Giant Chirp Mode-Locked Ring Laser Oscillators

In this blogpost we will deep-dive into the technology of Giant Chirp Mode-Locked Ring Laser Oscillators. A brilliant laser design concept which is the main column of simplification.

Mode-locking itself can be imagined as a single pulse circulating the laser cavity and we couple out some of it in every roundtrip. The saturable absorber is responsible for making sure there is only a single pulse circulating and it is the one which is the most intense.

Such oscillators exist in many forms: solid state linear cavity, thin disk or slab based cavity, optical fibre based ring cavity.  

Saturable absorbers
Mode-locked fibre ring oscillators have been around for a long time. Initially nonlinear polarization rotation (NPR) saturable absorber was used for mode-locking. Then later on semiconductor saturable absorber mirrors (SESAM) were used. NPR was not easy to start and environmental effects often disrupted the operation. SESAMs started the operation easily but on the long term they degraded so reliability was not great. In the last 10-15 years artificial saturable absorbers are used (nonlinear loop mirrors, graded index fibres or other Kerr-lens based devices) which can indeed provide reliable start-up and long term reliability. Regarding the future, there is a great interest towards Mamyshev oscillators.

Saturable absorbers are specialty filters which have lower loss for intense light than for low-intensity light. Let's get back to this topic later.

We will get soon to the giant-chirp. Bare with me. 

Ring oscillators
In mode locked ring oscillators a single laser pulse circulating. You can imagine four mirrors which create a ring (rectangle) light path. Inside the path there is a laser crystal. The pumped laser crystal upon tuning on the pumps emits randomly photons through spontaneous emission. If some of these emitted photons are taking the ring path in the cavity, they will get amplified and the laser starts operating. If we want to avoid infinite energy build up inside the oscillator, we make one of the mirrors a beam splitter and couple out the same amount of light which gets added to the beam by the laser crystal. 
..So far this is only a "continuous wave (CW)" laser. That means, no pulses, just continuously power (like a laser pointer) coming out of the laser ring cavity. To make a single pulse circulate inside the cavity we need to add a saturable absorber - an optical component which has a high loss (like absorption) for low intensity and low loss for high intensity light. Pulses are higher intensity than continuous power so this will be the preferred operation instead of continuous. 
You might ask where the pulses come from if there are no pulses initially? Well, good question. They originate from power fluctuations of the CW operation. 

Disclaimer: the reality is more complicated than this simple overview. 

Giant chirp
Now, that we mostly know how the mode-locked ring laser works, we can move on to giant chirp. 

Chirp is basically the property of a material / glass that the refractive index (speed of light in the material) is dependent on the wavelength. Let's assume our laser pulses has a wavelength spanning from 1063nm to 1065nm. If we send this pulse through a piece of glass, due to the refractive index (and speed of light) difference at 1063nm and 1065nm, the 1065nm part of the pulse will pass through the glass faster than the 1063nm. The speed of 1064nm will be between 1063 and 1065nm. That means, the pulse will get longer while passing through this medium = it will get a chirp. If the pulse passes through a very thick (kilometres long) piece of glass, it will get a giant chirp!

Now let's add an optical fibre (a kilometre long piece of glass) in our ring laser. The below animation shows what happens with the pulse inside the cavity.

Starting from the top right mirror, the short pulse get's more intense while passing through the gain medium (laser crystal or doped fibre). Then it gets stretched by the giant chirp introduced in the 1km long piece of fibre. Next we couple out some light from the cavity (you know, the prevent infinite energy build up). Finally the saturable absorber makes the remaining pulse short again. From here we just repeat. 

How many times do we repeat per second? This depends on the total path length of the ring laser cavity! So, the light speed is about 300 000 000 meters per seconds in air, meaning that the pulse goes around 300 000 times per second if we just calculate with the physical path length. At the same time we also couple out this many pulses per second at the beamsplitter.

If you are interested in some literature, even though NPR-based, Smirnov et al made a nice review in 2011 about such lasers: (PDF) Mode-Locked Fibre Lasers with High-Energy Pulses

But what is all this good for? 
On the one hand side we have a very manageable pulse spacing. 300 000 pulses per second is 3.3microseconds between them. Electronics operate at this speed very well while with a much shorter cavity 10s or MHz repetition rates would be real hard to manage. Furthermore, imagine that eventually these pulses will be used to machine some materials. The most typical spot size in focus of a galvo scanner and f-theta based system is 30um. If you place 300 000 laser pulses to 30um distance from each other, you get a 9meter per second scan speed requirement for the galvo. This is already getting close to the fastest speed these devices can scan the laser. With a 10MHz laser you would certainly end up using an expensive polygon scanner or similar if overlap of pulses if not desired. Most ultrashort pulse lasers of course rely on Electro-Optic Modulators and they just pick pulses from the oscillator. This little trick adds a US$10k module and control electronics to the assembly. 

A further advantage of the giant chirp is that the pulses will need stretching anyway for the chirped pulse amplification. If the pulses are not chirped, as they get amplified they will exceed the laser damage threshold of the glass they are being amplified in or produce other undesired nonlinear effect like stimulated Raman scattering, stimulated Brillouin scattering, supercontinuum generation, four-wave-mixing, self and cross phase modulation, self-focusing … So, having the pulses already stretched saves us the material cost of the pulse stretcher too. 
   
And this is still not the end of the list of advantages. The material dispersion (chirp) being mostly linear can be easily compensated with chirped volume Bragg gratings or diffraction grating pulse compressors. Some other tricks to get rid of expensive laser components use non-linear chirp. When the pulse is attempted to be compressed with a linear chirp, the efficiency will be very low. 

Finally, there is one more advantage which we have not mentioned yet. The laser gain medium converts the amount of pump light to our lasers wavelength at a certain efficiency. This is usually around 1-10% with ultrashort pulse lasers. A 10W of pump light will produce every second 0.1W - 1W of average output power at the laser output. Power is a physical quantity defined by energy per unit time - Joules per second. If the oscillator produces 1W average power from 10W pump power, then this 1 Joule of energy is distributed evenly among the pulses emitted per second. With 300 000 pulses we get a pulse energy of (1Joule / 300 000) = 3.3uJ but with 10MHz (10 000 000) pulses we get only 0.1uJ. This is very small energy but still 33x more with the giant chirp-oscillator. If the application requires 100uJ of pulse energy then a 10MHz laser oscillator will at least need two amplifier stages to achieve 100uJ. With the giant-chirp oscillator we will only need a single stage. Here we go again, we saved again a lot of money on the material cost of the amplifier.  

Summary
All-in-all the giant-chirp concept saves us a LOT of trouble. We save on material cost of the pulse stretcher, amplifier, pulse picker and all the related electronics. We also save a lot of labour cost on the assembly of all these components and testing the final product. Furthermore, our whole operations are positively affected: we have less money sitting in our inventory, less work for the purchasing department and the reduced complexity is paired with better quality and reliability.