I have built an f-2f interferometer. A dichroic mirror splits the supercontinuum into two arms of the Mechzehnder type f-2f interferometer.
One arm (red) has a spectrum of 1150-1700nm which after frequency doubling @1600nm in SHG crystal with an acceptance bandwidth of 4nm, the resulting second harmonic (800nm) has a bandwidth of 2nm.
The other arm (blue) has a spectrum of 600-1150 nm.
The two arms are made to combine on a beam splitter, spatial and temporal overlaps are confirmed by beam alignment in space and using a tunable delay respectively.
A 10 nm filter is used after a spatially and temporally overlapped beam to select only a 10 nm spectral region around the second harmonic signal (800nmm +/-5 nm) and the fundamental signal (800nmm +/-5 nm)
Assuming both beams aligned perfectly, what would the resulting interference look like?
Since, as I mentioned, the second harmonic has a bandwidth of only 2 nm and the blue arm beam has a 10 nm spectral width due 10 nm filter, what would be the effect of different spectral widths on the interference signal (carrier-envelope offset)? Do spectral widths need to be the same for CEP detection?
How to write the equations for interference in the time domain with different spectral widths
Thanks in advance!
The time-domain equations can be found in our 20-year old paper: IEEE J. Sel. Top. Quantum Electron. 9, 1030-1040 (2003). Regarding your question on the phase-matching bandwidth, I am not quite sure whether I understand the situation correctly. You are trying to measure the CEO frequency of an oscillator with MHz repetition rate? Then the conversion bandwidth should not matter much. You should just try to get about the same power (order of magnitude) in both interferometer arms. The situation would be different if you are trying to measure the CEP of a kHz laser system with spectral interferometry.
Thanks for the comment. Yes I am trying to measure the CEP of 32 MHz laser.
I am beating two 800nm signals coming from the two arms of Mech Zehnder intercerometer. Fundamental 800nm signal has 10nm bandwidth and frequrncy doubled has 2.5 nm bandwith. What would be the effect of different spectral widths on the CEP signal
So you are using a mode-locked oscillator. Then you simply have to place a fast photo diode in the output beam of your interferometer and look for the beat note on an rf spectrum analyzer, which should appear somewhere in the range from 0 to 16 MHz. And among all interferometer types, the Mach-Zehnder is probably the most difficult configuration as you have to very carefully adjust the length mismatch between the two arms. Michelson is much easier.
I used a 400 MHz photodiode and ESA to look for the beat signal but couldn't see any. I am sure that the lengths of the two arms are matched because I used a 30GHz photodiode and adjusted the length such that pulses from both arms overlap using a fast-sampling scope.
My mode-locked laser has a rep. rate 32 MHz and pulse width of 130 fs. I use around 0.5 m of PCF (NKT-nl-3.2-945) to generate supercontinuum. I wonder if my supercontinuum is not coherent that's why I can't see the beat signal.
Also, I spatially overlap the beams from two arms (visible spots before 10nm filter around 800nm ). Still, when I look at the spots after the 10nm filter, they appear misaligned slightly even though visible spots before the filter are perfectly aligned in near and far-field view. Is this because of dispersion, and should I care for overlapping spots after the 10nm filter because I want to beat the modes at 800nm to get the beat signal?
Have you checked the polarizations? You need to project the SHG and the fundamental into the same axis by using a polarizer in front of the diode.
Then your fiber seems to be rather long. We used something like 1 cm for best results at 800 nm.
And of course, 130 fs are rather long, too. So you may have lost the coherence here. I think there are only very few Yb:lasers that have been successfully stabilized with an f-2f interferometer.
You can write the interference condition for monochramatic light for example for electric fields E1, E2 or by using Gauss shaped laser beams or wave envelopes with spectral gauss shapes, time duration Gauss shaped and spatially radoial distributed according to Gauss distribution shaped(radial)/
- Badges
- Science topic
- Similar topics
- Filing