Storing Images in Warm Atomic Vapor

Reversible and coherent storage of light in an atomic medium is a promising method with possible applications in many fields. In this work, arbitrary two-dimensional images are slowed and stored in warm atomic vapor for up to $30\mu \mathrm{s}$, utilizing electromagnetically induced transparency. Both the intensity and the phase patterns of the optical field are maintained. The main limitation on the storage resolution and duration is found to be the diffusion of atoms. A technique analogous to phase-shift lithography is employed to diminish the effect of diffusion on the visibility of the reconstructed image.

Figure 1

The experimental setup and sequence. (a) The energy level scheme of the ${}^{87}\mathrm{Rb}$ $D1$ transition and the transitions of the pump and the probe. (b) The experimental setup: ECDL—external cavity diode laser, $\lambda /2$—half-wave plate, PBS—polarizing beam splitter, AOM—acousto-optic modulator, $\lambda /4$—quarter-wave plate, MS—magnetic shield, HC—Helmholtz coils. (c) The experimental sequence. The probe is shaped as a temporal Gaussian with $\sigma =5\mu \mathrm{s}$. After about half of the Gaussian pulse leaves the vapor cell (“leaked probe”), the remaining half is stored in the medium by turning off the pump beam. After a certain storage duration, the pump is turned back on, and the second half of the probe leaves the cell (“restored probe”). The figure shows probe detector traces for storage durations of 5, 15, and $25\mu \mathrm{s}$ and the pump trace for the $25\mu \mathrm{s}$ case.

Figure 2

Images of digits that were slowed and stored for different durations in warm vapor. The left column shows the original images that were sent to the cell. The middle column shows the image of the digit 2 that was slowed for about $6\mu \mathrm{s}$. The right column shows the images after they were stored and retrieved. The images of the digits 2, 6, and 9 were stored for 2, 6, and $9\mu \mathrm{s}$, respectively. The decay rate in the experiments was about $14000{\mathrm{s}}^{-1}$, and the total intensity in all of the experimental data was normalized for the clarity of presentation.

Figure 3

Storage of an image of three resolution lines for up to $30\mu \mathrm{s}$. The left-hand side shows experimental results (left column) and theoretical prediction (right column), when the image was sent to the cell with a constant phase. A good agreement between the experiment and the calculation was obtained. The right-hand side shows the improved resolution achieved when a $\pi$ phase shift was applied to the two outer lines in the image.

Figure 4

The visibility of the lines in Fig. 3 as a function of storage duration. The visibility of the constant-phase image drops rapidly due to diffusion, while the shifted-phase image shows only a small decrease in the visibility. The results are well described by our theoretical model, taking into account inaccuracies in the phase pattern of the phase-shift experiment. An error of $0.01\pi$ in the phase of the leftmost line gives the best fit to the experimental data (upper dashed curve). A curve for an error of $0.1\pi$ is also plotted (middle dashed curve) to demonstrate the effect of phase inaccuracies. The visibility presented here is obtained from the two leftmost lines, because of a relatively large error in the phase of the rightmost line (estimated to be $0.2\pi$).