optical heterodyne detection for spectrometer

Sure, that's exactly how I measure the linewidth of our diode lasers.

(LW about 1 MHz on the second time scale.... Mostly jitter, I think of the optics in the signal path, but it could be something else. On shorter time scales the LW looks to be ~1-200kHz, but that's close to the 'scope resolution.) And then of course I'm measureing the LW of two lasers, and assuming it's shared equally.

George H.

Hi,

I came across optical heterodyne detection on wikipedia:

Basically hitting a photodiode with two closely matched light sources, and then the electrical signal from the photodiode will output the frequency difference between the two light sources. Could something like this be rigged up to function as a spectrometer? If one signal was a reference light, a light from a supercontinuum laser that is swept over the desired light frequency range, then it might be similar to a spectrometer with a rotating grating and a photodiode. Would there be a way to operate in continuous mode so no frequency sweep is required?

cheers, Jamie

Yes, it has been done by Jun Ye at JILA in Boulder, CO. He and his colleagues combined a pair of femtosecond frequency combs with slightly different spacings on a photodiode and used an rf frequency analyzer to measure the optical spectrum change due to molecular absorption of one of the beams. A femtosecond comb is a phase locked set of optical waves where each "tooth" in the comb has a width below 1 Hz, the teeth are separated by say 100 MHz, and the frequencies range from say 3E14 Hz to 6E14 Hz (1000 nm to 500 nm in wavelength). The second femtosecond comb has the "teeth" separated by 100 MHz + 50 Hz. A pair of "teeth", one from each comb, are phase locked to be 50 MHz apart, then the beat from the next pair of teeth is 50 MHz + 10 Hz, etc. So the beats from the (6E14-3E14)Hz/100 MHz = 3E6 pairs of teeth in the two frequency combs are mapped into the range between

Bret Cannon