This paper describes the status of x-ray sources suitable for granular x-ray lithography systems (systems that require low initial investment). The key factors that determine the feasibility of point sources utilizing x-ray generation by heated plasmas (pinched gas, laser, x-pinch) will be described. In particular, the relationship between x-ray source power, required device overlay, and wafer throughput requirements for a production worthy system will be presented. In addition, relevant issues concerning the suitability of the different x-ray point source technologies for x-ray lithography applications will be discussed. Considerations for x-ray collimators will be presented in an appendix.

Projection lithography is driven to shorter wavelengths to meet the demand for smaller critical dimensions in advanced computer chips. This trend logically extends to the extreme ultraviolet (EUV) region, where reduction imaging can be achieved using all-reflective optics. The wavelength region of primary interest is from 11 nm to 14 nm, where multilayer reflective coatings have demonstrated reflectivity greater than 60%. The leading candidate for a practical, compact source of EUV radiation is a laser plasma source, which provides sufficient conversion efficiency, about 1%, in the relevant bandwidth. This paper discusses the concept of EUV lithography bassed on a laser plasma source and describes a laboratory EUV lithography tool currently being characterized.

We describe several experiments on Nova that use laser-produced plasmas to generate x-rays capable of backlighting dense, cold plasmas ((rho) approximately 1-3 gm/cm³, kT approximately 5-10 eV, and areal density (rho) l approximately 0.01-0.05 g/cm²). The x-rays used vary over a wide range of hv, from 80 eV (x-ray laser) to 9 keV. This allows probing of plasmas relevant to many hydrodynamic experiments. Typical diagnostics are 100 ps pinhole framing cameras for long pulse backlighter and a time-integrated CCD camera for a short pulse backlighter.

Since its invention, a major concern of those using a microscope has been to improve the resolution without the introduction of artifacts. While light microscopy carrier little risk of the introduction of artifacts, because the preparative techniques are often minimal, the resolution is somewhat limited. The advent of the electron microscope offered greatly improved resolution but since biological specimens require extensive preparation, the possiblity of causing structural damage to the specimen is also increased. The ideal technique for structural studies of biological specimens would enable hydrated material to be examined without any preparation and with a resolution equal to that of electron microscopy. Soft x-ray microscopy certainly enables living material to be examined and while the resolution does not equal that of electron microscopy it exceeds that attainable by light microscopy. This paper briefly reviews the limitations of light and electron microscopy for the biologist the various ways that soft x- rays might be used to image hydrated biological material. Consideration is given to the different sources that have been used for soft x-ray microscopy and the relative merits of laser-plasma sources are discussed.

The development of multilayer optics over the past 25 years has led to unique new instrumentation for plasma diagnostics, astronomy, and synchrotron radiation research. Accurate calibration and measurement techniques for the x-ray and extreme UV regions are required for this developemnt, and for the proper interpretation of the results acquired by the instrumentation. The developement of calibration and metrology techniques based on present- day plasma and synchrotron radiation sources is leading to a better understanding of multilayer performance and their limitations.

Plasmas produced from laser-irradiated gas puff xenon targets, created by pulsed injection of xenon with high-pressure solenoid valve, offer the possibility of realizing a debrisless x-ray point source for the x-ray lithography applications. In this paper we present results of the experimental investigations on the x-ray generation from a gas puff xenon target irradiated with nanosecond high-power laser pulses produced using two different laser facilities; a Nd:glass laser operating at 1.06 micrometers , which generated 10-15 J pulses in 1 ns FWHM, and Nd:glass slab laser, producing pulses of 10 ns duration with energy reaching 12 J for a 0.53 micrometers wavelength or 20 J for 1.05 micrometers . To study the x-ray emission different x-ray diagnostic methods have been used. X-ray spectra were registered using a flat CsAP crystal spectrograph with an x-ray film or a curved KAP crystal spectrograph with a convex curvature coupled to an x-ray CCD readout detector. X-ray images have been taken using pinhole cameras with an x-ray film or a CCD array. X-ray yield was measured with the use of semiconductor detectors (silicon photodiodes or diamond photoconductors).

A large volume long pulse excimer laser ((lambda) equals 308 nm) is used to generate a soft x- ray plasma source with long heating time at a power density in the order of 10¹² - 10¹⁴ W/cm². The characteristics of the plasma source for different laser pulse time evolution in the range 10-120 ns and for different target materials are investigated. In particular the most suitable source conditions (spectral energy distribution, time duration, etc.) for specific applications are analyzed.

We have configured a new type of target for laser plasma x-ray generation. This target consists of an in-vacuum flowing stream of liquid water droplets. We have successfully produced plasmas using this target, and have measured its extreme ultraviolet (EUV) emission spectrum. Bright lines from Li-like and He-like oxygen dominate in the plasma radiation in this region. Most importantly, not target debris related effects were observed for this type of target. A nearby Mo/Si multilayer EUV mirror suffered no reflectivity reduction at 13 nm after exposure to 10⁵ laser shots on target. This observation constitutes a major breakthrough in the utilization of laser plasma radiation for practical applications, in particular, for EUV projection lithography of advanced microelectronic circuits. The simplicity and versatility of a continuously-fed target with naturally smooth surface and no associated debris problems meshes strongly with the critical engineering required for envisioned production line EUV projection lithography installations. Additionally, through the use of water based solutions as targets, it should be possible to tailor the EUV emission spectrum to match the source requirements for other potential applications, such as the x-ray microscopy.

We describe the use of small liquid droplets as target for laser-plasma soft x-ray and EUV generation. Using a table-top laser, this < 15 micrometers diameter plasma source typically produces line radiation with approximately 1(DOT)10¹² photons/ster-line-pulse in the water-window wavelength range. The major adcvantages of the source is that it is practically debris free. We discuss applications of the source to soft x-ray microscopy and lithography.

High-reflectance multilayer mirrors and gratings have been developed and implemented in the x-ray and XUV regions. The imaging and spectroscopic instruments have high throughput and can be positioned at a large distance from the radiation source where damage from the plasma debris and the radiation flux does not occur.

We describe the use of microchannel plate (MCP) optics to focus soft x-rays produced by the Rutherford Appleton Laboratory (RAL) high brightness plasma source. In this source the x- ray emitting plasma is generated by a high repetition rate, picosecond pulsed excimer laser system. A low cost, low debris, high intensity soft x-ray 'beamline' results from the combination of this bright, point-like, plasma source with a planar square-pore MCP optic. Fluxes of 9 X 10⁹ photon/mm²/s at 33.7 angstrom wavelength and 4 X 10⁷ photon/mm²/s at 7.8 angstrom wavelength have been recorded at the MCP focus. This paper also describes the exploitation of the RAL source for the characterization of prototype MCP optics. Investigations into both planar MCPs with square pores and spherically 'slumped' MCPs with circular pores are reported. The small (10 micrometers diameter) size of the plasma source coupled with an ability to make absolute flux measurements greatly facilitates the calibration process.

Results on laser plasma EUV characteristics for various target materials and irradiation conditions are presented. Spectra of high-Z elements in the 12.5-15.4 nm range from plasmas generated with a high-power KrF laser at 2 X 10¹² W/cm² were measured. The highest conversion efficiency (CE) of 0.85% in 2% BW was found for Re near 13.6 nm, corresponding to a maximal EUV power of 550 mW in 2% BW at 50 Hz. The use of two successive laser pulses, investigated with 2.5 ns pulses of 0.53 micrometers radiation at (0.5-1) X 10¹³, resulted in an increase of the CE by a factor of 1.8-2.3 for the second pulse at specific delay values (6 and 12 ns for W). The total CE gain amounted to 1.4-2. The first demonstration of an alternative concept of a laser plasma raget for EUVL was performed, based on the usage of a centrufugal force for elimination of particluates. The principle of the approach is generation of laser plasmas at the edge of a fastly-rotating disc. The effect of re- direction of particulates was observed in experiments with a (phi) 50 mm Ta disc at 36.000 rpm at laser power densitites between 10⁹ and 10¹¹ W/cm².

This paper describes a high-intensity, high pulse-repetition-rate picosecond-pulse excimer laser system and plasma x-ray source, which generates up to 3 W of average x-ray power, into 2(pi) steradians, in a spectral band from 10-16 angstrom. The XeCl excimer laser system output, at 308 nm, consists of a train of 16 pulses, each approximately 45 ps in duration and spaced by 2 ns. The energy of each pulse in the train is approximately 25 mJ, and the pulse-train repetition rate is 60 Hz. Each pulse in the train is focused to a spot of < 10 micrometers diameter on a metal tape target, resulting in an intensity of 1 X 10¹⁵ W cm^-2. Spectral and spatial characteristics of the x-ray emission have been studied, and the laser energy to x-ray dose conversion efficiency has been measured in an experiment which simulates the x-ray lithography process. Lithographic efficiencies of 5.9% and 10.9% have been measured for copper and stainless steel targets, respectively.

Applications for laboratory soft-x-ray/VuV sources would benefit from the ability to collect a large energy bandwidth of radiation emanating from the very small source and redirect it into a well collimated beam without losing most of the incident radiation. Such optics would be beneficial to x-ray spectroscopy, x-ray lithography, diffractometry, and other applciations. We have been working to apply technology originally developed for astronomical x-ray telescopes to production of low cost replicated collimation optics for such x-ray/VuV instruments. Most of the steps in the production of these optics have previously accomplished with the larger astronomical optics but we want to reduce the size of these optics by at least an order of magnitude which introduces problems. In addition, very few copies of an x-ray telescope are made while we want to make hundreds of copies of our optics. This paper briefly discusses the design and fabrication of these small collimation optics and is a report on work in progress.

For monochromatic imaging applications the advantages of combining two bent crystals in one system are demonstrated in comparison to a single crystal. The investigation shows that considerable improvements in resolution and spectral selectivity can be achieved by successive reflections from two bent crystals. The x-ray imaging device can be designed to a compact optical device mounted with the detector to a single port of the experimental chamber. This type of arrangement is of particular interest to large laser facilities such as those at LLNL, ILE, and CEA where a high x-ray photon flux is available but the space available for diagnostics is restricted. A design for an experimental setup planned for imaging of indirect driven fusion experiments at Lauwrence Livermore National Laboratory will be discussed here as an example. In general, improvements of spatial resolution by a factor of about 4 and spectral selectivity by a factor of about 10 can be achieved.

Images of a 130fs laser produced Ta-plasma were obtained with a spherically curved Mica crystal in four narrow spectral bands simultaneously. These are to our knowledge the first published x-ray images of a fs plasma. Also Mica was used for the first time as the imaging crystal. The resolved plasma size in the experiment (50 micrometers (DOT) 90 micrometers ) agrees with ray-tracing calculations and corresponds to the actual source size in our current setup. The four spectral bands of imaging are due to the strong Mica reflectivity in higher Bragg diffraction orders. All contributing bands have an energy higher than 1keV. Several laser shots were accumulated to achieve the required flux on the detector. Source spectroscopy was conducted to verify the source emission and the plasma density during the interaction with the 130fs laser pulse. Based on our experimental data the luminosity of crystal imaging is compared to a pinhole camera.

General formulae for the applying the vertical dispersion principle in x-ray spectroscopy of multiply charged ions are summarized, the characteristics of the experimental schemes based on flat and bent crystals are discussed. The unique properties of the novel spectroscopic methods, i.e. their extremely high dispersion, high spectral and 1D spatial resolution, and good collection efficiency, make them very attractive for ultrahigh-resolution spectroscopy. The examples of successful use of the vertical dispersion modifications of the double-crystal and the Johann sepctrometer in diagnostics of several types of laser-generated plasma are presented.

Soft x-ray contact microscopy (SXCM), using a pulsed x-ray source, offers the possibility of imaging the ultrastructure of living biological systems at sub-50nm resolution. We have developed a pulsed plasma x-ray source for this application, generated by the large volume XeCl laser 'Hercules'. Various unstable optical resonator configurations were employed to achieve a high laser intensity to increase the conversion efficiency to 'water window' x-rays (280-530eV). Optimum plasma conditions for SXCM are discussed, including the effect of pulse duration on image resolution. Soft x-ray contact images of Chlamydomonas dysosmos (unicellular alga) and the cyanobacteria Leptolyngbya are shown. In addition, the potential of producing a 'movie film' of the development of x-ray images within the photoresist (acting as the recording medium) is discussed, following the resist development while viewing by atomic force microscopy.

X-ray images of the various live bacteria, such as Staphylococcus and Streptococcus, and micromolecule such as chromosomal DNA from Escherichis coli, and Lipopolysacchride from Burkholderia cepacia, are obtained with soft x-ray contact microscopy. A compact tabletop type glass laser system is used to produce x-rays from Al, Si, and Au targets. The PMMA photoresists are used to record x-ray images. An AFM (atomic force microscope) is used to reproduce the x-ray images from the developed photoresists. The performance of the 50nm spatial resolutions are achieved and images are able to be discussed on the biological view.

Soft x-ray contact microscopy successfully images hydrated biological material with resolution superior to light microscopy. Nanosecond laser pulses record the image before movement or radiation damage can occur, eliminating concerns of fixation-induced artifacts (cf electron microscopy). X-rays make the recording polymethyl methacrylate (PMMA) photoresist more soluble, thus a contour map of x-ray absorbency is produced in which relative heights, measured by atomic force microscopy, reflect specimen carbon density. Until now quantification of the carbon-density differences was impossible, neither has the minimum carbon density difference which is detectable been determined. Since biological specimens are composed of structures differing only marginally in carbon density the discrimination between carbon densitites is critical. Using SI₃N₄ windows coated with differing carbon thicknesses we have followed the rate of PMMA dissolution in order to produce calibration curves from which specimen carbon density can be determined. These experiments have also attempted to determine the minimum detectable carbon density difference. When using relatively thick (< 5micrometers ) specimens image interpretation can be difficult as spatially separated structures in the original specimen become superimposed in the x-ray image. To provide spatial resolution in three dimensions we are developing a soft x-ray stereo imaging system. Using two laser plasma x-ray sources and contoured photoresists we have obtained two simulanteous images of the same specimen from different angles. E-beam lithography, cutting and imprinting have been tested as means of producing contoured photoresists. The merits of each will be discussed and preliminary stereo images of hydrated biological specimens presented.

Soft x-ray contact microscopy provides the biologist with a technique for examining the ultrastructure of living cells at a much higher resolution than that possible by various forms of light microscopy. Readout of the developed photoresist using atomic force microscopy (AFM) produces a detailed map of the carbon densitites generated in the resist following exposure of the specimen to water-window soft x-rays (2-4nm) produced by impact of a high enrgy laser onto a suitable target. The established high resolution imaging method of transmission electron microscopy (TEM) has inherent problems in the chemical pretreatment required for producing the ultrathin sections necessary for this technique. Using the unicellular green alga Chlamydomonas the ultrastructural appearance of the cells following SXCM and TEM has been compared. While SXCM confirms the basic structural organization of the cell as seen by TEM (e.g. the organization of the thylakoid membranes within the chloroplast; flagellar insertion into the cytoplasm), there are important difference. These are in the appearance of the cell covering and the presence of carbon-dense spherical cellular inclusions.

We have developed an advanced Kirkpatrick-Baez (AKB) microscope to diagnose a laser- produced-plasma. The AKB microscope optics are two pairs of hyperbolical and elliptical cylindrical-mirrors to avoid a spherical aberration and field obliquity. Ray trace calculation was applied to optimize the characterization. The microscope has attained a spatial resolution of less than 3 mm at 2.5-keV x-ray in the field of 800 mm from experiments.

The collision of laser-produced plasmas has been diagnosed by x-ray spectroscopy and imaging. The two colliding plasmas are produced on Al thin foils at a distance of 200 to 900 micrometers irradiated at (lambda) equals 0.53 micrometers with laser intensities of 3 X 10¹³ to 6 X 10¹³ W/cm². Interpretation of the plasmas was visualized by replacing one of the foils material by magnesium. The main diagnostics were x-ray crystal optics based on flat, cylindrical, and toroidal crystals viewing the inter-target space. A multifluid eulerian monodimensional hydrodynamic code coupled with a radiative-atomic package provided simulations of the experiments. Hydrodynamic 2D simulations calculating the lateral expansion of the plasma enabled a reliable treatment of reabsorption along the line of sight of the spectrographs. The size and the time duration of the collision, the plasma parameters in the collision region (T_e, T_i, and n_e) and interpenetration were measured. The hydrocode simulations give a good understanding of the behavior of the collision in function of intertarget distance and laser intensity.

Colliding Au, CD, and Ti-CR plasmas have been generated by illuminating two opposing foils each with an approximately 100J, 0.5 nsec, 2(omega) Nd-glass laser beam from the Trident laser facility at Los Alamos. The plasmas are being used to study plasma interactions which span the parameter regime from interpenetrating to collisional stagnation. X-ray emission during the laser target interaction and the subsequent collision is used to diagnose the initial plasma conditions and the colliding plasma properties. X-ray instrumentation consists of a 100 ps gated x-ray pinhole imager, a time-integrated bremsstrahlung x-ray spectrograph and a gated x-ray spectrograph used to record isoelectronic spectra from the Ti-Cr plasmas. The imager has obtained multiframe images of the collision and therefore, a measure of the stagnation length which is a function of the ion charge state and density and a strong function of the electon temperature. Other isntrumentation includes a Thomson scattering spectrometer with probe beam, neutron detectors used to monitor the CE coated foil collisions, and an ion spectrometer. We will describe the current status of the experiments and current results with emphasis on the x-ray emission diagnostics. We will also briefly describe the modeling using Lasnex and ISIS, a particle-in-cell code with massless fluid electronics and inter-particle (classical) collisions.

We have studied theoretically and experimentally the x-ray production above 1 keV from femtosecond laser plasmas generated on periodically modulated surface targets. Laser energy coupling to plasma surface waves has been modeled using a numerical differential method. Almost total absorption of incident laser radiation is predicted for optimized interaction conditions. Silicon gratings have been irradiated by a 120 fs Ti: sapphire laser at irradiances in excess of 10¹⁶W/cm². X-ray intensities above 1.5 keV (K-shell lines) have been measured as a function of the incidence angle. Results show a distinct x-ray emission maximum for the first order diffraction angle and are in good qualitative agreement with our theoretical predictions.

Spectroscopic investigations of the x-ray emission of plasmas heated by 120 ps, frequency doubled pulses from the JANUS Nd:glass laser are presented. High Z K-shell spectra emitted from slab targets heated to near 10¹⁷ W cm^-2 intensity are investigated. High resolution ((lambda) /$DELTA(lambda) > 5000) x-ray spectra of multicharged ions of He-like Ti, Co, Ni, Cu, and also H-like Sc in the spectral range 1.5-3.0 angstrom are obtained in single laser shots using a spherically bent Mica crystal spectrograph with a 186 mm radius of curvature. The spectra have 1D spatial resolution of about 25 micrometers and indicate that the size of the emission zone of the resonance transitions is < 25 micrometers . Simultaneous x-ray images of the plasma from a charge-coupled device pinhole camera confirmed that the plasma x-ray emission is from a similar sized source. Survey spectra ((lambda) /(Delta) (lambda) equals 500 - 1000) taken with a flat LiF (200) crystal spectrometer with a charge-coupled device detector complement the high resolution data. 2D LASNEX modeling of the laser target conditions indicate that the high K-shell charge states are produced in the hot dense region of the plasma with electron temperature > keV and density approximately 10²² cm^-3. These experiments demonstrate that with modest laser energy, plasmas heated by high-intensity 120 ps lasers provide a very bright source of hard approximately 8 keV x-ray emission.

We propose single photon counting and photon correlation measurements to characterize ultrashort x-ray pulses, in particular to measure their coherence, spectral, and temporal characteristics. Photon coincidence measuremens allow us to measure higher order correlation between photons, from which temporal profiles of the pulses can be recovered.

We use classical trajectory Monte Carlo (CTMC) simulations to study the ionization of small rare gas clusters in short pulse, high intensity laser fields. We calculate, for a cluster of 25 neon atoms, the ionization stage reached and the average kinetic energy of the ionized electrons as functions of time and peak laser intensity. The CTMC calculations mimic the results of the much simpler barrier suppression model in the limit of isolated atoms. At solid density our results give much more ionization in the cluster than that predicted by the barrrier suppression model. We find that when the laser intensity reaches a threshold value such that on average one electron is ionized from each atom, the cluster atoms rapidly move to higher ionization stages, approaching NE⁺⁸ in a few femtoseconds. This 'ignition' process creates an ultrafast pulse of energetic electrons in the cluster at quite modest laser intensities.

Laser-generated, hard x-rays are produced in a > 10¹⁸ W/cm² focus of an ultrashort-pulse laser system. The application of ultrashort-duration, laser-generated x-rays to diagnostic medical imaging is discussed. Time-gated detection allows removal of scattered radiation, improved image quality and possible reduction of patient exposure. Methods for improvement of x-ray yield, design of appropriate drive lasers, and applications to mammography and angiography are also discussed.

We describe a new technique for investigating transient structural changes in solid materials involving short pulses of hard x-rays emitted by laser produced plasmas. Using a subpicosecond terawatt scale laser system, this technique promises to be simple and broadly applicable to both single-crystal and polycrystalline samples of thickness up to millimeters and mean atomic number up to approximately 30. Spatial resolutions of order 10 micrometers should be possible. Of particular interest are the so-called energetic materials, chemical high explosives, whose initiation and reaction dynamics in relation to physical microstructure are poorly known.

Harmonics generation of laser radiation ((lambda) equals 1.06 (mu) , (tau) equals 3 ps) in low- temperature atmospheric plasma and target plasma is investigated. Results of odd harmonics generation (up to eleventh, (lambda) equals 96 nm) are reported.

Achieving the goals set by the US National Semiconductor Roadmap requires that sub 0.18 micrometers design rules be incorporated into semiconductor device structures by the year 2001. A promising approach makes use of shorter wavelength radiation than is presently used in lithography, namely x-rays. Arrays of glass fibers have the potential of controlling parameters important to point-source lithogrpahy including local divergence, global divergence, and field uniformity. The high absorption rate and scattering of x-rays in air at energies less than 3 keV necessitates that experiments must be conducted in an evacuated environment. Consequently, there has been little research on the transmission of x-rays through glass polycapillary fibers at energies of 3 keV and lower. An experimental setup to test capillaries under such conditions has been developed at XOS. It has sufficiently long optical paths in the vacuum chamber to be useful in evaluating the parameters critical for semiconductor lithography. Experimental and simulated transmission characteristics of polycapillary fibers as well as a discussion on the feasibility of using them in a collimator for x-ray lithography are presented in this paper.

We perform pump-probe measurements in which intense ultrashort optical pulses are the pump pulses that initiate a chemical reaction and ultrafast x-ray pulses are the probe pulses that monitor the response of the system. We present experimental results on the observation of a chemical reaction process, photoinduced dissociation of gas phase SF₆ molecules, detected by ultrafast x-ray absorption spectroscopy with 3 ps time resolution near the sulfur K edge at a photon energy of 2.48 keV (4.98 A). High contrast light pulses of 400 fs duration (500 mJ energy and 0.53 micrometers wavelength) from the INRS terawatt laser were focused on high atomic number targets at an intensity of 5 X 10¹⁷ W/cm² in order to generate an x-ray continuum around the sulfur K edge. The SF₆ molecule exhibits intense near shape resonances at the sulfur K and L edges, due to the multiple scattering and interference of the emitted photoelectrons by the fluorine atoms that symmetrically surround the central sulfur atom. The shape resonance of the molecule is clearly resolved in the absence of any pump pulse, and the variation of the x-ray absorption spectrum was measured as a function of the delay between the optical pump and x-ray probe pulses. As expected from theory, the reaction process is faster than can be resolved with the 3 picosecond duration x-ray pulses used in this initial experiment. This fast response can, in principle, be used to measure the duration of ultrashort x-ray pulses.