With the IR detector and spectrometer in place, we searched for the IR emissions from the Sodium sample. These emissions would result from the transition from the virtual 5s state to the 4p states. However, after many attempts in varying configurations, we could not detect sufficient Infra-Red radiation. We verified that the detector functioned and tried again, but still could not get good data.
We sent a beam into the sample from the dye laser and then adjusted the dye laser to a wavelength at which we saw the brightest visible emissions. Using the spectrometer, we measured this wavelength at 608.55 nm, which corresponds to a double-photon energy of 32865 cm-1 and a dye laser dial setting of 24330. This is a virtual state of slightly lower energy than the real 5s state.
A scan over the expected emission wavelengths for transitions to the real 3p states yielded an intense peak at 639.7 nm in the emission spectrum. The peak was so intense that it pinned the maximum value Jim's program allowed, resulting in a meaningless graph.
This single value indicated transition to a single virtual state, since a transition to the real 3p states would yield 629.35 nm (3/2 state) or 628.61 nm (1/2 state). We believed we were seeing four wave mixing. To verify this, we detuned the laser above and below the 608.55 nm beam and looked for proportional shifting of the peak emission.
We tuned the laser to a dial setting of 24500, corresponding to a laser wavelength of 612.5 nm. Our emission spectrum was:
The centre of this emission peak is at 640.03 nm, a shift toward longer wavelength. However, the shift does not correspond to the decrease in energy for the virtual state, so the second virtual state must be shifted as well.
Now, if we see a shift downward when we tune the laser to a dial setting of 24200, corresponding to a laser wavelength of 605.5 nm, we will know our virtual state is moving.
Our peak is centred on 638.8 nm, the downward shift we expected. Again, it is not equal to the energy shift in our virtual 5s state, so the intermediate state must be a virtual 3p state.
We next tuned our laser so that we would excite to the 3d state for Sodium. Tuned to a dial setting of 27432.2, or 686.1 nm, our dye laser caused emissions at 589.2 nm and 589.8 nm.
On further inspection, we noticed that each of these values was exactly one Angstrom away from the calculated for the transitions back in the Theory. If this is a slight miscalibration of our dye laser, then we are right on the 3d and 3p states!
We did successive runs detuning the laser above and below the values used above, so as to ensure we were entering a virtual 3d state.
Slightly detuned, we noticed that the peaks were at the same wavelengths as when we stimulated to the real 3d state. This indicates that we are in a virtual 3d state, but dropping to the real 3p states. We are getting simple Raman scattering.
Further detuned, we only see one clear peak. However, detuning in each direction does not affect the wavelength, which remained that for a 3p 3/2 transition to ground.
Detuning even further in either direction resulted in no viable output. The above graphs show the noise recorded. At first glance it looks as if the graph on the left, tuned several Angstroms below the wavelength necessary to excite to the real 3d state, looks to have a peak at about 589.3 nm. However, the magnitude of this peak is on the same order of the negative peak at 588.25 nm, and that is obviously due to noise. It is safe to assume, then, that we are not seeing a shifted 589.2 nm peak (our calibration-shifted 589.1 nm peak) but we are seeing coincidental noise.