GB2189073A - Laser oscillator - Google Patents

Abstract

A laser oscillator comprises external feedback means, e.g. a mirror or multiple dielectric coatings, such that an increasing rate of change of phase gives rise to a resistance change which causes negative feedback, thereby stabilising the phase and frequency of the laser. As the light output increases the optical efficiency is reduced due to reduced electron-photon coupling.

Description

SPECIFICATION Laser oscillator The present invention relates to laser systems, and especially to a method of providing improved local oscillators using lasers.

Embodiments of the invention will now be described with reference to the accompanying drawings, in which: Figure 1 shows various graphs used to explain the invention. In these graphs, Fig. 1(a) represents how uncertainty in phase and number of electrons gives a circle of confusion for simultaneous measurements" with coherent states, Fig. 1(b) indicates how squeezed states can squeeze the uncertainty in one direction, e.g. squeezed in the number direction but enhanced in the phase direction to preserve the total "area of confusion" as required by quantum theory, and Fig. 1(c) is a squeezed state schematic for reduced uncertainty in phase but more uncertainty in amplitude.

Figure 2 relates to saturation of the laser: the tangent indicates the effective threshold for stimulated emission, and indicates how as the laser saturates so the threshold becomes effectively negative.

Figure 3 represents space change smoothing: electrons are emitted from the cathode with a Poisson distribution but the potential minimum increases if there is an excess emission and this decreses in potential minimum repels further electrons, so fewer electrons surmount that minimum. Thus the excess emission from the cathode is smoothed by the time the electrons get past the potential minimum.

Figure 4 shows characteristics of an asymmetric laser, single mode. is the near field pattern of optical density and p is the electron density. It will be seen that as the optical intensity increases there is spatial hole burning near the peak intensity which reduces the effective refractive index and causes the light to move further away from the central peak of electron density, thus decoupling the optical and electron distributions which reduces the overall production of photons. The laser light tends to shift over to the right.

The arrangement to be described gives a technique which gives a local oscillator with a photon stream output that is quieter in amplitude fluctuations or in phase fluctuations than in a conventional photon stream. It is believed that this is the first practical possibility of making squeezed photon states whereby although normally the uncertainty of number and phase is a "circle of confusion" given by An A the circle can in principle be an ellipse and the principal axis of the ellipse can be in any orientation dependent upon the details of the laser. See the explanatory graphs, Fig. 1(a), Fig. 1(b) and Fig. 1(c).

We start by observing that according to a simplified theory put forward by R. Loudon (The Quantum Theory of Light, Oxford University Press) the statistical output of a laser is expected to be Poissonion only if a (gas) laser is driven well above threshold. The theory then can be modified to apply to injection lasers and suggests that the output of the injection laser (in a fixed time T) wil have a variance in the photon count of (nmean+Pi)112 where Pt is the number of photons that could have been supplied by the threshold current if the laser had been 100% efficient. The mean photon count output is nmean. Low threshold lasers will then have outputs closer to the classical Poisson limit than high threshold lasers. This enables one to make low threshold single mode injection lasers.

However, this theory then strongly suggests that what one needs is a laser with a negative threshold so that the variance in the photon count is smaller than Poisson-a sub-poissonian stream. Fig. 2 indicates how such a negative threshold could arise given a dynamic bias to the laser. The laser saturates giving, effectively a negative threshold. Thus one looks further at the saturated laser noise.

We now discuss the meaning of a sub-poissonian stream of particles. This is well known in thermionic emission where electrons are emitted from the cathode as a Poisson stream, and if they are emitted with a sufficiently dense current, then Child's Law holds (Beck and Ahmed, Physical Electronics, Arnold) and a space charge minimum results close to the cathode, Fig. 3. If an extra electron is emitted above the mean then the minimum of the electric potential is depressed and there is negative feedback which prevents the electrons from being emitted from the cathode in the same numbers. The effect of the negative feedback is marked and is known as the space charge reduction factor r and r can approach around 0.01.The shot noise in the electron current is then reduced from in2=2e lo Af to in2= 2eel0 Af where f is the bandwidth of the system.

Thus the electron stream is sub-poissonian at least for characteristic timescales on which the charge can re-adjust itself. This space charge smoothing is well documented and easy to measure, has been measured in an attempt to see if feeding a quiet electron stream into a laser had any difference over a shot noise limited stream. The experiment was not sensitive enough to answer that par ticular question.

We now consider the use of such a negative feedback in a good single mode laser, in this case an Inverted Rib Waveguide (IRW) laser. It is extremely difficult to align the current injection stripe precisely above the inverted rib guide. If the laser is sufficiently misaligned then it is observed that the laser output power saturates. The mechanism is roughly as follows. As further photons are emitted, so these protons stimulate emission and deplete the electrons in the neighbourhood of the peak of the light. The change in the electron density changes the effective refractive index and the light peak is moved sideways into a less favourable position for stimulated emission, Fig. 4. The differential efficiency of the laser is reduced.There is no necessary additional absorptive noise added in this process, just that as a bunch of photons are produced by the current, there is a negative feedback mechanism which reduces the coupling efficiency between photon and electron stream and so reduces the intensity of the light. This seems wholly analogous to the space-charge reduction mechanism of thermionic current smoothing.

The mechanism has a finite bandwidth so that the smoothing can be expected to occur up to a limiting frequency created by the electron-photon resonances within the laser. With well designed lasers these resonances can be in the 10GHz range so that this is useful for active smoothing of the photon amplitude.

This principle can be extended to the control of phase fluctuations. We do not expect (from very fundamental reasons in quantum theory) to defeat the uncertainty principle and produce a light output that is both quieter in phase as well as in amplitude fluctuations than a classical coherent stream of photons. However, it is possible to squeeze the area of uncertainty in one direction or another, Fig. 1.

The scheme outlined above seemed to favour reducing amplitude fluctuations and if that scheme succeeded then one would expect the phase fluctuations to be increased. However, if there is a change in the electron density locally as a result of photons stimulating absorption or emission then there is also a change in the refractive index. A reduction in the electron density raises the refractive index and changes the ideal frequency at which the laser would oscillate. The link between amplitude and phase fluctuations appear to have its origins in the fundamental Kramers-Kronig relations.If the gain is changed, as it must be if an electron and hole recombine, then the imaginary part of the refractive index changes and this forces (through Kramers-Kronig relationships) a change on the real refractive index, which changes the rate of change of phase arid so causes a fluctuation in the phase. However, the external optical "circuit" of the active medium can also control the phase. In principle there is no reason why, over a strictly limited bandwidth, it is not possible to have the rate of change of reactance with frequency either decreasing or increasing.

The external circuit (e.g. a mirror or multiple dielectric coatings) then should be designed so that this reactance change caused by an increasing rate of change of phase, forces a negative feedback so that the frequency and hence phase is stabilised. If the phase is stabilised it follows from fundamental quantum theory that the amplitude will be destabilised.

Thus starting with the quiet amplitude laser it is envisaged that an appropriate phase sensitive feedback into the laser will decrease the phase noise at the expense of increasing the amplitude noise once more.

This idea is applicable to asymmetric IRW laser-based structures. The photon stream shifts sideways as the photons increase into a region of lower electron concentration which raises the refractive index slightly. The optical length of the laser increases and so the frequency of oscillation will decrease. Thus although the amplitude may stabilise, the rate of change of phase will change, and so the phase fluctuations will increase in line with basic quantum theory. A mirror for use with the laser is so designed so that as the frequency decreases, the amplitude feedback into the laser is increased, so that the laser keeps more readily to the initial frequency. The saturation mechanism of the amplitude is negated, but the phase stabilisation will be enhanced. In other words an external circuit on the laser can change amplitude stabilisation into frequency or phase stabilisation.

There may be better lasers than the IRW laser. The general principle is a system whereby the optical efficiency is reduced as the light output increases. A simple saturable absorption in the form of a non-linear loss is inadequate as loss is simply a noisy process.

The essence of this invention is that the mechanism is essentially a reactive mechanism, reducing the coupling between electrons and photons as more photons are emitted.

Claims (2)

1. A laser oscillator circuit, which includes an external feedback circuit such as a mirror or multiple dielectric coatings so designed that the resistance change caused by an increasing rate of change of phase in the laser forces a negative feedback so that the frequency and hence the phase of the laser is stabilised, the arrangement being such that the optical efficiency is reduced as the light output increases due to a reduction in the coupling between electrons and photons as more photons are emitted.

2. A laser oscillator circuit, substantially as described with reference to the accompanying drawings.