Decoupling Pulse Duration and Pulse Energy

Posted on 2008-09-24. Permanent link: http://www.rp-photonics.com/spotlight_2008_09_24.html

Ref.: encyclopedia articles on Q-switched lasers, master oscillator fiber amplifier

When the pulse repetition rate of an actively Q-switched laser is increased, while the (continuously applied) pump power remains constant, the pulse energy is reduced. An important side effect of this is that the pulse duration is increased. Therefore, the peak power is reduced even more, as it depends on both the pulse energy and the pulse duration. As a rule of thumb, which is normally quite precise for repetition rates above the inverse upper-state lifetime of the gain medium, a doubling of the repetition rate will reduce the pulse energy to one half the original value, double the pulse duration, and reduce the peak power by a factor of 4.

That relation between pulse repetition rate and duration is undesirable for some laser applications e.g. in the area of laser material processing. A change in repetition rate alone usually has the simple effect of linearly rescaling the processing time, assuming that the workpiece fully recovers between subsequent pulses. The additional change of pulse duration, however, can have effects which are difficult to predict.

A possible solution is to generate the pulses not with a Q-switched laser, but rather with a master oscillator power amplifier (MOPA) setup. If the amplifier is a fiber amplifier, the device may also be called master oscillator fiber amplifier (MOFA). (Some even call it a fiber laser.) The seed laser of such a device may be a laser diode. The seed pulse duration can simply be controlled via the electronic driver. Possible gain saturation effects in the amplifier can be compensated by adjusting the duration and possibly also the temporal shape of the seed pulses. Effectively, it becomes possible to decouple the pulse duration from the pulse energy. This can be very useful in cases where the repetition rate needs to be changed all the time, or some finite pulse trains are needed with larger temporal gaps in between, and each pulse (including the first one in each burst) needs to have well controlled parameters.

Unfortunately, that change of system architecture can also have disadvantages. A possible issue at least for fiber amplifiers is that fiber nonlinearities and the possibility of damage set limits to the peak power which can be generated that way. Also, such a MOPA has a high gain, and is thus very sensitive to backreflected light.

It depends on the circumstances which of the two concepts is better suited. A conventional Q-switched laser may be preferable when a variation of repetition rate is not required, or when a very high pulse energy is needed. On the other hand, the flexibility of a fiber-based MOPA may be very desirable particularly for the development of new applications. One may even easily experiment with different temporal pulse shapes.