We have used the Doppler-free, perturbation-facilitated optical-optical double-resonance technique to investigate the vibrational, rotational, and hyperfine structure of the 3 (3)Pi double minimum state of NaK. Since this electronic state arises from an avoided crossing with the nearby 4 (3)Pi state, we observe striking patterns in the data that provide a sensitive probe of the electronic wave function in the various regions of the double well potential. A single-mode cw dye laser excites 2(A) (1)Sigma(+)(v(A),J) approximately 1(b) (3)Pi(Omega=0)(v(b),J) mixed singlet-triplet "window" levels from thermally populated rovibrational ground state levels, 1(X) (1)Sigma(+)(v(X),J+/-1). Further excitation by a single-mode cw Ti:sapphire laser selects various 3 (3)Pi(0)(v(Pi),J(Pi)) rovibrational levels, which are detected by observing direct 3 (3)Pi(0)-->1(a) (3)Sigma(+) fluorescence in the green spectral region. Using the inverse perturbation approximation method, we have determined a 3 (3)Pi(0) potential curve that reproduces the measured energies to approximately 0.24 cm(-1). In addition, the hyperfine and spin-orbit constants, b(F) and A(v), have been determined for each region of the potential curve.

The excited 3 (3)Pi and 4 (3)Pi electronic states of the NaK molecule exhibit an avoided crossing, leading to the anomalous behavior of many features of the rovibrational energy levels belonging to each state. A joint experimental and theoretical investigation of these states has been carried out. Experimental measurements of the vibrational, rotational, and hyperfine structure of numerous levels of the 3 (3)Pi state were recently obtained using the Doppler-free, perturbation-facilitated optical-optical double resonance technique. Additional measurements for the 4 (3)Pi state as well as bound-free emission spectra from selected 3 (3)Pi, 4 (3)Pi, and mixed 3 (3)Pi to approximately 4 (3)Pi rovibrational levels are reported here. A model is also presented for calculating the mixed rovibrational level energies of the coupled 3 (3)Pi-4 (3)Pi system, starting from a 2x2 diabatic electronic Hamiltonian. The 3 (3)Pi and 4 (3)Pi potential curves and the coupling between them are simultaneously adjusted to fit the observed rovibrational levels of both states. The energy levels of the potential curves determined by the fit are in excellent agreement with experiment. The nonadiabatic coupling is sufficiently strong to cause an overall shift of 2-3 cm(-1) for many rovibrational levels as well as somewhat larger shifts for certain pairs of 3 (3)Pi to approximately 4 (3)Pi levels that would otherwise be very close together.

Earlier high-resolution spectroscopic studies of the fine and hyperfine structure of rovibrational levels of the 1 3delta state of NaK have been extended to include high lying rovibrational levels with v < or = 59, of which the highest levels lie within approximately 4 cm(-1) of the dissociation limit. A potential curve is determined using the inverted perturbation approximation method that reproduces these levels to an accuracy of approximately 0.026 cm(-1). For the largest values of v, the outer turning points occur near R approximately 12.7 angstroms, which is sufficiently large to permit the estimation of the C6 coefficient for this state. The fine and hyperfine structure of the 1 3delta rovibrational levels has been fit using the matrix diagonalization method that has been applied to other states of NaK, leading to values of the spin-orbit coupling constant A(v) and the Fermi contact constant b(F). New values determined for v < or = 33 are consistent with values determined by a simpler method and reported earlier. The measured fine and hyperfine structure for v in the range 44 < or = v < or = 49 exhibits anomalous behavior whose origin is believed to be the mixing between the 1 3delta and 1 1delta states. The matrix diagonalization method has been extended to treat this interaction, and the results provide an accurate representation of the complicated patterns that arise. The analysis leads to accurate values for A(v) and b(F) for all values of v < or = 49. For higher v (50 < or = v < or = 59), several rovibrational levels have been assigned, but the pattern of fine and hyperfine structure is difficult to interpret. Some of the observed features may arise from effects not included in the current model.

A two-step excitation experiment has been employed to measure the collisional rate coefficients and to study the velocity distribution of Cs͑6P 1/2 ͒ atoms that have undergone a single hyperfine state-changing collision with Ar. In addition, argon pressure broadening rates and shifts of the cesium 6P 1/2 ͑FЈ͒ → 8S 1/2 ͑FЉ͒ transitions have been determined. In the experiment, a single mode, cw Ti:sapphire laser is tuned to line center of the 6S 1/2 ͑F =4͒ → 6P 1/2 ͑FЈ = 3 or 4͒ transition. Then, the frequency of a single mode cw dye laser is scanned over the 6P 1/2 → 8S 1/2 manifold to probe the populations of the 6P 1/2 hyperfine levels. Absorption of probe laser photons is monitored by detecting 8S 1/2 → 6P 3/2 fluorescence. The experiment is conducted at room temperature, where the Cs density is low ͑n ϳ 3.4ϫ 10 10 atoms cm −3 ͒, and thus the probability of a Cs-Cs collision is negligible during the Cs͑6P 1/2 ͒ radiative lifetime. The Ar pressure is varied from 0 to 1.52 Torr, and Cs-Ar collisions cause population to be transferred from the directly excited 6P 1/2 ͑FЈ͒ level to the other 6P 1/2 hyperfine level. The data are analyzed using a density matrix formalism to yield the rate coefficients for Cs͑6P 1/2 ͒-Ar hyperfine state-changing collisions. In addition, the one-dimensional velocity changing collision kernel for Cs͑6P 1/2 ͒ atoms prepared with v z = 0 that undergo FЈ =3↔ FЈ = 4 hyperfine state-changing collisions with argon is reported.

No abstract