Experimental studies of impinging liquid jet-array cooling modules are described. Heat fluxes have reached 7.2 MW/m² over an area of 10.3 cm². Heat loads are applied to a metallic faceplate by a plasma-sprayed resistance heater. Faceplates are a few millimeters thick and area made from either copper or molybdenum alloys. The faceplate is cooled by an array of 14 small diameter water jets operating at speeds of 46 m/s. Cooling is believed to be entirely convective, without boiling. The construction of the plasma sprayed heaters presents several challenges, including unpredictable thermal and electrical properties, high thermal resistances, and fracture at high temperatures. Thermal resistances are quantified and our experience with the heaters is described.

A criterion is developed for determining the range of microscale heat transfer effects in systems exposed to laser radiation. Criteria in terms of the time and spatial scale for using microscale heat transfer mechanisms have been reported. The significance of microscale heat transfer effects during high incident laser radiation is investigated. Numerical simulations show the microscale heat transfer effects are scaleable with respect to the incident laser radiation. The significance of the microscale mechanism is shown to be dependent on the rate of heating. There is good agreement with values reported in the literature and the present model's predicted temperatures and normalized reflectance changes for a thermoreflectance problem.

Task T-222 of the International Thermonuclear Experimental Reactor (ITER) program addresses the manufacturing and testing of permanent components for use in the ITER divertor. Thermal-hydraulic and critical heat flux performance of the heat sinks proposed for use in the divertor vertical target are part of subtask T-222.4. As part of this effort, two single channel, medium-scale, bare copper alloy, hypervapotron mock-ups were designed by Sandia National Laboratories and McDonnell Douglas Aerospace (MDA), fabricated at MDA and tested at Sandia' Plasma Materials Test Facility using the EB-1200 electron beam system. The objectives of our effort were to develop the design and manufacturing procedures required for construction of robust HHF components, verify thermal-hydraulic, thermomechanical and CHF performance under ITER relevant conditions, and perform analyses of HHF data to identify design guidelines, failure criteria and possibly modify any applicable CHF correlations. This paper describes the design, fabrication and finite elements modeling of two types of hypervapotrons, a common version already in use at JET and a new attached- fin design. HHF test data on the attached-fin hypervapotron will be used to compare the CHF performance under uniform heating profiles on long heated lengths to that of localized, highly peaked, off-nominal profiles.

An experimental setup to measure the thermal contact conductance across a silicon-copper (Si-Cu) interface is described, and the results obtained are presented. The resulting thermal contact resistance data are used in estimating the thermo-mechanical and optical performance of optical substrates cooled by interfaced copper cooling blocks. Several factors influence the heat transfer across solid interfaces. These include the material properties, interface pressure, flatness and roughness of the contacting surfaces, temperature, and interstitial material, if any. Results presented show the variation of thermal contact conductance as a function of applied interface pressure for a Cu-Si interface. Various interstitial materials investigated include indium foil, silver foil and a liquid eutectic (Ga-In-Sn). As expected, thermal contact resistance decreases as interface pressure increases, except in the case of the eutectic, in which it was nearly constant. The softer the interstitial material, the lower the thermal contact resistance. Liquid metal provides the lowest thermal contact resistance across the Cu-Si interface, followed by the indium foil, and then the silver foil.

X-ray optical elements (such as single-crystal silicon monochromators) illuminated with high-power synchrotron- radiation beams produced by insertion devices and, to a lesser extent bending magnets, require cooling. When operating a silicon crystal at room temperature, channels for the coolant are often fabricated directly beneath the diffracting surface. Then a separate silicon distribution manifold/plenum is manufactured, and the components are bonded together using an adhesive or some intermediate material. In many cases, such monochromators suffer from strains induced by the bond. A silicon-to-silicon direct- bonding technique (i.e., without any intermediate material) has been developed that appears to be an attractive method for creating a bond with less strain between two pieces of silicon. This technique is well understood for the case of thin wafers (approximately 0.5 mm thickness) and is used by the semiconductor industry. Recently, bonding of 16-mm-thick 10-cm-diameter silicon crystals has been successfully performed inducing very little strain. A short review of the silicon-to-silicon direct-bonding process will be presented with an emphasis on its application to room temperature high-heat-load x-ray optics along with the present status of direct bonding efforts at the APS.

Third-generation, high-intensity, x-ray synchrotron radiation sources are capable of producing high heat-flux x- ray beams. In many applications finding ways to handle these powers is viewed as a burden. However, there are some technological applications where the deep penetration length of the x-rays may find beneficial uses as a volumetric heat source. In this paper we discuss the prospects for using high power x-rays for volumetric heating and report some recent experimental results. The particular applications we focus on are welding and surface heat treatment. The radiation source is an undulator at the Advanced Photon Source. Results of preliminary tests on aluminum, aluminum metal matrix composites, and steel will be presented.

A new open-loop heat flux control technique has been developed to conduct transient thermal testing of thick, thermally-conductive aerospace structures. This technique uses calibration of the radiant heater system power level as a function of heat flux, predicted aerodynamic heat flux, and the properties of an instrumented test article. An iterative process was used to generate open-loop heater power profile prior to each transient thermal test. Differences between the measured and predicted surface temperatures were used to refine the heater power level command profiles through the iterative process. This iteration process has reduced the effects of environmental and test system design factors, which are normally compensated for by closed-loop temperature control, to acceptable levels. The final revised heater power profiles resulted in measured temperature time histories which deviated less than 25 degree(s)F from the predicted surface temperatures.

This paper addresses some options and techniques in the thermal management of masks used in deep x-ray lithography. The x-ray masks are thin plates made of low-atomic-number materials on which a patterned thin film of a high-atomic- number metal has been deposited. When they are exposed to an x-ray beam, part of the radiation is transmitted to replicate the pattern on a downstream photoresist, and the remainder is absorbed in the mask in the form of heat. This heat load can cause deformation of the mask and thus image distortion in the lithography process. The mask geometry considered in the present study is 100 mm X 100 mm in area, and about 0.1 to 2 mm thick. The incident radiation is a bending magnet x-ray beam having a footprint of 60 mm X 4 mm at the mask. The mask is scanned vertically about +/- 30 mm so that a 60 mm X 60 mm area is exposed. The maximum absorbed heat load in the mask is 80 W, which is significantly greater than a few watts encountered in previous systems. In this paper, cooling techniques, substrate material selection, transient and steady state thermal and structural behavior, and other thermo-mechanical aspects of mask design are discussed. It is shown that, while diamond and graphite remain attractive candidates, at present beryllium is a more suitable material for this purpose and, when properly cooled, can provide the necessary dimensional tolerance.

The high power and/or power density of the X-ray beams of the European Synchrotron Radiation Facility induces engineering constraints for the design of the beamlines, in order to reduce the temperature and the thermal distortion of optical components. The requirements in beam stability, ever more stringent, lead to new engineering constraints, generally in contradiction with high cooling performances: the vibrations created by the cooling fluid -or flow induced vibrations- must now be integrated at the design stage. This document describes the efforts made at the ESRF to better master this aspect, and gives qualitative guidelines which could be used at the design stage of high power optical elements.

We present a graphite-filter/beryllium-window design for the high power wiggler beam lines at CHESS. This design, which has been in operation since 1993 and proves to be very successful, employs a 0.25 mm thick highly-oriented pyrolytic graphite (HOPG) foil as a prefilter to reduce the heat load on the first vacuum beryllium window. Unlike most graphite filter designs in other synchrotron sources that rely entirely on radiative cooling, the HOPG filter in our design is brazed onto a water-cooled copper flange which substantially lowers its maximum temperature and prolongs its operational lifetime. We discuss the safety criteria that we have established based on theoretical calculations and infrared temperature measurements on the graphite filter and the beryllium window.

A beryllium (Be) window for an Advanced Photon Source diagnostics beamline has been designed and built. The window, which has a double concave axisymmetrical profile with a thickness of 0.5 mm at the center, receives 160 W/mm² (7 GeV/100 mA stored beam) from an undulator beam. The window design as well as thermal and thermomechanical analyses, including thermal buckling of the Be window, are presented.

An analysis is presented of the optical distortions and stress levels of side-mounted, uncooled windows for interceptors. The optical distortions addressed herein are caused by aerodynamic heating of the window. The boresight error, boresight bias rate, boresight error slope and blur circle were analyzed.

High-power/high-energy (HEL) systems include an optical train consisting of mirrors and windows, which must be capable of transporting and directing the beam without seriously degrading the nominal performance of the laser. Since catastrophic failure modes are not a major threat at beam-power levels of current interest, the system's performance as measured in terms of achievable target irradiances can degrade as a result of thermal lensing, that is, the wavefront distortion caused by thermally induced phase aberrations. The purpose of this paper is to present an analytical investigation that addresses the problem of evaluating the impact of laser-driven mirror distortions; in this context it is shown how to obtain simple figures of merit (FoM) for rating the thermal lensing performance of mirror-faceplate material candidates. The performance of cooled HEL mirrors reflects their ability to minimize irradiance-mapping wavefront distortions, which leads to defining a thermal distortion coefficient (Xi) equals (alpha) (1+v) that controls the out-of-plane growth of the faceplate. It is then straightforward to derive equations for characterizing the RMSsed surface deformation and to assess the merits of mirror-faceplate material candidates in a pulsed or a CW environment. Figures of merit for CW operation must take into account the requirement that the faceplate should be as thin as possible but still able to minimize coolant-induced pressure ripples; the modulus of elasticity, therefore, must be properly factored into FoM expressions. Since water-cooled HEL mirror heat-exchangers exhibit relatively modest Biot numbers (Bi < 1), the thermal conductivity of the faceplate is not a critical material parameter. Numerical evaluations demonstrate that the ranking of faceplate-material candidates does not depend on the laser mode of operation or the efficiency of the heat exchanger. It is the thermal expansion coefficient (alpha) that determines the performance if optical distortions are of concern. For this reason, diamond shows much promise, which will attract attention as CVD-diamond fabrication technologies mature.

The use of refractive lenses for focusing x-ray beams has been the subject of publications since the early 1980s. Detailed calculations have been made for different shapes for the refractive lens: cylindrical, spherical, parabolic, and for a Fresnel-type refractive lens. The main drawback to the use of a single refractive lens to focus x-rays is that the index of refraction (n equals 1 - (delta) ) is very close to 1, which results in a lens with a very long focal length. Recently Snigerov and others have suggested and experimentally demonstrated, using cylindrical-shaped lenses, that this problem of long focal lengths can be overcome by using many lenses in series. Each lens refracts the photon through a small angle, but the sum of these sequential changes in direction can be moderately large. This increase in effective refraction angle reduces the focal length of the lens to a few meters or less and makes the multi-element lens a much more useful instrument for focusing x-rays. This paper, annualizes the expected performance of a lens consisting of a series of aligned hollow spheres in a beryllium substrate. The use of hollow spheres rather than hollow cylinders produces focusing of the x rays into a small focal spot in contrast to the single-directional focusing of the hollow cylinders, which produces a line focus. The use of beryllium as the substrate results in lower photo cross sections for both scattering and absorption relative to the value of the refractive index as compared to higher-Z materials and results in higher transmission values than for lenses with thin webs between the lens elements without distorting the surfaces of the neighbor lens element. This plus beryllium's low density, keep the absorption and scattering in the web at a minimum. The calculations suggest that one will be able to make Be lenses with short focal lengths (1 to 2 m) with useable transmissions (10 to 30%). Two multi-element lenses have been constructed: one with 20 1-mm-diameter hollow spheres in an aluminum substrate, and one with 50 hollow spheres, 1 mm in diameter, in a beryllium substrate. Some construction details and calculations of the expected performance, are given for these two multi-element lenses.

High energy wigglers produce extremely high total powers. For example, the insertion device for one beamline of the Basic Energy Sciences Synchrotron Research Center is an elliptical multipole wiggler which can generate circularly polarized X-rays on axis and produces a total power of approximately 8 kW. This insertion device will be used to simultaneously provide x-rays to three branch lines, a branch equipped with a normal double crystal monochromator feeding a scattering and spectroscopy station, and two branches with single-bounce horizontally deflecting monochromators for Compton scattering and High Energy Diffraction. The crystal optics for this type of device require substantially different heat load solutions than those used for undulator beamlines. We will discuss how the beam is split and shared among the beamline branch lines and present the crystal cooling strategies employed for both the double crystal monochromator and horizontally deflecting single-bounce monochromators.

Intense synchrotron radiation from high power wiggler sources has long been a difficult high-heat-load problem to the design of properly cooled x-ray optics. Large, high power and very intense beams thermally distort crystal optics, reducing throughput and broadening rocking curves. An internally cooled silicon monochromator has been fabricated which demonstrated the capability of diffracting wiggler radiation of unprecedented power without significant degradation of the beam. Cooling water flows through rectangular cooling channels 1 mm wide, 1 mm below the diffracting surface, fed by a manifold bonded to the underside of the diffracting crystal. In an attempt to improve high power performance, a second monochromator was fabricated with a pin-fin cooling structure instead of channels. Both used a novel silver diffusion bond to ensure leak-tight UHV performance. Recent test results at wiggler station F2 show a linear behavior of the x-ray flux with increasing storage ring current up to a total power of 3 kW and a peak surface power density of 5 W/mm². The improved monochromators have led to an increase of x-ray flux by a factor of six over previous contact-cooled designs and show that internal water-cooling can be an effective solution to high-heat-load problems at high power wiggler stations.

Assuring the survivability and adequate performance of monochromator crystals under the high heat-load conditions present at modern synchrotron light sources, without unduly compromising throughput, is an ongoing challenge in x-ray optics design. Various partial solutions were generated over the last few years but there is still a significant room for improvement. In this paper a new crystal design, allowing for a significant increase of throughput without sacrificing performance, is presented. Utilizing novel geometry the crystal optimizes the utilization of the tried and proven schemes of cryogenic cooling and thin-crystal design, to provide superb heat-load handling capability.

First optical elements at third-generation, hard x-ray synchrotrons, such as the Advanced Photon Source, are subjected to immense heat fluxes. The optical elements include crystal monochromators, multilayers and mirrors. This paper presents a mathematical model of the thermal strain of a three-layer (faceplate, heat exchanger, and baseplate), cylindrical optic subjected to a narrow beam of uniform heat flux. This model is used to calculate the strain gradient of a liquid-gallium-cooled x-ray monochromator previously tested on an undulator at the Cornell High Energy Synchrotron Source. The resulting thermally broadened rocking curves are calculated and compared to experimental data. The calculating rocking curve widths agree to within a few percent of the measured values over the entire current range tested (0 to 60 mA). The thermal strain gradient under the beam footprint varies linearly with the heat flux and the ratio of the thermal expansion coefficient to the thermal conductivity. The strain gradient is insensitive to the heat exchanger properties and the optic geometry. This formulation provides direct insight into the governing parameters, greatly reduces the analysis time, and provides a measure of the ultimate performance of a given monochromator.

At the Advanced Photon Source (APS), undulator insertion devices are capable of producing x-ray beams with a total power of about 5 kW and normal incidence heat fluxes of about 170 W/mm² at 30 m from the source. On beamlines in which the first optical element is a mirror, the reflected beam from the mirror still carries considerable power and power density. Depending on its location, the monochromator downstream of the mirror might be subject to 300 W total power and 5 W/mm² normal incidence heat flux. Thus, it is still necessary to carefully design a monochromator that provides acceptable performance under these heat loads. A contact-cooled u-shaped monochromator may be used in this case. The main feature of the u-shaped monochromator is that, by carefully selecting the geometry and cooling locations, it passively corrects for some of the thermally induced crystal distortions. We present experimental and computational results of a contact cooled u-shaped monochromator tested on an APS undulator beamline. The results are encouraging and compare favorably with liquid- gallium internally cooled crystals.

We fabricated a water-cooled silicon monochromator crystal with small channels for the special case of a double-crystal fixed-exit monochromator design where the beam walks across the crystal when the x-ray energy is changed. The two parts of the cooled device were assembled using a new technique based on low melting point solder. The bending of the system produced by this technique could be perfectly compensated by mechanical counter-bending. Heat load tests of the monochromator in a synchrotron beam of 75 W total power, 3 mm high and 15 mm wide, generated by a multipole wiggler at SSRL, showed that the thermal slope error of the crystal is 1 arcsec/40 W power, in full agreement with finite element analysis. The cooling scheme is adequate for bending magnet beamlines at the ESRF and present wiggler beamlines at the SSRL.

A wide variety of construction methods and mechanical designs for hard x-ray monochromators is in use at synchrotron radiation facilities around the world. These cover the gamut in complexity and sophistication. This paper will try to show that sophistication and complexity are not inevitable partners, and that the use of elastic monolithic design can remove much of the complexity in x-ray monochromators of all levels of sophistication.

On-stream performance of crystal diffraction components at synchrotron radiation facilities is certainly limited by the issues of heating and photochemistry discussed by other contributors to this conference. At the same time, the delivered performance is also limited by the quality of the starting material, strains and distortions introduced by the initial shaping and surface preparation, and by whatever re- forming operations (e.g. one and two dimensional stressing or bending) are needed to realize the desired x-ray optical function. Each of these issues, initial crystal quality, surface preparation, and final figure, can be assessed by rather simple arrangements using conventional x-ray diffraction tube sources in combination with appropriate beam conditioning elements and effective 2D imagers. These table-top arrangements represent a modest investment in comparison with realistic costs of beamline operation, but they are often neglected. An evident benefit of such off- line quality control is that on-line performance becomes predictable at least in the limit of low heating and radiation damage, so that the effects of these inevitable difficulties are more clearly isolated, thereby opening the way toward their effective management.

The time-delayed scattering in 14.4 keV Mossbauer resonance at ⁵⁷ Fe has been intensively studied using pulsed synchrotron radiation source and delayed coincidence technique. Here, results of x-ray interferometric experiments are shown, which were obtained at KEK with this nuclear resonance. There, the interference arises from the beams re-emitted by different nuclei with some time interval after the absorption of a single incident photon. High visibility interference oscillations verified the coherent superposition as well as the complete coincidence in the time domain of the beams after the nuclear resonance. The collective nuclear excitations from a single photon are discussed with regard to the no-reduction of the detection probability. In addition, the phase information transfer in successive absorption and re-emission with time-delay, and the dispersion effect at the nuclear resonance are shown.

We report on the recent development of high energy resolution X-ray optics for nuclear scattering experiments at the Nuclear Resonance beamline of the European Synchrotron Radiation Facility. The design of the monochromators with energy resolution of 4.4 meV, 1.7 meV, and 1.1 meV is described. Nuclear inelastic absorption experiments demonstrate their performance.

A monochromator for use at 13.84 keV with a calculated bandpass of 5.2 meV was designed, built, and tested. Tuning was performed by rotating the inner crystal of a pair of nested silicon channel cut crystals. The inner crystal employs the (884) reflection, and the outer crystal employs a collimating asymmetric (422) reflection (dynamical asymmetry factor, b, equal to -17.5). Tests were done with a double crystal Si(111) pre-monochromator situated upstream of the high resolution monochromator and a Si(777) backscattering crystal situated downstream. For this optical arrangement an ideal value of 6.3 meV as calculated by x-ray dynamical diffraction theory applies for the FWHM of the convolution of the net monochromator reflectivity function with that of the Si(777) reflection. This calculated value is to be compared to the value of 7.1 meV measured by tuning the high resolution monochromator. Measured efficiencies were less than ideal by a factor of 3.2 to 4.9, where the larger flux reduction factors were found with higher positron storage ring currents.

A new technique for making a high resolution X-ray analyzer is presented. The analyzer consists of a silicon wafer with <111> orientation, a Pyrex glass wafer and a concave polished Pyrex glass substrate. The energy resolution of the analyzer was studied on the inelastic scattering beamline of the Synchrotron Radiation Instrumentation Collaborative Access Team on sector 3 of the APS using the Si(777) back reflection at 13.84 keV. Details are presented and compared with other techniques, and we discuss contributions of the measured energy resolution.

We have studied the diffraction properties of Si-TaSi₂ single crystals as high flux x-ray monochromators in a wide range of x-ray energies from 4 keV to 80 keV on beamline 7-2 at the Stanford Synchrotron Radiation Laboratory. Uniform rocking curves with predominant Lorentzian shape were observed in symmetric Bragg geometry. The peak reflectivity of the (111) reflection varied between 25% and 69% and the full width at half height between 40 arcsec and 133 arcsec. Similar results were measured for the (220) reflection. An interesting possibility arises from the anisotropy of the material: the resolution and consequently the flux can be varied by a factor two or more by simply rotating the crystal in its diffraction plane. The gain factors measured ranged from 3.2 at 6 keV to 43 at 80 keV for the 111- reflection and from 2.3 at 6 keV up to 128 for 60 keV for the 220-reflection, respectively. The agreement with the theory based on the mosaic model was partly good, but generally unsatisfactory.

A number of different applications for high resolution Bragg Focusing Optics are reviewed. Applications include Sagittal Focusing, Energy Dispersive optics for x-ray absorption and diffraction, a curved analyzer-multichannel detector method for efficient acquisition of powder and small angle scattering data, the use of Backscattering Analyzers for very high resolution inelastic scattering, and curved crystals for high energy applications.

Various theoretical methods for calculating diffraction profiles of perfect crystals are available in literature. Although these methods hold within certain validity ranges due to their inherent approximations, they constitute the current state-of-the-art of numerical computation of diffraction profiles. In this paper we summarize the theory of Zachariasen for flat crystals, the multi-lamellar approximation for bent crystals and the Penning-Polder approximation for bent Laue crystals. Some examples of their results are presented. Another method to calculate the diffraction profile consists in solving the Takagi-Taupin equations. The finite difference method, that provides a numerical solution of these equations, is briefly discussed. A new method for solving numerically these equations using the finite element method is proposed. This method is very flexible, because it can consider a crystal with an arbitrary shape and cover the case of critical regime (i.e., inhomogeneities and deformations) with fine elements. In addition, it can couple naturally the diffraction calculation with thermal or mechanical crystal deformations. These deformations are generally induced by the x-ray beam (heat load), the crystal bender (mechanical stress) or are intrinsic to the crystal (inhomogeneities, impurities, dislocations, etc.). An example of the feasibility of this method is shown.

We report on a design and on some experimental results for the performance of a new high energy resolution monochromator. It is a large channel-cut Si crystal with a 197 mm separation between the two faces designed to operate in a near-backscattering regime. The device was tested as a second monochromator on Sector 3 of the Synchrotron Radiation Instrumentation Collaborative Access Team at the Advanced Photon Source using the Si(777) reflection at a photon energy of 13.84 keV. The same monochromator can be used for other energies with reflections of the type (hhh). Special care has been taken to equalize the temperature of the two faces by employing a Peltier heat pump. A Si(111) double-crystal pre-monochromator designed to withstand the high heat load of the undulator radiation was used upstream on the beamline. The measured throughput efficiency of the Si(777) channel-cut monochromator was less than ideal by a factor of 1.9. Dynamical diffraction theory was used to calculate the throughput of an ideally perfect crystal.

Synthetic diamond crystals for monochromators of synchrotron radiation beams were grown by temperature gradient method under high pressure and high temperature. Crystalline quality of the diamond was investigated by double-crystal X- ray rocking curve measurement, polarized optical microscopy and X-ray topography. FWHM of rocking curve of the crystal was found to be clearly correlate with macroscopic strain observed by polarized optical microscope. It was also found that large line defects in the crystals have a more significant influence on the FWHM than nitrogen impurity and plane defect. Most of line defects and nitrogen impurity could be removed by using high-quality seed and by adding nitrogen getter to metal solvent, respectively. Using this method, high-quality and large diamond-crystals of more than 9 mm, which is useful for the monochromator, could be obtained.

The performance of a fixed exit double crystal monochromator in terms of stability and reproducibility of the outgoing X- ray beam becomes the crucial point at modern synchrotron beamlines dealt with the high resolution X-ray optics. Due to the high heat load the monochromator crystals have to be cryogenically cooled. The cooling loop of the second crystal may have an impact on the performance of the monochromator. We therefore suggest to use a Si_1-xGe_x single crystal as the first cooled crystal of the monochromator. With that the second crystal is held at room temperature. To verify the proposed solution an experiment was performed where the lattice parameters of pure Si and SiGe crystals as a function of temperature were measured.

Time-dependent x-ray diffraction has been measured from laser-irradiated semiconductor crystals. Laser pulses with 100 fs duration and 800 nm wavelength excite the sample inducing phase transitions. 5 keV x-rays from the Advanced Light Source are diffracted by a sagittally-focusing Si (111) crystal and then by the sample crystal, InSb (111), onto an avalanche photodiode. By detecting individual pulses of synchrotron radiation, which have a duration of 70 ps, the diffracted intensity is observed to decrease because of photoabsorption in a disordered surfaced layer. Rocking curves measured after the laser irradiation show a tail, which results from a strained region caused by expansion of the crystal lattice.

Focusing of hard X-rays by refraction has been a long time been considered as unfeasible due to strong absorption and weak refraction of X-rays in matter. Recently it has been shown that compound refractive lenses can overcome the problem. It was demonstrated that the best candidates for lenses are low Z, high density materials. Linear and 2D lenses from aluminum, boron carbide, beryllium, pyrographite and Teflon were produced and tested. Focusing of 2 - 3 microns was achieved at an energy range from 9 to 30 keV. Compound refractive lenses have low sensitivity to heatload and are extremely well suited for focusing of undulator radiation. Two-plane focusing lenses have been optimized, built and installed in the white beam of the undulator on the machine diagnostic beamline of the ESRF to be used as an X-ray emittance diagnostic. The future potentials of the refractive lenses will be discussed as well.