13 November 1998
Position Paper on Atomic Structure

EXECUTIVE SUMMARY:

• DIRECTIONS: Goals: (1) To obtain a base of precise measurements of strategically selected atomic structure and collision data of sufficient scope and; (2) to develop theoretical and computational methods of sufficient generality and accuracy; (3) so that any data that are needed can be routinely generated.
  
  
• CONNECTIONS: These quantities are needed for quantitative formulations in areas such as chemical reactions, molecular bonding, scattering of particles, the absorption and emission of light, plasma modeling and diagnosis, materials science, etc.
  
  
• APPLICATIONS: Specific applications can be found in plasma-aided manufacturing (atomic layer deposition, surface etching, materials hardening, bulk material synthesization, semiconductor and microelectronics processing, plasma remediation of toxic exhaust), combustion, nanostructures, plasma display
  panels, photovoltaic technology, quantum computing, high temperature superconductor technology, the interpretation of energy production data from astrophysical sources, etc.
  

JUSTIFICATION:

The ultimate directions and goals of Atomic Structure and Atomic Collision research during the next decade should be to achieve a situation whereby the properties of all atoms and ions can routinely be specified to high accuracy. Quantities such as energies, transition probabilities, ionization potentials, electron affinities, electric polarizabilities, magnetic g-factors, interaction cross sections, hyperfine structure, etc, have essential connections to the quantitative formulation of chemical reactions, molecular bonding, materials science, scattering of particles, the absorption and emission of light, and plasma dynamics. These have applications in plasma-aided manufacturing, fusion energy production, combustion, nanostructures, photovoltaic technology, quantum computing, the interpretation of energy production data from astrophysical sources, etc. Experimental measurements cannot be made for all possible atomic systems, so the measurements must be made for carefully selected key systems, and methods must be developed and tested to permit accurate prediction of all systems.

The present situation in Atomic Structure lies far from achieving these goals, but we have now entered a promising era when unprecedented experimental precision can be achieved in the study of virtually any ion or atom, and tremendous computer power can be brought to bear on the comprehensive characterization and representation of these measured data. While technical advances have been made in computation, real fundamental theoretical questions about atomic structure have remained unanswered since 1924.

In his recently published book What Remains to be Discovered, John Maddox (the former Editor of Nature) lists a wide-ranging account of our present scientific ignorance. Among the crucial unanswered questions he includes "How can we reconcile Relativity and Quantum Mechanics?" The plethora of different computational methods that are variously optimized to specific classes of systems indicates that not simply new computations, but new theoretical ideas are needed, based on comprehensive and precise arrays of measured data. Many existing computational methods are based on self-consistent central field methods that treat many short-range interactions only perturbatively.

The answers to these questions require disciplined systematic and programmatic studies of atomic properties, which are sometimes less dramatic than powerful technical demonstrations of known principles. Ugo Fano has publicly cautioned against tendencies toward the "use of very expensive and technically sophisticated equipment to demonstrate that nature does indeed behave as we all know it does." The disciplined programmatic studies of atomic structure should not be sacrificed in favor of eye-catching one-time-only demonstrations of the validity of known physical relationships. (In a sports metaphor, the important problems will require efforts more akin to those of Cal Ripken that to those of Roger Maris).

Present computational methods work reasonably well (although not to spectroscopic accuracies) for low Z atoms, but many of these perturbatively treated interactions scale with high powers of Z. Thus they can dominate the dynamics of heavy atoms, although they are not being included in the specification of the wave function. Heavy atoms not only contain more electrons, they also possess inner shell electrons that are much more severely affected by relativity, quantum electrodynamics, and correlation. Consequently, discrepancies between theory and experiment for these systems often exceed 30%. Present theoretical methods are based on the optimization of the total binding energy of all of the electrons, which is largely derived from portions of the wave function far from the nucleus. The short-range interactions are included only perturbatively, and thus these wave functions cannot be expected to precisely describe the atom in the region near the nucleus.

One way to study the internal workings of a heavy neutral atom is to study highly ionized states of the same atom. If a theoretical model is to predict the global properties of a neutral lead atom, it should also predict the properties of a lead atom with 40 or 50 electrons stripped off. Much work remains to be done in the determination of transition probabilities. For multiply charged ions, the data consist almost exclusively of lifetime measurements, and virtually no branching ratio measure-ments have been made. Thus theoretical methods for the computation of branching ratios in complex ions are virtually untested. Electric polarizabilities of ground state ions are needed for the specification of long range interactions. Many types of devices exploit unusual conditions in a specific atom or ion (cancellation effects, fortuitous configuration interaction, wavelength congruencies) which can lead to metastable or anomalously short-lived levels, population inversions, etc. Such cases provide very severe challenges to theoretical methods.

Thus systematic and disciplined studies aimed at developing the means to comprehensively, accurately, and reliably quantifying atomic structure will be needed to meet the challenges of the next decade.