The Duke storage ring driven FEL scientific program is reviewed. The lattice design for Duke storage ring is described.

The newly commissioned Vanderbilt Free Electron Laser Center for Biomedical and Materials Research is a multidisciplinary users facility intended as an international resource. It provides extremely intense, continuously tunable, pulsed radiation in the mid-infrared (2-10 j.tm). Projects already underway include the linear and nonlinear interaction of laser radiation with optical materials, semiconductors, and mammalian tissue, the spectroscopy of species adsorbed on surfaces, measurement of vibrational energy transfer in DNA and RNA, the dynamics of proteins in cell membranes, the biomodulation of wound healing by lasers, image-guided stereotactic neurosurgery, and the use of monochromatic X-rays in medical imaging and therapy. The purpose of this article is to introduce the machine to the user community and to describe some of the new experimental opportunities that it makes possible. Details of several research projects are presented.

Laser emission is observed at a wavelength near 225 microns from a microtron-based free electron laser. The microtron accelerator produces macropulses with charge of 1 (mu) C at energies of 19 - 20 MeV and a 30 Hz repetition rate. The 16 microsecond(s) ec macropulse consists of 3 GHz micropulses approximately 6 mm long. The electron beam is steered and focused into a 10 m, 20 cm period, 750 helical undulator whose axis is collinear with that of a 15 m optical cavity formed by two 10 m radius of curvature copper mirrors. The far-infrared radiation is coupled through a 6 mm diameter hole located 5 mm off axis in one mirror and steered 10 m from the free electron laser and detected with a stressed Ge:Ga detector. A positive round trip laser gain is inferred from the temporal profile of the signal. The FIR temporal profile varies periodically with tuning of the rf micropulse frequency with a period of 25 kHz. This evidence of micropulse radiation interference and evidence of threshold behavior indicate optical cavity oscillation in the FIR. The intra-cavity power is estimated to be about 100 Watts. Characterization of the AT&T Bell Laboratories free electron laser is in progress and present status is reported.

The Stanford Picosecond FEL Center will provide intense picosecond optical pulses at wavelengths tunable from 1.5 micrometers to 15 micrometers . The ability to deliver picosecond pulses at wavelengths throughout the vibrational IR makes the Center's superconducting linac-driven FEL a powerful tool for studies of ultra-fast phenomena in molecular systems. In addition, the planned installation of a synchronized picosecond visible laser system will permit a variety of two-color experiments. Special efforts are being made to increase the cost-effectiveness of the FEL facility through the development of standardized optical setups.

The Los Alamos free-electron laser (FEL) facility has been modified by the replacement of the thermionic electron gun and bunchers with a 1300 MHz rf photoinjector. Two more accelerator tanks have been added to increase the beam energy to 40 MeV. Preliminary studies at 15 MeV have demonstrated excellent beam quality with a normalized emittance of 40 (pi) mm-mrad. The beam quality is now sufficient to allow harmonic lasing in the visible. We have recently begun lasing at a wavelength near 3 micrometers .

The results of several types of time-resolved experiments on rf-linac driven free-electron lasers (FELs) using streak-camera techniques are presented. In the past these techniques generally traded off time resolution, time span, and timing jitter to address either submicropulse or submacropulse effects. More recently, we have taken advantage of synchroscan streak cameras that were phase-locked to the reference 108.3 MHz rf signal combined with an orthogonal slow ramp deflection. One can then obtain submicropulse, submacropulse, and phase information during a single 100-microsecond(s) long macropulse. Samples of results include electron beam bunch lengths, cavity length tuning, phase slew/jitter, drive- laser phase stability, and visible FEL output temporal effects. Several of these demonstrations are the first of their kind on a FEL system (to our knowledge).

A short review of recent developments in laser photocathode is presented. The importance of different processes, such as the single and multiphoton photoelectric effects are analyzed on the basis of recent experimental results. The conditions for the formation of small divergence, short pulse duration, and high current density electron beam from metallic photocathodes are discussed. Other types of photocathodes can be used as the semiconductor (GaAs, chalcopyrite) in order to produce polarized electrons. The main steps and the diagnostics for applications in photoinjector development is discussed.

Two new far-infrared free-electron lasers (FELs) are operating in a newly constructed beam switchyard and are providing radiation to an adjoining users' facility. These have been constructed to the highest standards, resulting in substantially improved operating characteristics. For example, greatly improved electron beam recirculation now permits FEL operation to more closely match theoretical prediction and has extended operating pulse width and repetition rate. The design characteristics and operating properties of these lasers are discussed.

The design and construction of a second-generation free-electron laser (FEL) system at Los Alamos is described. Comprising state-of-the-art components, this FEL system will be sufficiently compact, robust, and user-friendly for application in industry, medicine, and research.

The intrinsic errors of actual magnetic undulators yield a degraded performance of the free electron laser (FEL) with respect to that obtained from ideal models of the magnetic undulator. The impact of these errors have been theoretically and computationally investigated for simple error models, in which the field errors have uniform or sinusoidal spatial extent and amplitudes which are statistically independent for each magnet pole piece. These simple models have been recently extended to include more complicated spatial structures and statistically correlated field errors in the analysis of the FEL performance. The measured data of the National Institute of Standards and Technology (NIST) undulator will be fit by the more advanced models. We present numerical simulation of the FEL in the presence of measured field errors.

Even at the conceptual level, the strong coupling between the laser subsystem elements, such as the accelerator, wiggler, optics, and control, greatly complicates the understanding and design of a free-electron laser (FEL). Given the requirements for a high-performance FEL, the coupling between the laser subsystems must be included in the design approach. To address the subsystem coupling, we implemented the concept of an integrated numerical experiment (INEX). Unique features of the INEX approach are consistency and numerical equivalence of experimental diagnostics. The equivalent numerical diagnostics mitigate the major problem of misinterpretation that often occurs when theoretical and experimental data are compared. A complete INEX model has been applied to the 10-micrometers high-extraction-efficiency experiment at Los Alamos and the 0.6-micrometers Burst Mode experiment at Boeing Aerospace. In addition, the agreement between INEX and the experiments is very good. With the INEX approach, it now appears possible to design high-performance FELs for numerous applications. The first full-scale test of the INEX approach is the Los Alamos High-Brightness Accelerator FEL (HIBAF) experiment. Implementation and experimental validation of the INEX concept are discussed.

Characteristics of the fundamental and harmonic emission from free-electron lasers (FELs) is examined in the spontaneous emission regime. The radiation at both odd and even harmonic frequencies is treated for electron beams with finite emittance and energy spread. For wigglers with many wiggle periods, calculation of the SE by integrating an ensemble of electrons along their exact trajectories becomes exceedingly cumbersome. Therefore, a different technique is used in which the far-field radiation pattern of a single electron is manipulated in transform space to include the effects of emittance. The effects of energy spread can be included by a weighted sum over the energy distribution. The program execution time for wigglers of arbitrary length is negligible. The transverse radiation patterns including the transverse frequency dependences, are given. How this radiation is modeled in FEL simulation codes is discussed.

We investigate the advantages of a device named the HARP/FEL (for Harmonic Amplifier/Free-Electron Laser), which may be described as a two-element, optical klystron FEL with the prebuncher stage oscillating at a frequency different from the output-stage frequency. In analysis based on the single particle treatment of Harvey and Palmer (where the one-dimensional, free-space theory is examined), if the prebuncher-wiggler period ((lambda) w1) differs from the output-coupler-wiggler period ((lambda) w2), then the gain and saturated power of the output coupler are at a strong maximum when ((lambda) w1/(lambda) w2) is an integer. Physically, this synchronism condition arises when the ratio of the bunching wavenumbers is also an integer, a conditions that ensures that both FEL modes are resonant and coherently coupled via the electron-beam bunching. The gain- enhancement mechanism is precipitated by injecting electron bunches into the output coupler with a period that is a subharmonic of the output coupler's ponderomotive potential. If the bunches are sufficiently localized, then each one will be confined to a single potential well and efficient energy coupling occurs between the electrons and the fields. Through integration of the FEL equations of motion, we have analyzed how the HARP's saturated power, saturation length, and susceptibility to e-beam energy spread compare to a free-electron laser and an optical klystron when operated at the same frequency with the same e-beam. Experimental evidence for the HARP mechanism will be published in a separate paper.

A free-electron laser enclosed in a waveguide of small transverse dimension may offer the opportunity to develop a compact source of coherent radiation in the far-infrared (100 (mu) - 100 (mu) ) region of the spectrum. The presence of the waveguide significantly affects the operation of the laser, qualitatively altering its spontaneous emission spectrum, its tuning characteristics, and its gain curve. We discuss the effects of the waveguide, including the opportunities it offers in terms of tapering, tuning, and slip reduction, and the limitations it imposes. We discuss the implications for compact free-electron lasers.

Stimulated emission of synchrotron radiation is possible when an electromagnetic wave polarized in the orbit plane propagates across a static magnetic field along which relativistic electrons gyrate. In principle, optical gain can be observed up to gyration harmonic numbers n of the order of n_cr equals 3γ³, where γ is the relativistic energy factor. For n > n_cr the gain falls off as exp(-2n/n_cr). A universal gain formula is presented, valid for all values of n/n_cr. It is shown that synchrotron radiation laser (SRL) gain is large than FEL gain when a similar electron beam is considered. For millimeter wavelength SRLs at 100 and 300 GHz, conceptual designs are presented which use a 2 MeV sub-Ampere electron beam with a magnetic field as low as 6 kG. For shorter wavelength SRLs, including visible wavelength devices, beam cooling may be required to obtain the necessary low energy spread. A means for achieving the cooling is described. A conceptual design for a 633 nm SRL is presented which would use a 10 MeV, 0.2 Ampere beam in a 100 kG uniform magnetic field.

Stimulated backscattered harmonic radiation generated by the interaction of an intense pump laser field with an electron beam or plasma is analyzed using a nonlinear, relativistic, fluid theory valid to all orders in the pump laser amplitude. The backscattered radiation occurs at odd harmonics of the doppler shifted incident laser frequency. The growth rate and saturation level of the backscattered harmonics are calculated and thermal limitations are discussed. This mechanism may provide a practical method for producing coherent radiation in the XUV regime.

Los Alamos has designed and proposes to establish an XUV-IR free-electron laser (FEL) user facility for scientific research and industrial applications based on coherent radiation ranging from soft x rays as short as 1 nm to far-infrared wavelengths as long as 100 micrometers . As the next-generation light source beyond low-emittance storage rings with undulator insertion devices, this proposed national FEL user facility should make available to researchers broadly tunable, picosecond-pulse, coherent radiation with 10⁴ to 10⁷ greater spectral flux and brightness. The facility design is based on two series of FEL oscillators including one regenerative amplifier. The primary series of seven FEL oscillators, driven by a single, 1-GeV rf linac, spans the short-wavelength range from 1 to 600 nm. A second 60-MeV rf linac, synchronized with the first, drives a series of three Vis/IR FEL oscillators to cover the 0.5 to 100-micrometers range. This paper presents the motivation for such a facility arising from its inherently high power per unit bandwidth and its potential use for an array of scientific and industrial applications, describes the facility design, output parameters, and user laboratories, makes comparisons with synchrotron radiation sources, and summarizes recent technical progress that supports the technical feasibility.

Three new types of free electron lasers (FELs) that are being examined in new ranges of parameter design space are: compact systems, XUV systems, and high power devices. Shorter wiggler wavelengths, shorter or longer lasers, higher currents, and higher quality electron beams are a few of the trends in the FEL community. The primary predictor of FEL oscillation is the small signal gain. We present 3-D small-signal calculations for more realistic parabolic-profile electron beams in the limit of moderately wide to wide electron beams. This limit complements the thin electron beam limit and needs be included in any analytical approximation that encompasses all 3-D effects. The system of equations for the optical modes are of Hamiltonian form and are solved as the analytical eigenmodes of the stationary quantum mechanical harmonic oscillator. We show the complete solution to the initial value problem in the special case of a cold, resonant electron beam, including the damped modes heretofore neglected. From this we derive the asymptotic solution as a superposition of Hermite [square symmetry] or Laguerre (circular symmetry) modes. We give expressions for the mode size, the spatial growth rate, and injection fraction for the dominant mode.

Abstract not available.

We describe the Ion-Ripple Laser as an advanced scheme for generating coherent radiation. A relativistic electron beam obligately propagating through an ion ripple excites electromagnetic radiation which is coupled to slow electrostatic waves with peak growth rate at the resonance frequency ω≈ 2λ²_Ok_wc via backward Raman scattering. This new scheme may provide novel tunable sources of coherent high-power radiation. By proper choice of device parameters, sources of microwaves, optical, and perhaps even x rays can be made. By employing fluid theory the dispersion relation for wave coupling is derived and used to calculate the radiation frequency and linear growth rate. The nonlinear saturation mechanism is due to trapping of the beam electrons by the ponderomotive potential. For an energetic electron beam, the peak growth rate is ω_i = ω^5/2_peε_isinθ/λ_O^3/4(2k_ωc)^3/2 and the efficiency is η= ω_pe/(2k_ωcλ_O^3/2). A 1 2/2 D-PIC simulation code was developed to verify the ideas, scaling laws, and nonlinear mechanism. From the observed power spectrum, backward Raman scattering is show to be responsible for the radiation. The growth rates and efficiencies given by the simulation match the ones of theory for different wiggler wave lengths and beam λ. Both of them show a slow decrease with momentum spread. Momentum spread also broadens the radiation spectrum.

Synchrotron radiation of 33 keV ((lambda) approximately equals 0.4 angstrom) has been tentatively employed for coronary angiography studies; the results were successful, but lack clinical quality. Synchrotron radiation has been proposed as a source for microlithography and the efficient production of integrated circuits, using energies of 0.5 - 2.5 keV ((lambda) approximately equals 5 - 20 angstroms). This application, although not yet carried out in earnest, appeared promising enough so that a number of synchrotron sources dedicated to this purpose are now being set up worldwide. Such applications require storage rings for GeV electrons, costing $20 - 60 million each. Less expensive, lower energy linacs, affordable by individual hospitals or smaller institutions, can produce kev channeling radiation suitable for microlithography and angiography. Channeling-radiation intensities, especially for x-ray energies of some tens of keV and higher will surpass synchrotron-radiation intensities. We have carried out quantitative studies confirming the above conclusions regarding the comparison of channeling radiation and synchrotron radiation. We found that < 5 MeV electron linacs costing less than $1 million can generate few keV channeling radiation intensities comparable to that of synchrotron radiation, suitable for microlithography and calcium-based angiography, as well as x-ray diffraction structure analysis and elemental analysis by x-ray fluorescence. (For iodine-based angiography, 20 MeV linacs are required.) If transition radiation were used for the same purposes, more expensive linacs of an order of magnitude higher electron energy would be needed.

In experiments using targets consisting of many thin metal foils, we have demonstrated that a narrow, forward-directed cone of transition radiation in the 0.8 to 60 keV spectral range can be generated by electron beams with moderate energies (between 17 and 500 MeV). We have measured the spectral and spatial photon densities of these radiators using low current electron beams. Using high currents, we have measured the total power of these radiators. A total power of 15.2 mW was measured from a beryllium radiator, 8.1 mW from the aluminum radiator, 4.8 mW from the titanium radiator, and 3.6 mW from the copper radiator. These values matched calculated predictions within experimental error. Both x-ray lithography for the production of integrated circuits and Laue diffraction for the study of biological materials are possible applications of this radiation. In particular, using the Al and Be radiators we have exposed photoresist-coated silicon wafers. Exposure times of the bare resist were as short as 120 s for 5 cm² of wafer area (this resist had a 230 mj/cm² energy dose per unit area). The shortest time for mask/wafer exposure was 180 seconds for 5 cm². Using an Intel mask, we obtained lithographs with features of 0.5 micrometers . In addition, we have calculated that the quasi-monochromatic bandwidth of transition radiation would be feasible for Laue diffraction of biological crystals. By designing transition radiators to emit x rays at the foil material's K-, L-, or M-shell photo-absorption edge, the x-ray spectrum is narrowed. The sources is then quasi-monochromatic and uses an electron beam whose energy is considerably lower than that needed for synchrotron sources.

We describe our recent experimental study of the production of x rays by an electron beam interacting with a crystal lattice, i.e., parametric x-ray (PX) generation. In this radiation process, the virtual photon field associated with a relativistic electron traveling in a crystal is diffracted by the crystal lattice in the same way that real x rays are diffracted by crystals. The radiation produced satisfies the Bragg conditions associated with the diffraction of the virtual photons which are nearly parallel to the velocity of the electrons. This phenomenon is associated with a more general class of radiation production mechanisms which include transition radiation (TR), diffraction radiation (DR), and Smith-Purcell radiation. In each case, radiation is produced when the particle's fields are altered by interacting with a material whose dielectric constant varies along or near the particle's trajectory. The usual acceleration mechanism for the production of radiation is not involved in these phenomena. In the case of a crystal, the periodic electric susceptibility interacting with the particle's field produces parametric x rays. We will also present a theoretical overview of this phenomenon which can be used to generate monochromatic, linearly polarized, directional x rays. Accelerators with energies ranging from a few MeV to hundreds of MeV may be used as drivers for novel parametric x-ray generators for various applications requiring the unique properties of these sources.

Gas puff Z-pinches have been operated successfully with peak currents of 100 kA to more than 10 MA. Very little effort has been developed to study long life rep-rated systems. The scaling laws of the gas puff Z-pinch show that up to a few MA (peak pinch current), the neon k-line yield is proportional to I⁴. In these machines the peak current is proportional to the square root of the energy stored in the system. As a result the x-ray yield is proportional to the square of the stored energy. A 1 pulse/sec with about one Megampere current can be realized with a capacitor-driven gas puff Z-pinch storing 20 to 40 kJ. Such a system will deliver about one kW Ne k line radiation (about 10 A) and a few hundred Watts Kr 1 line radiation. At this power level this machine becomes very attractive for x-ray applications such as x-ray lithography, x-ray microscopy, and spectroscopy. Reliability, components life time, debris, and state of the art are discussed.

We studied laser-produced plasma as an x-ray source for x-ray microscopy. Using water- window x rays, contact x-ray images of living sea urchin sperm were taken by a 500 picosecond x-ray pulse. The resist relief was examined by atomic force microscope and informations characteristic of x-ray microscopy were obtained. The finest feature noticed in the x-ray image was 0.1 micrometers .

Atomic number and laser intensity dependences of laser plasma soft-x-ray (SXR) emissions are presented, and the nonuniformities of line-shaped plasma SXR emission are studied with time- and spectrum- resolving. The SXR emissions of the half-cylindrical-shell target are investigated aiming at producing a more available plasma condition for x-ray lasering.

Recent work on the X-pinch discharge has demonstrated that the resulting plasma formed at the junction of the crossed wires is an efficient x-ray source. The reproducible spatial localization of the bright source renders the X-pinch ideal for x-ray microscopy and lithography. Here we report on an X-pinch discharge driven by a small conventional capacitor bank. Observation shows that the plasma formed is qualitatively similar to that generated by higher voltage pulsed power machines although a more localized x-ray emitting region is observed. Strong asymmetry in the plasma is also present along the anode cathode axis. In the soft x-ray region, a long tail is seen on the anode side of the crossing point whereas a shorter emitting plasma is present on the cathode side. Experiments were conducted to investigate the formation mechanisms for the evolution of this asymmetry. Detailed comparison of the plasma evolution and the resulting x-ray radiation is presented.

The explosion of a single wire with a high current pulse has long been established as an excellent source of x-ray radiation. However, experimental analysis of single wire explosions shows that the radiation is emitted from an inhomogeneous plasma consisting of high-density high-energy emitting points located randomly in a low density background column. The X- pinch, in which two or more crossed wires are exploded by a large current pulse driven by a low inductance source, is a good source of soft x-ray radiation in the energy range of 100 eV to 10 keV. The geometry of the crossed wire load ensures that the plasma emitting the bulk of the harder radiation is reproducibly located at the crossing point of the wires. This renders the source ideal for applications such as x-ray microscopy and lithography. Previous work on X- pinches has been carried out using elaborate high-voltage pulsed power generators delivering a brief high current pulse. Results are presented here of a first comprehensive study of X-pinch discharges driven by a low energy (4 kJ), low voltage (30 kV) 9 (mu) F capacitor bank. X- ray emission from X-pinches made from a variety of wires of different sizes and materials is characterized and compared with x-ray emission from single wire discharges. Radiation output is measured for different spectral regimes and information is presented concerning both the spatial and temporal emission of the radiation. Conditions for optimal soft x-ray yield on such low power machines are established.

In our earlier paper, we reported on the high-temperature capillary arc discharge (HTACD) operating as a source of soft x-ray radiation whose spectra were studied in the range to 15 nm. Recently, we have extended the range of spectroscopic study to 3 nm and found that when the discharge chamber is filled with air at pressure of 300 Pa practically, only the resonance line 4.027 nm of ion CV emerges from the source due to the longer wavelength emission cut-of in the air and the source turns out to be nearly monochromatic.