Broad linewidth continuous wave laser

Hello! I have a situation in which I need (try) to scan a laser frequency over a given range in search of a resonance in an atom. In our case the linewidths are very large 10-100 GHz (and for some reasons we can't cool the atoms down, so we are kinda stuck with these linewidths for now). Also the atoms are produced in small amounts (and in a beam, not a cell), so we would like to be as effective as possible. As the atomic beam is continuous we are using a continuous wave laser (a pulsed laser would miss a large number of the produced atoms). The bottom line is that I need to scan ~100 GHz linewidths with a CW laser. Of course if the linewidth of the laser is very small, say 1 MHz, I am able to scan ~##1/10^5## at a time. I was wondering how wide can a CW laser be or how can I make it wider such that I can cover more atoms at a time? Thank you!

BillKet said:

The bottom line is that I need to scan ~100 GHz linewidths with a CW laser.

Maybe not. Consider a third way.

I believe the problem you have can actually be better resolved by using a white light source, that has broadband energy content.

You make a rapidly changing spectral comb by rapidly modifying the optical path, each line on the comb can be an analyser. To separate the individual channels from the record requires an FFT. There is a transform gain in the process that can significantly increase the dynamic range of your instrument, which can more than compensate for a limited digitisation depth.

I think you might start by studying the way that an FTIR works, then consider an instrument based on that pool of concepts.
https://en.wikipedia.org/wiki/Fourier-transform_infrared_spectroscopy

Baluncore said:

Maybe not. Consider a third way.

I believe the problem you have can actually be better resolved by using a white light source, that has broadband energy content.

You make a rapidly changing spectral comb by rapidly modifying the optical path, each line on the comb can be an analyser. To separate the individual channels from the record requires an FFT. There is a transform gain in the process that can significantly increase the dynamic range of your instrument, which can more than compensate for a limited digitisation depth.

I think you might start by studying the way that an FTIR works, then consider an instrument based on that pool of concepts.
https://en.wikipedia.org/wiki/Fourier-transform_infrared_spectroscopy

Thanks a lot for this! Before I look more into it, can this method be used with frequency doubling/tripling? My transitions is around 250 nm.

BillKet said:

Thanks a lot for this! Before I look more into it, can this method be used with frequency doubling/tripling? My transitions is around 250 nm.

I guess it will come down to the implementation of your optical path. You will have to search the web to evaluate the components or technology that best suits your unusual application.

There has been FTIR for three or four decades. I am now out of touch with the field. Now there is also FTUV, here is an example.
https://www.envea.global/s/emissions/engine-exhausts/beryl-92m/

https://www.researchgate.net/figure...isplacement-x-along-the-detector_fig1_5599558

BillKet said:

Before I look more into it, can this method be used with frequency doubling/tripling?

I don't believe so. Even if you had sufficient phase matching, the spectral intensity of white light sources is much, much lower than a laser. Since frequency doubling and tripling are nonlinear processes, the doubling/tripling efficiency falls with lower intensity.

However, there are spectral lamps that can act as broadband sources in the near UV. Try looking for commercial spectral lamps, such as xenon arc or xenon-mercury arc sources. However, keep in mind that your spectral intensity will be quite low. I don't know about FTUV, but FTIR experiments typically have lots of time for averaging to improve the SNR caused by low intensity. I think you don't have that kind of time because your atomic beam is moving. I could be wrong.

UV frequency combs exist; however, they are whole experiments by themselves (for example). And quite technically demanding experiments at that.

For inspiration, you might want to read about results in slowing molecular beams on CaF and on YO (they also did this on SrF but their paper doesn't really give any useful info on the technique). I don't know if their techniques will work for you, because it sounds like your beam is much hotter (and likely faster) than theirs (based on the 100GHz linewidth).

The only thing I can think of off the top of my head is a Ti:Sapph comb frequency doubled (I'm not even sure how possible that is) and SFG'd (sum frequency generation) with a 633nm diode laser. Sounds like a massive headache though. And seriously expensive. Maybe you could homebrew a low-bandwidth Ti:Sapph comb? Sorry, I'm really not well versed in comb engineering.