3.2 Q-SWITCHING The power output of many lasers can be greatly increased by a process known as Q-switching. For lasers which are normally pulsed, a single high power pulse of shorter duration can be produced while continuously operating lasers can be made to give a train of pulses. For many applications such as drilling, welding high speed photography and optical radar, high power outputs, or short pulse duration times, are needed. The distinction between power and energy must be appreciated. Q-switched lasers always give lower energy outputs compared with when they are not Q-switched because of absorption and other sources of loss in the Q-switch itself. Nevertheless the power output is very much higher because the pulse duration is shorter. Equation (3.5) gives the elementary relation between power output, energy in the pulse and pulse duration. Pulse energy Gules) itput (watts) = —————__ Power output (watts) = 3° duration (seconds) G5) In the case of the ruby laser, for example, the pulse energy is typically 10 joules delivered in one millisecond. Q-switching reduces the pulse duration to 10 nanoseconds and so increases the power output from one kilowatt to one gigawatt (10° watts). Q-switching as its name implies, involves changing the Q of the laser cavity so that the feedback by the mirrors is suppressed and so depletion of the population of the upper laser energy level is not permitted until its population has reached a high value. The laser is Pumped with the resonator kept at a very low Q value; the Q of the cavity is then made to suddenly increase allowing gain from stimulated emission to take place, with the result that the laser energy escapes in a very short, highly intense pulse. For obvious reasons, this Process is also known as Q-spoiling of giant-pulse operation. Scanned with CamScannerQ-switching is carried out by placing a closed shutter, which may take various forms, in the cavity and so effectively eliminating the resonator from the laser medium. After the laser is pumped the shutter is very rapidly opened so restoring the cavity to the system, There are two important requirements to be met if Q-switching is to be possible. These are: (a) The rate of pumping must be faster than the spontaneous decay rate from the upper energy level. If the pumping is too slow, the upper level will empty faster than it can be filled and sufficient population inversion will not be achieved. (b) The Q-switch must switch rapidly in comparison with the build up in stimulated emission, otherwise the latter will be a gradual process and a longer pulse time than necessary will be obtained so reducing the power. In practice a Q-switch time of at least 10 ns is desirable. There are many ways of attaining Q-switch operation. The following methods work well and are to be found common use. Scanned with CamScanner3.2.2 The Electro-optic Q-switch ‘A very fast Q-switch can be made using the Kerr or Pockels effect described at the beginning of this chapter. If the output from the laser 1s not naturally polarized, a polarizer is placed in the cavity together with an electro-optic crystal (see Figure 3.6) maintained at such a voltage that the plane polarized light incident on the crystal is converted into circularly polarized light by the transmission through the crystal. The laser mirror reflects this circularly polarized light and in doing so reverses the direction of polarization. Hence on re-emerging from the electro- optic crystal, the light is again plane polarized but at 90° to its direction. It is therefore not transmitted by the polarizer. When the voltage across the plates is reduced to zero, usually done by shorting them out, the electro-optic crystal has no effect other than a small source of loss within the cavity. Polarizer f | I switched output Kerr cell y Fig. 3.6 Electro-optic Q-switch using the Kerr effect. The change in voltage which must obviously be synchronized with the pumping can be accomplished in less than 10 ns and so very effective Q-switching takes place. Voltages required for Kerr effect operation are usually higher than for the Pockels effect. In addition the undesirable properties of nitrobenzene and carbon disulphide, which are among the few materials having a sufficiently high Kerr constant, have led to the commercial clectro-optic Q-switches being based on the Pockels effect in ADP (ammonium dihydrogen Phosphate) or KDP (potassium dihydrogen phosphate). @ The electrodes on the faces of the crystal in longitudinal mode effect Pockels cell. Unitorr operating in the longitudinal mode are usually in the form of thick ae clean, Metallic layers would suffer damage due to the high power densities ical resistance of such thin films also prevents fast switching. Scanned with CamScanner33 MODULATION OF THE LASER OUTPUT Amplitude modulation of the laser output is important in communication applications and considerable efforts have been made to find methods of obtaining high frequency modulation. The following methods have proved successful although the electro-optic modulator appears to be the only type which will enable the wide bandwidth capabilities of optical communications systems to be realized. 3:31 Internal Modulation The easiest and most obvious way to modulate the output of a laser is to modulate the pump power. For helium-neon lasers the highest modulation frequencies are limited to about 100 kHz on account of the rise timer of the oscillator and the finite time required to obtain an excited neon atom by resonant energy transfer from an excited helium atom. The relation between pump power and output is non-linear, therefore for linear modulation the percentage modulation of the output cannot be very high. Semiconductor diode lasers on the other hand can be modulated at frequencies up to 10'° Hz by modulating the excited current. 3.3°2\ Mechanical Modulators External modulation can be achieved by feeding the output of a laser into a Michelsoor Twyman-Green interferometer. If the path length in one of the arms is varied, the output will be modulated accordingly. Path length variation is obtained by attaching one of the mirrors to a piezo-electric crystal which has the property of changing its size under the action of an applied electric field. Thus, modulation of the voltage across the crystal produces a corresponding modulation in the intensity of the output beam. Modulation frequencies of up to 5 MHz with a high percentage of modulation are possible. 33.3 Acoustic Modulation Under the action of a compressional acoustic wave some materials suffer a periodic change in refractive index. When the acoustic wavelength is of the order of the wavelength of light the material behaves as a diffraction grating so that an incident laser beam is split into a number of orders. In liquids this is known as the Debys-Sears effect and in solids as the Raman-Nath effect. There are basically two ways in which these effects can be exploited for modulation. If the laser beam is incident parallel to the plane of the acoustic fringes, plus one and minus one order diffraction beams can be obtalned, as indicated in Figure 3.7, together with a directly transmitted zero order beam. By appropriate placing of an aperture, these first order beams can be truncated by equal amounts. The second lens recombines the attenuated first order beams with the unaffected zero order beam to produce a plane wavefront of different intensity to that emerging from the acoustic grating. By varying the acoustic frequency, Scanned with CamScannerStop \, Gs Acoustic waves’ Vs Taser input Zero order IS PD rah yA Transducer Stop |<- Lens \ Attenuated beam Fig. 3.7 Acoustic modulation by attenuation of first order diffracted beams. the angle of diffraction can be changed and hence the degree of attenuation of the first order beams which in turn modulates the output. An alternative approach is to allow the laser beam to fall on the acoustic grating at an angle as shown in Figure 3.8. The angle is related to the acoustic (A,) and optical (A,) wavelength by the following equation: 2A, sin 8 Acoustic wavefront A Lo! Laser beam x frequency v Piezo-electric crystal 3.6) Directly transmitted beam of frequency v Bragg reflected beam of frequency v ~ Ay Fig. 3.8 Acoustic modulation using reflection at the Bragg angle. Bragg reflectionis said to occur in which the fringe planes act as mirrors and consequently a large percentage of the incident light is reflect Partially reflecting ted. As the acoustic grating is effectively moving perpendicular to the fringe direction, the Teflected wave is teduced in frequency by the frequency of the acoustic wave by virtue of the Doppler effect. The reflected and transmitted waves are then recombined to produce a carrier and a single Scanned with CamScannersideband. This single sideband operation is often preferred as it is more economical in bandwidth requirements. Water and fused silica respectively are often used in liquid and solid modulators while a piezo-electric transducer is used to generate the acoustic wave. Scanned with CamScanner3,35 Electro-optic Beam Deflection “Digit beam deflection in which the beam is directed to one of a large number of discrete points can be achieved by using a combination of electro-optic and birefringent crystals, Figure 3.9 indicates the basic principles of the process. Electro-optic crystals Laser input > | Alternative >| outputs > Birefringent crystals Fig. 3.9 A two-stage electro-optic beam deflection system. Figure 3.9 shows a two-stage deflector in which the laser beam can be deflected to one of four collinear positions. The first component can be a Pockels cell which can be switched Scanned with CamScannerbasic unit of the system. Many of these units are placed in tandem until the required number of final beam positions are obtained. Using n such units 2” positions can be addressed. The most important features of such a system are the resolution, the total number of positions and the switching time. The resolution or smallness of the deflected spot is greatly enhanced by using convergent light, which comes to a focus on the final address plane. However, there is a limit on this size because large angles of convergence resulting in very small diffraction spot sizes cannot be tolerated because of the large amount of unwanted background light produced. Further, this background light is also the limiting factor in the maximum number of digital positions which can be addressed. Deflectors capable of addressing the array of 256 x 256 positions have been built using 16 stages. Scanned with CamScanner