= Kusch-Foley magnetic moment experiment
{c}
<Polykarp Kusch> and <Henry Michael Foley> compared <hyperfine structure> resonance frequencies for different electronic states of <gallium>, <indium> and <sodium>. Their 1948 paper reports mutually consistent departures from the ratios expected with orbital factor $1$ and spin factor $2$. Comparing several states constrained the alternative explanation of perturbed atomic levels.<The Magnetic Moment of the Electron by Polykarp Kusch and Henry Michael Foley (1948)>[Abstract: atomic states, frequency comparisons and interpretation]{ref}
The <Rabi molecular beam magnetic resonance experiment> explains the readout principle: a resonant transition changes an atom's magnetic response, and hence its transmitted beam signal. Kusch describes rapidly interchangeable ovens, several <radio frequency> oscillators and careful field homogenization. Measurements taken during monotonic field drift were reduced to a common field.<The magnetic moment of the electron by Polykarp Kusch>[pp. 303–304: resonance analysis, ovens, oscillators and field drift]{ref}
\Image[/kusch-foley-resonance-comparison.png]
{title=Compare differently weighted atomic magnetic moments}
{description=Original functional diagram of the <Kusch-Foley magnetic moment experiment> and a separate ideal <LS coupling> calculation. The upper flow identifies the functions of beam preparation, resonance readout and atomic modeling; it is not a surviving apparatus drawing. The lower weights are calculated for $L=1$, $S=1/2$, $g_L=1$ and chosen $g_s=2.0024$, without nuclear <spin>. The frequencies in the teaching model cannot be substituted directly for measured hyperfine lines.}
Real <hyperfine structure> lines are not simply bare electronic $g_J$ times field. Nuclear <magnetic moments>, known interaction constants and nonlinear field dependence enter the inference.<The magnetic moment of the electron by Polykarp Kusch>[pp. 303–305: hyperfine modeling and atomic-state perturbations]{ref} Field homogeneity controls line broadening; alternating species addresses drift. Neither control removes an error in the atomic model. The beauty of the comparison is the separation of two problems: cancelling a shared calibration scale and testing whether several physical systems imply the same <electron spin g-factor>.
Back to article page
