The theoretical and experimental understanding of the low-lying S and P electronic states of francium is beginning to reach a level comparable to that of the other alkali-metal atoms @1‐4#. Their energies, dipole matrix elements and hyperfine constants show that the heaviest alkali, even though it is radioactive, is subject to quantitative analysis that makes it a very promising candidate for precision measurement tests of fundamental symmetries in nature @5#. To further increase our understanding of this atom, the low-lying D states have to enter the picture. The very different angular momentum properties of these states add some additional complications to the quantitative understanding of the simplest of the heavy elements. Optical double resonance is a well-established tool of laser spectroscopy @6#. It allows the study of an excited state through an intermediate, well-characterized one. The development of laser trapping and cooling techniques opened further its applicability. The magneto-optical trap ~MOT !@ 7# operates in a regime where some population is in the upper state, making it an ideal intermediate state for optical doubleresonance studies @8‐11#. Francium is a short-lived radioactive alkali that we are able to study now thanks to laser trapping and cooling techniques @12#. The number of atoms captured in a MOT and cooled to a fraction of a mK is enough for optical doubleresonance spectroscopy. In this paper, we present the location of the second excited state in the D series, the 7 D state of 210 Fr, and a measurement of its hyperfine splitting. To find the state, we use a semi-empirical approach based on an extrapolation of a quantum defect fit~QDF! of the previously measured high-lying states of the nD series @13#. Since the trapping laser is intense, the upper state of the cycling transition has a significant ac Stark splitting ~AutlerTownes!. We resolve this line splitting intrinsic to the cooling process and study in a simplified model its influence on the precision of our measurements. We have to rely on such models to study many of the possible systematic effects present in the optical double resonance in a MOT since we do not have the possibility to study them directly, given the availability of Fr. The following section of the paper describes the trapping of Fr and experimental techniques of exciting and detecting the 7D states. Section III relates ab initio and semi-empirical calculations of the energies of the 7 D states; it also describes numerical modeling of the Autler-Townes splitting of the transition. Section IV describes the measurements of the hyperfine structure of the 7D states and studies of the different contributions to the measurement uncertainty. We report the results of the measurements of the hyperfine splittings, the energies, and the fine structure in Sec. V. The conclusions are presented in Sec. VI. The Appendix describes how the detuning frequency of the trap laser may be obtained from the observed Autler-Townes splittings.