Hello everyone,
I created the following ultrafast laser design. Ultrafast laser design is rather difficult so I need you to challenge my idea, why this would not work. I hope you can help.
This is the optical schematics (optical simulation model).
The dark red beam is the pump which creates a ring formed (telecentric) pumped area on the Yb:CaF2 crystal with 10mm outer diameter and 9.4mm inner diameter. There is one nonlinear Michelson interferometer on each side of the crystal. The one on the left side is for passive mode locking and the one on the right is for coupling out pulses. Basically the left side works as a output coupler for low intensity light and the right side works as a mirror for low intensity light. As the pulse energy of the circulating pulse increases, eventually the right side couples out most of the pulse and the left side reflectance increases. Note, the pulse will have to circulate many times before it reaches the energy that it can couple out.
This passive mode locking technique was used earlier in this paper and later on it was called Additive-pulse Mode Locking with minor changes in the concept.
Let me know if you need more information to understand the concept. 
With best regards,
Daniel
The design is wonderful!
However, I think there will be some restrictions right?! Like why do you use a lot of GDD GTI mirror? Is that to compensate the dispersion?
How about small incremental variation if some unwanted natural vibration and thermal effect come into play? How to compensate this effects as well?
And how do you can obtain TEM00 mode instead of ring shape beam profile?
2 Likes
Excellent questions there! 
With femtosecond lasers often only fused silica optics are used because of the extremely low GVD and nonlinearity. Just 1mm of the AMTIR-1 glass has around 1000fs2 GDD compared to fused silica’s 57fs2. I used the GTI mirrors to compensate for this dispersion. (2 passes on 1mm AMTIR-1) Otherwise the second order dispersion would not just accumulate in multiple passes and broaden the pulse, but would cause a chrip in single pass and the electric field with the reference arm would not overlap.
The low GTI mirror in the reference arm compensates for the rest of GDD (coating, Yb:CaF2 crystal, beamsplitter). The total GDD is slightly negative so as the pulse energy grows, together with the self-phase modulation (intensity dependent GDD if you like) it comes to zero GDD. At the target pulse energy basically there is a soliton pulse going around in the cavity until it couples out.
Vibration might be an issue indeed. I’m hoping to build a 3D printed mount which dampens the vibrations. Keeping the cavity as short as possible is part of the mitigation too. The pulse will devlop very quickly in the cavity, the roundtrip time is about a nanosecond.
Thermal effects can cause issues in multiple ways. Thermal lensing is one possible issue: in the Yb:CaF2 it’s due to the pumping and it’s different with/without laser operation. The pumped area (ring) is in direct contact with the cooling all around it which reduces the effect similarly as in thin-disk lasers. Since I posted this I noticed that the AMTIR-1 has a weak two-photon absorption around 1050nm and low thermal conductivity which causes extreme thermal lens so I replaced it with a different nonlinear glass.
On the long term thermal index change and thermal expansion will cause slow length changes compareable with the wavelength so the laser will have some warmup time. 15-30minutes warmup is a usual thing with other solid state lasers too, but for other reasons.
I wonder if the nonlinear index changes with temperature 
Converting the ring to TEM00 goes with an axicon beam compressor (two axicons) but I would not expect perfect Msquared=1. Mainly because the axicon aperture is limited at the tip. Anyway, I don’t think TEM00 is necessary anyway - at this pulse energy and length even a 30-50um spot is perfectly suitable to ablate any material (when the beam is focused for ablation).
If interested you can find some more insights here:
But of course further questions are very welcome too.
