Sensitive and versatile measurement techniques have been developed to better characterize a wide variety of surfaces. These include sensitive optical noncontact and mechanical contact profilers, optical devices to measure area topography, profilers that can measure lengths up to 100 mm, and sensitive scatterometeis to measure small amounts of scattering from opaque and transparent materials. They can be used for characterizing, among others, high quality optical surfaces for applications in the infrared, visible, ultraviolet, extreme ultraviolet and x-ray regions, precision machined surfaces, and wafers used in the semiconductor industry. Fabrication shops can now have instruments for process control during manufacturing, so that smoother surfaces with less subsurface damage can be made. The ultrasensitive characterization techniques have also made possible experiments to better understand surface structure, ranging from atomic structure to macrostructural topography. The papers in this conference illustrate many of the latest developments in the area of surface characterization, and experiments performed using some of the specialized instrumentation.

With increasing competition in the manufacturing industries product quality is becoming even more important. The shortcomings of human inspectors in many applications are well know, however, the eye/brain combination is very powerful and difficult to replace. At best, any system only simulates a small subset of the human's operations. The economic justification for installing automatic inspection is often difficult without previous applications experience. It therefore calls for confidence and long-term vision by those making the decisions. Over the last ten years the use of such systems has increased as the technology involved has matured and the risks have diminished. There is now a complete spectrum of industrial applications from simple, low-cost systems using standard sensors and computer hardware to the higher cost, custom-designed systems using novel sensors and processing hardware. The underlying growth in enabling technology has been in many areas; sensors and sensing techniques, signal processing and data processing have all moved forward rapidly. This paper will examine the currrent state of automatic inspection and look to the future. The use of expert systems is an obvious candidate. Parallel processing, giving massive increases in the speed of data reduction, is also likely to play a major role in future systems.

Interferometers are widely used for measuring the profiles of 'super-smooth' surfaces. The most sensitive of these instruments employ a 'common-path' design with the two interfering beams being reflected from different areas of the test surface, making the interferometer insensitive to vertical motion of the surface. When these systems are applied to the measurement of surface 'roughness' the shortest surface wavelength that can be measured depends upon both the wavelength of the measurement radiation and the numerical aperture of the probe beam optics. The longest surface wavelength depends upon the optical configuration of the interferometer. A profiling interferometer will be briefly described that has a sensitivity to surface height of better than 0.01 nm and a surface wavelength range from 0.5 to 15 micrometres. The results of measurements on a number of surfaces using this instrument will be shown and the methods used for analysing these will be discussed.

The principle of the high precision optical surface sensor (HIPOSS) is based on the focus detection of the critical angle method. This sensor has two sensing ranges in its characteristic curve, a precision range and a wide range. The precision range is used for an ordinal ultra-high precision profilometer, of which sensitivity is better than 0.2nm rms with about 2μm measuring scope. On the other hand, relatively large linear measuring range of around 20μm with 1nm resolution can be achieved by use of the wide range. Besides the two and three dimensional profilometer, each range of the HIPOSS can be used as a compact displacement sensor or a dynamic sensor of higher than 10kHz response such as in-process measurement for single point diamond turning control. The HIPOSS was designed to eliminate several errors especially from the surface inclination. Taking advantage of the function, an inclination angle of the surface can also be detected by the HIPOSS with about 1 minute of arc sensitivity.

The statistical parameters of surfaces to be measured for industrial applications vary over several orders of magnitude. Surfaces with large slopes or edges are particularly difficult to be recorded. Some measuring methods developed in our laboratory are compared and the range of applications are discussed. For polished and fine ground glass and metal surfaces a heterodyne profilometer with a vertical resolution of 0.5 nm, lateral resolution of 0.6 μm, and large scanning length is discussed. The interferometer can be changed from single- to double-pass operation by rotation of a quarter-wave-plate. For rougher surfaces a profilometer of the photometric-balance type with resolution Rq < 4 nm and dynamic range of 20 μm and an interference microscope with automated fringe evaluation is described. An integral white light roughness sensor covers the roughness range 0.04µm to 10µm and measures independently mean roughness and autocorrelation width.

An optical device is described that performs height data acquisition by focusing a white light beam at a sample surface and processing the backscattered light. The principle of the operation is based on longitudinal chromatic aberration of the focusing lens and spectral analysis of the image irradiance. The surface microtopography is reconstructed after automatic point by point scanning. A personal computer interfaced to the probe controls the operation and produces the roughness parameters. Inspecting optical surfaces, the height data are considered as optical path differences and processed according to standard methods of optical testing, eventually working out the absolute shape parameters. The main system characteristics are discussed, and performance examples on selected objects are demonstrated.

A new three-dimensional, non-contact laser interferometric microscope is described which uses computerized phase measuring interferometry to achieve sub-nanometer vertical resolution. Areas profiled range from 7.8mm x 5.7mm to 0.078mm x 0.057mm with a pixel sampling interval ranging from 27.0μm down to 0.27μm. Test surface reflectivity can range from less than 1% up to 100%. Turret mounted, parfocal coherent illumination (Fizeau) and incoherent illumination (Mirau) interferometric objectives permit rapid magnification change. Laser illumination coupled with selectable incoherent and coherent illumination allows variable interference fringe ranges from a few microns to millimeters length allowing rapid fringe acquisition and measurement flexibility. Tip and tilt of the entire instrument head about the plane of the test surface eliminates feature walk off from the field of view at high magnifications. Three dimensional surface plots plus user selectable two dimensional profiles extracted from three dimensional data are displayed. Two dimensional autocovariance and spectral density analysis is available. Numeric output includes rms, Ra, peak-valley, and radius of curvature. A track-ball directed interactive cursor scans analyzed data to give single pixel coordinates relative to a user defined origin.

The derivation of surface-finish statistics from profile measurements are reviewed with emphasis on the corrections and limitations imposed by the measurement process itself. These issues are important for the comparison of the results of different types of measurements, connecting results with functional surface properties, and evaluating different finishing techniques.

The performance of stylus profilometers can be defined by modelling their abilities to respond to sinusoidal profiles, and can be compared by mapping their limits in amplitude-wavelenyth (AW) space. The performance of traditional stylus profilometers fall within well-defined limits; but new applications are requiring new capabilities beyond these traditional limits. At low amplitudes and wavelengths the tip radius of practical styli has been a limiting factor. Development of the scanning tunnelling and atomic force microscopes has opened up this area of AW space, which extends to the resolution of individual atoms. At low amplitudes and long wavelengths, temporal stability and quality of the datum are critically important. Advances into this area of AW space, which is important to X-ray optical and other super-smooth surfaces, have been made at NPL with the Nanosurf-2 instrument. A description of the instrument and its design philosophy are given, along with examples of precision surfaces that have been measured with it.

A wide range of non-contact optical profilometers has been described in recent years, based on a variety of principles such as interferometry, local slope measurement, and focus detection. The performances of these instruments can be compared by considering their ability to measure sinusoidal profiles, and mapping the limitations in amplitude-wavelength (AW) space. Relevant parameters depend on the design of the particular instrument. The numerical aperture of an objective lens, the maximum density of fringes that may be resolved, and the directional (angular) stability of a laser beam are examples of parameters that directly affect performance. Other relevant factors in optical, just as in stylus, instruments are the ranges and resolutions of scanning motions, and the quality of a straight datum. AW maps are presented and compared for a number of different types of optical probes and instruments, and comparisons are made with conventional stylus instruments.

For many years we have developed a method to reconstruct the profiles of statistically rough surfaces. This method is based on a microdensitometer analysis of electron micrographs of shadowed surface replicas. From the quantized profiles it is possible to compute the statistical moments -particularly the second order one called autocovariance function (ACF)- that characterize the surface. In general ACF's for pseudorandom surfaces are not decreasing monotonic functions and some complications arise when the definition of a auto-covariance length for those surfaces is considered. Solutions are proposed to overcome them; in particular, it seems preferable to deduce certain statistical parameters from the spectrum instead of the ACF. Moreover a new approach based on the minimal spanning tree (MST) -which is a graph constructed on the set of points representing the position of features on a surface- is proposed to study statistically order and disorder in the distribu-tion of these features.

Several techniques are available for characterizing the surface microstructure of smooth components. Often it is necessary to compare the results from two or more of these techniques. This can lead to problems unless it is understood that all measurement techniques are bandwidth limited, and each technique has a characteristic transfer function. We will discuss several techniques in this regard, including the optical profilometer (WYKO), the optical scatterometer (both angle-resolved and TIS), and the mechanical profilometer (Talystep). A number of samples having different microstructure properties were characterized using these techniques, and results will be discussed.

It is necessary to take into account scattering phenomena in optical systems since one is interested in a very detailed energy balance. We recall the principles of the vector theories that have been developed to predict the spatial and spectral distribution of light scattered from an optical surface or a multilayer coating. With the help of a scattering apparatus, we emphasize the key scattering parameters (roughness, autocorrelation length, isotropy degree, crosscorrelation laws) that are necessary to characterize surface defects and materials microstructure in thin film form. In spite of numerous parameters involved in the calculation, particular techniques enable us to extract all these parameters and lead to good agreement between theory and experiment. When the coating is made of a high number of layers, investigation of experimental results is quite more difficult. However, with the help of a correctly chosen model, we can strongly reduce the number of scattering parameters, and this allows an easy comparison between theory and experiment.

An apparatus built in Marseilles, which precisely measures scattered light over all space from optical surfaces, is described. Both coated and uncoated surfaces are studied. For superpolished surfaces, the total amount of scattered light does not exceed some 10-6 of the incident flux. In order to have access to precise information about the roughness spectrum, numerous experimental precautions must be taken. In particular, it is difficult to measure the roughness of a transparent substrate because of scattering from the bulk material and from the back surface. An interesting solution consists in coating the surface with a thin layer of aluminum making it thus opaque. This technique can also be applied to a surface covered with a multilayer coating, so that information concerning the roughness of the surface in contact with air can be obtained. Some results on this subject are presented to better specify the possibilities and limits of this technique. We particularly deal with scattering studies on mirrors and Fabry-Perot filters which are used for optical telecommunication demultiplexing.

Light scattering and non-contact sensing of rough surfaces with a laser focus are two optical methods which recently have also become available as commercial instruments. Optical measurements are compared with mechanical stylus measurements and the results are discussed under consideration of the physics of the different measuring principles.

The use of light scattering or diffraction for the determination of the surface roughness of machined metal components has been investigated by many researchers. As a result, a number of theoretical formulations relating the r.m.s. height Rq to the scattered light intensity distribution have been derived for periodic and random surfaces. Experimental results for rougher surfaces have been reported and from them the individual researchers have derived empirical relationships between the standard deviation of the angularly scattered light distribution and the r.m.s. surface slopes. Each of these relationships are claimed to be valid for a certain class of surfaces. As a result of this, commercial optical surface finish sensors, based upon light scattering have been produced. The manufacturers of these instruments have defined some new surface finish parameters like Sn and Rop, thereby contributing to the so-called "parameter rash". Based upon scalar diffraction theory, formal relationships are derived, between a The specular reflectance and the r.m.s. height parameter Rq, and b The standard deviation of the scattered light distribution andthe r.m.s. slope parameter Δq.

I investigated the ability of two scattering theories to predict the scattering of light very near the specular beam. One scattering theory is based on a scalar approximation and the other takes into account the vector nature of light. The scalar theory is further developed using an approximation that the surface height variation is much less than the wavelength of light. Surface height profiles were determined using optical interferometric techniques. The angle dependent scattering was calculated and found to be in fair agreement with experimental values.

An optimized fiber-optic surface roughness sensor is described in which two modulated light-emitting diodes coupled to optical fibers at two different wavelengths played a role of the light source while the light intensity backscattered from the investigated surface was detected by a pin photodiode.

Deposition of thin (0.1 to 1.0 micron thick) metal films on smooth surfaces can cause a reduction in microroughness and related optical scatter. The effect is more apparent for films which have been deposited using energetic processes, such as ion bombardment and sputtering, compared to evaporated films. The amount of smoothing is also dependent upon the initial roughness of the surface and the film material. Results from examining surfaces using several characterization techniques will be discussed.

Universal acceptance of a proposed standard measurement method can depend not only upon how soundly it is based in scientific theory but on its cost and technical implementation as well. A total integrated scatterometer for the optical shop is described with emphasis on economy, rapid measurement, repeatability, and ergonomic packaging as controlling design criteria. The advent of low-cost microvolt resolution in digital multimeters allows the use of large-area silicon photovoltaic cells for detection of the scatter and specular light from the sample. The thin cell profile permits placement of the scatter detector closer to the sample port for minimal scatter obscuration. The large cell area accepts the blur circle from an inexpensive molded acrylic dome for scatter collection. A dedicated pocket computer and printer calculates, displays and prints sample RMS roughness, average, and standard de-viation for multiple measurements; it also controls laser user-access, prints a tutorial, and identifies sample, operator and date/time. The laser is a 2mW HeNe (633 nm); safety issues are addressed. The specular beam reflects off the specular detector and onto an alignment target screen, ensuring sample alignment and measurement repeatability. The inverted design provides a gravity-loading sample stage that is completely accessible; custom sample mounts are readily added. Component sources are provided. Performance and correlation to other scatterometer and roughness measurement techniques such as optical and mechanical profilers are presented.

Low scatter mirror substrates for laser-gyro applications were investigated. Bare samples, prepared by different especially developed polishing processes were examined. Nomarski differential interference contrast microscopy was used to get an overview of the surfaces. It is demonstrated that this system yields a rms roughness resolution better than 0.1 nm. To objectify the results of the Nomarski system the total integrated scattering of the bare substrates was measured. For this purpose an Ulbricht integrating sphere with a special designed diaphragm to block unwanted light scatter by the rear side of the transparent substrate was used. Investigations of different surfaces gave results for the total integrated scattering from 10.10-6 to 3850.10-6 ; the corresponding rms roughnesses are 0.19 nm and 3.61 nm, respectively. Furthermore an especially developed polishing process which produces surfaces with a roughness of about 0.1 nm and probably lower is being discussed.

A characteristic feature of a supersmooth surface is its low scatter. The scatter is proportional to the square of the rms surface roughness. Therefore, light scattering is a suitable and nondestructive method for characterization of smooth surfaces. It is possible to detect scattering created by height differences of a few atomic layers but the lateral sensitivity is limited to the order of the wavelength, ~0.5μm. The new F 1048-87 ASTM standard test method for measuring the effective surface roughness of optical components is based on total integrated scattering (TIS). The amount of scattering, caused by the surface roughness, is of primary interest for optical applications, while the roughness itself is of greater concern in the fields of microelectronics and magnetic memory storage. This paper will highlight the use of a low noise TIS instrument for characterization of sub-Å roughness on semiconductor wafers, for thin film characterization, and for detection of traces of contamination on silicon surfaces.

The structure of Si-surfaces has been investigated by elastic lightscattering which provides rapid, contactless, and nondestructive information on any kind of defects like dust, particles, bloom, microscratches and micro etch pits. Moreover we used the diffuse scattered light as a very sensitive method to determine the microroughness of polished Si wafers in the submicron (0.16 μm - 10 μm) range. Using different wavelengths this technique allows to distinguish between superficial and subsurface defects. Angle resolved measurements under wafer rotation (azimuthal scans) demonstrate anisotropic etching behavior. Alkaline solutions (potassium hydroxide) produce a strong anisotropic surface structure on (100)- and (111)-oriented Si wafers. In contrast an acidic etch gives an isotropic and smoother surface. Chemo-mechanical polishing reduces the diffuse scattered light for all scattering vectors by about six orders of magnitude. According to Total Integrated Scattering (TIS) measurements (carried out for calibration) the smoothness of a perfectly polished wafer is below 1 Å (RMS). Polishing leads to regular atomic steps at the surface. If the misorientation is less than 0.1° the existence and regularity of these steps can be proved by this lightscattering method. During subsequent processing steps, however, the smoothness of the polished surface deteriorates in most cases again. It is demonstrated that oxidation and ion implantation increase the lightscattering level substantially due to surface roughening and/or subsurface damage.

Spectral structure in the diffuse reflectance of light from oxidized copper films on glass is reported. It has been experimentally verified that the growth of the oxide in the parent metal is essential for the development of interface roughness between the oxide and copper. This causes strong scattering of the reflected radiation. The scattering exhibits sepctral intensity variations which are oppositely phased to the regular interference pattern. It has been found that the diffuse reflectance scales very closely with the averaged electric field intensity at the oxide/metal interface. A model is presented, which is used to calculate diffuse reflectance spectra. It is shown that fitting of the calculated diffuse reflectance spectra to the experimental results gives quantitative information about the roughness of both interfaces. For oxidized copper films, excellent agreement has been found between the calculations and the values obtained by surface profilometry and TIS measurements on the bare copper film obtained after etching in hydrochloric acid.

In the context of developing optics for X-ray astronomy, measurements of a variety of flat mirrors and telescopes have revealed that the total integrated scatter (TIS) formalism, well known in the optical regime, also describes the surface scattering of soft X-rays. The topography of mirror surfaces can be well studied by angle resolved X-ray scattering, from which e.g. the power spectrum and auto-covariance function of surface irregularities can be inferred. X-rays can also be used as diagnostic tool to identify surface slope changes and periodic like structures. A quantitative formalism is derived to separate the spectrum of spatial surface wavelengths into components which can be treated in a geometrical optics appraoch and those which need exact diffraction treatment. This formalism has been applied to predict the X-ray point spread function of telescopes from mechanical profile measurements. The agreement between prediction and X-ray measurement is remarkably good.

X-UV and X-ray scattering by a LiF crystal is measured. The angular distribution of the scattered radiation (ADSR) reveals characteristic features, side-peaks or asymmetry, The surface of the sample is statistically characterized by a microdensitometer analysis of electron micrographs resolving the short spatial wavelengths of the surface roughness. This analysis shows that the surface has a large microroughness with an autocovariance function which is gaussian in its initial portion. The first-order perturbation vector theory of the roughness-induced scattering leads to an interpretation of the ADSR features in terms of the modulation of the surface power spectral density function associated with the microroughness, by an optical factor. The possibility of obtaining short-scale roughness characterization from X-UV or X-ray measurements is discussed.

Increasing attention is being paid to the use of non-contact sensors for the measurement of surface texture quality before removal of the component from the surface finishing machine. The information obtained in this way is of direct value in optimising process technology and for tool life prediction. Since faults in the process of surface generation can initiate changes in roughness, waviness and flaws, all the surface texture features need to be quantified. Whilst measurement of the angular spread of a beam of light scattered by a surface provides information on the first two parameters, isolated defects such as digs and small scratches are not so readily measured. An alternative approach, based on the precise measurement of the contrast of an image of the surface under test when illuminated with partially coherent white light, has shown considerable promise in measuring all relevant parameters. Providing reference surfaces bearing a range of calibrated texture features are available for direct photometric comparison, the method has been shown to be capable of measuring a wide range of texture features. The paper will report on the latest position concerning standards for the measurement of flaws, such as scratches and digs, and also on the potential of this image comparator technique for measuring surface roughness, waviness and microtopography, in an on-machine situation.

The following describes a new procedure used for the classification of local defects on optical surfaces by use of the observer's eye contrast threshold. A new apparatus which works in transmission or reflection was constructed and tested. We will describe the principle and give the results of the experiments carried out with this work station. This study was financed by the Centre Technique des Industries Mecaniques and the Centre National de la Recherche Scientifique. Technical follow up was provided by the Groupement des industries francaises de l'optique.

Quasi-calibrated, easy-to-use visual methods for assessing the roughness (or smoothness) of optical surfaces and thin film coatings in the optical shop are investigated as to their correlation with instrumental surface roughness measurement methods. An intriguing relation between visual relative ranking of optical surfaces with the aid of a Nomarski microscope and the fractal dimension of their scattered light field is described in some detail.

LASSI (Laser Spot Scanning Interferometer) is a highly precise and versatile surface profilometer developed for various measurement applications in surface lapping, etching and polishing processes. The principle of measurement is based on a differential interferometer in which two parallel light beams split from a He-Ne laser are scanned across a sample surface. The phase difference of the reflected beams changes proportionally with the height variation between the two spots illuminated on the surface. In using a phase-locked method to determine the phase differences height variations of a surface can be measured with manometer precision. The range of applications of an instrument built on this principle encompasses step height, roughness, slope, and profile measurement of surface microtopographies. Special versions of this tool have also been developed for in situ monitoring of etch or deposition rates in sputter-etching and wet etching processes. In this paper, the principle of measurement will be described and some theoretical aspects of the measurement technique will be discussed. The various LASSI tools and their applications are reviewed.

An evanescent field arising from photons tunneling through a total internal reflection boundary is employed as a high vertical resolution (less than 1 nm, detector-limited) height transducer which maps sample microtopography into a grayscale imagespace. A video densitometer and xyz oscilloscope restore the grayscale image to a 3D topographic image; restoration and viewing perspective manipulation are real-time. Depth of field is about 1 µm and lateral resolution is typical to an oil immersion microscope (about 0.4 µm).

A reflected image can constitute the input image for a coherent optical processing system. We have set up an optical-digital hybrid computer which in its optical part uses two distinct lasers, one to process transparent images, the other to process reflected images. As examples of achievable operations two kinds of processes are illustrated: in the first one a reflected image is optically filtered with a continuously varying filter obtained with the technique of rotating filters. In the second one a pattern recognition method, the joint Fourier transform (JFT) correlation, is applied to both stationary and time varying surface textures.

The shape of surfaces can be an important factor determining their properties. Likewise, the preparation of surfaces determines their shape. Understanding such processes on the microscopic level requires a variety of sophisticated probes, each with their limitations and advantages. The scanning tunneling microscope (STM), although a newcomer in this area, has the power to investigate surface properties with very high resolution. In this paper I discuss the application of STM techniques to the roughness of silver surfaces, where the roughness of the surface is generated by a variety of preparation methods, by adsorption on the surface, and by creation of structures with the STM itself. Examples are also given of local probes that provide spectroscopic information. Limitations and problems associated with the investigation of rough surfaces are covered, as well as an outlook to the applicability of STM to other areas such as adhesion, friction and engineering given. Finally I shall discuss some of the first results on the emission of photons from the tip-surface region of the STM. In the tunnel mode a large photon intensity is found in the optical range for silver films. These studies demonstrate an interesting link between STM and optical science.

Ultrasonic reflectance/scattering measurements have been made on metal samples possessing a large range of surface roughness values. The root-mean-square roughnesses Rq ranged from 0.3 to nearly 40 µm on the mostly periodic surfaces. The echo amplitude from short incident pulses of ultrasound in the frequency range of 1 to 30 MHz was used, in the manner of a comparator, to measure relative roughnesses with an area-averaging approach defined by the ultrasonic beam spot size. Ultrasonic wavelengths ranged from about 50 to 300 µm at these frequencies, and the beam spot sized varied from 0.2 to 5 mm in diameter. Both air and fluid coupling techniques were used between the sensor (transducer) and surface, on both static and rapidly (in excess of 5 m/sec surface speed) moving parts. On static surfaces, a resolution of better than 1.0 µm Rq was achieved at the higher ultrasonic frequencies. By focusing the ultrasonic beam at 30 MHz, a profilometry capability was demonstrated on a 1 µm Rq sinusoidal specimen of 800 µm wavelength.

Absorber coatings for solar-thermal energy conversion must have a low reflectance in the wavelength range of the solar irradiation and a high reflectance in the range of thermal radiation. There are several physical principles providing such a spectral behaviour. One of them is a surface microroughness with an rms-roughness smaller than 0.5 µm and a correlation length of the same order of magnitude or smaller. Only few measurement techniques can be used for the characterization of such a microstructure. This paper will discuss the applicability of non-stereoscopic scanning electron microscopy (SEM) with digital image processing for the evaluation of the correlation length of the microstructure. Emphasis will lie on the investigation of pyramidal structures, for which angular dependent SEM can yield additional information about the mean slope of the structure. The results are compared with the outcome of the evaluation of spectrooptical measurements in the wavelength range from 0,36 - 15 µm by means of statistical scattering theories. These measurements are carried out with a Zeiss PMQ3-spectrophotometer-system with a BaSO4-coated integrated sphere (0.36-2.5 µm), and with a Bruker FTIR spectrometer equipped with a diffuse-gold coated integrating sphere (2,0 - 15 µm), respectively.

It is proposed that surface enhanced Raman scattering (SERS) of a monolayer of a dye deposited on a rough metal surface can be used to characterize the film roughness. The advantage of these techniques is the use of a very sensitive methods (Raman scattering signal) to measure microscopic roughness in metal substrate. The Raman enhancement is due to the amplification by the Ag core of the electric field associated with the driven modes of the dye layer. In the present case the due is crystal violet (CV) and the measured enhancements are calculated to be of the order of 104 to 106. Since, the enhancement of the Raman spectra depends on the structure of the metal island film and specially the roughness, this effect can be applied to investigate the surface roughness of the substrate and characterize its structure.