Experimental test of an online ion-optics optimizer

Abstract A technique has been developed and tested to automatically adjust multiple electrostatic or magnetic multipoles on an ion optical beam line – according to a defined optimization algorithm – until an optimal tune is found. This approach simplifies the process of determining high-performance optical tunes, satisfying a given set of optical properties, for an ion optical system. The optimization approach is based on the particle swarm method and is entirely model independent, thus the success of the optimization does not depend on the accuracy of an extant ion optical model of the system to be optimized. Initial test runs of a first order optimization of a low-energy ( 60 keV) all-electrostatic beamline at the NSCL show reliable convergence of nine quadrupole degrees of freedom to well-performing tunes within a reasonable number of trial solutions, roughly 500, with full beam optimization run times of roughly two hours. Improved tunes were found both for quasi-local optimizations and for quasi-global optimizations, indicating a good ability of the optimizer to find a solution with or without a well defined set of initial multipole settings.