Typically, ultrafast pulse laser is realized by mode-lock technology and commercial products are available now, however, for most products, they only give pulse-width data, and the spectra-width data is “neglected”, of course, in most industrial application, that's not important, but for some quantum study, that matters. For example, some picosecond (
Hi Hongyan,
For a given pulsewidth, the phase of the pulse is what is going to determine the spectral bandwidth. If the phase varies linearly over time then the pulse will be transform limited (narrowest spectrum possible), if the phase varies quadratically, then the pulse is said to be positively or negatively chirped, depending on the sign of the curvature. The Fourier transform of such pulse gives a wider spectrum, the more dispersion the pulses has and the wider the spectrum gets. This kind of second order dispersion can usually be compensated using a pulse compressor (gratings, prims, chirped DBR,...). Coming out of a fiber, it is likely that the pulses will be positively chirped (easy compensation), but could also have some third order dispersion (cubic phase variation) which is more challenging to compensate.
I hope this helps.
Thank you! Alexandre, your answer is quite physically clear and help me a lot. Besides the complex dispersion and phase control, could I principly force a mode-locking laser giving narrow spectra pulse with a narrowing filter mirror, for example, Volume Bragg Mirror with less than 0.02nm FWHM line-width?
I think a narrow-band filter would blow up pulse duration.
However, if you're familiar with volume Bragg gratings, you might also be familiar with chirped volume Bragg gratings, which can act as ultrafast stretchers / compressors. One might be designed to compensate for phase dispersion.
Hongyan, the reason for pulses not to be transform limited is, basically Group Velocity Dispersion - the refractive index of any material depends on the frequency of the light, so upon propagation spectral components of the pulse accumulate rather different phases, so the pule-shape is distorted. There are several "tricks" to get your pulse close to transform limited. Chirped grating is one of them. There are commercially available chirped mirrors. Those are multi-layer mirrors, designed such that every wavelength is reflected form a diffrent mirror depth. Those are used to correct for propagation phase.
Additionally, you can have more versatile optical systems: either prism or grating based optical pulse compressors. Those are designed to disperse the pulse into spectral components, adjust relative optical path length of each of the components, and, fianlly recombine all those back to a single beam, to have the spectral components build the shortest pulse. An ultimate system, for such a pulse compression is pulse-shaper (papers from A. Weiner ~1995), where you can really adjust OPD for many narrow spectral bands costituing your pulse.
Another issue is pulse characterization. Well spectrum is easy, any spectrometer will do. Temporal duration is a different issue. Autocorrelation is one method, it is easy to set up, but it provides a rather limited information on your pulse shape. The most conveniet one of those I am aware, is FROG. Just read papers by R Trebino. They do offer commercial FROG packages.
could anybody tell me, why time domain description is more convenient in non linear optics?
Rajput: First of all time domain is more convenient description for very specific situations: for instance when you have no material resonances (no absorption etc) in the spectral range you are working. Then, in most cases, the response of the matter is instantaneous, but, obviously, following the excitation, so transition to freq. domain will make you add all kind of causality terms (similar to those you get in Kramers-Kronig relations). It is just unnecessary mess, for no benefit. Of course, if the situation dose involve any resonant terms - freq. domain becomes the natural choice, while, it gets additional benefit, since you can omit most of the contributions, except those involving resonances.
- Badges
- Science topic