Several nonlinear optical processes concurrently occur in single micrometer-sized droplets that are irradiated by the second-harmonic output of a single-mode Q-switched Nd:YAG laser. The authors review the following observations of stimulated Brillouin scattering (SBS) and stimulated Raman scattering (SRS) in droplets: (1) the image of the droplets that are emitting SBS and SRS; (2) the long delay and decay time of the SRS pulse; (3) the correlated temporal profiles of SBS and SRS pulses; (4) the high-resolution spectra of SBS; (5) the frequency splitting of a degenerate spherical normal mode in the SRS spectra; and (6) the fine structures in the angular distribution of SRS.

Computed results for some linear optical problems relevant to nonlinear optics in droplets are presented and discussed. (1) The electric energy density distributions inside homogeneous spheres (water droplets) illuminated with plane waves are computed using Lorentz-Mie theory and geometrical optics ray tracing. (2) The electric energy densities inside spheres illuminated by Gaussian beams are computed using an angular spectrum of plane waves approach, a technique applicable to scattering of Gaussian beams by axisymmetric objects. (3) The Q's and resonance locations of spherically symmetric, radially-inhomogeneous spheres are computed numerically. (4) The Q's and resonance locations of perturbed, homogeneous, droplets are computed using the T-matrix method. (5) The Q's and resonance locations of inhomogeneous, spherically-asymmetric droplets are computed using the T-matrix method.

This paper summarizes a recent theoretical study of stimulated Brillouin scattering, emphasizing the physical interpretation of the acoustic normal modes and the SBS gain formula.

Nonlinear optical processes in spherical microdroplets whose studies have been pioneered by Chang and co-workers are intriguing from the point of view of applications because of the extremely low thresholds they exhibit for the generation of stimulated output, but also from the theoretical point of view, because they require a synthesis between the methods of nonlinear optics and Mie scattering theory. The basis for such a synthesis has been given in the semiclassical theory of Kurizki and Nitzan, which implies that the nonlinearly amplified component of the scattered field is spatially orthogonal ('out of phase') with the linearly scattered field component (the ordinary Mie solution). This approach allows the calculation of amplification coefficients in various nonlinear processes, as a function of the Mie resonant denominators and spatial overlap of the input and output spherical waves. Recently, this approach has been extended to quantized solutions for parametric amplification and oscillation via four-wave mixing in microdroplets. The very low thresholds for oscillation at Mie resonances are predicted to correspond to strong squeezing, i.e., suppression of the photocurrent noise associated with the detection of the output waves below the noise level associated with an ideal laser. Even well below threshold, the spatially orthogonal output waves allow for strong squeezing at periodically recurring distances from the droplet.

Systematic theoretical calculations are presented investigating (a) the external excitation of resonances within a spherical droplet using a focused laser beam, (b) the internal excitation of resonances within a spherical droplet due to an embedded source, and (c) the effect of droplet nonsphericity on resonance excitation.

Significant cavity QED enhancement have been observed of both spontaneous and stimulated emission in 10- to 14-micrometers diameter liquid microdroplets containing fluorophors. Direct emission lifetime reduction was observed as well as greatly enhanced lasing gain. Use of droplets provides an excellent test of cavity QED theory for spherical geometries. The mode number and order of the spherical cavity resonances responsible for the observed emission peaks were identified and their Q values calculated from Lorenz-Mie theory. An experimental procedure, based on equating stimulated gain to cavity losses, is used to estimate the magnitude of the enhancement. Spherical cavity Qs of only 2600 were found to lead to cavity enhancements of over 100 in the emission cross-section of Rhodamine 6G in agreement with theory.

Intensity ratios of the resonant Na((lambda) equals 589 nm) to H(alpha ) ((lambda) equals 656 nm) emission lines have been measured for 10 GW/cm2 KrF laser radiation ((lambda) = 248 nm, tpulse equals 17 ns) interacting with 50 micrometers diameter saltwater droplets as a function of position in the breakdown region parallel to the incident laser beam. Results indicate the H(alpha ) emission intensity is highly non-uniform through the plasma plume and is highest in the plume ejected back toward the laser beam. In contrast, the Na emission line intensity shows a much weaker spatial dependence as a function of position within the plasma plume. Calibration data are also reported for the ratio of Na to H(alpha ) emission intensities as a function of Na concentration (0-2000 ppm) for approximately equals 20 micrometers diameter monodisperse droplets located at a fixed observation point in the plasma plume. Future research will extend the ratioing technique to remotely measure droplet salt concentrations in polydisperse saltwater sprays.

Stimulated Raman spectroscopy (SRS) can simultaneously measure droplet sizes and the associated component concentrations in a fuel injection. As spray evaporation is crucial in determining the performance parameters of a diesel engine, such as cold start and particulate emission formation, the new application of the method for spatially and temporally resolved measurements is a useful new diagnostic, extending our understanding of spray processes. Droplet sizes can be obtained from single shot SRS spectra by measuring the separation between morphology-dependent resonances (MDR) that correspond to standing wave modes confined near the droplet circumference. Power spectrum analysis allows the measurement of more than one droplet from a spectrum using a pumped laser sheet in the fuel spray. The MDRs are responsible for the simultaneous stimulation of multiple Raman spectral lines over and above those seen in bulk liquids. The SRS method for concentration measurement is effectively self-calibrating in that the relative intensity of two adjacent lines is used to measure concentration. Any particular fuel has a unique ratio of SRS antisymmetric to symmetric C-H stretch intensity. If individual components in a fuel blend are characterized beforehand, one can monitor the evolution of the spray during injection by measuring signal intensity ratios which yield the volume fraction of the component of interest. The SRS technique is being used to examine a number of spray dynamics phenomena such as fuel atomization, droplet evolution and front-end volatility effects, which are of current interest in diesel development studies.

Time-resolved Raman spectroscopy of a levitated droplet is demonstrated as the droplet absorbs vapor from the surrounding atmosphere. As the concentration of the absorbed species inside the droplet increases, its contribution to the Raman spectrum increases. Acquisition time for each spectrum is about five seconds, allowing the study of reactions within individual aerosol droplets.

The authors describe the results of recent droplet superheating experiments following excitation by pulsed lasers. Under low-irradiance excitation by either a short-pulse (0.4 microsecond(s) ec) or a long pulse (10 microsecond(s) ec) CO2 laser, two superheating thresholds may be identified, the deformation and disintegration thresholds. For higher-irradiance excitation by a Nd:YAG laser (pulse length approximately 10 nsec), the authors report observation of an evaporative instability in laser-irradiated microdroplets. The threshold behavior of these phenomena is discussed and a qualitative description of the effects is given.

Two different techniques for modeling the effects of diffraction in a homogeneously broadened, single-longitudinal mode laser are intercompared. Under comparable model conditions, the authors discuss similarities in the field profile and dynamics; however, in the region of small transverse mode spacing where mode-mode competition is relevant, the two models diverge. Both models display regular and irregular pulsing, and hysteresis when the cavity frequency is resonant with the atomic frequency.

The reconciliation of theoretical predictions with experimental results suggests that a more accurate description of the phenomenon is made by progressively adding what could at first be considered terms of secondary importance into the theoretical exposition of the problem. One such "secondary" physical effect which has enjoyed much renewed interest within the last decade, in the field of non—linear optics, is diffraction. Its study is important because diffraction is a natural consequence of wave propagation. Although the initiating mechanisms are well known (i.e., finite beam width to wavelength ratio), the consequences are often not as well understood. In order to simplify the analysis diffraction effects are often neglected

By using the hydrodynamic symmetric description, the quasi-static model of self-focusing processes has been extended to include the transient effect of nonlinear medium response. This paper presents a new operator method to study the transient nonlinear refractive index and two-photon absorption processes in which the laser pulse-width is assumed to be much shorter than or comparable to the response time of the nonlinear medium. The numerical computation uses a finite-difference implementation of a characteristic method. The asymmetrical focusing structure in full spatial and temporal coordinates are obtained. The pulse-width self- compression and the peak shifting of the wave field envelope, corresponding to the asymmetrical and anomalous broadening of the spectrum, are verified in the transient case. The influence on the self-focusing process of various parameters, including the saturation effect, is discussed.

This paper presents the results of a numerical calculation of laser pulse propagation in a vapor of three-level atoms. The calculation is semi-classical in that the macroscopic electric fields are determined classically using Maxwell's equations, while the microscopic polarization is determined quantum mechanically using Schrodinger's equation. The pump laser field is tuned nearly resonant with the first atomic transition of the three-level folded-ladder system. It has previously been shown that in a three-level vapor in which the dipole coupling of the second transition is sufficiently smaller than the dipole coupling of the first transition, the dynamics of the laser-matter interaction are dominated by self-action effects of the pump pulse. This calculation shows dramatically different propagation effects when the second transition dipole moment is greater than the first transition. The larger dipole moment allows the final state to compete with the ground state for population that is pumped into the intermediate state by the pump pulse. This enhances the Stokes gain and increases the Raman efficiency while hindering the processes that produce self-induced transparency and other self-action effects. The periodic fluctuations of the intermediate state population leads to the parametric breakup of the electric fields.

Using a semiclassical, effective-state approach to laser pulse propagation in a dilute gas of polyatomic molecules, a numerical model of multiphoton excitation which includes quasi- continuum transitions is used to calculate the effects of coherent propagation on multi-photon absorption. The initial laser field is assumed to be a quasimonochromatic Gaussian envelope which is nearly resonant with a portion of the lower vibrational states of the molecule. The molecular model consists of a single ground state and several bands of discrete states which are successively connected via electric dipole transitions. This model correctly mimics some of the major characteristics of multiphoton absorption, such as power threshold, fluence dependent absorption, and complete population inversion at high fluences. The model also allows for the calculation of the polarization induced in the medium by the electric field, and the subsequent influence that the polarization has on the field. As the pulse propagates, the Rabi oscillations induced in the lower level states become encoded onto the pulse, creating a series of peaks and troughs similar to the behavior of pulse propagation in two-level systems. Another coherent propagation effect which can change the excitation process is coherent sideband generation. The temporal distortions and the generation of new frequencies results in changes in the population distribution among the states, as well as changes in the total population transferred to the quasi-continuum. The pulse reshaping occurs while the total fluence of the pulse is decreasing due to energy loss to the quasi-continuum. The effective increase in total population transfer due to propagation effects is offset by the decrease in fluence. The resulting transfer of energy from the pulse to the molecule depends strongly on the initial characteristic of the pulse. These processes play an important role in any system where the efficient transfer of laser energy to the medium is required, such as laser-induced chemistry and laser isotope separation.

Intrapulse stimulated Raman scattering (ISRS) is an important physical phenomenon responsible for the self-induced frequency shift of solitons in optical fibers. ISRS is generally treated as a perturbation term in the nonlinear Schrodinger equation, an approximation that breaks down for ultrashort optical pulses. The authors obtain the solitary-wave solutions of the nonlinear Schrodinger equation by including both ISRS and self-steepening. These solutions do not correspond to optical pulses but represent optical fronts or optical shocks. The properties of ISRS-supported optical shocks are discussed in detail.

Using the Hertz vectors, it is possible to reduce the Maxwell field equations inside index- guided optical devices to two couple scalar partial differential equations. With the proper selection of gauge functions, the mixed partial derivatives with respect to the propagation direction and either transverse direction can be eliminated. This results in a form similar to the equations used for the scalar beam propagation method. However, space-varying functions multiplied by partial derivative operators rule out convenient use of the same Fourier transform methods commonly used for the scalar case. A numerical method for solving time independent propagation of coherent light through guided nonlinear media including vector coupling using these equations is presented.

A numerical simulation is described that accurately reproduces experimentally observed statistical properties of spontaneous Raman soliton generation. Comparisons of the soliton yield and the distributions of soliton delay times and peak intensities indicate a close agreement between theory and experiment.

The interaction of a dipole radiator with a phase-conjugate mirror through a quantized radiation field leads to dramatic enhancement of the nonlinearity if the radiator is a slightly anharmonic oscillator. Quantum noise from the conjugator saturates the linear and nonlinear response of a two-level radiator when the reflectivity exceeds unity.

Results on squeezed states of the electromagnetic field are reported. The authors study a two- mode squeezed state in which the modes of a cavity are degenerate in frequency, but possess orthogonal polarizations. The model includes stochastic coupling to the vacuum fluctuations outside the cavity and the strong correlation between the modes is generated by a two-photon transition within the medium filling the cavity.

The influence of the atomic lifetimes on the photon statistics of the micromaser is discussed. The generation of trapping states and number states requires very long lifetime atoms as well as low temperature.

The higher-order terms in the expansion of ln(1 + pu) with respect to the pumping regularity p are considered. The effects of higher-order terms, due to regular pumping, on the photon statistics are found to be significant in maser cases, and can be neglected in laser cases. The right order approximation must be used to ensure correct results.

The influence of the atomic lifetime and Raman field strength on the gain and laser threshold in noninverted atomic systems is studied. Results show that to achieve useful gain in such a system, it is necessary to use a sufficiently intense Raman field, atoms with a long lifetime for the lower active levels, and a small cavity loss.

The degenerate parametric oscillator (DPO) is analyzed with the nonlinear material treated quantum mechanically. The source of squeezing in the DPO is the same as in the two-photon correlated-spontaneous-emission laser (CEL)--the level coherence. In this sense, they are similar. The amount of squeezing in DPO is the same as in the two-photon CEL at two-photon resonance, but the gain in DPO is usually much smaller than that in the two-photon CEL at two-photon resonance.

A nondegenerate three-level atomic system is considered, and the possibility of lasering without population inversion via numerical computations is investigated. The atomic system consists of one upper level and two closely spaced lower levels, which are coherently mixed by a microwave field interaction. A numerical calculation is performed, based on a variable step formulation of the classic Adams formulae for solving first-order differential equations. This method of solution has been outlined by C. W. Gear and by L. F. Shampine and M. K. Gordon. The authors adapt this algorithm for use with complex double precision arrays and apply the resulting program to the Maxwell-Schrodinger equations of this atomic system. This enables the authors to carry out a strong field analysis of the atom-field interaction, since they are not restricting their analysis to perturbation theory.

The dynamical behavior of a Raman ring laser is examined under conditions of pump depletion. In the plane-wave and adiabatic limits, a mapping is obtained for two complex field amplitudes. Chaos is observed when injected pump and Stokes frequencies are detuned from a Raman resonance. Hyperchaos, characterized by two positive Liapunov exponents, is found when in addition the two injected fields are of comparable magnitudes.

The authors propose simpler coupled laser arrays employing class-A lasers, such as dye and gas lasers, in which both polarization and population dynamics can be adiabatically eliminated. Chaotic behavior is investigated on the basis of numerical analysis, paying special attention to the interplay between coexisting attractors.

A quantum amplifier model is used to simulate semiconductor laser operation in the presence of quantum fluctuations. The model explicitly takes into account of the finite dipole dephasing time of the lasing medium and the longitudinal non-uniformity of carrier and photon distributions. Quantum fluctuations are considered by regarding each of the spontaneous, stimulated emission and absorption processes as random processes obeying Poisson statistics. The simulation results agrees with conventional traveling wave rate equation approach in the absence of noise. However, with the presence of quantum fluctuations, the laser output experiences pulsation at the period of cavity round-trip time. The pulsation behavior is heavily dependent on the finite dipole dephasing time and the nonlinear gain constant of the lasing medium. By appropriately adjusting the mode-lock condition of a high speed laser diode, ultrashort pulse train can be obtained from the laser diode with external cavity repetition frequency.

Ultrafast gain recovery dynamics of a diode laser are explored when a picosecond pulse propagates through a semiconductor amplifying medium. Theoretical modelling employs a set of nonlinear dynamical equations to predict the temporal dependence of the population difference and the diode laser output intensity in the wake of an applied signal pulse. Experimentally, this system of study is a cw-pumped 670 nm InGaAlP diode laser probed by a tunable train of 10 psec pulses emitted from a synchronously pumped mode-locked pyridine 1 dye laser. In addition to discussing the details of the gain, the authors also report the phase locking of the diode laser to the applied signal over a broad range of wavelengths.

Acceleration of a relativistic electron beam by the longitudinal field component of a focused laser beam is optimal when the laser beam is radially polarized. A variational calculation shows that the electron energy gain in a laser beam of given power cannot greatly exceed that available from a Gaussian beam. Particle trajectories calculated in a radially polarized Gaussian beam indicate that transverse forces focus and recollimate appropriately phased off- axis electrons over the acceleration length.

It is shown that high energy gains are possible in a gas-loaded laser linac using a nondiffracting beam. The authors derive expression for optimum gas index as a function of particle injection energy and other parameters, including the case of longitudinal grading, for both Gaussian and non diffracting beams. The properties of pulsed laser beams are discussed.

The rate at which one atom in a photodetector is ionized by a beam of classical light is considered. A detailed examination of the probability of transition is presented, using first- order time-dependent perturbation theory and the rotating wave approximation. This method usually leads to the well-known rule called 'golden' by Fermi. The authors have made a more accurate analysis based on numerical quadrature techniques. They discover a curious oscillation in the probability for photoionization for short times and a temporal shift in the curve as compared to the results given by the golden rule. The shift is approximated analytically and found to correspond with the numerical result to a high precision. Results for the probability of photoionization are shown without using the rotating wave approximation.

The optical properties of a new class of composite nonlinear materials composed of coated grains, such as cadmium sulfide with a silver coating, are examined. These materials exhibit intrinsic optical bistability and resonantly enhanced conjugate reflectivity. The threshold for intrinsic optical bistability is low enough for practical applications in optical communications and optical computing. Some problems associated with the fabrication of these materials are addressed. Based on preliminary results, switching times are expected to be in the subpicosecond range.

Composite nonlinear optical materials consisting of silver nanoparticles, approximately 11 nm in diameter, in a polymer host were prepared. The polymer used in the preparation of the composites was a polyurethane bearing a tricyanovinyl moiety. The composite preparation method and polymer pretreatment method as well as pretreatment of the silver colloids to control particle agglomeration are described. The (chi) (3)xxxx of a composite was measured by degenerate four-wave mixing as 3.65 X 10-10 (esu), 7.91 X 10-11 (esu) and 5.86 X 10-11 (esu) at 532 nm, 562 nm, and 570 nm, respectively. The interfacial chemical interactions between the silver colloid and polymer host in the composites were explored using Raman, infrared, UV- visible, and NMR spectroscopy.

Mixed metal and metal-coated microparticles have been fabricated and their nonlinear optical properties examined. Degenerate four-wave mixing experiments of the various particles have not shown any enhancement in the phase conjugate signal except for certain thinly coated gold microparticles. These particles had an enhancement up to a factor of ten but their phase conjugate signal was found to be dependent on four photons instead of the usual three. All of the other microparticles showed a normal three photon dependent phase conjugate signal.

Suspensions of microparticles in either a liquid or solid host medium are known to possess very large optical nonlinearities (large (chi) 3), while also exhibiting the unusual phenomenon of enhanced backscattering. The authors have experimentally studied the relationship between these two effects using wavelength scale latex spheres in aqueous solutions. The results are presented along with a possible explanation.

A dielectric host medium with spherical high-Tc superconducting particles embedded is considered. The nonlinear electromagnetic field response of the composite superconductor material is calculated using a time-dependent Landau Ginzburg model. Specifically, the authors predict a strong frequency dependence of the nonlinear susceptibility, field intensity dependence of the transition temperature, and third harmonic generation. The mechanism for the nonlinearity is the breaking of quasi-particle bonds by the electromagnetic field.

This paper describes the results of preliminary measurements of the fluorescence properties of a laser dye in a silica gel matrix, produced through solgel processing techniques. These measurements enable observation of the effects of photodegradation of the dye by intense pulsed laser radiation with intensities characteristics of dye laser pumping. Time decay and spectral properties of the fluorescence from dye-doped gels were measured before and after photodegradation, yielding information about the effects of the gel matrix on the dye, its photodegradation, and recovery of the dye fluorescence with time.

In atoms and molecules, nonresonant second-order and third-order optical processes mediated by virtual electronic excitations are described by their second-order (beta) ijk(-(omega) 3;(omega) 1,(omega) 2) and third-order (gamma) ijkl(-(omega) 4;(omega) 1,(omega) 2,(omega) 3) optical susceptibilities, respectively. In general, the real population of the initial state for the virtual electronic excitations can be either the usual singlet ground state S0 or an optically pumped excited state Sn. The authors have found for quasi-one (1-D) and quasi-two (2-D) dimensional chain-like and disc-like structures, compared to the ground state, the nonresonant excited state (beta) ijkSn(-(omega) 3;(omega) 1,(omega) 2) or (gamma) ijklSn(-(omega) 4;(omega) 1,(omega) 2,(omega) 3) can markedly increase, and even change sign, when the first (S1) or second (S2) (pi) -electron excited state is optically pumped and populated for timescales sufficiently long to allow nonresonant measurements of (beta) ijkSn(-(omega) 3;(omega) 1,(omega) 2) or (gamma) ijklSn(-(omega) 4;(omega) 1,(omega) 2,(omega) 3).

Several electroactive polymers, such as polyacetylene, polythiophene, poly [p-phenylene vinylene] and poly [2,5-thienylene vinylene] have shown promise as NLO-active materials over the past few years. However, as several theoretical and experimental research groups have pointed out in recent publications and symposia, it is not evident that long conjugation lengths are necessary for enhanced (chi) (3) activity. As recently demonstrated, copolyamides which incorporate polyenylic or PTV oligomeric repeat units show (chi) (3)/(alpha) values of ca. 10-13 esu-cm at 532 nm(band-edge). In this paper, the authors discuss how ladder subunits related to the electroactive polymers POL and PTL can be incorporated into polymer films as (a) copolymer repeat units, (b) pendant groups attached to poly [p-hydroxystyrene] and (c) guest-host composites in polycarbonate. Sharp optical absorptions are found in all cases as well as promising (chi) (3) properties.

Head-to-tail mainchain chromophoric polymers have been of interest because of their high chromophore densities [about 30 X 1021 chromophores/cm3 in the case of poly(4-N-ethylene-N-ethylamino)-(alpha) -cyanocinnamate], their proposed tendency to align in an electric field more readily than unconnected chromophores, and the proposed enhancement of the macroscopic hyperpolarizability coefficients. These proposed properties have yet to be proven for the solid state. In fact, results on mainchain nonlinear optical polymers indicate that a much lower degree of alignment resulted than would be expected for unconnected dipoles. It is possible that the head-to-tail configuration of mainchain polymers has too high of an energy barrier to rotation and alignment to allow proper alignment of the dipoles. Therefore, a new class of mainchain polymers, namely head-to-head polymers connected with various flexible spacers (which may allow the dipoles to align) was developed. The synthesis and characterization of these new materials is described. The nonlinear optical properties of several of this new class of mainchain nonlinear optical polymers were compared to the properties of the head-to-tail mainchain polymers.

The synthesis and characterization of a new class of dipolar inorganic coordination polymers, as well as the examination of the third-order NLO properties of several organometallic acetylide complexes, are discussed. The dipolar materials were prepared for Cr, Mn, and Fe. The crystal structure of two Mn complexes demonstrate that coordination polymers are being formed in the solid-state. Solution third harmonic generation (1907 yields 636 nm) was used to determine the molecular hyperpolarizibilities ((gamma) ) for several organometallic halide and acetylide complexes. The values of (gamma) range from <3 X 10-36 to 50 X 10-36 esu.

Chaos iii lasers has a significant conceptual meaning, l)ecause fuiidarneiital models of lasers, which are (leriVed from well- recognized semiclassical laser equations (i. e. . Maxwell-Bloch equations ), possess inherent instal)ility leading to chaos. rF1ese models provide promising prototypes for investigating complex dynamical behaviors in strong connection with experimental demonstrations. These include Casperson instai)ility in inhomogeiieously l)roadened single-mode lasers l)ased on mode splittiiig' leading to chaos featuring period(1oul)Ii1g, breakup of a 2-torus and intermittency2 and Lorenz chaos in homogeneously broadened singlemode lasers featuring self-sustained Rabi precessions.3 However. these types of laser chaos have been restricted to the rather iionpractical far-infrared lasers (class (1 lasers). Titus, the question arises as to whether such chaotic behaviors take place in more practical laser devices, such as C02, soli(l-state, and semiconductor laser diodes (LDs). In such lasers, polarization decay tiiiie is much shorter thait other time scales involving lasing. As a result, polarization dynamics are adiabatically eliminated in Maxwe!l-Bloch (MB) laser equations. This implies that MB instability can never been expected iii ordinary situations 1)ecause of the lack of degrees of freedom. This essential difficulty can l)C removed of COIIF5C l)V intro(lucing other degrees (:)f freedoms in the form of external modulation, light injection. introduction of saturable absorber, external feedback and so on.4 Among these lasers, which are ofteii categorized as class B lasers, LDs have an important inherent characteristic which cannot be expected in other lasers. That is an anomalous dispersion effect at the lasing frequency and the free-carrier plasma effect. These effects result in carrier-density dependent refractive ill(lex, which is expressed by the so-called o-parameter. A resultant large (3) nonlinearity gives rise to a variety of nonlinear phenomena in LDs in addition to the lasing action itself. This paper reviews various phenomeiia reported so far. paying special attention to experimental demonstrations and their physical interpretations.

This paper reviews the importance of nonlinear gain and its impact on the performance of semiconductor lasers. The physical mechanisms which can lead to an intensity dependence of the optical gain in the above-threshold regime are described briefly. A specific nonlinear-gain mechanism, referred to as intraband gain saturation, is discussed in detail by considering its effect on the important laser characteristics such as the modulation bandwidth, intensity noise, and the laser linewidth. Particular attention is paid to the effects of cross saturation in nearly single-mode semiconductor lasers. Even a weak side mode can lead to saturation and rebroadening of the main-mode linewidth due to mode coupling induced by the nonlinear gain.

Multiple quantum well structures of several materials such as GaAs/GaAlAs and InGaAs/InP have interesting nonlinear optical properties for devices. One example is the adjustment of absorption edge as well as the refractive index by control of the well width or application of electric field. The theoretical understandings of these properties in terms of the intrinsic material constants and externally controlled parameters are to be discussed. We also review s