Chirped-Pulse Amplification Lasers

The chirped-pulse amplification laser concept is a qualitative leap in the techniques of high-peak-power laser systems. All laser systems simply repackage energy as a coherent package of optical power; chirped-pulse amplification lasers repackage the laser pulse itself during the amplification process.

The problem:

In a conventional master-oscillator power-amplifier (MOPA) laser system, a tiny prototype laser pulse is passed through a series of optical amplifiers. In high-peak-power laser systems, this tiny pulse is amplified until it begins to incur one of several nonlinear problems associated with intense light:

B-integral problems associated with the index of refraction of transparent materials: At high optical intensities, materials begin to explicitly show the dependence of their index of refraction on intensity. The index of refraction determines the phase velocity of light and the optical path length experienced -- the effective phase-distance that light travels in propagating through matter -- and so intense light begins to suffer phase delays, relative to weaker light. The accumulated phase-lag suffered by light in travelling through a medium is given by the B-integral:

Thus a beam with a nonuniform intensity distribution, say a Gaussian intensity cross-section, passing through a window of glass or even a distance of air suffers a delay of phase at the centre of the beam which differs from that at the beam edges. This alteration of the phase of the lightwave is a distortion of the light beam which can seriously degrade its ultimate capacity to be brought to a focus.
Self-focussing: This intensity-dependent phase-delay can be very similar to that imposed deliberately on weak light by the use of a lens -- the thicker central portion of a (positive) lens has a greater optical path length and retards the phase of light passing there, relative to light passing through the lens-edge, and by this brings light to a focus. Focussing like this therefore occurs even in flat plates of glass, as the intensity distribution of the intense light causes lateral differences in the index of refraction, and therefore similar differences in the optical path length. In a high-power solid-state laser system this can cause the whole beam, or parts of it, to collapse to a focus even within the laser rod itself, with catastrophic results.

Filamentation: Instead of the whole beam collapsing, it is possible that a beam which is not perfectly smooth in its intensity profile will break up into beamlets and each of those may self-focus. Between whole-beam self-focussing and filamentation, the difference is only the relative growth rates of different spatial-frequency components of the beam, and the initial amplitude of those components.
Self-phase modulation: Equally, in time there can be similar phase distortions. By virtue of the changing intensity of a pulse of light, it becomes possible for the phase-fronts in the weaker leading edge of the pulse to 'run away' from phase fronts in the more intense part of the pulse. Because of intensity-dependent differences in the index of refraction, it therefore becomes possible to add to the period between phase fronts, thereby reducing the frequency of the lightwave. Similarly, on the trailing edge of the pulse, phase fronts of weaker light may tend to 'catch up' to the intense parts of the pulse, with the effect of decreasing the optical period and increasing the frequency.
In this, new frequencies of light are created in the pulse, and the pulse acquires a component of changing frequency -- it becomes optically chirped.

Nonlinear optical damage: Aside from the subtler nonlinear effects caused by propagation of high-intensity light, the maximum power or fluence of a laser system may be limited by the small absorption of laser energy by components of the laser system. This leads to a damage threshold for laser mirrors and other components which is seen ultimately to depend in interesting ways on the wavelength and the duration of the laser pulse.

[General References]

The solution:

In typical high-power short-pulse laser systems, it is the peak intensity, not the energy or the fluence, which causes pulse distortion or laser damage. The chirped pulse amplification laser therefore dissects a laser pulse according to its frequency components, and reorders it into a time-stretched lower-peak-intensity pulse of the same energy. This benign pulse can then be amplified safely to high energy, and then only afterwards reconstituted as a very short pulse of enormous peak power -- a pulse which could never itself have passed safely through the laser system.

This optical sleight-of-hand is accomplished using matched pairs of diffraction gratings. A laser pulse comprises many frequencies, according to its Fourier transform. In striking a diffraction grating, the pulse spreads out in angle, each frequncy component leaving at a slightly different dispersed angle. In travelling between two such gratings, different frequencies take different paths, and the total distances of the different paths through this diffraction-grating system are not the same. Thus it is that at the output of a grating expander the frequency components arrive at staggered times, ordered by their frequencies. The result is a temporally stretched-out pulse of steadily rising frequency -- a positively chirped pulse.

In proportion to the time-dilation of the pulse, the peak intensity of the pulse drops. Made more tractable in this way, the pulse can be amplified to substantial energies without encountering intensity-related problems. When amplification is complete, a complementary arrangement of diffraction gratings can almost exactly compensate, recompressing the pulse to its original prototype duration, but with now with orders of magnitude greater peak power than would have been possible without the CPA technique.