Our Results

Switching Data

Time response of the switch at 2.5 nW switch beam power. The top pane shows the power through the on state aperture (red). The lower pane shows the power through the off state aperture (blue). The data plots are out of phase showing that when the optical power in one spot is high, there is no power in the other spot.© Lucas Illing and Andrew Dawes, 2005

To accurately measure how our switch responds to the switch beam, we measure the power generated in one of the two spots as we turn the switch beam on and off. To select only one of the spots, we use an aperture that is a little smaller than the spot we want to measure. The aperture lets most of the spot through but blocks light coming from anywhere else in the pattern. Using two apertures, one for an ``on'' spot and one for an ``off'' spot, we can make sure that when the switch turns on it transfers power from the on spot to the off spot.

The figure to the right shows a plot of the power through the on and off state apertures (top and bottom respectively). We see that for a switching beam power of 2.5 nW the on and off states differ by an amount that is roughly 30 times the size of the noise fluctuations in the measurement. We define this ratio as the contrast-to-noise ratio, it is a measure of how sure we can be that the switch has turned on or off.

Switch risetime for 2.5 nW. The fit line indicates a risetime of 4 ms.© Lucas Illing and Andrew Dawes, 2005

In order to determine the amount of energy needed to actuate (turn on) the switch, we look closely at the amount of time the switch takes to go from off to on. Shown here is a detailed plot of the on-state power as the switch beam is turned on. From this a fit-line indicates a turn on time (or risetime) of 4 ms. From this measurement we can calculate the number of photons that entered the switch from the switch beam within that amount of time and determine that the switch requires 40000 photons (or 10 fJ of energy) to turn on.

Another way to quantify the sensitivity of our switch involves calculating the energy density rather than the total energy. This is a simple matter of dividing the total energy by the cross-section area of the pump laser beams. This is a useful measure of the sensitivity of our switch because it is possible to reduce the size of the laser beams in our experiment to an optimal dimension. In order to compare the number of photons in different sized beams, it is useful to give results in terms of densities. Additionally, we choose to describe the density in units of photons per l²/2p because the area l²/2p corresponds to the cross-section of a beam whose size has been reduced as much as possible. Using this measure, our higher-power switching data shows switch operation with 4.4 x 10^-2 photons per l²/2p.

Switching with only 230 pW

Time response of the switch at 230 pW switch beam power. The top pane shows the power through the on state aperture (red). The lower pane shows the power through the off state aperture (blue). Although the data plots are out of phase showing that when the optical power in one spot is high, there is less power in the other spot, for this lower switch beam power, roughly half of the output power is transferred from "off" to "on."© Lucas Illing and Andrew Dawes, 2005

To further understand our switch, we observe the behavior when the switching beam power is reduced to 230 pW. At this power, the switch is found to transfer roughly half of the power from the off spot to the on spot when the switching beam is applied. This can be seen in the plot to the right where the power in the on spot only reaches half of the value observed for higher switch beam power. Similarly, the off state power is only reduced from 1 to 0.5 rather than completely turned off.

Switching with 230 pW, we see a actuation time of 3 ms which translates to switch turn-on with 2700 photons or a switching energy density of 3 x 10^-3 photons per l²/2p. This photon number and energy density indicate that with modest optimization through a beam-size reduction of 10-30 times, we could possibly operate this switch in the few-photon regime.

Rise time data and fit for 230 pW switch-beam power. The switch actuates in 3 ms at this power.© Lucas Illing and Andrew Dawes, 2005