Recently Conradi and Norberg have proposed that a small density of molecular hydrogen in $a$-Si: H films provides the relaxation mechanism which is responsible for a minimum in the proton spin-lattice relaxation time $T_1$ at about 30 K. Although we are unable to observe an NMR line attributable to the ${\mathrm{H}}_2$, we are able to observe the conversion of the ${\mathrm{H}}_2$ molecules from the ortho state to the para state at 4.2 K. The process is bimolecular with a rate constant of 0.010 ${\mathrm{h}}^{-1\mathrm{}}$. The existence of a large number of sites able to trap such a small molecule may provide an important insight into the defect structure of these films.

• Received 2 November 1981

DOI:https://doi.org/10.1103/PhysRevB.25.1435