Observation of Mechanical Faraday Effect in Gaseous Media

We report the experimental observation of the rotation of the linear polarization of light propagating in a gas of fast-spinning molecules (molecular superrotors). In the observed effect, related to Fermi’s prediction of “polarization drag” by a rotating medium, the vector of linear polarization tilts in the direction of molecular rotation. We use an optical centrifuge to bring the molecules in a gas sample to ultrafast unidirectional rotation and measure the polarization drag angles of the order of ${10}^{-4}\mathrm{rad}$ (with an experimental uncertainty about ${10}^{-6}\mathrm{rad}$) over the propagation distance of the order of 1 mm in a number of gases under ambient conditions. We demonstrate an all-optical control of the drag magnitude and direction and investigate the robustness of the mechanical Faraday effect with respect to molecular collisions.

Figure 1

Scheme of the experimental setup. Top: femtosecond pulses with the central wavelength of 792 nm (upper, red) and 398 nm (lower, blue) are used for creating the centrifuge and the probe pulses, respectively. The pulses are shaped, delayed with respect to one another, combined in a collinear geometry, and focused in a gas cell. Bottom: after passing through the gas sample, probe pulses are filtered out from the centrifuge light and sent to the time-gated polarization analyzer, implemented with a boxcar integrator. PC, Pockels cell; $\lambda /2$, zero-order half wave plate; $P$, polarizer; $M$, metallic mirror; DM, dichroic dielectric mirror; $L$, lens; BC, Berek compensator (Newport 5540); WP, Wollaston prism (Thorlabs WP10-A); BD, balanced detector (Thorlabs PDB220A2). Alternatively, the probe pulses may be sent to a Raman spectrometer to characterize the rotation of the centrifuged molecules.

Figure 2

Experimentally measured decay of the polarization drag signal for two rotational frequencies of oxygen superrotors and two gas pressures. Cubic splines are drawn through the corresponding data points for guiding the eye. A “broken” centrifuge does not produce any molecular superrotors (see text for details). Vertical error bars represent the standard error of the mean over $\sim {10}^5$ laser pulses.

Figure 3

Experimentally measured dependence of the polarization drag signal on the initial rotational frequency of oxygen superrotors 2 ns after their release from an optical centrifuge (blue dots, connected by the cubic spline to guide the eye). The red and magenta dots correspond to the drag signal in ambient air 200 ps and 2 ns after the centrifuge, respectively. Vertical error bars represent the standard error of the mean over $\sim {10}^5$ laser pulses.