Advantage of narrowing the bandwidth of a fs laser?

I am reading about the possibility of narrowing the bandwidth of a femtosecond laser to picosecond pulses. I understand how it works, but I cannot get enough information about the advantages of doing it.
Since it is not my field, I would like to know why people are interested in converting a femtosecond laser to a picosecond.
Which are the main applications for this? What kind of labs/research would be interested in doing it?
I am not referring to those fs lasers that allow to switch from one mode to the other, but to those setups where a fs is passed through a tunable filter like an AOTF, a LCTF or a grating.

Thanks for your help!

Diane

I am not entirely sure what your question is, the normal reason to convert a fs pulse to a ps pulse is to avoid damaging your laser system (fs pulses have high enough electric fields that there are a lot of nonlinear processes that happen in places that you do not want them to happen). It is usually done by a processes called 'chirping', which is simply putting the beam through a dispersive element (AOTF, grating, glass, special mirrors, etc if you are starting with a fs pulse just about anything will work) that stretches the pulse in time by spreading out the wavelengths of the pulse along the direction of propagation. If you are careful you can also reverse the process to get the original short pulse back (ex, if you want to amplify fs pulse you usually need to stretch it, then amplify the stretched pulse, and recompress it afterward).

Normally you do not want to get rid of bandwidth in the process, however there are nonlinear processes which make this possible, ex you can take advantage of spectral narrowing (which is normally considered a parasitic process, and the system is designed to minimize it) to get a narrower (in spectrum) pulse than you started with, without loosing the energy that is contained on the wavelengths that are eliminated. I have never head of an application where this would be beneficial, because you might as well just start with a laser source that has a narrower bandwidth in the first place! But I suppose that if you have your heart set on using a broadband source (say perhaps, you want the particular properties of a cep stable ti:saph laser) but only need wavelengths in a narrow range (ex, seeding a nd:yag amplifier) it could be worth it to you to use a nonlinear fiber to get spectral narrowing to compress all of the energy you have into the wavelength range that is used by the amplifier.

It also desensitizes your system to wavefront tilt if you do not need the fS pulse. I just got hit with a similar question in a job interview. I asked why, he asked. If your experimental system needs colliding fS pulses, the leading edges of the fS pulses across the beam profile need to be uniform and parallel in time domain for maximum interaction during a measurement. If you can slow things down, you have a better chance of having a interaction.

Otherwise you slow pulses down to amplify them as Krazer states. Running a ultrashort pulse , unchirped, through a amplifier will usually result in a damaged amplifier.

Steve