We compute the localization properties of light in a random system of linear dielectric slabs and a random system of Kerr-like nonlinear dielectric slabs. Particular attention is paid to systems that are periodic on average, and studies are made of the localization of light in these systems at frequencies both within and outside the band gaps of the average periodic system. It is found that the average periodic properties have the greatest effect on the exponential localization of light in the linear system, and exhibit only a small transient influence on the power law localization observed in the nonlinear system. The average periodic properties of the nonlinear system are observed to affect localization only in systems of finite extent.

We study the strong localization of light in a two-dimensional, randomly disordered, dielectric medium that is periodic on average. The medium consists of an infinite array of infinitely long, parallel, dielectric rods of square cross section embedded in vacuum. The intersections of the axes of these rods with a perpendicular plane form a square lattice. The dielectric constant of each rod, which is lossless, is an independent random variable whose values are distributed about an average value (epsilon) _(alpha ). The photonic band structure of the average periodic system has been calculated, and the edges of the band gaps have been found by a careful study of the convergence of the calculation. Maxwell's equations are integrated numerically in space and time to yield the electromagnetic field radiated by a line source emitting light at frequency (omega) whose current amplitude has a Gaussian shape in time. By looking in the disordered system at the electromagnetic energy stored between consecutive pairs of cylinders of radii m(Delta) R and (m + 1)(Delta) R, where m equals 0,1,2..., centered on the line source we show that the system displays strong localization of light. The localization length is found to be much smaller for frequencies in the gaps of the photonic band structure of the average periodic medium than for frequencies outside the gaps.

The transfer matrix method has been used for the calculation of transmission and reflection properties of periodic and/or disordered dispersive photonic band gap (PBG) materials. We have studied the transmission properties of: (1) PBG materials constructed of low resistive Si wafers forming the newly proposed layer-by-layer structure and exhibiting PBG at around 100 GHz; (2) a two-dimensional square lattice consisting of metallic wires; (3) materials having structural gaps close to a polariton gap.

The photonic band structure of 2D photonic crystals consisting of cylindrical dielectric rods on a square and on a triangular lattice is computed by a method which is commonly used in studying low-energy electron diffraction in conventional crystals. The matrices involved in the present method are substantially smaller than those required for the plane-wave method. Unlike the multiple-scattering method, a root searching procedure is not needed. In addition, with this method, the attenuation lengths for frequencies within the forbidden bands are computed automatically with the band structure calculation. Reflection and transmission characteristics of crystals of finite thicknesses are also computed.

The effects of an externally applied magnetic field on the photonic band structure of a periodic medium composed of vacuum and metallic or semiconducting materials are studied. Specifically, a waveguide formed from perfectly conducting parallel plates which contain between them a periodic array of vacuum and n-type semiconductor (or metal) slabs is examined in the presence of a uniform external static magnetic field applied parallel to both the plates and slab surfaces. The photonic band structure for electromagnetic waves propagating in a direction perpendicular to the slab faces is found to be sensitive to the applied magnetic field. Two effects are observed in the presence of a nonzero applied magnetic field: (1) a gap opens in the frequency region 0 < (omega) /c < (pi) /d, where d is the plate separation, and (2) a nonreciprocal structure is generated which yields a band structure that is asymmetric in wavevector space.

A new method of calculating reflection and transmission of electromagnetic waves through a 2D or 3D periodic dielectric structure has been developed. This method employs the coupled- wave approach, similar to the one used for 1D periodic structures, and is most suitable for rectangular lattices with a rectangular dielectric center. Different orders of reflection and transmission have thus been found within the photonic band gaps as well as outside the gaps.

This paper reports recent progress in FEL field in China. The experimental results of Raman- type and Compton-type, as well as electromagnetic wave pumped FEL and Cherenkov FEL, are given. The modeling and simulation works on both accelerator and laser are reviewed.

A key requirement for the development of commercial fusion power plants utilizing inertial confinement fusion (ICF) as a source of thermonuclear power is the availability of reliable, efficient laser drivers. These laser drivers must be capable of delivering ultra-violet (UV) optical pulses having energies of the order of 5 MJ to cryogenic deuterium-tritium (D/T) ICF targets. Excimer lasers are a leading candidate to fill these demanding ICF driver requirements. However, since excimer lasers are not storage lasers, the excimer laser pulse duration is determined primarily by the length of the excitation pulse delivered to the excimer laser amplifier. Pulsed power associated with efficiently generating excimer laser pulses has a time constant that falls in the range, 30 (tau) _p < (tau) _pp < 100 (tau) _p. As a consequence, pulse compression is needed to convert the long excimer laser pulses to pulses of duration (tau) _p. These main ICF driver pulses require longer, lower power precursor pulses delivered to the ICF target before the arrival of the main pulse. Computer simulations have shown that a `chirped,' self-seeded, stimulated Brillouin scattering (SBS) pulse compressor cell using SF<sub>6</sub> at a density, (rho) approximately 1 amagat can efficiently compress krypton fluoride laser pulses at (lambda) equals 248 nm. In order to avoid the generation of output pulses substantially shorter than (tau) <sub>p</sub>, the optical power in the chirped input SBS `seed' beams was ramped. Compressed pulse conversion efficiencies of up to 68% were calculated for output pulse durations of (tau) _p approximately 6 ns. Techniques for generating a variety of temporally complex output pulse shapes are discussed.

Electronically excited oxygen, O₂(¹(Delta) ), is the power source that drives chemical oxygen iodine lasers. A new type of singlet oxygen generator has been both analytically and experimentally characterized. This new generator uses uniformly sized droplets of basic hydrogen peroxide (BHP) as the liquid phase reaction surface. The gaseous chlorine and O₂(¹(Delta) ) flow between a multitude of droplets in the generator. The fraction of chlorine that is converted into O₂(¹(Delta) ) depends on the coupled gas and liquid flow fields, diffusive mass transport, and homogeneous gas, liquid, and interface chemical reactions. A 1D flow/chemistry code has been developed and used to investigate the effects of the choice of parameters associated with the two-phase chemistry and transport on the chlorine utilization, O₂(¹(Delta) ) yield, and the efficiency of converting chlorine into O₂(¹(Delta) ). Predicted sensitivities to flow conditions, and the chemistry of the reactive media are presented. In particular, modeling results that identify dominant physical processes, and appropriate mathematical models are discussed. Analysis and a review of available information on the chlorine reaction with BHP in O₂(¹(Delta) ) generators indicates that the performance of the UDOG should be dominated by the gas-liquid interfacial and internal liquid chemical reactions and diffusion processes. A major part of the modeling effort has been to investigate this assumption and question it. A surface reaction model provides much better agreement with measurements made in a uniform droplet O₂(¹(Delta) ) generator.

A mathematical model for the production of singlet delta oxygen, O₂(¹(Delta) ) from the reaction of a gas containing chlorine (Cl₂) with the hydroperoxy ion (HO₂⁺) in liquid basic hydrogen peroxide (BHP) is reviewed. A solution for the generator's Cl₂ utilization, O₂(¹(Delta) ) yield and efficiency is obtained, which is applicable for both long and short gas exposure times over a wide range of Cl₂ pressures and HO₂⁺ molarities. It is shown that generator performance is characterized by six parameters. The dependence of these parameters upon generator geometry, Cl₂ pressure, diluent ratio, HO₂⁺ molarity, the gas and liquid flow velocities, and the gas and material properties is discussed. The yield is similar in form to the well-known transport solution for the fraction of O₂(¹(Delta) ) exiting a duct except that an effective gas residence time replaces the gas transit time. When the surface concentration of HO₂⁺ is constant, i.e., in the well-stirred limit, this solution reduces to that obtained previously and, in this case, generator performance can be characterized by a set of universal performance curves dependent upon only three parameters. The model is used to examine the behavior of a rotating disk O₂(¹(Delta) ) generator at high pressure and Cl₂ flow rates.

A new rear cone decentration model has been implemented in the cylindrical resonator optical quality (CROQ) code, which simulates the performance of the ALPHA annular resonator. The new model avoids the traditional way of using equivalent phase tilts to simulate the decentration effect of optical elements in an optical train. Instead, it uses the approach of local field shift. As a result, the new model requires fewer Fourier components to represent the azimuthal variation of fields and thus largely extends the range of rear cone decentration that can be modeled by the CROQ code. The required memory space and CPU time for each decentration case are also significantly reduced.

CO overtone laser pulses with energy up to 4.4 J and efficiency up to 1% with completely suppressed lasing in the fundamental band were obtained from mixtures of CO:N₂ excited by an e-beam sustained discharge with a duration of 1.5 microsecond(s) . A comprehensive model of this laser was formulated. The results of numerical simulations satisfactorily correlate with experiments. The conditions for obtaining efficiency as high as 20% are predicted.

An unstable resonator with a semitransparent output coupler is feasible for lasers with moderate gain and a large cross section of active medium. The resonator fundamental mode can be obtained up to Fresnel numbers of 10 and more. Output beam quality does not differ much from the Gaussian beam, but at the same time intensity distribution is rather flat and mode-medium coupling is better. This approach does not require the mirrors with tapered reflectivity profile. In practice the design of this type of resonator often requires the developer to place a spherical mirror inside to fold the optical path and that inevitably causes astigmatism. We describe the results of the investigation of resonator sensitivity to intracavity astigmatisms. The requirements for the resonator setup to obtain nearly unperturbed fundamental mode operation and a convenient resonator design to meet these requirements are discussed.

Lasers which oscillate within inhomogeneously broadened gain media exhibit spectral hole burning and concomitant reduction in output power compared with equivalent homogeneously broadened laser gain media. By increasing the cavity length, it may be possible to demonstrate at least a partial transition from an inhomogeneous laser cavity mode spectrum to a homogeneous spectrum. In neutral xenon lasers the inhomogeneous spectral broadening mechanism arises from Doppler shifts of individual atoms in thermal motion within the electric discharge comprising the laser gain medium. Optical transitions corresponding to these noble gas atoms have natural linewidths. Simulations of the output power characteristics of the xenon laser were carried out as a function of laser cavity parameters, including the cavity length, L. These calculations showed that when the intracavity mode spacing frequency, c/2L < (Delta) (nu) _n, the inhomogeneously broadened xenon mode spectrum converted to a homogeneously broadened oscillation spectrum with an increase in output power. These simulations are compared with experimental results obtained for the long laser oscillation characteristics of the (5d[5/2]₂ degree(s) yields 6p[3/2]₁) transition corresponding to the strong, high-gain 3.508 (mu) line in xenon.

A novel scheme to generate a subnanosecond pulse in VUV spectral region is proposed, by applying the dual-wavelength pumping method in conjunction with the polarization switching technique to the stimulated rotational Raman scattering process. The optimum conditions for the short pulse generation are presented through the numerical calculations. The proposed process is applicable to a seed pulse generation for subnanosecond pulse amplification by the pump laser in the VUV region.

We present the results of a simulation program able to describe the dynamics of a Nd:MgO₂:LiNbO₃ P.E. waveguide laser. A special feature of this simulator is the ability to take into account the spatial hole burning effect on the population inversion in the transverse directions and the time modulation of the pump absorption due to the population inversion dynamics. Modal pump and signal waveguide fields and population inversion are expanded using Gauss-Hermite functions in the lateral direction and generalized Laguerre functions in depth. Only one function is used to represent the pump and the lasing fields and a maximum of eight functions in lateral and eight in depth direction to represent the population inversion. The need of this accurate modeling is shown by comparison with the results of a simplified model.

A numerical method for the analysis of the fast axial flow glow discharge C02 laser has been developed. The method is based on the self-consistent solution to the one-dimensional steady-state glow discharge equations, the gas dynamic equations and the vibrational relaxation equations. The discharge equations include the continuity ones for the electrons, the positive and negative ions and Poisson's equation for the electric field. The three-mode relaxation model for the vibrational kinetics and the plane-parallel optical resonator model have been used. This approach does not require previous assignment of the discharge power distribution and enables obtaining the discharge structure included the near-electrode regions in addition to the laser characteristics.

There is proposed a simple design of an unstable laser resonator with a continuous rectangular aperture of an output beam. A mathematical model was developed to simulate the discharge laser cavity and it was used to study numerically the performance of a fast-flow technological C02 laser with such type of resonator which, as it has been shown, possesses good perspectives.

The paper describes the design of the antireflection coating and the high reflectance stack for variable reflectance mirrors. A deposition model which takes into account the collisions of the evaporated molecules was studied in order to obtain the radial thickness profile. For a certain multilayer system, deposed with a fixed mask, a map relating the parameters of the reflectance profile to the deposition geometry was calculated. The effect produced on this map by the deposition using a rotating mask with shaped holes was also investigated.

Efficient lasing on multiple-selected lines has been recently demonstrated experimentally on a cw hf chemical laser. Multiple-selected-line operation is required to both enhance atmospheric transmission and to ensure high-power extraction efficiency on multiple vibrational hf levels. Seventy-five percent of the multiline power was measured on two selected hf lines using a confocal, unstable cavity with a high-efficiency (97%) intercavity diffraction grating. This measured fraction of the multiline power is consistent with theoretical calculations, which include the effect of rotational nonequilibrium. The two-selected-line hf unstable cavity was not prone to parasitic oscillations. A novel multiple-selected-line integral master oscillator power amplifier (IMOPA) concept was also evaluated. Line selection on two hf lines was demonstrated with the IMOPA, although the hole diameter had to be made sufficiently large to prevent parasitic oscillations within the amplifier. It was concluded from our experiments and theoretical calculations that, although the IMOPA concept was demonstrated at relatively low power (400 W), parasitics may be a problem at much higher values of the single-pass gain.