Knowledge Bank

New high resolution Fourier Transform spectra of the $0-3$, $0-4$, $1-3$, $1-4$ and $1-5$ bands of the $B\prime \mathrm{<mrow\; data-mjx-texclass="ORD">2</mrow>}\Sigma +-X2\Sigma +$ transition of MgH, together with lines from sunspot spectra coupling $B\prime$ -state levels $v\prime =0\&1$ to X-state levels $v\prime \prime =3-8$, combined with older $B\prime -X$ band $dataa$ involving $v\prime \prime (X)=3-9$ and $v\prime (B\prime )v\prime \prime =0-9$ and ground-state $2-1$ and $1-0$ infrared data, yield a description of the ground state for $v\prime \prime =0-2$ and $v\prime \prime =3-9$. The $v\prime \prime =2-3$ gap was bridged using transitions from the $1-2$ and $1-3$ bands of the $A2\Pi -X2\Sigma +$ spectrum. In order to avoid complications due to perturbations in the excited state, all of these electronic band data were re-arranged and treated as fluorescence series into the ground state. The resulting data set consisted of a total of 4140 transitions for six isotopomers of MgH and MgD. We have performed two types of combined isotopomer analyses of these data. (i) A fit to empirical Dunham-type expansions which included hydrogenic Born-Oppenheimer breakdown correction terms required 54 expansion parameters, plus the 701 fluorescence series'' origins. However, the resulting empirical centrifugal distortion constants will be unreliable for extrapolation to J values significantly higher than those included in the data set. (ii) An equally good fit to an analytic model potential energy function plus adiabatic and non-adiabatic radial Born-Oppenheimer breakdown correction functions required only 21 potential and correction-function parameters, plus the 701 fluorescence series'' origins. The resulting potential function has the MLJ $form,b$ and was constrained to have the correct $C6/R6$ long-range behaviour, so it should prove highly reliable for extrapolations in v or J. The parameter-fit and potential-fit computer programs used for this analysis were DSParFit and $DSPotFit.c$