurations overlap one another to a considerable de- gree, making theoretical calculations based on the mixing of configurational basis functions extremely difficult [4]. This is particularly true in the case of cerium, which, ex- cept for Gd, is the only lanthanide with two partially filled subshells (4f and 5d). Experimental verification of the ex- istence of the predicted negative ion structure is therefore necessary to judge the validity of theoretical approxima- tions. In particular, since knowledge of the electron affini- ties of rare-earth atoms is limited, there is keen interest in experimental data concerning the electron affinities of the lanthanides [1]. Semiempirical estimates of the electron affinities of cer- tain lanthanides have been made in the past [5 –7]. A more recent theoretical calculation by O’Malley and Beck [8] was based on a relativistic configuration interaction cal- culation method which begins with a zeroth-order mul- ticonfigurational Dirac-Fock solution. This calculation yielded an electron affinity of cerium as 0.428 eV with a ground-state configuration for Ce 2 of [Xe] 4f 5d 3 6s and a prediction of 14 excited states of Ce 2 , six of which be- long to the ground-state configuration, and the other eight with a configuration [Xe] 4f 5d6s 2 6p [8]. The difficult nature of these calculations for Ce is highlighted by the fact that these calculations are refinements of a previous calcu- lation by the same method, in which the electron affinity (EA) of Ce was reported to be 0.259 eV [9]. Improvements in the calculations were reportedly due to a better treat- ment of second order effects and a more suitable choice of a neutral threshold [8]. Previous experimental investigators have reported production of stable lanthanide negative ions (to include cerium) using accelerator mass spectrometry (AMS) techniques [10,11]. The reported negative ion production yields for La 2 and Ce 2 were much higher than for the other atomic lanthanides, indicating that either the electron affinities of lanthanum and cerium are greater than other rare-earth atoms or La 2 and Ce 2 have more than one bound state [10,11]. The theoretical predictions of O’Malley and Beck [8] bear this out. Also, subsequent experimental studies have confirmed the existence of an excited bound state for La 2 [12]. Nadeau et al. have reported measurements of the electron affinities of Tm, Yb, and Dy using an electric field dissociation technique [13], although some of these results are disputed [14]. The relative yields of sputtered negative ions can be used 073003-1 0031-9007 02 88(7) 073003(4)$20.00 © 2002 The American Physical Society 073003-1