We briefly review some of the history and technology of scanning microscopy. The common threads that connect these different scanning microscopies are higlighted. In the case of far-field scanning microscopy, the resolving power is governed by the Abbe criterion. In the case of near-field scanning microscopy, the Abbe limit can clearly be surpassed. The invention of the Scanning Tunneling Microscope has provided some of the confidence necessary to broaden and enhance the capabilities of near-field scanning microscopies. The paper concentrates on the history and technology of selected scanning near-field and far-field microscopies.
The scanning tunneling microscope(STM) has the capability of mapping the topography of a surface with unprecedented resolution in both vertical and lateral directions. STM has been used to obtain real space images of individual atoms on a surface. In this paper we review the principles and technical aspects of STM, and give some examples of its applications both in air and vacuum. These include images of graphite, silicon, and indium on silicon. Several techniques of digital image processing developed to facilitate the analysis of STM data are also described.
Scanning Tunneling Microscopy (STM) can be used to perform lithography on a surface and to image the results of lithography performed using other techniques. We have imaged a 1 nm structure created with the STM on graphite, a 300 nm structure electrodeposited with the STM on gold, and a gold diffraction grating created using a diamond scribe. In addition we present atomic resolution images of graphite showing a surface reconstruction.
The Atomic Force Microscope can resolve features on conducting or nonconducting surfaces down to the atomic level. The heights of features are recorded as a sharp tip scans over the surface in parallel scans. The interaction between the tip and the surface is the interaction potential between atoms. Individual carbon atoms separated by 0.146 nm have been resolved on graphite. Ordered structure on the "native" oxide of silicon has been observed. Rows of molecules that are separated by 0.5 nm have been resolved in an organic monolayer. The key to the operation of an AFM is the development of a system for sensing tracking forces that are small enough to avoid damaging the surface. The images in this report were obtained by sensing with electron tunneling the deflection 1 - 10 nm) of springs (k 0.1 - 100 N/m) fabricated from silicon oxide or fine wires.
Our version of the atomic force microscope (AFM), and variants which measure electric and magnetic force are described from an engineering viewpoint. Discussion centers on design and performance issues, with emphasis on application to magnetic recording. The basic force probe consists of an L-shaped wire cantilever, whose end is etched to a radius of less than 1000 A, mounted on a piezoelectric bimorph When the tip is held close to a sample, forces between tip and sample modify the dynamical properties of the resonant cantilever; the force gradient changes the effective spring constant, and thus the resonant frequency. Vibration of the tip is measured by heterodyne interferometry, a technique about as sensitive as tunnelling but easier to set up and insensitive to slow drifts. By driving the tip near resonance, the change in the resonance frequency is translated into an amplitude shift; force gradients of 10-4N/m and force increments of 10-13N have been measured in this way. This atomic force probe has achieved lateral resolutions of 50 A, and there seems to be no fundamental impediment to achieving atomic resolution. The magnetic force microscope (MFM) is very similar. It uses a steel wire tip to measure magnetic field or field gradient with a spatial resolution of better than 1000 Å. Data are shown, which demonstrate the usefulness of magnetic imaging for profiling the field patterns in recording heads and disk media.
In the last few years there has been a resurgence in research on optical microscopes. One reason stems from the invention of the acoustic microscope by Quate and Lemons,1 and the realization that some of the same principles could be applied to the optical microscope. The acoustic microscope has better transverse definition for the same wavelength than the standard optical microscope and at the same time has far better range definition. Consequently, Kompfner, who was involved with the work on the early acoustic microscope, decided to try out similar scanning microscope principles with optics, and started a group with Wilson and Sheppard to carry out such research at Oxford.2 Sometime earlier, Petran et a13 had invented the tandem scanning microscope which used many of the same principles. Now, in our laboratory at Stanford, these ideas on the tandem scanning microscope and the scanning optical microscope are converging. Another aspect of this work, which stems from the earlier experience with the acoustic microscope, involves measurement of both phase and amplitude of the optical beam. It is also possible to use scanned optical microscopy for other purposes. For instance, an optical beam can be used to excite electrons and holes in semiconductors, and the generated current can be measured. By scanning the optical beam over the semiconductor, an image can be obtained of the regions where there is strong or weak electron hole generation. This type of microscope is called OBIC (Optical Beam Induced Current). A second application involves fluorescent imaging of biological materials. Here we have the excellent range definition of a scanning optical microscope which eliminates unwanted glare from regions of the material where the beam is unfocused.3 A third application is focused on the heating effect of the light beam. With such a system, images can be obtained which are associated with changes in the thermal properties of a material, changes in recombination rates in semiconductors, and differences in material properties associated with either acoustic or thermal effects.4,5 Thus, the range of scanning optical microscopy applications is very large. In the main, the most important applications have been to semiconductors and to biology.
The limitations of the scanning optical microscope in producing high contrast images of nearly uniform objects are discussed. It is shown that by resorting to linear and differential imaging concepts, a number of techniques can be devised 'which overcome such inherent limitations. The design and operation of a scanning differential amplitude, and linear differential interference contrast microscopes are discussed, and a number of experimental results are presented. These microscopes are capable of stable operation and sensitivity down to 10-6 in detecting a refractive index change, and 1.5>10-4nm height variation, in a 1 kHz bandwidth.
Over the last few years scanning optical microscopy (SOM) has been largely developed as a tool to explore the physical properties of materials. In particular the optical beam induced current (OBIC) mode of the SOM has been used to map the electronic properties of semiconducting devices. A new type of scanning microscopy method, as well as some results obtained by it, will be reported in this paper. Though similar, to a certain extent, to the standard scanning optical microscopy, this new investigation technique, from now on refered to as infrared beam induced contrast (IRBIC), differs from it in substance. The chopped light from a quartz halogen lamp is focused by a conventional microscope rearranged on the specimen surface, and a pin-hole is positioned so as to reduce the probe size (not the resolving power) to 1.5um. The resulting beam power density is of the order of 1mW*cm-2. Such a low power density presents some disadvantages in comparison with the traditional laser sources, but, on the other hand, it allows a very high sensitivity in the investigation of the defect electrical activity. With this experimental set-up the specimen front surface is probed with band-gap radiation. Its back surface is illuminated by continuous light in the infrared, coming through a monochromator from a glow-bar. The radiation wavelength can be selected continuously so as the photon energy ranges over the whole valence-to-conduction energy gap. When the specimen is probed, the photoinduced carriers are separated by the built-in field due to the depletion zone of a p-n junction or a Schottky barrier, and the photocurrent is amplified by the lock-in technique. The application of a back-surface radiation of less than the band-gap energy modifies, in some way, the photoconductive response to the band-gap probe since the secondary illumination changes the occupancy of the traps in the forbidden gap active in the photoconductive process. This phenomenon, known as "quenching" of extrinsic photoconductivity, when applied to scanning optical microscopy allows localized investigations on energy levels. As a matter of fact, the beam induced contrast at a defect depends on the secondary illumination wavelength, that is on the trap level excited, as we observed in defective Si.
The optical profile measuring method described earlier was analyzed with the assumption of infinite lens apertures. A more detailed analysis shows that the presence of finite apertures induces a focal shift in addition to the well known effects on spot size and shape. The calculations lead to a modified design for the profilometer with better performance.
This paper describes a semiconductor metrology system based upon ultra-violet wavelength confocal microscopy. The system is capable of linear metrology of resist features down to 0.5 microns linewidth with low dependence on substrate type. Short term precision of better than 5nm standard deviation can be obtained with this system. The optical design for 325nm operation is described together with details of the data acquisition system. Experimental data compares the performance of ultra-violet and visible light versions of the system for resist metrology, showing the benefit of using a wavelength at which the resist is absorbent. Conclusions are drawn about optimal regimes for metrology as well as the extension of this technology to yet shorter ultra-violet wavelengths.
The elastic scattering efficiency from subwavelength-size surface holes or protrusions depends sensitively on the dielectric properties of their immediate environment. Scanning near-field optical microscopy exploits this effect to create optical images whose resolution is not restricted by the diffraction limit. Typically, 20 nm laterally and, when operated in a topographic mode, 0.1 nm in height can be resolved. Images were obtained both in transmission and reflection.
Superresolution optical images have been generated with near-field scanning optical microscopy (NSOM). The underlying concept is presented and several modes of operation are discussed. A resolution based on edge sharpness of 70 nm or better has been demonstrated with two different instruments. Images have been obtained which characterize the resolution as a function of two critical parameters: the aperture size and the aperture-sample separation. The near-field images also illustrate novel features resulting from several forms of contrast. Finally, the potential of NSOM is compared with conventional lens-based forms of microscopy as well as with more recent scanned tip methods.
The use of surface plasmon resonance measurements for the microscopical imaging of surfaces and thin films has been investigated. Theoretical results are presented for sensitivity optimization and focussed beam measurement. Experimental results are presented for scanned beam and plane-wave imaging of silver surfaces and superimposed dielectric layers. Thickness sensitivity of about 3 angstroms has been achieved, with lateral resolution better than 20 microns.
Scanning capacitance microscopy is a mechanically scanned microscopy which uses variations in the capacitance between the sample and a scanning tip as a probe of sample topography. In this paper we discuss the theory of the capacitance probe and its application to scanning capacitance microscopy and fringe field profilometry .
Acoustic microscopy has been growing steadily since the development of the first microscope in the mid-seventies. Since that date, acoustic microscopes have been used as both qualitative tools for imaging, and quantitative tools for the characterization of materials. The frequency range, and thus resolution, of acoustic microscopes covers several decades which makes the methodology useful for examining materials with spatial resolutions ranging from millimeters to Angstroms. In this paper, we will review the principle of operation of the microscope, show some imaging results, and present some quantitative measurements of samples using both amplitude only and amplitude and phase measuring microscopes. We will also show some recent results of quantitative measurements of anisotropy and surface residual stress that are possible with new types of lenses.
A review is made of progress in high resolution near field thermal imaging. The Scanning Thermal Profiler, initially demonstrated as a surface profiler, has profiled aluminum films with sub 100 nanometer resolution. More recently, the thermal sensor has been used to map current induced heating of a surface with comparable resolution. Finally, photothermal imaging of laser heated surfaces has been demonstrated, with image structure well below 100 nanometers.
Thermal wave imaging methods of investigating both surface and near subsurface regions of a specimen are considered. Several mechanisms for subsurface probing are distinguished and issues related to spatial resolution for the various mechanisms are discussed.
This paper discusses the reasons for applying fluid film bearings in high precision applications. The information presented shows how these bearings can be designed to yield maximum and constant film stiffness in the mid-position or alternatively increasing stiffness as the bearing departs from the mid position. Design features to minimise transmitted noise in bearings are described to enable the designer to produce a mechanically quiet system. The data presented is related to both gas and liquid fed bearings. The design data presented is limited to journal bearings due to limitations of time to compute the charac-teristics of flat pad bearings. The general design criteria however applies to both flat pad and journal bearings.
Early uses of scanning tunneling microscopes indicate that many areas of science could greatly benefit if centimeter scale objects could be probed on an atomic scale. This could be accomplished by the development of a linear slide having one Angstrom resolution and a one-tenth meter range of travel. Magnetic suspension technology has been proposed as an ideal candidate for the suspension of thee linear slide. This technology is attractive since it combines low noise, excellent positioning accuracy, zero wear, and requires no lubrication. The paper examines the feasibility of developing a magnetic suspension with a required resolution exceeding one Angstrom and establishes the requirements for this suspension. The achievable resolution is shown to be determined by position measurement accuracy, suspension gain, suspension bandwidth, and disturbance force levels. Expected disturbance force levels from ground motion, air currents, and acoustic effects are projected. Given the expected disturbance force levels, measurement accuracy, and required controller performance, a magnetic suspension could be developed which would provide resolution better than the one Angstrom required.
The technology for "seeing" with sound has an important and interesting history. Some of nature's creatures have been using sound waves for many millenia to image otherwise unobservable objects. The human species, lacking this natural ability, have overcome this deficiency by developing several different ultrasonic imaging techniques. acoustic microscopy is one such technique, which produces high resolution images of detailed structure of small objects in a non-destructive fashion. Two types of acoustic microscopes have evolved for industrial exploitation. They are the scanning laser acoustic microscope (SLAM) and the scanning acoustic microscope (SAM). In this paper, we review the principles of SLAM and describe how we use elements of SLAM to realize the scanning tomographic acoustic microscope (STAM). We describe the data acquisition process and the image reconstruction procedure. We also describe techniques to obtain projection data from different angles of wave incidence enabling us to reconstruct different planes of a complex specimen tomo-graphically. Our experimental results show that STAM is capable of producing high-quality high-resolution subsurface images.
PIPE is a pipelined image processing device that was designed for real-time robot vision applications. It accepts images from a video camera 60 times per second, and contains hardware for digitizing, displaying, and performing operations on these images at video rates. Each stage of the pipeline contains arithmetic and logic units, convolvers, image buffers, and look-up tables. The purpose of this paper is to introduce PIPE and some of its applications to the scanning microscopy community. Three kinds of applications are described. The first is stereo analysis, whose purpose is to automatically extract range from two cameras mounted side by side. The second application is motion analysis and tracking. This application involves detecting motion, measuring its velocity, using it to obtain three-dimensional information, and tracking it through time. The final application is inspection of two-dimensional patterns.
The theory and numerical considerations that are used in the computation of the scattered electromagnetic fields near the surface of a silver strip on a glass substrate are presented. These calculations provide theoretical guidance for the measurement of the width of the strip by means of near-field optical scanning. The dimensions of the strip cross section, e. g. 300 nm by 100 nm, can be a fraction of the wavelength of the incident light, 632.8 nm. The flux 1 nm above the surface shows sharp spikes at the edges of the strip. The features of the fields near such a surface could be used for accurate width measurements up to about 30 nm above the strip. The effects of other variables are also shown in the figures.
Soft x-ray microscopy allows high resolution imaging of wet biological samples. 100A resolution has been demonstrated with contact microscopy while to date scanning microscopes have achieved 750A resolution. The present state of soft x-ray microscopy is reviewed with emphasis on scanning systems. We compare the various combinations of soft x-ray sources and focusing optics for scanning microscopes and summarize the present performance of the three operating synchrotron / zone plate scanning microscopes. Finally, we describe the SLanford tabletop scanning soft x-ray microscope which uses a laser-produced plasma source with normal incidence mirror focusing optics.
We demonstrate the usefulness and high sensitivity of the atomic force microscope (AFM) for imaging surface dielectric properties and for potentiometry through the detection of electrostatic forces. The attractive force with an applied voltage between tip and sample is generally much larger than the van der Waals force. On the other hand, electric forces as small as 10-10 N have been measured, corresponding to a capacitance of 10-19 farad. The sensitivity of our AFM should ultimately allow us to detect capacitances as low as 8 x 10-22 farad. We have used this technique to detect the presence of dielectric material over Si, and have made measure-ments of the voltage over a p-n junction with sub-micron spatial resolution.
A scanning microscopy workstation has been developed to digitally control a scanning laser microscope and to provide a powerful image processing system to display and process scanning microscope images. The modular hardware interface has been designed so that the workstation can be easily adapted to control and accept data from other kinds of scanning microscopes and instruments that measure spatially-resolved data. The workstation accepts data files in a wide variety of formats, from single line scans to large rectangular images. All image processing algorithms are designed to accept these formats and treat images as arrays of data points, not just pictures.
A new three-dimensional, non-contact laser interference microscope* is described which uses computerized phase measurement 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 0.271μm. Test surface reflectivity can range from less than 1% up to 100%. Turret mounted, parfocal objectives permit rapid magnification change. Laser illumination yields interference fringes throughout the entire depth of focus of each objective magnification permitting rapid fringe acquisition. Tip and tilt of the entire instrument head about the plane of the test surface eliminates feature walkoff from the field of view at high magnification. 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.